EXPERIMENTALNEUROLOGY 105,221-232 (1989)

Loss of Nerve Growth Factor Receptor-Containing Neurons

in Alzheimer’s Disease: A Quantitative Analysis across
Subregions of the Basal Forebrain
ELLIOTT J. MUFSON,* MARK BOTHWELL,? AND JEFFREY H. KORDOWER$
*Christopher Center for Parkinson’s Research, Institute for Biogerontology Research, Sun City, Arizona 85351; TDepartment of Physiology and
Biophysics, University of Washington, Seattle, Washington 98195; and $Department of Anatomy and Cell Biology,
University of Illinois School of Medicine, Chicago, Illinois 60612

INTRODUCTION
Magnocellular neurons comprising the Chl-Ch4 re-
gions of the basal forebrain provide topographic cholin- Over the last decade, the continuum of magnocellular
ergic innervation to the cerebral cortex, thalamus, and cholinergic neurons located within the basal telencepha-
basolateral nucleus of the amygdala. Most quantitative lon has been the focus of numerous investigations with
studies analyzing the status of these neurons in Alzhei- regard to its role in normal memory and dementia (see
mer’s disease (AD) have employed Nissl-stained prepa- reviews (4, 7, 26)). On the basis of acetylcholinesterase
rations. These studies principally analyzed large neu- (AChE) and choline acetyltransferase (ChAT) histo-
rons of a prespecified cell diameter. Since basal fore- chemical markers, these basal forebrain neurons have
brain neurons atrophy in Alzheimer’s disease, an been subdivided into four major cholinergic cell groups
immunocytochemical marker for these neurons would designated Chl-Ch4 (32). The ovoid cholinergic neu-
appear to be a better alternative for determining rons within the medial septum (Chl) and the round hy-
whether there is regionally specific degeneration of
perchromic neurons within the vertical limb of the diag-
cholinergic neurons across subregions of the basal fore-
brain. Brain sections from seven AD and five aged- onal band (Ch2) project mainly to the hippocampus. The
matched control patients were immunocytochemically fusiform neurons comprising the horizontal limb of the
stained with a monoclonal antibody raised against the diagonal band (Ch3) provide cholinergic input to the ol-
receptor for nerve growth factor (NGF), a probe which factory bulb and the large multipolar neurons of the nu-
has previously been demonstrated to extensively and cleus basalis of Meynert (Ch4) provide a source of topo-
exclusively colocalize with cholinergic basal forebrain graphically organized cholinergic innervation to the en-
neurons in humans (17, 25, 35). NGF receptor-immu- tire cortical mantle, to select diencephalic sites, and to
noreactive neurons within the hippocampal projecting the basolateral nucleus of the amygdala (21, 23, 29, 32).
nuclei of the medial septum (Chl) and vertical limb of The Ch4 region has further been subdivided into ante-
the diagonal band (Ch2) were minimally affected in AD romedial, anterolateral, intermediate, and posterior sub-
as compared to control cases. In contrast, the Ch4 re- sectors on the basis of connectivity and cytoarchitecture
gion demonstrated a significant loss of NGF receptor- (32). Recently using ChAT and nerve growth factor
immunoreactive neurons in AD that inversely corre- (NGF) receptor immunohistochemistry (35, 37), we
lated (-0.786) with the duration of the disease process. have shown that all subsectors of the Ch nomenclature
All four subregions of Ch4 were affected in the AD cases
originally applied to the basal forebrain of the monkey
with the anterolateral (76.4%), intermediate (62.1%)
and posterior divisions (76.5%) demonstrating the (32) have analogous nuclei in the human brain.
Impaired telencephalic cholinergic neurotransmission
greatest reduction in NGF receptor-immunoreactive
neurons. Nissl-counterstained sections failed to reveal in Alzheimer’s disease (AD) (2-4, 7, 8, 26,40, 42, 53, 55,
magnocellular neurons which were not immunoreac- 57) and other disorders in which dementia can be a clini-
tive for the NGF receptor, suggesting that reductions in cal feature (1, 2, 50, 50, 53, 54) has been well docu-
immunocytochemically stained neurons reflects neuron mented. It has been suggested that pathological alter-
loss and not the failure of viable neurons to synthesize ations within cholinergic basal forebrain neurons may,
NGF receptors. These data indicate that cholinergic to some extent, underlie cognitive dysfunction. Some
basal forebrain neurons which project to the amygdala, studies have reported extensive (>50%) neuron loss or
as well as to the temporal, frontobasal, and frontodorsal shrinkage (1, 3, 8, 39, 42, 53, 55, 56) of these large cells.
cortices, are most affected in AD. u 1989 Academic press, I~~. In contrast, others have reported limited or no basal
forebrain neuronal loss accompanying a neurochemical

221 0014.4886/89 $3.00

Copyright 0 1989 by Academic Press, Inc.
All rights of reproduction in any form reserved.
222 MUFSON, BOTHWELL, AND KORDOWER

deficit in ChAT (39,40,57). The majority of these mor- years) and seven AD patients (mean age, 77 years; range,
phological studies have analyzed Nissl-stained material, 71-86 years) were removed at autopsy (postmortem de-
using an a priori criterion of neuron size as a working lay: normal cases, mean 4 h; range, 3-5 h; AD cases,
definition for basal forebrain neurons. However, since mean, 5 h; range, 3-11 hours (see Table 1)) as part of the
atrophy of basal forebrain cells occurs with the degener- brain donation program at the Institute for Biogerontol-
ative process (24, 39, 42), many shrunken, but viable, ogy Research (Sun City, AZ). After removing a brain
basal forebrain neurons would be counted as lost using from the calvaria, it was cut into l-cm coronal slabs us-
a perikaryal-size cell counting procedure. Detailed quan- ing a calibrated brain slice apparatus and each slab was
titative analyses of the status of cholinergic basal fore- hemisected. The right hemisphere was placed in a solu-
brain neurons using histological markers for the cholin- tion of 4% paraformaldehyde (Fischer) in 0.1 M phos-
ergic degredating chemical AChE or the synthesizing phate buffer (pH 7.4) for 24 h at 4°C and cryoprotected
enzyme ChAT have been difficult. The heavy neuropil in graded sucrose concentrations of lo-40% in phos-
staining which results from procedures visualizing phate buffer (pH 7.4) prior to sectioning. Each brain was
AChE severely compromises the accuracy of neuron then cut into several series of adjacent 40-pm-thick sec-
counting. Moreover, immunohistochemical localization tions on a freezing microtome and stored in a cryopro-
of ChAT is capricious in the human brain (34,35). The tectant solution (52) until processed. Sections through-
recent characterization of an antibody directed against out the entire basal forebrain extending from the fusion
the receptor for NGF has led to the demonstration that of the subcallosal gyrus to the level of the lateral genicu-
NGF receptor-containing neurons extensively codistri- late nucleus were examined. Asymmetries between
bute and/or colocalize with ChAT-containing neurons hemispheres in the basal forebrain do not appear to oc-
within the monkey (22) and human (17,25,35,37) basal cur in the normal or AD brain (3,8) and thus data gener-
forebrain. Indeed, in rats (48), monkeys (22,45), and hu- ated from one hemisphere should reflect that of both
mans (17,25,35,37), the basal forebrain is virtually the sides of the brain. The clinical diagnosis of AD was con-
only region throughout the neuraxis which contains so- firmed with Thioflavin-S and Bielschowsky silver stain-
mata immunoreactive for the NGF receptor. The reli- ing procedures (19,60).
ability of the NGF receptor antibody as a consistent
marker for cholinergic basal forebrain neurons in the NGF Receptor Immunohistochemistry
human brain makes its use an optimal approach for as-
sessing, in a quantitative fashion, the degree of neuronal Sections were processed for the localization of NGF
loss in this region in AD. This is especially true given receptor-containing elements with the labeled antibody
our recent findings that, in AD, even the most atrophied procedure of Hsu and co-workers (18) according to the
and dystrophic basal forebrain neurons still express protocol of Kordower and co-workers (22,24,35). After
NGF receptors (24). extensively washing the cryoprotectant from the tissue,
Only a few studies have assessed whether there is endogenous peroxidase containing elements were re-
differential degeneration of cholinergic neurons across moved with a 20-min incubation in 0.1 M sodium per-
subregions of the basal forebrain and those, in some re- iodate in phosphate-buffered saline (pH 7.4; PBS). After
spects, have failed to reach a consensus (3, 8, 53). The three lo-min washes in a dilution media solution con-
present study uses a monoclonal antibody raised against sisting of 0.1 M Tris-buffered saline (pH 7.2; TBS) and
the NGF receptor for a quantitative assessment of neu- 0.05% Triton-X, background staining was blocked with
ron loss within Chl, Ch2, and Ch4 subfields of the basal a l-h incubation in dilution media containing 3% normal
forebrain. These data have renewed importance given horse serum and 2% bovine serum albumen (BSA). Tis-
the topographic projections from these regions coupled sueswere then incubated for 48 h (24 h at room tempera-
with the potential for novel therapeutic strategies such ture/24 h at 4°C) in the monoclonal NGF receptor pri-
as the administration NGF (10,13,28,58,59), the intra- mary antibody (NGFR5; 1:lOOO). The NGFR5 antibody
cerebral implantation of NGF-containing tissues (27, was diluted in dilution media, 1% normal horse serum,
43,47), or intracerebral transplantation of fetal cholin- and 1% BSA. This solvent alone served to process con-
ergic primordia (e.g. (9, 14, 30)) to provide focal circum- trol incubated sections. After further washes in dilution
scribed effects within the adult mammalian central ner- media, the tissue was incubated in the biotinylated horse
vous system (CNS). antimouse secondary antibody (1:200; Vector labs) for
60 min. Following further dilution media washes, the
MATERIALS AND METHODS sections were incubated in the avidin-biotin complex
(Vector labs) at a concentration of 1:200. The sections
Preparation of Tissues were then washed in TBS. One series was reacted in a 0.1
Brains from five neurologically normal and pathologi- M phosphate-buffered saline solution containing 0.05%
cally intact patients (mean age, 74 years; range, 68-78 3,3’-diaminobenzidine (DAB) and 0.005% HzOp result-
 NGF RECEPTORS IN AD 223

TABLE 1
Clinical and Demographic Characteristics of Patient Cases

Age Brain wt. Autolysis Disease duration

Case (year) Sex (EC) Diagnosis (h) (year)

1 82 M 1400 AD 8.0
2 74 F 970 AD 4.0 2
3 73 M 1130 AD 4.5 12
4 71 F 1010 AD 4.0 10
5 72 M 1100 AD 3.5 4
6 86 F 1050 AD 11.0 10
7 83 F 840 AD 5.0 3
8 78 F 1040 CON 3.0 NA
9 77 M 1170 CON 3.5 NA
10 70 M 1250 CON 4.5 NA
11 68 M 1510 CON 4.0 NA
12 77 F 1020 CON 3.25 NA

Note. AD, Alzheimer’s disease; CON, control; NA, not applicable.

ing in a specific brown reaction product. A second series
was completed in a modification of the nickel enhance-
ment procedure of Hancock (15) which produces a black
reaction product. Sections processed in this manner
were rinsed in a buffer containing 0.2 M imidazole-1.0
M sodium acetate (pH 7.2). This buffer then served as
the solvent for the chromagen solution containing 2.5%
nickel II sulfate, 0.05% DAB, and 0.005% H202 (final
pH, 7.2). The reaction was terminated in rinses with the
imidazole-acetate buffer. Sections were mounted on gel-
atin-coated slides, air-dried, dehydrated, and cover-
slipped with Permount. Additional sections were
mounted, air-dried, defatted with a 1:l mixture of chlo-
roform/99% alcohol for 24 h, and then counterstained
for Nissl substance with cresyl violet acetate (l%, pH
3.3). Others were counterstained with Thioflavin-S and
coverslipped with Apathy’s medium for the visualization
of amyloid-bearing plaques and tangles in the area of
NGF receptor-immunoreactive neurons (60). Both
bright- and dark-field microscopy were used to evaluate
the distribution of NGF receptor-containing profiles,
whereas epifluorescence aided in the analysis of Thi-
oflavin-S-stained material.
The NGFR5 antibody is the product of a hybridoma
generated from a mouse immunized with homogeneously
pure NGF receptor and identified according to its ability
to specifically immunoprecipitate NGF receptors (31).
The specificity of NGFR5 appears to be identical to that
of human NGF receptor monoclonal antibody ME20.4
(44) except that ME20.4 blocks NGF binding to NGF
receptors while NGFR5 does not. In unfixed frozen sec-
tions, immunocytochemical analyses of a variety of tis-
sues show that the ME20.4 and NGFR5 antibodies pro-
duce identical results. The NGFR5 antibody is generally
employed since it maintains reactivity with the NGF re-
ceptor after fixation of tissues, whereas the ME20.4 anti-
body does not. In addition to the concordance of immu-
nostaining patterns of ME20.4 and NGFR5, the follow-
ing lines of evidence support the specificity of NGFR5
for the NGF receptor. First, the NGFR5 immunoprecip-
itates homogeneously pure NGF receptors (31). Second,
NGFR5 specifically labels a protein comigrating with
purified NGF receptor on Western blots of protein from
human melanoma cell lines expressing the NGF recep-
tor and does not detect any protein in cell lines that do
not express the NGF receptor (31). Third, NGFR5 does
not react with a variety of fibroblast cell lines that do
not express the NGF receptor but does react with these
cells (in immunoperoxidase and immunorosetting as-
says) following transfection of these cells with the
cloned human NGF receptor gene (20). Last, in every
tissue examined, including the basal forebrain, the level
of NGF receptor immunostaining correlates with the
level of NGF receptor mRNA.
Morphometric Analysis
The basal forebrain was divided into subfields as ini-
tially classified by Mesulam and co-workers in monkeys
(32) and subsequently adapted to the human brain (3,
35). Counts of NGF receptor-containing basal forebrain
neurons were performed on right hemisphere sections
using a microplotter 4000 (DiLog Instruments, Talla-
hassee, FL) cell counter interfaced with a Hewlett-
Packard X-Y plotter connected to the stage of a Nikon
microscope. Counts were carried out at a total magnifi-
cation of 200X. The total number of neurons was esti-
mated by counting the numbers of NGF receptor-immu-
noreactive neurons in every third section and interpolat-
ing these data over the rostrocaudal extent of the cell
population. Only neurons with a clear immunoreactive
cell soma were counted. In addition, selective loss of
224 MUFSON, BOTHWELL, AND KORDOWER
 NGF RECEPTORS IN AD 225

NGF receptor-containing neurons within each subdivi- tentially represents differences in amyloid accumulation
sion of the basal forebrain was determined by comparing between these two cell types.
differences in the total number of immunoreactive neu- NGF receptor immunohistochemistry clearly delin-
rons counted over several sections through each Ch sub- eated the Chl-Ch4 regions within the basal forebrain.
region for all AD and control cases. Each section ana- While the staining pattern was dense in some regions,
lyzed was treated as an individual data point. Total loss especially in the control cases, individual neurons were
of NGF receptor-containing Ch4 neurons in AD brains easily discerned. Qualitative analyses revealed substan-
as a function of disease duration was expressed as a per- tial reductions in NGF receptor-immunoreactive neu-
centage of control values. The Ch3 subfield was excluded rons throughout much of the Ch4 region in the AD cases.
from the present analysis due to the limited number of Although there was variability between cases with re-
NGF receptor-containing neurons found in the human gard to the Ch subfield most affected, loss of NGF re-
horizontal limb of the diagonal band. Statistical compar- ceptor-immunoreactive neurons was most apparent in
isons were determined using the Student’s t test. the anterolateral (Ch4al; Fig. 2), the intermediate
(Ch4i), and the posterior (Ch4p) subdivisions. In con-
RESULTS trast, the most rostra1 portion of the basal forebrain
(Chl-Ch2) appeared less affected in AD with regard to
the number of NGF receptor-immunoreactive neurons
Thioflavin-S- and Bielschowsky-stained sections con-
(Fig. 3), although some of the neurons in this region ap-
firmed the clinical diagnosis of AD in all seven cases.
peared atrophied and dystrophic like their Ch4 counter-
Numerous neuritic plaques and neurofibrillary tangles
parts. Moreover, a concomitant dimunition in NGF re-
were seen within the amygdala, hippocampal complex,
ceptor fiber staining was observed in the AD cases
temporal cortex, and nucleus basalis of AD cases (Fig.
within the external capsule, the major neocortical effer-
1). Similar profiles were only occasionally observed in
ent fiber pathway of Ch4 neurons (Fig. 4). In both AD
these regions in control patients. NGF receptor immu-
and control cases, sections counterstained for Nissl sub-
nohistochemically stained sections counterstained for
stance revealed a virtual 1:l correspondence between
Thioflavin-S were used to determine whether NGF re-
NGF receptor-immunoreactive somata and basophillic
ceptor-containing neurons showed evidence of neurofi-
magnocellular basal forebrain neurons.
brillary degeneration. Fluorescence microscopic analy-
sis revealed NGF receptor-immunoreactive neurons in- The control cases revealed an average of 190,000 NGF
vested with neurofibrillary material in a manner receptor-immunoreactive neurons within the right basal
suggestive of a continuum of degenerative events. For forebrain. The quantitative assessments which are sum-
example, some neurons were Thioffavin-S negative or marized in Table 2 confirmed the qualitative observa-
only minimally fluorescent for this amyloid marker. tions seen in AD cases relative to normal controls. The
Others were more heavily invested suggesting a further number of NGF receptor-immunoreactive neurons
involvement in the disease process (Figs. 1D and 1E). within the hippocampal projecting Chl-Ch2 were com-
In addition, many neurofibrillary tangle-bearing profiles parable in Alzheimer’s disease cases and age-matched
observed in the basal forebrain were ghosts; that is, neu- controls (t(l1) = 3.81; P > 0.70). In contrast, significant
rons appeared to be largely devoid of cytoplasm and or- reductions in the number of NGF receptor-immunoreac-
ganelles and seemed to consist mainly of neurofibrillary tive neurons were observed in Ch4am (t(17) = 3.23; P
remnants suggestive of end stage degeneration (Fig. 1D < 0.005), Ch4al (t(15) = 6.642; P < O.OOOl), Ch4i (t(17)
(39)). Virtually no neuritic plaque formation was ob- = 4.295; P < 0.005), and Ch4p (t(7) = 4.57; P < 0.0025).
served within the basal forebrain. It is interesting to note The total number of NGF receptor-immunoreactive
that the neurofibrillary tangles observed within the mag- neurons throughout the basal forebrain in the Alzhei-
nocellular basal forebrain were ovoid in shape and iiuo- mer’s cases inversely covaried with the length of the dis-
rested as a dense core mass (Fig. 1C). This is in contrast ease process. As illustrated in Fig. 5, there was a correla-
to the more wispy appearance of neurofibrillary tangles tion of -0.786 between the number of NGF receptor-im-
observed within the entorhinal cortex (Fig. 1A) and po- munoreactive neurons and disease duration.

FIG. 1. Thioflavin-S-stained sections showing neurofibrillary tangles (NFTs) and neuritic plaques in AD patients. (A) NFTs within l;vpr
II of the entorhinal cortex (Case 5). (B) Neuritic plaques located within the superior temporal cortex (Case 1). (C!) NFTs located within +hr
nucleus basalis (Ch4am, Case 6). (D) NGF receptor-immunoreactive nucleus basalis (Ch4al) neurons (DAB chromagen) counterstainrd with
Thioflavin-S (Case 6). SoIid white arrows indicate NGF receptor containing neurons, open arrows indicate ghost tangles (see text), and curved
arrow denotes tangle bearing NGF receptor-containing neuron. (E) Another example of a tangle bearing NGF receptor-containing rwwor~.
Note the dark rim of NGF receptor immunoreactivity surrounding the fluorescent NFT staining. (A-E, 380X; Ch4am, anteromedial division of
the cholinergic cell group 4; Ch4a1, anterolateral division of the cholinergic cell group 4.)
226 MUFSON, BOTHWELL, AND KORDOWER
 NGF RECEPTORS
DISCUSSION
The present study demonstrates that there is a sig-
nificant loss of NGF receptor-containing neurons
throughout all subregions of the nucleus basalis (Ch4) in
AD. In contrast, AD cases did not reveal a significant
loss of these neurons within the hippocampal projecting
Chl-Ch2 regions. Nissl-counterstained material failed
to reveal magnocellular basophilic neurons within the
basal forebrain that were not immunoreactive for the
NGF receptor beyond the small percentage of such cells
that are typically observed in the normal aged human
brain (35). We have previously demonstrated that virtu-
ally all healthy and atrophied basal forebrain cholinergic
neurons continue to express NGF receptors in AD and
remain immunoreactive for cholinergic enzymatic
markers (24). These data, taken together, strongly sug-
gests that the absence of NGF receptor-immunoreactive
somata in AD reflects neuron death, and not the failure
of viable basal forebrain neurons to synthesize and ex-
press NGF receptors to within detectable limits.
Previous studies using Nissl-stained material have
generated mixed results regarding the extent of neurons
lost within the Ch4 region in AD. The initial findings of
Whitehouse and co-workers (53, 55, 56), which have
been corroborated by some (e.g. (2,3,8, 37, 57)) but not
all (39,40) investigations, have been generally supported
by the present immunocytochemical data. Prior analy-
sesacross the different subregions of the basal forebrain
have produced inconsistent results. Doucette and Ball
(8) have reported that the more posterior, temporal lobe
projecting, basal forebrain neurons were most affected
in AD, while the anterior portions of the basal forebrain
were relatively unaffected. Whitehouse reported severe
(>50%) neuron loss throughout each subregion of the
basal forebrain, including the Chl-Ch2 subregions (53).
Arendt and co-workers (3) using Nissl-stained material
report extensive loss (62.5%) of putative cholinergic
neurons within the Chl-Ch2 subfields with 45-64% re-
ductions seen in the four Ch4 subfields. The present im-
munocytochemical data revealed a pattern of results
different from those previously described using Nissl-
stained material. For example, the number of NGF-re-
ceptor immunoreactive neurons within the Chl-Ch2
subfields were comparable between Alzheimer’s and
aged-matched control cases, although some neurons
within this region appeared atrophic. In the present AD
cases, the anterolateral (76.4%), intermediate (62.1%),
and posterior (76.5%) subdivisions of the Ch4 region dis-
played the most severe loss of NGF receptor-immunore-
FIG. 2. Comparison of the loss of NGF receptor-immunoreactive
tally normal patient (Cases 9). (C, D) AD patient (Case 3). At the higher magnifications, the dystrophic appearance
decrease in NGF receptor-immunoreactive neuropil can be observed.
IN AD 227
active neurons. Multiple factors may account for the
differences between the previous (3, 8, 53) and present
data. First, studies examining Nissl-stained material
used perikaryal size as a criterion for inclusion into the
basal forebrain. Significant atrophy of neurons occurs in
many brain regions in aging and Alzheimer’s disease (see
review (11)) including the cholinergic basal forebrain
(24,39,42). Indeed, perikaryal shrinkage is probably an
early pathological event in the process of degeneration.
If a neuron atrophies below the prespecified criteria then
it is counted as lost. Both healthy and dystrophic cholin-
ergic basal forebrain neurons still express NGF recep-
tors in Alzheimer’s disease (24). By using immunocyto-
chemical procedures, both large and atrophied choliner-
gic neurons within the basal forebrain are counted and
can therefore generate a more accurate reflection of sta-
tus of basal forebrain neurons in diseases of cognitive
dysfunction such as AD. A second source of variability
is the AD cases themselves. AD demonstrates heteroge-
neous pathological events upon postmortem analyses
with gender, severity of the disease process, age of dis-
ease onset, and other factors (e.g, agonal state) playing
a role in the final neuropathological state (see reviews
(4,7,26)). In this regard, the present study demonstrates
a high inverse correlation between disease duration and
the number of NGF receptor-immunoreactive neurons
within the basal forebrain. Thus differences between
studies most likely reflect variability in antemortem fea-
tures across patient populations as well.
The anterolateral and intermediate subdivisions of
Ch4 demonstrate severe pathology within the basal fore-
brain in our AD cases. In previous independent investi-
gations carried out in nonhuman primates utilizing ret-
rograde tract-tracing techniques, we have demonstrated
that, in addition to frontobasal and frontodorsal cortical
projections, these subregions of Ch4 project preferen-
tially to the basolateral nucleus of the amygdala (23,32).
This topographic input has recently been confirmed in
monkeys with anterograde tracing procedures (21). The
cholinergic input to the amygdala has been generally ne-
glected when assessing the role acetylcholine plays in
the normal and pathological processing of cognitive in-
formation. This is true in spite of the fact the cholinergic
deficit observed in the amygdala is at least equal to that
seen in cortex with respect to both magnitude and pace
of onset (38). The amygdala has been implicated in the
processing of cognitive information in primates (33)) al-
though it is still unclear whether it is fundamentally in-
volved in primary memory function ((33) but see (46))
or performs a more integrative role related to the reward
neurons within the Ch4al region of the nucleus basalis. (A, B) Neurologi-
of Ch4al neurons and the
(A and C, 19X; B and D, 171X; AC, anterior commissure).
228 MUFSON, BOTHWELL, AND KORDOWER

FIG. 3. Comparison of NGF receptor-immunoreactive neurons within the Ch2 region of a neurologically normal (A, Case 8) and an AD
(B, Case 5) patient. Note that as compared to the extensive loss of NGF receptor containing neurons in Ch4al (see Fig. Z), the Ch2 region is
relatively intact in AD. C and D are higher magnifications of NGF receptor-immunoreactive neurons and fibers in the Ch2 of a normal (C) and
an AD (D) case, respectively (180X). Note that while present, many of the Ch2 neurons in the AD case appear dystrophic. (Ch2, cholinergic
cell group 2).

of the memory process (see (12) for discussion). It is in- that the low levels of NGF observed in this region are at
teresting to note that the amygdala contains relatively least partially responsible for the potentiated neurode-
low levels NGF as compared to those of either the hippo- generation observed within the amygdaloid-projecting
campus or the cerebral cortex. It is tempting to speculate regions of the cholinergic basal forebrain.
 NGF RECEPTORS IN AD 229

FIG. 4. Dark-field photomicrograph of NGF receptor-containing fibers coursing within the external capsule of a neurologically normal (A,
Case 8, 4X) and an AD (B, Case 5, 11X) patient. Note the striking loss of NGF receptor-containing fibers between the control and the AD
patient. B is presented at a magnification higher than that of A to better visualize the few remaining fibers seen in the AD case.
230 MUFSON, BOTHWELL, AND KORDOWER

TABLE 2
Selective Loss of NGF Receptor-Containing Neurons within the Subfields of the Basal Forebrain

Group Chl-2 Ch4am Ch4al Ch4i Ch4p

Controls (5) 1366 + 257’ 2425k265 3979 k 496 1713+237 204Ok232

AD(7) 1138+254 1573+180 945f153 650f 92 48Ok120

Reduction (%) 16.7 (NS) 35.1* 76.4** 62.1** 76.5*

Note. AD, Alzheimer’s disease; NS, not significantly different between groups.
’ Mean and SEM.
*P< 0.005.
** P < 0.001.

In addition to the amygdaloid projecting regions, ex- process. Indeed the present data suggest that efforts to
tensive pathology was also observed in the posterior sub- rescue and restore telencephalic cholinergic neurotrans-
division of Ch4 which provides cholinergic innervation mission with growth factors might best be focused upon
to the temporal cortex. These findings, coupled with the the hippocampal projecting Chl-Ch2 neurons which are
extensive characteristic Alzheimer’s degeneration seen atrophied, but present, in these end-stage Alzheimer’s
within the amygdala (5,6), hippocampal-entorhinal COF- cases. In contrast, neuron replacement strategies might
tical complex (17), and temporal cortex (51), serve to un- best be focused within the target areas of the anterolat-
derscore the severity of temporal lobe pathology in AD. eral, intermediate, and posterior Ch4 neurons since
In conclusion, the present study demonstrates that these cells are lost in the greatest number in AD.
there is a regionally specific loss of NGF receptor-con-
taining neurons within the basal forebrain in AD. In the ACKNOWLEDGMENTS
AD cases presently analyzed, neurons projecting to the
basolateral amygdala, as well as to the frontobasal, fron- This research was supported by NS 25-65501 Al, NS26146-OlAl,
todorsal, and temporal cortices, were most affected. In The American Health Assistance Foundation, Arizona Disease Con-
contrast, hippocampal projecting basal forebrain neu- trol Commission Grant 8277-OOOOOO-Ol-OYK-8365, and a fund from
rons were affected to a lesser degree. While the presence P. Riely. We thank L. Sue and G. Binskin for histologic assistance, S.
Styren for photographic assistance, and Dr. H. Civin for neuropatho-
or absence of neurons should not be equated with the
logic examination.
functional efficiency of neural systems, experimental
strategies aimed to rescue basal forebrain neurons (10,
13,28,45,47,58,59) or provide circumscribed neuronal REFERENCES
(e.g. (9, 14, 30)) or neurotransmitter (16) replacement
might focus their attention at these basal forebrain tar- 1. ARENDT, T., V. BIGL, AND A. ARENDT. 1984. Neuron loss in the
get regions which appear severely affected by the disease nucleus basalis of Meynert in Creutzfeld-Jakob disease. Acta
Neuropathol. 65: 85-88.
2. ARENDT, T., V. BIGL, A. ARENDT, AND A. TENNSTEDT. 1983.
Loss of neurons in the nucleus basalis of Meynert in Alzheimer’s
disease, paralysis agitans, and Korsokoff s disease. Acta Neuro-
pathol. 61: 101-108.
100 3. ARENDT, T., V. BIGL, A. TENNSTEDT, AND A. ARENDT. 1985.
NEURON # Neuronal loss in different parts of the nucleus basalis is related to
(lo31 80 neuritic plaque formation in cortical target areas in Alzheimer’s
disease. Neuroscience 14: 1-14.
60 4. BARTUS, R. T., R. L. DEAN, B. BEER, AND A. S. LIPPA. 1982. The
cholinergic hypothesis of geriatric memory dysfunction. Science
217: 408-417.
5. BRADY, D. R., AND E. J. MUFSON. 1989. Amygdaloid pathology
in Alzheimer’s disease. Anat. Rec. 223: 17A.
6. BRASHEAR, H. R., M. S. GODEC, AND J. CARLSEN. 1988. The dis-
tribution of neuritic plaques and acetylcholinesterase staining in
1 2 3 4 10 10 12 the amygdala in Alzheimer’s disease. Neurology 38: 1694-1699.
DISEASE DURATION (Yrs.)
7. COYLE, J. T., D. L. PRICE, AND M. R. DE LONG. 1983. Alzheimer’s
FIG. 5. The numbers of remaining neurons within the Ch4 region disease: A disorder of cortical cholinergic denervation. Science
inversely correlate with the disease duration. 219: 1184-1190.
 NGF RECEPTORS IN AD 231

8. DOUCETTE, R., AND M. BALL. 1987. Left-right symmetry of neu- localization with cholinergic enzymes. Sot. Neurosci. Abstr. 14:
ronal cell counts in the nucleus basalis of control and Alzheimer’s 1222.
disease brain. Brain Res. 422: 357-360. 26. KORDOWER, J. H., AND D. M. GASH. 1986. Animals and experi-
9. DUNNETT, S. B., W. C. Low, S. D. IVERSON, U. STENEVI, AND A. mentation: An evaluation of animal models of Alzheimer’s and
BJ~RKLUND. 1982. Septal transplants restore maze learning in Parkinson’s disease. Zntegr. Psych. 4: 64-80.
rats with fimbria-fornix lesions. Brain Res. 225: 335-348. 27. KORDOWER, J. H., M. S. FIANDACA, M. F. D. NOTTER, J. T. HAN-
10. FISCHER, W., K. WICTORIN, A. BJ~~RKLUND, L. R. WILLIAMS, S. SEN, AND D. M. GASH. 1989. Scientific basis of dopaminergic
VARON, AND F. H. GAGE. 1987. Amelioration of cholinergic neu- brain grafting. In W. C. KOLLER AND G. PAULSON, Eds., Thera-
ron atrophy and spatial memory impairment in aged rats by peutic Approaches to Parkinson’s Disease. Decker, New York, in
nerve growth factor. Nature (London) 329: 65-68. press.
11. FLOOD, D. G., AND P. C. COLEMAN. 1988. Neuron number and 28. KROMER, L. F. 1987. Nerve growth factor treatment after brain
sizes in aging brain: Comparison of human, monkey, and rodent injury prevents neuronal death. Science 235: 214-216.
data. Neurobiol. Aging 9: 453-463. 29. LEVEY, A. I., A. E. HALLENGER, AND B. H. WAINER. 1987. Cho-
12. FRIEDMAN, H. R., AND P. S. GOLDMAN-RAKIC. 1988. Activation linergic nucleus basalis neurons may influence the cortex via the
of the hippocampus by working memory: A 2-deoxy-D-glucose thalamus. Neurosci. Lett. 74: 7-13.
study on behaving Rhesus monkeys. J. Neurosci. 8: 4693-4706.
30. Low, W. C., P. R. LEWIS, S. T. BUNCH, S. B. DUNNETT, S. R.
13. GAGE, F. H., D. M. ARMSTRONG, L. R. WILLIAMS, AND S. VARON. THOMAS, S. B. IVERSON, AND A. BJ~RKLUND. 1982. Functional
1988. Morphological response of axotomized septal neurons to recovery following neural transplantation of embryonic septal
nerve growth factor. J. Comp. Neurol. 269: 147-155. nuclei in adult rats with septohippocampal lesions. Nature (Lon-
14. GAGE, F. H., A. BJ~RKLUND, U. STENEVI, S. B. DUNNETT, AND don) 300: 260-262.
P. T. KELLY. 1984. Intrahippocampal septal grafts ameliorate 31. MARANO, N., B. DIETZSCHOLD, J. J. EARLRY, JR., G. SCHATTE-
learning impairments in aged rats. Science 225: 533-536. MAN, S. THOMPSON, P. GROB, A. H. Ross, M. BOTHWELL, B. F.
15. HANCOCK, J. C. 1982. A serotonin-like immunoreactive fiber sys- ATKINSON, AND H. KOPROWSKI. 1987. Purification and amino
tem in the dorsal columns of the spinal cord. Neurosci. Lett. 3 1: acid terminal sequence of human melanoma nerve growth factor
247-252. receptor. J. Neurochem. 48: 225-232.
16. HARBAUCH, R. E., D. W. ROBERTS, D. W. COOMBS, R. L. SAUN- 32. MESULAM, M.-M., E. J. MUFSON, A. I. LEVEY, AND B. J. WAINER.
DERS, AND T. M. REEDER. 1984. Preliminary report: Intracranial 1983. Cholinergic innervation of cortex by basal forebrain: Cyto-
cholinergic drug infusion in patients with Alzheimer’s disease. chemistry and cortical connections for the septal area, diagonal
Neurosurgery 15: 515-518. band nuclei, nucleus basalis (substantia innominata) and hypo-
HEFTI, F., J. HARTIKKA, P. SALVIATERRA, W. J. WEINER, AND thalamus in the rhesus monkey. J. Comp. Neurol. 214: 170-197.
17.
D. MASH. 1986. Localization of nerve growth factor receptors on 33. MISHKIN, M. 1978. Memory in monkeys is severely impaired by
cholinergic neurons of the human basal forebrain. Neurosci. Lett. combined, but not separate removal of the amygdala and hippo-
69: 275-281. campus. Nature (London) 273: 297-298.
18. Hsu, S. M., AND L. RAINE, AND H. FANGER. 1981. Use of avidin- 34. MUFSON, E. J., D. C. MASH, AND L. B. HERSH. 1988. Neurofibril-
biotin-peroxidase complex (ABC) in immunoperoxidase tech- lary tangles in cholinergic pedunculopontine neurons in Alzhei-
niques: A comparison between ABC and unlabeled antibody mer’s disease. Ann. Neural. 24: 623-629.
(PAP) procedures. J. Histochem. Cytochem. 29: 577-580. 35. MUFSON, E. J., M. A. BOTHWELL, L. B. HERSH, AND J. H. KOR-
19. HYMAN, B. T., G. W. VAN HOESEN, A. R. DAMASIO, AND C. L. DOWER. 1989. Nerve growth factor receptor immunoreactive pro-
BARNES, 1984. Alzheimer’s disease: Cell specific pathology iso- files in the normal aged human basal forebrain: Colocalization
lates the hippocampal formation. Science 225: 1168-1170. with cholinergic neurons. J. Comp. Neural. 285: 196-217.
20. JOHNSON, D., A. LANAHAN, C. R. BUCK, A. SEHGAL, C. MORGAN, 36. MUFSON, E. J., D. M. GASH, BOTHWELL, M., HERSH, L. B., AND
E. MERCER, M. BOTHWELL, AND M. CHAO. 1986. Expression and J. H. KORDOWER. 1988. Alterations in nerve growth factor recep-
structure of human NGF receptor. Cell 47: 545-554. tor immunoreactive profiles in the basal forebrain of Alzheimer’s
21. KITT, C. A., S. J. MITCHELL, M. R. DELONG, B. WAINER, AND and Parkinson’s patients. Sot. Neurosci. Abstr. 14: 1222.
D. L. PRICE. 1987. Fiber pathways of the basal forebrain in mon- 37. MUFSON, E. J., AND KORDOWER, J. H. 1989. Nerve growth factor
keys. Brain Res. 406: 192-206. receptor expression in human basal forebrain neurons: Patho-
22. KORDOWER, 3. H., R. T. BARTUS, M. BOTHWELL, G. SCnATTE- logic alterations in Alzheimer’s and Parkinson’s disease. In Pro-
MAN, AND D. M. GASH. (1988). Nerve growth factor receptor im- ceedings of the First Znternational Symposium on Alzheimer’s
munoreactivity in the nonhuman primate (Cebus apella): Distri- Disease., in press.
bution, morphology, and colocalization with cholinergic en- 38. PALMATIER, M. A., B. K. HARTMAN, AND E. M. JOHNSON. 1984.
zymes. J. Comp. Neurol. 277: 465-487. Demonstration of retrogradely transported endogenous nerve
23. KORDOWER, J. H., R. T. BARTUS, F. F. MARCIANO, AND D. M. growth factor in axons of sympathetic neurons. J. Neurosci. 4:
GASH. 1989. Telencephalic cholinergic system of the New World 751-756.
monkey (Cebus apella): Morphological assessment and analysis 39. PEARSON, R. C. A., M. V. SOFRONIEW, A. C. CUELLO, T. P. S.
of the projection to amygdala. J. Comp. Neurol. 279: 528-545. POWELL, F. ECKENSTEIN, M. M. ESIRI, AND G. K. WILCOCK.
24. KORDOWER, J. H., D. M. GASH, M. BOTHWELL, L. B. HERSH, 1983. Persistance of cholinergic neurons in the basal nucleus in
AND E. J. MUFSON. 1989. Nerve growth factor and choline acetyl- a brain with senile dementia of the Alzheimer’s type demon-
transferase remain colocalized in the nucleus basalis (Ch4) of strated by immunohistochemical staining for choline acetyl-
Alzheimer’s disease patients. Neurobiol. Aging 10: 287-294. transferase. Brain Res. 289: 375-379.
25. KORDOWER, J. H., D. M. GASH, M. BOTHWELL, L. B. HERSH, 40. PERRY, R. H., J. M. CANDY, E. K. PERRY, D. IRVING, G. BLESSED,
AND E. J. MUFSON. 1988. Nerve growth factor receptor immuno- A. F. FAIRBAIRN, AND B. E. TOMLISON. 1982. Extensive loss of
reactivity in the human and monkey brain: Distribution and co- choline acetyltransferase activity is not reflected by neuronal
232 MUFSON. BOTHWELL. AND KORDOWER

loss in the nucleus of Meynert in Alzheimer’s disease. Neurosci. pugilistica: Loss of basal forebrain neurons and cortical choliner-
Lett. 33:311-315. gic markers. Ann. Neurol. 12: 99-104.
41. RASOOL, C. G., C. ABRAHAM, B. H. ANDERSON, M. HAUGH, J. 51. VAN HOESEN, G. W., AND A. R. DAMASIO. 1987. Neural correlates
KAHN, AND D. J. SELKOE. 1984. Alzheimer’s disease: Immunore- of cognitive impairment in Alzheimer’s disease. Pages 871-898
activity of neurofibrillary tangles with antineurofilament and an- in Handbook of Physiology, Vol. 5, Chap. 22.
to-paired helical filament antibodies. Bruin Res. 310: 249-260. 52. WATSON, JR., R. E., S. J. WIEGAND, R. W. CLOUGH, AND G. E.
HOFFMAN. 1986. Use of cryoprotectant to maintain long-term
42. RINNE, J. O., L. PAU.&RVI, AND U. RINNE. 1987. Neuronal size
peptide immunoreactivity and tissue morphology. Peptides 7:
and density in the nucleus basalis of Meynert in Alzheimer’s dis-
155-159.
ease. J. Neural. Sci. 79: 67-76.
53. WHITEHOUSE, P. J. 1987. Clinical and neurochemical conse-
43. ROSENBERG, M. B., T. FRIEDMANN, R. C. ROBERTSON, M. Tus- quences of neuronal loss in the nucleus basalis in Parkinson’s
ZYNSKI, J. A. WOLFF, X. 0. BREAKEFIELD, AND F. H. GAGE. 1988. and Alzheimer’s disease. Pages 393-398 in M. D. YAHRAND K. J.
Grafting genetically modified cells to the damaged brain: Restor- BERGMANN, Eds., Advances in Neurology, Vol. 45. Raven Press,
ative effects of NGF expression. Science 242: 1575-1578. New York.
44. Ross, A. H., M. GROB, M. BOTHWELL, D. E. ELDER, C. S. ERNST, 54. WHITEHOUSE, P. J., J. C. HEDREEN, C. L. WHITE III, AND D. L.
N. MARANO, B. F. D., GHRIST, C. C. SLEMP, M. HERLYN, B. AT- PRICE. 1983. Basal forebrain neurons in the dementia of Parkin-
KINSON, AND H. KOPROWSKI. 1984. Characterization of nerve son’s disease. Ann. Neurol. 13: 243-248.
growth factor receptor in neural crest tumors using monoclonal 55. WHITEHOUSE, P. J., D. L. PRICE, A. W. CLARK, J. T. COYLE, AND
antibodies. Proc. Natl. Acad. Sci. USA 81: 6681-6685. M. R. DELONG. 1981. Alzheimer’s disease: Evidence for selective
45. SCHATTEMAN, G. C., L. GIBBS, A. A. LANAHAN, P. CLAUDE, AND loss of cholinergic neurons in the basal forebrain. Ann. Neural.
M. BOTHWELL. 1987. Expression of NGF receptor in the develop- 10:122-126.
ing and adult primate central nervous system. J. Neurosci. 8: 56. WHITEHOUSE, P. J., D. L. PRICE, R. G. STRUBLE, A. W. CLARK,
860-873. J. T. COYLE, AND M. R. DELONG. 1983. Alzheimer’s disease and
46. SQUIRE, L. R., AND S. ZOLA-MORGAN. 1988. Memory: Brain sys- senile dementia: Loss of neurons in the basal forebrain. Science
tems and behavior. Trends Neurosci. 11: 170-175. 21: 1237-1239.
41. SPRINGER, 57. WILCOCK, G. K., M. M. ESIRI, D. M. BOWDEN, AND C. T. SMITH.
J. E., T. J. COLLIER, M. F. D. NOTTER, R. Lou, AND
1983. The nucleus basalis in Alzheimer’s disease: Cell counts and
J. R. SLADEK, JR. 1989. CNS grafts of NGF-rich tissue as an
cortical biochemistry. Neuropathol. Appl. Neurobiol. 9: 175-17s.
alternative source of trophic support for axotomized cholinergic
neurons. Prog. Brain Res. 78: 401-407. 58. WILL, B. R., AND F. HEFTI. 1985. Behavioral and neurochemical
effects of chronic intraventricular injections of nerve growth fac-
48. SPRINGER, J. E., S. KOH, M. W. TAYRIEN, AND R. LOY. 1987. tor in adult rats with fimbrial lesions. Behav. Brain Res. 17: 17-
Basal forebrain magnocellular neurons stain for nerve growth 24.
factor receptor: Correlation with cholinergic cell bodies and
59. WILLIAMS, L. R., S. VARON, G. M. PETERSON, K. WICTORIAN,
effects of axotomy. J. Neurosci. Res. 17: 111-118.
W. FISHER, A. BJBRKLUND, AND F. H. GAGE. 1986. Continuous
49. TONIOLO, G., S. B. DUNNETT, F. HEFTI, AND B. WILL. 1985. infusion of nerve growth factor prevents basal forebrain neuronal
Acetylcholine-rich transplants in the hippocampus. Influence of cell death after fimbria fornix transection. Proc. Natl. Acad. Sci.
intrinsic growth factors and application of NGF on choline ace- USA83: 9231-9236.
tyltransferase activity. Brain Res. 345: 141-146. 60. YAMAMOTO, T., AND A. HIRANO. 1986. A comparative study of
50. UHL, G. R., M. MCKINNEY, J. C. HEDREEN, C. L. WHITE III, J. T. modified and thioflavin-S stains on Alzheimer’s neurofibrillary
COYLE, P. J. WHITEHOUSE, AND D. L. PRICE. 1984. Dementia tangles. Neuropathol. Appl. Neurobiol. 12: 3-9.