EXPERIMENTALNEUROLOGY
116,76-84
(1692)
Detailed Distribution of Nerve Growth Factor in Rat Brain Determined by a Highly Sensitive Enzyme Immunoassay TAKESHI NISHIO,*,~ *Department
ICHIRO AKIGUCHI,*
AND SHOEI FURUKAWA~
of Neurology, Faculty of Medicine, Kyoto University, Kawahara-Cho 54, Shogoin, Sakyo, Kyoto, Japan; and tDivision Neuroimmurwlogy, National Institute of Neuroscience, NCNP, Ogawa-Higashi 4-1-1, Kodaira, Tokyo, Japan
of
Is NGF-responsiveness limited to only these neurons in the central nervous system (CNS)? In the rat brain stem, a moderate level of NGF mRNA (11, 28, 32) and high-affinity NGFRs (23, 25) have been detected. Furthermore, a low level of NGF protein is unambiguously detectable in the cerebellum (13, 14, 31), although no cholinergic neurons are contained in this region. These facts suggest that NGF acts as a neurotrophic molecule on some sorts of central neurons in addition to the MCNs or striatal neurons. Histological visualization of NGF and NGF mRNA in the CNS may contribute greatly to the elucidation of the role of NGF in the CNS. However, the immunohistochemical data on NGF distribution (2, 8, 32) have shown no obvious correlations to the quantitative data determined by a reliable enzyme immunoassay (EIA) (13,14,31); and ins&u hybridization has been successful for visualization of NGF mRNA only in regions such as the hippocampus and/or cerebral cortex where the NGF mRNA level is relatively high (3,10,24,29,34). This lack of detection is probably because the levels of NGF and NGF mRNA are extremely low in most of the CNS (30). Therefore, as a challenge to detect low levels of NGF, we developed a new EIA system that enabled us to measure the NGF content of as small as 2 mg (wet weight) of brain tissue; and by using this method we determined the detailed distribution of NGF in the adult rat brain.
We modified a previously reported enzyme immunoassay method to make it more sensitive for quantification of nerve growth factor (NGF), and succeeded in measuring the NGF content in as small as 2 mg (wet weight) of rat brain tissue. Rat brain was cut into about 600 pieces of the same size, and the NGF content in each piece was determined by this method. The findings were as follows: (i) In the cerebral cortex, NGF contents were unevenly distributed, ranging from less than 0.1 to 1.8 rig/g wet wt. The level was highest in the caudal parietal and rostra1 occipital cortices and lowest in the lateral parietal cortex. (ii) Areas comprising the limbic system such as the cingulate gyrus, pyriform cortex, amygdala, anterior and medial thalamus, hippocampus, septum, and diagonal band of Broca contained high levels of NGF. (iii) In the brain stem and cerebellum, the levels were low; however, a relatively high level was registered in the cerebellar nuclei, lateral vestibular nucleus, ventral cochlear nucleus, superior olive, and pontine reticular nuclei. These findings, taken together with previously published information, suggest that the neurons in the anterior and medial thalamus, pontine reticular nuclei, superior olive, ventral cochlear nucleus, and cerebellar Purkinje cells may be additional populations of NGF-responsive neurons in the rat brain. 0 1992 Academic Press, Inc.
INTRODUCTION MATERIALS It has been so far established that nerve growth factor (NGF) acts as a neurotrophic factor for the magnocellular cholinergic neurons (MCNs) in the basal forebrain (30, 33). The evidence for the retrograde axonal transport of NGF from the neocortex to the basal forebrain (26, 27), and the presence of high levels of NGF and NGF mRNA in the projection area of the MCNs (11,13, 14, 28, 31, 32) and of high-affinity NGF receptors (NGFR’s) in the cell bodies of the MCNs (23, 25) suggest a role for NGF as a trophic factor on MCN in the basal forebrain. Although the data are not so numerous as in the case of MCN, striatal intrinsic choline@ neurons should also be included in this category (19,23,25). 0014-4886/92 $3.00 Copyright 0 1992 by Academic Press, All rights of reproduction in any form
Materials. Wistar rats were purchased from Japan SLC, Inc. Bovine serum albumin (BSA) was obtained from Sanko Junyaku; gelatin, from Bio-Rad; EDTA, from Dojin; and aprotinin, from Sigma. Other chemicals were reagent grade. NGF was purified from mouse submaxillary glands as previously described (9). Preparation of rut brain samples. The brains of 7week-old Wistar rats (n = 3), which were taken from animals killed by deep narcosis induced by diethyl ether, were coronally sectioned into 0.9-mm-thick slices with a cryostat at -10°C. Each coronal slice was perpendicularly cut into squares (1.5 X 1.5 mm) with a stainless 76
Inc. reserved.
AND METHODS
DETAILED
DISTRIBUTION
14nm -I
I-
FIG. 1. Schematic representation of the preparation of samples. Wistar rat brain was coronaily sectioned into 0.9-mm-thick slices with a cryostat at -10°C. Each coronal slice was perpendicularly cut into squares (1.5 X 1.5 mm) with a razor blade, resulting in small pieces having maximal dimensions of 0.9 X 1.5 X 1.5 mm (2.0 mm3).
razor blade (Fig. 1). The brain samples of 0.9 X 1.5 X 1.5 mm (2.0 mm3) were prepared. The sectioning of the coronal slices was adjusted by cutting the brain initially at the plane including hregma and sectioning into 1.5 X 1.5-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 (21) and the dissections of brain areas were reliably reproducible. ~eter~i~utio~ of o~ti~~~ co~~it~~s for NGF extruction. Such small and numerous brain sections cannot be homogenized easily and reliably and the background signals measured in our EIA system were lower with the extraction of NGF by freeze-thaw procedure than with the extraction by homogenization. Therefore, we tried to extract NGF with buffer A (0.1 M Tris-HCI buffer ]pH 7.61, containing 2% [w/v] BSA, 2% [w/v] gelatin, 1.0 M NaCl, 2 m&f EDTA, and 80 trypsin-inhibitory units of aprotinin/liter) from each piece of the rat brain by a freeze-thaw procedure. Each piece was soaked in 100 ~1 of buffer A in a 96-multiwell plate (Falcon), frozen at -8O”C, and thawed completely at room temperature. The effect of frequency of this freezing-thawing
OF NGF IN RAT BRAIN
77
procedure on NGF extraction was monitored by measurement of NGF released into the solution, and the degree of recovery of NGF by this procedure was compared with that obtained by homogenization. Several pieces from a defined area in the rat brain (the caudal parietal cortex) were homogenized by 30 strokes with a glass pestle (Wheaton) in buffer A with or without exogenous NGF. The homogenates were centrifuged at 100,OOOg for 10 min and NGF content was determined by the EIA described below. The recovery of NGF in the extract was calculated as the ratio of (NGF content with exogenous NGF) - (NGF content without exogenous NGF) /exogenously added NGF. HA for NGF. NGF contents were essentially measured by the highly sensitive two-site EIA previously described (9). However, it was drastically modified to minimize sample volume as follows: 150 ng anti-mouse NGF rabbit IgG in 5 ~1 of 0.05 M Tris-HCl buffer (pH 8.5) was placed on the center of a U-bottom well of 96multiwell polystyrene plate (Falcon). After incubation for 30 min at room temperature, the IgG solution was removed, and the wells were washed twice with 0.15 ml of 0.1 M Tris-HCl buffer (pH 7.6) containing 0.4 M NaCi, 0.1% BSA, 0.02% NaN,, and 1 mM MgCl, (buffer B). Each of the IgG-coated wells was incubated with gentle shaking for 4 h at room temperature with 10 ~1 of sample solution prepared in buffer A, after blockade of nonoccupied space with 0.15 ml of buffer B containing 0.05% (v/v) bovine milk by a l-h incubation. Then, each well was washed twice with 0.15 ml of buffer B. Twenty microliters of buffer B cont~ning biotinylated antibodies (10 nglml), which were affinity-purified polyclonal rabbit antibodies against murine NGF, was applied to each well, and the plate was incubated for 12 h at 4°C with gentle shaking. After two washings, 20 ~1 of /3-D-galactosidase-conjugated streptavidin (Biogenex Lab) dissolved in buffer B was applied to each well, and the plate was incubated for 1 h at room temperature with gentle shaking. The enzyme reaction was started at room temperature by the addition of 30 ~1 of 6 pM 4methylumbelliferyl-P-D-galactoside in buffer B and stopped by that of 0.1 ml of 0.1 M glycine-NaOH buffer (pH 10.3). The amounts of 4-methylumbelliferone formed were measured by fluorometry with a spectrofluorometer (model 650-60, Hitachi). The wavelengths 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 IgG-coated wells or noncoated wells. NGF contents in the samples were evaluated by reference to the standard curve. The brain tissue pieces cut into smaller than a 1.5 X 1.5-mm square, i.e., those at the marginal areas, were sketched immediately and their wet weight was calculated. NGF content evaluated
78
NISHIO,
.i
. 1. . lb .1&l ioilo Content of p-NGF
loo00
AKIGUCHI,
rtioo
f$/assay
FIG. 2. Standard curve of the enzyme immunoassay for mouse &NGF. Experimental procedures are described in the text. The detection limit in this assay was as low as 10 fg of NGF per assay well (asterisk).
by EIA was expressed as amount weight,
per gram tissue wet
RESULTS
Improved methods for detecting NGF. The assay system after modification could detect as low as 10 fg of NGF per assay well (Fig. 2). The relationship between the efficiency of NGF extraction and the frequency of freezing-thawing is shown in Fig. 3. The ordinate values in Fig. 3 are expressed as percentages of the values obtained by homogenization with a glass pestle. The release of NGF into the solution was first observed after the third freeze-thaw cycle and increased abruptly after five or six cycles. A plateau value was reached after eight cycles. As recovery of NGF by homogenization was over 90%, we considered eight cycles of freezing-tha~ng as sufficient for a practical and quantitative extraction of NGF from the brain pieces. So, we routinely extracted the factor from the brain pieces by this number of freeze-thaw cycles. There were no recognizable regional differences in the efficiency of NGF extraction by the freezing-thawing procedure (data not shown). NGF content ~di~d EIA.
in small bruin sections determined
AND
FURUKAWA
3) of NGF content per each brain piece. In the cerebral cortex, NGF content was distributed unevenly with great regional differences in NGF level (about a maximal 20-fold difference), ranging from the level of less than 100 pg/g-wet tissue [lateral frontopolar cortex (B3), lateral striatum (E-15, F-14, G-12, 18, 23), lateral parietal cortex (G-15, J-8, and -9)] to about 1800 pg/gwet tissue [caudal parietal cortex (K-4)] pg/g-wet tissue. The pieces K-l, -2, -3, -4, L-2, M-l, -2, and -3, comprising the dorsomedial portions of the caudal parietal cortex and rostra1 occipital cortex, exhibited high NGF levels (800 to 1800 pg/g-tissue). The pieces B-12, C-17, E-20, -23, F-24, H-28, I-30, -31, J-31, K-29, -30, and -31, representing the pyriform cortex, showed moderately high levels (500 to 1000 pg/g). The pieces F-7, -11, G-9, -15, -20, H-9, -14, -19, I-8, -13, -18, J-9, -14, and -19, taken from the lateral parietal cortex, showed low NGF levels (less than 200 pg/g). The pieces E-l, -5, G-l, and H-l, reflecting the cingulate gyrus, showed moderately high levels (500 to 700 pg/g). The pieces J-10, -11, K-11, L-7, -12, -17, -18, -22, -23, M-12, -13, and -22 from the hippocampus exhibited high NGF levels (900 to 1400 pg/g-wet tissue). In coronal slice J, the pieces including the dentate gyrus and areas CA3-CA4 (J-10 and -11) exhibited higher NGF levels than the pieces, J-5 and -6, comprising the areas CAl-CAB. The pieces E-9, -17, F-8, and -12, indicating the septal nucleus and diagonal band of Broca, showed moderate levels of NGF (600 to 1200 pg/g). However, NGF levels in the pieces G-22 and H-21, reflecting the nucleus basalis of Meynert and globus pallidus were low (100 pg/g-wet tissue). The levels
3 802 c 5 60 8
2
by the
The ~stribution of NGF contents in the pieces from the whole rat brain is shown in Fig. 4, and its superficial distribution in the cerebral cortex unfolded into a single plane is shown in Fig. 5. The value of NGF content in each piece obtained from each animal deviated by 10 to 30% from the mean value. Each circle at the center of a square represented the mean value (n =
o~@loa 0
2
4
6
810
Frequency of freezing and thawfng FIG. 3. Relationship between efficiency of NGF extraction and frequency of freezing and thawing. The ordinate values are expressed as percentages of the values obtained by homogenization.
DETAILED
DISTRIBUTION
in the striatum (pieces D-9, -13, E-11, -14, F-13, -14, -17, -18, G-12, and -18) were also generally low (less than 200 pg/g), although pieces E-10 and F-9 contained moderate levels of NGF (450 and 900 pg/g-wet tissue, respectively). Pieces from the anterior and medial portions of thalamus, I-14, -15, -19, and J-15, showed relatively high NGF levels (300 to 400 pg/g) compared with the pieces I-20, J-16, -17, -20, -21, K-15, -16, -17, and -21 from the other portions of the thalamus (less than 100 pg/g). Pieces I-24, J-25, and K-25 of the hypothalamus exhibited low NGF levels (less than 300 pg/g), and those comprising the amygdaloid body, I-26, K-27, and -28, contained moderate levels (600 to 700 pg/g). The brain stem and cerebellum generally showed low NGF levels (less than 100 pg/g). However, the pieces N-18 and P-13, containing the pontine reticular nucleus; P-17, including the superior olive; and Q-11, -12, and -15, comprising the cerebellar nuclei, lateral vestibular nucleus, ventral cochlear nucleus, and paraflocculus, showed relatively high (150 to 250 pg/g) NGF levels compared with other pieces in the brain stem and cerebellum.
DISCUSSION
Knowledge about the distribution of NGF in the CNS is very important for us to elucidate the role of this growth factor in the CNS, although there have been some studies on the immunohistochemical visualization of NGF in embryonic murine, fetal, and adult rat brain (2,8,32). Because of the extremely low level in the CNS, the most reliable method so far to detect NGF in the CNS has been the EIA. We modified our earlier EIA (9) to measure the extremely low level of NGF contained in a very small amount of brain tissue. The newly modified assay system showed a detection limit of as low as 10 fg of NGF per assay well (Fig. 2). The points responsible for this marked improvement in sensitivity were as follows: (i) specific (affinity-purified) anti-NGF antibodies were used for second antibodies to sandwich the antigen; (ii) second antibodies were biotinylated and used in the streptavidin-biotin system; (iii) the system was minimized in terms of sample volume by the use of Ubottom 96-multiwell plates; (iv) extraction was successfully achieved by freezing-thawing, which enabled us to
OF
NGF
IN
RAT
BRAIN
79
extract NGF quantitatively from a brain piece as small as one of 2-mg wet wt (Fig. 3). The specificity of the anti-NGF antibodies in the present study was tested previously by immunodiffision analysis, immunoadsorption method, and polyacrylamide gel electrophoresis (9) and these antibodies specifically recognized NGF. Recently, other members of neurotrophic factors with about 50% homology of amino acid sequence, brain-derived neurotrophic factor (BDNF), neurotrophin-3 (NT-3), which are called NGF family, have been found (12,16). Although we cannot exclude the possibility that the antibodies in use may cross-react to these members of NGF family and our EIA system may count falsepositive signals, the possibility seems very small because of the following points. (i) NGF have five hydrophilic regions expected to be potential antigenic determinants (20), and small parts of them are homologous in amino acid sequence to BDNF or NT-3 (12). Even though NGF shares some antigenic similarities with those molecules, the signals detected false-positively in this assay system would be small or negligible, because this assay system requires at least two epitopes for detection, i.e., two-site EIA. (ii) The levels of mRNA of BDNF and NT-3 in adult rat cerebellum are high whereas NGF mRNA level in cerebellum is very low (18). If our assay system count false-positive signals of BDNF or NT-3, the cerebellar signals detected in the present study would be expectedly high. But, cerebellar NGF levels were much lower in our study than hippocampus or cerebral cortex levels. Therefore, the antiNGF antibodies in use scarcely cross-react to BDNF or NT-3 and the assay system in the present study specifically detects NGF. NGF and NGF mRNA levels in the major brain regions, e.g., the cerebral cortex, hippocampus, or septal nucleus, have been measured by other investigators (11, 13-15,28,31,32). The cerebral cortex and hippocampus contain high levels of both NGF and NGF mRNA, whereas the septal nucleus contains a moderate amount of NGF but a low NGF mRNA level (13,14). Based on evidence such as the retrograde axonal transport of radiolabeled NGF (26,27) and the distribution of bigh-affinity NGFRs (23,25), NGF has been considered to act as a tropbic factor in the CNS only on MCNs in the basal forebrain and on striatal intrinsic choline& neu-
FIG. 4. Distribution map of NGF contents in the whole rat brain. The circle at the center of each square represents the mean value of NGF content in the piece determined by EIA. Note the blue circles represent a lower range of values than the red circles. Abbreviations used in the figure are as follows: a, amygdala; bn, basal nucleus; c, caudate nucleus; Ci, cingulate gyrus; cn, ventral cochlear nucleus; cp, caudate-putamen; DG, dentate gyrus; Di, diagonal band of Broca; En, entorhinal cortex; FC, frontal cortex; fn, facial nucleus; FPC, frontopolar cortex; gp, globus pallidus; Hl, CA1 area; H2, CA2 area; H3, CA3 area; H4, CA4 area; ht, hypothalamus; i, interpose nucleus; ic, internal capsule; ice, inferior colliculus; 1, lateral cerebellar nucleus; lg, lateral geniculate body; Iv, lateral vestibular nucleus; m, medial cerebellar nucleus; mg, medial geniculate body; mr, midbrain reticular formation; Oc, occipital cortex; p, pyramidal tract; PC, parietal cortex; pfl, paraflocculus; pg, periaqueductal gray; pn, pontine nucleus; pr, pontine reticular formation; pt, pontine tegmental nucleus; Py, pyriform cortex; r, red nucleus; re, gigantocellular reticular nucleus; RS, retrosplenial cortex; S, septal nucleus; SC, superior colliculus; sn, substantia nigra; so, superior olive; stn, spinal tract nucleus; t, trigeminal nerve; Te, temporal cortex; th, thalamus; tm, trigeminal motor nucleus.
80
NISHIO,
AKIGUCHI,
FIG,
AND FURUKAWA
I--Continued
DETAILED
DISTRIBUTION
OF NGF IN RAT BRAIN
b
FIG. 4-Continued
81
82
NISHIO,
AKIGUCHI,
AND FURUKAWA
Parietal
Occipital
100
Is
200
300
400
500
000
700
000
900
1000
1200
1000 i&g-time
FIG. 5. Distribution map of NGF contents in the surface of the cerebral cortex, unfolded into a single plane. The circle at the center of each square represents NGF content.
Ions. The present study generally supports this theory, because both areas where NGF should be synthesized, i.e., the cerebral cortex and hippocampus, and the area to which NGF should be transported, i.e., the nuclei of the basal forebrain, contained high NGF contents. In addition, this study also confirmed the fact that the NGF levels in the diencephalon, brain stem, andcerebellum are low, as previously reported (13,14,31). However, this investigation is noteworthy because the distribution of NGF contents in the brain was determined in detail and the following new facts have emerged: (i) In the cerebral cortex, NGF was unevenly distributed ranging from less than 0.1 to 1.8 rig/g wet tissue. The level was highest in the caudal parietal and rostra1 occipital cortices and lowest in the lateral parietal cortex. (ii) Areas comprising the limbic system such as the cingulate gyrus, pyriform cortex, amygdala, anterior and medial thalamus, hippocampus, septum, and diagonal band of Broca contained relatively high (300 to 400 pg/g) to high (900 to 1400 pg/g) NGF levels. (iii) In the brain stem and cerebellum, the level was low; however, a relatively high level was found in the cerebellar nuclei, lateral vestibular nucleus, pontine reticular nucleus, superior olive, and ventral co&ear nucleus. The NGF levels of the ~pp~ampus and cerebral cortex were much higher than those of basal forebrain. Thus, the NGF-producing sites seemed to contain a much higher level than the sites to which NGF is transported. However, in the rat peripheral nervous system, the relationship between NGF-producing sites and terminal sites for NGF transportation is the converse.
Namely, sympathetic ganglia contain a much higher level of NGF than NGF-producing sites such as the heart and submaxillary glands (30). One explanation for the former fact is that NGF in these NGF-rich regions, i.e., the hippocampus or cerebral cortex, may act as a trophic factor for the interneurons within the hippocampus or cerebral cortex. Recently, Lu et al. (17) reported a local delivery mechanism of NGF in the brain during embryonic stages. ~though we used young adult rats, a part of NGF may act locally on the hippocampal or cortical interneurons. Another possible explanation is that yet unidentified NGF-responding neurons that project to the hippocampus or cortex may take up NGF molecules at their nerve terminals. The limbic system is functionally related to emotion, memory, and the autonomic nervous system (4). NGF may be supplied as a tropbic factor for not only cholinergic neurons in the basal forebrain but also other neurons in the limbic system. Because the NGF mRNA level in the diencephalon is reported to be moderately high (11, 13, 28, 32), the diencephalon may produce NGF and receive nerve terminals from NGF-responsive neurons. The areas including the anterior and medial thalamus, which receive the afferent projections from the mammillary body, amygdala, frontotemporal cortex, and substantia innominata (4), contained relatively high NGF levels in the diencephalon (pieces I-14, -15, -19, and J-15 in Fig. 4). The positive immunoreactivity of NGF is shown in the embryonic murine anterior thalamus (8), and NGFR-immunoreactivity is expressed in the postnatal rat thalamus (7, 35). These support the
DETAILED
DISTRIBUTION
idea that NGF acts as atrophic factor for the neurons in the anterior and medial thalamus. The brain stem contains a moderate level of NGF mRNA (11,28,32); and some populations of neurons in the brain stem, such as the superior olive and cochlear nuclei, dorsolateral lemniscus, periventricular column, raphe nucleus, and reticular nucleus, express NGFRs (23, 25, 35). Ebbot and Hendry (6) reported the retrograde transport of NGF from the cerebellum to the tegmental reticular nucleus of pons, dorsal nucleus of the lateral lemniscus, and dorsal raphe. Our study revealed that although the NGF level in the brain stem was low, the pontine reticular nucleus (pieces N-18 and P-13 in Fig. 4), the superior olive (piece P-17), and the ventral cochlear nucleus (piece Q-15) contained a relatively high NGF level. Therefore, the neurons in these nuclei may be NGF-responsive, because the brain area of high NGF level should be either a NGF-producing site or a NGF-transported site, i.e., an area where NGF-responsive neurons exist. The cerebellum is likely to be a terminal site for NGF transportation rather than a NGF-producing site, because the cerebellar NGF mRNA level in the adult rat is extremely low (11,13,14,28,32), whereas NGF is unambiguously detected in the cerebellum. Accordingly, the cerebellar neurons projecting efferent fibers may be one of the candidates for NGF-responsive neurons. All cerebellar efferent fibers are projected from Purkinje cells in the cerebellum (4) and high- and low-affinity NGFRs are expressed on these cells (522). Our data revealed a relatively high NGF level in the cerebellar nuclei and lateral vestibular nucleus (pieces Q-11 and -12 in Fig. 4), nuclei that receive afferent projections from Purkinje cells. These findings suggest that Purkinje ceils may respond to NGF. Immunohistochemistry and in situ hybridization experiments have shown that NGF is predominantly expressed by neurons (1, 3, 10, 24, 29, 34). Synthesis of NGF is regulated by neuronal activity via non-NMDA glutamate receptors (36). These facts suggest the cellular source of NGF synthesis may be neurons in the CNS. Although our data do not allow identification of the cellular source of NGF synthesis, neurons or glial cells, we are now approaching this issue by the application of our EIA system to cryostat-sectioned tissue samples. In conclusion, the neurons in the anteromedial portion of the thalamus, pontine reticular nuclei, superior olive and ventral cochlear nuclei, and cerebellar Purkinje cells are proposed to be additional NGF-responsive neurons in the rat brain.
NGF
2.
AYER-IZLIEVRE, C. S., T. EBENDAL, L. OLSON, 1983. Localization of NGF-like immunoreactivity tissue. Med. Biol. 61: 296-304.
3.
AYER-LELIEVRE, C., L. OLSON, T. EBENDAL, A. SEIGER, AND H. PERSSON. 1988. Expression of the p-nerve growth factor gene in hippocampal neurons. Science 240: 1339-1341. CARPENTER, M. B. 1985. Core text of neuroanatomy, 3rd ed., Williams & Wilkins, Baltimore.
4.
ALoE, young 5640.
L. 1989. Adrenalectomy decreases nerve growth rat hippocampus. Proc. Natl. Acad. Sci. USA
factor
BRAIN
AND A. SEIGER. in rat neurons
COHEN-CORY, S., C. F. DREYFUS, AND I. B. BLACK. 1989. Expression of high- and low-aftlnity nerve growth factor receptors by Purkinje cells in the developing rat cerebellum. Ezp. Neural. 105: 104-109.
6.
EBBOT, S., AND I. HENDRY. 1978. Retrograde transport growth factor in the rat central nervous system. Brain 160-163.
of nerve
Res. 139:
7.
ECKENSTEXN, F. 1988. Transient expression of NGF-receptorlike immunoreactivity in postnatal rat brain and spinal cord. Brain Res. 446: 149-154.
8.
FINN, P. J., I. A. FERGUSON, P. A. WILSON, J. VAHAVIOLOS, AND R. A. RUSH. 1987. Immunohistoehemicel evidence for the distribution of nerve growth factor in the embryonic mouse. J. Neurocytol. 16: 639-647.
9.
FURUKAWA, S., I. KAMO, Y. FURUKAWA, S. AKA~AWA, E. SATOYOSHI, K. ITOH, AND K. HAYASHI. 1983. A highly sensitive enzyme immunoassay for mouse fi nerve growth factor. J. Neurochem. 40: 734-744.
10.
GALL, C. M., AND P. J. ISACKSON. neuronal production of messenger Science 240: 1339-1341.
11.
GOEDERT, M., A. FINE, S. P. HUNT, AND A. ULLFZICH. 1988. 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. Mol. Brain Res. 1: 85-92.
12.
HOHN, A., J. LEIBROCK, K. BAILEY, AND Y.-A. BARDE. 1990. Identification and characterization of a novel member of the nerve growth factor/brain derived neurotrophic factor family. Nature 344: 339-341. KORSCHING, S., G. AUBURGER, R. HE~MANN, J. SCOTT, AND H. THOENEN. 1985. Levels of nerve growth factor end its mRNA in the central nervous system of the rat correlate with choline&c innervation. EMBO J. 4: 1389-1393. LARGE, T. H., S. C. BODARY, D. 0. &EGG, G. WESKAMP, U. @ITJ3N, AND L. F. REXCHARDT. 1988. Nerve growth factor gene expression in the developing rat brain. Science 234: 352-355.
13.
14.
15.
1989. Limbic seizures increase RNA for nerve growth factor.
LARKFORS, L., T. EBENDAL, S. R. WH~MORE, H. PERSSON, B. HOFFER, AND L. OLSON. 1987. Decreased level of nerve growth factor (NGF) and its messenger RNA in the aged rat brain.
Brain Res. 427: 55-60. 16.
LEIBROCK, J., F. L~~SPEICH, A. HOHN, M. HOFER, B. HANGERER, P. MASIAKOWSKI, H. THOENEN, AND Y.-A. BARDE. 1989. Molecular cloning and expression of brain-derived neurotrophic factor. Nature 341: 149-152.
17.
Lu, B., C. R. BUCK, C. F. DREYFUS, pression of NGF and NGF receptor brain: Evidence for local delivery Neurol. 104: 191-199.
18.
MAISONPIERRE, P. C., L. BELLUSCIO, B. FRIEDMAN, R. F. ANDERSON, S. J. WIEGAND, M. E. FORTH, R. M. LINDSAY, AND G. D. YANCOPOULOS. 1990. NT-3, BDNF, and NGF in the developing rat nervous system: Parallel as well as reciprocal pattern of expression. Neuron 5: 501-509.
19.
MARTINEZ, H. J., C. F. DREYFUS, G. MILLZR BLACK. 1985. Nerve growth factor promotes
in
86: 5636-
RAT
5.
REFERENCES 1.
IN
83
OF
AND I. B. BLACK. 1989. EXmRNAs in the developing and action of NGF. Exp.
JONAKAIT, choline@
AND I. B. develop-
84
20.
21.
22.
23.
24. 25.
26.
27.
28.
NISHIO,
AKIGUCHI,
ment in brain striatal cultures. Proc. Natl. Acad, Sci. USA 82: 1777-7781. ’ MEIER, R., M. B&~~-ANDRE, R. GOT-Z, R. HEIJMANN, A. SHAW, AND H. TI-IOENEN. 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. EMBO J. 5: 1489-1493. PALKOVITS, M., AND M. J. BROWNSTEIN. 1988. Maps and guide to microdissection of the rat brain. Elsevier, Amsterdam/New York. PIORO, E. P., AND A. C. CUEUO. 1988. Purkinje cells of adult rat cerebellum express nerve growth factor receptor immunoreactivity: Light microscopic observations. Brain Res. 455: 82-186. RAMCH, G., AND G. W. KREUTZBERG. 1987. The localization and distribution of high affinity b-nerve growth factor binding sites in the central nervous system of the adult rat, a light microscopic autoradiographic study using [iZI] b-nerve growth factor. Neuroscience 20: 23-36. RENNERT, P. D., AND G. HJIINRICH. 1986. Nerve growth factor mRNA in brain: Localization by in situ hybridization. Biochem. Biophys. Res. Commun. 138: 813-819. RICHARDSON, P. M., V. M. K. VERGE-ISSA, AND R. J. RIOPELLE. 1986. Distribution of neuronal receptors for nerve growth factor in the rat. J. Neurosci. 6: 2312-2321. SCHWAB, M. E., U. @-TEN, Y. AGID, AND H. THOENEN. 1979. Nerve growth factor (NGF) in the rat CNS: Absence of specific retrograde axonal transport and tyroeine hydroxylase induction in locus coeruleus and substantia nigra. Brain Res. 188: 473483. SEILER, M., AND M. E. SCHWAB. 1984. Specific retrograde transport of nerve growth factor (NGF) from neocortex to nucleus basalis in the rat. Brain Res. 300: 33-39. SHELTON, D. L., AND L. F. REICHARDT. 1986. Studies on the expression of the nerve growth factor (NGF) gene in the central
AND FURUKAWA
29.
30.
31.
32.
33. 34.
35.
nervous system: Level and regional distribution of NGFmRNA suggest that NGF functions as a trophic factor for several distinct populations of neurons. Proc. Natl. Ad. Sci. USA 83: 2714-2718. SPILLANTINI, M. G., E. ALOE, E. ALLEVA, R. DE SIMONJZ, M. GOEDERT, AND R. LEVI-M• NTALCINI. 1989. Nerve growth factor mRNA and protein increase in hypothalamus in a mouse model of aggression. Proc. Natl. Ad. Sci. USA 86: 6555-6559. THEONEN, H., C. BANDTLOW, AND R. HEUMANN. 1987. Thephysiological function of nerve growth factor in the central nervous system: Comparison with the periphery. Rev. Physid. Biochem. Phurmacol. 109: 146-178. WESKAMP, G., AND U. OTTEN. 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. WHITTEMORE, S. R., T. EBENDAL, L. L-Rs, L. OLSON, A. SEIGER, I. STROMBERG, AND H. PER~SON. 1986. Developmental and regional expression of @nerve growth factor messenger RNA and protein in the rat central nervous system. Proc. Natl. Ad. Sci. USA 83: 817-821. WHITTEMORE, S. R., AND A. SEIGER. 1987. The expression, localization and functional significance of p-nerve growth factor in the central nervous system. Brain Res. Rev. 12: 439-464. WHITTEMORE, S. R., P. L. FFUEDMAN, D. LARHAM MAR, H. PF,RSSON, M. GONZALEZ-CARVAJAL, AM) V. R. HOLET~. 1988. Rat 8nerve growth factor sequence and site of synthesis in the adult hippocampus. J. Neurosci. Res. 20: 403-410. YAN, Q., AND E. M. JOHNSON. 1988. An immunohistochemical study of the nerve growth factor receptor in developing rats. J. Neurosci. 8: 3481-3498.
36. ZAFFU,
F., B. HENGERER,
J. LEIBROCK,
H. THOJZNJN,
AND D.
1990. Activity dependent regulation of BDNF and NGF mRNAs in the rat hippocampus is mediated by nonNMDA glutamate receptors. EMBO J. 9: 3645-3560. LINDHOLM.
116,76-84
(1692)
Detailed Distribution of Nerve Growth Factor in Rat Brain Determined by a Highly Sensitive Enzyme Immunoassay TAKESHI NISHIO,*,~ *Department
ICHIRO AKIGUCHI,*
AND SHOEI FURUKAWA~
of Neurology, Faculty of Medicine, Kyoto University, Kawahara-Cho 54, Shogoin, Sakyo, Kyoto, Japan; and tDivision Neuroimmurwlogy, National Institute of Neuroscience, NCNP, Ogawa-Higashi 4-1-1, Kodaira, Tokyo, Japan
of
Is NGF-responsiveness limited to only these neurons in the central nervous system (CNS)? In the rat brain stem, a moderate level of NGF mRNA (11, 28, 32) and high-affinity NGFRs (23, 25) have been detected. Furthermore, a low level of NGF protein is unambiguously detectable in the cerebellum (13, 14, 31), although no cholinergic neurons are contained in this region. These facts suggest that NGF acts as a neurotrophic molecule on some sorts of central neurons in addition to the MCNs or striatal neurons. Histological visualization of NGF and NGF mRNA in the CNS may contribute greatly to the elucidation of the role of NGF in the CNS. However, the immunohistochemical data on NGF distribution (2, 8, 32) have shown no obvious correlations to the quantitative data determined by a reliable enzyme immunoassay (EIA) (13,14,31); and ins&u hybridization has been successful for visualization of NGF mRNA only in regions such as the hippocampus and/or cerebral cortex where the NGF mRNA level is relatively high (3,10,24,29,34). This lack of detection is probably because the levels of NGF and NGF mRNA are extremely low in most of the CNS (30). Therefore, as a challenge to detect low levels of NGF, we developed a new EIA system that enabled us to measure the NGF content of as small as 2 mg (wet weight) of brain tissue; and by using this method we determined the detailed distribution of NGF in the adult rat brain.
We modified a previously reported enzyme immunoassay method to make it more sensitive for quantification of nerve growth factor (NGF), and succeeded in measuring the NGF content in as small as 2 mg (wet weight) of rat brain tissue. Rat brain was cut into about 600 pieces of the same size, and the NGF content in each piece was determined by this method. The findings were as follows: (i) In the cerebral cortex, NGF contents were unevenly distributed, ranging from less than 0.1 to 1.8 rig/g wet wt. The level was highest in the caudal parietal and rostra1 occipital cortices and lowest in the lateral parietal cortex. (ii) Areas comprising the limbic system such as the cingulate gyrus, pyriform cortex, amygdala, anterior and medial thalamus, hippocampus, septum, and diagonal band of Broca contained high levels of NGF. (iii) In the brain stem and cerebellum, the levels were low; however, a relatively high level was registered in the cerebellar nuclei, lateral vestibular nucleus, ventral cochlear nucleus, superior olive, and pontine reticular nuclei. These findings, taken together with previously published information, suggest that the neurons in the anterior and medial thalamus, pontine reticular nuclei, superior olive, ventral cochlear nucleus, and cerebellar Purkinje cells may be additional populations of NGF-responsive neurons in the rat brain. 0 1992 Academic Press, Inc.
INTRODUCTION MATERIALS It has been so far established that nerve growth factor (NGF) acts as a neurotrophic factor for the magnocellular cholinergic neurons (MCNs) in the basal forebrain (30, 33). The evidence for the retrograde axonal transport of NGF from the neocortex to the basal forebrain (26, 27), and the presence of high levels of NGF and NGF mRNA in the projection area of the MCNs (11,13, 14, 28, 31, 32) and of high-affinity NGF receptors (NGFR’s) in the cell bodies of the MCNs (23, 25) suggest a role for NGF as a trophic factor on MCN in the basal forebrain. Although the data are not so numerous as in the case of MCN, striatal intrinsic choline@ neurons should also be included in this category (19,23,25). 0014-4886/92 $3.00 Copyright 0 1992 by Academic Press, All rights of reproduction in any form
Materials. Wistar rats were purchased from Japan SLC, Inc. Bovine serum albumin (BSA) was obtained from Sanko Junyaku; gelatin, from Bio-Rad; EDTA, from Dojin; and aprotinin, from Sigma. Other chemicals were reagent grade. NGF was purified from mouse submaxillary glands as previously described (9). Preparation of rut brain samples. The brains of 7week-old Wistar rats (n = 3), which were taken from animals killed by deep narcosis induced by diethyl ether, were coronally sectioned into 0.9-mm-thick slices with a cryostat at -10°C. Each coronal slice was perpendicularly cut into squares (1.5 X 1.5 mm) with a stainless 76
Inc. reserved.
AND METHODS
DETAILED
DISTRIBUTION
14nm -I
I-
FIG. 1. Schematic representation of the preparation of samples. Wistar rat brain was coronaily sectioned into 0.9-mm-thick slices with a cryostat at -10°C. Each coronal slice was perpendicularly cut into squares (1.5 X 1.5 mm) with a razor blade, resulting in small pieces having maximal dimensions of 0.9 X 1.5 X 1.5 mm (2.0 mm3).
razor blade (Fig. 1). The brain samples of 0.9 X 1.5 X 1.5 mm (2.0 mm3) were prepared. The sectioning of the coronal slices was adjusted by cutting the brain initially at the plane including hregma and sectioning into 1.5 X 1.5-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 (21) and the dissections of brain areas were reliably reproducible. ~eter~i~utio~ of o~ti~~~ co~~it~~s for NGF extruction. Such small and numerous brain sections cannot be homogenized easily and reliably and the background signals measured in our EIA system were lower with the extraction of NGF by freeze-thaw procedure than with the extraction by homogenization. Therefore, we tried to extract NGF with buffer A (0.1 M Tris-HCI buffer ]pH 7.61, containing 2% [w/v] BSA, 2% [w/v] gelatin, 1.0 M NaCl, 2 m&f EDTA, and 80 trypsin-inhibitory units of aprotinin/liter) from each piece of the rat brain by a freeze-thaw procedure. Each piece was soaked in 100 ~1 of buffer A in a 96-multiwell plate (Falcon), frozen at -8O”C, and thawed completely at room temperature. The effect of frequency of this freezing-thawing
OF NGF IN RAT BRAIN
77
procedure on NGF extraction was monitored by measurement of NGF released into the solution, and the degree of recovery of NGF by this procedure was compared with that obtained by homogenization. Several pieces from a defined area in the rat brain (the caudal parietal cortex) were homogenized by 30 strokes with a glass pestle (Wheaton) in buffer A with or without exogenous NGF. The homogenates were centrifuged at 100,OOOg for 10 min and NGF content was determined by the EIA described below. The recovery of NGF in the extract was calculated as the ratio of (NGF content with exogenous NGF) - (NGF content without exogenous NGF) /exogenously added NGF. HA for NGF. NGF contents were essentially measured by the highly sensitive two-site EIA previously described (9). However, it was drastically modified to minimize sample volume as follows: 150 ng anti-mouse NGF rabbit IgG in 5 ~1 of 0.05 M Tris-HCl buffer (pH 8.5) was placed on the center of a U-bottom well of 96multiwell polystyrene plate (Falcon). After incubation for 30 min at room temperature, the IgG solution was removed, and the wells were washed twice with 0.15 ml of 0.1 M Tris-HCl buffer (pH 7.6) containing 0.4 M NaCi, 0.1% BSA, 0.02% NaN,, and 1 mM MgCl, (buffer B). Each of the IgG-coated wells was incubated with gentle shaking for 4 h at room temperature with 10 ~1 of sample solution prepared in buffer A, after blockade of nonoccupied space with 0.15 ml of buffer B containing 0.05% (v/v) bovine milk by a l-h incubation. Then, each well was washed twice with 0.15 ml of buffer B. Twenty microliters of buffer B cont~ning biotinylated antibodies (10 nglml), which were affinity-purified polyclonal rabbit antibodies against murine NGF, was applied to each well, and the plate was incubated for 12 h at 4°C with gentle shaking. After two washings, 20 ~1 of /3-D-galactosidase-conjugated streptavidin (Biogenex Lab) dissolved in buffer B was applied to each well, and the plate was incubated for 1 h at room temperature with gentle shaking. The enzyme reaction was started at room temperature by the addition of 30 ~1 of 6 pM 4methylumbelliferyl-P-D-galactoside in buffer B and stopped by that of 0.1 ml of 0.1 M glycine-NaOH buffer (pH 10.3). The amounts of 4-methylumbelliferone formed were measured by fluorometry with a spectrofluorometer (model 650-60, Hitachi). The wavelengths 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 IgG-coated wells or noncoated wells. NGF contents in the samples were evaluated by reference to the standard curve. The brain tissue pieces cut into smaller than a 1.5 X 1.5-mm square, i.e., those at the marginal areas, were sketched immediately and their wet weight was calculated. NGF content evaluated
78
NISHIO,
.i
. 1. . lb .1&l ioilo Content of p-NGF
loo00
AKIGUCHI,
rtioo
f$/assay
FIG. 2. Standard curve of the enzyme immunoassay for mouse &NGF. Experimental procedures are described in the text. The detection limit in this assay was as low as 10 fg of NGF per assay well (asterisk).
by EIA was expressed as amount weight,
per gram tissue wet
RESULTS
Improved methods for detecting NGF. The assay system after modification could detect as low as 10 fg of NGF per assay well (Fig. 2). The relationship between the efficiency of NGF extraction and the frequency of freezing-thawing is shown in Fig. 3. The ordinate values in Fig. 3 are expressed as percentages of the values obtained by homogenization with a glass pestle. The release of NGF into the solution was first observed after the third freeze-thaw cycle and increased abruptly after five or six cycles. A plateau value was reached after eight cycles. As recovery of NGF by homogenization was over 90%, we considered eight cycles of freezing-tha~ng as sufficient for a practical and quantitative extraction of NGF from the brain pieces. So, we routinely extracted the factor from the brain pieces by this number of freeze-thaw cycles. There were no recognizable regional differences in the efficiency of NGF extraction by the freezing-thawing procedure (data not shown). NGF content ~di~d EIA.
in small bruin sections determined
AND
FURUKAWA
3) of NGF content per each brain piece. In the cerebral cortex, NGF content was distributed unevenly with great regional differences in NGF level (about a maximal 20-fold difference), ranging from the level of less than 100 pg/g-wet tissue [lateral frontopolar cortex (B3), lateral striatum (E-15, F-14, G-12, 18, 23), lateral parietal cortex (G-15, J-8, and -9)] to about 1800 pg/gwet tissue [caudal parietal cortex (K-4)] pg/g-wet tissue. The pieces K-l, -2, -3, -4, L-2, M-l, -2, and -3, comprising the dorsomedial portions of the caudal parietal cortex and rostra1 occipital cortex, exhibited high NGF levels (800 to 1800 pg/g-tissue). The pieces B-12, C-17, E-20, -23, F-24, H-28, I-30, -31, J-31, K-29, -30, and -31, representing the pyriform cortex, showed moderately high levels (500 to 1000 pg/g). The pieces F-7, -11, G-9, -15, -20, H-9, -14, -19, I-8, -13, -18, J-9, -14, and -19, taken from the lateral parietal cortex, showed low NGF levels (less than 200 pg/g). The pieces E-l, -5, G-l, and H-l, reflecting the cingulate gyrus, showed moderately high levels (500 to 700 pg/g). The pieces J-10, -11, K-11, L-7, -12, -17, -18, -22, -23, M-12, -13, and -22 from the hippocampus exhibited high NGF levels (900 to 1400 pg/g-wet tissue). In coronal slice J, the pieces including the dentate gyrus and areas CA3-CA4 (J-10 and -11) exhibited higher NGF levels than the pieces, J-5 and -6, comprising the areas CAl-CAB. The pieces E-9, -17, F-8, and -12, indicating the septal nucleus and diagonal band of Broca, showed moderate levels of NGF (600 to 1200 pg/g). However, NGF levels in the pieces G-22 and H-21, reflecting the nucleus basalis of Meynert and globus pallidus were low (100 pg/g-wet tissue). The levels
3 802 c 5 60 8
2
by the
The ~stribution of NGF contents in the pieces from the whole rat brain is shown in Fig. 4, and its superficial distribution in the cerebral cortex unfolded into a single plane is shown in Fig. 5. The value of NGF content in each piece obtained from each animal deviated by 10 to 30% from the mean value. Each circle at the center of a square represented the mean value (n =
o~@loa 0
2
4
6
810
Frequency of freezing and thawfng FIG. 3. Relationship between efficiency of NGF extraction and frequency of freezing and thawing. The ordinate values are expressed as percentages of the values obtained by homogenization.
DETAILED
DISTRIBUTION
in the striatum (pieces D-9, -13, E-11, -14, F-13, -14, -17, -18, G-12, and -18) were also generally low (less than 200 pg/g), although pieces E-10 and F-9 contained moderate levels of NGF (450 and 900 pg/g-wet tissue, respectively). Pieces from the anterior and medial portions of thalamus, I-14, -15, -19, and J-15, showed relatively high NGF levels (300 to 400 pg/g) compared with the pieces I-20, J-16, -17, -20, -21, K-15, -16, -17, and -21 from the other portions of the thalamus (less than 100 pg/g). Pieces I-24, J-25, and K-25 of the hypothalamus exhibited low NGF levels (less than 300 pg/g), and those comprising the amygdaloid body, I-26, K-27, and -28, contained moderate levels (600 to 700 pg/g). The brain stem and cerebellum generally showed low NGF levels (less than 100 pg/g). However, the pieces N-18 and P-13, containing the pontine reticular nucleus; P-17, including the superior olive; and Q-11, -12, and -15, comprising the cerebellar nuclei, lateral vestibular nucleus, ventral cochlear nucleus, and paraflocculus, showed relatively high (150 to 250 pg/g) NGF levels compared with other pieces in the brain stem and cerebellum.
DISCUSSION
Knowledge about the distribution of NGF in the CNS is very important for us to elucidate the role of this growth factor in the CNS, although there have been some studies on the immunohistochemical visualization of NGF in embryonic murine, fetal, and adult rat brain (2,8,32). Because of the extremely low level in the CNS, the most reliable method so far to detect NGF in the CNS has been the EIA. We modified our earlier EIA (9) to measure the extremely low level of NGF contained in a very small amount of brain tissue. The newly modified assay system showed a detection limit of as low as 10 fg of NGF per assay well (Fig. 2). The points responsible for this marked improvement in sensitivity were as follows: (i) specific (affinity-purified) anti-NGF antibodies were used for second antibodies to sandwich the antigen; (ii) second antibodies were biotinylated and used in the streptavidin-biotin system; (iii) the system was minimized in terms of sample volume by the use of Ubottom 96-multiwell plates; (iv) extraction was successfully achieved by freezing-thawing, which enabled us to
OF
NGF
IN
RAT
BRAIN
79
extract NGF quantitatively from a brain piece as small as one of 2-mg wet wt (Fig. 3). The specificity of the anti-NGF antibodies in the present study was tested previously by immunodiffision analysis, immunoadsorption method, and polyacrylamide gel electrophoresis (9) and these antibodies specifically recognized NGF. Recently, other members of neurotrophic factors with about 50% homology of amino acid sequence, brain-derived neurotrophic factor (BDNF), neurotrophin-3 (NT-3), which are called NGF family, have been found (12,16). Although we cannot exclude the possibility that the antibodies in use may cross-react to these members of NGF family and our EIA system may count falsepositive signals, the possibility seems very small because of the following points. (i) NGF have five hydrophilic regions expected to be potential antigenic determinants (20), and small parts of them are homologous in amino acid sequence to BDNF or NT-3 (12). Even though NGF shares some antigenic similarities with those molecules, the signals detected false-positively in this assay system would be small or negligible, because this assay system requires at least two epitopes for detection, i.e., two-site EIA. (ii) The levels of mRNA of BDNF and NT-3 in adult rat cerebellum are high whereas NGF mRNA level in cerebellum is very low (18). If our assay system count false-positive signals of BDNF or NT-3, the cerebellar signals detected in the present study would be expectedly high. But, cerebellar NGF levels were much lower in our study than hippocampus or cerebral cortex levels. Therefore, the antiNGF antibodies in use scarcely cross-react to BDNF or NT-3 and the assay system in the present study specifically detects NGF. NGF and NGF mRNA levels in the major brain regions, e.g., the cerebral cortex, hippocampus, or septal nucleus, have been measured by other investigators (11, 13-15,28,31,32). The cerebral cortex and hippocampus contain high levels of both NGF and NGF mRNA, whereas the septal nucleus contains a moderate amount of NGF but a low NGF mRNA level (13,14). Based on evidence such as the retrograde axonal transport of radiolabeled NGF (26,27) and the distribution of bigh-affinity NGFRs (23,25), NGF has been considered to act as a tropbic factor in the CNS only on MCNs in the basal forebrain and on striatal intrinsic choline& neu-
FIG. 4. Distribution map of NGF contents in the whole rat brain. The circle at the center of each square represents the mean value of NGF content in the piece determined by EIA. Note the blue circles represent a lower range of values than the red circles. Abbreviations used in the figure are as follows: a, amygdala; bn, basal nucleus; c, caudate nucleus; Ci, cingulate gyrus; cn, ventral cochlear nucleus; cp, caudate-putamen; DG, dentate gyrus; Di, diagonal band of Broca; En, entorhinal cortex; FC, frontal cortex; fn, facial nucleus; FPC, frontopolar cortex; gp, globus pallidus; Hl, CA1 area; H2, CA2 area; H3, CA3 area; H4, CA4 area; ht, hypothalamus; i, interpose nucleus; ic, internal capsule; ice, inferior colliculus; 1, lateral cerebellar nucleus; lg, lateral geniculate body; Iv, lateral vestibular nucleus; m, medial cerebellar nucleus; mg, medial geniculate body; mr, midbrain reticular formation; Oc, occipital cortex; p, pyramidal tract; PC, parietal cortex; pfl, paraflocculus; pg, periaqueductal gray; pn, pontine nucleus; pr, pontine reticular formation; pt, pontine tegmental nucleus; Py, pyriform cortex; r, red nucleus; re, gigantocellular reticular nucleus; RS, retrosplenial cortex; S, septal nucleus; SC, superior colliculus; sn, substantia nigra; so, superior olive; stn, spinal tract nucleus; t, trigeminal nerve; Te, temporal cortex; th, thalamus; tm, trigeminal motor nucleus.
80
NISHIO,
AKIGUCHI,
FIG,
AND FURUKAWA
I--Continued
DETAILED
DISTRIBUTION
OF NGF IN RAT BRAIN
b
FIG. 4-Continued
81
82
NISHIO,
AKIGUCHI,
AND FURUKAWA
Parietal
Occipital
100
Is
200
300
400
500
000
700
000
900
1000
1200
1000 i&g-time
FIG. 5. Distribution map of NGF contents in the surface of the cerebral cortex, unfolded into a single plane. The circle at the center of each square represents NGF content.
Ions. The present study generally supports this theory, because both areas where NGF should be synthesized, i.e., the cerebral cortex and hippocampus, and the area to which NGF should be transported, i.e., the nuclei of the basal forebrain, contained high NGF contents. In addition, this study also confirmed the fact that the NGF levels in the diencephalon, brain stem, andcerebellum are low, as previously reported (13,14,31). However, this investigation is noteworthy because the distribution of NGF contents in the brain was determined in detail and the following new facts have emerged: (i) In the cerebral cortex, NGF was unevenly distributed ranging from less than 0.1 to 1.8 rig/g wet tissue. The level was highest in the caudal parietal and rostra1 occipital cortices and lowest in the lateral parietal cortex. (ii) Areas comprising the limbic system such as the cingulate gyrus, pyriform cortex, amygdala, anterior and medial thalamus, hippocampus, septum, and diagonal band of Broca contained relatively high (300 to 400 pg/g) to high (900 to 1400 pg/g) NGF levels. (iii) In the brain stem and cerebellum, the level was low; however, a relatively high level was found in the cerebellar nuclei, lateral vestibular nucleus, pontine reticular nucleus, superior olive, and ventral co&ear nucleus. The NGF levels of the ~pp~ampus and cerebral cortex were much higher than those of basal forebrain. Thus, the NGF-producing sites seemed to contain a much higher level than the sites to which NGF is transported. However, in the rat peripheral nervous system, the relationship between NGF-producing sites and terminal sites for NGF transportation is the converse.
Namely, sympathetic ganglia contain a much higher level of NGF than NGF-producing sites such as the heart and submaxillary glands (30). One explanation for the former fact is that NGF in these NGF-rich regions, i.e., the hippocampus or cerebral cortex, may act as a trophic factor for the interneurons within the hippocampus or cerebral cortex. Recently, Lu et al. (17) reported a local delivery mechanism of NGF in the brain during embryonic stages. ~though we used young adult rats, a part of NGF may act locally on the hippocampal or cortical interneurons. Another possible explanation is that yet unidentified NGF-responding neurons that project to the hippocampus or cortex may take up NGF molecules at their nerve terminals. The limbic system is functionally related to emotion, memory, and the autonomic nervous system (4). NGF may be supplied as a tropbic factor for not only cholinergic neurons in the basal forebrain but also other neurons in the limbic system. Because the NGF mRNA level in the diencephalon is reported to be moderately high (11, 13, 28, 32), the diencephalon may produce NGF and receive nerve terminals from NGF-responsive neurons. The areas including the anterior and medial thalamus, which receive the afferent projections from the mammillary body, amygdala, frontotemporal cortex, and substantia innominata (4), contained relatively high NGF levels in the diencephalon (pieces I-14, -15, -19, and J-15 in Fig. 4). The positive immunoreactivity of NGF is shown in the embryonic murine anterior thalamus (8), and NGFR-immunoreactivity is expressed in the postnatal rat thalamus (7, 35). These support the
DETAILED
DISTRIBUTION
idea that NGF acts as atrophic factor for the neurons in the anterior and medial thalamus. The brain stem contains a moderate level of NGF mRNA (11,28,32); and some populations of neurons in the brain stem, such as the superior olive and cochlear nuclei, dorsolateral lemniscus, periventricular column, raphe nucleus, and reticular nucleus, express NGFRs (23, 25, 35). Ebbot and Hendry (6) reported the retrograde transport of NGF from the cerebellum to the tegmental reticular nucleus of pons, dorsal nucleus of the lateral lemniscus, and dorsal raphe. Our study revealed that although the NGF level in the brain stem was low, the pontine reticular nucleus (pieces N-18 and P-13 in Fig. 4), the superior olive (piece P-17), and the ventral cochlear nucleus (piece Q-15) contained a relatively high NGF level. Therefore, the neurons in these nuclei may be NGF-responsive, because the brain area of high NGF level should be either a NGF-producing site or a NGF-transported site, i.e., an area where NGF-responsive neurons exist. The cerebellum is likely to be a terminal site for NGF transportation rather than a NGF-producing site, because the cerebellar NGF mRNA level in the adult rat is extremely low (11,13,14,28,32), whereas NGF is unambiguously detected in the cerebellum. Accordingly, the cerebellar neurons projecting efferent fibers may be one of the candidates for NGF-responsive neurons. All cerebellar efferent fibers are projected from Purkinje cells in the cerebellum (4) and high- and low-affinity NGFRs are expressed on these cells (522). Our data revealed a relatively high NGF level in the cerebellar nuclei and lateral vestibular nucleus (pieces Q-11 and -12 in Fig. 4), nuclei that receive afferent projections from Purkinje cells. These findings suggest that Purkinje ceils may respond to NGF. Immunohistochemistry and in situ hybridization experiments have shown that NGF is predominantly expressed by neurons (1, 3, 10, 24, 29, 34). Synthesis of NGF is regulated by neuronal activity via non-NMDA glutamate receptors (36). These facts suggest the cellular source of NGF synthesis may be neurons in the CNS. Although our data do not allow identification of the cellular source of NGF synthesis, neurons or glial cells, we are now approaching this issue by the application of our EIA system to cryostat-sectioned tissue samples. In conclusion, the neurons in the anteromedial portion of the thalamus, pontine reticular nuclei, superior olive and ventral cochlear nuclei, and cerebellar Purkinje cells are proposed to be additional NGF-responsive neurons in the rat brain.
NGF
2.
AYER-IZLIEVRE, C. S., T. EBENDAL, L. OLSON, 1983. Localization of NGF-like immunoreactivity tissue. Med. Biol. 61: 296-304.
3.
AYER-LELIEVRE, C., L. OLSON, T. EBENDAL, A. SEIGER, AND H. PERSSON. 1988. Expression of the p-nerve growth factor gene in hippocampal neurons. Science 240: 1339-1341. CARPENTER, M. B. 1985. Core text of neuroanatomy, 3rd ed., Williams & Wilkins, Baltimore.
4.
ALoE, young 5640.
L. 1989. Adrenalectomy decreases nerve growth rat hippocampus. Proc. Natl. Acad. Sci. USA
factor
BRAIN
AND A. SEIGER. in rat neurons
COHEN-CORY, S., C. F. DREYFUS, AND I. B. BLACK. 1989. Expression of high- and low-aftlnity nerve growth factor receptors by Purkinje cells in the developing rat cerebellum. Ezp. Neural. 105: 104-109.
6.
EBBOT, S., AND I. HENDRY. 1978. Retrograde transport growth factor in the rat central nervous system. Brain 160-163.
of nerve
Res. 139:
7.
ECKENSTEXN, F. 1988. Transient expression of NGF-receptorlike immunoreactivity in postnatal rat brain and spinal cord. Brain Res. 446: 149-154.
8.
FINN, P. J., I. A. FERGUSON, P. A. WILSON, J. VAHAVIOLOS, AND R. A. RUSH. 1987. Immunohistoehemicel evidence for the distribution of nerve growth factor in the embryonic mouse. J. Neurocytol. 16: 639-647.
9.
FURUKAWA, S., I. KAMO, Y. FURUKAWA, S. AKA~AWA, E. SATOYOSHI, K. ITOH, AND K. HAYASHI. 1983. A highly sensitive enzyme immunoassay for mouse fi nerve growth factor. J. Neurochem. 40: 734-744.
10.
GALL, C. M., AND P. J. ISACKSON. neuronal production of messenger Science 240: 1339-1341.
11.
GOEDERT, M., A. FINE, S. P. HUNT, AND A. ULLFZICH. 1988. 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. Mol. Brain Res. 1: 85-92.
12.
HOHN, A., J. LEIBROCK, K. BAILEY, AND Y.-A. BARDE. 1990. Identification and characterization of a novel member of the nerve growth factor/brain derived neurotrophic factor family. Nature 344: 339-341. KORSCHING, S., G. AUBURGER, R. HE~MANN, J. SCOTT, AND H. THOENEN. 1985. Levels of nerve growth factor end its mRNA in the central nervous system of the rat correlate with choline&c innervation. EMBO J. 4: 1389-1393. LARGE, T. H., S. C. BODARY, D. 0. &EGG, G. WESKAMP, U. @ITJ3N, AND L. F. REXCHARDT. 1988. Nerve growth factor gene expression in the developing rat brain. Science 234: 352-355.
13.
14.
15.
1989. Limbic seizures increase RNA for nerve growth factor.
LARKFORS, L., T. EBENDAL, S. R. WH~MORE, H. PERSSON, B. HOFFER, AND L. OLSON. 1987. Decreased level of nerve growth factor (NGF) and its messenger RNA in the aged rat brain.
Brain Res. 427: 55-60. 16.
LEIBROCK, J., F. L~~SPEICH, A. HOHN, M. HOFER, B. HANGERER, P. MASIAKOWSKI, H. THOENEN, AND Y.-A. BARDE. 1989. Molecular cloning and expression of brain-derived neurotrophic factor. Nature 341: 149-152.
17.
Lu, B., C. R. BUCK, C. F. DREYFUS, pression of NGF and NGF receptor brain: Evidence for local delivery Neurol. 104: 191-199.
18.
MAISONPIERRE, P. C., L. BELLUSCIO, B. FRIEDMAN, R. F. ANDERSON, S. J. WIEGAND, M. E. FORTH, R. M. LINDSAY, AND G. D. YANCOPOULOS. 1990. NT-3, BDNF, and NGF in the developing rat nervous system: Parallel as well as reciprocal pattern of expression. Neuron 5: 501-509.
19.
MARTINEZ, H. J., C. F. DREYFUS, G. MILLZR BLACK. 1985. Nerve growth factor promotes
in
86: 5636-
RAT
5.
REFERENCES 1.
IN
83
OF
AND I. B. BLACK. 1989. EXmRNAs in the developing and action of NGF. Exp.
JONAKAIT, choline@
AND I. B. develop-
84
20.
21.
22.
23.
24. 25.
26.
27.
28.
NISHIO,
AKIGUCHI,
ment in brain striatal cultures. Proc. Natl. Acad, Sci. USA 82: 1777-7781. ’ MEIER, R., M. B&~~-ANDRE, R. GOT-Z, R. HEIJMANN, A. SHAW, AND H. TI-IOENEN. 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. EMBO J. 5: 1489-1493. PALKOVITS, M., AND M. J. BROWNSTEIN. 1988. Maps and guide to microdissection of the rat brain. Elsevier, Amsterdam/New York. PIORO, E. P., AND A. C. CUEUO. 1988. Purkinje cells of adult rat cerebellum express nerve growth factor receptor immunoreactivity: Light microscopic observations. Brain Res. 455: 82-186. RAMCH, G., AND G. W. KREUTZBERG. 1987. The localization and distribution of high affinity b-nerve growth factor binding sites in the central nervous system of the adult rat, a light microscopic autoradiographic study using [iZI] b-nerve growth factor. Neuroscience 20: 23-36. RENNERT, P. D., AND G. HJIINRICH. 1986. Nerve growth factor mRNA in brain: Localization by in situ hybridization. Biochem. Biophys. Res. Commun. 138: 813-819. RICHARDSON, P. M., V. M. K. VERGE-ISSA, AND R. J. RIOPELLE. 1986. Distribution of neuronal receptors for nerve growth factor in the rat. J. Neurosci. 6: 2312-2321. SCHWAB, M. E., U. @-TEN, Y. AGID, AND H. THOENEN. 1979. Nerve growth factor (NGF) in the rat CNS: Absence of specific retrograde axonal transport and tyroeine hydroxylase induction in locus coeruleus and substantia nigra. Brain Res. 188: 473483. SEILER, M., AND M. E. SCHWAB. 1984. Specific retrograde transport of nerve growth factor (NGF) from neocortex to nucleus basalis in the rat. Brain Res. 300: 33-39. SHELTON, D. L., AND L. F. REICHARDT. 1986. Studies on the expression of the nerve growth factor (NGF) gene in the central
AND FURUKAWA
29.
30.
31.
32.
33. 34.
35.
nervous system: Level and regional distribution of NGFmRNA suggest that NGF functions as a trophic factor for several distinct populations of neurons. Proc. Natl. Ad. Sci. USA 83: 2714-2718. SPILLANTINI, M. G., E. ALOE, E. ALLEVA, R. DE SIMONJZ, M. GOEDERT, AND R. LEVI-M• NTALCINI. 1989. Nerve growth factor mRNA and protein increase in hypothalamus in a mouse model of aggression. Proc. Natl. Ad. Sci. USA 86: 6555-6559. THEONEN, H., C. BANDTLOW, AND R. HEUMANN. 1987. Thephysiological function of nerve growth factor in the central nervous system: Comparison with the periphery. Rev. Physid. Biochem. Phurmacol. 109: 146-178. WESKAMP, G., AND U. OTTEN. 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. WHITTEMORE, S. R., T. EBENDAL, L. L-Rs, L. OLSON, A. SEIGER, I. STROMBERG, AND H. PER~SON. 1986. Developmental and regional expression of @nerve growth factor messenger RNA and protein in the rat central nervous system. Proc. Natl. Ad. Sci. USA 83: 817-821. WHITTEMORE, S. R., AND A. SEIGER. 1987. The expression, localization and functional significance of p-nerve growth factor in the central nervous system. Brain Res. Rev. 12: 439-464. WHITTEMORE, S. R., P. L. FFUEDMAN, D. LARHAM MAR, H. PF,RSSON, M. GONZALEZ-CARVAJAL, AM) V. R. HOLET~. 1988. Rat 8nerve growth factor sequence and site of synthesis in the adult hippocampus. J. Neurosci. Res. 20: 403-410. YAN, Q., AND E. M. JOHNSON. 1988. An immunohistochemical study of the nerve growth factor receptor in developing rats. J. Neurosci. 8: 3481-3498.
36. ZAFFU,
F., B. HENGERER,
J. LEIBROCK,
H. THOJZNJN,
AND D.
1990. Activity dependent regulation of BDNF and NGF mRNAs in the rat hippocampus is mediated by nonNMDA glutamate receptors. EMBO J. 9: 3645-3560. LINDHOLM.
