Brain Research, 586 (1992) 27-35 © 1992 Elsevier Science Publishers B.V. All rights reserved 0006-8993/92/$05.00

27

BRES 17837

Nerve growth factor promotes collateral sprouting of cholinergic fiber,,; in the septohippocampal cholinergic system of aged rats with fimbria transection H e Y u n s h a o , Y a o Zhibin, G u Yaoming, K u a n g G u o b i and C h e n Yici Department of Anatomy, Sun Yat-sen University of Medical Sciences, Guangzhou (People~ Republic of China) (Accepted 28 January 1992)

Key words: Nerve growth factor; Cholinergic fiber; Sprouting; Septum; Hippocampus; Brain aging

Nerve growth factor (NGF) was injected intraventricularly into aged (24 months) rats with unilateral fimbria transection. Controls received intraventricular injections of cytochrome c. A quantitative analysis of acetylcholinesterase (AChE)-positive fibers was used to evaluate whether the NGF treatment can stimulate regeneration and reinnervation of the cholinergic axons in the septohippocampal system of aged rats with fimbria transection. A marked increase in the density of AChE.positive fibers was observed in the lateral septum, the dorsal fornix and the dorsal hippocampus of the NGF-treated animals, as compared to the controls. In the lateral septum, the increase was observed in the 2-month NGF-treated animals but not in the 15-day NGF-treated animals. In the dorsal fornix at the level of the dorsal hippocampus, the increase was observed on both the lesioned and unlesioned sides of both the 15-day and 2-month NGF-treated animals, in the denervated (lesioned side) hippocampus, the increase took place in the dorsal hippocampus but not in the ventral hippocampus of both the 15-day and 2-month NGF-treated animals. There was no recovery of AChE-positive fibers on the lesioned side of the fimbria distal to the lesion site even in the 2-month NGF-treated animals. These results demonstrate that intraventricular injections of NGF can stimulate collateral sprouting of intact cholinergic axons in the septohippocampal system and promote cholinergic reinnervation of the denervated hippocampus of aged rats with fimbria transection,

INTRODUCTION

Nerve growth factor (NGF) has recently been shown to be atrophic factor for basal forebrain cholinergic neurons of the adult rat 4'l°'tl'ts. The basal forebrain projections from the medial septum (MS) and diagonal band of Broca (DB) to the hippocampus provide an anatomical model system for the study of the role of NGF. Transection of the septohippocampal pathway via the fornix and fimbria (FF) leads to biochemical and anatomical changes, which include a pronounced loss of cholinergic and non-cholinergic neurons in the MS and DB, and reductions in choline acetyltransferase (CHAT) activity and in acetylcholinesterase (ACHE) fiber staining in the hippocampus t'5-8, lntraventrieular infusions of NGF were found to prevent the lesion-induced loss of septal cholinergic neurons and

attenuate the reductions in ChAT activity in young adult rats 'j't4'ts't'j'a°''~land aged rats j~'~s, Several investigators have reported that intraventricular administration of NGF can promote local growth of cholinergic neurites t¢',3°, but the nature and origin of this growth of cholinergic neurites have not boon determined and it is not known whether adminstration of NGF can promote the growth of cholinergic ncurites in the lesioned septohippocampal system of aged animals. In the present study, we used a quantitative histochemical technique which allows the visualization of AChE-positive fibers to determine whether NGF promotes growth of cholinergic neurites and analyze the possible source and mechanism of the growth of cholinergic neurites in the lesioned septohippocampal system of aged rats.

Correspondence: H. Yunshao, Department of Anatomy, Sun Yat-sen University of Medical Sciences, Guangzhou 510089, People's Republic of China.

28 MATERIALS AND METHODS Surgical pr~x'edur('.~ mul intn,',vuricuh~r injections Aged (24 months) female albino rats of the Sprague-Dawley strain, 4(X)-5()) g in weight, underwent a partial unilateral transection of the septohippocampal pathway. Briefly, rats were anesthetized with sodium pentobarbital (40 mg/kg) and placed in a stcrcotaxic device. A specially designed knife was positioned vertically on the left side and at a starting point I mm lateral to midline, and 2.2 mm posterior to the bregma, lowered to depth of 5.5 mm from the skull surface, driven laterally to 2 mm of the midline, then lowered further to 6.5 mm from the ~kull surfitee, and driven laterally 4 mm from the midline. After the lesion, a cannula device was stereotaxically inserted into the lateral ventricle on the lesioned side for repeated injections, as described in detail elsewhere =-'. Half of the animals were injected intraventricularly twice weekly with 15 ~g of NGF (2.5 5) from mouse suhmandibular glands dissolved in 5 pl of a artificial ccrebrospinal fluid '~". The other half served as controls; they received equal amounts of cytochrome c. The injections after the operation were not performed under anesthesia, A total of 20 animals were divided into 4 groups. Two groups ~,'cr¢ treated for 15 days with NGF or cytochrome c (controls). The other two groups were treated for 2 memths with NGF or cytochrome c (controls), Groups of 4 animals were used for the treatment of 15 day.~ and groups of h animals fi)r 2 monlhs, l~ul after the 2-month treatment 4 animals per group remained and the others died of illness.

Fig. 1. A simple square lattice system on a micrograph of the AChE-positive fibers of the hippocampus. The AChE-positive fiber density is expressed by the number of intersects (arrow) or the fibers with the vertical and horizontal lines.

Acetyh'holmc..m.rase hi.~tnchemi,~tO' Fifteen days and 2 months after the operation, animals were anesthetized and perfused through the heart with Illll ml of (I.I M phosphate buffer followed by 4(XI ml ~)l"4C/ paraformaldehyde. The brains were rcntoved and cryoprotected with 1(I-25e~ sucrose solution in phosphate buffer, 15-#re.thick sections were cut for AChE histochemist~, llt order to obtailt coltsistenl secti()ns, the dorsal hippoc=mtl)US was CUt ill the coronal plane and the ventral hippt~cam. pus m the horizontal phme, A('hE.p,~sitive fibers were sh,~wn with the metlmd described by Iledreen el al, I~, ('ry~stat s¢ctiOll,~ were stained on slides, The pmcedt, re Ceulsists el' the follt~wiltl~ serk,,s t~l' steps: (I) section,~ were rinsed in tw,~ ¢llatt~es of I), I M acetat~ I~uffer, pll h,ll, (2) Incul~,titql wa,, perfornlcd in the followin~t medium: 2,~ it1~ of aeetylllti¢~cholin¢ i,~dide, 32,~ nil of II, I M ~odium acetate huffer, pl I t~,(), 2 nil ¢~t'(I, I M ,,~diu111 cilrale, ~ nil ed' (I,03 M cupric sull'ate, ¢),~ I111,tl' distilled !1 ,(), I nil ,if II,llll~ M pt~tits~iuttt ferrieyanide, hleubation was ut ro~fll temperature for 31l rain, (.I~ After .~ ~.'hanges of acetate but'fur (I ntin each), sections were treated with It; amntonium sulfide solution t'~r I rain followed by 5 changes of ILl M sttdiunt nitrate (I ntin each), (4) Sections ~,'ere then eXl~)sed to I':; silver nitrate for I rain t'olh~wing hy ,~ changes of ILl M sodium nitrate (I rain each). (5) Sections were dehydrated, cleared and eoverslipped. Specificity of staining was tested using the AChE inhibitor, eserine sulfide (Ill 4 M),

E¢ ahmtt.. ~q',.l('hE.pr~,~.itir~ [ilwr ch'tt,~ily For a quantitative analysis of AChE-positive fiber density, secIrons were selected from the fi)llowing regkms: ( I ) the lateral septum a~ 3fill pm r(;stal to the anterior comn)isure decussation and below the corpu.,, callosum; (2) the dorsal forxin at 2 mm l~sterinr to the .,,eptal pole of the hippy)can)pus; (3) the ('AI and CA3 regk)ns and the dentate gyms opposed to the CAI region of the dorsal hippocampu~ at 2.5 mm posterior to the septal pole and those ot" the ventral hippt~campus at 25 mm dorsal tt~ the temporal pole, The AChE-positive fiber density in the regions except the dt)rsal fl~rnix was assessed with a simple square lattice system ~n pht)tomicrographs (Fig. I ), The system consists elf paralleled and perpendicular line5 drawn at I cm intervals on transparent plastic film, The number ~)f fibers intersecting each line was counted and recorded, The fiber density was expressed as the number of intersects in li),()(X) pro" real area, which was calculated based on the multiple of measured ph¢~tographs, The number of transected points of ACHEpositive fibers in the dorsal fornix was directly counted because those fiher~ were perpendicular to the c~)ronal plane.

Stati,~'tical analysis Statistical differences were assessed by means of ANOVA, in order to assess differences between controls (eytoehrome c) and NGF-treated groups, as well as between lesioned and unlesioned sides. The Newman=Keuls test or Scheff F-test was used respectively whell a varialtee was hornogelleOUS or Iteter(Igeneous,

RESULTS

TIw AChE.l~nSitit'e fiber density in the lateral ,~'('pttml AChE-positive fibers in the medial septum were dense with a deep background and those in the lateral septum were sparse with a clear background in normal animals, Thus, the htteral septum is favorable to analysis of AChE-positive fiber density, Fifteen days after the lesion, there was no difference between the AChE-positive fiber densities on the lesioned and unlesioned sides of the lateral septum of the cytochrome c-treated animals. There was a slight increase in the AChE-positive fiber density which did not reach significance in the 15-day NGF-treated animals, as compared to the cytochrome ('-treated animals (Table !). There was a dramatic increase in the AChE-positive fiber densities on both the lesioned and unlesioned sides of the 2-month NGF-treated animals, as compared to the cytochrome c-treated animals, and the increase was grcater on the lesioned side than that on the unlesioned side (Fig. 2, Table It. Many AChE-positive fibers were present in a lateral zone of the lateral septum and in the ependyma and appeared toward the lateral ventricle in the 2-month NGF-treated animals, but few of the fibers were there in the control animals.

29 TABLE I

Effect of NGF on numbers of A ChE-positire fibers b~ the lateral septum of aged rats

unilateral fimbria transection

with

Each value represents the number {+S.E.M.) of intersects of fibers with the square lattice system per 10,000 /~m z real area. Controls, cytochrome c-treated controls. 15-day administration Lesioned {left)side Unlesioned {right) side

2-month administration

Controls

NGF

Controls

NGF

282+ 16 237 + 26

341 + 15 254 + 65

312 + 14 # 212 + ! i

534 + 14 *# 377 + 35

* P < 0.01 compared to c3,tochrome c-treated controls~ # P < 0.01 compared to unlesioned side. ~

.~,~.z - ~&~... ,-, .

..~

~,W'~(%.

.

.~,, -' .,,. t

,,.

~. •

' '

, ' "

..., ~.,. ,:.. ,~ ,~,-. ,:.,.,:.,, . ..

:

:

~,"6,~. ~:,,'W"

,

)~

.... )~.-..;-~.-~...~,,~.~,~,_~'.D~:~:~,~J'~',.,o .,~ n ~, ~.~:',, .~ ~ . % , ~ , r . t..">.'~;2~" ~ : : ~

:.,,.,~,;, , : ; , . ~ ~ ~ ~

~-.-, -. v

~. :.,. :,. :;~:.

~:,'(

.

,,

.

,t~

'

,),

,.,-~,

.

a

'

.,~,~.:~.~

£-.,

~.

~

,

.,'. , s ' - , -

~',,"

~,.~v,

~~'t~.,

?Y ,~ ...... ",".~'~ ~:~/~

,~~)t:~~~:,~,?

~.~':~.~zl,

.

.

.

.

'. '

.

~,

~, 'N

• k,.

. . . .

"'

.~'~

'

,,,;.",

.

'"

"~ '

• \

,

..,

":-l','(',,;~

,,~f,,: -)"

"

'., ~',,

,,'

•

,

,",-. , ~,

,

',~

\ ,

:.

. , .q~ , , '"'.k'

A'"~

~ ....

'~~ ' , :"'~'.m,., "d, , ' . , '~. %.,.

~

,

",r

"

,

;'.'I.

4""

',,'.. , : ' , .

'

-,

" ' "'~

~' ' ~ " ~ ,

.,,,

,,P',)

" :

, i ~

~.,,'t'.;"~-",. t !' ,'~'~," ~,I,).~:.

,,~,t'~,

"), , . ~. , , " "~ ,"' .i7;

t

, ~,,,'.~

,

IN')',~:" ', i ~ '-'.."~'.). > .

~,.

~,)~

..,,~':,,~,:. ,~ ~ . ~

.

J"

..~,

,

,,,.

' " -i" /

,~.

,!,~

Fig. 2. Micrographs of AChE-positive fibers in the lateral septum 2 months after fimbria transection, The densities of AChE.positive fibers o n the unlesioned (A) and lesioned (B) sides of NGF-treated animals are greater respectively than those of cytochrome c-treated animals, Arrows. surface of the ependyma. Bar = 25 p.m,

TABLE !I

Effect of NGF on the numbers of A(.3zE-positire fibers in the dorsal fon~ix of aged rats with mailateral h,sio, of the lateral fimbria Each value represents the number ( + S.E.M.) of transected points in the whole fornix. Controls, cytochrome c.treated animals. 15-day administration

Lesioned (left) side Unlesioned (right) side

2-month administration

Controls

NGF

Controls

NGF

331 + 14 288 + 12

521 + 25 * 458 + 5 *

443 + 12 400 + 43

536+ 3t) * 567 + 39 *

* P < 0.01 compared to cytochrome c-treated controls,

3O # ,~

~

, ~ .

~r ~

..... %,~, ,: a ,e:tr..~ ~," A , . , ,,°t,O4P O", , .. ..... ,,,. L~;?"~ ' ' ~.,.

q,

•

.~

::,? " ~,

• •

.-

•

'e"

.

•

,.

o

.0,/

4:'

,

-

• %'4,,.1'. ,

,

e

W~,

*

.

•

" i¢,,, ,. , " ff~ ,.:.' , 0/$~?~,.

~,

...a~,.

~

•

~%q~.~l

"

d, ,b. ~ ,, :-~.~,~. ~:,,

,

"lik.'St?,'-~r~' • d w

. ~mt¢_,~

•

~.,

~:~ ~ , ~ . u . ," ~ ,~4~t,~; ! ~.':, ,. :.. j..'~'I('~:o~'~Y,',

~ .~.¢~'~ "e.. *. • ,

•

B

~,.*.

.., • .

"t

¢

~ ¢

~,

~,'~ • ,~°

.

- |" '"

" . " ,..m '

.a, .

t

" .

Fig. 3. Micrographs of AChE-positive fibers in the lesioned side of the dorsal fornix 2 months after fimbria transection. Note the apparent increase in the AChE-positive fiber number in the dorsal fornix and in the dorsal hippocampal commissure (arrow) of NGF-treated animals (A) as compared to cytochrome c treated animals (B). CC. corpus callosum; DF, dorsal fornix; DHC, dorsal hippocampal commissure. Bar = 25 ~m,

The AChE-positit'e fiber number hi the dorsal fornix in the d o r s a l fornix were arranged so t'egularly that they were shown as transected points (Fig. 3) in the coronal section. On both the lesioned and unlesioned side of the NGFtreated animals for 15 days and 2 months, the ACHEpositive fiber number were markedly greater than those of the cytoehrome c-treated animals (Fig. 3, Table ll). The cytochrome c-treated animals for 2 months had greater AChE-positive fiber number than the cytochrome c-treated animals for 15 days (Table ll). A l m o s t all o f A C h E - p o s i t i v e

fibers

AChE.positil,e fibers in the dorsal hippocampai commis. slurp

"['he AChE-positive

fibers were few ill the dorsal

h i p p o : a m p a l commissure o f the cytochrom¢ c-treated

animals, but increased markedly in the NGF-treated animals for 2 months (Fig. 3). The fibers in the dorsal hippocampal commissure appeared to project from the dorsal fornix to the denervated hippocampus.

The AChE-positit,e fiber density in the dorsal hippocampus

in the hippocampal formation, AChE-positive fibers were heterogeneous in density. The greatest density of AChE-positive fibers was present in the pyramidal cell layer, lower densities in the oriens layer, the stratum radiatum and the lacunosum moleculare layer, and few AChE.positive fibers in the alveus and the hippocampal fissure. The AChE-positive fibers were arranged in all orientations in the above mentioned regions, but AChE-positive fibers arranged in some orientation were

TABLE !11

Effect of NGF ot~ rim, hers of AChE.l~Jsitit'e fibers in the hippocampal formation of aged mrs with mlilat¢,ralfimbria tramectkm Each value represents the number (+S.E.M.) of intersects of fibers with the square lattice system per 10,(XX) ttm: real area, Controls, cytochrome ('-treated animals; DH, dor,~al hippocampus; VH, ventral hippocampus: DO, dentate gyrus, 15-day administration Lesioned side of the CAI of DH Unlesioned side of the CAI of DH Lesioned side of the CA3 of DH Unlesioned side of the CA3 of Dtl Lesioned side of the dorsal DG Unlesioncd side of the dors,d DG Lesioned side of the CA! of VF! Unlesioned side of the CAi of VH Lesioned side of the CA3 of VF! Unlesioned side of the CA3 of VH Lesioned side of the ventral DG Unlesioned side of the ventral DG

2-month administration

Controls

NGF

Controls

NGF

34+10.' 98+ 15 35+ 4 *' 121 + It) 57+21" 159+21 21 + 3 *' 173+71 9+ 2." 151+ 9 14+ 3*' 161+ 9

!i8+ 9 " 1214+ 15 47+ 5 *' 121 + 15 150+ 5 ' 146+ 7 58+ 10 ~ 121+ 5 15+ 2 # 184+ 7 20+ 9 # 143+ 8

1(19+ 3* 139+ 6 58+12 *' 137+ 10 109+ 7.'* 171+ 7 38+ 9 *' 133+ I0 17+ I ~ 130+25 27+ 5*" 146+16

167+12"* 193+ 8 ** 1(19+17 ** 136+ 7 170+!5" 192+16 65+ 10 * 130+22 19+ l*" 116+14 30+ 6 * 139+ 9

* P < 0,()1 compared to controls; P < 0,01 compared to unlesioned side of the same animals; * P < 0.01 compared to the same side of 15-day treated animals,

.,

31

•~

.. '~, ' R ~ , ,

"

'

¢ "~,

.

.

""

¢ "

""

" "

~

~

'

'

:"

~4-.

,

•. • .,\

-~

.

"~,

. . , , , ,

-~

•

t

.

~ ~ ,~:~"/~"~'"

-=~." .

.

~

"-

""-'-

"

:'.~-,,,:'~..',r'.~,'~_-x4~"

~

.

q~'

.

,

,

~.~'

""

,'."~

.~'e~.

-,

,

'

:'.

.

,

~2.;,i:,~ ,,,,~~:~,. ~ ~ ~.: 'N:,.., " ,:;:~~'-~.

~..,;. ;~.:~ ..

."

~,. ~ , ' . ~

.'~.:'C

•

•

.

,.;---.~' ¥~:- 4 " ~ ,,:

, . ~ .'..:.,'..

&

7,~,. .

,..

I.

"

..

,x~e~

.

,

•

• .-_~-~-.~."

~ - ~ .

,

,r

y.

"

...,

..,., ~ ,.,

.

•

,

"

"

..~..,. . ' ~ •

~

~_ - ~

,,.. ,' ,. ~

.

.~ . ,...,'-"

., .~,

,~r

-~.-"

..

,/.:

, ,

..,.

,.

Fig. 4. Micrographs of AChE-positive fibers in the CAI region on the lesioned side of the dorsal hippocampus, Note marked increases in the AChE-positive fibers in both 15-day (A) and 2-month (B) NGF-treated animals compared to those in 15-day (C) and 2-month (D) cytochrome c-treated animals, Bar - 10 ttm,

predominant in some regions, such as tangentially oriented fibers in the molecular layer and radially oriented fibers in the stratum radiatum. On the lesioned side of the 15-day cytochrome ctreated animals, there was a dramatic decrease in the AChE-positive fiber density while compared to the unlesioned side (Table liD. The decrease in the ventral hippocampal formation was greater than that ila the dorsal hippocampal formation. Of all examined regions of the hippocampal formation, including the CA1 and CA3 regions and the dentate gyrus, the decrease was greatest in the CA3 region, which almost lacked the AChE-positive fibers. There was a significant recovery of the AChE-positive fibers in the CA1 region and dentate gyrus of 2-month cytochrome c-treated animals, when compared to the 15-day cytochrome ctreated animals (Figs. 4 and 5; Table lit). On the lesioned side, a significant increase in AChE-positive fiber densities was observed in the CA1

region and dentate ~rus of the dorsal hippocampal formation of the 15-day NGF-treated animals, relative to the controls. Two.month NOF treatment resulted in a substantial increase in those two regions and in a significant increase of AChE-positive fibers in the CA3 region of the dorsal hippocampus. Although there was a tendency toward an increase in the AChE.positive fiber density, such an increase did not reach significance in all regions of the ventral hippocampai formation of both the 15-day and 2-month NGF-treated animals, respectively relative to the controls (Fig. 5, Table Ill)). AChE.positive

f i b e r s in t h e f i m b r i a

The AChE-positive fibers on the unlesioned side of the fimbria were regularly arranged toward the hippocampus (Fig. 6A). A complete lack of AChE-positive fibers was shown on the lesioned side of the fimbria distal to the lesion site. There was not a bit of recovery

32 200 DH

DISCUSSION

150

--

NGF promotes growth of cholinergic neurites of the septohippocampal system

s

~,/i~

100

Several earlier studies provided evidences for the regenerative capacity of cholinergic nerve after experimental lesion in the adult rat brain ~'7"~2. After transplanting fetal hippocampal tissue into the site of timbria transection, transected axons do not only innervate the transplant but also the host dorsal hippocampUS 20'2t'2~'29. Several studies s'a° have reported that intraventricular NGF adminstrations result in a dramatic increase in ChAT-IR fibers in the ipsilateral septum and a large accumulation of AChE staining in the dorsal lateral quadrant of the ipsilateral septum. Hefti et al. reported recently that the action of long-term NGF administration on neurite growth is restricted to limited growth close to the lesion site u~. Those studies have only provided some incomplete evidence for the neurite growth promoted by exogenous NGF administration. In the present study, 2-month NGF treatment resulted in a significant increase of AChE-positive fibers in the lateral septum. The increase on the lesioned side was larger than that on the unlesioned side. The reason for the difference may be that the eholinergic neurite growth is influenced by gradients of NGF because NGF was injected on the lesioned side, The fact that NGF treatment resulted in the cholinergic neurite growth on the unlesioned side shows that NGF may act on the cholinergic neurons in the septum contralateral to the injected side. The exact source of the regenerating fibers is not known. They may be

II1_

r///

50

i1~ //f

j J

IIJ

j

-. -. -. • ~ / ,

iii I " I" I "

,

T

I I I / / /

. . . .

iii

o

.

//,4

"/14

IO0 VH

50

Isd

2m

,I,sd

2m

¢A!

¢A3

Fig. 5. Effects of the 15-day (15d) and 2-manth (2m) NGF treatments (hatched bars)on AChE-positive fiber densities in the CAI and CA3 regions of the dorsal (DID and ventral (Wit) hippocampus. , . @. P < ().(Nil, compared to those of cytochrome c treatments (open bars). P < 0.01. compared between the 15-day and 2.month cytochrome c treatments. The ordinate represents the number ( + S.E.M.) of intersects of fibers with the square lattice system p~r IO.(XX) pm:' real area,

from the complete lack of AChE-positive fibers even in the animals with the 2-month NGF treatment (Fig. 6B). There also was not any emergence of AChE-posi= tive fibers in the lesion site.

• "

~:'-"7-'~'~

•

~

,

t~'.F.-o

p

. .

~,dt

~ ,,_,-.

,,,'" .%

"""

..

•

"

e''

r

,....~.....

• "

,

u , ,-. • .i •

•

. . . .

e"lIb,

•

-

g,

.pq~ . q B ~ ,

dP"

17.

• B

,: •

-

~. o

Fig. 6. Micrographs of AChE-positive fibers in the fimbda distal to the lesion site. Note rich AChE-positive fibers in the unlesioned side of the fimbria (A) and a complete lack of AChE-positive fibers in the lesioned side of the fimbria (B) of 2-months NGF-treated animal. Bar = 20 p.m.

33 from intact axons in the septum contralateral to the lesion and from intact axons and the proximal segment of the lesioned axons in the septum ipsilateral to the lesion. The local cholinergic neurite growth limited directly by gradients of NGF was reported by other investigators ~6a~'26'3°. The present study showed not only the local cholinergic neurite growth, but also a more extensive cholinergic neurite growth, which took place at the distal segment of intact axons adjacent to the denervated field, such as those in the dorsal fornix, and at the axons and terminals remaining in the denervated field, such as those in the dorsal hippocampus.

Mechanism of growth of the cholinergic neurites promoted by NGF As mentioned above, several studies provide some evidence for the cholinergic neurite growth promoted by NGF, but it is not known at present whether this cholinergic neurite growth represents true regeneration of the severed cholinergic axons or collateral sprouting from intact axons ~+'. Regeneration of the severed cholinergic axons is a desirable way because it is probable for the regeneration to reestablish a homogeneous innervation of the severed cholinergic neurons 2. Although present in the peripheral nervous system aad in the CNS of lower vertebrates, the true regeneration of axons has long been regarded as impossible in the mature mammalian CNS. But, there is now evidence that true axonal regeneration can occur under especially favorable conditions ~7'~"'2'~.in the present study, there was no evidence that the cholinergic neurite growth promoted by NGF administration was tho true rogeneration of the severed axons. After the fimbria, via which septohippocampal cholinergic fibers end diffusely throughout the hippocampus, was transected, AChE-positive fibers were completely lacking on the lesioned side of fimbria distal to the transected site and had not even slightly recovered even in animals with 2.month NGF treatment. Thus the finding rules out the possibility of the true regeneration of severed axons through their original path. The increase of AChE.positive fibers promoted by NGF has .to be attributed to the result of collateral sprouting of the intact cholinergic axons. In the present study, the NGF-induced increase of AChE-positive fibers in the lateral septum emerged 2 months after the lesion and that in the dorsal fornix and dorsal hippocampus after the lesion. This shows that the collateral sprouting stimulated by NGF takes place more easily in the denervated field. In addition, there was a significant increase of the AChE-positive fibers in the dorsal hippocampus of the control animals 2-months after the transection, compared to the 15-day control animals.

The recovery phenomena after lesion have been , served and considered as collateral sprouting response by other investigators+''7. Some studies showed that sympathetic sprouting was stimulated by some trophic factors, such as NGF, from the denervated target 2.27, which attract and stimulate the reinnervation of the collateral sprouts of intact cholinergic axons. Thus, the greater capacity of the collateral sprouting stimulated by exogenous NGF in the denervated and adjacent field may be attributed to involvement of the endogenous NGF.

Tile source of the cholinergic fibers into the &'nerrated hippocampus Many studies have reported the recovery of choiinergic marker in the denervated hippocampus after lesion and NGF treatments 3'7's'~2'~4'~6. Recovery of cholinergic fibers in the denervated hippocampus was also observed by some investigators ~6'2'~. However, the source of the fibers in the denervated hippocampus has not been known clearly. A compensatory collateral sprouting by the ventral septohippocampal pathway was suggested by some investigatorszs'~2. It is surprising that the recovery of AChE-positive fibers in the ventral hippocampus was not observed in our experiment. The reason for this discrepancy is not clear. The above hypothesis was suggested mainly according to analysis of the neurochemical results. It is possible that there is a discrepancy between the morphologic and neurochemical results. On the present data, we suggest 3 possible sources of the fibers reinnervating the denervated hippocampus: (I) the collateral sprouting of cholinergic axons, which come via the dorsal fornix pathway, remained in the hippocampus (number I in Fig, 7); (2) the collateral sprouting of cholinergic axons in the dorsal fornix, which cross the hippocampal commissure into the denervated hippocampus (number 2 in Fig. 7); (3) long-distance growth of the collateral sprouting of the lesioned and unlesioned cholinergic axons in the septum (number 3 in Fig. 7). it has been confirmed that the septohippocampal projection innervates the dorsal hippocampus via the dorsal fornix and the whole hippocampus via the fimbria 1'~'2'~. After the fimbria pathway was transected in the present animal models, there was less loss of AChE.positive fibers in the dorsal hippocampus than that in the ventral hippocampus, especially in the CA3 region, in which there were few AChE-positive fibers. A significant recovery of AChE-positive fibers was observed in the dorsal hippocampus rather than in the ventral hippocampus of the NGF-treated animals. This shows that the fibers that remained in the denervated hippocampus are essential to the reinnervation by the cholinergic collateral

34 dene~'ated hippocampus. If the long-distance growth of the collateral sprouting is possible, the denervated hippocampus may obtain homotypic reinnervation by the collateral sprouts of the segment of the severed axons, which move around rather than across the lesion site. The gradual increase in AChE-positive fibers observed in the CA3 region of the dorsal hippocampus shows a tendency toward the long-distance growth. T h e present study demonstrates that the exogenous N G F treatments to aged rats enhance reinnervation of the denervated hippocampus. The results lead to the prediction that exogenous N G F treatment may promote collateral sprouting of the remaining cholinergic axons to improve the cholinergic function of the basal forebrain in AIzheimer's disease.

//. /I

Acknowk,dgements. We thank Mrs. Xie Yao and Mrs. Yuan Ounfang far technical assistance, Mrs. Zhang Xiaoping far assistance with figures. The project was supported by National Natural Science Foundation of China.

REFERENCES

MSDB ~" , Fig. 7. A summaryof possihle mechanisms of the cholinergic neuritc growth (broken line) promoted by exogenous NGF. It is impossible fi~r the severed axons in the fimbria to regenerate across the lesion site (cross), The collateral sprouting stimulated by exogenous NGF enay take place at the dencrvated dorst,I hippocampus (I), the intact dorsal fl~rnix(2) and the lateral septum (3), Thus, 3 possible soe,rces of th~ fibers relnnervating the denervated hippoeampus were sag. gested: (I) the collateral sprouting of cholinerglc axons (I) which come via the intact dorsal fornix pathw:eyremained in the hippocam. pits; (2) the collateral sproutintl el' cholinert~ic axons in the dorsal fornix (2)i (3) Ionl;.distancc growth of the collateral sprouting of the lesioned (arrow) and unlesioned (not shown) axons in the septum. LS, lateral septum: LV, lateral ventricle: FF, fornix-fimbria; DF, dorsal fornix; DH, dorsal hippocampus; VH, ventral hippocampus: MSDB, medial septum and diagonal band of Broca.

sprouting promoted by NGF. Thus, the collateral sprouting of the remaining cholinergic axons is the most possible source of the cholinergic reinnervation of the denervated hippocampus. A marked increase in the AChE-positive fiber number in the dorsal fornix of the NGF-treated animals shows that there is abundant collateral sprouting of the cholinergic axons in the dorsal fornix. A large emergence of AChE-positive fibers in the hippocamual commisure of the NGFtreated animals is direct evidence for the collateral sprouts innervating the denervated hippocampus. A marked increase of AChE-positive fibers stimulated by exogenous NGF was observed in the lateral septum, it is not known whether the collateral sprouts stimulated by NGF can grow for a long distance to reach the

I Armstrong, D.M., Terry, R.D., Deteresa, R.M., Bruce, G., Hersh, LB. and Gage, F.H., Response of septal cholinergic neurons to axotomy, J. Comp. Neurol., 264 (19~,7)421-436. 2 Crutcher, K.A., Sympathetic sprouting in the central nervous system: a model for studies of axonal growth in the mature mammalian brain, Brain Res. Rev., 12 (1987)203-233. 3 Daitz. H.M. and Powell. T.P.S., Studies on the connections of the fornix syst0m. J. Neuro~ Neurosur¢ Psychiatry, 17 (1954) 75-82, 4 Fusee. M.. Oderfeld-Nowak, B., Vantini, G., Schiaw), N., Gradkowska. M., Zar0mba, M. and Leon, A., Nerve growth llJctor affects uninjured, adult rat septohippocamp,l cholinergic neu, tons, M.uroscience, 33 ( If)Hg) 47-52, Gage, F.ll., Armstrong, D,M,, Williams, L.R. and Varon, S,, Morphological response of axotomized septal n~urons to nerv,., ~rowth I'actor, J, ('¢mq~, Nt,uro/,, 269 (I~)HH) 147-155, ('~ Gage, F.H., Wictorin, K,, Fischer, W., Williams, L,R,, Veron, S, and BjOrklund, A., Life tend death of cholinergic neurons: in the septal and diagonal band region following complete fimbria fornix transection, Nemx~scit,nce, 19 (1986) 241-256, 7 Gage, F.H., Bj6rkhmd, A. and Stenevi, U., Cells of origin of the ventral cholinergic septohippocampal pathway undergoing compensatory collateral sprouting following fibria-fornix transection, Nt.ur~xwi, Lt,tt,, 44 (1984) 21 !-216, 8 Gage, F,H,, Bjgrklund, A,, Stenevi, U, and Dunnett, S,B,, Functional correlates of compensatory collateral sprouting by aminergic tend cholinergie afferents in the hippocampal formation, Brain Rt,s,, 268 (1983) 39-47. 9 Hagg, 1",, Manthorpe, M., Vahlsing, H,L, and Varon, S., Delayed treatment with nerve growth factor reverses the apparent loss of cholinergic neurons after acute brain damage, Exp, Neurol., 101 (1988) 31)3~312, 10 Hagg, T,, Hagg, F,, Vahlsing, H,L,, Manthorpe M. and Varon S., Nerve growth factor effects on cholinergic neurons of neostria, turn and nucleus accumbens in the adult rat, Neumscience, 30 (1989) 95-103, i I itagg, T., Fass-Holmes, B., Vahlsing, H.L.. Manthorpe, M., Conner, J,M, and Varon, S,, Nerve growth factor (NGF) reverses axotomy-induced decreases in choline acetyltransferase, NGF receptor and size of medial septum cholinergic neurons, Brain Res,, 505 (1989) 29-38. 12 He, Y,S,, Yao, Z,B, and Chen, Y.C,, Effect of nerve growth factor on the lesioned septohippocampal cholinergic system of aged rats, Brain Res., 552 (1991) 159-163.

35 13 Hedreen, J.C., Bacon, S. and Price, D.L., A modified histochemical technique to visualize acetylcholinesterase-containing axons, J. Histochem. Cytochem., 33 (1985) 134-140. 14 Hefti, F., Dravid, A. and Hartikka, J., Chronic intraventricular injections of nerve growth factor elevate hippocampal choline acetyitransferase activity in adult rats with septo-hippocampal lesions, Brain Res., 293 (1984) 305-311. 15 Hefti, F., Nerve growth factor (NGF) promotes survival of septal cholinergic neurons after fimbriai transections, J. Neurosci., 6 (1986) 2155-2162. 16 Junard, E.O., Montero, C.N. and Hefti, F., Long-term administration of mouse nerve growth factor to adult rats with partial lesions of the cholinergic septohippocampal pathway, Exp. Neurol., !10 (1990) 25-38. 17 Kao, C.C., Comparison of healing process in transected spinal cords grafted with autogenous brain tissue, sciatic nerve, and nodose ganglion, Exp. Neurol., 44 (1974) 424-439. 18 Kao, C.C., Chang, L.W. and Blood-Worth, J.M.B., Axonal regeneration across transected mammalian spinal cords: an electron microscopic study of delayed microsurgical nerve grafting, .Exp. Neurol., 54 (1977) 591-615. 19 Kromer, L.F., Nerve growth factor treatment after brain injury prevents neuronal death, Science, 235 (1987) 214-216. 20 Kromer, L.F., Bj6rklund, A. and Stenevi, U., Regeneration of the septo-hippocampal pathways in adult rats is promoted by utilizing embryonic hippocampal implants as bridges, Brain Res., 210 (1981) 173-200. 21 Kromer, L.F., Bjiirklund, A. and Stenevi, U., Intracephalic embryonic neural implants in the adult rat brain, l. Growth and mature organization of brain stem, cerebellar, and hippocampal implants, J. Comp. Neurol., 218 (1983) 433-459. 22 Milner, T.A, and Amaral, D.G., Evidence for a ventral septal projection to the hippocampal formation of the rat, EXp. Brain Res., 55 (1983) 579-585.

23 Milner, T.A., Loy, R. and Amaral, D.G., An anatomical study of the development of the septo-hippocampal projection in the rat, De~:. Brain Res., 8 (1983) 343-371. 24 Montero, C.N. and Hefti, F., Because of lesioned septal cholinergic neurons by nerve growth factor: specificity and requirement for chronic treatment, J. Neurosci., 8 (1988) 2986-2999. 25 Montero, C.N. and Hefti, F., lntraventricular nerve growth factor administration prevents lesion-induced loss'of septal cholinergic neurons in aging rats, Neurobiol. Aging, 10 (1989) 739-743. 26 Rosenberg, M.B., Friedmann, T., Robertson, R.C. et al., Grafting genetically modified cells to the damaged brain: restorative effects of NGF expression, Science, 242 (1988) 1575-1578. 27 Stanfield, B.B. and Cowan, W.M., The sprouting of septal afferents to the dentate gyrus after lesions of the entorhinal cortex in adults rats, Brain Res., 232 (1982) 162-167. 28 Svendgaard, N.-A., Bj6rklund, A. and Stencvi, U., Regeneration of central cholinergic neurones in the adult rat brain, Brain Res., 102 (1976) 1-22. 29 Tuszynski, M.H., Buzsaki, G. and Gage, F.H., Nerve growth factor infusions combined with fetal hippocampal grafts enhance reconstruction of the lesioned septohippocampal projection, Neuroscience, 36 (1990) 33-44. 30 Williams, L.R., Varon, S., Peterson, G.M., Wictorin, K., Fischer, W., Bjgrklund, A. and Gage, F.H., Continuous infusion of nerve growth factor prevents basal forcbrain neuronal death after timbria fornix transection, Proc. Natl. Acad. ScL U.S.A., 83 (1986) 9231-9235. 31 Williams, L.R., Jodelis, K.S. and Donald M.R., Axotomy-dcpendent stimulation of choline acetyltransferase activity by exogenous mouse nerve growth factor in adult rat basal forebrain, Brain Res., 498 (1989) 243-256.