Neuroscience Vol. 44, No. 1, pp. 349-156,

0306-4522/91$3.00+ 0.00 Pergamon Press plc 0 1991IBRO

1991

Printed in Grest Britain

SPROUTING OF CHOLINERGIC AXONS DOES NOT OCCUR IN THE CEREBRAL CORTEX AFTER NUCLEUS BASALIS LESIONS z. I-kNDBRSON Department of Physiology, Worsley Medical and Dental Building, University of Leeds, Leeds LS29NQ, U.K. Abstract-Different doses of the excitotoxin quisqualate were used to make lesions in the caudal part of the ferret nucleus basalis, i.e. the part that projects to the visual cortex. The higher doses of the excitotoxin destroyed all nerve growth factor receptor-immunoreactive cells in the caudal nucleus basalis and gave rise to up to 75% loss of acetylcholinesterase-containing axons in the visual cortex. In sections stained for Nissl substance there was generalized tissue damage around the injection sites and extensive loss of all neuron types in areas surrounding the caudal nucleus basalis. Lower doses of the excitotoxin damaged only a proportion of the nerve growth factor receptor-immunoreactive neurons in the caudal nucleus basalis and produced a much lower depletion of acetylcholinesterase-positive fibres in the visual cortex. The only damage seen in sections stained for Nissl substance was a loss of magnocellular neurons in the vicinity of the injection sites. A quantitative morphological approach was used to show that either one week or three months after the lesions there was a linear correlation between the proportion of acetylcholinesterase-positive axons lost in the visual cortex and the proportion of nerve growth factor receptor-immunoreactive cells that had disappeared from the caudal nucleus basalis. Since the correlation lines for the short-term (one week) survival and the long-term (three months) survival experiments coincided, this indicated that no collateral sprouting of choline@ axons had occurred in the visual cortex of the lona-tenn survival animals regardless of size of the lesion in the nucleus basalis.

One of the major problems faced by neurology is that damage inflicted on the central nervous system by disease or trauma tends to be permanent, especially in the adult. Nevertheless, limited functional recovery can occur depending upon the circumstances, and the mechanisms of this recovery have been the subject of intensive inquiry using anatomical, cytochemical and biochemical methods. It is now believed that recovery from lesions involving cell death and anterograde degeneration may depend either on anatomically demonstrable sprouting from surviving axons or to compensatory metabolic processes within these axons. Of special interest has been the response to injury of the choline& projection to the cerebral cortex. This pathway is derived from neuronsin the nucleus basalis24,25*32*” and is one of the neuronal systems that is selectively affected in Alzheimer’s disease.4,36 The cause of the degeneration of cortical cholinergic axons in Alzheimer’s disease is not known, but according to the “threshold theory” (applicable also to Parkinson’s disease) clinical symptoms do not appear until a large proportion of the input is lost, apparently because of compensatory responses provided by remaining axons.B In animal models of Alzheimer’s disease, depletion of choline@ markers in the cerebral cortex is usually achieved by making excitotoxic lesions in the nucleus basalis. Thus, a

Abbreviations: AChE, acetylcholinesterase;

ChAT, choline acetyltransferase; NGF-R, nerve growth factor receptor.

few days after the injection of glutamate receptor agonists such as ibotinate or kainate into the rat nucleus basalis, up to 80% of choline acetyltransferase (ChAT) activity is lost in the cerebral cortex.‘.“,” Recovery of the levels of ChAT activity in the cerebral cortex is apparent when the animals are allowed to survive for a few months rather than a few days, although there is some discrepancy as to the actual extent of this compensatory activity.‘,5,7,‘0*3s The compensatory mechanisms responsible for the restoration of ChAT activity in the cerebral cortex after nucleus basalis lesions are not known, but so far, sprouting of choline@ axons in the cerebral cortex has not been demonstrated using morphological techniques. 2627In these anatomical studies, however, no consideration was made of the effect of the lesion size or the extent of nonspecific damage in the nucleus basalis on the occurrence of axonal sprouting in the cerebral cortex. In order to determine whether sprouting can occur in the basalocortical pathway under certain circumstances, I have performed a quantitative morphological analysis of the damage to this pathway after injections of different concentrations of the excitotoxin quisqualate into the ferret nucleus basalis. Recently it has been shown that injection of quisqualate into the nucleus basalis produces less nonspecific damage at the injection site than the more commonly used excitotoxins such as N-methyl-D-aspartate, kainate and ibotinate, at doses that produce equivalent ChAT depletions. 7*2* This study was made on the protection 149

Z. HENDERSON

150

from the nucleus basalis to the ferret visual cortex because this pathway arises from a compact region in the basal forebrain, ” further reducing the chances of nonspecific damage at the lesion site. EXPERIMENTAL

PROCEDURES

Under sodium pentobarbitone anaesthesia (50 mg/kg, i.p.) each of 18 pigmented or albino ferrets (Biomedical Services, Leeds University) received seven injections, totalline 3.5 ul. of 0.05-1.5 M auisaualate into the left caudal nucle;s b&&is. The injections were made stereotaxically at coordinates (Horsley and Clarke method): AP = 3 mm/ L=S.Smm and L=6.5mm; AP=4mm/L=Smm and L=6mm, and AP=Smm/L=4mm, L=Smmand L= 6 mm, at a depth of 8-10 mm from the pia. In the ferret, this part of the nucleus basalis, which is wrapped around the posterior part of the globus pallidus, supplies most of the choline@ projection to the visual cortex.‘5 The injections were made through the bevelled tip of a 10 ~1 Hamilton syringe. The quisqualate (Sigma or Tocris Neuramin, U.K.) was made up in phosphate-buffered saline and stored in aliquots at - 70°C until just before use. In all except for two short-term experiments, pairs of animals of the same sex, strain and weight were injected under identical conditions. One of each pair was allowed to survive for one week and the other for three months after the lesion. At the alloted survival time each animal was given a lethal dose of sodium pentobarbitone (80mg/kg, i.p.) and perfused through the ascending aorta with saline followed by chilled 2% paraformaldehyde and 15% saturated picric acid in 0.1 M phosphate buffer. nH 7.4. The brains were sunk in 30% sucrose in 0.01 M phosphate buffer and 50-pm coronal sections were cut on a freezing microtome. The extent of the damage done to the choline@ projection to the visual cortex was assessed at the level of the cerebral cortex by staining the cholinergic axons histochemically for acetylcholinesterase (AChE). In the nucleus basalis the damage was assessed by staining cell bodies for nerve growth factor receptor (NGF-R) or ChAT immunoreactivity, or for Nissl substance using Cresyl Violet. Staining for AChE was carried out using the Lewis mod&&ion of the Koelle thiocholine method.** The sections were washed in succinate buffer (PH 5.3) and incubated for up to 48 h at room temperature in medium containing 6 mM acetylthiocholine, 9 mM cupric ions and 16 mM glycine in acetate buffer at pH 5.3 (a more detailed description of the preparation of the medium is given in Ref. 17). Ethopropazine (0.1 mM) was included in the medium to inhibit pseudocholinesterase. After the incubation period the sections were washed several times in distilled water. They were then developed in a freshly prepared solution of 2.2% sodium sulphide in 0.2 N acetic acid (PH 5-6) for a few minutes in a fume cupboard. After further wases in water the sections were dehydrated and mounted under coverslips in Canada Balsam. For the immunocytochemistry, free-floating sections were stained with antibodv to ChAT*i or NGF-R.30 The antiChAT antibody was a kind gift from Dr B. Wainer and the antibody to NGF-R (clone ME20-4) was purchased from Amersham International. During incubation with antibody the sections were kept in continuous agitation on a shaker. The incubation steps were carried out at room temperature unless mentioned otherwise. In between antibody steps the sections were rinsed several times with wash solution consisting of 0.1% Triton X-100 in phosphate-buffered saline (OH 7.4). also the diluent for the antibody solutions. The aitibody. procedure was carried out using the avidin-biotin-horseradish oeroxidase kit supplied by Vector Labs. The sections first went into the appropriate 2% normal serum for 1 h and were then incubated in the

primary antibody solution (Ii300 dilution for the antrChAT and l/600 dilution for the anti-NGF-R) overnight at 4°C. The primary antibody was omitted in control incubations. After incubation in primary antibody the sections went into l/SO biotinylated anti (rat or mouse)-l&i for 1 h, and then into a l/25 dilution of the avidin biotin-horseradish peroxidase complex also for I h. The sections were then reacted for horseradish peroxiddse activity for 30 min in 0.1% diaminobenzidine and 0.004% H,O, in 0.1 M phosphate buffer (pH 7.4). The sections were dehydrated, cleared and mounted under coverslips in DPX. With the AChE histochemical method employed, It was possible to resolve each individual AChE-positive axon even in areas where the density is high, e.g. in the lower layers of the cortex (Fig. IA, B). The relative densities of AChEpositive axons in the visual cortex on the control and lesioned sides of the brain were therefore estimated using a modification of the method of Stichel and Singer.3’ Under high power and with the aid of camera lucida, a line was drawn from the pia to the white matter of the visual cortex. All the AC&E-positive axons that were seen to cross this line were then drawn and counted. The relative density of the AChE-positive axons was taken to be the number of axons that cross each micrometre of the line. Three to seven samples were obtained from equivalent parts of the visual cortex on each side of the brain and averaged. The proportional drop in density of AChE-positive axons on the lesioned side was then determined from the ratio of the two averages. The proportional loss of neurons in the nucleus basalis was determined from the numbers of NGF-R-positive neurons in sections through the caudal nucleus basalis. Under low power optics and with the aid of a camera lucida, the cells were drawn, counted and plotted as shown in Fig. 28. The area underneath each curve between the coordinates AP = 3 mm and AP = 5.5 mm was determined by counting squares. The ratio of the areas corresponding to the lesioned and control sides of the brain was then taken to indicate the proportional loss of NGF-R-positive neurons after the lesion. In order to confirm that the depletion of NGF-R-positive neurons in the lesioned nucleus basalis reflects actual cell loss rather than loss of NGF-R staining, a fluorescent dye tag was used to follow the fate of the lesioned cells in five additional experiments. In the first experiment, multiple injections of a total of 2~1 of 2% True Blue were made through a Hamilton syringe into the primary visual cortex of the ferret. This was done under sodium pentobarbital anaesthesia, with the aim of producing a distribution of tracer at the injection site similar to that described in Henderson.is Three days later the animal was perfused with fixative and sections were cut at 50pm. The sections were mounted under coverslips in phosphate buffer. Cells labelled with True Blue as a result of retrograde axonal transport from the iniection site were photographed under UV optics. The sections were then removed from under the coverslips, susoended in buffer and stained either for ChAT or for NdF-R. The sections were then mounted under coverslips in glycerol. In order to identify double-labelled cells, the immunocytochemically stained cells were observed under bright-field optics. With the aid of the camera lucida their images were then superimposed onto the photographs of the True Blue-labelled cells. Thus it could be determined which of the cells with the immunocytochemical deposit had been retrogradely labelled with the True Blue tracer. Injections of True Blue were then made into the visual cortex of four other animals, and three days later injections of 0.05-O. 1 M quisqualate were made into the posterior nucleus basalis on the same side of the brain (as detailed above). A week later the brains were fixed, sectioned, photographed and stained for NGF-R, ChAT, Nissl or AChE and processed as described in the preceding paragraphs.

No collateral sprouting in cerebral cortex after nucleus basalis lesions

Fig. 1. (A, B) Sections of ferret visual cortex stained for AChE from the lesioned (A) and control (R) sides of the brain. These were from a short survival experiment involving the injection of 0.1 M quisqualate into the nucleus basalis, resulting in the loss of 63% NGF-R-positive cells from the caudal nucleus basalis and 46% AChE-positive axons from the visual cortex. Scale bar = 0.5 mm. (C-F) !kctions of caudal nucleus baaahs from a short survival experiment, involving the injection of 0.1 M quisqualate into the nucleus basalis, and resulting in the loss of 62% NGF-R-positive cells in the caudal nucleus baaalis and 49% AChEpositive axons in the visual cortex. (C, D) Neurons stained for NGF-R on the lesioned (C) and control (D) sides of the brain (33,F) Views from adjacent sections stained for Nissl, corresponding to the boxed regions in C and D, mspectively. Mild gliosis plus loas of large cells is visible at the leaion site (E). nbm, nucleus basalis magnocelhtlaries; ep, entopeduncular nucleus; put, putamen. Scale bar = 1mm for C, D and 0.25 mm for E, F.

151

152

Z.

A

1~ 5

HENDERSO~V

A Long survival 0 Short ssrrirol

1

2

9 q--y--l

im

%NGF

lKl1601uI _

120-

3

-R+vc

ncwons lost in N&l

rZOlahorl surriralI &lM quiqsatr 0 Conlnl tide A Lesion rids

35 A-P

i

distance

I I t

1

5.5 of section

6

trOtI7 car bar ZWO~~tTI)

Fig. 2. (A) A graph depicting the correlation between the loss of AChE-positive axons in the visual cortex and the loss of NGF-R-positive neurons in the caudal nucleus basalis one week (y = 0.37x + 27.6, r = 0.7, P < 0.05, dashed line) and three months (Y = 0.37x + 22.3. r = 0.8. P < 0.02. continuous line) afte;.injections of quisqualat~ into the’ posterior nucleus basalis in the ferret. The two correlation lines

almost coincide, suggesting that regenerative sprouting of AChE-positive axons did not occur in the long-term. (B) Example from experiment F20, showing how the numbers of NGF-R-positive cells in the caudal nucleus basalis were plotted for each experiment.

RESULTS

In the normal ferret brain the distribution of cells that stain positively for ChAT immunoreactivity is the same as that in other species.i6 These cells are located in the striatum (caudate nucleus, putamen and nucleus accumbens), olfactory tubercule, medial septal nucleus and nucleus of the diagonal band, and in the substantia innominata and around the globus pallidus where they collectively constitute the nucleus In rat and primate basalis magnocellularis. 16~19~25 species it has been established that the forebrain neurons that stain for NGF-R immunoreactivity correspond specifically to the cholinergic cells that project to cortical areas, i.e. the cerebral cortex and the hippocampus. 37The distribution of neurons that stain for NGF-R have, therefore, a more restricted distribution than the ChAT-positive neurons and they are confined mainly to the medial septal and diagonal band nuclei and the nucleus basalis.2,‘9 This pattern of distribution of NGF-R in forebrain neurons is also found in the ferret, using the

ME20-4 antibody directed against the human NGf-R receptor.‘* As expected, injections of quisqualate into the ferret nucleus basalis reduced the numbers of NGFR-positive (Fig. lC, D) and ChAT-positive neurons there within a week. For the quantitative analysis, the reduction in the numbers of the NGF-R-positive rather than the ChAT-positive neurons was taken as an indicator of the damage inflicted by the quisqualate on the basalocortical pathway. This is because the NGF-R immunoreactivity provides a means for distinguishing cortically projecting cholinergic neurons from neighbouring cholinergic cells that are associated with the circuitry of the striatum. The experiments involving the retrograde labelling of nucleus basalis neurons with True Btue also confirmed that the decline in the numbers of NGF-Rpositive neurons was due to cell death rather than to loss of expression of NGF-R immunoreactivity in intact neurons (Fig. 3). In the control experiment the injection of True Blue into the ferret visual cortex labelled ChAT-positive and NGF-R-positive neurons that were distributed primarily in the caudal part of the nucleus basalis, as predicted from a previous study.” In the True Blue experiments in which quisqualate was injected into the caudal nucleus basalis three days after application of the retrograde tracer, either no retrogradely labelled cells were observed in the nucleus basalis (two experiments) after a week, or their numbers were considerably reduced (two experiments). This loss of cells was correlated with a parallel reduction in the numbers of neurons that stained for NGF-R in the caudal nucleus basalis. In the last two experiments the retrogradely labelled cells that remained in the nucleus basalis all showed positive staining either for NGF-R or ChAT (Fig. 3). The NGF-R-positive retrogradely labelled cells rctained normal levels of staining intensity. whereas some of the ChAT-positive retrogradely labelled cells showed reduced levels of staining in comparison with neurons in the control side of the brain (Fig. 3). The damage done to the nucleus basalis as seen in section stained for Nissl substance was variable and depended on the dose of quisqualate injected. At quisqualate concentrations of 0.12 M or more. the damage consisted of extensive gliosis and the disappearance of all neurons around the injection sites. There was also loss of neurons in nearby regions. especially in the amygdala and the paleocortex, but there was no damage due to backflow of the neurotoxin up the needle track. This extensive type of damage seen in the Nissl preparations usually correlated with up to 100% loss of NGF-R-positive cells in the posterior nucleus basalis and 75% reduction in the numbers of AChE-positive axons in the visual cortex. With the lower concentrations (0.05-O. I M) of quisqualate used the only visible damage at the injection sites was a mild gliosis and a loss of large (magnocellular) neurons in the nucleus basalis and in neighbouring parts of the globus pallidus and

No collateral sprouting in cerebral cortex after nucleus basalis lesions

153

Fig. 3. Surviving neurons (arrows) near quisqualate (0.0541 M) injection sites in the caudal nucleus basalis. These cells were retrogradely labelled with True Blue (A, C, E) after injection of the tracer into the visual cortex prior to the lesion being made. The photographed cells come from two experiments in which there was 5946% loss of NGF-R-positive cells from the caudal nucleus besalis. (B, D, F) The True Blue-Welled cells after staining for NGF-R (B, D) and ChAT (F). (G) ChAT-positive cells in the caudal nucleus basalis on the control side of the brain. Scale bar = 0.25 mm. entopeduncular nucleus (Fig. lE, F). This damage was associated with smaller reductions in the

experiment, an injection of 0. I M quiaqualate cerebral cortex did not damage any cortical

into the cells. In

numbers of cells and fibres in the nucleus basalis and cerebral cortex, respectively. In one control

another control experiment, the injection of phosphate-buffered saline into the nucleus basalis did not

Z. HENDERSON

154

damage any of the “magnocellular” neurons of the basal forebrain. Thus in the ferret, low concentrations of quisqualate selectively injure large neuron cell types although these need not exclusively be choline@ cells. The effects that caudal nucleus basalis lesions had on the cholinergic cortical circuitry were monitored by looking at the changes in densities of AChEpositive axonal networks in the ferret visual cortex (Fig. IA, B). In a previous study it was shown that the distribution of axonal AChE in the ferret visual cortex is a good indicator of the distribution of the choline@ axons there. ” The depletion of AChEpositive axons extended throughout area 17 and was observed in all cortical layers (Fig. IA, B). There was, however, always a residual 20% or more of AChEpositive axons (Fig. 2A). The remaining axons most likely belong to nucleus basalis cells that lie outside the range of the injected excitotoxin rather than to cortical interneurons, since surgical isolation of the cat visual cortex leads to complete loss of axonal staining for AChE.3 The quantitative analyses on the relative densities of AChE-positive fibres and numbers of NGF-Rpositive cells in the different experiments showed that there was a direct linear correlation between the loss of AChE-positive axons in the visual cortex and the loss of NGF-R-positive neurons in the caudal nucleus basalis (Fig. 2A). This correlation applied to both the short- and the long-term experiments (Fig. 2A). The two correlation lines were compared: if after the long survival times there was a greater density of AChEpositive axons in the visual cortex than could be counted for by the number of NGF-R-positive neurons in nucleus basalis, this would indicate that collateral sprouting of cortical choline@ axons had taken place in the long-term. Since there was a virtual coincidence between the two correlation lines (Fig. 2A), it was assumed that there had been no change in axon density after three months over what was expected from the numbers of NGF-R-positive cells remaining in the posterior nucleus basalis, over the whole spectrum of damage caused by the quisqualate lesions. DEXXJSSION

In this study in the ferret, a quantitative morphological approach was used to monitor changes in the basalocortical pathway at both the level of the cerebral cortex and the nucleus basalis in response to lesions made in the nucleus basalis with the excitotoxin quisqualate. The short- and long-term damage to the cortical cholinergic innervation was assessed from the numbers of AChE-positive axons in the visual cortex and NGF-R-positive neurons in the caudal nucleus basalis which projects to the visual cortex. Nonspecific damage at the lesion site was determined by examining Nissl-stained sections. Thus it was shown that low doses of quisqualate selectively

injure large neuron cell types located mainly in the nucleus basalis, whereas higher doses destroy all neuron types located in the nucleus basalis and in the surrounding areas. In high but not low doses, therefore, quisqualate produces damage at the injection which mimics that routinely obtained with kainic acid and ibotenic acid.7.26Nevertheless, the results of the quantitative analysis of this study indicated that sprouting of cholinergic axons in the cerebral cortex does not occur in the long-term regardless of the type and extent of damage produced in the nucleus basalis by quisqualate injections. In conclusion, it is highly likely that the long-term recovery of ChAT in the cerebral cortex following excitotoxin lesions in the nucleus basalis’~s~i0~‘5 is due to the increase in the synthesis of ChAT by parent neurons of remaining cholinergic axons, as proposed by Ojima et al.*’ The lack of collateral sprouting of choline+ axons in the cerebral cortex and the possibility for an increase in ChAT synthesis in response to basal forebrain lesions has certain parallels with the compensatory responses that occur after damage to the dopaminergic nigrostriatal pathway which depend on metabolic changes rather than on axonal sprouting. 3’ Both collateral sprouting of cholinergic axons and recovery in the levels of ChAT have been reported after injury to the septohippocampal pathway(‘.*,” which is functionally related to the basalocortical pathway. This may be due to the different methods used for producing lesions in the hippocampal pathway which produces damage to axons rather than cell bodies, or because of differences in anatomical organization.823 Evidence obtained with the AChE histochemical method points to the collateral sprouting occurring from a ventral pathway which is not affected by sectioning the fimbria, the main fibre bundle through which the septohippocampal pathway courses.9 Nevertheless, this does not rule out metabolic compensation being an important factor for the recovery of ChAT activity in the hippocampal as well as in the basalocortical system. Why the septohippocampal or basalocortical pathway should show any signs of recovery at all after a lesion is suggested by experiments involving the intracerebral application of the NGF, a putative trophic substance of basal forebrain cholinergic neurons.37 In adult animals, repeated intraventricular injections of NGF immediately after a partial fimbria-fornix lesion hastens the recovery of ChAT activity in the septum and hippocampusi4 and prevents the permanent demise of a proportion of septal cholinergic neurons.‘3.20 Interestingly, this treatment has no effect upon the histochemical staining for AChE in the hippocampus.14 This suggests that the NGF boosts the synthesis of ChAT by the septal neurons rather than promotes sprouting from their axons, which is in support of the metabolic compensatory mechanisms operating in this system. The application of NGF also improves the recovery of cholinergic markers in the cortex after nucleus basalis

No collateral sprouting in cerebral cortex after nucleus basalis lesions lesions.“*” Thw, the long-term recovery processes after basal forebrain or Gmbria-fomix lesions most likely depend upon the basal forebrain choline+ neurons gaining access to an adequate supply of endogenous

NGF,

normally

ebral or hippocampal

synthesized

in the cer-

cortex and retrogradely trans-

ported to the cell bodies of choline@ basal forebrain.37

15s neurons in the

~cknowfedge~~~-us

research WBSsupported by MRC grant No. G8621410N. I am grateful to Ann Stamper and David Johanson for their technical assistance.

RJIFERENCFS

1. Arendash G. Hr., Millard W. J., Dunn A. J. and Meyer E. M. (1987) Long term ne~o~tholo~~l and neur~h~~l effects of nucieus basalis lesions in the rat. Science 23g, 952-956. 2. Batchelor P. E., Armstrong D. M., Blaker S. N. and Gage F. H. (1989) Nerve growth factor receptor and choline acetyltransferase colocalisation in neurons within the rat forebrain: response to &nbria-fomix transection. J. camp. Neural. 2&1, 187-204. 3. Bear M. F., Carnes K. M. and Ebner F. F. (1985) An investigation of cholinergic circuitry in cat striate cortex using acetylcholinesterase histochemistry. J. camp. Neural. 234, 41 l-430. 4. Bowen D. M., Benton J. S., Spillane J. A., Smith C. T. T. and Allen S. J. (1982) Choline acetyltransferase activity and ~sto~~oIo~ of frontal neocortex from biopsies of demented patients. J. neural. Sci. S7, 191-202. 5. Casamenti F., Di Patre P. L., Bartolini L. and Pepeu G. (1988) Unilateral and bilateral nucleus basalis lesions: differences in neurochemical and behavioral recovery. Neuroscience 24, 209-216. 6. Dravid A. R. and van Heusen E. B. (1984) Recovery of enxytne markers for cholinergic terminals in septo-temporal regions of the hippocampus following selective Embrial lesions in adult rats. Bruin Res. 324, 119-128. 7. Dunnett S. B., Whitshaw I. Q., Jones G. H. and Bunch S. T. (1987) Behavioural, biochemical and histochemical effects of different neurotoxic amino acids injected into nucleus basalis magnocellularis of rats. Neuroscience 20, 653-670. 8. Gage F. H., BjiirkIund A. and Stenevi V. (1983) Reinnervation of the partially deafferented hippocampus by ~rn~nsato~ collateral sprouting from spared cholinergic and noradrenergic aBerents. Brain Res. 268, 27-37. 9. Gage F. H., BjSrkland A. and Stenevi U. (1984) Cells of origin of the ventral cholinergic septohippocampal pathway undergo compensatory collateral sprouting following fimbria-fomix transection. Neur&i. &tr. 44, 21 l-216. 10. Gardiner I. M.. De Belleroche J.. Premi B. K. and Hamilton M. H. (1987) Effects of lesions of the nucleus basalis of rat on acetylchbline release in cerebral cortex: time course of compensatory events. Bruin Res. 4@7, 263-271. 11. Haroutonian V., Kanof P. D. and Davis K. L. (1986) Partial reversal of lesion induced deficits in cortical choline@ markers by nerve growth factor. Brain Res. 386, 397-399. 12. H~outunj~ V., Kanof P. D. and Davis K. L. (1989) Attention of nucleus basalis of Meynert lesion-ind~ cholinergic deficit by nerve growth factor. &&I Res. 487, 20&204. 13. Hefti F. (1986) Nerve growth factor promotes survival of cholinergic neurons after fimbrial transection. f. Neurosci. 6, 2155-2162.

14. Hefti F., Dravid A. and Hartikka J. (1984) Chronic intraventricular injections of nerve growth factor elevate hippocampal choline acetyltransferase activity in adult rats with partial septohippocampal lesions. Bruin Res. 293, 305-311.

1.5. Henderson 2. (1987) Source of the eholinergic input to ferret visual cortex. Bruin Res. 412,363-369. 16. Henderson 2. (1987) A small proportion of choline&c neurons in the nucleus basalis rna~~llnla~s of ferret appear to stain positively for tyrosine hydroxylase. Brain Res. 412, 363-369. 17. Henderson 2. (1987) Cholinergic innervation of ferret visual system. Neuroscience 20, 581-592. 18. Henderson 2. and Evans S. (1991) Presence of a cholinergic projection from ventral striatum to amygdala that is not immunoreactive for NGF receptor. Neurosci. Left. (in press). 19. Kordower J. H., Bartus R. T., Bothwell M., Shatteman G. and Gash D. M. (1988) Nerve growth factor receptor immunoreactivity in the nonhnman primate (Cebus apefla): distribution, morphology, and colocalisation with cholinergic enzymes. J. camp. Nearol. 277, 46-7. 20. Kromer L. F. (1987) Nerve growth factor treatment after brain injury prevents neuronal death. Sctence 235,214216. 21. Levey A. I., Armstrong D. M., Atweh S. F., Terry R. D. and Wainer B. H. (1983) Monoclonal antibodies to choline acetyltransferase: production, specificity and immunohistochemistry. J. Neurosci. 3, l-9. 22. Lewis P. R. (1961) The effect of varying the conditions in the Koelle technique. Bibl. Anar. 2, 11-20. 23. Lewis P. R. and Shute C. C. D. (1967) The cholinergic limbic system: projections to the hippocampal formation, medial cortex, nuclei of the ascending cholinergic reticular system, and the subfomical organ and the supra-optic crest. Brain 90, 521-540. 24. Mesulam M.-M., Mufson E. J., Levey A. I. and Wainer B. H. (1983) Choline@ inne~ation of wrtex by the basal forebrain: cytochemistry and cortical connections of the septal area, diagonal band nuciei, nucleus basalis (substantia innominata) and hypothalamus in the rhesus monkey. J. camp. Neural. 214, 170-197. 25. Mesulam M.-M., Mufson E. J., Wainer B. H. and Levey A. I. (1983) Central cholinergic pathways in the rat: an overview based on an alternative nomenclature (ChlCh6). Neuroscience 10, 1185-1201. 26. Muf.son E. J., Kehr A. D., Wainer B. H. and Mesulam M. M. (1987) Cortical effects of neurotoxic damage to the nucleus basalis in rats: persistent loss of extrinsic choline@ input and lack of transsynaptic effect upon the number of somatos~tin-conning, cho~n~tera~-positive and choline& cortical neurons. Brain Res. 417, 385-388. 27. Ojima H., Sakurai T. and Yamasaki T. (1988) Changes in choline acetyltransferase ~~or~ti~ty and the number of immunoreactive flbres remaining after lesion to the magnocellular basal nucleus of rats. Nearosci. fatr. 95, 31-36.

28. Robbins T. W., Ever&t B. J., Ryan C. N., Marston H. M., Jones G. H. and Page K. J. (1989) Comparative effects of quisqualic and ibotenic acid induced lesions of the substantia innominata and globus pallidus on the acquisition of conditional visual discrimination: differential effects of cholinergic mechanisms. Neuroscience 28, 337-353. 29. Roth M. (1986) The association of clinical and neurological findings and its bearing on the classification and aetiology of Alxheimer’s disease. Br. med. B&f. 42, 42-50,

156

Z. HENDERSON

30. Ross A. H., Grob P., Bothwell M., Elder D. E., Ernst C. S., Marano N., Christ B. F., Slemp C. C.. Herlyn M.. AtkInson

31. 32. 33. 34. 35. 36. 31.

B. and Koprowski H. (1984) Characterization of nerve growth factor receptor in neural crest tumours using monoclonal antibodies. Proc. nafn. Acad. Sci. U.S.A. 81, 6681-6685. Shultz W. (1984) Recent physiological and pathophysiological aspects of Parkinsonian movement disorders f-rfc err 34, 2213-2223. Shute C. C. D. and Lewis P. R. (1967) The ascending cholinergic reticular system: neocortical. olfactory and subcortical projections. Brain 90, 497-520. Stichel C. C. and Singer W. (1987) Quantitative analysis of the choline acetyltransferase-immunoreactive axnnal network in the cat primary visual cortex. I. Adult Cats. J. romp. Neural. ZSS, 91-98. Wenk H., Big1 V. and Meyer U. (1980) Cholinergic projections from magnocellular nuclei of the basal forebrain to cortical areas in rats. Brain Res. Rev. 2, 295-316. Wenk G. L. and Olton D. S. (1984) Recovery of neocortical choline acetyltransferase activity following ibotenic acid injection into the nucleus basalis of Meynert in rats. Brain Res. 293, 184-186. Whitehouse P. J., Price D. L., Clark A. W., Coyle J. T. and DeLong M. R. (1981) Alzheimer’s disease: evidence for a selective loss of choline@ neurons in the nucleus basalis. Ann. Neural. 20, 122-126. Whittemore S. C. and Sieger A. (1987) The expression, localisation and functional significance of B-nerve growth factor in the central nervous system. Brain Rex Rev. 12, 439464. (Accepted

13 March 1991)