Nitric Oxide Protects Neuroblastoma Cells from Apoptosis Induced by Serum Deprivation through cAMP-response Element-binding Protein (CREB) Activation*

The transcription factor cAMP-response element-binding protein (CREB) mediates survival in many cells, including neurons. Recently, death of cerebellar granule neurons due to nitric oxide (NO) deprivation was shown to be accompanied by down-regulation of CREB activity (1). We now provide evidence that overproduction of endogenous NO or supplementation with exogenous NO renders SK-N-BE human neuroblastoma cells more resistant to apoptosis induced by serum deprivation. Parental cells underwent apoptosis after 24 h of serum deprivation, an outcome largely absent in clones overexpressing human neuronal nitric oxide synthase (nNOS). This protective effect was reversed by the inhibition of NOS itself or soluble guanylyl cyclase, pointing at cGMP as an intermediate effector of NO-mediated rescue. A slow-releasing NO donor protected parental cells to a significant extent, thus confirming the survival effect of NO. The impaired viability of serum-deprived parental cells was accompanied by a strong decrease of CREB phosphorylation and transcriptional activity, effects significantly attenuated in nNOS-overexpressing clones. To confirm the role of CREB in survival, the ectopic expression of CREB and/or protein kinase A largely counteracted serum deprivation-induced cell death of SK-N-BE cells, whereas transfection with a CREB negative mutant was ineffective. These experiments indicate that CREB activity is an important step for NO-mediated survival in neuronal cells.

immune regulation, and neurotransmission (2)(3)(4)(5). Intracellular NO is produced by a family of nitric oxide synthases (NOS), which includes two constitutive and Ca 2ϩ -dependent forms, neuronal NOS (nNOS) and endothelial NOS (eNOS), and a Ca 2ϩ -independent, inducible form (iNOS) (2,6). In the brain, NO behaves as an intercellular and intracellular diffusible messenger involved in multiple functions from developmental and adult neural plasticity to the control of neurotransmitter release and memory consolidation (7)(8)(9)(10). Although excessive production of NO results in neurodegeneration (11), its regulated activity can be neuroprotective (12)(13)(14). For example, a survival-promoting role of NO has been documented for primary neuronal cultures and neuronal cell lines in which the inhibition of NOS counteracts growth factor-mediated survival, whereas the inclusion of NO donors promotes it (15)(16)(17). Recently, we have also demonstrated that the sustained inhibition of NO production triggers apoptosis in differentiated cerebellar granule neuron cultures (1 h). Interestingly this is accompanied by down-regulation of important survival factors such as Akt/PKB kinase (14) and CREB transcription factor (1). This action occurred through the cGMP/PKG system, the main known target of NO in neuronal cells (8). Increasing evidence points to a key role for CREB activation in the survival and differentiation of neuronal cells (18,19,20). The demonstration that such an activation may be elicited also by NO, in addition to classical neurotrophins, is critical to the understanding of NO function in the brain.
To address this point, we have utilized the human neuroblastoma cell line SK-N-BE, which represents a good model for studying the effect of NO activity in neuronal survival and differentiation. Previous observations have shown that spontaneous apoptosis of SK-N-BE cells overexpressing the p75 neurotrophin receptor could be rescued by either high endogenous NOS activity or the presence of NO donors in the culture medium (21).
In this study we show that (i) the overexpression of the nNOS isoform protects neuroblastoma cells from death induced by trophic factor withdrawal; (ii) this protective action occurs in parallel with enhanced CREB activity; and (iii) the overexpression of activated CREB mimics the survival-promoting effect of high NO activity. Overall, our results support the notion that NO acts as a survival-promoting agent in the neural cell through CREB activations. * The present work was supported by grants from the National Research Council in the framework of the targeted project "Biotechnologies" and the special project "Molecular and Genetic Approaches to the Study of Pathologies" of the University of Bologna and European Union Project Grant QLRT-1999-00573 (to G. D. V.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
ʈ To whom correspondence should be addressed: Dept. of Biology, University of Bologna, Via Selmi 3, 40126 Bologna, Italy. Tel.: 39-051-2094134; Fax: 39-051-251208; E-mail: acontest@alma.unibo.it. 1 The abbreviations used are: NO, nitric oxide; NOS, NO synthase; nNOS, neuronal NOS; eNOS, endothelial NOS; iNOS, Ca 2ϩ -independent, inducible NOS; PK, protein kinase (type A, B, or G); caPKA, constitutively active PKA; CRE, cAMP-response element; CREB, CREbinding protein; P-CREB, Ser 133 -phosphorylated CREB; mCREB, dominant negative CREB; ERK, extracellular signal-regulated kinase; MAPK, mitogen-activated protein kinase; MES, 4-morpholineethanesulfonic acid; PBS, phosphate-buffered saline; TUNEL, terminal deoxynucleotidyltransferase-mediated dUTP nick end-labeling; ELISA, enzyme-linked immunosorbent assay; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; ANOVA, analysis of variance; L-NAME, N G -nitro-L-arginine methyl ester; ODQ, 1H- [ This article has been withdrawn by the authors upon request by the Journal. In December 2019, the Journal raised questions regarding three figures of this article. The authors were not able to retrieve the original data from experiments conducted about 2 decades earlier. The analysis performed by the Journal indicated that some cells were removed from panel 4 of Fig.  2A. The authors do not agree with the Journal's analysis, which they state overinterpreted faint light spots as cells. In Fig. 4A, analysis performed by the Journal indicated the following. Lanes 1 and 2 of the actin immunoblot (with serum) were duplicated in lanes 4 and 5. Lane 3 of the P-MAPK (with serum) immunoblot was duplicated in lane 5 of the P-MAPK (with serum) and in lane 2 of the P-MAPK (without serum) immunoblots. Lane 4 of the P-MAPK (with serum) immunoblot was reused in lanes 1 and 5 of the P-MAPK (without serum) immunoblot. Lanes 3 and 4 of the P-MAPK (without serum) are the same. Lanes 1 and 5 of the P-Akt (without serum) are the same. Lanes 2 and 4 of the P-Akt (without serum) immunoblot are the same. The authors do not agree with the Journal's analysis and state that simple magnification shows differences that exclude duplication and reuse of the same bands. In Fig.  5B, the analysis by the Journal indicated that the P-CREB immunoblot was inappropriately manipulated to remove features and/or bands. The authors state that some background dirt may have been removed from a film produced with techniques dated by 20 years. In the authors' opinion, the questions raised by the Journal do not undermine the scientific value of the article, which has been cited by many articles, some of them using similar models of neuroblastoma cells and resulting in conceptually concordant findings

EXPERIMENTAL PROCEDURES
Cell Culture-The human neuroblastoma SK-N-BE cells (22) were seeded at a density of 10 5 cells/cm 2 in plastic culture plates and grown to confluency in RPMI 1640 medium (Invitrogen) containing 10% heatinactivated fetal calf serum (Invitrogen), 2 mM glutamine, 100 units/ml penicillin, and 50 g/ml streptomycin (Sigma) at 37°C in a 5% CO 2containing humidified atmosphere. Upon confluency, the cells were dispersed with trypsin, split, and subcultured. Cultures were routinely observed under phase-contrast inverted microscope and under fluorescence microscope after vital staining with fluorescein diacetate.
Treatments-Cell were shifted to serum-free medium 24 h after seeding at the density of 10 5 cells/cm 2 . The various pharmacological agents tested (L-NAME, DETA NONOate, ODQ, forskolin) were added in a volume of 1 l/ml and left in the medium for 12-24 h.
Plasmids and Transfection Experiments-The plasmid pDNA3-nNOS was kindly supplied by Dr. E. Clementi (San Raffaele Scientific Institute, Milano, Italy). Rsv containing cDNA for mCREB, CREB, PKA, and the lacZ reporter gene (rsv-lacZ) or, in some transfection experiments, the green fluorescent protein (GFP) were used. The previously described (1) plasmid pCRE-Luc was also used. Stable and transient transfection experiments were performed using polethylenimine (PEI 25K) as a DNA carrier according to the previously described method (23). The efficiency of transfection obtained with our protocol was between 5 and 10%. Briefly, semi-confluent cultures of SK-N-BE cells were stably transfected with 5-10 g of plasmid DNA. Selection was carried out in a medium supplemented with 600 g/ml G418, and colonies were isolated 4 weeks after transfection. Stably transfected single clones were isolated, propagated in complete RPMI 1640/G418 medium, and then examined for the expression of human nNOS.
Luciferase Measurement-After serum withdrawal, when the plasmids with a luciferase reporter were used, cells were washed twice with ice-cold phosphate-buffered saline and lysed by incubation in 50 mM Tris-MES (pH 7.8), 1 mM dithiothreitol, and 1% Triton X-100 for 5 min on ice. The lysate was cleared of cellular debris by centrifugation (24). Luciferase assays were performed with a TD-20/20 luminometer (Promega, Madison WI).
Determination of Nitrite Concentration-Nitrite, a stable product of NO oxidation, was determined by measuring nitrite accumulation in culture medium through a Griess reaction as described previously (25). In brief, culture supernatants (600 l) were collected and mixed with 60 l of sulfanilamide (dissolved in 1.2 M HCl) and 60 l of N-naphthylethylenediamine dihydrochloride. After 5 min at room temperature, samples were measured at the wavelength of 560 nm. Nitrite concentrations were calculated against a NaNO 2 standard.
Immunofluorescence Cytochemistry-Cells were grown on coverslips for 48 h after subculture. Cells were washed twice with PBS and fixed for 10 min with ice-cold 4% paraformaldehyde in 0.1 M phosphate buffer at pH 7.4. After washing with PBS, cells were permeabilized and treated with blocking solution (5% normal goat serum and 0.4% Triton X-100 in PBS) for 30 min. Cells were incubated overnight with anti-nNOS rabbit polyclonal antibody (Santa Cruz Biotechnology; dilution 1:500). Coverslips were washed four times, 5 min each, with PBS/Triton X-100 solution. Cells were incubated with fluorescein isothiocyanateconjugated goat anti-rabbit secondary antibody (Santa Cruz Biotechnology; dilution 1:500) for 1 h. Coverslips were washed twice with PBS, mounted on glass slides, observed, and photographed with a Zeiss fluorescence microscope.
Hoechst Staining-Cultures were washed twice with PBS and fixed in paraformaldehyde solution. Cells were stained with 2 g/ml Hoechst dye 33342 for 10 min at room temperature, observed, and photographed with the fluorescence microscope. Cells with condensed nuclei were quantified by averaging cell counts in four randomly selected fields per dish. Cell counts were made blindly. To evaluate the proportion of dying (presumably apoptotic) cells, the ratio of condensed to total nuclei was calculated on photographic prints as described previously (26).
Visualization of Apoptotic Cells through TUNEL-After fixation in paraformaldehyde, cells were washed with PBS and permeabilized with 0.1% Triton X-100 in 0.1% sodium citrate, washed again with PBS, and incubated for 60 min at 37°C in the dark with the in situ cell death fluorescence detection kit (Roche Molecular Biochemicals). In situ labeled nuclei were observed and photographed under the fluorescence microscope.
Apoptosis Assay-A sandwich ELISA method was used to assess apoptosis (Cell Death ELISA, Roche Molecular Biochemicals). The assay measures the enrichment of histone-associated DNA fragments in the cytoplasm of apoptotic cells. Detection of bound nucleosomes from the samples is made using a monoclonal anti-DNA antibody with a peroxidase (POD) label. Bound anti-DNA-POD is quantified using the peroxidase substrate 2,2Ј-azino-di-[3-ethylbenzthiazoline sulfonate] (ABTS), the product of which is measured by absorbance at 405 nm.
MTT Assay-Cell viability was determined by the tetrazolium dye colorimetric test (MTT test). The MTT absorbance was read using a test wavelength of 550 nm and a reference wavelength of 630 nm.
Western Blotting-For analysis of Ser 133 -phosphorylated CREB, total CREB, phospho-MAPK, Ser 437 -phosphorylated and total Akt, and ␤-actin proteins, Western blot was performed on 2 ϫ 10 6 cells per experimental point. Cell pellets were added at 4°C with a lysis buffer containing 1% deoxycholate, 1 g/ml aprotinin, 2 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, and 1 mM sodium orthovanadate for 10 min. Cell lysates were sonicated and either immediately processed by Western blot or kept frozen until assayed. Protein concentration of samples was estimated by the method of Lowry et al. (27). Equivalent (50 g) amounts of proteins per sample were subjected to electrophoresis on a 10% sodium dodecyl sulfate-acrylamide gel. The gel was then blotted onto a nitrocellulose membrane, and equal loading of protein in each lane was assessed by brief staining of the blot with 0.1% Ponceau S. Blotted membranes were blocked for 1 h in a 4% suspension of dried skimmed milk in PBS and incubated overnight at 4°C with the following: 1) a rabbit polyclonal anti-CREB serum; 2) a rabbit serum directed against the phosphorylated Ser 133 form of CREB (both from Upstate Biotechnology, Lake Placid, N.Y.; 1:1000 dilution); 3) a polyclonal anti-Ser 473 -phosphorylated Akt serum; 4) a polyclonal anti-Akt serum (both from New England Biolabs Inc. Beverly, MA dilution 1:1000); 5) a polyclonal anti-phospho MAPK serum (New England Biolabs, Beverly, MA; dilution 1:1000); or 6) a polyclonal anti-␤-actin serum (Sigma; dilution 1:700). Membranes were washed and incubated for 1 h at room temperature with peroxidase-conjugated anti-rabbit immunoglobulin G (IgG) (dilution 1:1000). Specific reactions were revealed with the ECL Western blotting detection reagent (Amersham Biosciences).
Statistics-Data were expressed as mean Ϯ S.E., and statistical significance was assessed by using one-way ANOVA followed by Bonferroni's test. Differences were considered significant starting from p Ͻ 0.01 value.

Overexpression of nNOS in Neuroblastoma Cell Line Protects from Serum Withdrawal-induced Apoptosis-
To study the survival-promoting action of NO in neuronal cells and gain insight into the cell signaling systems transducing it, we generated stable SK-N-BE neuroblastoma cell clones expressing nNOS protein. The level of protein expression in the various clones was assessed through Western blotting. Four of the ϳ30 G418resistant cell clones obtained were further analyzed. Two of them (clones 4 and 1) were characterized by high levels of protein expression (up to 4.5-fold as compared with parental cells), one (clone 9) displayed an expression level of nNOS comparable with that of the parental cells, and the last clone (clone 5) was characterized by an intermediate level of expression, (Fig. 1A). The relative nNOS expression in the various cell lines correlated well with their respective NO-generating catalytic activity, as assessed by the accumulation in the culture medium of NO 2Ϫ , a stable oxidized product of the released NO. Nitrite accumulation increased by up to 3.5-fold in the medium of nNOS-overexpressing clones during a 48 h interval after seeding (Fig. 1A). This accumulation depended on the catalytic enzyme activity being blocked by the inhibitor L-NAME (2.5 mM, data not shown). The increased NO production was a direct consequence of the overexpression of the neuronal enzyme isoform. Indeed, eNOS expression was not different from parental cells in nNOS-overexpressing clones, whereas iNOS was barely detectable in both parental and transfected cells (Fig. 1A, right  panel). Direct visualization of nNOS expression performed at the cellular level through immunocytochemistry confirmed the higher expression of the protein in an nNOS-overexpressing clone as compared with parental cells (Fig. 1B).
To investigate how nNOS overexpression could affect cell survival, cultures of the four clones, along with those of the parental cells, were shifted to a serum-free medium, and death was assessed 24 h later through different techniques. Staining with Hoechst 33258, which reveals apoptotic cells for their nuclear shrinkage and bright fluorescence, showed a substantially higher level of apoptosis in serum-deprived parental cells than in clones 1 and 4 ( Fig. 2A). Instead, clones 5 and 9 showed a density of condensed nuclei not significantly different from parental cells (Fig. 2, A and B). Cell death was measured by counting nuclei with condensed chromatin and expressing them as a percentage of total nuclei in randomly selected culture fields. Similar results were obtained when an ELISA assay to evaluate DNA fragmentation was used to assess apoptosis (Fig. 2B, right panel). As shown in Fig 2B (right panel), the level of fragmented chromatin was much higher in serumdeprived parental cells than in clones 1 and 4, whereas clones 5 and 9 did not significantly differ from parental cells. These results were eventually substantiated by a complementary approach (MTT assay) to quantitatively assess cell viability (data not shown).
To further confirm the apoptotic nature of cell death induced by serum deprivation, we applied the TUNEL technique to the serum-deprived cultures. This experiment very clearly demonstrated that the large number of apoptotic nuclei present in parental cells 24 h after serum withdrawal was much decreased in clones 1 and 4 but substantially unaffected in clones 5 and 9 (Fig. 3A). To confirm that this effect was indeed related to the increased activity of overexpressed nNOS, we tested the effect of L-NAME, a specific NOS inhibitor, on clone 4 cells following serum starvation. Fig. 3B shows that the inhibitor abolished the protective effect mediated by nNOS in these cells. Furthermore, the survival effect of nNOS overexpression appeared to be related to the ability of NO to activate guanylate cyclase as ODQ, a specific inhibitor of this enzyme, largely replicated the effect of L-NAME (Fig. 3B).
nNOS Overexpression Counteracts the Decrease of CREB Activity Consequent to Serum Withdrawal-To investigate the molecular mechanisms through which high NO activity protects neuroblastoma cells from serum deprivation-induced apoptosis, we focused on the activation of the survival-promoting transcriptional factor CREB. In a previous study (1), we had demonstrated that blocking NO production promoted apoptosis of cerebellar granule neurons concomitantly with a down-regulation of CREB activity. In the presence of serum, ectopic expression of nNOS had no effect on the constitutive levels of native CREB protein (not shown) and had minimal effect on its active, phosphorylated form (Fig. 4A, left panels). On the contrary, serum deprivation caused an almost total disappearance of the phosphorylated form of CREB in parental cells and clone 9 and a minor sparing in clone 5, whereas the effect was significantly less drastic in clones 1 and 4 (Fig. 4A, right panels). To verify that the difference between cell clones was not due to a generic response to serum deprivation, we monitored the phosphorylation state of other survival factors such as Akt/PKB and p42/p44 MAPK (ERK1/2) kinases. The results depicted in Fig. 4A show that serum deprivation affected the phosphorylation state of these kinases in parental cells and all clones in the same way. This suggests that the degree of CREB phosphorylation following serum starvation in our cells is intrinsically related to their specific levels of nNOS activity. Because the survival effect of nNOS overexpression was substantially hampered by the inhibition of NOS activity and primary NO targets such as guanylate cyclase, we examined whether these pharmacological treatments could affect CREB function. Indeed, the portion of CREB phosphorylation still present in serum-deprived clone 4 cells was strongly reduced by the concomitant exposure to L-NAME or ODQ (Fig. 4B). Next, we investigated whether the ability of nNOS overexpression to modulate the levels of intracellular P-CREB in serum-free conditions could affect its transcriptional activity. The regulation of the CRE-mediated transcription was examined by transfecting parental cells as well as the various clones with a CRE luciferase reporter construct. Serum deprivation for 24 h significantly decreased CRE-mediated transcription in SK-N-BE parental cells as well as in clones 9 and 5, whereas no significant alterations were elicited by serum withdrawal in clones 1 and 4 (Fig. 4C).

Nitric Oxide Donor and Forskolin Partially Rescue Parental SK-N-BE Cells from Serum Deprivation by Sustaining CREB
Activity-When parental SK-N-BE cells were shifted to a serum-free medium, 25-40% of them died during the first 24 h, whereas, as shown before, nNOS-overexpressing clones were considerably more resistant. To verify whether exogenous NO could mimic the protective effect caused by an increased production of endogenous NO, serum-deprived parental cells were treated with a slow-releasing (t1 ⁄2 of release 20 h) NO donor, DETA NONOate. Inclusion of the donor in the medium was protective at 5 and 10 M concentration as determined by the cell death ELISA assay (Fig. 5A). A stronger protective effect was observed when 10 M forskolin, a known inducer of CREB activity, was added to the culture medium, (Fig. 5A). Similar results were also corroborated through a MTT assay (not shown) Moreover, parallel cultures treated with the same compounds were analyzed by Western blot to determine the level of CREB phosphorylation. As shown in Fig. 5B, the dramatic decrease of CREB phosphorylation consequent to serum withdrawal was completely blocked by forskolin and significantly prevented by DETA NONOate.
CREB Activity Antagonizes Serum Deprivation-induced Death in SK-N-BE Cells-Because the NO-dependent protective effect against serum deprivation appears to be related to CREB activation in SK-N-BE cells, we further investigated whether increased CREB activity might be sufficient to protect cells from serum deprivation-induced apoptosis. To this purpose, we transfected SK-N-BE cells with a panel of vectors encoding, respectively, CREB, CREB dominant negative (mCREB), and a constitutively active PKA (caPKA).
First, all constructs were tested for their ability to reconstitute a CREB-dependent transcriptional response in SK-N-BE cells on a CRE-luciferase reporter (Fig. 6A). In addition to CREB, caPKA was sufficient to induce luciferase activity, likely due to the increased rate of phosphorylation of endogenous CREB (Fig. 6A). Cotransfection of CREB and caPKA generated a stronger response, whereas the use of a mCREB was sufficient to completely inhibit CREB transcriptional activity (Fig. 6A). Forskolin-mediated activation of CREB transcriptional activity was in the same range of the CREB ϩ PKA combination (Fig. 6A), thus suggesting the attainment of a good transcription level under this latter condition.
Second, to demonstrate a direct effect of enhanced CREB transcriptional activity on neuronal survival, CREB, mCREB, and/or caPKA (28,29) were transfected in SK-N-BE cells together with a ␤-galactosidase-expressing plasmid used as a marker of transfected cells. The cells were deprived of serum for 24 h, starting 12 h after transfection, and stained for ␤-galactosidase activity. The level of cell death was determined by scoring blue cells in cultures with or without serum and calculating the ratio between the two conditions. Overexpression of CREB in serum-deprived cells had a small but significant effect on survival (Fig. 6B). A similar effect was also observed in cells transfected with caPKA, thus suggesting that the constitutive activation of endogenous CREB is sufficient to promote substantial survival of neuroblastoma cells. In line with these results, cell survival was further increased by co-transfection of CREB and caPKA, whereas it was prevented by mCREB (Fig.  6B). However, we could not exclude the possibility that serum deprivation may affect the transcription efficiency of the rsv promoter that controls the expression of the exogenous proteins. To address this point, we transfected cells with an rsv-LUC construct and tested the luciferase activity in the presence or absence of serum not observing any significant difference in luciferase activity between the two conditions (data not shown). This reinforces our view that CREB is a genuine determinant of survival in SK-N-BE cells.
Finally we noticed that CREB overexpression, in addition to survival, might induce in SK-N-BE cells morphological alter-ation typical of neuronal differentiation (30). To investigate this point, SK-N-BE cells were transfected with rsv vectors expressing CREB, mCREB, and caPKA together with a green fluorescent protein reporter plasmid (pGFP). As shown in Fig.  6C, overexpression of CREB or a combination of CREB and PKA appeared to affect cell morphology. In fact, transfected cells, grown in the absence of serum, displayed neurites of some extension and were provided with growth cones (Fig. 6C, right upper and left lower panels) as compared with the normal morphology of parental cells grown in serum-containing medium (Fig. 6C, left upper panel). By contrast, the overexpression of mCREB induced a round, poorly differentiated phenotype in the presence of serum (Fig. 6C, right lower panel). DISCUSSION In this study we have investigated the molecular mechanisms through which NO protects neuronal cells from apopto- sis. We provide evidence that NO protects SK-N-BE neuroblastoma cells from serum deprivation-induced apoptosis by activating the transcription factor CREB. Our experimental data further point to guanylate cyclase, a widespread target for NO in many cell types, as an intermediate step of the pathway linking NO to CREB activation.
The role of NO in the regulation of apoptosis has been elusive until recently due to the fact that it may result in either induction (31) or prevention (32,14,1) of apoptosis in different cells. High levels of NO have been shown to be cytotoxic through the inhibition of ATP synthetic enzymes (33) or through DNA damage (34). However, NO can also prevent apoptosis in several cell types, including hepatocytes (35), human B lymphocytes (36), PC12 cells (32), splenocytes (37), eosinophils (38), ovarian follicles (39), and neuronal cells (14). Several mechanisms for the antiapoptotic effect of NO have been identified, including the up-regulation of protective proteins such as heat shock protein 70 (40), heme oxygenase (41), or Bcl-X L (42). The NO-mediated increase of cGMP accounts for the protective effects in some cells (38,39,43). In particular, the antiapoptotic action of NO is mediated, at least in part, through cGMP in PC12 cells (32), and we have recently shown that cGMP and its kinase (PKG) are intermediate effectors of NO survival effect in cerebellar granule neurons (1). Using an array of different methods, we show here that the overexpression of nNOS protects neuroblastoma cells from serum withdrawal-induced apoptosis. Although the extent of cell death and the protection afforded by increased NO production appeared somewhat different as evaluated through the various methods adopted, the overall results clearly pointed toward the same conclusion, i.e. increased endogenous NO production or supplementation with exogenous NO conferred significant protection to serum-deprived neuroblastoma cells. The NO donor was less effective than ectopic nNOS overexpression in protecting from serum withdrawal and, on the other hand, L-NAME was not able to completely reverse the survival effect of nNOS overexpression. This may suggest that ectopic nNOS was acting in some other way in addition to providing more NO to cultured cells. More likely, this was due to the intrinsic limitation of the available pharmacological tools. In fact, the concentration of an NO donor could not be increased over 10 M in our system to avoid toxicity due to excessive NO bursts, and the standard L-NAME concentration cannot be enough to completely block extra nNOS in the presence of high arginine concentration in the medium. The degree of protection offered by increased nNOS endogenous activity appeared not to be strictly proportional to the degree of protein overexpression. Indeed, an intermediate-expressing clone such as clone 5 was not efficient in protecting against cell death. Therefore, it appeared that a sort of threshold of nNOS expression and activity had to be reached in order to grant significant protection. Although this might imply that endogenous levels of NO production provide no protection in our system, it cannot be taken to deny the physiological relevance of a neuroprotective role for NO. Rather, considering the many physiological ways that rapidly and substantially increase NO-production, the results obtained in our system may reliably reflect a physiological role for NO in vivo.
Programmed cell death, or apoptosis, can be induced in response to various cytotoxic stimuli, including Fas, tumor necrosis factor-␣ (TNF␣), and serum or growth factor withdrawal (44,45). These stimuli activate a series of tightly controlled intracellular signals that, in many cell types, induce the activation or inhibition of transcription factors. Our results point at the transcription factor CREB as being involved in the protective effect of NO toward apoptosis of SK-N-BE cells con-sequent to serum deprivation. Here, indeed, we provide evidence that nNOS-overexpressing cell clones are protected from serum deprivation-induced apoptosis through a positive regulation of intracellular CREB transcriptional activity. We performed Western blot analysis to examine the profile of the phosphorylation state of CREB in parental and nNOS cell clones. The P-CREB levels were not much different, under normal serum conditions, between nNOS-overexpressing clones and parental cells, probably due to the high basal level of CREB activity induced by growth factors present in the serum. In contrast, P-CREB levels were almost completely abated after 24 h of serum withdrawal in parental cells, although they were substantially preserved in nNOS-overexpressing clones. A similar conclusion was also reached by studying the transcriptional activity of CREB in parental cells and nNOS-overexpressing clones under conditions of serum deprivation. In clone 5, the relatively high level of nNOS expression was accompanied by the maintenance of some residual CREB activity upon serum withdrawal. However, CREB transcriptional activity, which is required to promote survival, was not significantly different from that of parental cells under the same conditions. CREB activation through phosphorylation has been implicated in neuronal survival (18 -20), and it has recently been demonstrated that disruption of CREB function in the brain leads to neurodegeneration (46). Moreover, CREB is required for survival of, and axon extension by, sensory and sympathetic neurons, because in mice harboring a null mutation of the CREB gene these neurons exhibit excess apoptosis and degeneration and display impaired axonal growth and altered connections (47). Although growing evidence supports the link between the activation of CREB and neuronal survival, the identity of upstream initiators and intermediate signaling pathways are not univocal. In neural cells, CREB phosphorylation can be induced by a multiplicity of extracellular signals, including synaptic glutamate, growth factors, and membrane depolarization (19,20,48,49), and is mediated through several kinases including PKA, PKC, PKG, Akt/PKB, CaMK, MAPK, and the pp90 ribosomal S6 kinase family (Rsks) (50,51,19,52,53). Concerning the present results, the role of NO as survival promoter through CREB is concurrently demonstrated by the increased survival of nNOS-overexpressing clones as well as by its inhibition through L-NAME and the partial replication of the survival effect obtained by supplementing parental cells with exogenous NO through a slow releasing NO donor.
To dissect the signaling pathways involved in nNOS-mediated protection through CREB activation, we looked at the effect of NO on cell viability in the presence of the specific inhibitor of NO-dependent soluble guanylate cyclase, ODQ. The choice to investigate the cGMP-linked signaling pathway was suggested by the fact that activation of guanylate cyclase is a widespread target of NO (8). ODQ inhibited the protective effect of nNOS overexpression and significantly decreased P-CREB level, thus suggesting that both of these NO-mediated effects occurred through cGMP in neuroblastoma cells. PKG can phosphorylate CREB directly in vitro and as a consequence of NO-cGMP activity in vivo (52). Phosphorylation of CREB by PKG activated by NO-dependent generation of cGMP contributes to the late phase of long-term potentiation in the hippocampus (54). We have recently demonstrated a PKG-dependent NO survival action exerted through CREB activation in cerebellar granule neurons (1). Because CREB phosphorylation depends on cGMP in our cellular system, this may occur through the direct involvement of PKG. In other cellular systems, NO can also activate other survival pathways targeting CREB phosphorylation such as Akt/PKB and ERK (55,14). From our present data, possible roles for Akt/PKB and MAPK appear to be marginal at best in neuroblastoma cells.
Because the previously discussed results provided circumstantial evidence linking the NO-survival action to CREB activity, it was implied that CREB overexpression could, by itself, eventually result in protection from serum withdrawal-induced cell death. It was previously demonstrated that CREB is cleaved by caspases in apoptotic death induced in neuroblastoma and PC12 cells (56). Furthermore, CREB overexpression promotes the survival of PC12 cells (20). However, direct evidence of survival promoted by CREB overexpression is, to date, lacking in neuroblastoma cell lines. We demonstrate here that transfection of parental SK-N-BE cells with an expression vector containing a CREB and/or a constitutively active PKA gene increased the resistance of these cells toward serum withdrawal-induced apoptosis. In these experiments, we have used a constitutively active PKA as the most direct and potent way to phosphorylate endogenous CREB and have found it to have a significant effect by itself. The increased effect displayed by experiments of co-transfection of CREB together with PKA was most likely due to the fact that, under these conditions, both native and transfected CREB could be used as a PKA substrate. Moreover, a CREB dominant negative mutant protein decreased cell survival by effectively competing with wild-type CREB activity and thereby blocking the physiological response. The fact that direct activation of CREB through forskolin or PKA may turn out to be stronger than that obtained by activation through NO, is not surprising. These activators are by far the most potent compounds able to activate CREB at very high rates (57). In line with this, transfection with wild-type CREB also leads to lower activation and survival rates, probably because the excess substrate can only be partly used to increase the total amount of the activated form.
Overall our results point at CREB transcriptional activity as an essential step for NO-mediated survival in neuroblastoma cells. The challenge for future researches will be to study genes that are activated by CREB in neuroprotection.