sign in sign up
<<
Home , Purkinje cell

The Journal of Neuroscience, February 1991, 17(2): 462-471

NGF and Excitatory Neurotransmitters Regulate Survival and Morphogenesis of Cultured Cerebellar Purkinje Cells

Susana Cohen-Cory,1v2 Cheryl F. Dreyfus,2sa and Ira B. Black2za ‘Laborator\/ of Neurobiolonv, The Rockefeller Universitv, and ‘Division of Developmental Neurology, Cornell University Medical College, New Yo;kT New York 10021 .

The development of cerebellar Purkinje cells is subject to et al., 1989;Allendoerfer et al., 1990) suggestthat NGF regulates regulation by multiple epigenetic signals. To define mech- the ontogeny of diverse neuronal populations. anisms by which trophic and presynaptic stimulation may In particular, previous studies indicate that cerebellar Pur- potentially regulate Purkinje ontogeny, we studied the kinje cells expressNGF-Rs during early development in several effects of NGF and excitatory transmitters on Purkinje cell species(Eckenstein, 1988; Schatteman et al., 1988; Yan and survival and morphological maturation in dissociated cell Johnson, 1988). The selective expression of NGF-Rs by Pur- culture. Purkinje cells were identified by expression of vi- kinje cells Suggeststhat NGF may be critical for cerebellar on- tamin D-dependent calcium-binding protein and by their togeny. Purkinje cells, the major efferent of the cere- characteristic morphology. NGF receptors were selectively bellum, play a central role in the formation of precisecerebellar localized to Purkinje cells by both ligand and monoclonal cortical networks and are critical for integration of motor func- antibody binding, suggesting responsivity to the trophic tion. We have previously found that the biologically active, agent. Simultaneous exposure to depolarizing agents and high-affinity NGF-Rs are localized to developing Purkinje cells NGF specifically enhanced Purkinje cell survival in culture. (Cohen-Cory et al., 1989). The transient expressionof NGF-Rs NGF, in combination with either high potassium or veratridine in vivo coincides with the period of active differentiation and markedly increased survival of Purkinje cells. Furthermore, ,suggesting that NGF may regulate the ontogeny NGF together with the excitatory neurotransmitters aspar- of this cell type. We have now examined the potential role of tate or glutamate promoted a 2-fold increase in survival. In NGF during development by studying Purkinje cell survival addition, NGF increased Purkinje cell size and promoted and differentiation in culture. neurite elaboration. These effects required simultaneous ex- Purkinje cells constitute a particularly advantageousmodel posure to NGF and either aspartate, glutamate, or pharma- to characterize the role of NGF in development: cologic depolarizing agents. Effects on survival or neurite The cellular events that lead to the morphological and functional elaboration were not evoked by exposure to trophic factors differentiation of Purkinje cells (Altman, 1972;Altman and Bay- or transmitters alone. Our results suggest a novel mecha- er, 1985) and the role that several environmental signalsplay nism for regulation of development, in which trophic factor in cerebellar ontogeny (for review, see Ito, 1984) have been and afferent stimulation interact to promote survival and extensively studied. Multiple epigenetic factors are known to morphogenesis of developing Purkinje cells. regulate Purkinje cell development. For example, afferent innervation is of major importance for differentiation of de- Increasing evidence suggeststhat NGF plays an important role veloping Purkinje cells (Rakic and Sidman, 1973; Berry and in the CNS (for review, see Thoenen et al., 1987; Whittemore Bradley 1976b; Bradley and Berry 1976a,b;Sotelo and Arsenio- and Seiger, 1987; Dreyfus, 1989). While initial studiesindicated Nunes, 1976). Interactions amongseveral epigeneticsignals may that basalforebrain cholinergic neuronsrespond to NGF (Hefti be essentialfor proper regulation of neuronal development (for et al., 1984; Martinez et al., 1985), more recent evidence suggests review, seePurves, 1988). Consequently, we examined the ac- that NGF may have widespread functions in the CNS. The tions of both NGF and putative afferent transmitters in the transient expressionof NGF receptors (NGF-Rs) in multiple regulation of Purkinje cell development. Responsesto trophic brain areas (Large et al., 1986; Escand6n and Chao, 1989; Lu factor and excitatory neurotransmitters were analyzed in dis- sociated cell culture. We have found that NGF in combination with the putative Received May 30, 1990; revised Sept. 14, 1990; accepted Sept. 21, 1990. excitatory neurotransmittersaspartate and glutamate markedly We wish to thank Dr. S. Christakos for providing vitamin D-dependent calcium- increasedsurvival and promoted morphological maturation of binding protein antibodies, Dr. P. Bernd for the lz51-NGF, and Dr. E. M. Johnson, Jr., for the NGF receptor antibody used in this study. We also thank Mr. P. Peirce Purkinje cells. for excellent photography and Ms. E. Vianna for technical assistance. We are most grateful to Dr. T. Wiesel for his valuable suggestions. This work was supported Materials and Methods by the Lucille P. Markey Charitable Trust and by NIH Grants NS 10259 and HD 23315. I.B.B. is a recipient of a M&night Research Project Award. Tissue dissociation and cell culture. Cerebella from embryonic day 18 Correspondence should be addressed to Susana Cohen-Gory, Department of (E 18) SpragueDawleyrats were usedto establishdissociated cell cul- Physiology and Biophysics, College of Medicine, University of California, Irvine, tures.Timed pregnant rats were killed by exposureto CO, vapor. Fetuses Irvine, CA 927 17. wereremoved by cesareansection and transferred to a sterilepetri dish a Present address: Department of Neuroscience and , Robert Wood with phosphate-bufferedsaline (PBS). Fetal cerebellawere dissected Johnson Medical School, UMDNJ, 675 Hoes Lane, Piscataway, NJ 08854-5635. from surroundingbrain tissue.In brief, the skullwas removed, and the Copyright 0 1991 Society for Neuroscience 0270-6474/91/l 10462-10$03.00/O midbrainwas dissected to exposethe cerebellumand underlyingbrain The Journal of Neuroscience, February 1991, 17(2) 463 stem. Only cerebellar tissue medial to the cerebellar peduncles was antibodies (Fig. 2). CaBP is a cell-specific antigen exclusively dissected, and the were completely removed. Pooled tissue expressedby Purkinje cells within the (Jandeet al., from 24 litters wasminced into OS-l.O-mm piecesand mechanically dissociatedin 2 ml of nutrient medium.Culture conditions were mod- 1981; Christakos et al., 1987). Phase microscopy confirmed ified after those described previously (Hatten, 1985). Cultures were immunopositivity of the entire cell body and processes,ofCaBP- maintained in nutrient medium containing Minimal Essential Medium positive cellsin culture. Immunopositive cellsappeared as phase- (MEM) with Earle’s salts supplemented with 2 mM glutamine (Gibco, bright neuronlike cells, exhibiting morphology characteristic of Grand Island,NY), glucose(6 mg’ml), heat inactivatedhorse serum (10% vol/vol), and low concentrations of penicillin (0.5 U/ml) and developing Purkinje cells (Fig. 2). streptomycin(0.5 &ml). Cellswere countedand platedon poly-D- lysine-coated petri dishes (35 mm) at a final density of 1 x lo6 cells/ Purkinje cells in culture express NGF receptors dish. Cultures were maintained in nutrient medium at 37°C in a 95% While previous studiesindicated that Purkinje cellsexpress NGF sir/5% CO, humidifiedincubator. As indicated,NGF (200U/ml), po- receptorsin vivo, expressionin culture had yet to be delineated. tassium chloride (25 mM), veratridine (VRT; 0.5 PM), aspartic acid (10 In fact, Purkinje cells identified both morphologically and by PM), and glutamic acid (10 PM) were added to the medium at the time of plating. The concentration of the drugs used in this study was de- CaBP immunoreactivity expressedNGF-Rs by immunocyto- termined from dose-response experiments. We chose doses yielding chemical criteria. Using the optimal culture conditions, the anti- optimal effects without toxicity. The medium was replaced once after rat NGF-R monoclonal antibody 192-IgG revealed the presence 4-5 d in culture, with the addition of fresh drugs. of positive cells that were morphologically identified as deve- Preparation of NGF. The 0 subunit of NGF was prepared from adult mouse salivary glands as previously described (Mobley et al., 1976) loping Purkinje cells (Fig. 3). In all casesexamined, NGF-R andbioactivity wasdetermined following published procedures (Fenton, antigenicity colocalized with CaBP immunoreadtivity (not 1970). shown). Expressionof NGF-R was confirmed by lZSI-NGF bind- Immunocytochemistry. After 7 d in culture, dissociates were rinsed in ing studies. Incubation with radioactive NGF specifically la- PBS andfixed for 3 hr at 4°Cin 4%paraformaldehyde in 0.1 M phosphate buffer (pH, 7.4). Polyclonal antibodies against neuron-specific enolase beledbinding siteson CaBP-positive cells(Fig. 4A). Silver grains (NSE; Polysciences, Inc. Laboratories; 1: 1000 dilution) were used for were restricted to CaBP-immunopositive cells, suggestingthat immunostaining the neuronal populations in culture. To detect the Pur- expression of NGF-R is limited to the Purkinje population. kinje cell population specifically, a polyclonal antibody to vitamin D-de- Specific lZ51-NGFbinding was completely abolished in control pendentcalcium-binding protein (CaBP), generously provided by Dr. cultures incubated with lOOO-fold excessunlabeled NGF (Fig. S. Christakos, was used at a 1:20dO dilution (Christakos et al., 1987). The 192-IaG monoclonalantibodv directedagainst the rat NGF-R. 4B). The expression of NGF-Rs by Purkinje cells in culture generously-providedby Dr. E. M. johnson(Yai and Johnson,1988); suggeststhat these cells retain the potential ability to respond was used at a 1: 10 dilution. The immunostaining method involved the to NGF in vitro. use of the avidin-biotin complex (ABC) technique (Vectastain ABC kit, Vector Labs.)to visualizestaining. Biotinylated secondary antibodies NGF and depolarizing stimuli enhance Purkinje cell survival wereobtained from Vector Laboratories. NGF binding assay. The lz51-NGF binding protocol used in this study Dissociated cell culture allows the study of neuronal survival has been previouslydescribed (Bemd et al., 1988).In brief, El 8 cere- in a rigorously controlled seriesof environments. However, nor- bellar cell cultures were grown for 7 d, nutrient medium was removed, mal afferents and cell-cell interactions are disrupted. Because and cultureswere washed with 3 changesof MEM with 0.6%glucose impulse activity apparently regulatesPurkinje cell development for 2 hr at 37°Cto removeany endogenous NGF. For 1251-NGFbinding, cultures were incubated with 0.2 tn.4 Y-NGF for 1 hr at 37°C in MEM, in vivo, cells were grown under depolarizing conditions, and 0.6%glucose, and 5 mg/ml bovine serumalbumin. Following incuba- NGF effectswere analyzed. Neuronal cell number was markedly tion, cultures were rinsed rapidly 6 times in ice-cold PBS and fixed with increasedin cultures grown for 7 d in high-potassium medium 4% paraformaldehyde.Cultures were immunostained,dipped in L4 (25 mM KCl). Treatment with high potassiumalone increased emulsion(Ilford), and exposedfor autoradiography. Morphometric analysis and assessment of cell survival. Cultures were the cell number by 40% (Figs. 1,5). Moreover, NGF significantly examinedusing a Nikon Labophote100 microscope. Immunopositive increasedPurkinje cell number by 75% over control with no cellcounts were obtained by analyzingan areaequivalent to 10%of the potassium(Fig. 5). However, NGF had no effect in the absence dish surface. The number of CaBP-positive cells was expressed as the of high potassium. The effect of NGF and potassium in com- percent of cells in experimental versus control cultures. Morphometric bination wasspecific and significantly different from that elicited measurementswere performed using representative, isolated CaBP-im- munopositive cells in sparse cultures. Cell diameter and neurite length by either agent alone (Fig. 5). were measuredwith an eyepiecemicrometer. Numbers of neuritesand To determine whether depolarization per se was responsible branchpoints per cell weredetermined by counting.We defined“neu- for the potassium-elicited effect, cultures were grown in the rons with spiny processes” as those CaBP-positive cells with processes presenceof veratridine (VRT). The alkaloid depolarizesneurons 0.5-1.0 pm in lengthprotruding from centralneurites. A drawingtube coupled to the Nikon microscope was used for tracing CaBP-positive by opening sodium channels, a mechanismdistinct from that cells. An analysisof variance(ANOVA) programwas used for the sta- of potassium.VRT depolarization increasedsurvival of all neu- tistical analysis of data. rons, mimicking potassium. VRT treatment alone increased Purkinje cell survival by 130%(Fig. 6). Moreover, a further and Results significant 70% increasein survival was elicited by NGF in the To evaluate the potential role of NGF during Purkinje cell on- presenceof VRT (Fig. 6). Consequently, depolarization with togeny, we establisheda culture system allowing analysis of either high potassium or VRT was required for NGF actions responsesto trophic factor stimulation. Dissociated rat cere- on cell survival. These observations raised the possibility that bellar cell culturesmaintained in serum-containingmedium fos- afferent excitatory transmitters and trophic agentsact in concert. tered survival and differentiation of Purkinje cells. The use of mixed cultures (neuronal and non-neuronal cells)promoted dif- NGF and excitatory neurotransmitters enhance Purkinje ferentiation of Purkinje cells in a relatively rich environment cell survival (Fig. 1). This system allowed detailed analysisof the effects of In the developing rat cerebellum, synaptogenesislargely tran- NGF on survival and differentiation of individual neurons.Pur- spires during the first postnatal weeks, coincident with high kinje cells were visualized immunocytochemically with CaBP NGF-R messageexpression (Buck et al., 1988; Lu et al., 1989). 464 Cohen-Coty et al. l NGF Regulates Purkinje Cell Development

Figure 1. Visualization of neurons in culture by immunocytochemical staining with NSE antibody. Cerebellar cells were grown for 7 d in culture under standard conditions (A) or in the presence of NGF (B), potassium (C), or potassium with NGF (0). Note the increase in neuronal cell number and neurite length by treatment with potassium or potassium with NGF. Scale bar, 100 pm.

However, at birth, prior to the establishmentof cerebellarex- citatory afferent ,developing Purkinje cells are already Neuritic morphology sensitiveto L-aspartateand L-glutamate(Woodward et al., 1971; Potassium and NGF. Treatment with potassium alone signifi- DuPont et al., 1987). To ascertain whether normal innervation cantly increasedneurite outgrowth in all Purkinje cells (Table is necessaryfor trophic regulation of Purkinje cell development, 1, Fig. 1). Moreover, treatment with both potassiumand NGF we analyzed the role of both aspartateand glutamateon survival. elicited specificchanges, increasing neurite number per cell (Ta- The combination of NGF and either aspartate or glutamate ble 1; cf. control cultures) and neurite branching (Table 1, Fig. markedly enhanced survival of Purkinje cells (Figs. 7, 8). A 1). In contrast, NGF alone had no significant effect on neurite. specific, 100% increasein survival was observed after exposure number or neurite branching. Treatment with NGF and high to aspartate and NGF (Fig. 7). A similar effect was obtained potassiumincreased the proportion of cells with spiny neurites with glutamate and NGF (Fig. 8). Unlike effectsnoted with the (Table 1). pharmacologicalagents, the specifictransmitters L-aspartateand Aspartate/glutamateand NGF. In agreementwith the survival L-glutamatealone had no effect on the survival of CaBP-positive results,treatment with aspartateor glutamate had only a mod- cells (Figs. 7, 8). Moreover, upon exposure to higher dosesof erate effect on neurite morphology. However, treatment with glutamate (0.05-l mM), a toxic effect was observed on all cells excitatory neurotransmittersand NGF markedly enhancedneu- (not shown). Consequently, in vivo, excitatory neurotransmit- ritic length and complexity (Fig. 9). NGF, in combination with ters, in concert with the trophic protein, may regulatePurkinje aspartate, significantly increased neurite branching (Table 2). cell survival. The treatment similarly increased the proportion of Purkinje cells with spiny neurites (Table 2). NGF and depolarizing agents regulate cell size and shape To define NGF actions on Purkinje cell development further, Cell size we examined the effects of NGF and depolarizing agentson the NGF in combination with high potassium (Table l), aspartate morphological maturation of cells in culture. Severalparameters (Table 2, Fig. 9) or glutamate (Fig. 9) significantly increased were evaluated. the size of the of CaBP cells. These effects were not ob- The Journal of Neuroscience, February 1991, 1 f(2) 465

Figure 2. Immunocytochemicalvi- sualizationof Purkinjecells in culture. Cerebellarcell cultures were grown for 7 d in the absence(A) or presenceof eitherNGF (B), potassium(C), or po- tassiumwith NGF (0) and immuno- stainedwith CaBPantibodies. Note the markedenhancement in neuriteelab- oration by treatment with potassium with NGF. Scalebar, 40 pm.

served by treatment with aspartate, glutamate, or potassium alone. A culture system to study Purkinje cell growth To begin defining the trophic mechanismsthat may regulate Discussion Purkinje cell development, we studied NGF in a system readily Localization of the biologically active, high-affinity NGF bind- accessibleto experimental manipulation. Dissociated embry- ing sitesto developing Purkinje cells suggestedthat the factor onic cell cultures provide a particularly advantageousmodel, regulates ontogeny of this specific neuronal population. The since survival and differentiation of Purlcinje cells parallel present study indicates that NGF and depolarizing signalsin- that observed in vivo. In culture, Purkinje cells develop and fluence survival and morphological differentiation of Purkinje expressseveral traits normally present in vivo (seebelow) and cells in culture. can be visualized immunocytochemically with CaBP antibodies

Table 1. Effects of potassium and NGF on Purkinje cell size and neurite elaboration

Control NGF K’ K’ + NGF Cell diameter&m) 18.5f 0.5 20.4 f 0.6 20.9 f 0.5 22.4 + 0.6* Neuritenumber 4.5 f 0.2 5.6 -t 0.2 5.6 + 0.3 6.3 f 0.2* Neuritelength &m) 218 +24 339 f 36 521 +- 49** 592 + 40** Branchpoint number 3.8 rtr 0.4 5.1 + 0.4 5.2 f 0.4 7.4 + 0.5*** Neuronswith spinyprocesses 8% 30% 38% 78% El 8 rat cerebeilar cells were grown in culture and immunostained for CaBP. Fifty cells, in 3 separate dishes prepared on the same day, were analyzed per experimental condition. Doses of 25 rnM KC1 qd 200 U/ml NGF were used. Statistical analysis was by l-way ANOVA and Scheffe P test. Each value represents mean f SEM. *, Significantly different from control; **, significantly different from control and NGF, ***, significantly different from control, NGF, and potassium. p < 0.05. 466 Cohen-Gory et al. l NGF Regulates Purkinje Cell Development

t c

Figure 3. Immunocytochemical visualization of NGF-R-positive cells in culture. Cerebellar cultures were grown for 7 d in the presence of potassium and NGF and immunostained with the anti-rat 192-IgG antibody. NGF-R-immunopositive cells were recognized as Purkinje cells by morphology and similarity to CaBP-positive cells in sister cul- tures (not shown). Note the punctate staining over the cell body and developing . Scale bar, 20 pm.

(Christakoset al., 1987).However, someofthe epigeneticsignals known to be important for Purkinje cell development are absent in culture. For example, afferent innervation is known to play a critical role during development, and in cell culture normal circuitry is disrupted. Consequently, we examined the role of afferent neurotransmitters as well as NGF in the regulation of Figure 4. lz51-NGF binding to Purkinje cells in culture. ‘Y-NGF bind- Purkinje cell ontogeny. ing was identified as silver grains over CaBP-immunopositive cells in During normal cerebellar development, differentiation of Pur- culture. A, Bright-field illumination reveals the presence of silver grains kinje cellsis influenced by 2 distinct afferent systems:the parallel over an immunoreactive Purkinje cell. B, In control cultures exposed and climbing fibers. In classical studies, Ram&y Cajal provided to excess unlabeled NGF at the time of incubation with radioactive the first evidence that climbing fibers guide Purkinje dendritic NGF, only background silver grains were detected. Cell cultures were prepared from E 18 cerebella and grown for 7 d in serum-supplemented growth (Ram6n y Cajal, 1911); subsequentevidence indicated medium alone. For binding studies, the ‘Y-NGF binding protocol was that climbing fibers play an important role in neurite branching used (see Materials and Methods); cultures were then fixed and im- (Bradley and Berry 1976a,b;Sotelo and Arsenio-Nunes, 1976). munostained for CaBP and treated for autoradiography. The exposure More recently, Mason and Blazeski (1989) have shown that time was 17 d.

Table 2. Effects of aspartate and NGF on Purkinje cell size and neurite branching

Control NGF ASP ASP + NGF Cell diameter (pm) 17.6 +- 0.3 18.1 t 0.3 19.1 & 0.4 21.0 + 0.3** Neurite number 4.3 k 0.1 4.6 k 0.1 4.5 + 0.2 4.8 + 0.2 Neurite length (pm) 263 + 38 302 k 36 352 + 36 404 f 36 Branch point number 3.8 f 0.4 4.4 + 0.3 5.0 * 0.4 5.8 f 0.3* Neurons with spiny processes 1% 2% 12% 26% El8 rat cerebellar cells were grown in culture and immunostained for CaBP. Fifty cells, in 3 separate dishes prepared on the same day, were analyzed per experimental condition. Doses of 10 PM aspartic acid and 200 U/ml NGF were used. Statistical analysis was by l-way ANOVA and ScheffeF test. Each value represents mean + SEM. *, significantly different from control; **, significantly different from control, NGF, and aspartate.p < 0.05. The Journal of Neuroscience, February 1991, 7 7(2) 467

300 ** 200

I

= * = e e E E 200 8 z E t - * 100 - x = 5 0 2 100

‘j

0 0 Control rKF K K+NGF control VRT VRT+NGF

Figure 5. Effects of potassium and NGF on Purkinje cell survival. El8 Figure 6. Effects of VRT and NGF on Purkinje cell survival. El 8 rat rat cerebellar cells were grown in culture in the presence or absence of cerebellar cells were grown in culture and immunostained for CaBP. elevated potassium, NGF, or potassium with NGF and immunostained Numbers of positive cells were obtained as described in Figure 5. In for CaBP. Numbers of positive cells were obtained by analyzing 10% this case, duplicates of 4 independent experiments were analyzed per of the dish area. Triplicates of 4 independent experiments were analyzed condition. Doses of 0.5 PM VRT and 200 U/ml NGF were used, Values per condition. Data is expressed as percent CaBP cell number compared depicted are as follows: Control, 100 ? 5.42%; NGF, 122.71 + 6.29%; to controls grown alone. Each experimental value represents mean cell VRT, 234.85 k 8.24%; VRT + NGF, 286.85 & 8.02%. The actual number + SEM. Statistical analysis was by l-way ANOVA and Scheffe number of cells in representative control cultures was 1527 + 64.17. *, F test. Doses of 25 mM KC1 (KY-and 200 U/ml NGF were used. Values Differs from control and NGF; **, differs from control, NGF, and VRT deuicted are as follows: Control, 100 +- 4.89%: NGF. 108.25 + 4.17%: by p < 0.05. K,-137.71 ? 12.49%; K + NGF; 175.43 ? 15.44%. The actual number of cells in representative control cultures was 1923.2 + 35.2 1. *, Differs from control and NGF, **, differs from control, NGF, and K by p < 0.05. NGF and pharmacologic depolarizing agents enhancePurkinje cell survival The selective expressionof NGF-Rs by Purkinje cells in culture initial afferent interactions between climbing fibers and undif- led usto analyze the putative effectsofNGF on the development ferentiated Purkinje cells occur during late embryonic stages, of this population. Since presynaptic innervation regulates prior to dendritic differentiation. During this early stageof cer- neuronal differentiation in several systems (Black, 1978) and ebellar development, the local expressionof NGF messageand potentially guidesPurkinje cell development, we examined the the expressionof NGF receptorsby Purkinje cells coincide with roles of NGF and depolarizing signals.As an initial approach, the establishmentof initial contacts. Consequent- we examined the possibility that NGF regulatesPurkinje cell ly, theseepigenetic signalsmay interact during development. survival. In control experiments, Purkinje cell survival declined Study ofcerebellar mutants (Rakic and Sidman, 1973; Sotelo, after 1 week in culture. At this time, only a fraction (1 O-20%) 1975; Sadlerand Berry, 1989) and of ablation (Berry of the cells plated survived. Indeed, decreasingsurvival with and Bradley, 1976a;Crepe1 et al., 1980) have also indicated that time is characteristicof cerebellarmacroneurons grown in mixed excitatory stimulation by parallel fibers is required for normal cultures (Hockberger et al., 1989) and has been associatedwith maturation of Purkinje cells. Indeed, parallel fiber stimulation a neuronotoxic activity releasedto the culture medium (Grau- coincides with the peak of NGF-R messageexpression (Buck et Wagemanset al., 1984). Treatment with either high potassium al., 1988). Therefore, we examined the interaction of NGF and or VRT alone increasednumbers of all neuronal types, as in- depolarizing influencesin the regulation of Purkinje cell devel- dicated by an increasein NSE-positive cells (Fig. 1). This result opment. Initially, we examined receptor expressionin vitro. confirmed the well-known effects of depolarizing agents in the promotion of survival of many different populations in culture Purkinje cells expressNGF receptorsin culture (Lasher and Zagon, 1972; Nishi and Berg, 1981; Gallo et al., During early development in vivo, Purkinje cells expresslow- 1987; Hockberger et al., 1987, 1989; Moran and Patel, 1989). and high-affinity NGF-Rs, suggestingpotential responsivity to In addition, however, simultaneousexposure to depolarizing NGF (Cohen-Cory et al., 1989). We now find that Purkinje cells agentsand NGF specifically increasedPurkinje cell numbersin retained the ability to express NGF-Rs in culture. A highly our cultures. NGF, in combination with either high potassium sensitive lZ51-NGFbinding assayrevealed NGF-Rs localized to or VRT, selectively increasedsurvival of CaBP-positive neu- Purkinje cells in control cultures. Moreover, receptor number rons, an effect significantly different from the general action of was enhancedafter treatment with depolarizing agents(see be- potassium or VRT. Yet, NGF alone did not increasesurvival low), allowing the detection of NGF-Rs by monoclonal antibody of Purkinje cells. The observation that depolarizing signalsare binding. The selective localization of NGF-Rs to Purkinje cells required for NGF-induced survival suggeststhat neuronal de- suggestedthat receptor expressionis specifically limited to this polarization may influencereceptivity to the trophic agent.Con- population. In turn, selective expressionsuggested that Purkinje sequently, afferent innervation in combination with trophic sup- cells might be responsiveto NGF in culture. port may govern survival. 466 Cohen-Gory et al. * NGF Regulates Purkinje Cell Development

200

= e r; 8 8

a& 100 = 8

0 control NJ Asp Asp+NGF control NJ GIU GbNGF

Figure 7. Effects of aspartate and NGF on Purkinje cell survival. El 8 Figure 8. Effects of glutamate and NGF on Purkinje cell survival. E 18 rat cerebellar cells were grown in culture, immunostained for CaBP, and rat cerebellar cell cultures were immunostained for CaBP and analyzed analyzed as described for Figure 5. In this case, duplicates of 2 inde- as described for Figure 5. Duplicates of 2 independent experiments were pendent experiments were analyzed per condition. Doses of 10 PM as- analyzed per condition. Doses of 10 PM glutamate (G/u) and 200 U/ml partate (Asp) and 200 U/ml NGF were used for the time in culture. NGF were used. Values depicted are as follows: Control, 100 + 8.96%; Values depicted are as follows: Control, 100 k 4.49%; NGF, 115.0 k NGF, 107.33 k 14.19%; Glu, 124.41 k 9.52%; Glu + NGF, 184.23 + 6.72%; Asp, 99.75 + 8.42%; Asp + NGF, 194.0 k 15.70%. The actual 19.02%. The actual number of cells in representative control cultures number of cells in representative control cultures was 1688 k 99.42. was 1691.4 + 150.95. *, Differs from control by p -C0.05. **, Differs from control, NGF, and aspartate by p < 0.05,

Although it is apparent from our data that NGF, in combi- only in combination with NGF. The fact that naturally occurring nation with depolarizing agents,regulates Purkinje cell survival, transmitters had morerestricted effectssuggests that theseagents 2 other possibilities warrant consideration. First, NGF could act through specific receptors, expressedonly by restricted pop- act to induce cell proliferation, potentially resulting in increased ulations in culture. Alternatively, aspartate and glutamate may Purkinje cell number. This is unlikely, however, becausePur- act through a mechanismdistinct from depolarization. A pu- kinje neurons are postmitotic at the time of dissociation. A tative role for excitatory neurotransmitters as trophic agents second possibility is that NGF could act to induce neuronal during development remains a possibility. differentiation and, therefore, expression of CaBP. While we Differences of effects of pharmacologic depolarizing agents have not definitively excluded the latter possibility, published and excitatory neurotransmitters may also be attributable to results indicate that CaBP expressionis restricted to Purkinje different mechanismsof action. Elevated potassium is known cells in the cerebellum(Christakos et al., 1987).Studies currently to act through a more generalmechanism to alter survival (Iash- in progressare addressingthese different mechanisms. er and Zagon, 1972; Kingsbury et al., 1985; Hockberger et al., 1989; Moran and Patel, 1989). Potassium and VRT, conse- NGF and excitatory neurotransmittersenhance Purkinje cell quently, may stimulate multiple neuronal and non-neuronal survival populations, leading to manifold indirect effects. In contrast, Parallel and climbing fibers constitute the 2 major excitatory since NGF-Rs are restricted to Purkinje cells (seeabove), the afferents of Purkinje cells; L-glutamate and L-aspartateare gen- trophic factor probably acts directly on this specific population. erally regardedas the respective excitatory transmitters. There- fore, we examined the effects of aspartateand glutamate on cell NGF and excitatory neurotransmittersregulate cell size and survival. Treatment with aspartateor glutamate, in combination shape with NGF, markedly enhanced survival of Purkinje cells. Our While afferents are known to regulate the morphological differ- observation suggeststhat the combination of excitatory trans- entiation of Purkinje cells (Rakic and Sidman, 1973; Berry and mitter and NGF is required to enhancesurvival. Although Pur- Bradley 1976b; Bradley and Berry, 1976a,b; Sotelo and Arsenio- kinje cells are sensitive to excitatory amino acid iontophoresis Nunes, 1976; for review, seeIto, 1984), underlying mechanisms at the time of dissociation (Woodward et al., 1971; DuPont et remain unclear. To define mechanismsby which the 2 epigenetic al., 1987), the transmitters alone do not alter survival (Hock- signals,NGF and innervation, may regulate differentiation, we berger et al., 1989; S. Cohen-Gory, C. F. Dreyfus, and I. B. examined the effects of trophic and excitatory stimulation on Black, unpublishedobservations). Our evidence suggestsa novel morphological maturation. mechanism in which trophic factors and afferent stimulation In fact, NGF increasedcell size and promoted neurite elab- interact to promote survival of developing Purkinje cells. oration of developing Purkinje cells. The effects required si- The effects of excitatory neurotransmitters on the survival of multaneous exposure to NGF and either aspartate, glutamate, Purkinje cells apparently differ from those of the general phar- or pharmacologic depolarizing agents. Since only minimal macologic depolarizing agents.While potassiumand VRT alone differencesin maturation were observed with any agent alone, increasedcell survival, aspartate and glutamate elicited effects our results suggestthat presynaptic transmitters and NGF act The Journal of Neuroscience, February 1991, f f(2) 469

CONTROL \

GLUTAMATE ASPARTATE

GLUTAMATE + NGF ASPARTATE + NGF 1 / -\ / r 9i .i-;:ai\ I 7 Figure 9. Morphological maturation of Purkinje cells elicited by excitatory neurotransmitters and/or NGF. Camera lucida tracings of representative CaBP-positive cells were obtained from each experimental condition. Note the marked enhancement in neurite elaboration and in cell size after treatment with glutamate with NGF, or aspartate with NGF. L-Aspartate and L-glutamate were used at 10 PM final concentration, and NGF was used at 200 U/ml. Scale bar, 40 pm. 470 Cohen-Gory et al. f NGF Regulates Purkinje Cell Development in concert to induce differentiation. The mornhological features pression of the nerve growth factor receptor gene in multiple brain I Y exhibited by these neurons in culture are similar to thefeatures areas. Dev Brain Res 44:259-268. Christakos S, Rhoten WB, Feldman SC (1987) Rat calbindin D28K of Purkinje cells developing in viva (Altman, 1972; Berry and purification, quantitation, immunocytochemical localization and Bradley, 1976a; Morris et al., 1985). comparative aspects. In: Methods in enzymology, Vol 139’ (Means Regulation of neurite elaboration bv NGF in combination AR, ed), pp__ 534-551. San Diego: Academic. with aspartate, the putative climbing fiber transmitter, or with Cohen-Cory S, Dreyfus CF, Black IB (1989) Expression of high- and low-affinity nerve growth factor receptors by Purkinje cells in the glutamate, the putative parallel fiber transmitter, suggeststhat developing rat cerebellum. Exp Neurol 105: 104-109. afferent stimulation and NGF interact to regulate dendritic de- Crepe1 F, Delhaye-Bouchaud N, DuPont JL, Sotelo C (1980) Dendritic velopment in vivo. Consequently, the temporal appearanceof and axonal fields of Purkinje cells in developing x-irradiated rat cer- the 2 afferent signal categoriesmay fine tune the regulation of ebellum. A comparative study using intracellular staining with horse dendritic arborization. radish peroxidase. Neuroscience 5:333-347. Dreyfus CF (1989) Effects of nerve growth factor on cholinergic brain neurons. Trends Pharmacol Sci 10:145-149. A potential mechanismfor the interaction of depolarizing DuPont JL, Gardette R, Crepe1 F (1987) Postnatal development of signalsand trophic injluences the chemosensitivity of rat cerebellar Purkinje cells to excitatory ami- During development, individual neurons are exposed to many no acids. An in vitro study. Dev Brain Res 34:59-68. Eckenstein F (1988) Transient expression of NGF-receptor-like im- environmental signalsthat eventually determine cell fate. In- munoreactivity in postnatal rat brain and spinal cord. Brain Res 446: teractions among signals, and appropriate temporospatial ex- 149-154. pression, may regulate normal, orderly neuronal development Escandon E, Chao MV (1989) Developmental expression of the chick- and function. The idea that neural activity modulates trophic en nerve growth factor receptor gene during brain morphogenesis. interactions during development hasgained considerableatten- Dev Brain Res 47: 187-196. Fenton EL (1970) Tissue culture assay of nerve growth factor and the tion over the last few years (for review, see Purves, 1988). A specific antiserum. Exp Cell Res 59:383-393. number of models basedon evidence from both the PNS and Gallo V, Kingsburg A, Balazs R, Jorgensen 0 (1987) The role of CNS suggestthat trophic support may be regulated by neural depolarization on the survival and differentiation of cerebellar granule activity. One possiblemechanism of interaction between excit- cells in culture. J Neurosci 7:2203-22 13. Grau-Wagemans MP, Selak I, Lefebvre PP, Moonen G (1984) Cere- atory amino acidsand trophic agentshas recently been proposed bellar macroneurons in serum-free cultures: evidence for intrinsic for the development of hippocampal neuronsin culture (Matt- neuronotrophic and neuronotoxic activities. Dev Brain Res 15: 1 l- son et al., 1989). 19. The apparent interaction of neurotransmitter and trophic Hatten ME (1985) Neuronal regulation of astroglial morphology and agentsin the regulation of neuron development may potentially proliferation in vitro. J Cell Biol 100:384-396. Hefti F, Dravid A, Hartikka J (1984) Chronic intraventricular injec- be attributable to multiple mechanisms.Our recent studies,in tions of nerve growth factor elevate hippocampal choline acetyltrans- fact, suggesta novel mode of interaction. Preliminary work from ferase activitv in adult rats with martial sentohinnocamnal- . . A lesions. our group indicates that depolarizing influences increaserecep- Brain Res 293:305-3 11. tor expressionby Purkinje-cells in culture (S. Cohen-Gory, k. Hockberger P, Tseng H, Connor JA (1987) Immunocytochemical and electrophysiological differentiation of rat cerebellar granule cells in C. Elliot, C. F. Dreyfus, and I. B. Black, unpublished obser- explant cultures. J Neurosci 7: 1370-1383. vations). This raisesthe possibility that presynaptic stimulation Hockhereer PE. Tseng HY, Connor JA (1989) Development of rat in vivo may foster responsivity to NGF, a contention we are cerebellar Purkinje cells. Electronhvsioloaical orooerties following acute now activelv investigatine. isolation and in long term cult&& J N&ro&!%2258-227 1. Jande SS, Maler L, Lawson DEM (198 1) Immunohistological mapping References of vitamin D-dependent calcium binding protein in the brain. Nature Allendoerfer KL, Shelton DR, Shooter EM, Shatz CJ (1990) Nerve 2941765-767. growth factor receptor immunoreactivity is transiently associated with Kingsbury A, Gallo V, Woddhams PL, Balasz R (1985) Survival, the subplate neurons of the mammalian cerebral cortex. Proc Nat1 morphology and adhesion properties of cerebellar intemeurones cul- Acad Sci USA 87: 187-l 90. tured in chemically defined and serum-supplemented medium. Dev Altman J (1972) Postnatal development of the cerebellar cortex of the Brain Res 17:17-25. rat. II. Phases in the maturation of the Purkinje cells and of the Large TH, Bodary SC, Clegg DO, Weskamp G, Otten U, Reichardt LF molecular layer. J Comp Neurol 145:399-464. (1986) Nerve growth factor gene expression in the developing rat Altman J, Bayer SA (1985) Embryonic development of the rat cere- brain. Science 234:352-355. bellum. III. Regional differences in time of origin. migration and Lasher J, Zagon R (1972) The effect of potassium on neuronal differ- settling of Purkiije cells. J Comp Neurol 23 1:42165. - entiation in culture of dissociated newborn rat cerebellum. Brain Res Bemd P. Martinez HJ. Drevfus CF. Black IB (1988) Localization of 71761-770. high-affinity and low-affinity nerve growth factor reckptors in cultured Lu B, Buck CR, Dreyfus CF, Black IB (1989) Expression of NGF and basal forebrain. Neuroscience 26: 12 l-l 29. NGF receptor mRNA in the develonina brain. Evidence for local Berry M, Bradley P (1976a) The growth of the dendritic trees in the delivery and action of NGF. Exp Ne&o

phate activated glutaminase activity in primary cultures of cerebellar Sotelo C (1975) Anatomical physiological and biochemical studies of granule neurons and astroglial cells during development. Dev Brain the cerebellum from mutant mice. II. Morphological study of cere- Res 46:97-105. bellar cortical neurons and circuits in the weaver mouse. Brain Res Morris RJ, Beech JN, Barber PC, Raisman G (1985) Early stages of 94: 1944. Purkinje cell maturation demonstrated by Thy- 1 immunocytochem- Sotelo C, Arsenio-Nunes ML (1976) Development of Purkinje cells istry on postnatal rat cerebellum. J Neurocytol 14:427452. in the absence of climbing fibers. Brain Res 111:389-395. Nishi R, Berg DK (198 1) Effects of high K+ concentrations on growth Thoenen H, Bandtlow C, Heumann R (1987) The physiological func- and development of ciliary neurons in cell culture. Dev Biol tion of nerve growth factor in the central : comparison 87:301-307. with the periphery. Rev Physiol Biochem Pharmacol 109:145-178. Purves D (1988) Body and brain. A trophic theory of neural connec- Whittemore SR, Seiger A (1987) The expression, localization and tions. Cambridge, MA: Harvard UP. functional significance of p-nerve growth factor in the central nervous Rakic P, Sidman RL (1973) Organization of the cerebellar cortex system. Brain Res Rev 12:439-464. secondary to deficit of granule cells in weaver mutant mice. J Comp Woodward DJ, Hoffer BJ, S&ins GR, Bloom FE (1971) The onto- Neurol 152:133-162. genetic development of synaptic junctions, synaptic activation and Ramon y Cajal S (19 11) Histologie du systeme nerveux de l’homme responsiveness to neurotransmitter substances in rat cerebellar Pur- et des vertebres, Vol 11. Paris: Maloine. kinje cells. Brain Res 34:73-97. Sadler M, Berry M (1989) Topological link-vertex analysis of the Yan Q, Johnson EM (1988) An immunohistochemical study of the growth of Purkinje cell dendritic trees in normal, , and weaver nerve growth factor (NGF) receptor in developing rats. J Neurosci 8: mice. J Comp Neurol 289:260-283. 348 l-3498. Schatteman Gd, Gibbs L, Lanahan AA, Claude P, Bothwell M (1988) Expression of NGF receptor in the developing and adult primate . J Neurosci 8:860-873.