• Embed Doc
  • Copy Link
  • Readcast
  • Collections
  • CommentGo Back
 
Subpopulations of Proliferating Cells of the Adult Hippocampus RespondDifferently to PhysiologicNeurogenic Stimuli
GOLO KRONENBERG,
1,2
KATJA REUTER,
1
BARBARA STEINER,
3
MORITZ D. BRANDT,
1
SEBASTIAN JESSBERGER,
1,3
MASAHIRO YAMAGUCHI,
4
 AND
GERD KEMPERMANN
1,3
*
1
Max Delbru¨ck Center for Molecular Medicine (MDC) Berlin-Buch, 13125 Berlin, Germany
2
Department of Psychiatry, Freie Universita¨t Berlin, 14050 Berlin, Germany
3
 VolkswagenStiftung Research Group, Department of Experimental Neurology,Charite´ University Hospital, Humboldt University, 10117 Berlin, Germany
4
Department of Physiology, Graduate School of Medicine, University of Tokyo,Bunkyo-ku, Tokyo 113-0033, Japan ABSTRACTTo study how adult hippocampal neurogenesis might originate from the proliferation of stem or progenitor cells in vivo, we have used transgenic mice expressing green fluorescentprotein (GFP) under the nestin promoter to identify these cells. Having described anastrocyte-like type 1 cell with low proliferative activity, a characteristic morphology, vascularend feet, and passive electrophysiological properties, we focused here on the large populationof nestin-GFP-expressing type 2 cells, which lack all these features. Type 2 cells were highlyproliferative and showed signs suggestive of their involvement in the neuronal lineage. Theycould be subclassified by the absence (type 2a) or presence (type 2b) of a coexpression of theearly neuronal marker doublecortin. A third type of proliferating cells was doublecortinpositive but nestin-GFP negative (type 3). We believe that type 2a, 2b, and 3 cells mirror amarker progression during earliest neuronal development. This view is supported by theincreasing coexpression of the early granule cell-specific marker Prox-1. The low proliferativeactivity of type 1 cells showed little change over time or under “neurogenic interventions,”such as a challenge by environmental complexity (ENR) or voluntary physical activity (RUN).However, RUN led to a significant increase of type 2 cells labeled with the proliferationmarker bromodeoxyuridine (BrdU). ENR did not cause increased cell proliferation or anincreased number of BrdU-labeled type 2 cells, but both ENR and RUN resulted in morenewly generated cells lacking nestin-GFP immunoreactivity and expressing Prox-1. Thesefindings allow us to break down what was broadly perceived as “proliferation” in earlierexperiments into the relative contribution of several cell types, representing the earliest stepsof neuronal development. J. Comp. Neurol. 467:455–463, 2003.
©
2003 Wiley-Liss, Inc.
Indexing terms: nestin; progenitor cell; stem cell; neurogenesis; doublecortin; environmental enrichment
Grant sponsor: VolkswagenStiftung; Grant sponsor: Deutsche Forschungs-gemeinschaft; Grant number: KE 615/2-1; Grant sponsor: GlaxoSmithKline(G.Kr.); Grant sponsor: Deutschen Forschungsgemeinschaft (M.D.B.).*Correspondence to: Gerd Kempermann, Max Delbru¨ck Center for Mo-lecular Medicine (MDC) Berlin-Buch, Robert-Ro¨ssle-Str. 10, 13125 Berlin,Germany. E-mail: gerd.kempermann@mdc-berlin.deReceived 28 April 2003; Revised 24 July 2003; Accepted 6 August 2003DOI 10.1002/cne.10945Published online the week of November 3, 2003 in Wiley InterScience(www.interscience.wiley.com).
THE JOURNAL OF COMPARATIVE NEUROLOGY 467:455–463 (2003)
©
2003 WILEY-LISS, INC.
 
 Adult hippocampal neurogenesis is a prominent aspectof neural stem cell biology, because new neurons in theadult hippocampus are derived from neural stem or pro-genitor cells that persist in the adult brain (Palmer et al.,1997). The controlled addition of new neurons to the adulthippocampus has received widespread scientific interestbecause of its many implications for our understanding of hippocampal function and pathology (D’Sa and Duman,2002; Gould and Gross, 2002; Kempermann, 2002; Parentand Lowenstein, 2002). Therefore, more detailed knowl-edge about hippocampal precursor cells in vivo is of par-ticular importance. Many concepts developed to describeneural stem cell biology have made a distinction between“stem cells” and transiently amplifying “progenitor cells”(Gage et al., 1995; Seaberg and van der Kooy, 2003).Supposedly, the latter category is defined by a lineagerestriction and limited self-renewal. Neural stem cells, incontrast, would not show a determination to a neuronal orglial phenotype and would have the capability of unlim-ited self-renewal. At present, however, it is not clear howthese findings from ex vivo and in vitro studies relate toconditions in vivo. Adult hippocampal neurogenesis orig-inates from a precursor cell population located in thesubgranular zone (SGZ) of the adult mammalian dentategyrus. There is an ongoing debate on whether these pre-cursor cells are truly stem cells in the strictest sense of theterm (Palmer et al., 1997; Seaberg and van der Kooy,2002).We have previously shown that, by the standard of nestin expression, an intermediate filament expressed inneural stem or progenitor cells, two types of dividing pre-cursor cells can be found in the adult SGZ (Filippov et al.,2003). Type 1 cells had astrocytic features, such as theexpression of glial fibrillary acidic protein (GFAP), theextension of a long radial-glia-like process spanning theentire granule cell layer, vascular end feet on capillaries of the SGZ, passive membrane properties, and potassiumcurrents. Seri et al. (2001) had shown earlier that GFAP-expressing cells of the adult SGZ can repopulate the neu-rogenic system of the dentate gyrus after cytotoxic abla-tion of actively dividing cells, so it is assumed that type 1cells are the stem cells of the adult hippocampus, althoughfuture studies will have to address this issue in moredetail. In our earlier study, we also identified a secondtype of nestin-expressing cell that completely lacked as-trocytic features. With few exceptions, these cells werelocated in the SGZ, had rather plump and mostly tangen-tial processes, and had an irregularly shaped nucleus withdense chromatin. Importantly, they were highly prolifer-ative and accounted for the majority of dividing nestin-expressing cells in the SGZ. Their proliferative activityandexpressionofnestinmakesthemalikelycandidateforthe proposed population of transiently amplifying progen-itor cells.The present study was designed to describe nestin-expressing cells in the SGZ in more detail. Could thepopulation of type 2 cells be further differentiated? Isthere any evidence that these cells show a neuronal pre-determination? If so, what is their relationship to furtherstages of early granule cell development? How do thesecells react to stimuli known to induce adult hippocampalneurogenesis and cell proliferation in the SGZ? Answersto these questions will help to identify the stages of neu-ronal development within adult hippocampal neurogen-esis, which are sensitive to regulatory influences. For ourstudy, we made use of transgenic mice that express greenfluorescent protein (GFP) under a neurally specific en-hancer region of the nestin promoter (Yamaguchi et al.,2000).
MATERIALS AND METHODS Animals
The generation of transgenic mice expressing GFPdriven by regulatory elements of the nestin gene has beendescribed elsewhere (Yamaguchi et al., 2000; Sawamoto etal., 2001). At the beginning of the different experiments,all animals were 12 weeks old. For the time-series exper-iment, animals received a single intraperitoneal injectionof bromodeoxyuridine (BrdU; 50
g/kg body weight) andwere killed after survival times of 2 hours, 24 hours, 3days, 7 days, or 4 weeks. The second experiment wasdesigned to study the early effects of environmental ma-nipulation on proliferating cells. Twenty-four mice wererandomly assigned to one of three conditions. Controls(CTR) and runners (RUN) were kept in standard labora-tory cages with two or three animals per cage. RUN hadunlimited access to a running wheel in their cage (Tecni-plast,Hohenpeißenberg,Germany).Theenrichedenviron-ment (ENR) was similar to that used in our previousstudies (Van Praag et al., 1999a). Briefly, it consisted of one larger cage of approximately 80
80 cm floor area,equipped with mesh wire ladders and a rearrangeabletunnel system. The enriched environment did not containa running wheel, and the ENR animals received the samefood as the other groups. After 4 weeks, animals of thesecond experiment received a single injection of BrdU andwere perfused 24 hours later. All animals were kept on a 12-hour light/dark schedulewith ad libitum access to food and water. All applicablelocal and federal regulations for animal welfare were fol-lowed. The animal protocol was approved by Landesamtfu¨r Arbeitsschutz, Gesundheitsschutz und technischeSicherheit Berlin (LaGetSi).
Immunohistochemistry
 Animals were deeply anesthetized with ketamine andperfused transcardially with cold 4% paraformaldehyde in0.1 M phosphate buffer, pH 7.4. The brains were dissectedfrom the skulls and postfixed overnight. Before sectioningfrom a dry ice-cooled copper block on a sliding microtome(Leica, Bensheim, Germany), the brains were transferredto 30% sucrose in 0.1 M phosphate buffer, pH 7.4, untilthey sank. Brains were cut in the coronal plane in 40-
m-thick sections and cryoprotected. For BrdU staining, DNAwas denatured in 2 N HCl for 30 minutes at 37°C. Sectionswere then rinsed in 0.1 M borate buffer, pH 8.5, andthoroughly washed in Tris-buffered saline (TBS), pH 7.4.Primary and secondary antibodies were diluted in TBSsupplemented with 0.1% Triton X-100, 0.1% Tween 20,and 3% donkey serum (TBS-plus). For immunohistochem-istry, sections were pretreated with 0.6% H
2
O
2
and incu-bated with the secondary antibody for 2 hours at roomtemperature. ABC reagent (Vectastain Elite; Vector Lab-oratories, Burlingame, CA) was applied for 1 hour at aconcentration of 9
l/ml. Diaminobenzidine (DAB; SigmaGermany) was used as a chromogen at a concentration of 0.25 mg/ml in TBS with 0.01% H
2
O
2
and 0.04% nickel-chloride. For immunofluorescence, one-in-twelve series of 
456 G. KRONENBERG ET AL.
 
each brain were triple-labeled. Sections were incubatedwith the primary antibody for 48 hours at 4°C, washedwith TBS and TBS-plus, then incubated with the second-ary antibodies for 4 hours at room temperature. Sectionswere then washed with TBS and coverslipped in polyvinylalcohol with diazabicyclooctane (DABCO) as anti-fadingagent. The primary antibodies were applied in the follow-ing concentrations: anti-BrdU (rat, 1:500; Harlan Seralab,Indianapolis, IN), anti-GFAP (guinea pig, 1:1,000; Ad- vanced Immunochemistry, Los Angeles, CA), anti-S100
(rabbit, 1:2,500; SWant, Belinzona, Switzerland), anti-NeuN (mouse, 1:100; Chemicon, Temecula, CA), anti-doublecortin (goat, 1:200; Santa Cruz Biotechnologies,Santa Cruz, CA), anti-polysialic acid-NCAM (PSA-NCAM;mouse, 1:400; Chemicon), anti-Prox-1 (rabbit, 1:5,000;kindly provided by Samuel Pleasure, University of Cali-fornia, San Francisco), and anti-Map-2ab (mouse, 1:400; Abcam). For secondary antibodies, we used: anti-ratrhodamine-X, anti-guinea pig Cy5, anti-rabbit fluoresceinisothiocyanate (FITC), anti-mouse FITC, anti-goat Cy5,and anti-mouse IgM rhodamine-X (all 1:250; Jackson Im-munoresearch, West Grove, PA; distributor: Dianova,Hamburg, Germany).
Quantification and imaging
 As cell counts are highly sensitive to methodologicalerrors, most notably resulting from unnoticed changes inthe volume of the structure to be studied, stereologicalmethods that either determine the sample and reference volumes (disector principle) or are independent of the volume (fractionator principle) have to be used (Cogge-shall and Lekan, 1996). To determine the number of BrdU-labeled cells per dentate gyrus, we therefore used amodified version of the fractionator principle, partly basedon a method devised by Williams and Rakic (1988). Thismethod has been used in our previous studies and is alsodiscussed elsewhere (Kempermann et al., 1997a). The ra-tionale for this modification lies in the fact that BrdU-positive cells are comparatively rare and are not evenlydistributed in the granule cell layer. More than 60% of theBrdU-positive cells are found in the SGZ and the innerone-third of the granule cell layer (Kempermann et al.,2003). When an automated stereology system, such as Ste-reoInvestigator, is used, this leads to an extremely highnumber of counting frames, most of which do not containlabeled cells. We therefore stained a one-in-six series of sec-tions covering the entire hippocampus in its rostrocaudalextension with the peroxidase/diaminobenzidine (DAB)method. The optical fractionator principle was modified inthat only the labeled cells in the uppermost focal plane (at
40) were excluded to avoid oversampling, but BrdU-positive cells were otherwise counted exhaustively in thegranule cell layer and the SGZ (Kempermann et al., 2003).The total number of BrdU-labeled cells was then esti-mated by multiplying the resulting counts by 6, becauseevery sixth section had been used. The numbers obtainedin our study are thus absolute numbers per dentate gyrus(including SGZ) and are independent of the volume of thedentate gyrus. In a similar way, we determined the abso-lute numbers of nestin-positive type 1 and type 2 cells, of DCX-positive cells, and of Prox-1-positive cells.For phenotypic analyses of all antibody combinations of interest, 50 cells per animal were examined with immu-nofluorescent triple labeling and spectral confocal micros-copy (Leica TCS SP2). The absolute number of cells of acertain phenotype was then calculated by multiplying theabsolute number of cells of the reference phenotype asdetermined with the peroxidase method with the percent-age of the phenotype of interest. Volocity imaging software (Improvision, Heidelberg,Germany) was used for three-dimensional reconstructionof confocal z-stacks (usually three or four optical sectionsper micrometer of z-distance) and to document doublelabeling in cells depicted in xy-, xz-, and yz-planes. Digitalfalse-color images were processed with Adobe Photoshop7.0 for Macintosh. Only general adjustments of color, con-trast, and brightness were made. The images were nototherwise manipulated.
Statistical analysis
 All numerical analyses were performed using Statview5.0.1.ANOVAwasfollowedbyFisher’sposthoctestwhereappropriate.
RESULTSNestin-GFAP-positive type 2 cells expressearly neuronal markers
 As described earlier, two types of nestin-GFP-expressing cells can be distinguished in the adult SGZ(Filippov et al., 2003). Type 1 cells have a characteristicmorphology: Their cell body resides in the SGZ, and theyextend a long slender process that branches in the innerone-third of the granule cell layer and extends into themolecular layer. These cells share astrocytic features,whereas type 2 cells do not (Fig. 1A–D).Here, type 2 cells were in the focus of our interest. Type2 cells were consistently negative for the astrocytic mark-ers glial fibrillary acidic protein (GFAP) and S100
. If these cells do not express glial markers, is there anyevidence that they are predetermined to become neurons?We examined nestin-GFP-expressing type 2 cells forcoexpression of transient neuronal markers DCX andPSA-NCAM, mature neuronal markers NeuN and Map-2ab, and granule cell markers Prox-1 and calbindin. Dou-ble staining with antibodies against DCX and PSA-NCAMas well as with the granule cell marker Prox-1 (see below)was detected in type 2 cells (Fig. 1E,F). This is suggestiveof a preneural state or neural lineage determination of these putative progenitor cells. We did not detect type 2cells positive for any of the mature neuronal and granulecell markers (Fig. 1G). As expected, type 1 cells werenegative for all of these markers, further substantiatingtheir astrocyte-like identity.It is interesting that not all type 2 cells expressed mark-ers of immature neurons, such as DCX or PSA-NCAM.The further analysis was based on DCX expression as areference marker. DCX-negative nestin-GFP-expressingcells will be referred to as type 2a cells; DCX-positivenestin-GFP-expressing cells as type 2b cells.In general, nestin-GFP-expressing cells of the SGZ havebeen identified as highly proliferative. Therefore, our nextquestion was whether both type 2a and type 2b cells wereactively dividing. This was addressed by studying triplelabelings with antibodies against GFP, DCX, and the cell-cycle marker Ki-67. We found that both DCX-negativetype 2a and -positive type 2b subgroups of nestin-GFP-expressing cells were undergoing cell division. There wasa third type of cell, which was negative for nestin-GFP but
457PROGENITOR CELLS IN THE ADULT HIPPOCAMPUS
of 00

Leave a Comment

You must be to leave a comment.
Submit
Characters: ...
You must be to leave a comment.
Submit
Characters: ...