Corticotropin releasing factor enhances survival of cultured GABAergic cerebellar neurons after exposure to a neurotoxin

Developmental Brain Research 151 (2004) 119 – 128 www.elsevier.com/locate/devbrainres

Research report

Corticotropin releasing factor enhances survival of cultured GABAergic cerebellar neurons after exposure to a neurotoxin Paul Madtes Jr. a,b, Kyung-Hoon Lee a,c, James S. King a,d, Richard W. Burry a,d,* a Department of Neuroscience, The Ohio State University, Columbus, OH 43210, USA Biology Department, Mount Vernon Nazarene University, Mount Vernon, OH 43050, USA c Department of Anatomy, Sungkyunkwan University, Suwon, South Korea d Neuroscience Graduate Program, The Ohio State University, Columbus, OH 43210, USA

b

Accepted 13 April 2004 Available online

Abstract Corticotropin-releasing factor (CRF), in addition to its role as a hormone in the stress response, functions as a neuromodulator in the cerebellum, where it enhances both the spontaneous and amino acid induced firing rate of Purkinje cells. In the cerebellum, CRF and its two types of receptors (CRF-R1 and CRF-R2) are present during cerebellar development at ages that precede the onset of afferent ingrowth and synaptogenesis, suggesting a distinct role during early cerebellar development. The present study was undertaken to determine whether CRF enhances the survival of cerebellar neurons, in particular GABAergic neurons. Primary cultures of cerebellar neurons obtained from embryonic day 18 mice were composed primarily, but not exclusively, of GABAergic neurons. Although CRF-R1 is present in most neurons in this culture system, when CRF was added to the medium, no significant change in neuronal survival was observed when compared to control cultures. It is possible that the role for CRF is not seen in growth-promoting culture medium at the plating density chosen for this study and may only be evident when the cells have been exposed to conditions that reduce the likelihood of survival, such as exposure to neurotoxins such as AraC. We propose that, because AraC increases the number of cleaved caspase-3 positive cells, indicating apoptosis, it is possible that a CRF effect involves an inhibition of the apoptotic pathway. Cultures treated with AraC had a decrease in the total number of GABAergic neurons and an increase in apoptotic cells as measured with the apoptotic marker cleaved caspase-3. Co-treatment with CRF rescued many GABAergic neurons. It is interesting to note that apoptotic cells do not exhibit GABA or c-fos positive immunolabeling. Thus, these data support the concept that CRF plays a neuroprotective role in the survival of GABAergic cerebellar neurons in culture after exposure to a neurotoxin. D 2004 Elsevier B.V. All rights reserved. Theme: Development and regeneration Topic: Neurotrophic factors: biological effects Keywords: Apoptosis; Cerebellum; Corticotropin-releasing factor; CRF; GABAergic; Neuroprotective

1. Introduction Although the early studies on corticotropin-releasing factor (CRF) defined a major role for this hormone in the hypothalamo – pituitary –adrenocortical axis in response to stress, more recent data indicate that CRF is involved in a range of effects in the central nervous system (for review, see Refs. [18,55,59]). In the adult cerebellum, CRF acts as a * Corresponding author. Department of Neuroscience, College of Medicine and Public Health, The Ohio State University, 4190 Graves Hall, 333 W. 10th Street, Columbus, OH 43210, USA. Tel.: +1-614-2922814; fax: +1-614-688-8742. E-mail address: [email protected] (R.W. Burry). 0165-3806/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.devbrainres.2004.04.009

neuromodulator by enhancing the responsiveness of Purkinje cells to excitatory amino acids by decreasing the duration of an afterhyperpolarizing potential [4,5,28]. In the mouse cerebellum, CRF is present at embryonic day (E) E10 [6] and its receptors are present from E11 through adulthood [6,40,45,51]. Thus, CRF and its receptors are present during ages that precede by several days the onset of synaptogenesis. Our working hypothesis is that CRF has a role that is distinct from that occurring in the mature brain, possibly acting as a trophic factor that influences cell survival or differentiation. A previous study using cerebellar cultures has shown that CRF has a mitogenic effect on immature astrocytes, which could be inhibited by astressin, a competitive CRF receptor antago-

120

P. Madtes Jr. et al. / Developmental Brain Research 151 (2004) 119–128

nist [35], suggesting CRF could play a gliotrophic role during development. The intent of the present study was to determine if CRF plays a role during neuronal development. As a model system, cerebellar cultures from E18 mice were prepared. At this age, the primary neurons present in the cultures were shown to be GABAergic [57]. One possible influence CRF might have on developing neurons is to increase survival. We tested this possibility by culturing cerebellar neurons in the absence and in the presence of CRF. Little change in cell survival was noted. It also is possible that the role for CRF is more adaptive, being only evident when the cells have been exposed to conditions that reduce the likelihood of survival, such as exposure to neurotoxins. If so, CRF would exhibit a neuroprotective role in the cerebellum only after the cells had been stressed or traumatized. Such a role is quite feasible since it has been reported that a member of the CRF family of peptides, urocortin, protects cultured hippocampal cells from E16 rats from oxidative and excitotoxic cell death [54]. One well-established agent that acts as a neurotoxin is cytarabine (cytosine 1 h-D-arabinofuranoside, AraC), which is frequently used in treatment of acute leukemia [64,65]. One of the side effects of chemotherapy, in addition to bone marrow toxicity, is delayed toxicities comprising cognitive deficits, aphasia and progressive dementia [64]. Consequently, there is evidence that AraC-induced neurotoxicity is a problem for these patients. AraC potentiates cell death by inducing an apoptosis pathway in developing and in mature neurons and glia [1,16,19,23,32,38,48,52,63,68] by damaging cellular DNA [32,52,68]. This mechanism is especially interesting for the present study as CRF has been shown to suppress apoptosis by acting upstream of the activation of caspases in neurons [21,56]. Since AraC stresses neurons via a pathway that is blocked by CRF Y79 human retinoblastoma cells, we decided to expose our cultures to AraC to investigate whether CRF may exert a role in cerebellar neuronal survival by inhibiting the effect of AraC. The findings presented here suggest that CRF has a neuroprotective role in the survival of GABAergic cerebellar neurons in culture after AraC-induced stress. In addition, data here suggest that the AraC-induced apoptotic pathway acts to eliminate this neuronal phenotype (i.e., loss of GABA) prior to completion of apoptosis.

2. Methods 2.1. Primary cell cultures Primary cell cultures of cerebella from E18 mice were established following the methods of Fischer [26] and Schilling et al. [57]. Briefly, 18 days after mating, pregnant C57Bl/6 mice were anesthetized using an intraperitoneal injection of 2.5% Avertin (0.2 ml/10 g). The

embryos were removed, decapitated and placed in phosphate-buffered saline (PBS). The cerebella were dissected from the brainstem and the meninges were removed. The PBS then was replaced with Basal Medium Eagles solution (BME Complete Medium) with transferrin (100 Ag/ml), aprotinin (1 Ag/ml), selenite (30 nM), triiodothyronine (0.2 ng/ml), insulin (10 Ag/ml), NaHCO3 (26.18 mM), glucose (0.25%), pyruvate (1 mM), glutamate (2 mM) and bovine serum albumin (BSA) (1.0 mg/ml) plus 5% heat-inactivated horse serum. The cerebella then were triturated 15 times before the cells were passed through a 40-Am cell strainer (Falcon). After centrifugation at 150  g for 5 min, the pellet was resuspended in BME Complete Medium with 5% heat-inactivated horse serum. Cells then were counted using a hemocytometer (Brightline) and plated at a density of 100  103 cells/cm2 into a four-chamber (1.8 cm2) slide (Lab-Tek) that was pretreated with poly-D-lysine (10 mg/100 ml). The cultures were incubated in 10% CO2 at 37 jC. One day after plating, the cells were fed with serum-free BME Complete Medium containing the various agents as delineated below. Every 3 days, half the volume of medium was removed and replaced with fresh complete medium containing the appropriate agent. Each culture was treated for either 1 day or for 6 days with one of four conditions: (a) BME complete medium only (control), (b) BME complete medium with 1 AM CRF, (c) BME complete medium with 5 AM AraC, and (d) BME complete medium with 5 AM AraC and either 0.1 or 1 AM CRF. At 6 days of treatment (DT), cultures were fixed for immunocytochemistry. 2.2. Immunocytochemistry Cells in culture were identified using double- or triplelabel fluorescence immunocytochemistry. Briefly, cultures were fixed with 4% paraformaldehyde, 0.1 M phosphate buffer (pH 7.1), 50 mM sucrose and 0.4 mM CaCl2 for 30 min, and then rinsed with PBS+ (PBS containing 0.1% saponin, 0.02% sodium azide and 1 mg/ml BSA). Cells were incubated at room temperature overnight in a single solution containing one of the following combinations: (1) a primary antibody for calbindin (generated in mouse) and a primary antibody for GABA (generated in rabbit), (2) a primary antibody for GABA (generated in rabbit) and a primary antibody for CRF-R1 (generated in goat), (3) a primary antibody for GABA (generated in mouse) and a primary antibody for cleaved caspase-3 (generated in rabbit), (4) a primary antibody for GABA (generated in mouse), a primary antibody for cleaved caspase-3 (generated in rabbit), and a primary antibody for c-fos (generated in goat). Subsequently, cells were rinsed with PBS+ and incubated in species-specific secondary antibodies for 1 – 3 h at room temperature. After rinsing with PBS+, slides were coverslipped with Immu-Mount and sealed with clear fingernail polish to avoid dehydration.

P. Madtes Jr. et al. / Developmental Brain Research 151 (2004) 119–128

2.3. Image acquisition and analysis Data were obtained using one of two microscopes: (1) a Zeiss Axioskop 2 fluorescence microscope equipped with a digital camera and Axiovision 3.0 Image Acquisition Software (Zeiss) or (2) a Zeiss 510 META laser scanning confocal on a Zeiss Axiopan 2 microscope in the Ohio State Campus Microscopy and Imaging Facility. For images obtained using the Axioskop 2, the acquisition time and gain were standardized to obtain the optimal gray-scale distribution of each image. Using GABA-labeled cells as the basis for monitoring changes in cell survival, a series of consecutive digital images (180 Am2/sample, spanning the width of the chamber) were acquired for each slide, to ensure that a uniform area was analyzed and an adequate sample size for quantification was obtained. This approach resulted in a minimum of 200 cells per condition. The cell counts are expressed as mean percentages of control numbers F standard errors of the mean. Significance was determined using Student’s t-test, two-tailed. For images obtained using the 510 META confocal microscope, representative cells were sampled. Confocal images were imported into Adobe Photoshop 7.0 and pseudocolorized as follows: GABAgreen, cleaved caspase-3-red and c-fos-blue. 2.4. Chemicals and antibodies The following chemicals were obtained from Sigma: aprotinin, Avertin (tribromoethanol), BSA, cytosine arabinoside, glucose, glutamate, insulin, lysine, paraformaldehyde, pyruvate, saponin, selenite, sodium bicarbonate, sodium chloride, sodium phosphate (monobasic), sucrose, triiodothyronine and transferrin. The following chemicals were obtained from various sources as indicated: BME (Gibco), human, rat CRF (Bachem), horse serum (Gibco), Immu-Mount (Thermo Shandon) and sodium azide (EM Science). Primary antibodies were obtained from various sources as indicated and used at the following dilutions: Sigma-mouse calbindin (1:500), goat c-fos (1:500), rabbit cleaved caspase-3 (1:500), mouse GABA (1:300), rabbit GABA (1:1000), mouse GFAP (1:500) and Santa Cruz-goat CRF-R1 (1:3000). Fluorescent secondary antibodies were used as follows: Jackson Labs-Cy-2 (1:500 for mouse GABA and 1:800 for rabbit GABA) and Cy-3 (1:800 for CRF-R1 and for cleaved caspase-3); Molecular Probes-Alexa Fluor 647 (1:800 for c-fos).

3. Results To determine if CRF enhances the survival of GABAergic neurons in growth-promoting culture medium, we used primary cell cultures using cerebellar cells from E18 mice because neurons that are known to be GABAergic have been born by this age and these cultures have been reported to contain a variety of cerebellar neurons, including GABAergic

121

interneurons, e.g., basket cells, stellate cells and/or Golgi cells, which label with GABA antibodies [2,57]. Immunohistochemistry indicated our cultures contained numerous GABAergic neurons and only a few scattered Purkinje cells, identified by calbindin immuno-positive label (data not shown). Antibodies to GFAP revealed no glial cells in the cultures, although rare large flat cells were observed under phase contrast optics (data not shown). Based on these observations, we concluded these cultures would permit study of the effects of CRF on neurons that were immunoreactive to the GABA antibody, presumably cerebellar interneurons. Representative images of the cultures in growth-promoting medium, taken from the same visual field, are shown in Fig. (1A –D). The phase image (Fig. 1A) shows that these cultures have a significant number of neurons present, and no glial cells are observed in this micrograph. It is evident that the neuronal cell bodies tend to cluster by 6 DT. In addition, there is extensive growth of neurites on the substrate between clusters with many neurites found in bundles between aggregates of neuronal cell bodies. Based on these observations, it is evident that these culturing conditions will permit growth of healthy cerebellar neurons. The general appearance of normal GABAergic cells in culture is seen in Fig. 1B, again using the same field as in Fig. 1A. Not all neurons labeled with the GABA antibody (as indicated by the small arrows), indicating only a subpopulation of neurons in culture are being studied. The cell bodies of the GABA-positive cells are clustered, with extensive processes that are highly branched. The label is most intense over the cell bodies and over primary processes. To verify the presence of CRF receptors on cells under these culturing conditions, cerebellar cultures fixed after 6 DT were labeled with antibodies to CRF-R1. A representative image, from the same field as shown in Fig. 1A, is presented in Fig. 1C. (The insert is an enlargement of the cluster of cell bodies indicated by the large arrow.) The cell bodies of the CRF-R1-positive cells are clustered, with extensive processes that are highly branched. The label appears to be most intense in a perinuclear distribution in the cell bodies of positive cells, consistent with synthesis in the endoplasmic reticulum and subsequent transport to the Golgi apparatus. Fine label is seen at the plasma membrane of cell bodies and in multiple neurite processes. Cells were positive for CRF-R1 under all experimental conditions (data not shown), indicating the treatments did not change the presence of CRF receptors. When the images of GABA and CRF-R1 were merged (Fig. 1D), the vast majority of GABA-positive cells were found to be positive for CRFR1, indicating they have the ability to respond to CRF. In order to determine the maximal effect of CRF on neuronal survival, we examined cultures over a time course of several days. While it is probable that steps necessary for survival are initiated within minutes of exposure to CRF, treatments were done for 6 days to maximize any phenotypic changes (i.e., number of cells surviving, complexity of neurite outgrowth, maturity of

122

P. Madtes Jr. et al. / Developmental Brain Research 151 (2004) 119–128

Fig. 1. Images of representative cells from 6 DT cerebellar cultures. Control cultures of the same field are shown as a phase image (A), a GABA-labeled fluorescent image (B) and a CRF-R1-labeled fluorescent image (C). Note: not all cells are GABA positive, as indicated by small arrows. The insert in (C) is an enlargement of the area indicated by the large arrow. Image D represents an overlay of GABA-labeled cells (B) and CRF-R1-labeled cells (C). Magnification = 200 .

cells) that might occur. Preliminary studies showed the maximum phenotypic change was seen at 6 DT. By this time, the cultures were mature and healthy, and treated cultures showed the various effects reported here. In addition, initial studies indicated that plating densities higher than 100  103 cells/cm2 masked any effect due to CRF, possibly due to endogenous factors released from closely packed neurons. While higher plating densities permit survival of many types of cells (including Purkinje cells) [2,57], it also precludes study of a specific effect of CRF on cell survival since endogenous factors appear to mask the effects of exogenous factors. Specifically, the plating density of 150  103 cells/cm2 partially reduced the observed effects of CRF and no CRF-induced effects were present when the plating density was 200  103 cells/cm2 or higher (data not shown). To test whether CRF influenced the survival of GABAergic neurons, cultures were treated for 6 days without or with 1 AM CRF. The overall health of the cultures appears to be comparable based on the number and appearance of cells in the phase microscope, and the number GABAergic neurons present detected with immunocytochemistry (Fig. 2A and B versus C and D). Thus, CRF had no effect on cell survival in growth-promoting medium at the plating density chosen for this study. An alternative role for CRF would be to play a more adaptive role, influencing neuronal survival under conditions in which the neurons have been subjected to insult, i.e., stressors such as neurotoxins. We elected to treat our

cultures with a known neurotoxin, AraC, to determine whether it would stress the neurons. AraC has been shown to induce neuronal cell death, via an apoptotic mechanism [16,32,38,48,52]. As shown in Fig. 2E and F, the effect of AraC on GABAergic neurons was pronounced. AraC treatment greatly decreased the overall number of neurons in the culture, including GABAergic neurons, at 6 DT; the few neurons that remained were isolated or formed small clusters with one or two other neurons. In addition, the neurites that were present were very simple, with few, if any, branches. There was a tendency for the neuronal processes to form a circular pattern, suggesting the processes remained in the vicinity of the cell body of the neuron from which they originated. Apparently, the outgrowth of neurites was affected. The paucity of immunolabeled GABAergic and unlabeled neurons in these cultures (Fig. 2E and F) indicates the effect of AraC is not limited to GABA-positive neurons. To determine if CRF might enhance neuronal survival when the neurons had been stressed, these cultures were treated with AraC in the presence of CRF for 6 days. This resulted in a significant increase in cell survival and neurite outgrowth (Fig. 2G and H). GABA-positive neurons are more numerous than in cultures treated with AraC only. In addition, there is a tendency for the cell bodies to cluster, although it is not as extensive as control cultures. Furthermore, the neurites in CRF and AraC-treated cultures take on a more complex appearance, i.e., there is more branching, even to the point that arborization occasionally appears in the terminal portion of the outgrowth (Fig. 2H, arrows) than

P. Madtes Jr. et al. / Developmental Brain Research 151 (2004) 119–128

123

of GABAergic cells to 28% of control levels, which was not a significant change, compared to AraC alone. One question that arises is a possible mechanism for this neuroprotective role of CRF. Since AraC is known to induce an apoptotic pathway [16,32,38,48,52], it seemed reasonable to test whether any AraC-induced apoptosis was occurring in these cultures. Because the induction of apoptosis is expected to be a very rapid event, occurring possible within hours of the initial exposure, cultures were treated for 1 day, then fixed and prepared for immunocytochemistry with cleaved caspase-3, an intermediate in the apoptotic pathway induced by AraC [8,56,60]. GABAergic neurons were also labeled for c-fos as an indicator of cell activity since this molecule has been implicated in cell signaling [7,14,33]. Cells that label for cleaved caspase-3 (Fig. 4, arrows) were more abundant when the cultures were exposed to AraC (Fig. 4C). In contrast, in healthy neurons, GABA is localized in the cytoplasm of the cell body and neurites (Fig. 4, green) while c-fos is found in nuclei (Fig. 4, blue). Cells that were positive for cleaved caspase-3 (Fig. 4, red) did not have neurites. Finally, GABA-positive cells have c-fos-positive nuclei, but neurons that are positive for cleaved caspase-3 rarely appear to be positive for GABA and c-fos. Following this short incubation with AraC, the number of cleaved caspase-3 positive cells increased, consistent with the loss of neurons see at 6 days. In addition, labeling for GABA and c-fos as well as the presence of neurites was not see in cleaved caspase-3 positive cells

Fig. 2. Images of representative cells from 6 DT cerebellar cultures. The same field is shown as both phase images (left column) and GABA-labeled fluorescent images (right column). Control cultures are shown as (A) and (B), cultures treated with CRF alone (C) and (D), AraC alone (E) and (F), and AraC and CRF together (G) and (H). Arrows in (H) indicate a more complex neurite appearance, i.e., more branching, even to the point of arborization. Magnification = 200 .

in AraC-treated cultures. Again, the neuronal distribution in the phase image (Fig. 2G) indicates the effect of AraC is not limited to GABA-positive neurons. To determine if these effects result in actual differences in the number of cells present, cell counts were made for GABAergic neurons (Fig. 3). After treatment for 6 days with 1 AM CRF, the number of GABAergic neurons was not significantly different from control levels. In contrast, AraC treatment caused a dramatic decrease in the number of GABAergic cells, down to 19% of control (untreated) levels ( p < 0.001) (Fig. 3). However, co-treatment with 1 AM CRF increased the number of GABAergic cells to 53% of control levels, which is significantly higher than the AraC treatment alone ( p < 0.01). To determine whether the neuroprotective effect of CRF was dependent upon the amount of CRF present, cultures were co-treated with 0.1 AM CRF. Cotreatment with AraC and 0.1 AM CRF increased the number

Fig. 3. Effect of CRF on GABAergic neurons in culture. Primary cerebellar cultures were treated with CRF alone, AraC alone or both, for 7 days. Cell counts from over 200 cells per condition are presented as a percentage of control conditions (complete medium). Significance was determined using Student’s t-test, two-tailed: * represents p < 0.001, with respect to control, and # represents p < 0.01, with respect to AraC, and + represents p < 0.07, with respect to AraC + 0.1 AM CRF.

124

P. Madtes Jr. et al. / Developmental Brain Research 151 (2004) 119–128

Fig. 4. Confocal images of neurons immunolabeled for GABA (green), cleaved caspase-3 (red) and/or c-fos (blue). The immunolabel for cleaved caspase-3 and c-fos within cells in cultures of GABAergic neurons was localized after 1 DT. GABA is localized in the cytoplasm whereas c-fos labels nuclei. Most GABApositive cells have c-fos positive nuclei. Cleaved caspase-3 immunolabeling (white arrows) rarely appears in neurons labeled with GABA and c-fos. Magnification = 630 .

indicating a loss of neuronal characteristics in neurons prior to entering the apoptotic pathway.

4. Discussion Our working hypothesis is that CRF acts during embryonic and early postnatal development as a neurotrophic factor. In particular, since CRF has been reported to have a cytoprotective effect in a variety of cells in a neuroprotective role we propose that it acts to enhance neuronal survival [22,29,42,53]. To investigate this hypothesis, we used cerebellar cell cultures since these have been demonstrated to serve as a model system for studying development [9 –12]. First, we examined the effect of CRF on healthy cultures from E18 mouse cerebella containing populations of neuronal cells including post-mitotic Purkinje cells; interneurons including basket cells, stellate cells and Golgi cells; and immature granule cells [2,57]. Because neuronal survival following the dissociation procedure is greatest for cell populations that have not yet established extensive connections with other neurons and are still undergoing mitosis, migration, and/or differentiation, the neurons that are most likely to

survive under the culturing conditions used here include the GABAergic interneurons, i.e., basket cells, stellate cells and Golgi cells. The birth dates for these cells (E12– 15) fit the time frame from which the cells used in this study arise [46,47]. Although these neurons are born at E12 –15, data indicate they continue to divide as they migrate from the ventricular zone to the cerebellar cortex [47]. Thus, they remain in an immature state even at E18 and survive well in culture. A previous study using cells from the cerebella of E18 mice investigated the nature of cells plated at high density in medium containing serum. This approach yielded astrocytes without neurons. It was found that these cultures, when grown in the presence of CRF for 14 DIV, showed an increased number of astrocytes compared to cultures grown without CRF [35]. These findings demonstrated that, in the glial culture system, CRF induced a proliferation of immature astrocytes that was verified with enhanced levels of BrdU labeling. It was concluded that CRF plays a gliotrophic role during cerebellar development. To preclude any trophic effect from glia in our study of GABAergic neurons, glial cells were eliminated from the cultures by using low plating density and supplement containing growth-promoting medium that contained no serum [2,57].

P. Madtes Jr. et al. / Developmental Brain Research 151 (2004) 119–128

In our cultures, the number of GABAergic neurons found following treatment with CRF in neuron-conducive medium was not significantly different from control culturing conditions without CRF. Following the growth for 6 days in culture, the cells in both conditions extended numerous neurites and the cell bodies began to cluster. At the end of the treatment, the neurons in both conditions appeared normal and healthy, having numerous processes. These observations suggest that CRF in growth promoting medium does not have an effect on survival under these conditions. However, it is possible that the number of neurons that survive the dissociation and plating process involved in establishing the cultures with growth-promoting medium is at a maximum. Thus, no enhanced survival would be evident during CRF treatment alone. Our hypothesis is that CRF enhances survival; however, the culture conditions under which that occurs may not be in growth promoting medium alone. Possibly, the survival rule can only be detected when the cells have been insulted (i.e., stressed or traumatized). For example, conditions that reduce the likelihood of survival, such as exposure to neurotoxins, might induce apoptosis. If CRF were to inhibit the apoptotic pathway, CRF would enhance survival, exhibiting a neuroprotective role in the cerebellum only after the cells had been stressed or traumatized. This proposal seems plausible in light of reports that, after exposure to various agents or conditions, CRF [24,29,53] and urocortin [54] reduce the number of cells lost, an effect which might occur via a caspase-dependent apoptotic pathway [56], although the mechanism of induction is not known. One specific means by which neurons can be insulted is by exposure to a neurotoxin. This is especially important in light of the fact that, presently, a common approach to treating patients who have leukemia is to use chemotherapy [64 – 66]. One well-established agent frequently used chemotherapeutically in the treatment of acute leukemia is cytarabine (cytosine 1 h-D-arabinofuranoside), which acts as a neurotoxin [64,65]. Consequently, there is a significantly greater possibility that AraC-induced neurotoxicity will increasingly be a problem for patients. AraC is known to induce stress in cells by activating an apoptotic pathway [1,16,23,48,52]. AraC induces apoptosis of cultured cerebellar granule cells by over-expression of glyceraldehyde-3-phosphate dehydrogenase [38] and it acts on cerebellar granule neurons early during differentiation, possibly via a cyclin-dependent kinase (cdk)-dependent pathway involving caspase [16]. AraC also is especially useful since it acts via a p53-dependent, JNK-independent mechanism and does not inhibit a trophic factor-maintained pathway required for survival [1]. To address the possibility that AraC might stress neurons in our cultures, the cells were treated with AraC. It was found that the overall number of neurons was decreased. Furthermore, the number of GABAergic cells was significantly reduced to only 19% of the number in control cultures. Co-treatment with AraC and 1 AM CRF resulted

125

in a significant increase in the number of GABAergic cells. The indication that a 10-fold lower concentration of CRF present in the culture medium did not influence the neuronal survival to the same degree supports the notion that the survival response is specific to CRF. Further, the presence of CRF receptors on these cells shows that CRF ligand could mediate such an effect. The finding that CRF increases neuronal survival in GABAergic neurons after exposure to a neurotoxin suggests CRF may act as a neuroprotective agent reducing cell loss due to apoptosis [56]. This is consistent with a previous report that showed that urocortin, a member of the CRF family of peptides, exhibited a neuroprotective effect against glutamate-induced neurotoxicity in cultured hippocampal neurons, acting via CRF-R1 type receptors. The fact that most of the GABAergic neurons in our culture contain CRF-R1 suggests CRF is likely acting through this receptor type in our culture system. The mechanism(s) by which CRF affects neuronal survival during development are unknown. The neuroprotective effect for CRF found in the present study suggests that CRF may influence cerebellar development by attenuating apoptosis. One possible mechanism by which CRF may do this has been suggested recently when it was found that a cryoprotective effect of CRF in Y79 human retinoblastoma cells is exerted by suppressing pro-apoptotic pathways, probably at a site upstream from procaspase-3 [56]. This mechanism apparently involves CRF-R1 receptors. Since our cultures have been shown to contain cells that express CRF-R1 receptors extensively, and since AraC is thought to be involved with an apoptotic pathway, it is not inconceivable that the neuronal survival effect reported here might employ a similar mechanism. Immunocytochemistry of AraC-treated cultures showed increased labeling for cleaved caspases-3, indicative of apoptosis. Whereas normal, healthy cells exhibited robust immunolabeling for GABA over the cytoplasm and robust immunolabeling for c-fos over the nucleus, in cleaved caspase-3positive cells, no GABA or c-fos immunolabel was found. As expected, the cells appear to go through a series of stages, beginning with low levels of immunolabel for cleaved caspase-3 in the peripheral aspects of the cytoplasm and ending with label covering the entire cell, indicating destruction of the nuclear membrane, just before cell death. The observation that no cell that had robust labeling for GABA and c-fos also had intense labeling for cleaved caspase-3 suggests that the apoptotic pathway eliminates the expression of the GABAergic phenotype early in the process. A neurotrophic role during development has been proposed for a variety of molecules including GABA [27], glutamate [33,34,36,50,61], serine and glycine [31], and insulin-like growth factor-I (IGF-I) [13,62,67]. In addition, one key group of molecules that are critical to development is the neurotrophins. Neurotrophins regulate the survival and differentiation of a variety of cerebellar cells, including granule cells, Purkinje cells, basket cells and stellate cells

126

P. Madtes Jr. et al. / Developmental Brain Research 151 (2004) 119–128

(for review, see Ref. [43]). Specifically, BDNF, neurotrophin (NT)-3 or NT-4/5 increase cell survival [37,41] and process outgrowth [49,58]. Furthermore, the finding that neurotrophins also are linked to apoptotic pathways [3,25] indicates the link between CRF, cell signaling, and cell survival during development might be very complex. Since G-protein-coupled receptors (GPCRs) have been recognized as important mediators of cellular growth and differentiation [20,44], and CRF acts through a GPCR, it is possible that CRF might employ similar pathways typical of other G protein-coupled receptors, e.g., the neurotrophins. The apparent connection between neurotrophin signaling and neuronal development is especially intriguing [30,39], including a link to apoptotic pathways [3,25], and suggests that CRF may act via one of the signaling pathways to influence the development of GABAergic neurons, possibly involving c-Jun kinases, which act with different roles during development and during stress in cerebellar granule neurons [15]. Clearly, the role that a particular neurotrophic molecule plays depends on a variety of factors and their interactions. The findings presented here demonstrate that CRF is capable of influencing cerebellar neuronal development by increasing neuronal survival in GABAergic neurons after the cells have been stressed by treatment with a neurotoxin, i.e., AraC. The mechanism by which this role is carried out remains to be elucidated; however, the data here show that CRF diminishes the effect of an induced apoptotic pathway. While CRF might act as a trophic molecule during cerebellar development, its effects may be limited to stimulating survival following stressful stimuli. Our findings also suggest that, when using cytarabine (AraC) as a therapeutic agent, consideration be given to its neurotoxic effects, and the possibility that other agents might be used to reduce or eliminate those effects [16,17].

Acknowledgements The authors express gratitude to Dr. Georgia A. Bishop for critical readings of this manuscript. The authors also thank Mr. Dan Boulter for technical assistance. This work was supported by a grant from the NIH (NS-08798).

References [1] C.N. Anderson, A.M. Tolkovsky, A role for MAPK/ERK in sympathetic neuron survival: protection against a p53-dependent, JNK-independent induction of apoptosis by cytosine arabinoside, J. Neurosci. 19 (1999) 664 – 673. [2] S.L. Baader, K. Schilling, Glutamate receptors mediate dynamic regulation of nitric oxide synthase expression in cerebellar granule cells, J. Neurosci. 16 (1996) 1440 – 1449. [3] C. Bechade, C. Mallecourt, F. Sedel, S. Vyas, A. Triller, Motoneuronderived neurotrophin-3 is a survival factor for PAX2-expressing spinal interneurons, J. Neurosci. 22 (2002) 8779 – 8784.

[4] G.A. Bishop, Neuromodulatory effects of corticotropin releasing factor on cerebellar Purkinje cells: an in vivo study in the cat, Neuroscience 39 (1990) 251 – 257. [5] G.A. Bishop, C.W. Kerr, The physiological effects of peptides and serotonin on Purkinje cell activity, Prog. Neurobiol. 39 (1992) 475 – 492. [6] G.A. Bishop, J.S. King, Corticotropin releasing factor in the embryonic mouse cerebellum, Exp. Neurol. 160 (1999) 489 – 499. [7] J.C. Bittencourt, P.E. Sawchenko, Do centrally administered neuropeptides access cognate receptors?: an analysis in the central corticotropin-releasing factor system, J. Neurosci. 20 (2000) 1142 – 1156. [8] A. Bobba, A. Atlante, S. Giannattasio, G. Sgaramella, P. Calissano, E. Marra, Early release and subsequent caspase-mediated degradation of cytochrome c in apoptotic cerebellar granule cells, FEBS Lett. 457 (1999) 126 – 130. [9] R.W. Burry, A unit density method of grain analysis used to identify GABAergic neurons for electron microscopic autoradiographs, J. Neurosci. Methods 5 (1982) 369 – 381. [10] R.W. Burry, Protein synthesis requirement for the formation of synaptic elements, Brain Res. 344 (1985) 109 – 119. [11] R.W. Burry, R.S. Lasher, A quantitative electron microscopic study of synapse formation in dispersed cell cultures of rat cerebellum stained either by Os-UL or by E-PTA, Brain Res. 147 (1978) 1 – 15. [12] R.W. Burry, J.J. Lah, D.M. Hayes, Redistribution of GAP-43 during growth cone development in vitro; immunocytochemical studies, J. Neurocytol. 20 (1991) 133 – 144. [13] D. Chrysis, A.S. Calikoglu, P. Ye, A.J. D’Ercole, Insulin-like growth factor-I overexpression attenuates cerebellar apoptosis by altering the expression of Bcl family proteins in a developmentally specific manner, J. Neurosci. 21 (2001) 1481 – 1489. [14] E.T. Coffey, K.E. Akerman, M.J. Courtney, Brain derived neurotrophic factor induces a rapid upregulation of synaptophysin and tau proteins via the neurotrophin receptor TrkB in rat cerebellar granule cells, Neurosci. Lett. 227 (1997) 177 – 180. [15] E.T. Coffey, G. Smiciene, V. Hongisto, J. Cao, S. Brecht, T. Herdegen, M.J. Courtney, c-Jun N-terminal protein kinase (JNK) 2/3 is specifically activated by stress, mediating c-Jun activation, in the presence of constitutive JNK1 activity in cerebellar neurons, J. Neurosci. 22 (2002) 4335 – 4345. [16] M.J. Courtney, E.T. Coffey, The mechanism of Ara-C-induced apoptosis of differentiating cerebellar granule neurons, Eur. J. Neurosci. 11 (1999) 1073 – 1084. [17] M.J. Courtney, K.E. Akerman, E.T. Coffey, Neurotrophins protect cultured cerebellar granule neurons against the early phase of cell death by a two-component mechanism, J. Neurosci. 17 (1997) 4201 – 4211. [18] E.B. De Souza, Corticotropin-releasing factor receptors: physiology, pharmacology, biochemistry and role in central nervous system and immune disorders, Psychoneuroendocrinology 20 (1995) 789 – 819. [19] C. D’Sa, B.J. Klocke, F. Cecconi, T. Lindsten, C.B. Thompson, S.J. Korsmeyer, R.A. Flavell, K.A. Roth, Caspase regulation of genotoxin-induced neural precursor cell death, J. Neurosci. Res. 74 (2003) 435 – 445. [20] N. Dhanasekaran, L.E. Heasley, G.L. Johnson, G protein-coupled receptor systems involved in cell growth and oncogenesis, Endocr. Rev. 16 (1995) 259 – 270. [21] K. Eckart, O. Jahn, J. Radulovic, M. Radulovic, T. Blank, O. Stiedl, O. Brauns, H. Tezval, T. Zeyda, J. Spiess, Pharmacology and biology of corticotropin-releasing factor (CRF) receptors, Recept. Channels 8 (2002) 163 – 177. [22] C.R. Elliott-Hunt, J. Kazlauskaite, G.J. Wilde, D.K. Grammatopoulos, E.W. Hillhouse, Potential signalling pathways underlying corticotrophin-releasing hormone-mediated neuroprotection from excitotoxicity in rat hippocampus, J. Neurochem. 80 (2002) 416 – 425. [23] Y. Enokido, T. Araki, S. Aizawa, H. Hatanaka, p53 involves cytosine arabinoside-induced apoptosis in cultured cerebellar granule neurons, Neurosci. Lett. 203 (1996) 1 – 4.

P. Madtes Jr. et al. / Developmental Brain Research 151 (2004) 119–128 [24] L. Facci, D.A. Stevens, M. Pangallo, D. Franceschini, S.D. Skaper, P.J. Strijbos, Corticotropin-releasing factor (CRF) and related peptides confer neuroprotection via type 1 CRF receptors, Neuropharmacology 45 (2003) 623 – 636. [25] U. Felderhoff-Mueser, M. Sifringer, S. Pesditschek, H. Kuckuck, A. Moysich, P. Bittigau, C. Ikonomidou, Pathways leading to apoptotic neurodegeneration following trauma to the developing rat brain, Neurobiol. Dis. 11 (2002) 231 – 245. [26] G. Fischer, Cultivation of mouse cerebellar cells in serum free, hormonally defined media: survival of neurons, Neurosci. Lett. 28 (1982) 325 – 329. [27] M.L. Fiszman, L.N. Borodinsky, J.H. Neale, GABA induces proliferation of immature cerebellar granule cells grown in vitro, Brain Res. Dev. Brain Res. 115 (1999) 1 – 8. [28] E.A. Fox, D.L. Gruol, Corticotropin-releasing factor suppresses the after hyperpolarization in cerebellar Purkinje neurons, Neurosci. Lett. 149 (1993) 103 – 107. [29] M.W. Fox, R.E. Anderson, F.B. Meyer, Neuroprotection by corticotropin releasing factor during hypoxia in rat brain, Stroke 24 (1993) 1072 – 1075. [30] W.J. Friedman, L.A. Greene, Neurotrophin signaling via Trks and p75, Exp. Cell Res. 253 (1999) 131 – 142. [31] S. Furuya, T. Tabata, J. Mitoma, K. Yamada, M. Yamasaki, A. Makino, T. Yamamoto, M. Watanabe, M. Kano, Y. Hirabayashi, Lserine and glycine serve as major astroglia-derived trophic factors for cerebellar Purkinje neurons, Proc. Natl. Acad. Sci. U. S. A. 97 (2000) 11528 – 11533. [32] H.M. Geller, K.Y. Cheng, N.K. Goldsmith, A.A. Romero, A.L. Zhang, E.J. Morris, L. Grandison, Oxidative stress mediates neuronal DNA damage and apoptosis in response to cytosine arabinoside, J. Neurochem. 78 (2001) 265 – 275. [33] R. Griffiths, C. Malcolm, L. Ritchie, A. Frandsen, A. Schousboe, M. Scott, P. Rumsby, C. Meredith, Association of c-fos mRNA expression and excitotoxicity in primary cultures of mouse neocortical and cerebellar neurons, J. Neurosci. Res. 48 (1997) 533 – 542. [34] R. Griffiths, L. Ritchie, K. Lidwell, A. Grieve, C.S. Malcolm, M. Scott, C. Meredith, Calcium influx via L-type voltage-gated channels mediates the delayed, elevated increases in steady-state c-fos mRNA levels in cerebellar granule cells exposed to excitotoxic levels of glutamate, J. Neurosci. Res. 52 (1998) 641 – 652. [35] B.K. Ha, G.A. Bishop, J.S. King, R.W. Burry, Corticotropin releasing factor induces proliferation of cerebellar astrocytes, J. Neurosci. Res. 62 (2000) 789 – 798. [36] K. Hashimoto, R. Ichikawa, H. Takechi, Y. Inoue, A. Aiba, K. Sakimura, M. Mishina, T. Hashikawa, A. Konnerth, M. Watanabe, M. Kano, Roles of glutamate receptor delta 2 subunit (GluRdelta 2) and metabotropic glutamate receptor subtype 1 (mGluR1) in climbing fiber synapse elimination during postnatal cerebellar development, J. Neurosci. 21 (2001) 9701 – 9712. [37] D. Ichikawa, A. Tabuchi, A. Taoka, T. Tsuchiya, M. Tsuda, Attenuation of cell death mediated by membrane depolarization different from that by exogenous BDNF in cultured mouse cerebellar granule cells, Brain Res. Mol. Brain Res. 56 (1998) 218 – 226. [38] R. Ishitani, D.M. Chuang, Glyceraldehyde-3-phosphate dehydrogenase antisense oligodeoxynucleotides protect against cytosine arabinonucleoside-induced apoptosis in cultured cerebellar neurons, Proc. Natl. Acad. Sci. U. S. A. 93 (1996) 9937 – 9941. [39] D.R. Kaplan, F.D. Miller, Neurotrophin signal transduction in the nervous system, Curr. Opin. Neurobiol. 10 (2000) 381 – 391. [40] J.S. King, G.A. Bishop, The distribution and cellular localization of CRF-R1 in the vermis of the postnatal mouse cerebellum, Exp. Neurol. 178 (2002) 175 – 185. [41] L. Larkfors, R.M. Lindsay, R.F. Alderson, Characterization of the responses of Purkinje cells to neurotrophin treatment, J. Neurochem. 66 (1996) 1362 – 1373. [42] F. Lezoualc’h, S. Engert, B. Berning, C. Behl, Corticotropin-releasing hormone-mediated neuroprotection against oxidative stress is associ-

[43] [44]

[45]

[46]

[47]

[48]

[49]

[50]

[51]

[52]

[53]

[54]

[55] [56]

[57]

[58]

[59]

[60]

[61]

[62]

127

ated with the increased release of non-amyloidogenic amyloid beta precursor protein and with the suppression of nuclear factor-kappaB, Mol. Endocrinol. 14 (2000) 147 – 159. D. Lindholm, S. Hamner, U. Zirrgiebel, Neurotrophins and cerebellar development, Perspect. Dev. Neurobiol. 5 (1997) 83 – 94. L.M. Luttrell, Y. Daaka, R.J. Lefkowitz, Regulation of tyrosine kinase cascades by G-protein-coupled receptors, Curr. Opin. Cell Biol. 11 (1999) 177 – 183. P. Madtes Jr., J.S. King, The temporal and spatial development of CRF binding sites in the postnatal mouse cerebellum, Neurosci. Res. 34 (1999) 45 – 50. A. Maklad, B. Fritzsch, Partial segregation of posterior crista and saccular fibers to the nodulus and uvula of the cerebellum in mice, and its development, Brain Res. Dev. Brain Res. 140 (2003) 223 – 236. S.M. Maricich, K. Herrup, Pax-2 expression defines a subset of GABAergic interneurons and their precursors in the developing murine cerebellum, J. Neurobiol. 41 (1999) 281 – 294. D.P. Martin, T.L. Wallace, E.M. Johnson Jr., Cytosine arabinoside kills postmitotic neurons in a fashion resembling trophic factor deprivation: evidence that a deoxycytidine-dependent process may be required for nerve growth factor signal transduction, J. Neurosci. 10 (1990) 184 – 193. K. Mertz, T. Koscheck, K. Schilling, Brain-derived neurotrophic factor modulates dendritic morphology of cerebellar basket and stellate cells: an in vitro study, Neuroscience 97 (2000) 303 – 310. M.E. Morrison, C.A. Mason, Granule neuron regulation of Purkinje cell development: striking a balance between neurotrophin and glutamate signaling, J. Neurosci. 18 (1998) 3563 – 3573. T.L. Overbeck, J.S. King, Developmental expression of corticotropinreleasing factor in the postnatal murine cerebellum, Brain Res. Dev. Brain Res. 115 (1999) 145 – 159. D.S. Park, E.J. Morris, L. Stefanis, C.M. Troy, M.L. Shelanski, H.M. Geller, L.A. Greene, Multiple pathways of neuronal death induced by DNA-damaging agents. NGF deprivation, and oxidative stress, J. Neurosci. 18 (1998) 830 – 840. W.A. Pedersen, D. McCullers, C. Culmsee, N.J. Haughey, J.P. Herman, M.P. Mattson, Corticotropin-releasing hormone protects neurons against insults relevant to the pathogenesis of Alzheimer’s disease, Neurobiol. Dis. 8 (2001) 492 – 503. W.A. Pedersen, R. Wan, P. Zhang, M.P. Mattson, Urocortin, but not urocortin II, protects cultured hippocampal neurons from oxidative and excitotoxic cell death via corticotropin-releasing hormone receptor type I, J. Neurosci. 22 (2002) 404 – 412. M.H. Perrin, W.W. Vale, Corticotropin releasing factor receptors and their ligand family, Ann. N.Y. Acad. Sci. 885 (1999) 312 – 328. M. Radulovic, C. Hippel, J. Spiess, Corticotropin-releasing factor (CRF) rapidly suppresses apoptosis by acting upstream of the activation of caspases, J. Neurochem. 84 (2003) 1074 – 1085. K. Schilling, M.H. Dickinson, J.A. Connor, J.I. Morgan, Electrical activity in cerebellar cultures determines Purkinje cell dendritic growth patterns, Neuron 7 (1991) 891 – 902. A. Shimada, C.A. Mason, M.E. Morrison, TrkB signaling modulates spine density and morphology independent of dendrite structure in cultured neonatal Purkinje cells, J. Neurosci. 18 (1998) 8559 – 8570. G.N. Smagin, A.J. Dunn, The role of CRF receptor subtypes in stress-induced behavioural responses, Eur. J. Pharmacol. 405 (2000) 199 – 206. A. Srinivasan, K.A. Roth, R.O. Sayers, K.S. Shindler, A.M. Wong, L.C. Fritz, K.J. Tomaselli, In situ immunodetection of activated caspase-3 in apoptotic neurons in the developing nervous system, Cell Death Differ. 5 (1998) 1004 – 1016. A.M. Szekely, M.L. Barbaccia, H. Alho, E. Costa, In primary cultures of cerebellar granule cells the activation of N- methyl-D-aspartatesensitive glutamate receptors induces c-fos mRNA expression, Mol. Pharmacol. 35 (1989) 401 – 408. I. Torres-Aleman, M. Villalba, M.P. Nieto-Bona, Insulin-like growth factor-I modulation of cerebellar cell populations is developmentally

128

P. Madtes Jr. et al. / Developmental Brain Research 151 (2004) 119–128

stage-dependent and mediated by specific intracellular pathways, Neuroscience 83 (1998) 321 – 334. [63] K.K. Vaux, N.C. Kahole, K.L. Jones, Cyclophosphamide, methotrexate, and cytarabine embropathy: is apoptosis the common pathway? Birth Defects Res., Part A Clin. Mol. Teratol. 67 (2003) 403 – 408. [64] C.C. Verstappen, J.J. Heimans, K. Hoekman, T.J. Postma, Neurotoxic complications of chemotherapy in patients with cancer: clinical signs and optimal management, Drugs 63 (2003) 1549 – 1563. [65] A. Virchis, M. Koh, P. Rankin, A. Mehta, M. Potter, A.V. Hoffbrand, H.G. Prentice, Fludarabine, cytosine arabinoside, granulocyte-colony stimulating factor with or without idarubicin in the treatment of high risk acute leukaemia or myelodysplastic syndromes, Br. J. Haematol. 124 (2004) 26 – 32.

[66] H. Vogel, D.S. Horoupian, Filamentous degeneration of neurons. A possible feature of cytosine arabinoside neurotoxicity, Cancer 71 (1993) 1303 – 1308. [67] P. Ye, Y. Xing, Z. Dai, A.J. D’Ercole, In vivo actions of insulin-like growth factor-I (IGF-I) on cerebellum development in transgenic mice: evidence that IGF-I increases proliferation of granule cell progenitors, Brain Res. Dev. Brain Res. 95 (1996) 44 – 54. [68] A.U. Zaidi, C. D’Sa-Eipper, J. Brenner, K. Kuida, T.S. Zheng, R.A. Flavell, P. Rakic, K.A. Roth, Bcl-X(L)-caspase-9 interactions in the developing nervous system: evidence for multiple death pathways, J. Neurosci. 21 (2001) 169 – 175.