Induction of motor neuron apoptosis by free 3-nitro-l-tyrosine


Address correspondence and reprint requests to Alvaro G. Estèvez, PhD, Department of Physiology and Biophysics, The University of Alabama at Birmingham, MCLM 850, 1918 University Boulevard, Birmingham, AL 35294-0005, USA. E-mail:


Peroxynitrite-dependent tyrosine nitration has been postulated to be involved in motor neuron degeneration in amyotrophic lateral sclerosis (ALS). Evidence supporting this supposition includes the appearance of both free and protein-linked 3-nitro-l-tyrosine (nitrotyrosine) in both sporadic and familial ALS, as well as of increased free nitrotyrosine levels in the spinal cord of transgenic mice expressing ALS-linked superoxide dismutase mutants at symptom onset. Here we demonstrate that incubation with clinically relevant concentrations of nitrotyrosine induced apoptosis in motor neurons cultured with trophic factors. Nitrotyrosine was bound to proteins, but it was not incorporated into α-tubulin, as previously demonstrated for other cell types. Neither inhibition of nitric oxide production nor scavenging of superoxide and peroxynitrite prevented increases in cell nitrotyrosine immunoreactivity or motor neuron death, suggesting that these effects are not due to the endogenous formation of reactive nitrogen species. In contrast, some populations of astrocytes incorporated nitrotyrosine into α-tubulin, but free nitrotyrosine had no effect on the viability and phenotype of astrocytes in culture, as evaluated by glial fibrillary acidic protein immunoreactivity, cell growth and morphology. Co-culture of motor neurons on astrocyte monolayers delayed, but did not prevent, nitrotyrosine-induced motor neuron death. These results suggest that free nitrotyrosine may play a role in the induction of motor neuron apoptosis in ALS.

Abbreviations used

amyotrophic lateral sclerosis


brain derived neurotropic factor






iron 5,10,15,20-tetrakis-4-carboxy phenyl porphyrin


glial-derived neurotrophic factor


glial fibrillary acidic protein




nitro-l-arginine methyl ester


manganese tetrakis (4-benzoyl acid) prophyrin


3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide


phosphate-buffered saline


sodium dodecyl sulfate


superoxide dismutase





Peroxynitrite-mediated tyrosine nitration has been suggested as a major mechanism for the induction of neuron death in several neurodegenerative disorders including amyotrophic lateral sclerosis (ALS) (Beckman et al. 1992, 1993, 1995; Ischiropoulos et al. 1992; Beckman and Crow 1993; Varner and Beckman 1994; Ischiropoulos 1998). The spinal cords of patients with sporadic and familial ALS show immunoreactivity for 3-nitro-l-tyrosine (nitrotyrosine) (Abe et al. 1995, 1997; Chou et al. 1996a, 1996b; Beal et al. 1997; Ferrante et al. 1997a; Tohgi et al. 1999). In more recent studies of transgenic mice carrying the G93A mutant Cu/Zn human superoxide dismutase (SOD1) expressed in the familial form of ALS, immunoreactivity was observed in the spinal cord of symptomatic animals (Ferrante et al. 1997b).

Although it remains a matter of debate whether protein-linked nitrotyrosine is present in transgenic mice expressing ALS-linked SOD mutants (Bruijn et al. 1997a, 1998), it is generally agreed that increased levels of free nitrotyrosine are observed in both transgenic mice and patients with ALS (Beal et al. 1997; Bruijn et al. 1997a; Ferrante et al. 1997b; Tohgi et al. 1999). The concentration of free nitrotyrosine in the ventral spinal cord of patients with ALS is double that found in control individuals (Beal et al. 1997). In addition, an increase in the proportion of nitrotyrosine from 2% of net free tyrosine in controls to 6% in the mice carrying mutated SOD occurs at the onset of clinical symptoms (Bruijn et al. 1997a).

The spinal cords of patients with ALS and transgenic mice expressing mutant SOD1 are the sites not only of progressive motor neuron degeneration but also of astrocyte alterations, suggesting a potential pathogenic role for astrocytes (Wong et al. 1995; Hirano 1996; Schiffer et al. 1996; Bruijn et al. 1997b; Kato et al. 1997; Hall et al. 1998). Activated astrocytes can influence the final outcome following a range of insults, as they can produce trophic factors, induce neuronal death or inhibit axonal growth depending on the conditions (Chen and Swanson 2003). Much effort is being focused on understanding the mechanisms that trigger reactive astrocytosis (Ridet et al. 1997; Cassina et al. 2001; McGraw et al. 2001). Mice expressing a G85R mutant SOD1 show abnormal protein inclusions in astrocytes even at an early, clinically presymptomatic period when motor neuron pathology is not yet apparent (Bruijn et al. 1997b). Reactive astrocytes expressing inducible nitric oxide synthase and displaying hyaline inclusions are evident in both familial and sporadic ALS (Kato et al. 1996; Almer et al. 1999) and stimulated motor neuron apoptosis in culture (Cassina et al. 2002).

The potential role of nitrotyrosine as an innocuous by-product of reactive nitrogen species formation or as an effector molecule capable of inducing motor neuron death, either directly or by altering astroglial cells, remains to be elucidated. We hypothesized that free nitrotyrosine may play a role in the motor neuron degeneration seen in ALS. To test this hypothesis, we studied the effects of clinically relevant concentrations of nitrotyrosine on the survival of rat motor neuron cultures and of motor neurons co-cultured over a spinal cord astrocyte monolayer.

Materials and methods

Cell culture

Motor neuron cultures were prepared as described previously (Henderson et al. 1995; Estévez et al. 1998b; Raoul et al. 1999). Briefly, rat embryo (E15) spinal cords were dissected, and the dorsal half of each cord was removed. Ventral cords were chopped into pieces and incubated in phosphate-buffered saline (PBS) supplemented with 0.05% trypsin for 15 min at 37°C, followed by mechanical dissociation. Motor neurons were then purified by centrifugation on a metrizamide cushion and immunoaffinity using IgG 192 against p75 neurotrophin receptor (Chandler et al. 1984). Motor neurons were plated at a density of 280 cells/cm2 in 35-mm dishes precoated with polyornithine and laminin in neurobasal medium supplemented with B27 supplement, glutamate, glutamine, 3-mercaptoethanol and antibiotics, as described previously (Pennica et al. 1996; Estévez et al. 2000). Cultures were maintained at 37°C in a 5% CO2 humidified atmosphere. On each occasion the cells were counted in time course experiments, two-thirds of the culture medium was replaced by fresh medium containing supplements for the total volume. More than 95% of the cells were found to be immunoreactive for the motor neuron markers Islet 1/2 (4D5 monoclonal antibody from the Developmental Studies Hybridoma Bank, Iowa City, IA, USA) (Ericson et al. 1992; Tsuchida et al. 1994) and p75 neurotrophin receptor, as described previously (Estévez et al. 1998b).

Primary cultures of astrocytes were prepared from newborn rats, as described previously (Saneto and Vellis 1987; Peluffo et al. 1997). Briefly, suspensions of cells obtained from spinal cords were plated at a density of 1.5 × 106 cells per 25-cm2 culture flask in Dulbecco's modified Eagle's medium (Hyclone, Logan, UT, USA) supplemented with 10% fetal bovine serum, penicillin (100 IU/mL) and streptomycin (100 µg/mL) (Gibco-Invitrogen Carlsbad, CA, USA). After 3 days, the culture medium was replaced every other day until the cells reached confluence. Then the monolayers were shaken for 48 h. After rinsing, the astrocyte monolayers were incubated for an additional 48 h with cytosine arabinoside (10 µm) (Sigma, St Louis, MO, USA) to eliminate the remaining proliferating microglia and oligodendrocytes. Astrocytes were plated at a density of 2 × 104 cells/cm2 in 35-mm dishes and allowed to achieve confluence before being used in experiments, unless otherwise indicated. The culture purity was > 98% as determined by immunoreactivity for glial fibrillary acidic protein (GFAP), as described previously (Peluffo et al. 1997).

In the co-culture experiments, motor neurons were plated as described above on the astrocyte monolayers. Nitrotyrosine (Sigma), and vehicle were added directly to the culture medium. The co-cultures were incubated in L15 supplemented with 22 mm sodium bicarbonate, 0.1 mg/mL conalbumin (Gibco-Invitrogen), 0.1 mm putrescine, 5 µg/mL insulin (Sigma), 30 nm sodium selenite (Sigma), 20 mm glucose (Gibco-Invitrogen), 20 nm progesterone (Sigma), 100 IU/mL penicillin (Gibco-Invitrogen), 100 µg/mL streptomycin (Gibco-Invitrogen) and 2% horse serum, as detailed elsewhere (Peluffo et al. 1997; Cassina et al. 2002).


Cultures were fixed with paraformaldehyde plus glutaraldehyde (Sigma) on ice for 15 min or incubated with methanol at − 20°C for 5 min before the staining was performed as described previously (Estévez et al. 1998b; Eiserich et al. 1999; Cassina et al. 2002). The antibodies used were rabbit anti-nitrotyrosine (1 : 500) (Ye et al. 1996; Viera et al. 1999), anti-nitrotyrosined tubulin (1 : 600) (Bisig et al. 2002), anti-dynein heavy chain (Santa Cruz Biotechnology, Santa Cruz, CA, USA), and mouse monoclonal antibodies against α-tubulin clone DM1A (1 : 1000) (Sigma) and against GFAP (1 : 500) (Sigma). Images were obtained using an epifluorescence IX70 microscope (Olympus, Melville, NY, USA) equipped with an Olympix digital camera controlled by the Ultraview software (Perkin-Elmer, Boston, MA, USA). Images were composed using Photoshop 7.0 (Adobe, Seattle, WA, USA).

Cell viability

Motor neuron survival was assessed by counting all phase-bright cells displaying intact neurites longer than four cell diameters in a prefixed area (1 cm2) of 35-mm dishes, or two well diameters in four-well plates. In the co-cultures, all cells meeting the previous criteria and showing immunoreactivity for the p75 neurotrophic receptor were counted (Cassina et al. 2002). All values are expressed as a percentage of the number of motor neurons present in parallel cultures maintained with brain derived neurotrophic factor (BDNF) (0.1 ng/mL) (R & D Systems, Minneapolis, MN, USA) (Estévez et al. 1998b).

Astrocyte viability was estimated using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) (Sigma) assay, which was performed as described previously (Estévez et al. 1995). In addition, after staining nine different fields with 4,6-diamidino-2-phenylindole (DAPI, 1 µg/mL; Molecular Probes, Eugene, OR, USA), we used fluorescence intensity and morphology to differentiate between and to count total and apoptotic nuclei. Finally, the rate of release of lactate dehydrogenase into the medium was determined by measuring the rate of decline of NADH at 340 nm from a solution containing 22.77 mm sodium pyruvate and 0.47 mm NADH in 0.1 m KHPO4 buffer (pH 7.4) at 37°C. The amount of enzyme reducing the absorbance in 0.001 units in 1 min was defined as one enzymatic unit.

Cell growth

Astrocytes were plated as described for the viability assay; 24 h after the culture medium was replaced by fresh medium containing the indicated concentrations of nitrotyrosine or tyrosine. Three days later, astrocyte cultures were harvested and counted using a hemocytometer. To determine DNA synthesis, the cultures were incubated for 3 days with 5 µCi/mL [3H]thymidine (Amersham Biosciences, Piscataway, NJ, USA), washed twice with PBS at 37°C, incubated with ice-cold 10% trichloroacetic acid on ice for 10 min, and then washed twice more with trichloroacetic for 5 min. The DNA was resuspended in 0.3% NaOH + 1% sodium dodecyl sulfate (SDS). Samples were measured in a liquid scintillation counter.

Western blotting and dot blotting

After 3 days of incubation with or without nitrotyrosine, growing and confluent astrocyte cultures were harvested in 50 µL lysis buffer (25 mm HEPES, 5 mm MgCl2, 5 mm EDTA, 2 mm dithiothreitol, 1 mm phenylmethylsulfonyl fluoride, 1% Sigma protease inhibitor cocktail, 1% Triton X-100, 1 mm Na3VO4, 1% SDS) for determination of GFAP. Some 10 µg protein (50 µg for GFAP) from whole-cell lysates was loaded per lane and then separated using SDS-polyacrylamide gel electrophoresis (10% gel). The protein was transferred to a polyvinylidene difluoride membrane for immunoblotting, as detailed elsewhere (Eiserich et al. 1999). The following dilutions of the primary antibodies were used: 1 : 500 for monoclonal against GFAP (Sigma), 1 : 2000 for rabbit polyclonal against nitrotyrosine (Ye et al. 1996; Viera et al. 1998; Brito et al. 1999), 1 : 1000 for mouse monoclonal against α-tubulin (clone B5-1-2; Sigma) and 1 : 1000 for mouse monoclonal against tyrosinated tubulin (clone TUB-1A2; Sigma). Primary antibodies were detected using goat anti-mouse and anti-rabbit IgG horseradish peroxidase-conjugated antibodies at a dilution of 1 : 10 000 (Bio-Rad, Hercules, CA, USA) and, in the case of GFAP, visualized using Supersignal West Pico chemiluminescent substrate (Pierce, Rockford, IL, USA). All other primary antibodies were detected using horseradish peroxidase-conjugated anti-rabbit and anti-mouse at 1 : 6000 dilution (Dako, Carpinteria, CA, USA) and visualized by enhamced chemiluminescence (Amersham Biosciences).

Motor neurons were plated at a density of 106 cells in 13-mm wells as described previously. Nitrotyrosine was added at the time of plating at a concentration of 500 µm. Sixteen hours after plating the cell were rinsed with PBS (supplemented with Ca2+ and Mg2+) and resuspended in lysis buffer [50 mm Tris, pH 7.5, 150 mm NaCl, 2 mm EGTA 1% triton X100, 1 mm phenylmethylsulfonyl fluoride, 1 × phosphatase inhibitor cocktail II (Sigma), 1 × protease inhibitor cocktail (Sigma)]. Cells were homogenized by sonication, centrifuged (10 000 g, 10 min, 4°C) and loaded on to a polyvinylidene difluoride membrane using a Bio Dot Microfiltration Unit (Bio-Rad). The membranes were processed as described for western blotting, using the anti-nitrotyrosine polyclonal antibody (1 : 500).

Statistical analysis

Reported values are the mean ± SD for the number of samples indicated for each experiment. Values were analyzed by anova followed by the Tukey test using Instat 3.0 (GraphPad Software, San Diego, CA, USA). p < 0.05 was considered significant.


Incubation of BDNF-treated motor neuron cultures with nitrotyrosine resulted in its incorporation in the motor neuron soma (Fig. 1b). The intensity and localization of nitrotyrosine immunofluorescence did not vary between cultures fixed with 4% paraformaldehyde and those fixed and permeabilized with methanol at − 20°C (not shown), indicating that the nitrotyrosine had bound itself to proteins. In addition, increased nitrotyrosine immunoreactivity was seen by dot blot analysis of proteins from free nitrotyrosine-treated motor neurons (Fig. 1e). Nitrotyrosine immunoreactivity was limited to a perinuclear area devoid of microtubules (Fig. 1b). The soma and proximal neurites of motor neurons cultured with BDNF showed immunoreactivity for dynein (Fig. 1c). Treatment with nitrotyrosine induced a decrease in the immunoreactivity for dynein, which was limited to the perinuclear area of the cells (Fig. 1d).

Figure 1.

Incorporation of 3-nitrotyrosine in cultured motor neurons. Three-day-old BDNF-treated motor neuron cultures were incubated for 24 h with 8% nitrotyrosine (32 µ m ). After fixation, cultures were processed for immunocytochemistry. (a) Control motor neurons in the presence of BDNF (0.1 ng/mL) visualized by immunocytochemistry against α-tubulin (FITC, green) did not show detectable levels of nitrotyrosine immunoreactivity (Cy3, red). (b)  Incubation with nitrotyrosine resulted in incorporation of the modified amino acid into motor neuron soma (red), but no co-localization with α-tubulin was observed. (c)  Motor neurons cultured with BDNF showed a widespread distribution of immunofluorescence for dynein. (d)  Incubation with nitrotyrosine decreased the intensity of dynein immunofluorescence and limited the localization to the soma. (e)  Incorporation of nitrotyrosine into proteins. Proteins from BDNF-treated motor neurons incubated without or with 500 µ m nitrotyrosine (NT) were analyzed by dot blot for nitrotyrosine . Controls were performed by blotting culture media with and without 500 µ m nitrotyrosine.

Incubation of BDNF-treated motor neurons with free nitrotyrosine at the time of plating induced cell death after 3 days in culture (Fig. 2a). Motor neuron death was concentration dependent for nitrotyrosine, with 50% of motor neuron death induced by approximately 5% (20 µm) of the net tyrosine (400 µm) being nitrotyrosine. The maximum effect was achieved with 12% nitrotyrosine in the culture medium, leading to an ∼ 80% death rate for motor neurons. To test whether nitrotyrosine affected neurite outgrowth rather than survival, 1-day-old motor neuron cultures were incubated with increasing concentrations of nitrotyrosine for a further 3 days (Fig. 2b). Motor neuron survival was similar under both conditions, suggesting that the effects of nitrotyrosine on motor neurons were not due to inhibition of cell attachment or neurite growth.

Figure 2.

Nitrotyrosine-induced apoptosis in purified motor neuron cultures. (a)  Motor neurons cultured with BDNF (0.1 ng/mL) were incubated with the indicated concentrations of nitrotyrosine (▪). (b)  Motor neurons were cultured with BDNF for 24 h before the addition of nitrotyrosine at the indicated concentrations (▪). The motor neuron survival rate was significantly reduced at nitrotyrosine concentrations of ≥ 4% ( p <  0.001, anova ). Incubation with tyrosine (○) did not affect motor neuron survival in either condition. (c)  Motor neuron cultures grown in the presence of BDNF (0.1 ng/mL) were incubated with 8% nitrotyrosine (NT; 32 µ m ) and caspase inhibitors (cmk-YVAD 50 µ m ; fmk-zVAD 25 µ m , fmk-DEVD 25 µ m ) were added after 1 day in culture. Caspase inhibitors significantly prevented motor neuron death triggered by nitrotyrosine (* p <  0.001, anova ). Motor neuron survival was determined 3 days after the addition of nitrotyrosine or tyrosine. Values are the mean ± SD of at least three independent experiments performed in duplicate or quadruplicate.

After incubation with nitrotyrosine, dying motor neurons showed nuclear condensation and fragmentation, shrinkage of the cytoplasm and disintegration of neurites, suggesting apoptosis. To determine whether other conditions in addition to apoptotic cell death could mediate the effects of nitrotyrosine, cultures treated with nitrotyrosine were incubated with caspase inhibitors (Fig. 2c). Inhibition of caspases with Ac-Tyr-Val-Ala-Asp-chloromethylketone (YVAD) (50 µm; caspase 1), Ac-Asp-Glu-Val-Asp-fluoromethylketone (DEVD) (25 µm; caspase 3) and z-Val-Ala-Asp-fluoromethylketone (zVAD) (50 µm; general caspase inhibitor) significantly diminished the motor neuron apoptosis induced by 32 µm nitrotyrosine (8% of total free tyrosine) in the presence of BDNF.

Nitrotyrosine-induced motor neuron death was delayed but not prevented when motor neurons were cultured with multiple trophic factors (Fig. 3). Iron 5,10,15,20-tetrakis-4-carboxyphenyl porphyrin (FeTCPP), a superoxide and peroxynitrite scavenger (Patel and Day 1999; Crow 2000), prevented motor neuron apoptosis induced by trophic factor deprivation in a dose-dependent manner (Fig. 4a) with an EC50 of 50 nm. In addition, FeTCPP maintained motor neuron survival for more than 15 days in culture (Fig. 4b). However, the apoptosis induced by 20% nitrotyrosine was not inhibited in trophic factor-treated motor neurons by incubation with the nitric oxide synthase inhibitor nitro-l-arginine methyl ester (L-NAME), and the superoxide and peroxynitrite scavengers FeTCPP and manganese tetrakis (4-benzoyl acid) prophyrin (MnTBAP) (Fig. 4c). Similar results were observed when motor neurons were cultured with the same combination of trophic factors or BDNF alone and 8% nitrotyrosine (not shown).

Figure 3.

Effect of nitrotyrosine on motor neurons incubated with multiple trophic factors. Motor neurons were plated with a combination of trophic factors (NTFs: BDNF 1 ng/mL; glial-derived neurotrophic factor (GDNF) 0.1 ng/mL; cardiotrophin I 10 ng/mL) plus 8% (32 µ m ; filled bars) or 20% (100 µ m ; empty bars) nitrotyrosine. The survival of trophic factor-treated motor neurons was not affected by incubation with either 32 µ m (gray bars) or 100 µ m (dotted bars) tyrosine (NTFs + Tyr) as determined after 7 days in culture compared with the survival of trophic factor-treated motor neurons after 24 h in culture. Motor neuron survival in the presence of nitrotyrosine was determined at the indicated times. Motor neuron survival after 1 day in culture in the presence of NTFs was considered to represent 100% survival. Values are the mean ± SD of at least two independent experiments performed in duplicate or quadruplicate. * p <  0.01 versus respective tyrosine control at 7 days ( anova ) .

Figure 4.

FeTCPP prevents motor neuron apoptosis induced by trophic factor deprivation. (a)  Motor neurons were cultured in the presence of the indicated concentrations of FeTCPP. Motor neuron survival was determined 3 days after plating. (b)  Motor neurons were cultured in complete neurobasal medium without trophic factors (○) or with 1 µ m FeTCPP (▪). Values are the mean ± SD of three independent experiments performed in duplicate. (c)  Motor neurons were cultured with a combination of trophic factors alone (NTFs: BDNF 1 ng/mL; GDNF 0.1 ng/mL; cardiotrophin I 10 ng/mL) or supplemented with 20% nitrotyrosine (NT 20%). Motor neurons exposed to nitrotyrosine were also incubated with FeTCPP (1 µ m ), MnTBAP (100 µ m ) or L-NAME (100 µ m ). Neither inhibition of the formation of, nor increased scavenging by, peroxynitrite prevented apoptosis induced by free nitrotyrosine. Motor neuron survival was determined after 3 days in culture. Values are the mean ± SD of three independent experiments performed in duplicate or quadruplicate.

To analyze the effect of nitrotyrosine on astrocyte-derived neurotrophic activity, cells were incubated with increasing concentrations of nitrotyrosine. Addition of nitrotyrosine to motor neurons co-cultured on top of spinal cord astrocyte monolayers at the time of neuronal plating induced significant motor neuron apoptosis (Fig. 5). Astrocyte monolayers partially prevented the toxicity of free nitrotyrosine (up to 20%) on motor neurons after 3 days. The same effect was observed when nitrotyrosine was added 24 h after plating of motor neurons on the astrocyte monolayers (not shown).

Figure 5.

Effect of astrocytes on motor neuron apoptosis induced by nitrotyrosine. Motor neurons were cultured on a spinal cord astrocyte monolayer in the presence of the indicated concentrations of nitrotyrosine (▪). After 3 days, the cultures were fixed and processed for immunocytochemistry against p75 neurotrophic receptor. All cells immunoreactive for p75 neurotrophic receptor displaying neurites longer than four soma diameters were counted as viable cells. Values are the mean ± SD of at least three independent experiments performed in duplicate or quadruplicate. Nitrotyrosine induced a significant decrease in motor neuron survival at concentrations of ≥ 10% (50 µ m ) of the culture media ( p <  0.05). Tyrosine (○) did not affect motor neuron survival.

The effect of nitrotyrosine on astrocyte survival and phenotype was also evaluated. Incubation of subconfluent astrocyte cultures with nitrotyrosine in proportions of up to 50% of net tyrosine did not affect cell viability after 3 days, as evaluated by MTT reduction, nuclear morphology and lactate dehydrogenase release (Table 1). Moreover, incubation of astrocyte monolayers with nitrotyrosine affected neither cellular morphology nor GFAP immunoreactivity or expression (Fig. 6). It was impossible to distinguish between GFAP-expressing cells in control and nitrotyrosine-treated cultures. In addition, incubation of subconfluent astrocyte cultures with nitrotyrosine had no effect on astrocyte growth or DNA synthesis, as assessed by cell number and [3H]thymidine incorporation respectively (Table 1), even though there was evidence from immunofluorescence and immunoblotting analyses that nitrotyrosine had been incorporated into α-tubulin (Figs 7 and 8). The incorporation of nitrotyrosine into tubulin was dose dependent and was detected following exposure of the cells to a concentration of 370 µm nitrotyrosine (∼ 50% of total tyrosine) in the culture medium, a far higher concentration than that characterizing any reported neuropathological condition. Nitrotyrosine induced a dose-dependent decrease in the levels of tyrosinated α-tubulin, and an increase in the total α-tubulin cell content (Figs 7b and c). Surprisingly, after 3 days' incubation not all astrocytes incorporated nitrotyrosine into α-tubulin in confluent cultures, at concentrations as high as 50% (Fig. 8). Astrocytes immunoreactive for nitrotyrosine were found next to astrocytes with an identical morphology but showing no immunoreactivity for the modified amino acid (Fig. 8). Furthermore, the level of nitrotyrosine incorporation into α-tubulin was not found to be dependent on a specific astrocyte morphology (Fig. 8) or on levels of GFAP expression (not shown). A larger proportion of astrocytes became immunoreactive for nitrotyrosine in non-confluent, nitrotyrosine-treated cultures than in confluent monolayers (not shown), wherein nitrotyrosine immunoreactivity was restricted to isolated cell patches (Fig. 8).

Table 1.  Effect of 3-nitrotyrosine on astrocyte viability and growth
TreatmentViabilityCell growth
Apoptotic nuclei
Cell count
  1. All data are mean ± SD of at least three independent experiments performed in duplicate. LDH, lactate dehydrogenase. nd, not determined.

Control100 ± 10 (n = 5)100 ± 3 (n = 6)0.4 ± 0.4 (n = 5)100 ± 15 (n = 9)100 ± 10 (n = 12)
Nitrotyrosine 10%103 ± 6 (n = 5)106 ± 4 (n = 6)0.2 ± 0.3 (n = 6)113 ± 9 (n = 12)114 ± 14 (n = 11)
Nitrotyrosine 50%103 ± 8 (n = 4)106 ± 9 (n = 6)0.3 ± 0.3 (n = 6)ndnd
Figure 6.

Effect of nitrotyrosine on the expression of GFAP by spinal cord astrocyte monolayer. Astrocyte cultures were grown to confluence, then incubated for 3 days with (a, c) or without (b, d) nitrotyrosine. The cultures were then fixed and stained using two different antibodies against GFAP. Images in (a) and (b) were captured using a 20 × objective, and those in (c) and (d) with a 100 × objective. Even at the higher magnification, no morphological differences were observed between astrocytes cultured with or without nitrotyrosine. (e) As determined by immunoblot analysis, incubation of astrocyte monolayers with nitrotyrosine (NO 2 Tyr) did not alter GAFP expression from that seen in non-treated cultures (CTL).

Figure 7.

Effect of free nitrotyrosine on the levels of protein-linked nitrotyrosine, tyrosinated α-tubulin and α-tubulin in growing astrocyte cultures. Astrocytes plated at 10 × 10 3 cells/cm 2 were incubated for 3 days with the indicated concentrations of nitrotyrosine before being processed for immunoblotting. (a) Membranes stained for nitrotyrosine show a distinct band at the expected molecular weight for α-tubulin (55 KDa). The positive control (C) is α-tubulin isolated from Trypanosoma cruzi cultures (MW 49.8 KDa) incubated with free nitrotyrosine. (b)  Membranes were stained for tyrosinated tubulin. A notable decrease in the C-terminus tyrosine of α-tubulin is evident with increasing concentrations of free nitrotyrosine. (c)  Membranes stained for α-tubulin. A notable increase in the level of α-tubulin is induced by free nitrotyrosine.

Figure 8.

Nitrotyrosine and α-tubulin immunoreactivity in non-confluent astrocyte cultures. Astrocytes were plated at a density of 2 × 10 4 cells/cm 2 in chambered slides without (Control) or with 370 µ m nitrotyrosine (50%; NO 2 Tyr) in the culture medium. Astrocyte cultures were allowed to achieve confluence before incubation without (Control) or with 370 µ m nitrotyrosine (NO 2 Tyr) in the culture medium. The monolayers were fixed 3 days after the addition of nitrotyrosine and processed for immunofluorescence. Cells were stained using monoclonal antibodies against α-tubulin visualized with FITC-conjugated goat anti-mouse IgG (Anti-tubulin), rabbit polyclonal antibodies against nitrotyrosine (Anti-NO 2 Tyr) and DAPI for counterstaining of the nuclei. The co-localization of nitrotyrosine and α-tubulin was determined by merging the nitrotyrosine and α-tubulin immunofluorescence images (Merge) and is evidenced by the yellow color of the cells. Detailed co-localization is particularly obvious at 100 × magnification.


Nitrotyrosine is incorporated into the carboxyl end of α-tubulin by the enzyme tubulin-tyrosine ligase (Eiserich et al. 1999; Bisig et al. 2002), leading to changes in the morphology and activity of cells (Eiserich et al. 1999; Chang et al. 2002). Intrastriatal injection of free nitrotyrosine and 6-hydroxydopamine produced a similar loss of tyrosine hydroxylase-containing neurons in mice (Mihm et al. 2001), suggesting that free nitrotyrosine can stimulate neuronal death. Because nitrotyrosine is incorporated into the carboxyl end of α-tubulin, which is the domain for the binding of the motor protein dynein, alterations in cellular transport and in the binding of dynein have been proposed as possible explanations for the apoptopic effects of free nitrotyrosine (Eiserich et al. 1999; Chang et al. 2002). Nitrotyrosine induced the same alterations in the cellular distribution of dynein in motor neurons (Fig. 1) as it had in other cell types (Eiserich et al. 1999) indicating that changes in interactions between tubulin and dynein might be responsible for the induction of motor neuron death. Lending support to this hypothesis, a recent study found that mutations in the dynein gene are linked to alterations in the retrograde transport and degeneration of motor neurons (Hafezparast et al. 2003). However, it was not possible to detect incorporation of nitrotyrosine in the α-tubulin of motor neurons. Indeed, the motor neurons immunostained simultaneously for nitrotyrosine and α-tubulin showed a discordant cellular localization (Fig. 1). These results were further confirmed using an antibody against carboxy terminal nitrotyrosine-tubulin (Bisig et al. 2002), which failed to show immunoreactivity in motor neurons (not shown). Nitrotyrosine is incorporated into the α-tubulin of chick embryo neuronal cultures, indicating that neurons incorporate nitrotyrosine into α-tubulin without inducing cell death (C. G. Bisig, S. A. Purro and C. A. Arce, unpublished observations). Furthermore, even though its incorporation into proteins was not detected after intrastriatal injection, nitrotyrosine still induced degeneration of tyrosine hydroxylase-containing neurons (Mihm et al. 2000), further suggesting that there is not a relationship between the induction of neuronal death by nitrotyrosine and its incorporation into α-tubulin. In fact, incorporation of nitrotyrosine into the α-tubulin of astrocytes did not affect morphology, growth or survival (Table 1, Fig. 8). Detection of incorporation of nitrotyrosine into proteins by western blot did not provide reproducible results (not shown), but nitrotyrosine immunoreactivity was detected in protein cell extracts by dot blot analysis. These results indicate that nitrotyrosine is bound to proteins, but the characteristics of the interaction are unclear.

Explanations for the lack of nitrotyrosine incorporation into α-tubulin in motor neuron cells include the possible expression by motor neurons of a carboxypeptidase that recognizes and rapidly removes nitrotyrosine from α-tubulin (Bisig et al. 2002), the expression by motor neurons of a denitrase activity that preferentially targets nitrotyrosine in tubulin (Kamisaki et al. 1998; Balabanli et al. 1999; Irie et al. 2003) or the preferential targeting of nitrotyrosinated α-tubulin to proteolysis, a phenomenon observed for other tyrosine-nitrated proteins (Souza et al. 2000). The explanation positing in situ nitrotyrosine formation by nitrating species can be rejected because neither inhibition of nitric oxide production nor increased scavenging of superoxide and peroxynitrite prevented the increase in nitrotyrosine immunoreactivity.

Trophic factor deprivation does not seem to be the mechanism for nitrotyrosine toxicity because the addition of greater concentrations of multiple trophic factors delayed but did not prevent motor neuron death induced by free nitrotyrosine (Fig. 3). In addition, even though increased scavenging of superoxide and peroxynitrite and inhibition of nitric oxide synthase activity succeeded in preventing trophic factor deprivation-induced apoptosis (Estévez et al. 1998b, 2000), they did not affect the motor neuron apoptosis induced by nitrotyrosine (Fig. 4). These results suggest that apoptosis induced by nitrotyrosine is governed by a different pathway than that induced by trophic factor deprivation and characterized by nitric oxide production and peroxynitrite formation (Estévez et al. 1998a, 2000; Raoul et al. 2002).

Astrocyte–motor neuron interactions, such as spatial proximity (O'Really et al. 1995; Levine et al. 1999) and the glutamate uptake systems in astrocytes (Rothstein et al. 1992; Rothstein et al. 1995; Howland et al. 2002), are altered in human patients and in animal models of ALS. Astrocytes are the first cell type to show alterations in animal models of ALS (Wong et al. 1995; Bruijn et al. 1997b), suggesting that alterations in the interactions between astrocytes and motor neurons may play a role in the pathogenesis of ALS (Cassina et al. 2001, 2002). Spinal astrocytes in culture release trophic factors, which keep motor neurons alive for several days (Eagleson et al. 1985; Peluffo et al. 1997). However, under inflammatory conditions, astrocytes may produce death factors, signaling motor neurons to undergo apoptosis (Cassina et al. 2002). Incubation of growing spinal cord astrocytes with nitrotyrosine at concentrations well above clinically relevant levels had no effect on cell morphology (not shown), growth or survival (Table 1, Fig. 6). Surprisingly, nitrotyrosine was incorporated into the α-tubulin of discrete populations of astrocytes (Fig. 8), in contrast to its homogeneous incorporation into cytoskeletal α-tubulin of other cell types (Eiserich et al. 1999; Bisig et al. 2002). This effect might be explained by the presence of uniquely differentiated populations of astrocytes [e.g. tyrosine tubulin ligase (TTL) content] or because astrocytes are at different stages of cell or metabolic cycles (e.g. denitrase activity). This interpretation is further supported by the observation that only focal distributions of cells incorporated nitrotyrosine into tubulin in confluent monolayers (Fig. 8), whereas the proportion of cells incorporating nitrotyrosine was greater in growing cultures (not shown). As previously discussed for multiple trophic factors, culture of motor neurons on an astrocyte monolayer decreased but did not prevent nitrotyrosine toxicity (Fig. 5). The effect of astrocytes might be explained by trophic support of the motor neurons or the simple scavenging and sequestration of nitrotyrosine from the culture media. However, astrocyte monolayers did not significantly decrease the levels of nitrotyrosine in phenol red-free culture media as measured by the absorbance at 430 nm (not shown), suggesting that direct interactions between astrocytes and motor neurons may instead explain the decreased toxicity of nitrotyrosine.

Increased levels of free 3-nitrotyrosine have been reported in a large number of neurodegenerative diseases (Beal 2002) including ALS (Beal et al. 1997; Tohgi et al. 1999). Increased levels of free nitrotyrosine were also observed in ALS-SOD1 mutant mice (Bruijn et al. 1997a). We show herein that incubation of highly purified motor neuron cultures with clinically relevant concentrations of nitrotyrosine triggers motor neuron apoptosis. Moreover, nitrotyrosine was not found to be toxic to astrocytes, and the co-culture of motor neurons with astrocytes delayed, but did not prevent, nitrotyrosine-induced apoptosis. However, astrocytes incorporated nitrotyrosine into α-tubulin, suggesting that if there is free nitrotyrosine in the spinal cord of patients with ALS and mice transgenic for ALS mutant human SOD1, protein-linked nitrotyrosine must be present (Abe et al. 1995, 1997; Beal et al. 1997; Ferrante et al. 1997a, 1997b). The regional progression of ALS indicates that the increase in free nitrotyrosine concentrations in the cerebrospinal fluid might not be the direct cause of motor neuron death, but rather may increase the vulnerability of motor neurons to other toxic stimuli. Supporting this interpretation of the results is the recently reported ability of the antioxidant FeTCPP to provide protective effects in mice transgenic for human ALS mutant SOD against the progression of the disease without affecting the time of its onset (Wu et al. 2003).


We thank Brandy M. Britton, Nicholas M. Crovo, Daniel A. Demaree, Mariana Pehar and Patricia Matthews for their invaluable assistance with the experiments and in the preparation of this manuscript. We especially thank Francisco Schöpfer and Bruce A. Freeman for their critical insight during the preparation of this manuscript. The authors also thank Urufarma for its support. This work was supported by Programa para el Desarrollo des las Ciencias Básicas and grants from the National Institutes of Health NS36761 and NS42834 (AGE) and from the ALS Association (AGE).