Mechanisms of Ageing and Development

123 (2002) 499–509

Postnatal development of rat dentate gyrus: effects ofmethylazoxymethanol administration

Sandra Ciaroni a,*, Tiziana Cecchini a, Paola Ferri a, Patrizia Ambrogini b,Riccardo Cuppini b, Gabriella Lombardelli c, Guidubaldo Peruzzi c,

Paolo Del Grande a

a Institute of Morphological Sciences, Uni�ersity of Urbino, loc. Crocicchia, I-61029 Urbino(PS), Italyb Institute of Physiological Sciences, Uni�ersity of Urbino, Urbino, Italy

c Institute of Pharmacology and Pharmacognosy, Uni�ersity of Urbino, Urbino, Italy

Received 9 February 2001; received in revised form 11 June 2001; accepted 27 July 2001

Abstract

In order to investigate the role of postnatal neurogenesis in granule cell number control in the rat dentate gyrus,we administered Methylazoxymethanol (MAM), a drug able to prevent cells from dividing, on P3, P5, P7, P9, whenthe most granule cells are produced. The effect of MAM on the number of proliferating precursors and of granulecells was examined at P16 and P90. We used 5-bromo-2�-deoxyuridine administration to label proliferating cells andimmunohistochemistry to characterize the cell phenotype using neuron markers TUC 4, PSA-NCAM, CalbindinD28K and glial marker GFAP. At 16 days of age in MAM-treated rats we observed a significant decrease ofBrdU-positive cells. Consistently, a decrease in density and number of granule cells was found compared to thecontrols. At 90 days the dentate gyrus of treated rats showed a complete recovery: no differences in the density, totalnumber of neurons, the BrdU- and TUC 4-positive cells were revealed with respect to the controls. No deficits wereevident in performance on the water maze in MAM-treated rats. These data suggest that the dentate gyrus is able tore-establish the proliferative zone and to rebuild the granule cell layer following neonatal MAM administration.© 2002 Elsevier Science Ireland Ltd. All rights reserved.

Keywords: Postnatal neurogenesis; Dentate gyrus; Neuronal precursors; Methylazoxymethanol

www.elsevier.com/locate/mechagedev

1. Introduction

The processes controlling neuron number dur-ing development have important effects on brain

function and consequently on behaviour(Williams and Herrup, 1988). The rodent dentategyrus formation begins in the embryonic periodand a second phase of neurogenesis occurs in theperinatal and juvenile periods, when a large num-ber of granule cells is produced in the tertiarydentate matrix. Finally, the neurogenesis contin-ues in the subgranular zone throughout adult life

* Corresponding author. Tel.: +39-0722-304270; fax: +39-0722-304243.

E-mail address: s.ciaroni@uniurb.it (S. Ciaroni).

0047-6374/02/$ - see front matter © 2002 Elsevier Science Ireland Ltd. All rights reserved.

PII: S 0 0 47 -6374 (01 )00359 -1

S. Ciaroni et al. / Mechanisms of Ageing and De�elopment 123 (2002) 499–509500

(Altman and Bayer, 1990a,b; Kuhn et al., 1996).Recent studies on the estimation of neuron num-ber in the dentate gyrus indicate a genetic controlunderlying the differences in neuron number ob-served in different strains of laboratory mice(Kempermann et al., 1997a; Abusaad et al., 1999).The nature, fate and survival of cells originatingfrom postnatal neurogenesis within the rodentdentate gyrus have been well studied in severalworks carried out during the past three decades.Some studies focused on the unusual feature ofthe dentate gyrus where, during the perinatal pe-riod, the granule cells undergo massive cell deathcoincident with or immediately preceding cell pro-liferation (Schlessinger et al., 1975). The closerelationship between these events was pointed outin several experimental conditions where both cellproliferation and death were affected in the samemanner (Gould et al., 1994; Cameron and Gould,1996) and where the neuronal death had a role indetermining the number of proliferating cells(Gould and Tanapat, 1997). Therefore, these find-ings raise the hypothesis that both cell prolifera-tion and death may be involved in the processwhich maintains a basal number of neurons.

Methylazoxymethanol is an alkylating agentthat prevents the proliferation of neuroblastswithout affecting either glial cells or quiescentcells (Cattaneo et al., 1995). It appears to exert itseffects on those cells undergoing their final mito-sis, and it is widely used to induce neuronalablation during the period of active neurogenesis(Haddad et al., 1972). In particular, repeated ad-ministrations of MAM during early perinatal ageproduce permanent reductions in cerebellar andolfactory bulb weight, whereas the hippocampusis able to compensate for neonatal MAM inducedlesion, increasing its own weight until it reachesthat of control values. In this study the authorspresume that the hippocampal compensation maybe related to the dentate gyrus capacity for prolif-eration, but they did not demonstrate it (Sullivan-Jones et al., 1994). Considering the unusualfeature of adult dentate gyrus neurogenesis, thiscapacity for recovery is not surprising and in anattempt to identify the variable responsible forthis recovery we examined, in the present study,the relationship between the drug-induced inabil-

ity of precursor cells to divide, and the prolifera-tion, migration, differentiation and reconstitutionof neuron number in adulthood.

2. Materials and methods

2.1. Animals

Sprague–Dawley male rats were used. All ani-mal use was conducted in accordance with theEuropean Union guidelines and with Italian laws.

On the day of birth P1, the rats were divided intwo experimental groups: MAM-treated (MAM)and controls (C). Rat pups were injected subcuta-neously into the neck area twice a day for 4non-consecutive days on P3, P5, P7, P9 witheither 4 mg/kg MAM or saline. The injectionswere given at 9:00 and 18:00 h. The site of injec-tion was rotated over the 4 days of treatment toinclude the left and right flank and dorsal areas.This protocol was chosen so as to administerdoses of MAM when the proliferative peaks weretaking place, and to induce a decrease in precur-sor cells without lethality. Within each litter, halfthe pups were treated with MAM and the otherswere treated with saline; within each treatmentgroup, half the rats were killed at P16, the othersat P90.

2.2. Histological procedure

The animals were intraperitoneally injectedwith BrdU (50 mg/Kg b.w.) twice a day (8.30 and18:30 h) over 3 consecutive days before death. Allanimals from each group were killed the day afterthe last BrdU administration. All the rats usedwere anaesthetized using sodium thiopental viai.p. (45 mg/Kg b.w.) and killed with an intracar-diac injection of the same anaesthetic. The brainwas then removed, dissected into two hemi-spheres, fixed by Carnoy for 48 h and embeddedin Histovax (Histo-lab. Ltd., Goteborg, Sweden;melting point=56–58 °C).

Serial coronal sections from the left hemisphereof each brain were cut at a thickness of 6 �m.Three, 1-in-20 series of sections (120 �m spacing)were mounted in polylysine coated slides for

S. Ciaroni et al. / Mechanisms of Ageing and De�elopment 123 (2002) 499–509 501

staining with cresyl-violet. Closely previousslides were processed for BrdU and closely con-secutive slides for TUC 4 labeling detection;other slides were randomly selected for GFAP,PSA-NCAM and Calbindin D28K evidence anddouble immunodetection for BrdU and TUC 4.Each analysis was performed in the whole den-tate gyrus.

2.3. Quantitati�e analysis in cresyl-�iolet staining

The estimation of the granule cell layer(GCL) volume, mean nucleus size of granulecells, mean cell density per mm3 and total num-ber were performed using a Leitz microscopeequipped with a camera lucida drawing tube.

To estimate GCL volume, one section every20 slices, i.e. every 120 �m, was selected and theoutlines of the granule cell layer were drawnwith a low-power objective lens (4× ). The areaswere calculated by means of OPTILAB softwarefor image analysis, and the total volume foreach hemisphere was estimated as the sum ofthe areas multiplied by the distance between theanalyzed sections.

The granule cell size was measured in all ex-perimental groups to obtain values to carry outstereological correction. To calculate the meangranule cell nucleus size OPTILAB software wasused after drawing at least 150 neurons for eachanimal at the final magnification of 985× .

To estimate the mean granule cell density permm3, one section every 40 slices, i.e. every 240�m, was selected. Granule nuclei were countedin 5–6 fields randomly chosen within both thetwo blades of the dentate gyrus, and cell densityin each animal was obtained by the formula:

d= (n/a)×K2,

where n is the total number of plotted nuclei, ais the total area of the sampling field and K isthe linear magnification of the camera lucidaprojection (625× ). The granule cell density permm3 was:

D=d×166.67,

where 166.67 is the number of the sections in-cluded in 1 mm of thickness. The value was

corrected following Abercrombie’s formula(Abercrombie, 1946), in which mean diametersof granular nuclei of each rat were used.

Total number of granule cells was estimatedas follows:

N=�i(Vi×Di),

where Vi is the volume of a 240-�m-wide zoneof dentate gyrus included between two sectionswhere the surface area was calculated, and Di isthe corrected granule cell density per mm3 deter-mined in the same zone.

2.4. Immunohistochemical staining

The detection of BrdU-labeled cells was car-ried out by an indirect immunohistochemicaltechnique. Rehydrated sections were treated with0.1 M HCl at 4 °C for 10 min, with 4 M HClat 37 °C for 30 min, and rinsed with borate-buffered saline (0.1 M, pH 8.5) for DNA denat-uration. After rinsing in 1% bovine albuminserum in phosphate buffered saline (PBS; 0.1 M,pH 7.4), sections were incubated with normalhorse serum (DBA, Vector, 1:10 in PBS) for 10min, then they were incubated with a mouseanti-BrdU primary antibody (DBA; 1:300 inPBS) overnight at 4 °C, followed by incubationwith FITC-conjugated horse anti-mouse (DBA,Vector, 1:50 in PBS) secondary antibody.

BrdU-positive nuclei were counted in theGCL in at least 18 non-consecutive sections:very small-sized nuclei were discarded from thecount, as it was considered that they could beglial nuclei. In infant rats BrdU-labeled cellswere subgrouped following a topographic crite-rion into: (a) the tertiary dentate matrix in thebasal polymorph layer; (b) the granule cell layer.In adult rats BrdU-labeled cells found in thesubgranular zone were considered together withthose in the rest of the GCL. The density ofBrdU-labeled cells per mm3 was evaluated asfollows: the GCL areas were measured in cresyl-violet stained sections closely consecutive tothose where BrdU-positive cells were counted,using OPTILAB software for image analysis.The total number in the whole dentate gyrus

S. Ciaroni et al. / Mechanisms of Ageing and De�elopment 123 (2002) 499–509502

was also calculated by multiplying the densityvalue by the volume of the dentate gyrus in theleft hemisphere.

The distribution of GFAP-positive cells wasdetected by incubating sections with normalsheep serum (Sigma, 1:10 in PBS) for 10 minand then with a rabbit anti-GFAP (Sigma, 1:100in PBS) overnight at 4 °C, and finally with aCY3-conjugated sheep anti-rabbit (Sigma, 1:100in PBS) for 1 h at room temperature.

Both in the animals of 16 days and those of90 days of age the expression of TUC 4, a tran-sient antigen expressed after the last mitotic di-vision and before the overt differentiation(Minturn et al., 1995) was investigated. So, sec-tions were incubated with normal goat serum1:10 in PBS for 10 min and then incubated witha rabbit primary antibody against TUC 4 (S.Hockfield’s gift; 1:5000 in PBS). After rinsing inPBS, sections were incubated with a CY3-conju-gated goat anti-rabbit secondary antibody(DBA, Vector, 1:500 in PBS) for 1 h at roomtemperature. To characterize proliferated cells inthe rats of 90 days of age, TUC 4 and BrdUwere shown together in the same sections. TheTUC 4-positive cells were revealed as describedabove and then sections were post-fixed with 4%paraformaldheyde in phosphate buffer 0.1 M for1 h; subsequently, the DNA denaturation wascarried out and the BrdU was evidenced as ex-plained above.

TUC 4-positive cells were counted in theGCL in at least 18 non-consecutive sections andthe density was evaluated: the same GCL areasselected for BrdU-positive cell density were used.Moreover, the total number in the GCL wascalculated in infant and adult rats.

The expression of PSA-NCAM, the embry-onic form of neuronal adhesion molecule (Sekiand Arai, 1993) was performed: after beingtreated with normal horse serum (DBA, Vector,1:10 in PBS) for 10 min, sections were incubatedwith a mouse anti-PSA-NCAM (R. Gerardy-Schanhn’s gift, 1:1000 in PBS) overnight at 4 °Cand then with a FITC-conjugated horse anti-mouse secondary antibody (DBA, Vector, 1:50in PBS) for 1 h at room temperature.

Mature granule cells were revealed by the de-tection of Calbindin D28K, a binding Ca2+

protein, used as a marker of fully differentiatedgranule cells (Sloviter, 1989). After incubationwith normal goat serum 1:10 in PBS for 10 min,sections were incubated with a rabbit anti-Cal-bindin D28K primary antibody (SWant, 1:1000in PBS) overnight at 4 °C, followed by a CY3-conjugated goat anti-rabbit secondary antibody(DBA, Vector, 1:500 in PBS).

To observe FITC fluorescence, a fluorescencemicroscope (VANOX, Olympus Italia s.r.l., MI)was used with a combination of BP 490 and EY455 excitation filters. To view CY3 fluorescencea BP 545 excitation filter was used.

2.5. Beha�ioral test

To reveal if MAM injections caused perma-nent learning impairment, the experiment wasperformed a week before the death of the ratskilled at P90.

The behavioral testing was carried out in awater-filled maze, as described by Giurgea andMouravieff-Lasuisse (1972). The maze was arectangular gray tank (120×50×40 cm3) withseveral vertical barriers inside. The tank wasfilled to a depth of 24 cm with the water main-tained at 14 °C. The entrance was illuminatedwith a 25 W bulb mounted 30 cm above themaze. The ladder at the opposite end was in-clined at a 45° angle. Behavioral testing con-sisted of two trials per day for 4 days. In eachtrial the animals were individually placed at theentrance of the maze and in order to escapefrom the maze they were required to swimaround the barriers until the exit ladder at theopposite end was reached. The numbers of er-rors (turning in the wrong direction, entry intodead-end compartments) were recorded and to-taled for each animal.

2.6. Statistic

Values are expressed as mean�SD; compari-son was made with the non-parametric Mann–Whitney U-test; the threshold of significancewas fixed at P�0.05.

S. Ciaroni et al. / Mechanisms of Ageing and De�elopment 123 (2002) 499–509 503

3. Results

3.1. Quantitati�e analysis in cresyl-�iolet staining

The rodent dentate gyrus consists of the outermolecular layer, the granule cell layer and thepolymorph layer, well distinguishable in 16 and90-day-old controls. At 16 days the dentate gyrushas not yet its mature appearance. The granulecells appeared in the inner part of GCL, smallerand darker than the lightly stained and roundedmature granule cells located in the outer portionof the granular layer. The mean nucleus sizes were73.0�3.3 �m2 in infant control rats and 74.2�3.1 �m2 in infant MAM rats. In adult rats thegranular nuclei measured 92.4�5.9 �m2 in con-trol rats and 99.3�10.8 �m2 in treated rats.

No evidence of alterations in the cytoarchitec-tonic organization of the dentate gyrus was ob-served in MAM-treated rats with respect tocontrols.

Quantitative analysis (Table 1) of dentate gran-ule cells at 16 days of age revealed that the densityand the total number were significantly lower intreated rats when compared with controls(Mann–Whitney U-test: P�0.05), consistentwith the finding that MAM prevents precursorproliferation (Cattaneo et al., 1995). In accor-

dance with previous reports (Boss et al., 1985;Ciaroni et al., 1999), the total number, from 16 to90 days, did not change in control rats. On thecontrary, in the same lifespan, the MAM-treatedgroup exhibited a significant increase in the totalgranule number (Mann–Whitney U-test: P�0.05). At 90 days of age, the density and the totalnumber of the granule cells caught the controlvalues up.

Moreover, the volume of the granule cell layerand the granule cell density per mm3 showed thesame trend in the two experimental groups: thevolume increased, and the cell density decreasedwith age although in MAM rats this decrease wasless marked than in controls (Mann–Whitney U-test: P�0.05).

3.2. Immunohistochemical analysis

The data from cresyl-violet stained sections in-dicated that the MAM injection caused the elimi-nation of dentate granule cells. Therefore, wedetermined its effect on granule cell neurogenesisusing: BrdU administration to label proliferatingcells, detection of GFAP to show glial cells, themarkers TUC 4 and PSA-NCAM for early post-mitotic and migrating neurons, and CalbindinD28K for differentiated neurons.

Table 1Quantitative analysis of granule cell layer in control and MAM-treated rats

Adult ratsVolume (mm3) Density per Total number Total numberVolume (mm3) Density perInfant ratsmm3mm3

697 884 787 214C 5 C 11.128 1.445 486 553 703 0701.487C 2645 897 670 604679 8920.950C 6 450 978

1.000 706 234 706 234 C 4C 7 1.483 431 547 639 984

694 670 713 115 MeanMean 1.472b1.026 456 359b 671 21913 462 70 909 SD 0.023SD 16 1050.092 31 548

1.490MAM 1642 477 631 608632 3591.016MAM 8 423 8980.940 597 970 562 092 MAM 2 1.423 453 229 644 945MAM 90.862 657 446 566 718 MAM 4MAM 10 1.437 468 187 674 222

629 259a 650 258b448 771b1.450bMean0.939 590 429aMean0.077 29 859 45 134 SD 0.035 22 530 21 798SD

a P�0.05, Mann–Whitney U-test with respect to C.b P�0.05, Mann–Whitney U-test with respect to the infant rats.

S. Ciaroni et al. / Mechanisms of Ageing and De�elopment 123 (2002) 499–509504

Fig. 1. Immunohistochemical staining of coronal sections of the dentate gyrus of 16-day-old MAM-treated rat. BrdU-labeled cellsin granule cell layer and basal polymorph layer (A). GFAP immunoreactivity appears in molecular and polymorph layers (B). TUC4-positive granule cells in the inner half of granule cell layer (C). PSA-NCAM-positive granule cells in the inner part of the granulecell layer when the two blades are joint (D). Bar represents 10 �m.

3.2.1. Infant ratsAt 16 days of age, in both control and treated

rats, BrdU immunostaining revealed numerouslabeled cells distributed in the GCL and in thebasal polymorph layer; in this latter area cellsrepresent the tertiary dentate matrix which re-duces to form the subgranular zone, a persistentsource of granule cell production during adult life(Fig. 1A; Fig. 2A). No differences on BrdU-label-ing pattern were detected in either experimentalgroups. At this stage the MAM rats showed adecreased density of BrdU-positive nuclei both inthe granule cell layer and in the basal polymorphlayer in comparison to controls.

In all experimental groups GFAP-immunos-taining was distributed almost exclusively in theplexiform and polymorph layers and showed thesame appearance, according to Eriksdotter-Nilsonet al. (1986) who did not find immunoreactivitychanges by MAM. In the granular layer GFAP-

positive cell bodies were not observed (Fig. 1B).TUC 4-positive cells were found in the inner

half of the granule cell layer (Fig. 1C). The samedistribution observed for TUC 4-labeled cells wasfound for PSA-NCAM, which is expressed tran-siently by newly generated granule cells (Fig. 1D)(Seki and Arai, 1993).

As expected, in the infant rats the CalbindinD28K-labeled cells were located in the upper partof the GCL, while the inner part was not labeled(Fig. 3A). This strip lacking in labeling representsan area of precursors and immature granule cellswhich, in cresyl-violet stained sections, appearsmaller and darker than the lightly stained androunded mature granule cells (Fig. 3B–E). Thethickness of the whole GCL showed a reductionof 17% in MAM-treated rats and it was due to aproportional decrease of both Calbindin D28K-positive and -negative areas, in agreement withantiproliferative activity of MAM.

S. Ciaroni et al. / Mechanisms of Ageing and De�elopment 123 (2002) 499–509 505

3.2.2. Adult ratsIn the GCL no significant difference was

found in the density of BrdU-positive cells ofcontrols or treated animals: they were located ina zone comprised between the cell layer of GCLfacing hilus and a two cell body sized layer be-low and often organized in clusters (Fig. 2B,Fig. 4A).

Both in control and treated rats a strongerGFAP-immunoreactivity than in infant rats wasdetected in the dentate gyrus: numerous im-munoreactive fibers crossing the GCL and fol-lowing a course perpendicular to the orientationof the GCL were seen. Heavy labeling of cellbodies was clearly visible outside the GCL (Fig.4B).

Combining BrdU-labeling with immunohisto-chemistry for TUC 4, we established the neu-ronal identity of proliferated cells in the dentategyrus of control and MAM rats: about 20% ofBrdU-positive cells expressed TUC 4 (Fig. 4C,D). All TUC 4-labeled cells in the dentate gyrusof control and treated rats were observed in theinner rows of the granular layer (Fig. 2B; Fig.4E). The expression of PSA-NCAM and TUC 4decreased markedly with respect to infant rats(Fig. 4F).

In the adult rats the immunostaining for Cal-bindin D28K delineated the granule cell layerfrom the molecular and polymorph layers (Fig.3F).

3.3. Beha�ioral test

In adult controls the number of errors pro-gressively decreased throughout the trials andreached a relatively low level since the fourthtrial. A similar time course was found in adultMAM rats showing that no impairment inlearning ability occurred in this group (Fig. 5).

4. Discussion

The production of the granule cells begins inthe prenatal period (E16), from stem cells inextrinsic matrix areas, which will form the skele-ton of the two blades of dentate gyrus, and con-tinues in the postnatal life by the transient tertiarydentate matrix establishing in the basal poly-morph layer from P2. The tertiary dentate matrixis the principal source of cells contributing to thegranular layer between P3 and P10. From P13 thetertiary dentate matrix reduces and is graduallyreplaced by the subgranular zone, by the 1stmonth of life the neuronal precursors reside in thesubgranular zone. The subgranular zone isthought to be mainly responsible for the produc-tion of granule cells throughout juvenile and adultperiods (Altman and Bayer 1990a,b; Kuhn et al.,1996). At 16 days the proliferating cells werefound in the polymorph layer, where they repre-sented the reducing tertiary dentate matrix and

Fig. 2. Quantitative analysis of labeled cells in the dentate gyrus. (A) 16-day-old rats, number of BrdU-positive cells in the GCL andthe tertiary dentate matrix (DGT) in the whole dentate gyrus of one hemisphere, where it is significantly decreased in MAM-treatedrats (Mann–Whitney U-test: P�0.05). (B) 90-day-old rats, total number of BrdU-positive and TUC 4-positive cells in GCL of thewhole dentate gyrus. No significant difference was found in MAM-treated with respect to control rats.

S. Ciaroni et al. / Mechanisms of Ageing and De�elopment 123 (2002) 499–509506

Fig. 3. Calbindin D28K immunoreactivity (A, B, D, F) and cresyl-violet (C, E) staining of coronal sections of the rat dentate gyrus.Calbindin D28K-positive cells in the upper part of GCL of 16-day-old control rat (A, B), of 16-day-old MAM-treated rat (D) and90-day-old MAM-treated rat (F). MAM administration gives rise to an increase in the cell packing density. Bar represents 50 �m(A), 10 �m (B–F).

increasing subgranular zone. The results of thisstudy indicate that the decrease in number ofproliferating cells in the dentate gyrus followingMAM injection in early postnatal days is notpermanent. The dentate granule cell layer in ourstudy showed a compensatory plastic response,while different effects of MAM on the morpho-logical characteristics of the cerebellar cortex werefound (Sullivan-Jones et al., 1994; Ferguson et al.,1996). This drug permanently reduces the numberof cerebellar granule cells both in the mouse andrat (Chen and Hillmann, 1986) although a partialregeneration of the external germinal layer hadbeen described (Lovell and Jones, 1980). The dif-

ferent response found in the dentate gyrus may berelated to the prolonged postnatal neurogenesisoccurring in this region throughout adulthood.Consistent with the schedule of neurogenesis, ourdata show that the MAM administration from P3to P9, when the tertiary dentate matrix is promi-nent and the majority of granule cells originate,results in a significant decrease in proliferatingprecursors labeled with BrdU in animals of 16days of age. This lower number of BrdU-positivecells gives rise to a decrease of immature neuronsin the inner granule cell layer and a significantdecrease in density and total number of granuleneurons.

S. Ciaroni et al. / Mechanisms of Ageing and De�elopment 123 (2002) 499–509 507

Fig. 4. Neurogenesis in the dentate gyrus of adult MAM-treated rats: cluster of BrdU-positive cells (A). GFAP-immunoreactivity:numerous fibers cross the GCL where GFAP-positive cell bodies are not observed (B). Double immunolabeling for BrdU-positivecells (C) and for the neural specific early postmitotic marker TUC 4 (D) confirms the neuronal identity of proliferating cells.Distribution of TUC 4-positive cells (E) and PSA-NCAM-positive cells (F). Bar represents 10 �m.

In control rats the total number of granularneurons appears to remain constant from 16 to 90days; whereas, in treated rats the number in-creases, reaching that found in controls at 90days. This increase is paralleled by a quantitativerecovery of proliferating stem cells that probablysupports it. Moreover, stem cells differentiate intoneurons as is shown by the occurrence of TUC4-positive cells in the same quantity as that ofcontrols at this time. One may speculate thatMAM kills proliferating cells and that from 16 to90 days the spared stem cells re-establish theproliferative zone. Moreover, the reconstituted

stem cells produce a number of neurons to rebuildthe whole structure of granule cell layer.

The two reparative processes must be separatelyconsidered from a conceptual and from a mecha-nistic point of view. The re-establishment of theproliferative zone suggests that there is a mecha-nism able to maintain the number of stem cells ata constant level throughout life. The rebuilding ofthe granule cell layer suggests the existence of amechanism stabilizing the number of granule cells.Neurogenesis is the result of a balance between‘redundant’ proliferation of neuron precursorsand newborn cell death occurring precociously

S. Ciaroni et al. / Mechanisms of Ageing and De�elopment 123 (2002) 499–509508

Fig. 5. Learning test in water-maze by Giurgea andMouravieff-Lasuisse in adult rats. The number of errors isrepresented throughout training. The error is defined as turn-ing in the wrong direction, entry into dead-end compartments.No significant difference was found in MAM-treated withrespect to control rats.

1996), and even if it is unlikely that this alone isresponsible for the survival of granular neurons, itmay contribute to survival control.

Neuron precursors undergo mitosis, newborncells migrate and both mature and immature adultgenerated cells extend their axons, can formsynapses and are incorporated into the hippocampalcircuitry, as recently described in retrograde labelingstudies (Hastings and Gould, 1999; Markakis andGage, 1999). It has been proposed that massiveneuron cell death during development reflects thefailure of these neurons to obtain the necessaryneurotrophic factors produced by the target cellsand required for neuron survival (Raff et al., 1993).

MAM injection and the possible elimination ofmany precursors could be responsible for the rescueof newly generated neurons, through an enhancedproduction of trophic factor(s) by new generatedneuron targets that are poorly innervated becauseof the reduced number of granule cells.

Acknowledgements

We thank Drs S. Hockfield for providing TUC4 antibody, and R. Gerardy-Schanhn for providingPSA-NCAM antibody. We also thank Dr J. Parentfor the useful technical advice and Mr S. Cecchinifor digital processing of imaging.

References

Abercrombie, M., 1946. Estimation of nuclear populationfrom microtome sections. Anat. Rec. 94, 239–247.

Abusaad, I., MacKay, D., Zhao, J., Stanford, P., Collier,D.A., Everall, I.P., 1999. Stereological estimation of thetotal number of neurons in the murine hippocampus usingthe optical dissector. J. Comp. Neurol. 408, 560–566.

Altman, J., Bayer, S.A., 1990a. Mosaic organization of thehippocampal neuroepithelium and the multiple germinalsources of dentate granule cells. J. Comp. Neurol. 301,325–342.

Altman, J., Bayer, S.A., 1990b. Migration and distribution oftwo populations of hippocampal granule cell precursorsduring the perinatal and postnatal periods. J. Comp. Neu-rol. 301, 365–381.

Boss, B.D., Peterson, G.M., Cowan, W.M., 1985. On thenumber of neurons in the dentate gyrus of the rat. BrainRes. 338, 144–150.

Cameron, H.A., Gould, E., 1996. Distinct populations of cellsin the adult dentate gyrus undergo mitosis or apoptosis inresponse to adrenalectomy. J. Comp. Neurol. 369, 56–63.

after the last cell division (Cameron et al., 1993;Okano et al., 1996; Gould et al., 1999). Therefore,the stabilizing mechanism may, in principle, affectcell birth or death. In particular newborn cellsurvival is susceptible to being affected by neuro-trophic factors: both BDNF and GDNF were foundto promote survival and to modulate hippocampalplasticity (Lindvall et al., 1994), and environmentalenrichment, which stimulates postnatal neurogene-sis and facilitates neuronal survival (Kempermannet al., 1997b), induces an increase in these neuro-trophic factors’ immunoreactivity (Young et al.,1999). Other growth factors are involved in neuro-genesis control (Ray et al., 1993; Gage et al., 1995;Palmer et al., 1995; Cameron et al., 1998) such asbFGF which is elevated after physical activity(Gomez-Pinilla et al., 1997, 1998): the physicalexercise doubles the total number of survivingnewborn cells in the dentate gyrus (van Praag et al.,1999). It is intriguing to note that MAM rats, inthe present study, showed a mild hyperactivity inseveral assessments. A mild hyperactivity has alwaysbeen observed in MAM animals (Ferguson et al.,

S. Ciaroni et al. / Mechanisms of Ageing and De�elopment 123 (2002) 499–509 509

Cameron, H.A., Woolley, C.S., McEwen, B.S., Gould, E., 1993.Differentiation of newly born neurons and glia in the dentategyrus of the adult rat. Neuroscience 56, 337–344.

Cameron, H.A., Hazel, T.G., McKay, R.D.G., 1998. Regulationof neurogenesis by growth factors and neurotransmitters. J.Neurobiol. 36, 287–306.

Cattaneo, E., Reinach, B., Caputi, A., Cattabeni, F., Di Luca,M., 1995. Selective in vitro blockade of neuroepithelial cellsproliferation by methylazoxymethanol, a molecule capableof inducing long lasting function impairments. J. Neurosci.Res. 41, 640–647.

Chen, S., Hillmann, D.E., 1986. Selective ablation of neuronsby methylazoxymethanol during pre and postnatal braindevelopment. Exp. Neurol. 94, 103–119.

Ciaroni, S., Cuppini, R., Cecchini, T., Ferri, P., Ambrogini, P.,Cuppini, C., Del Grande, P., 1999. Neurogenesis in the adultrat dentate gyrus is enhanced by vitamin E deficiency. J.Comp. Neurol. 411, 495–502.

Eriksdotter-Nilson, M., Jonsson, G., Dahal, D., Bjorklund, M.,1986. Astroglial development in microencephalic rat brainafter fetal MAM treatment. J. Dev. Neurosci. 4, 353–359.

Ferguson, S.A., Paule, M.G., Holson, R.R., 1996. Functionaleffects of methylazoxymethanol-induced cerebellar hypo-plasia in rats. Neurotoxicol. Teratol. 18, 529–537.

Gage, F.H., Coates, P.W., Palmer, T.D., Kuhn, H.G., Fisher,L.J., Suhonen, J.O., Peterson, D.A., Suhr, S.T., Ray, J., 1995.Survival and differentiation of adult neuronal progenitor cellstransplanted to the adult brain. Proc. Natl. Acad. Sci. USA92, 11 879–11 883.

Giurgea, C., Mouravieff-Lasuisse, F., 1972. Effet facilitateur dupiracetam sur un apprentissage repetitif chez le rat. J.Pharmacol. 3, 17–30.

Gomez-Pinilla, F., Dao, L., Vannarith, S., 1997. Physical exerciseinduces FGF-2 and its mRNA in the hippocampus. BrainRes. 764, 1–8.

Gomez-Pinilla, F., So, V., Kesslak, J.P., 1998. Spatial learningand physical activity contribute to the induction of fibroblastgrowth factor: neural substrates for increased cognitionassociated with exercise. Neuroscience 85, 53–61.

Gould, E., Tanapat, P., 1997. Lesion-induced proliferation ofneuronal progenitors in the dentate gyrus of adult rat.Neuroscience 80, 427–436.

Gould, E., Cameron, H.A., McEwen, B.S., 1994. Blockade ofNMDA receptors increases cell death and birth in thedeveloping rat dentate gyrus. J. Comp. Neurol. 340, 551–565.

Gould, E., Beylin, A., Tanapat, P., Reeves, A., Shors, T.J., 1999.Learning enhances adult neurogenesis in the hippocampalformation. Nat. Neurosci. 2, 260–265.

Haddad, R.K., Rabe, A., Dumas, R., 1972. Comparisons of effectof methylazoxymethanol acetate on brain development indifferent species. Fed. Proc. 31, 1520–1523.

Hastings, N.B., Gould, E., 1999. Rapid extension of axons intothe CA3 region by adult-generated granule cells. J. Comp.Neurol. 413, 146–154.

Kempermann, G., Kuhn, H.G., Gage, F.H., 1997a. Geneticinfluence on neurogenesis in the dentate gyrus of adult mice.Proc. Natl. Acad. Sci. USA 94, 10 409–10 414.

Kempermann, G., Kuhn, H.G., Gage, F.H., 1997b. Morehippocampal neurons in adult mice living in an enrichedenvironment. Nature 386, 493–495.

Kuhn, H.G., Dickinson-Anson, H., Gage, F.H., 1996. Neuroge-nesis in the dentate gyrus of adult rat: age-related decreaseof neuronal progenitor proliferation. J. Neurosci. 16, 2027–2033.

Lindvall, O., Kokaia, Z., Bengzon, J., Elmer, E., Kokaia, M.,1994. Neurotrophins and brain insults. Trends Neurosci. 17,490–496.

Lovell, K.L., Jones, M.Z., 1980. Partial external germinal layerregeneration in the cerebellum following methyla-zoxymethanol administration: effects on Purkinje cell den-dritic spines. J. Neuropathol. Exp. Neurol. 39, 541–548.

Markakis, E.A., Gage, F.H., 1999. Adult-generated neurons inthe dentate gyrus send axonal projections to field CA3 andare surrounded by synaptic vesicles. J. Comp. Neurol. 406,449–460.

Minturn, J.E., Geschwind, D.H., Fryer, H.J.L., Hockfield, S.,1995. Early postmitotic neurons transiently express TOAD-64, a neural specific protein. J. Comp. Neurol. 355, 369–379.

Okano, H.J., Pfaff, D.W., Gibbs, R.B., 1996. Expression ofEGFR-, p75NGFR-, and PSTAIR (cdc2)-like immunoreac-tivity by proliferating cells in the adult rat hippocampalformation and forebrain. Dev. Neurosci. 15, 199–209.

Palmer, T.D., Ray, J., Gage, F.H., 1995. FGF-2 responsiveneuronal progenitors reside in proliferative and quiescentregions of the adult rodent brain. Mol. Cell Neurosci. 6,474–486.

Raff, M.C., Barres, B.A., Burne, J.F., Coles, H.S., Ishizaki, Y.,Jacobson, M.D., 1993. Programmed cell death and thecontrol of cell survival: lessons from the nervous system.Science 262, 670–695.

Ray, J., Peterson, D.A., Schinstine, M., Gage, F.H., 1993.Proliferation, differentiation and long-term culture of pri-mary hippocampal neurons. Proc. Natl. Acad. Sci. USA 90,3602–3606.

Schlessinger, A.R., Cowan, W.M., Gottlieb, D.I., 1975. Anautoradiographic study of the time of origin and the patternof granule cell migration in the dentate gyrus of the rat. J.Comp. Neurol. 159, 149–176.

Seki, T., Arai, Y., 1993. Highly polysialylated neural cell adhesionmolecule (NCAM-H) is expressed by newly generated granulecells in the dentate gyrus of the adult rat. J. Neurosci. 13,2351–2358.

Sloviter, R.S., 1989. Calcium-binding protein (calbindin-D28K)and parvalbumin immunocytochemistry: localization in therat hippocampus with specific reference to the selectivevulnerability of hippocampal neurons to seizure activity. J.Comp. Neurol. 280, 183–196.

Sullivan-Jones, P., Ali, S.F., Gough, B., Holson, R.R., 1994.Postnatal methylazoxymethanol: sensitive periods and re-gional selectively of effects. Neurotoxicol. Teratol. 16, 631–637.

van Praag, H., Kempermann, G., Gage, F.H., 1999. Runningincreases cell proliferation and neurogenesis in the adultmouse dentate gyrus. Nat. Neurosci. 2, 266–270.

Williams, R.W., Herrup, K., 1988. The control of neuronnumber. Ann. Rev. Neurosci. 11, 423–453.

Young, D., Lawlor, P.A., Leone, P., Dragunow, M., During,M.J., 1999. Environmental enrichment inhibits spontaneousapoptosis, prevents seizures and is neuroprotective. Nat.Med. 5, 448–453.