Proc. Nail. Acad. Sci. USAVol. 86, pp. 2056-2060, March 1989Neurobiology

Gangliosides potentiate in vivo and in vitro effects of nerve growthfactor on central cholinergic neurons

(trophic factors/neural plasticity/neuronal degeneration/septal cell culture/nucleus basalis)

A. C. CUELLO*, L. GAROFALO, R. L. KENIGSBERG, AND D. MAYSINGERDepartment of Pharmacology and Therapeutics, McGill University, 3655 Drummond Street, Montreal, PQ, Canada H3G 1Y6

Communicated by Rita Levi-Montalcini, September 30, 1988

ABSTRACT The effects of nerve growth factor (3(-NGF)and ganglioside GM1 on forebrain cholinergic neurons wereexamined in vivo and in vitro. Following unilateral decortica-tion of rats, GM1 (5 mg/kg per day) administered intracere-broventricularly could protect forebrain cholinergic neurons ofthe nucleus basalis magnocellularis from retrograde degener-ation in a manner comparable to P-NGF. Administered incombination with (-NGF, GM1 produced a significant increasein choline acetyltransferase activity in the nucleus basalismagnocellularis and remaining cortex ipsilateral to the lesion.Concentrations of GM1 that were ineffective when adminis-tered alone in this lesion model, when given with .3-NGF,potentiated (i-NGF effects in both of the above brain areas. Indissociated septal cells in vitro, an increase in choline acetyl-transferase activity was noted at fi-NGF concentrations as lowas 0.1 pM and reached a plateau at 1 nM. A moderate (up to35%) stimulation of choline acetyltransferase activity wasobserved with 10 ,uM GM1. The application of 13-NGF incombination with 10 ,uM GM1 or 0.1 ,uM GM1, a concentra-tion that is ineffective in these cultures, produced a muchgreater increase in choline acetyltransferase activity than did18-NGF alone. These observations support the idea that exog-enously applied gangliosides can elicit trophic responses incholinergic neurons of the central nervous system. That GM1increases and even potentiates 13-NGF effects suggests thatsome of the trophic actions of this compound may be mediatedthrough endogenous trophic factors.

Nerve growth factor ,3 (P3-NGF) can be considered theprototype for substances exerting in vivo and in vitro trophiceffects on defined cells of the nervous system (1). In additionto its peripheral actions, j3-NGF acts on subsets of centrallylocated neurons (2, 3). Central forebrain cholinergic neuronscontain B3-NGF binding sites, and their terminal targetsproduce the trophic peptide, which can be transportedretrogradely to cell bodies ofthese cholinergic neurons (4). Inthe adult, these cholinergic neurons respond to exogenous,B-NGF after partial or total damage of the septal-hippocam-pal connections (5-7). In addition, /3-NGF has been found toaffect forebrain cholinergic neurons in vitro (8, 9).

Sialogangliosides, in particular GM1, exert trophic-likeactivity, both in vivo and in vitro, and resemble "bona fide"trophic factors in many ways (for review, see ref. 10). Whenapplied in vivo, they promote the anterograde regeneration ofacetylcholinesterase-reactive fibers in the hippocampus afterpartial fimbria transections (11). Administration of GM1 alsoprevents the retrograde cell shrinkage of cholinergic neuronsof the nucleus basalis magnocellularis (NBM) that followscortical infarction (12), as well as cell death in the medialseptum after unilateral hippocampal ablation (13). Thesefindings have prompted us to investigate the possible in vivo

and in vitro interactions between a selective and a specifictrophic factor (,B-NGF) and a biological substance (GM1)whose apparent trophic activity is less well defined. Ourpreliminary observations in the in vivo model (14) and theresults presented here strongly suggest that the gangliosideGM1 has an enabling or potentiating effect on P-NGF-mediated responses in central cholinergic neurons.

MATERIAL AND METHODSCortical Lesions and Drug Treatment. Male Wistar rats

(Charles River Breeding Laboratories, 300-350 g) were usedin these experiments and were subjected to a unilateralleft-side cortical devascularizing lesion as described (15). Thesurgical procedure produced a limited, well-defined infarc-tion of the neocortex without affecting underlying brainstructures (15).A group of lesioned rats (n = 5) received GM1 (0.5 or 5

mg/kg per day) intracerebroventricularly (i.c.v.) for 1 weekthrough a stainless steel cannula (23-gauge) permanentlyimplanted into the right lateral ventricle [coordinates fromBregma (16): AP, -0.8; L, 1.4; V, 3.5]. The cannula wasconnected by flexible polyethylene tubing to a subcutane-ously implanted osmotic minipump (Alzet 2001, Alza). An-other lesioned group (n = 5) received ,-NGF (12 ,ug per day)alone or in combination with GM1 (0.5 or 5 mg/kg per day)in the same manner. A group of lesioned rats (n = 6) receivedphysiological saline (0.9% NaCl) containing 0.1% bovineserum albumin; Sigma), and rats that had not undergonesurgery (n = 6) served as controls. Animals were decapitated30 days after the surgical procedure (i.e., 23 days aftertreatment with f3-NGF and/or GM1 ceased). Discrete brainareas were microdissected from fresh tissue slices as de-scribed (15).

Biochemical Analysis of Microdissected Tissue. Microdis-sected tissues were kept frozen at -80°C until required.Choline acetyltransferase (ChAT) activity was determined bya radiometric assay described by Fonnum (17), and proteincontent was determined by the Bradford method (18).Immunohistochemical Analysis of Brain Tissue. Three ad-

ditional rats from each group were anesthetized and perfusedas described (19). Sections (50 ,um) of the entire NBM wereobtained from frozen brain tissue blocks by using a slidingmicrotome (Reichert). Sections were then processed free-floating for ChAT immunocytochemistry (19).

Cell Culture. The septal area was dissected from brain ofday-17 embryonic rat fetus (Sprague-Dawley) as describedby Dunnett et al. (20). Dissociated septal cells were preparedby a modification ofa published method (21). Viable cell yieldwas determined by the trypan blue-exclusion test. Thedissociated cells were resuspended in culture medium to a

Abbreviations: ANOVA, analysis of variance; ChAT, choline ace-tyltransferase; NBM, nucleus basalis magnocellularis; f3-NGF,nerve growth factor /3; i.c.v., intracerebroventricular(ly).*To whom reprint requests should be addressed.

2056

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

June

24,

202

1

Proc. Natl. Acad. Sci. USA 86 (1989) 2057

final cell density of 0.6 x 106 viable cells per ml and seeded(1 ml of cell suspension per well) on poly(L-lysine) (Sigma)-coated culture wells (16 mm diameter) in 24-well plates(Falcon). One rat embryo yielded on average 0.5-0.75 x 106viable septal cells. Exposure of the cells to either 13-NGF orGM1 began immediately after seeding. Cultures were main-tained at 370C in a humidified incubator in a 95% air/5% CO2atmosphere. The cells were inspected regularly and mediumwas replenished every 3 days. Data were analyzed byanalysis of variance (ANOVA) and/or two-tailed t test.

Biochemical Determinations of Tissue Culture Preparations.The cells in culture were washed free of medium withCa2+/Mg2+-free phosphate-buffered saline (PBS, pH 7.4),collected in a total volume of 100 gl of homogenization buffer(10 mM EDTA/0.5% Triton X-100, pH 7.4), and homoge-nized on ice in a glass homogenizer. Aliquots of the homog-enate were taken for determination ofChAT activity (17) andprotein concentration (Bio-Rad protein assay).Immunocytochemistry of Tissue Culture Preparations. Cul-

tures were washed free ofmedium with Ca2+/Mg2+-free PBS,fixed for 20 min at room temperature in 4% paraformalde-hyde, and then washed four or five times with PBS/0.02%Triton X-100 before incubation with the primary antibodies[anti-glial fibrillary acidic protein, anti-neurofilament, anti-ChAT, anti-NGF receptor (22, 36)]. Immunocytochemicalstaining was obtained by adapting previously describedmethods (22) to our tissue culture conditions.

RESULTSIn Vivo Studies. These studies confirm that ChAT activity

decreases significantly in the ipsilateral NBM of mature rats30 days following a unilateral devascularizing lesion of theneocortex. The previously described (23) retrograde cellshrinkage and loss of neurites of these forebrain cholinergicneurons were clearly apparent in NBM sections from le-sioned rats (compare Fig. 1 a and b). As previously reported(15), when compared with the control group, no significantchanges in ChAT activity were found in any other microdis-sected brain areas (data not shown).The i.c.v. administration of 13-NGF (12 ,tg per day), for 7

days from the onset of the lesion, prevented a decrease inChAT activity in the NBM after partial cortical infarction.The magnitude of this protective effect was shown to becomparable to that obtained with the i.c.v. administration ofGM1 alone (5 mg/kg per day) (Table 1, experiment A). Thecombined administration of P-NGF and GM1 in the partiallydecorticated animals increased ChAT activity in the NBM,ipsilateral to the lesion, above control levels. Immunocyto-chemical analysis revealed not only full protection of thecholinergic neurons from retrograde cell shrinkage and loss ofdendritic extensions but also an apparent increase in thenumber of ChAT-immunoreactive processes in the neuropil(Fig. lc).

In this series of experiments ChAT activity in the remain-ing cortex of the lesioned animals was found to be similar tothat of the unlesioned side. Furthermore, treatment with,B-NGF orGM1 alone increased ChAT activity ofthe lesionedside over that of control. A more noticeable cooperativitybetween these two factors was observed in the remainingneocortex ipsilateral to the lesion, where the combinedtreatment brought ChAT activity to 237% of control values(Table 1).

In cortically lesioned animals treated with low doses ofGM1 (0.5 mg/kg per day, i.c.v., 7 days), a significantdecrease (32% of control values) in ChAT activity wasobserved in the NBM ipsilateral to the lesion (Table 1,experiment B). However, if these subthreshold amounts ofganglioside were administered concurrently with effectivedoses of 8-NGF to lesioned animals, a significant increase

-.4pp~~~~."Af->~

WAe tF~~~~~~~~,L

.A

a 4~.*%.

r

-.,

b.4

I .7

*s4

4 ':

Fr'p,

/

I

.4-P'i_Iss

V.ip4 ..,' 4

#Lro

FIG. 1. Appearance of ChAT-immunoreactive neurons in NBMin control (a), lesioned (b), and lesioned, GM1/p-NGF-treated (c)rats. Clustered (asterisks) and isolated (arrows) cholinergic cellbodies are indicated. Thinner, paired arrows in c indicate immuno-reactive processes. Note that cell shrinkage is prevented in factor-treated rats. (Interference contrast optics; bar = 25 ,um.)

(21%) in ChAT activity was observed in the area of theaffected forebrain cholinergic neurons. Even more remark-able were the changes observed in ChAT activity in theipsilateral cortex of the lesioned animals after simultaneousi.c.v. administration of subthreshold amounts of GM1 andeffective doses of 3-NGF. While this low dosage of GM1 perse did not alter ChAT activity in the remaining ipsilateralcortex, in combination with f3-NGF it produced a markedincrease of the ChAT enzymatic activity.In Vitro Studies. Dissociated septal cells in culture repre-

sent a mixed neuronal-glial cell population as determinedimmunocytochemically with the use of antiserum to glialfibrillary acidic protein and antiserum to neurofilament pro-

Neurobiology: Cuello et al.

Dow

nloa

ded

by g

uest

on

June

24,

202

1

Proc. Natl. Acad. Sci. USA 86 (1989)

Table 1. Effect of P-NGF administered for 7 days in combination with an effective (5 mg/kg perday; experiment A) or an ineffective (0.5 mg/kg per day; experiment B) dose of GM1 on ChATactivity in the NBM and cortex of mature rats, 30 days after unilateral decortication

Ipsilateral NBM Ipsilateral cortex

Group n ChAT activity % control ChAT activity % controlExperiment A

Control 6 57.67 ± 3.86 35.81 ± 2.39Lesion plus vehicle 6 31.16 ± 3.17 54* 35.85 ± 1.74 100Lesion plus GM1 5 61.94 ± 6.55 107 50.70 ± 2.44 142*Lesion plus f3-NGF 5 50.94 ± 3.75 88 47.63 ± 3.12 132*Lesion plus GM1 plus ,-NGF 5 69.41 ± 1.06 120* 84.82 ± 10.42 237*

Experiment BControl 6 69.06 ± 4.67 39.20 ± 3.77Lesion plus vehicle 6 44.87 ± 6.60 65* 38.20 ± 4.69 97Lesion plus GM1 5 46.92 ± 2.80 68* 36.93 ± 2.80 94Lesion plus /3-NGF 5 73.07 ± 3.30 109 59.06 ± 2.90 151*Lesion plus GM1 plus f3-NGF 5 83.87 ± 6.56 121* 72.98 ± 4.08 186*

Tissues were obtained 30 days after lesion (i.e., 23 days after treatment with GM1 and/or (3-NGFceased). Values for ChAT activity are means ± SEM and are expressed as nmol per mg of protein perhr; n indicates number of rats.*Significantly different from control at P < 0.01 (ANOVA followed by a post-hoc Dunnett's test).

tein (data not shown). The presence of ChAT-immunoreac-tive neurons in this culture system was confirmed as well(Fig. 2). Although immunoreactive sites for the 3-NGFreceptor were found on cells with neuronal appearance (Fig.2), a number of non-neuronal cells were found to possessthese receptors. Upon exposure of these cells in culture toeither p-NGF or GM1, ChAT activity was subject to modu-lation. ChAT activity in the septal cell cultures was dramat-ically increased after a 7-day exposure to exogenous /3-NGF.As observed by Hatanaka and Tsukui (9) this increase inenzyme activity was dose-related. In our experiments, theincrease in ChAT activity was detectable at f3-NGF concen-trations as low as 0.1 pM and was maximal at 1 nM. Exposureof the cells to nanomolar concentrations of ,B-NGF increasedChAT enzymatic activity to 200% of control levels (Fig. 3).GM1, on the other hand, produced a less dramatic increase

in ChAT activity. Enzyme induction was detected only whencells were exposed to 10 ,tM GM1 and was increased overcontrol by some 15-30%. Lower concentrations of GM1(e.g., 0.1 ,uM) were found to be ineffective when appliedalone to these cultures (Fig. 4).When effective concentrations ofGM1 (10 ,uM) were added

in combination with ,3-NGF, we observed a clear potentiationof the effects of this trophic factor. In combination with 10,M GM1, submaximally (0.1 pM) and maximally (1 nM)effective concentrations of S-NGF produced an increase in

- !*4 .io-'%%7'

.. :.OL.

J

- a

d.I :14

ip

ChAT enzymatic activity that was significantly greater thanthat obtained with j3-NGF alone. This potentiating effect ofGM1 was more dramatic when GM1 was applied in combi-nation with 0.1 pM B-NGF. We found that even 0.1 ,uM GM1,which was ineffective when applied alone, potentiated the,-NGF-induced increase in ChAT activity (Figs. 3 and 4).

DISCUSSIONThe trophic effects of p-NGF on central and peripheralneurons are in all likelihood receptor-mediated. Althoughthere is still no definite information on the cellular andmolecular events that take place following the interaction ofP-NGF with its receptor, several mechanisms have beenpostulated (for review, see ref. 24). The interaction of theganglioside with endogenous trophic factors, and with ,B-NGF in particular, could take place at any level. Neverthe-less, it is likely that this interaction occurs at the level of thecell membrane where GM1 is incorporated (25). In thisregard, it is interesting that immobilized GM1 is capable ofbinding P-NGF with low affinity (26). Thus, gangliosidescould provide additional binding sites for growth factors or,on the other hand, may modify the growth factor receptorstate, as is the case for the ganglioside GD2 and thevitronectin receptor (27).

,44

.. '.4. 0.A to*

1# Jerj.

/3

.7X7l

FIG. 2. Detection of ChAT (a)- and NGF receptor (b)-immunoreactive cells in septal cell culture. Note the similarity in incidence and inmorphology of ChAT- and NGF receptor-immunoreactive cells. Single arrows indicate immunoreactive cell bodies, and paired arrows indicateneurites. (Interference contrast optics; bar = 20 ,um.)

*.,,

II

AP-'_ -6

I

__0

*

I

wvp,lK.._

2058 Neurobiology: Cuello et al.

--w

I

04

T I

-W A

Dow

nloa

ded

by g

uest

on

June

24,

202

1

Proc. Natl. Acad. Sci. USA 86 (1989) 2059

0M-.z0

05

C)L

100

1-5 -1 11 -9 -710-5 10i 10 10 10C GM1 NGF(3 +10-5GM1 ) M

FIG. 3. Effect of 10 ,um GM1, alone or in combination withvarious concentrations of /3-NGF, on ChAT activity in cultures ofdissociated septal cells. Septal cells were grown for 7 days in theabsence of GM1 and 83-NGF [control (C), hatched bar], in thepresence of 10 AuM GM1 (cross-hatched bar), or in the presence of0.1pM, 10 pM, 1 nM, or 0.1 AM 18-NGF alone (stippled bars) or incombination with 10 AtM GM1 (open bars). Bars represent the means± SEM from quadruplicate culture wells from sister culture prepa-rations. Control absolute value was 4.5 nmol of acetylcholine per mgof protein per hr.

Certain conditions are required for gangliosides to show atrophic effect in vivo or in vitro. These have been referred toas "permissive conditions" for the in vivo effects (19) or a"window of opportunity" for the in vitro effects.t It isconceivable that under certain conditions the availability ofendogenous trophic factors is affected, and consequently theability of cells to respond to these factors is influenced. In thein vivo experimental model, the early initiation of gangliosidetreatment is essential (19). This is in agreement with theevidence that, in response to injury, the brain produces lowamounts of endogenous trophic factors immediately after theinsult (e.g., ref. 28). Therefore, in instances of extensiveneural lesions, a situation of extreme vulnerability mightarise. In such a case, irreversible anterograde and retrogradecellular damage would occur. In the case of central cholin-ergic neurons, retrograde degenerative changes can be in partreversed by the timely administration of P-NGF (5-7). Inview of the above, it can be proposed that in the present invivo situation the administration of gangliosides preventsanterograde and retrograde neuronal degeneration by poten-tiating the action of the limited quantity of endogenoustrophic factors produced in the first few days subsequent tothe lesion. The difficulties encountered in rescuing choliner-gic neurons by administration of exogenous gangliosides inaged rats (19) could be explained in the same way, since agingis accompanied by an apparent loss of f3-NGF receptors (29)

200

-I0c-z0O 150-aR

I-I-

0

I-0cu

100

I

-I-

I

....

I .1

10-7 10-13 10-11 109 10-7

C GM1 NGF 0I (+1IF7GMI M1)FIG. 4. Effect of a subthreshold concentration (0.1 ,uM) of GM1,

alone or in combination with various concentrations of 13-NGF, onChAT activity in cultures ofdissociated septal cells. Septal cells weregrown for 7 days in the absence of GM1 and f8-NGF [control (C),hatched bar], in the presence of 0.1 ,uM GM1 (cross-hatched bar), orin the presence of 0.1 pM, 10 pM, 1 nM, or 0.1 LM ,p-NGF alone(stippled bars) or in combination with 0.1 ,.M GM1 (open bars). Barsrepresent the means ± SEM from quadruplicate culture wells fromsister culture preparations. Control absolute value was 4.5 nmol ofacetylcholine per mg of protein per hr.

and a diminished production of endogenous factors afterinjury (30).The idea that gangliosides act cooperatively with .-NGF in

the present in vivo model is reinforced by the observationsthat the expression of j-NGF is increased in the target areasof the basal forebrain cholinergic neurons following mechan-ical lesions (for review, see ref. 30). Although 8-NGFapparently does not act as a trophic agent for dopaminergicneurons, the reported actions of GM1 over the nigrostriatalsystem (31) might be similarly explained. For this, theidentification of a specific trophic factor(s) that acts on thissubset of central neurons would be necessary.The effects of ,B-NGF and/or GM1 in the cortex suggest

that their administration may provoke an important reorga-nization of the cholinergic fibers of the remaining neocortex.Whether this is due to an increased production of thebiosynthetic enzyme or due to sprouting of cholinergicterminals is not known.The concept that gangliosides potentiate ,B-NGF-mediated

effects on cholinergic neurons is supported by in vitrostudies, in which the exogenous concentrations of the twofactors can be accurately controlled. In other cell culturesystems, the trophic actions of gangliosides are dependentupon the presence of p-NGF (32). Although GM1 does exertsome trophic action when administered alone in our in vitromodel, we have established that there is also an importantinteraction between specific (,B-NGF) and nonspecific (GM1)factors on cholinergic markers. It is interesting that botheffective and subthreshold concentrations of the ganglioside,when applied in combination with B-NGF, produced supra-maximal responses in ChAT activity, consistent with the invivo observations in the cerebral cortex. Although Hefti et al.

tVaron, S., Skaper, S. D. & Katoh-Semba, R., International Societyfor Neurochemistry, Satellite Symposium on Neuronal Plasticityand Gangliosides, May 29-31, 1985, Mantova, Italy, Abstr. 13, p.19.

Neurobiology: Cuello et al.

Dow

nloa

ded

by g

uest

on

June

24,

202

1

Proc. Natl. Acad. Sci. USA 86 (1989)

(33) reported that gangliosides did not potentiate the effectsof P-NGF on septal cells in culture, their experiments differedfrom ours, as they employed a mixture of gangliosides ratherthan pure GM1 and a single, high concentration off-NGF. Wehave observed that the synergism between GM1 and P-NGFoccurs at submaximal concentrations of P-NGF. However,the molecular mechanisms underlying these interactions havenot yet been properly examined. Nevertheless, it is possiblethat the contribution of glial cells is important both in vivo andin vitro. There is evidence that p-NGF is produced by glialcells (34). In vivo, neural damage results in reactive gliosis, aphenomenon that may contribute to increased availability ofendogenous trophic factors.

Validation of the ganglioside-trophic factor cooperativityhypothesis will require further investigations ofthe molecularmechanisms underlying their interactions in the central andperipheral nervous systems. There is already evidence of thecooperativity of P-NGF and gangliosides in the peripheralnervous system (35). The investigation of the interaction ofgangliosides with j3-NGF and other trophic factors mayprovide valuable insight for the establishment of therapeuticregimes in neurodegenerative diseases.

We dedicate this paper to the memory ofMr. Manuel Madanes. Wethank Drs. R. Levi-Montalcini, L. Aloe, and W. Mushynski forvaluable advice and materials. We thank J. Seguin for secretarialassistance and A. Foster for photography. This work was supportedby the Medical Research Council (Canada) and, in part, by FIDIALaboratories (Italy). L.G. received a Fonds de la Recherche Scien-tifique (Quebec) studentship, and R.L.K., a postdoctoral fellowship(Medicorp, Canada).

1. Levi-Montalcini, R. (1987) Science 237, 1154-1162.2. Levi-Montalcini, R. 4 Aloe, L. (1985) Proc. Nadl. Acad. Sci.

USA 82, 7111-7115.3. Korsching, S. (1986) Trends Neurosci. 9, 570-573.4. Theonen, H., J3andtlow, C. & Heumann, R. (1987) Rev.

Physiol. Biochem. Pharmacol. 109, 145-178.5. Hefti, F. (1986) J. Neurosci. 6, 2155-2162.6. Williams, L., Varon, S., Peterson, G., Wictorin, K., Fischer,

W., Bjorklund, A. & Gage, F. (1986) Proc. Nail. Acad. Sci.USA 83, 9231-9235.

7. Kromer, L. (1987) Science 235, 214-216.8. Hefti, F., Hartikka, J., Eckenstein, F., Gnahn, H., Heumann,

R. & Schwab, M. E. (1985) Neuroscience 14, 55-68.9. Hatanaka, H. & Tsukui, H. (1986) Dev. Brain Res. 30, 47-56.

10. Ledeen, R. W. (1985) Trends Neurosci. 8, 169-174.

11. Oderfeld-Nowak, B., Skup, M., Ulas, J., Jezierska, M., Grad-knowska, R. & Zaremba, M. (1984) J. Neurosci. Res. 12, 409-420.

12. Cuello, A. C., Stephens, P. H., Tagari, P. C., Sofroniew,M. V. & Pearson, R. C. A. (1986) Brain Res. 376, 373-377.

13. Sofroniew, M. V., Pearson, R. C. A., Cuello, A. C., Tagari,P. C. & Stephens, P. H. (1986) Brain Res. 378, 393-396.

14. Cuello, A. C., Maysinger, D., Garofalo, L., Tagari, P.,Stephens, P. H., Pioro, E. & Piotte, M. (1987) in Receptor-Receptor Interaction, eds. Fuxe, K. & Agnati, L. F. (Mac-millan, New York), pp. 62-77.

15. Stephens, P. H., Cuello, A. C., Sofroniew, M. V., Pearson,R. C. A. & Tagari, P. C. (1985) J. Neurochem. 45, 1021-1026.

16. Paxinos, G. & Watson, C. (1986) in The RatBrain in StereotaxicCoordinates (Academic, London), 2nd Ed.

17. Fonnum, F. (1975) J. Neurochem. 24, 407-409.18. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254.19. Stephens, P. H., Tagari, P. C., Garofalo, L., Maysinger, D.,

Piotte, M. & Cuello, A. C. (1987) Neurosci. Lett. 80, 80-84.20. Dunnett, S. B., Wishaw, I. Q., Bunch, S. T. & Fine, A. (1986)

Brain Res. 378, 357-373.21. Barbin, G., Selak, I., Manthorpe, M. & Varon, S. (1984)

Neuroscience 12, 33-43.22. Eckenstein, F. & Thoenen, H. (1982) EMBO J. 1, 363-368.23. Sofroniew, M. V., Pearson, R. C. A., Eckenstein, F., Cuello,

A. C. & Powell, R. (1983) Brain Res. 289, 370-374.24. Levi-Montalcini, R. & Calissano, P. (1986) Trends Neurosci. 9,

473-477.25. Toffano, G., Benvegnu, A., Bonetti, A., Facci, L., Leon, A.,

Orlando, F., Ghidoni, R. & Tettamanti, G. (1980) J. Neuro-chem. 35, 861-866.

26. Schwartz, M. & Spirman, N. (1982) Proc. Natl. Acad. Sci. USA79, 6080-6083.

27. Cheresh, D. A., Pytela, R., Pierschbacher, D., Klier, F. G.,Ruoslahti, E. & Reisfeld, R. A. (1987) J. Cell Biol. 105,1163-1173.

28. Nieto-Sampedro, M., Manthorpe, M., Barbin, G., Varon, S. &Cotman, C. W. (1983) J. Neurosci. 3, 2219-2229.

29. Koh, S. & Loy, R. (1988) Brain Res. 440, 396-401.30. Whittemore, S. R. & Seiger, A. (1987) Brain Res. Rev. 12, 439-

464.31. Toffano, G., Savioni, G., Moroni, F., Lombardi, G., Calza, L.

& Agnati, L. F. (1984) Brain Res. 296, 233-239.32. Doherty, P., Dickson, J. G., Flanigan, T. P. & Walsh, F. S.

(1985) J. Neurochem. 44, 1259-1265.33. Hefti, F., Hartikka, J. & Frick, W. (1985) J. Neurosci. 5, 2086-

2094.34. Assouline, J. G., Bosch, P., Lim, R., Kim, I. S., Jensen, R. &

Pantazis, N. J. (1987) Dev. Brain Res. 31, 103-118.35. Vantini, G., Fusco, M., Bigon, E. & Leon, A. (1988) Brain Res.

448, 252-258.36. Chandler, C. E., Parsons, L., Hosang, M. & Shooter, E. M.

(1984) J. Biol. Chem. 259, 6882-6889.

2060 Neurobiology: CueHo et al.

Dow

nloa

ded

by g

uest

on

June

24,

202

1