Signaling mechanisms of interferon gamma induced apoptosis in chromaffin cells: involvement of nNOS, iNOS, and NFκB


Address correspondence and reprint requests to M. J. Oset-Gasque, Department of Biochemistry and Molecular Biology, Faculty of Pharmacy, Complutense University of Madrid, E-28040 Madrid, Spain.


Previous work of our group stated that exogenously added and endogenous nitric oxide (NO) generated by cytokines induce apoptosis in chromaffin cells. In this work, we investigate the specific regulation of the NO synthase (NOS) isoforms, inducible NOS (iNOS) and neuronal NOS (nNOS), and their particular participation in cell death induced by interferon gamma (IFNγ). Lipopolysaccharide (LPS) and IFNγ increase iNOS expression, with no effect on nNOS expression. On the other hand, dexamethasone increases basal nNOS expression but decreases LPS + IFNγ-induced iNOS expression. IFNγ-induced cell death was abolished by W-1400, a specific iNOS inhibitor, but only partially by nNOS inhibitors [N-ω-propyl-l-arginine (N-PLA), 3-Bromo-7-nitroindazol (7-NI), l-methyl thiocitrulline and N-methyl l-arginine], indicating the main iNOS participation in chromaffin cell death. IFNγ and LPS induce nuclear factor κB (NFκB) translocation to the nucleus, a process implicated in activation of iNOS expression, as inhibition of NFκB translocation, by SN50, decreased iNOS expression. In addition, IFNγ and LPS induce 847Ser-nNOS phosphorylation, inhibiting nNOS activity. Both processes, nNOS phosphorylation and iNOS expression induced by LPS + IFNγ, are regulated by Janus Kinase/Signal Transducer and Activator of Transcription (JAK/STAT) pathway, as IFNγ increases 727STAT-3 phosphorylation and specific inhibitors of JAK/STAT pathway, such as AG490, inhibited both processes. Taken together, these results support the hypothesis of an inactivating phosphorylation of nNOS by IFNγ, via JAK/STAT, in bovine chromaffin cells. Low NO concentrations achieved by this event, would activate NFκB translocation, increasing iNOS expression and generating, this last, high apoptotic NO concentrations.

Abbreviations used



3′-5′-cyclic guanosine monophosphate


endothelial nitric oxide synthase


interferon gamma


inducible nitric oxide synthase


nuclear factor κB inhibitor


Janus Kinases


N-methyl l-arginine




Mitogen Activated Protein Kinase Kinase


nuclear factor κB


neuronal nitric oxide synthase


nitric oxide synthase




cAMP-dependent protein kinase


Signal Transducers and Activators of Transcription


Tumor necrosis factor α


N-(3-aminomethyl)benzyl) acetamidine

The free radical nitric oxide (NO) is a cellular messenger playing very important roles in physiological processes including regulation of vascular tone, neuronal transmission, and modulation of immunological and inflammatory reactions as well as cellular growth, survival, apoptosis, proliferation, and differentiation (see Madhusoodanan and Murad 2007). In the nervous system, NO has a very important function as a regulator of neurotransmission, synapse formation, synaptic plasticity, and brain development and also in pathological events of neuronal cell death underlying neurotoxicity and neurodegeneration [see Calabrese et al. (2007) for a review].

Endogenous NO is synthesized from l-arginine by three isoforms of NO synthase (NOS). Two isoforms are expressed constitutively, the neuronal (nNOS or type 1) and the endothelial (eNOS or type 3) and one inducible under pathological conditions and inflammation (iNOS or type 2), producing large amounts of NO for up to long periods (Bredt 1999). All three isoforms are found in the CNS. The induction of a high output system for NO in response to cytokines (Liu et al. 2002) or a massive production of NO following accumulation of glutamate (Nakamura et al. 2007) can result in cell death and in pathological events underlying neurotoxicity and neurodegeneration. In fact, inhibition of nNOS and iNOS activity ameliorates the progression of disease pathology in animal models of different neurodegenerative diseases [see Calabrese et al. (2007) for a review].

Apoptosis is recognized as a normal feature in the development of the nervous system and may also play a role in neurodegenerative diseases, excitotoxicity and aging (Sastry and Subba 2000). NO plays a dual role in apoptosis, being both pro and anti-apoptotic (Boyd and Cadenas 2002; Calabrese et al. 2007). As a pro-apoptotic factor, it has been proved that NO generates peroxynitrite, activates death receptors, inhibits mitochondrial ATP synthesis, and inactivates antioxidant enzymes; on the other hand, NO can act as an anti-apoptotic signal, stimulating 3′-5′-cyclic guanosine monophosphate (cGMP) production, or protein S-nitrosylation (Kim et al. 2001). These roles much depend on the concentration of NO: nNOS produces small amounts of NO that may become deleterious when stimulated for long periods of time, whereas, once activated, iNOS produce high amounts of NO for long periods of time, actually as long as the enzyme remains activated. This high output of NO could get deleterious to the surrounding cells.

The pathways involving NO reactivity are countless. Among them, major effects are due to nuclear factor κB (NFκB) and STAT3. These proteins have been proven to mediate the effects of cytokines on inflammation, translocating to the nucleus where they activate the transcription of the iNOS, and therefore controlling the amount of this enzyme (Kleinert et al. 2004).

In bovine chromaffin cells, the presence of a constitutively expressed nNOS has been demonstrated (Vicente et al. 2002). In these cells, the l-arginine/NO/cGMP pathway has an important inhibitory role in both basal and acetylcholine-stimulated catecholamine secretion (Oset-Gasque et al. 1994; Vicente et al. 2002). Moreover, previous results from our group indicate that the exposure of these cells to high concentrations of NO donors, peroxynitrite, or cytokines for long periods induced apoptosis in chromaffin cells (Vicente et al. 2006; Pérez-Rodríguez et al. 2007). However, the intracellular origin of endogenous NO generated by cytokine-treatment of chromaffin cells is unknown, as different NOS isoforms could be implicated. Thus, in this work we try to asses the role of intracellular NO in chromaffin cell apoptotic death by studying (i) the presence and specific regulation of different NOS isoforms in chromaffin cells and (ii) their specific participation in interferon gamma (IFNγ)-induced apoptosis on these cells. Given the importance of transcriptional factor NFκB as regulator of both cell death/survival process and transcriptional activation of iNOS, we also study the role of this transcription factor on both activation of different NOS isoforms and NO-induced apoptosis. Finally, we assess the possible involvement of the JAK/STAT pathway on these effects.

Materials and methods

Chromaffin cell culture and drug treatments

Chromaffin cells were isolated from bovine adrenal glands and cultured as described by Vicente et al. (2006) with some modifications. Briefly, glands were incubated with Ca2+-free Locke containing 0.1% protease (Sigma, St Louis, MO, USA) for 20 min (2 × 10 min). Medulla was detached from cortex, submitted to mechanical disgregation, and further incubated with 0.1% collagenase (Worthington) in Ca2+-free Locke for 25 min with shaking. After digestion, solution was filtered through a 190 μm pore nylon membrane, and chromaffin cells purified through a series of 10 min centrifugations (1 × 180 g, 6 × 50 g) using a 4% albumin gradient in the last one. Cell viability and purity checking and cell plating and treatments were performed as described (Vicente et al. 2006).

Measurement of nitrite production

Nitrites were determined with the spectrofluorimetric method of Misko et al. (1993), with minor modifications as described by Vicente et al. (2006). This method is based on the measurement of the fluorescent product 1-(H)-naphthotriazole formed by the reaction of nitrites with 2,3-diaminonaphtalene in acidic conditions. Samples were calibrated with a standard curve of freshly prepared nitrites and results were expressed as arbitrary fluorescence units.

Measurement of NOS activity (citrulline assay)

Cultured bovine chromaffin cells (106/condition) were incubated in the absence or presence of 200 ng/mL IFNγ and/or in the absence or presence of different NOS inhibitors, for 24 h either in a Dulbecco’s modified Eagle’s medium without phenol red containing 2.5 mM Cl2Ca (total activity) or in a medium without calcium containing 4 mM EGTA + 10 μM calmidazolium (1-[bis(4-chlorophenyl)methyl]-3-[2-(2;4-dichlorophenyl)-2-{2;4-dichlorobenzyloxy}ethyl]-1H-imidazolium chloride) (calcium-independent activity). Then, supernatants were taken out for nitrite measurements, cells were lysed with 400 μL of pure water and NOS activity was measured by quantification the production of [U-14C]l-citrulline from [U-14C]l-arginine as described (Vicente et al. 2002). Calcium-dependent NOS activity was calculated as the difference between the amount of synthesized [14C]l-citrulline in the presence of physiological extracellular calcium and that formed in the absence of extracellular calcium.

Measurement of cell viability by the XTT test

Changes in cell viability induced by IFNγ, in the presence and absence of NOS inhibitors, was measured by the Cell Proliferation Kit II (Hoffmann-La Roche Ltd Diagnostics, Basel, Switzerland) (XTT assay) as described (Figueroa et al. 2005). Results were expressed as ratios over the respective controls (not-treated cells).

Flow cytometric analysis of apoptosis

Measurement of apoptosis induced by IFNγ, in the presence and absence of NOS inhibitors, was carried out by Flow cytometry as described (Figueroa et al. 2005). Results were expressed as ratios over the respective controls (not-treated cells).

Preparation of cytosolic and nuclear extracts

Cytosolic and nuclear extracts were isolated by a modified procedure based on the method of Andrews and Faller (1991) as previously described (Pérez-Rodríguez et al. 2007). Aliquots of extracts were analyzed for protein content using the Bio-Rad (Hercules, CA, USA) protein reagent.

Electrophoretic mobility shift assays (EMSA)

The oligonucleotide sequence corresponding to the NFκB site was the proximal κB motive (nucleotides −92 to −65) of the rat NOS-2 promoter (tcga 5′-CCAACTGGGGACTCTCCCTTTGGGAACA-3′ and tcga 5′-TGTTCCCAAAGGGAGAGTCCCCAGTTGG-3′). electrophoretic mobility shift assays (EMSA) assays were performed as described (Pérez-Rodríguez et al. 2007).

Western blot analysis

Western blots were carried out as described by Pérez-Rodríguez et al. (2007). Band intensities were measured on a densitometric scanner, and normalized with respect to β-actin expression.

RT-PCR analysis

RNeasy Mini Kit (Quiagen Ltd, Sussex, UK) was used for total RNA isolation. RT was carried out for 1 h at 55°C with oligodeoxythymidylate primer using 5 μg of total RNA from each sample for complementary DNA synthesis.

Semiquantitative and real-time quantitative PCR to determine the levels of rat NOS and housekeeping Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNAs was performed by using the following specific primers synthesized at Sigma-Genosys.


Real-time PCR

The SYBR Green PCR Master Mix and the 7900 HT Fast Real-Time PCR system (Applied Biosystems, Foster City, CA, USA) were used to detect the real-time quantitative PCR products of reverse-transcribed cDNA samples, according to the manufacturer’s instructions. q-PCR conditions were: 95°C (10 min) followed by 40 cycles of 15 s at 95°C and annealing for 1 min at 60°C. Three independent quantitative PCR assays were performed for each gene and measured in triplicate. Three no-template controls were run for each quantitative PCR assay, and genomic DNA contamination of total RNA was controlled using RT minus controls (samples without the reverse transcriptase).

Semiquantitative PCR

Conventional PCR amplifications were conducted in a 25 μL solution containing 1 × PCR buffer, 0.2 mM dNTP mix (Invitrogen, Carlsbad, CA, USA), 1.5 mM magnesium chloride, 400 nM of each primer and 1 U of DNA polymerase and 2 μL of cDNA template, corresponding to 5 μg total RNA in a 20 μL final volume. Negative control of amplification was performed with 2 μL of water instead of cDNA template. Amplification conditions were: 2 min at 95°C, 11 cycles of 30 s at 95°C, 30 s at 61°C, decreasing 0.5°C every cycle, and 20 s at 72°C, followed by 23 cycles of 30 s at 95°C, 30 s at 55.5°C, and 20 s at 72°C, and a final extension of 2 min at 72°C. Reactions were carried out in a thermal cycler. 10 μL of the PCR products were resuspended in 6X loading buffer [30% glycerol, 0.5 μg/mL ethidium bromide (BrEt)] and electrophoresed through 1.5% agarose in 0.5X Tris/Borate/EDTA (TBE) buffer (45 mM Tris-borate; 1 mM Na2EDTA pH 8.0) with 0.5 μg/mL ethidium bromide (BrEt) for 1.5 h.


Data were expressed as mean ± SEM values of three or four independent experiments with different cell batches, each one performed in duplicate or triplicate. Statistical comparisons were assessed by using one-way anova (Scheffe’s F-test) followed in some instance by a two-way anova test. Differences were accepted as significant as < 0.05 or less.


Chromaffin cells express nNOS and iNOS both showing different regulation by cytokines and glucocorticoids

In order to study the possible participation of different NOS isoforms in chromaffin cell death-induced by IFNγ, we first investigate the expression of specific NOS isoforms in these cells, at both protein and mRNA levels. As shown in Fig. 1 western blot experiments for chromaffin cells treated with cytokines or dexamethasone for 24 h indicate the expression of two NOS isoforms, nNOS and iNOS, nNOS being basally expressed. Both NOS isoforms showed a different regulation by cytokines and dexamethasone. Thus, while nNOS expression was not significantly regulated by cytokines, it was activated by dexamethasone in a concentration-dependent manner (Fig. 1b). iNOS was induced by cytokines, especially by IFNγ, but not by dexamethasone (Fig. 1a).

Figure 1.

 Regulation of nNOS and iNOS expression by LPS, IFNγ, TNFα, and glucocorticoids at protein level. Bovine chromaffin cells were incubated for 24 h in the absence (control) or presence of 10 μM LPS, 10 nM IFNγ, or 10 nM TNFα (Prepotech, London, UK) (a) or indicated concentrations of dexamethasone (b). NOS expression levels were measured by western blot with antibodies anti-nNOS and anti-iNOS (BD-Biosciences, San Diego, CA, USA), both with positive controls. Data are expressed as a quantification of results over control and are mean ± SEM values of three experiments. Statistic compares the effect of cytokines (a) or dexamethasone (b) with respective control values (c). (*< 0.05; **< 0.01; ***< 0.001) (one-way anova test).

In order to go deeper in the study of NOS isoform expression and its specific regulation, PCR [semiquantitative for Fig. 2(a) and quantitative for Fig. 2(b and c)] experiments were set for the nNOS, iNOS, eNOS, and G3PDH genes. Semiquantitative RT-PCR experiments confirm the presence of specific mRNAs for nNOS and iNOS (Fig. 2a). Data also showed a small, mild presence of eNOS mRNA, this probably due to a small contamination in the chromaffin cell cultures with endothelial cells, although as counted at the moment of platting, this contamination in never higher as 5%. So, our data indicate that the main isoforms expressed on primary cultures of bovine chromaffin cells are nNOS and iNOS.

Figure 2.

 mRNA expression and regulation of different NOS isoforms in bovine chromaffin cells. (a) For conventional RT-PCR bovine chromaffin cells were incubated for 6 h in the presence or absence of LPS 10 μM plus IFNγ 10 nM (L + I). mRNA was isolated and treated with appropriate set of primers for nNOS, eNOS, iNOS, and GAPDH as a control, as described in Materials and Methods. (b and c) For real time RT-PCR bovine chromaffin cells were incubated for 6 h with (b) Increasing concentrations of LPS (2–50 μg/mL; 2–50 μM) or IFNγ (40–1000 ng/mL; 2–50 nM) or (c) Increasing concentrations of dexamethasone (1 nM–10 μM) in the absence (nNOS) or presence of LPS 10 μg/mL plus IFNγ 200 ng/mL (iNOS) and mRNA was extracted and treated with appropriate sets of primers for nNOS and iNOS, and GAPDH as a control, in an AB 7900 HT Fast Real PCR from Applied Biosystems, as described in Materials and Methods. Data are expressed as ratios over their respective controls and are mean ± SEM values obtained from three experiments each one performed in triplicate. Statistic compares the effect of cytokines (b) or dexamethasone (c) with their specific controls (*< 0.05; **< 0.01; ***< 0.001) (one-way anova test).

In Fig. 2(b and c), we present data on mRNA NOS expression obtained by quantitative real-time PCR techniques after chromaffin cell incubation with increasing concentrations of lipopolysaccharide (LPS) or IFNγ (Fig. 2b) or increasing concentrations of dexamethasone, in the absence (nNOS) or presence of LPS plus IFNγ (iNOS) (Fig. 2c). These data show that both LPS and IFNγ increased mRNA iNOS expression in a dose-dependent manner (Fig. 2b) while these treatments do not significantly affect nNOS mRNA expression (data not shown). Moreover, treatment with dexamethasone increased mRNA nNOS basal expression but decreased mRNA iNOS in a concentration-dependent manner (Fig. 2c). So, these results confirm the expression of specific nNOS and iNOS isoforms in chromaffin cells with different regulation by cytokines and glucocorticoids at both protein and mRNA levels.

Chromaffin cells express a calcium-dependent basal NOS activity and a calcium-independent IFNγ-induced NOS activity, which could be specifically inhibited by different NOS inhibitors

As it is known, constitutively-expressed nNOS activity is calcium dependent and iNOS activity-induced by cytokines is mainly calcium independent. In order to confirm the presence of both, calcium-dependent and independent NOS activity in chromaffin cells, NOS activity was measured by the quantification of the production of [U-14C]l-citrulline from [U-14C]l-arginine, as described (Vicente et al. 2002), both in the presence of physiological extracellular calcium and in the presence of EGTA plus calmidazolium (1-[bis(4-chlorophenyl)methyl]-3-[2-(2;4-dichlorophenyl)-2-{2;4-dichlorobenzyloxy}ethyl]-1H-imidazolium chloride), a CaM blocker, in the absence or presence of IFNγ. As shown in Fig. S1A, in the presence of 2..5 mM extracellular calcium, total basal NOS activity was 0.63 ± 0.04 nmol [14C]citrulline/106 cells in 24 h. This activity was reduced by 70% in the absence of calcium. The stimulation of chromaffin cells with 10 nM IFNγ increases by 40% the total NOS activity, this effect being stronger (about 75%) in the absence of calcium. By contrast, calcium-dependent NOS activity was not significantly affected by IFNγ (Fig. S1A). To evaluate the specificity and relative potency of different NOS inhibitors on both Ca2+-dependent (nNOS) and Ca2+-independent (iNOS) activities, we measured NOS activity in the presence of different concentrations. For this study N-(3-aminomethyl)benzyl) acetamidine (1400W) as a specific iNOS inhibitor (Garvey et al. 1997), N-ω-propyl-l-arginine (N-PLA) a very high specific nNOS inhibitor (Zhang et al. 1997; Fedorov et al. 2003) and S-methyl-l-thiocitrulline, N-methyl l-arginine (l-NMA) and 3-Bromo-7-nitroindazol (7-NI) as less specific nNOS inhibitors (Vicente et al. 2002) were used. These compounds produced a dose-dependent inhibition of NOS activity with IC50s for different inhibitors indicated in Fig. S1B. The order of potency of inhibitors on total NOS activity was: N-PLA > W-(1400) > thiocitrulline = 7-NI > l-NMA. Similar results were obtained by measurement of nitrites in extracellular medium (data not shown). In order to demonstrate inhibition specificity on calcium-dependent (nNOS) or calcium-independent (iNOS) activity of different inhibitors, NOS activity was measured in the presence or absence of calcium for three of these inhibitors: 1400W, N-PLA and thiocitrulline. Results obtained with these inhibitors and IC50 values are shown in Fig. S1C–E. These results clearly demonstrated that, in this cell type, W-1400 was a more specific inhibitor of Ca2+-independent NOS activity (IC50 = 0.27 ± 0.09 μM) than Ca2+-dependent one (IC50 = 11.5 ± 2.9 μM), while N-PLA and thiocitrulline were better inhibitors of Ca2+-dependent activity (IC50 = 0.098 ± 0.007 and 12.3 ± 3.6 μM, for N-PLA and thiocitrulline respectively) than of the Ca2+-independent activity (178.9 ± 56.8 and 157.8 ± 53.6 μM). Therefore, W-1400 (0.1-10 μM) could be used in our system as a specific inhibitor of iNOS and N-PLA (0.1-10 μM) and thiocitrulline (1-100 μM) as specific nNOS inhibitors.

IFNγ-induced chromaffin cell death is totally reverted by iNOS inhibitors but partially by nNOS inhibitors

As results above indicate the expression of two isoforms, nNOS and iNOS, in chromaffin cells, the specific participation of nNOS and iNOS on cell viability and apoptosis-induced by cytokines was investigated by challenging the bovine chromaffin cells with IFNγ in the absence or presence of different NOS inhibitors. 10 nM IFNγ (200 ng/mL) decreased cell viability in about a 65% (Fig. 3a). This effect was reverted, in a dose-dependent manner, by NOS inhibitors. Concentrations in the range of 1–5 μM of W-1400 were able to totally revert IFNγ effects, whereas higher concentrations (up to 1000 μM) were needed in the case of nNOS-specific inhibitors.

Figure 3.

 Effect of different concentrations of NOS inhibitors on IFNγ-induced increases in NO levels and decreases in cell viability. Bovine chromaffin cells were incubated for 24 h with the indicated concentrations of NOS inhibitors (W-1400, thiocitrulline, l-NMA, 7-NI, N-PLA) in the presence of 10 nM (200 ng/mL) IFNγ. NO levels (a and c) were measured by a 2,3-diaminonaphtalene spectrofluorimetric assay and cell viability (b and d) by the XTT spectrofluorimetric assay as described in Materials and Methods. Data were expressed as ratios over control and are mean ± SEM values obtained from three experiments each one performed in quadruplicate. [NO]basal = 0.72 ± 0.07 nmol/106 cells. *p < 0.05; **p < 0.01; ***p < 0.001 (one-way ANOVA test).

The increase in NO production induced by IFNγ (1.52 ± 0.11 times the NO basal levels), was also inhibited by NOS inhibitors but in a manner not always parallel to their ability to decrease cell viability. Thus, W-1400 restored nitrite production and cell viability to their basal levels, even at very small concentrations (Fig. 3a and b). N-PLA drastically reverted NO production bellow the basal levels but had a mild effect on cell viability at small concentrations; similar effects were observed with thiocitrulline and 7-NI (Fig. 3c and d). Thus, while 1400 W, a specific iNOS inhibitor, was able to inhibit completely both cell death and increase in NO levels-induced by IFNγ at 1–10 μM concentration, nNOS inhibitors needed very high concentrations (up to 1000 μM) to get the same or lower effects. At a concentration of 10 μM, the strength of these inhibitors was as follows: W-1400 > thiocitrulline = L-PLA > 7-NI > l-NMA. These results point out that the diminution in cell viability-induced by IFNγ was mainly due to iNOS activation.

In order to know whether IFNγ-induced increase in cell death is due to apoptosis, and the implication of NOS isoforms, we tested the action of these NOS inhibitors on apoptosis induced by IFNγ. Data from Fig. 4 show that apoptosis-induced by 10 nM IFNγ (about three times the basal values) was completely inhibited by 1400W 10 μM, while N-PLA, thiocitrulline and l-NMA, even at doses 100 times higher, were not able to induce a total reversion of apoptosis. From these results, it seems fair to conclude that the main NOS isoform involved on IFNγ-induced apoptosis is iNOS, with minor effect of nNOS.

Figure 4.

 Effect of NOS inhibitors on IFNγ-induced-apoptosis in chromaffin cells. Bovine chromaffin cells (2 × 106) were incubated for 24 h with the indicated concentrations of NOS inhibitors in the presence of 10 nM IFNγ. Apoptosis was measured by flow cytometry as described in Materials and Methods. Data were expressed as ratios over basal (3.4 ± 0.40% M1) and are mean ± SEM values obtained from three experiments each one performed in duplicate. Statistic compares the effect of NOS inhibitors in the presence of IFNγ with the apoptosis induced by IFNγ alone (*< 0.05; ***< 0.001) (one-way anova test).

Cytokines activate the translocation to the nucleus of NFκB which mediates activation of iNOS expression

It is well known in the literature that cytokines activate iNOS expression by the early activation of some nuclear transcriptional factors, including NFκB (Aktan 2004). Moreover, it is also known that physiological levels of NO, similar to those produced by the basal activity of nNOS or eNOS, prevent induction of iNOS mRNA expression through the suppression of NFκB activation (Colasanti et al. 1995; Togashi et al. 1997). Therefore, in order to study in depth the possible mechanism by which each NOS isoform could contribute to NO effect on cell viability and apoptosis of chromaffin cells, we study the implication of NFκB on IFNγ-induced apoptosis and on iNOS activation. Because it is known that NFκB translocation to nucleus and activation require phosphorylation and degradation of its specific inhibitor nuclear factor κB inhibitor (IκB), which complexed NFκB into the cytosol, we performed both electrophoretic mobility shift assays (EMSA) and western blot experiments to check out NFκB translocation to the nucleus and IκB degradation on the cytosol, respectively. Results from Fig. 5(a) show that LPS, IFNγ and TNFα were able to increase IκB degradation, as observed on the western analysis on the top of figure, and also to increase NFκB translocation to the nucleus where it binds to nuclear DNA (Fig. 5a and b).

Figure 5.

 Effect of cytokines and SN50 on NFκB translocation to the nucleus and apoptosis. Bovine chromaffin cells were incubated for 1 h with different cytokines (LPS 10 μM, IFNγ 10 nM, TNFα 10 nM) (a) or with LPS + IFNγ at indicated times (b) and IκB degradation in cytosol measured by western-blot (‘a’ up) or NFκB binding to DNA from nuclear extracts isolated and electrophoresed in EMSA (‘a’ down and b) was measured as described in Materials and Methods. The effect of SN50 (Calbiochem, San Diego, CA, USA) (1 and 10 μM) on NFκB activity induced by LPS (c) or apoptosis-induced by cytokines (d) was determined. Apoptosis data were expressed as ratios over control. Effects of SN50, in (c) and (d), were statistically significant against both, basal and respective cytokines without SN50, at > 0.001 (multi-variance analysis of anova test). (a–c) show a representative experiment of three. In (d) data are mean ± SEM values of three experiments performed by duplicate.

To check out whether cytokine-induced NFκB activation was specific, we studied this in the presence and absence of the membrane permeable peptide SN50 (1 and 10 μM), which contains the nuclear localization signal (NLS) of NFκB and blocks the intracellular recognition mechanism for the nuclear localization signal (NLS) on NFκB dimmers thereby preventing their nuclear translocation (Lin et al. 1995). Results in Fig. 5(c) show that SN50 was able to inhibit NFκB activation in a dose-dependent manner. However, apoptotic effect of LPS and cytokines was enhanced and not inhibited by SN50, that indicating that NFκB could be activating survival genes in addition to death genes, but that NFκB activation is not enough to stop chromaffin cells apoptosis induced by cytokines. In order to know if NFκB activation could be mediate iNOS expression induced by IFNγ we study the SN50 effect on the increase in iNOS mRNA and protein expression induced by LPS + IFNγ. Results from Fig. 6 indicate that SN50 was able to inhibit iNOS expression at transcriptional level in a dose-dependent manner, with 30% inhibitory effect at 10 μM concentrations (Fig. 6a). The same inhibitory effect (about 50%) was shown at level of protein expression (Fig. 6b). These data are in agreement with those from enzymatic activity where 10 μM SN50 was able to reduce calcium-independent NOS activity induced by LPS + IFNγ by 50% (C = 0.53 ± 0.07 nmol [U-14C]l-citrulline/106 cells; + 10 μM SN50 = 0.36 ± 0.04 nmol/106 cells).

Figure 6.

 Effect of SN50 on iNOS expression at both mRNA (a) and protein (b and c) level. Bovine chromaffin cells were incubated for 24 h with LPS 10 μM plus IFNγ 10 nM in the absence or presence of 1–10 μM SN50 and mRNA or proteins was extracted as indicated in Materials and Methods. Data in (a) and (c) are expressed as ratios over basal control and are mean ± SEM values of two experiments each one performed in triplicate. In (b) a representative western blot is shown. Statistic compares the effect of SN50 on iNOS mRNA or iNOS protein induced by LPS + IFNγ (**< 0.01; ***< 0.001) (one-way anova test).

nNOS inhibitors shorten the time of the IFNγ-induced translocation to the nucleus of NFκB and increase by themselves NFκB activation

Because our results above suggested some involvement of nNOS in IFNγ effects on chromaffin cell death and apoptosis, in order to know the possible participation of nNOS in the IFNγ effects on NFκB nuclear translocation we studied the effect of different NOS inhibitors on NFκB translocation to the nucleus, in basal conditions and in the presence of LPS + IFNγ. Chromaffin cells were challenged, at the indicated times, with LPS plus IFNγ in the presence and absence of nNOS (thiocitrulline, N-PLA) and iNOS (W-1400) inhibitors. As seen on Fig. 7(a), LPS + IFNγ induced NFκB translocation to the nucleus (2–3 times basal levels) at long times, 1–18 h. Addition of W-1400 had not a significant effect on NFκB activation at these times. However, the presence of the specific nNOS inhibitors (N-PLA and Thiocitrulline) shortened the time of translocation of NFkB to the nucleus to 15–30 min (Fig 7a).

Figure 7.

 Effect of NOS inhibitors on LPS plus IFNγ-induced (a) or basal (b) NFκB translocation to the nucleus in bovine chromaffin cells. Bovine chromaffin cells (5 × 106 cells/condition) were challenged for indicated times with LPS 10 μM + IFNγ 10 nM, in the absence or presence of the specific NOS inhibitors W-1400 0.1 μM, thiocitrulline 10 μM or N-PLA 10 μM (a) or with indicated NOS inhibitors at 1 mM concentrations (b). Nuclear extracts were isolated and electrophoresed in an EMSA, as described in Materials and Methods. A representative experiment of three is shown.

On the other hand, in the absence of IFNγ stimuli, nNOS inhibitors basally induced NFkB activation (Fig. 7b). Moreover, these inhibitors were able to induce, by themselves, a slight increase of 30% in iNOS gene expression (data not shown).

IFNγ induces 847Ser nNOS phosphorylation and 727Ser STAT-3 phosphorylation and translocation to the nucleus

One of the more probable mechanisms used by IFNγ to inactivate nNOS generating low NO levels, stimulate NFκB activation and thus up-regulate the iNOS mRNA expression, is the phosphorylation of nNOS. In order to know if this mechanism is mediated by IFNγ in chromaffin cells, we studied nNOS phosphorylation on serine 847 at different times and with several cytokines. Western blot results shown on Fig. 8(a) indicate that LPS, IFNγ and TNFα, and the combination LPS + IFNγ were able to increase nNOS phosphorylation in 847Ser while they do not have any significant effect or even decrease nNOS expression. nNOS phosphorylation was time-dependent, maximal effect being achieved after 1 h stimulation for all the stimuli evaluated (Fig. 8b and c, left panel). These results are consistent with results obtained on NFκB activation, shown above, and with maximal effect of NOS inhibitors on nNOS activity (Vicente et al. 2002). Thus, taken all together, these data suggest that IFNγ activates iNOS expression, inducing NFκB translocation to the nucleus and these effects might be mediated by nNOS phosphorylation.

Figure 8.

 Cytokine induction of 847Ser nNOS phosphorylation and 727STAT-3 phosphorylation in bovine chromaffin cells. Bovine chromaffin cells were challenged for 1 h with LPS 10 μM, IFNγ 10 nM, TNFα 10 nM or a combination of LPS plus IFNγ (a) or with IFNγ at indicated times (b) and proteins from cytosolic P-nNOS (a and b) or cytosolic P-STAT-3 and nuclear P-STAT-3 (b) was measured by western blot techniques using specific antibodies [Abcam (Cambridge, UK) and Santa Cruz Biotechnology, Inc. (Heidelberg, Germany), respectively]. A representative experiment of three is shown. (c) Quantification of results for three different experiments. Statistic compares results obtained at different times over basal control (*< 0.05; **< 0.01; ***< 0.001) (one-way anova test).

Because one of the more important signal transduction pathways by which IFNγ induces its apoptotic and inflammatory effects, as well as NFκB activation, is the canonical JAK/STAT pathway, largely responsible for the antiviral and growth-inhibitory activities of interferons and iNOS activation (Stempelj et al. 2007), we study the possibility that IFNγ was able to activate STAT-3 phosphorylation. Results in Fig. 8(b and c, middle panel) shown that IFNγ was able to activate STAT-3 phosphorylation in 727Ser, maximal effect being at 1 h stimuli, time at which p-STAT3 translocation to the nucleus was maximal too (Fig. 8b and c, right panel).

JAK/STAT pathway is involved in both nNOS phosphorylation and activation of iNOS expression induced by IFNγ

In order to know if JAK/STAT3 pathway activation induced by IFNγ could be mediated by IFNγ induction of nNOS phosphorylation and if activation of this pathway could mediate the increase in activation of iNOS expression induced by IFNγ, we tested the effect of specific inhibitors of JAK/STAT pathway on both, nNOS phosphorylation and iNOS mRNA expression induced by LPS + IFNγ, these effects being compared to the effects of another protein kinase inhibitors. Results from Fig. 9(a) show that nNOS phosphorylation induced by LPS + IFNγ was highly inhibited by the specific JAK family inhibitor JAKI1 (100 nM) and by AG490 (2 μM), a JAKs family tyrosine kinase inhibitor which inhibit STAT-3 activation mainly by JAK2, as well as by 5 μM PD98059, 1 μM KT5823 and 10 μM H89, concentrations at which these compounds specifically inhibit Mitogen Activated Protein Kinase Kinase (MEK or MAPKK), cGMP-dependent protein kinase (PKG) or cAMP-dependent protein kinase (PKA) activity, respectively, but not by 5 μM LY294002, a specific inhibitor of PI3K (Fig. 9a). These data suggest the participation of JAK pathway, as well as some other Ser-Tre kinases, in the activation of nNOS phosphorylation induced by IFNγ in chromaffin cells. However, only the inhibition of the STAT3 signaling pathway by 2 μM AG490 or 1 μM JAKI1, was able to inhibit by 90% and 50%, respectively, the mRNA iNOS expression, without affecting nNOS expression, while the above indicated PKA, PKG, and MEK inhibitors did not have a significant inhibitory effect on mRNA iNOS expression (Fig. 9b).

Figure 9.

 Effect of different protein kinase inhibitors on nNOS phosphorylation at 847Ser and on iNOS mRNA expression. Bovine chromaffin cells were challenged with a combination of LPS 10 μM plus IFNγ 10 nM in the absence or presence of protein kinase inhibitors (at concentrations indicated in the text) and 847Ser p-nNOS levels (1 h) (a) and mRNA iNOS expression (6 h) (b) were measured by western blot o real time PCR techniques as is described in Materials and Methods. Data were expressed as a quantification of results over control without LPS + IFNγ. Statistic compares the effect of protein kinase inhibitors on nNOS phosphorylation (a) or mRNA iNOS expression (b) induced LPS plus IFNγ alone. (***< 0.001) (one-way anova test).


Recent studies carried out by our group postulated that both, exogenous NO and endogenous NO, triggered by cytokines, induce apoptosis in chromaffin cells (Vicente et al. 2006; Pérez-Rodríguez et al. 2007). NO donors, therefore exogenous NO, induce apoptotic cell death mediated by NO and peroxynitrites, conclusions also supported in the literature (Nomura 2004). However, the effect of endogenous NO, which in these cells can be generated by cytokines (Turquier et al. 2002; Vicente et al. 2006) or glutamate (González et al. 1998; Arce et al. 2004), on apoptosis is more controversial, as NO can be synthesized in most cell types via different NOS isoforms (nNOS, eNOS, and iNOS), having these isoforms a tissue-specific localization and different effects on cell death and survival. Neurons mainly express nNOS (Förstermann et al. 1998), but also iNOS in pathological conditions such as trauma, inflammation and ischemia (Bredt 1999). In this work, we study the effect of endogenously NO generated by IFNγ on cell viability and apoptosis of chromaffin cells and study the possible involvement of both, nNOS and iNOS isoforms, in chromaffin cell death.

Chromaffin cells express both nNOS and iNOS

Our studies on the expression of specific NOS isoforms in chromaffin cells at both transcriptional and translational levels permit us to assess here the presence in bovine chromaffin cells of at least two NOS isoforms, nNOS and iNOS, having a different regulation. We outlined that iNOS expression, at protein level, is induced by cytokines while nNOS is not. The use of more accurate techniques in further essays, such as real-time PCR, permits us to define the regulation of both isoforms better. Thus, IFNγ or LPS, alone or in combination, induce a dose-dependent increase in the mRNA level of iNOS, effect that is abolished by dexamethasone in a dose–response manner, as described in many other cell types (Korhonen et al. 2002; Golde et al. 2003; Shinoda et al. 2003). On the other hand, nNOS expression is up-regulated by dexamethasone at both protein and mRNA levels. It has been described that dexamethasone inhibit the expression of inflammatory genes like iNOS, by a mechanism such as destabilizing iNOS mRNA (Korhonen et al. 2002), post-transcriptional level (Shinoda et al. 2003) or reduction in protein synthesis (Golde et al. 2003). However, the effects of dexamethasone on nNOS expression are more controversial. So, although there are many data on the literature indicating a decrease by glucocorticoids on nNOS expression in neuroblastoma cell lines (Schwarz et al. 1998) and in endotoxemic neonate rat brain (Wang et al. 2005) or rat lung, liver, and aorta of the rat (Knowles et al. 1990), there are also data in the literature showing an activation of nNOS expression by dexamethasone or an increase in its activity induced by glucocorticoids in cerebellar glial cells (Baltrons et al. 1995). Thus, it seems that nNOS expression is subjected to differential tissue-specific mechanisms. In this way, as the presence of glucocorticoid-responsive elements (GRE) in nNOS promoter has not been demonstrated, it is possible that the dexamethasone action on nNOS gene expressed in chromaffin cells is not a direct action but the result of interaction with other transcriptional factors like cAMP-response element binding protein (Zhong and Minneman 1993). On the other hand, the effect of glucocorticoids increasing nNOS expression and activity could produce very high NO levels which could repress NFκB activation and inhibit iNOS expression. These results suggest us a model where basal NO produced by nNOS could inhibit NFκB translocation to the nucleus, regulating iNOS expression.

iNOS is the NOS isoform mainly involved in chromaffin cell death

Our results here show that in chromaffin cells both, calcium dependent basal NOS activity (nNOS) and calcium-independent NOS activity-induced by IFNγ (iNOS), are found. In these cells W-1400, at doses between 1 and 10 μM, is a specific inhibitor of only iNOS and N-PLA (0.1-100 μM) or thiocitrulline (10–100 μM) are specific only of nNOS. Therefore, we used these inhibitor concentrations to study the participation of NO generated by nNOS and iNOS in chromaffin cell viability and apoptosis.

On the basis of results showing that iNOS inhibitor W-1400 completely reverted IFNγ-induced increase in NO levels and decrease in cell viability but nNOS inhibitors only partially reverted these effects, we conclude that iNOS is the main NOS isoform involved in cytokine-induced chromaffin cell death with minor participation of nNOS. The same results were obtained on apoptosis, that is, iNOS inhibitors were the strongest inhibitors of apoptosis induced by IFNγ.

NFκB activation and nNOS phosphorylation are involved in the mechanism of iNOS induction by IFNγ

The implication of transcriptional factor NFκB on iNOS regulation is well documented (Aktan 2004). We observed that cytokines, activators of apoptosis, do activate NFκB, promoting its translocation to the nucleus. However, the fact that an increase of apoptosis-induced by cytokines is produced when NFκB translocation to the nucleus is avoided, indicates that NFκB might be a survival factor, probably because it activates other survival gene in addition to iNOS activation.

In respect to NFκB participation on iNOS mRNA and protein expression, our studies with SN50, a specific inhibitor of NFκB translocation into the nucleus, indicate that SN50 was able to inhibit iNOS expression at both transcriptional and translational level, but only about 30–50% at SN50 doses used. These results suggest that NFκB participation in iNOS induction is not the only mechanism mediating iNOS gene expression and indicate that this transcriptional factor could also be mediating the expression of other survival genes which counteract death genes mediating apoptosis.

Regarding the possible participation of nNOS in IFNγ induction of iNOS and apoptosis through NFκB, some authors defend an interesting perspective of apoptotic regulation, that is, cytokine induction of iNOS is regulated by nNOS through NFκB (Colasanti et al. 1995; Togashi et al. 1997). Our results show that different nNOS inhibitors induced NFκB activation, showing that nNOS inhibition in basal conditions was enough to activate NFκB and thus iNOS expression. Therefore, basal NO levels, generated by constitutively expressed nNOS, could block NFκB activation and thus, iNOS gene expression. Taken these data all together with the increase in NFκB translocation to the nucleus induced by IFNγ, we can deduce that IFNγ activates iNOS expression, inducing NFκB translocation to the nucleus, this effect being regulated by nNOS.

Nitric oxide has been shown to affect the activity of NFκB and other transcriptional factors through S-nitrosylation. In the case of NFκB, NO mainly affects their transcriptional activities indirectly, by S-nitrosylation of IκB kinases (IKKs), that preventing the phosphorylation and degradation of IκB and thus, inhibiting the activation of NFκB pathway (Reynaert et al. 2004; Kenny and Chung 2006). In our model, NO donors inhibited NFkB activation and iNOS expression (data not shown). So, the low NO levels generated by cytokines or by NOS inhibitors could increase NFkB activation probably by avoiding NO-induced S-nitrosylation of NFkB.

Our results agree with a very interesting hypothesis brought up in the literature. This model proposes that iNOS expression is regulated by nNOS in the following manner: constitutive NO produced by nNOS would be enough to avoid NFκB translocation to the nucleus. However, once there, NFκB would activate iNOS expression, producing enormous amounts of NO and apoptosis. In this model, cytokines would inhibit nNOS via phosphorylation, avoiding NO formation and favoring NFκB translocation to the nucleus (Mariotto et al. 2004; Conti et al. 2007). To test the accuracy of this hypothesis on our cell model, chromaffin cells were subjected to cytokine challenge and nNOS phosphorylation was measured. nNOS isoforms have got different sites for serine or tyrosine phosphorylation. nNOS can be phosphorylated by PKA, PKG and Ca2+/calmodulin-dependent protein kinase II (Bredt et al. 1992). Phosphorylation of nNOS by both PKG and PKA diminishes its catalytic activity (Dinerman et al. 1994). Down-regulation of the nNOS activity by phosphorylation leads to a lower NO concentration, and therefore to NFκB activation and iNOS expression. Phosphorylation of nNOS in serine 847 has been widely studied, providing a down-regulation of activity in many cell types (Nakane et al. 1991; Hayashi et al. 1999). Here, we observed that cytokine challenge of chromaffin cells provided an increase in the nNOS phosphorylation in ser847, maximal at 1 h. These results occur in parallel to maximal NFκB translocation into the nucleus after the same stimuli, pointing at a possible interaction between both events. The serine threonine kinases PKA, PKG could be involved in nNOS phosphorylation as specific inhibitors of these enzymes were able to strongly inhibit this cytokine effect. Also, the dual kinase MEK, but not PI3K, could be involved in this Ser-nNOS phosphorylation, probably via extracellular signal-regulated kinases (ERKs) because specific inhibitors of p38MAPKs and c-Jun N-terminal kinases (JNKs) were much less effective on Ser-nNOS phosphorylation inhibition than MEK inhibitor PD98059 (data not shown).

JAK/STAT pathway is involved in the mechanism of iNOS induction by IFNγ

Another pathway involved in iNOS expression has been described for the JAK/STATs family (Bolli et al. (2003) and Stempelj et al. (2007). The human NOS II promoter contains consensus sequences for the binding of transcription factors, including IFNγ regulatory factor-1, STAT binding to interferon-gamma activated sites (GAS) elements, activator protein 1 (AP-1), NFκB, and others (Spitsin et al. 1996; Linn et al. 1997). In murine messagial cells it has been described that there might be a cross-talk between the STATs and NFκB transcription factors. Thus, STAT3, via direct interactions with NF-kappaB p65, serves as a dominant-negative inhibitor of NF-kappaB activity to suppress indirectly cytokine induction of the iNOS promoter in these cells (Yu et al. 2002). Although current knowledge on the role of STAT3 concerning regulation of the human iNOS gene is still fragmentary, the recent use of interfering RNA technology identified STAT3 as being crucial for up-regulation of iNOS, demonstrating that the interleukin-22/STAT3 pathway potentiates expression of iNOS in human colon carcinoma cells (Zieschéet al. 2007). As observed with nNOS phosphorylation, our experiments report an increase in STAT3-phosphorylation and p-STAT3 nuclear translocation maximal at 1 h after IFNγ stimulation of chromaffin cells. Moreover, JAKI1, a specific inhibitor of JAK family tyrosine kinases, and AG490, a selective inhibitor of STAT-3 phosphorylation by JAKs, were able to inhibit both nNOS phosphorylation and iNOS mRNA expression induced by IFNγ, while other protein kinase inhibitors were only able to inhibit nNOS phosphorylation without affecting iNOS mRNA expression.

So, on the basis of these results, we point at a possible p-nNOS-p-STAT3-NFκB cross-talk in chromaffin cells, JAK/STAT pathway potentiating iNOS expression probably by increasing nNOS phosphorylation and NFκB activity.

In short, at the view of results from this paper, we put forward that endogenous NO has a very important role on the regulation of apoptosis of chromaffin cells and we propose a model to explain the effects of endogenous NO, generated by IFNγ stimulation, on cell death of bovine chromaffin cells.

  • • In basal conditions, chromaffin cells express only nNOS. Physiological concentrations of NO produced by nNOS would be enough to stop NFκB activation, as it has been observed that nNOS inhibitors promote its activation.
  • • IFNγ, probably by promoting tyrosine kinase activity of JAK, induces nNOS phosphorylation on bovine chromaffin cells, diminishing NO basal levels, thus allowing NFκB translocation to the nucleus, and DNA binding. This would, in turn, activate iNOS gene expression, increasing iNOS protein expression and producing great amounts of NO, involved in apoptosis.
  • • Nuclear factor κB would thereby act as a survival or death factor, depending on the pathways involved and the cell conditions. This double role would mean that small activation of NFκB could permit cell death, whereas a great activation would stop it. Activation/inhibition mainly depends on the time curse of cellular and molecular events.
  • • The JAK-STAT3 pathway would be involved in these effects by modulating nNOS phosphorylation and iNOS expression, the last effect probably by modulating NFκB activity.

To sum up, both isoforms (nNOS and iNOS) and NFκB are involved in the IFNγ-induced apoptosis in chromaffin cells. The iNOS expression in chromaffin cells, shown by the first time in this paper, suggest that, as well as in neurons (Minc-Golomb et al. 1996), this enzyme could contribute significantly to the vulnerability of the neural cells to various inflammatory insults and once more support the use of chromaffin cells as a neural model to study molecular mechanisms of neuronal cell death underlying neurodegenerative diseases.


This study was supported by the grants BFI2003-03886 from Ministry of Science and Technology, (MCYT, Spain) and SAF2006-05563 from Ministry of Education (MEC, Spain). R. Pérez-Rodríguez has a contract from Spanish Ministry of Health (Instituto de Salud Carlos III) RETICS-RD06/0026 and A. M. Olivan has a fellowship from MEC. The authors thank Javier Morón-Oset for his help in improving the manuscript.