Chronic hyperammonemia induces tonic activation of NMDA receptors in cerebellum

Authors


Address correspondence and reprint request to Vicente Felipo, Laboratory of Neurobiology, Centro de Investigacion Principe Felipe, Avda Autopista del Saler, 16, 46012 Valencia, Spain. E-mail: vfelipo@cipf.es

Abstract

J. Neurochem. (2009) 112, 1005–1014.

Abstract

Reduced function of the glutamate–nitric oxide (NO)–cGMP pathway is responsible for some cognitive alterations in rats with hyperammonemia and hepatic encephalopathy. Hyperammonemia impairs the pathway in cerebellum by increasing neuronal nitric oxide synthase (nNOS) phosphorylation in Ser847 by calcium-calmodulin-dependent protein kinase II (CaMKII), reducing nNOS activity, and by reducing nNOS amount in synaptic membranes, which reduces its activation following NMDA receptors activation. The reason for increased CaMKII activity in hyperammonemia remains unknown. We hypothesized that it would be as a result of increased tonic activation of NMDA receptors. The aims of this work were to assess: (i) whether tonic NMDA activation receptors is increased in cerebellum in chronic hyperammonemia in vivo; and (ii) whether this tonic activation is responsible for increased CaMKII activity and reduced activity of nNOS and of the glutamate–NO–cGMP pathway. Blocking NMDA receptors with MK-801 increases cGMP and NO metabolites in cerebellum in vivo and in slices from hyperammonemic rats. This is because of reduced phosphorylation and activity of CaMKII, leading to normalization of nNOS phosphorylation and activity. MK-801 also increases nNOS in synaptic membranes and reduces it in cytosol. This indicates that hyperammonemia increases tonic activation of NMDA receptors leading to reduced activity of nNOS and of the glutamate–NO–cGMP pathway.

Abbreviations used
CaMKII

calcium-calmodulin-dependent protein kinase II

HE

hepatic encephalopathy

MK-8O1

(5R,10S)-(+)-5-Methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5,10-imine hydrogen maleate

nNOS

neuronal nitric oxide synthase

NO

nitric oxide

PCS

portacaval shunted

TIF

triton-insoluble fraction

Patients with liver cirrhosis may develop hepatic encephalopathy (HE) which leads to a wide range of cognitive and motor alterations (Weissenborn et al. 2001, 2005; Amodio et al. 2004). Chronic moderate hyperammonemia is a main contributor to the neurological alterations in HE (Felipo and Butterworth 2002; Felipo 2008). The mechanisms by which liver failure and hyperammonemia lead to the neurological alterations are beginning to be clarified in animal models of chronic hyperammonemia and HE. One of the mechanisms involved in cognitive impairment in HE is the impairment of the function of the glutamate–nitric oxide (NO)–cGMP pathway associated to the NMDA type of glutamate receptors. Activation of these receptors leads to an increased intracellular calcium in the post-synaptic neuron. Calcium binds to calmodulin and activates neuronal nitric oxide synthase (nNOS), increasing the formation of NO, which activates soluble guanylate cyclase and the formation of cGMP. Activation of this glutamate–NO–cGMP pathway is necessary for some types of learning and memory, including learning of a conditional discrimination task in a Y maze (Bernabeu et al. 1996, 1997; Yamada et al. 1996; Piedrafita et al. 2007; Llansola et al. 2009).

The function of this pathway is impaired in cerebellum in vivo in portacaval shunted (PCS) rats, the most common model of chronic HE (Monfort et al. 2001). Reduced function of this pathway is responsible for the impairment of the ability of PCS rats to learn the Y maze task. The mechanisms leading to impaired function of the pathway (and subsequently of learning) are beginning to be clarified. A main contributor to this impairment is chronic hyperammonemia. The function of the glutamate–NO–cGMP pathway (Hermenegildo et al. 1998) and learning ability (Aguilar et al. 2000; Erceg et al. 2005b) are also impaired in an animal model of ‘pure’ hyperammonemia without liver failure. Treatment with phosphodiesterase inhibitors restores the function of the pathway and learning ability in PCS rats (Erceg et al. 2005a) and in rats with ‘pure’ chronic hyperammonemia (Erceg et al. 2005a,b).

Two main mechanisms by which hyperammonemia impairs the function of the pathway in cerebellum are by increasing phosphorylation of nNOS in Ser847, which reduces nNOS activity, and by reducing the content of nNOS in synaptic membranes, which reduces its activation in response to activation of NMDA receptors (El-Mlili et al. 2008). Increased phosphorylation of nNOS in Ser847 in hyperammonemia is because of increased activity of calcium-calmodulin-dependent protein kinase II (CaMKII), which results from increased phosphorylation in Thr286 (El-Mlili et al. 2008). The reason for this increased activity of CaMKII in hyperammonemia has not been clarified. CaMKII is activated by calcium-calmodulin and its activity is increased following the activation of NMDA receptors.

It has been shown that acute hyperammonemia, when ammonia levels are in the millimolar range, leads to the activation of NMDA receptors in brain in vivo (Hermenegildo et al. 2000; Hilgier et al. 2004), which is responsible for ammonia-induced death of animals (Marcaida et al. 1992; Hermenegildo et al. 1996).

In animal models of chronic hyperammonemia and HE, as well as in patients with chronic liver cirrhosis and HE, ammonia levels are chronically increased but only in the micromolar (100–200 μM) range (Ehrlich et al. 1980; Miñana et al. 1988; Cauli et al. 2007). The effects of chronic moderate hyperammonemia on NMDA receptor function have not been studied in vivo. Studies in cultured neurons indicate that these effects would depend on the concentration of ammonia, the duration of exposure, and the concentration of the exogenously added agonist (Marcaida et al. 1995; Hermenegildo et al. 1998; Llansola et al. 2003).

We hypothesized that chronic moderate hyperammonemia could also lead to increased tonic activation of NMDA receptors in vivo which would be responsible for increased activity of CaMKII and, subsequently, for reduced activation of nNOS and of the glutamate–NO–cGMP pathway, and for impaired learning ability.

The aims of this work were to assess: (i) whether tonic activation of NMDA receptors is increased in cerebellum in chronic hyperammonemia in vivo, leading to reduced nNOS activity and glutamate–NO–cGMP pathway function; and (ii) whether this tonic activation is responsible for increased activity of CaMKII and, subsequently for increased phosphorylation and reduced activity of nNOS and of the glutamate–NO–cGMP pathway.

To assess whether tonic activation of NMDA receptors is increased in cerebellum in chronic hyperammonemia in vivo, we used brain microdialysis in freely moving rats and assessed whether blocking NMDA receptors increases nNOS activity and glutamate–NO–cGMP pathway function by measuring extracellular nitrites + nitrates (stable metabolites of NO) and cGMP, the final product of the pathway. To analyze the molecular mechanisms by which increased tonic activation of NMDA receptors in hyperammonemia impairs the function of the glutamate–NO–cGMP pathway, we performed ex vivo experiments in cerebellar slices from control and hyperammonemic rats.

Materials and methods

Hyperammonemic rats without liver failure

Male Wistar rats (120–140 g) were made hyperammonemic by feeding them an ammonium-containing diet for 4 weeks as previously described (Felipo et al. 1988).

Determination of the effects of blocking NMDA receptors on extracellular cGMP and nitrites + nitrates in rat cerebellum by in vivo microdialysis

Rats were anesthetized using halotane and a microdialysis guide was implanted in the cerebellum. The coordinates were 10.2 mm posterior to bregma and 1.2 mm to the right of the bregma, and 1 mm below the duramater surface (Paxinos and Watson 1996), as previously described (Fedele and Raiteri 1996; Hermenegildo et al. 2000). The guide was secured to the skull with miniature screws and dental cement, and the skin was sutured. After 48 h, a microdialysis probe was implanted. Probes were perfused with artificial CSF at a flow rate of 3 μL/min. The composition of artificial CSF was (in mM): NaCl, 145; KCl, 3.0; CaCl2, 2.26; buffered at pH 7.4 with 2 mM phosphate buffer and filtered through 0.45 μm pore size Millipore filters. After a 2-h stabilization period, consecutive samples were collected every 30 min with a fraction collector. Five to six samples were collected initially to establish basal concentration of cGMP or nitrites + nitrates and then MK-801 0.5 μM was infused through the microdialysis probe during the five following fractions to block NMDA receptors and assess whether its tonic activation is increased after chronic hyperammonemia. Preliminary experiments showed that 0.5 and 1 μM (5R,10S)-(+)-5-Methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5,10-imine hydrogen maleate (MK-801) induce similar effects, so that we choose the lower dose (0.5 μM) for these studies. Samples were stored at −80°C until analysis of cGMP or nitrites + nitrates content.

Determination of the effects of blocking NMDA receptors on the concentration of cGMP and of nitrites + nitrates in cerebellar slices

Control or hyperammonemic rats were sacrificed by decapitation and their brains were transferred into ice-cold Krebs buffer (in mM): NaCl 119, KCl 2.5, KH2PO4 1, NaHCO3 26.2, CaCl2 2.5 and glucose 11, aerated with 95% O2 and 5% CO2 at pH 7.4. Cerebella were dissected and transversal slices (300 μm) were obtained using a Vibratome, transferred to incubation wells and incubated for 30 min at 35.5°C in Krebs buffer.

The slices were incubated for 10 min in the presence or the absence of MK-801 (1 μM, an antagonist of NMDA receptors), collected and homogenized in 200 μL of acetate buffer (sodium acetate 50 mM, pH 5.8). Samples were centrifuged (14 000 g, 5 min) and cGMP or nitrites + nitrates were measured in the supernatant as described below. Pellets were re-suspended in 200 μL of 0.25 M NaOH and protein was measured using the bicinchonic acid method.

Analysis of the phosphorylation of nNOS in Ser847 and of CaMK-II in Thr286

Cerebellar slices were prepared from control or hyperammonemic rats as described above. Slices were collected and homogenized in Tris–HCl 50 mM (pH 7.4), 0.1% sodium dodecylsulfate, 1 mM EGTA, 1 mM EDTA, 0.1% 2-mercaptoetanol, 1% Triton X-100, containing phosphatase inhibitors: 1 mM sodium orthovanadate, 1 mM sodium fluoride, 5 mM sodium pyrophosphate, 10 μM okadaic acid, and protease inhibitors (Complete Set; Roche Diagnostics, Mannheim, Germany). The homogenates were boiled for 5 min and protein was determined by the bicinchonic acid method. Samples were subjected to sodium dodecylsulfate–polyacrylamide gel electrophoresis and immunoblotting as previously described (Corbalán et al. 2002) using antibodies against nNOS (1 : 1000) from BD Transduction Laboratories (Lexington, KY, USA); nNOS phosphorylated at Ser 847 (1 : 400) from abCam (Cambridge, UK); CaMKII phosphorylated at Thr 286 (1 : 1000) from Bio reagents (Golden, CO, USA); or CaMKII (1 : 1000) from Chemicon International (Hampshire, UK). The images were captured using the SCANJET 5300C from HP (Spain) and band intensities quantified using the Intelligent Quantifier Alpha Imager 2200 program (AlphaInnotech Corporation, San Francisco, CA, USA).

Analysis of subcellular localization of nNOS

To analyze the localization of nNOS, subcellular fractions were prepared essentially as described by Picconi et al. (2004). As described by these authors Triton-insoluble fractions (TIF) were obtained instead of the classical post-synaptic densities preparation because the amount of the starting material was very limited. This TIF fraction contains essentially post-synaptic proteins and does not contain pre-synaptic markers (Picconi et al. 2004).

Cerebellar slices were prepared from control or hyperammonemic rats as described above. Slices were rapidly homogenized in cold 0.23 M sucrose containing (in mM): 1 HEPES, 1 MgCl2, 1 NaHCO3, and 0.1 phenylmethylsulfonyl fluoride, pH 7.4, in the presence of a Complete Set of Protease inhibitors (Complete; Roche Diagnostics, Mannheim, Germany) and of phosphatase inhibitors: fostriecin 20 nM, okadaic acid 10 μM, deltametrin 20 nM, sodium fluoride 1 mM, sodium orthovanadate 1 mM, imidazole 2 mM. The homogenized tissue was centrifuged at 1000 g for 10 min at 4°C. The resulting supernatant was centrifuged at 3000 g for 15 min at 4°C. The supernatant contained the cytosol fraction and the pellet contained mitochondria and synaptosomes. The pellet was re-suspended in 100 μL of 75 mM KCl containing 1% Triton X-100 and was centrifuged at 100 000 g for 1 h at 4°C. The final pellet is the TIF fraction and was homogenized in 50 μL of 20 mM HEPES with phosphatase and protease inhibitors as above. nNOS was quantified in each fraction by immunoblotting as described above and was found in cytosol and TIF fractions, with negligible amounts in the other fractions.

Determination of nitrites and nitrates

Nitrites and nitrates were measured in microdialysis samples and in cerebellar slices by the Griess method (Verdon et al. 1995) using nitrate reductase. Fifty μL of slice supernatant or 90 μL of microdialysis samples were mixed with equal volumes of Griess reagent. After 10 min at 23°C, absorbance was measured at 540 nm.

Determination of cGMP

cGMP was measured in microdialysis samples and in cerebellar slices using the BIOTRAK cGMP enzyme immunoassay kit from Amersham (Amersham Pharmacia Biotech, Buckinghamshire, UK).

Statistical analyses

The data were analyzed by analysis of variance (anova) followed when appropriate by Newman-Keul’s post hoc test. When only two values were compared, the Student’s t-test was used. Significance levels were set at alpha = 0.05.

Results

Blocking NMDA receptors increases extracellular cGMP in cerebellum in vivo strongly in rats with chronic hyperammonemia but only very mildly in control rats

To assess whether reduced function of the glutamate–NO–cGMP pathway and extracellular cGMP in cerebellum in hyperammonemia are resulted from the increased basal tonic activation of NMDA receptors, we tested by in vivo brain microdialysis whether blocking NMDA receptors with MK-801 increases the formation of cGMP in cerebellum in vivo of hyperammonemic rats.

Basal levels of extracellular cGMP in cerebellum were reduced by 44% in hyperammonemic rats compared with control rats, in agreement with Erceg et al. (2005a).

We administered MK-801, an antagonist of NMDA receptors, through the microdialysis probe to control and hyperammonemic rats. As shown in Fig. 1, MK-801 increased (< 0.01) extracellular cGMP in hyperammonemic rats to 421 ± 71% of basal concentration. However, in control rats, MK-801 affects extracellular cGMP only very mildly, reaching only 141 ± 15% of basal. This indicates that tonic activation of NMDA receptors is high in hyperammonemic rats, but very mild in control rats.

Figure 1.

 Blocking NMDA receptors with MK-801 increases extracellular cGMP in cerebellum in vivo in hyperammonemic but not in control rats. Microdialysis probes were inserted in the cerebellum of control or hyperammonemic rats, perfused at 3 μL/min and samples were taken every 30 min. After taking five samples to determine basal levels of cGMP, MK-801 (0.5 μM) was administered in the perfusion stream as indicated by the horizontal bar. Values are given as percentage of basal and are the mean ± SEM from seven control and five hyperammonemic rats. Values significantly different (< 0.05) from basal concentration of cGMP are indicated by ‘a’. Values significantly different from control rats are indicated by asterisks. **< 0.01.

Blocking NMDA receptors increases extracellular nitrites + nitrates in cerebellum in vivo in rats with chronic hyperammonemia but not in control rats

To assess whether increased ‘tonic activation’ of NMDA receptors contributes to reduced nNOS activity in cerebellum in vivo in hyperammonemia, we assessed whether blocking NMDA receptors with MK-801 increases nNOS activity by measuring nitrites and nitrates in the extracellular fluid. MK-801 significantly (< 0.05) increased extracellular nitrites + nitrates in hyperammonemic rats, reaching 127 ± 7% of basal in fraction 8 and returning to basal levels thereafter. However, in control rats, MK-801 did not affect extracellular nitrites and nitrates, which remained at 111 ± 4% of basal (Fig. 2). This indicates that tonic activation of NMDA receptors in hyperammonemia reduces nNOS activity in cerebellum in vivo.

Figure 2.

 Blocking NMDA receptors with MK-801 increases extracellular nitrites and nitrates in cerebellum in vivo in hyperammonemic but not in control rats. Microdialysis probes were inserted in the cerebellum of control or hyperammonemic rats, perfused at 3 μL/min and samples were taken every 30 min. After taking six samples to determine basal levels of nitrites and nitrates, MK-801 (0.5 μM) was administered in the perfusion stream as indicated by the horizontal bar. Values are given as percentage of basal and are the mean ± SEM from seven rats per group. Values significantly different (< 0.05) from basal concentration of nitrites and nitrates are indicated by asterisks.

Blocking NMDA receptors increases cGMP and nitrites + nitrates in cerebellar slices from rats with chronic hyperammonemia but not in control rats

To analyze the mechanism involved in the effects of MK-801 on the glutamate–NO–cGMP pathway and on nNOS activity, we performed ex vivo experiments in cerebellar slices from control or hyperammonemic rats. We first tested whether this system reproduces the effects of MK-801 observed in vivo.

As shown in Fig. 3(a), basal levels of cGMP were significantly (< 0.05) reduced in slices from hyperammonemic rats to 73 ± 6% of controls, reproducing the decrease in extracellular cGMP found in cerebellum in vivo. Treatment of the slices with MK-801 did not affect cGMP levels in slices from control rats (102 ± 9% of basal) but completely normalized cGMP levels in slices from hyperammonemic rats, returning to 116 ± 15% of controls. This reproduces the increase in cGMP induced by MK-801 in cerebellum in vivo.

Figure 3.

 Blocking NMDA receptors restores the levels of cGMP and of nitrites + nitrates in cerebellar slices from hyperammonemic rats to normal levels. Cerebellar slices were freshly prepared from control or hyperammonemic rats and incubated in the absence or the presence of MK-801 (1 μM) as described in Materials and methods. The levels of cGMP (a) and of nitrites + nitrates (b) under basal conditions and 10 min after addition of MK-801 are shown. Values are the mean ± SEM from seven rats per group. Values significantly different (< 0.05) from basal are indicated by ‘a’. Values significantly different from control rats are indicated by asterisks, *< 0.05.

As shown in Fig. 3(b), basal levels of nitrites + nitrates were significantly (< 0.05) reduced in slices from hyperammonemic rats to 71 ± 8% of controls. Treatment of the slices with MK-801 did not affect nitrites + nitrates levels in slices from control rats (88 ± 11% of basal) but completely normalized nitrites + nitrates levels in slices from hyperammonemic rats, returning to 98 ± 12% of controls. This also reproduces the increase in nitrites + nitrates induced by MK-801 in cerebellum in vivo.

Blocking NMDA receptors reduces phosphorylation of nNOS in Ser847 and of calcium calmodulin-dependent kinase II in Thr286 in cerebellar slices from rats with chronic hyperammonemia but not in control rats

We assessed whether MK-801-induced restoration of NOS activity is because of normalization of nNOS phosphorylation in Ser847. As shown in Fig. 4(a) and (b), phosphorylation of nNOS in Ser847 is increased (< 0.05) by 31 ± 6% in slices from hyperammonemic rats. Treatment with MK-801 did not affect phosphorylation of Ser847 in slices from control rats (104 ± 8% of basal) but completely normalized phosphorylation of nNOS in Ser847 in hyperammonemic rats, reducing it to 90 ± 6% of controls.

Figure 4.

 Blocking NMDA receptors reduces phosphorylation of nNOS at Ser847 and of CaMKII in Thr286 in cerebellar slices from hyperammonemic rats to normal levels. Cerebellar slices were freshly prepared from control or hyperammonemic rats and incubated in the absence or the presence of MK-801 (1 μM) as described in Materials and methods. Slices were homogenized and the homogenates (50 μg of protein) were subjected to electrophoresis and phosphorylation of nNOS at Ser-847 (a and b) or of CaMKII at Thr286 (c and d) were analyzed by immunoblotting. A representative immunoblotting is shown in (a). The intensities of the bands were quantified and expressed as percentage of controls (b). Values are the mean ± SEM of 8 rats per group for nNOS and six rats per group for CaMKII. Values significantly different from controls are indicated by asterisks, *< 0.05. Values significantly different (< 0.05) from basal are indicated by ‘a’. C, control rats; HA, hyperammonemic rats.

We then assessed whether MK-801-induced normalization of nNOS phosphorylation in Ser847 is because of reduced phosphorylation (and activity) of CaMKII in Thr286. As shown in Fig. 4(c) and (d), phosphorylation of CaMKII in Thr286 is increased (< 0.05) by 30 ± 9% in slices from hyperammonemic rats. Treatment with MK-801 did not affect phosphorylation of CaMKII in Thr286 in slices from control rats (102 ± 7% of basal) but completely normalized phosphorylation of nNOS in Ser847 in hyperammonemic rats, reducing it to 99 ± 11% of controls.

These data show that tonic activation of NMDA receptors is increased in slices from hyperammonemic rats, leading to activation of CaMKII which phosphorylates nNOS in Ser847, reducing its activity and subsequently, formation of cGMP. Blocking NMDA receptors normalizes all the steps of the pathway in slices from hyperammonemic rats.

Blocking NMDA receptors increases the amount of nNOS in synaptic membranes in cerebellar slices from rats with chronic hyperammonemia but reduces it in control rats

Chronic hyperammonemia also alters the subcellular localization of nNOS decreasing it in synaptic membranes and increasing it in the cytosol (El-Mlili et al. 2008). This results in reduced activation of nNOS in response to the activation of NMDA receptors. To assess whether this effect is also because of increased tonic activation of NMDA receptors, we analyzed the effect blocking these receptors with MK-801on the subcellular distribution of nNOS. As shown in Fig. 5, the content of nNOS is reduced in synaptic membranes (73 ± 6% of control, < 0.05) and increased in the cytosol (176 ± 19% of control, < 0.01) from slices of hyperammonemic rats. Blocking NMDA receptors with MK-801 in slices from hyperammonemic rats increased nNOS content in synaptic membranes to the same level of control rats (101 ± 8%) and decreased it in cytosol to 124 ± 9% of basal control (non-significantly different from controls, = 0.1). In control rats, MK-801 decreased the content of nNOS in synaptic membranes (81 ± 7% of basal, < 0.01) and increased it in cytosol (150 ± 7% of basal, < 0.01).

Figure 5.

 Blocking NMDA receptors normalizes the amount of nNOS in post-synaptic membranes and in the cytosol in cerebellar slices from hyperammonemic rats. Cerebellar slices were freshly prepared from control or hyperammonemic rats and incubated in the absence or the presence of MK-801 (1 μM). Subcellular fractions were prepared as described in Methods and were subjected to electrophoresis to analyze the content of nNOS by immunoblotting. nNOS was found only in the post-synaptic (a) and cytosolic (b) fractions. Representative immunoblottings are shown in (a) for post-synaptic and (b) for cytosolic fractions. The intensities of the bands were quantified and expressed as percentage of controls. Values are the mean ± SEM of six rats per group. Values significantly different from controls are indicated by asterisks, *< 0.05. Values significantly different (< 0.05) from basal are indicated by ‘a’. C, control rats; HA, hyperammonemic rats.

Discussion

In cerebellar slices from rats with chronic hyperammonemia the function of the glutamate–NO–cGMP pathway is reduced, resulting in lower levels of nitrites + nitrates and of cGMP than that in slices from control rats. This is a consequence of reduced nNOS activity, as a result of increased phosphorylation of nNOS in Ser847 by calcium calmodulin-dependent kinase II, which activity is increased in hyperammonemia because of increased phosphorylation in Thr286 (El-Mlili et al. 2008). Basal levels of extracellular cGMP are also reduced in cerebellum of hyperammonemic rats in vivo (this work and Erceg et al. 2005a).

The mechanism by which chronic hyperammonemia increases the activity of CaMKII and, subsequently, reduces nNOS activity and the function of the glutamate–NO–cGMP pathway has not been studied previously. It has been reported that activation of NMDA receptors increases phosphorylation of CaMKII in Thr286 and its activity (Fukunaga et al. 1996; Strack and Colbran 1998; Cammarota et al. 2002).

We therefore hypothesized that increased activity of CaMKII in hyperammonemia could be resulted from increased tonic activation of NMDA receptors. If this were the case, the activity of the kinase would be reduced by blocking NMDA receptors with MK-801.

We show now that blocking NMDA receptors in slices from hyperammonemic rats with MK-801 normalizes the phosphorylation of CaMKII in Thr286 and the activity of this kinase, returning to the same values found in slices from control rats. Normalization of the CaMKII activity leads subsequently to normalization of nNOS phosphorylation and activity and of the levels of nitrites + nitrates and of cGMP.

These data show that tonic activation of NMDA receptors is increased in chronic hyperammonemia. Moreover, increased activation of NMDA receptors is responsible for the increased activity of CaMKII. Blocking NMDA receptors normalizes the phosphorylation of nNOS in Ser847, the activity of nNOS, and the levels of NO as reflected by nitrites and nitrates and of cGMP. This indicates that increased activation of NMDA receptors is the main responsible for the reduced activity of nNOS in chronic hyperammonemia and, subsequently, of reduced activation of guanylate cyclase and decreased levels of cGMP.

In contrast, blocking NMDA receptors in control slices did not affect the function of the glutamate–NO–cGMP. The levels of cGMP and of nitrites + nitrates as well as the phosphorylation of nNOS in Ser 847 and of CaMKII in Thr286 remain unaltered in slices from control rats.

This indicates that tonic activation of NMDA receptors is mild in control rats and is strongly increased in chronic hyperammonemia.

Increased tonic activation of NMDA receptors reduces the function of the glutamate–NO–cGMP pathway in hyperammonemic rats at least by two different mechanisms: (i) reduced activity of nNOS because of increased phosphorylation in Ser847, and (ii) retention of nNOS in the cytosol and reduced amount of nNOS in the synaptic membrane.

As mentioned above, the increased phosphorylation of nNOS in Ser847 is because of increased activity of CaMKII and both increases are abolished by blocking NMDA receptors, supporting that increased tonic activation of NMDA receptors is responsible for increased activity of CaMKII and, subsequently, for the reduced activity of nNOS and function of the glutamate–NO–cGMP.

The mechanisms modulating the transport of nNOS to the synaptic membrane are not well known. Ohnishi et al. (2008) showed, in pheochromocytoma cells, that acute activation of NMDA receptors by high (but not by low) concentrations of NMDA increases membrane expression of nNOS. In the same cells, low concentrations of NMDA added together with pituitary adenylate cyclase activating polypeptide increased nNOS in the membrane by a different mechanism involving both protein kinases A and C. Blocking NMDA receptors prevents the effects of pituitary adenylate cyclase activating polypeptide. This suggests that, under normal conditions, acute moderate activation of NMDA receptors would induce a translocation of nNOS to the synaptic membrane and that this process is modulated by other neurotransmitters through different signaling mechanisms. Low tonic activation of NMDA receptors would be enough to cooperate with other neurotransmitters in maintaining a certain amount of nNOS in the membrane.

Taking into account these results of Ohnishi et al. (2008), we could interpret the effects of MK-801 on nNOS amount in cytosol/membrane in slices from control rats as a reduction in the transport of nNOS from cytosol to the membrane owing to low tonic activation of NMDA receptors in combination with the action of other neurotransmitters. Blocking NMDA receptors with MK-801 would prevent this transport, increasing nNOS in the cytosol, and reducing it in synaptic membranes.

MK-801 induces the opposite effect in hyperammonemic rats, it increases the amount of nNOS in synaptic membranes, returning it to normal levels. The mechanisms involved are not clear by now. We believe that one possibility is that in chronic hyperammonemia, increased tonic activation of NMDA receptors would lead to induction of an adaptive response to prevent excessive amount of nNOS in the synaptic membrane. This adaptive response would involve induction of an NMDA receptor-dependent signaling pathway which would retain nNOS in the cytosol. Blocking NMDA receptors with MK-801 would inhibit this adaptive pathway and restore the normal transport of nNOS to the synaptic membrane, increasing its amount and restoring normal levels.

At first sight, it could be surprising that increased activation of NMDA receptors result in reduced function of the pathway glutamate–NO–cGMP in chronic hyperammonemia. Under physiological conditions or in acute pathological situations, activation of NMDA receptors leads to increased function of the pathway. In fact, excessive activation of the glutamate–NO–cGMP pathway mediates the neurotoxic effects of excessive activation of NMDA receptors in some acute pathological situations.

Excessive activation of NMDA receptors is neurotoxic (Novelli et al. 1987; Choi 1987; Miñana et al. 1996), playing a main role in the neuronal damage in cerebral ischemia (Bennett et al. 1990) and in some neurodegenerative diseases (Dong et al. 2009).

Activation of the glutamate–NO–cGMP pathway mediates excitotoxic neuronal death as a result of excessive activation of NMDA receptors. Excitotoxic neuronal death requires increased intracellular Ca2+ (Manev et al. 1989; Choi 1987), activation of NOS (Dawson et al. 1991, 1993; Cazevieille et al. 1993; Lafon-Cazal et al. 1993), and activation of soluble guanylate cyclase (Montoliu et al. 1999). The above studies show that blocking any of the steps of the glutamate–NO–cGMP pathway prevents excitotoxic neuronal death.

Acute hyperammonemia induced by injection of large doses of ammonia leads to excessive activation of NMDA receptors and of the glutamate–NO–cGMP pathway (Hermenegildo et al. 2000; Hilgier et al. 2004) which is responsible for ammonia-induced death of animals (Marcaida et al. 1992; Hermenegildo et al. 1996).

However, it has been shown that many processes and parameters are affected in different (sometimes opposite) ways by acute and chronic hyperammonemia (Monfort et al. 2000; Cagnon and Braissant 2007). We have previously shown that the function of the glutamate–NO–cGMP pathway (the increase in cGMP induced by activation of NMDA receptors) is reduced in cerebellum of rats with chronic hyperammonemia in vivo (Hermenegildo et al. 1998) and in cerebellar slices from hyperammonemic rats (El-Mlili et al. 2008). This indicates that in chronic hyperammonemia there is an adaptive response to reduce the function of the glutamate–NO–cGMP pathway and prevent the toxic effects of its excessive activation. As commented above, we have previously shown that reduced activity and activation of nNOS in response to activation of NMDA receptors contributes to the reduced function of the pathway (El-Mlili et al. 2008). However, the underlying mechanism remained unclear and it was not clear whether activation of NMDA receptors is enhanced or reduced in chronic hyperammonemia.

We show now that activation of NMDA receptors is also tonically increased in a model of chronic moderate hyperammonemia in rats which reproduces the levels of hyperammonemia present in patients with liver cirrhosis and HE (Miñana et al. 1988).

The reduced function of the glutamate–NO–cGMP pathway in chronic hyperammonemia, because of increased tonic activation of NMDA receptors could be considered as an adaptive response to prevent the neuronal damage induced by excessive activation of NMDA receptors. In the above acute situations using large doses of ammonia NMDA receptors and the glutamate–NO–cGMP pathway are rapidly and strongly activated leading to neuronal damage and death. However, in chronic injuries such as in chronic moderate hyperammonemia, the neurons have time to adapt to the injury by reducing the function of the glutamate–NO–cGMP pathway and this prevents neuronal death.

However, this adaptive response has also detrimental effects. Under physiological conditions, activation of this glutamate–NO–cGMP pathway modulates important cerebral processes including long-term potentiation (Son et al. 1998; Monfort et al. 2002) and some forms of learning and memory. Activation of NMDA receptors, NOS, and soluble guanylate cyclase and formation of cGMP are required for some types of learning and memory (Collingridge 1987; Danysz et al. 1995; Bernabeu et al. 1996, 1997; Yamada et al. 1996; Piedrafita et al. 2007; Llansola et al. 2009).

The adaptive response leading to reduced function of the glutamate–NO–cGMP pathway in chronic hyperammonemia, although prevents excitotoxic neuronal death or damage, would also result in impairment of the physiological functions which are modulated by this pathway, including long-term potentiation (Muñoz et al. 2000) and learning ability (Aguilar et al. 2000).

In addition to the glutamate–NO–cGMP pathway, NMDA receptors modulate other signal transduction pathways, including activation of mitogen-activated protein kinases, calcineurin, phospholipases A and C, cAMP response element binding, etc. (Ding et al. 1997; Llansola et al. 2001; Rodrigo et al. 2009). These pathways modulate important cerebral processes including different types of learning or the biological clock. Increased tonic activation of NMDA receptors in hyperammonemia would also lead to altered function and/or modulation of these pathways which could contribute to some of the neurological alterations present in chronic hyperammonemia and HE.

In summary, the results reported show that chronic hyperammonemia results in increased tonic activation of NMDA receptors in cerebellum. This increased activation of NMDA receptors would alter many different signal transduction pathways associated to NMDA receptors and the cerebral processes modulated by these pathways, thus contributing to the neurological alterations in chronic hyperammonemia and HE. These results also suggest that reducing activation of NMDA receptors would have beneficial effects in the treatment of cerebral alterations in chronic hyperammonemia and HE. One of the consequences of increased tonic activation of NMDA receptors in chornic hyperammonemia is a reduction in the function of the glutamate–NO–cGMP pathway owing to two main mechanisms: (i) increased activity of CaMKII which phosphorylates nNOS in Ser847, reducing its activity; and (ii) reduced amount of nNOS in the synaptic membrane, leading to reduced activation of nNOS in response to activation of NMDA receptors.

Acknowledgements

This work was supported by grants from Ministerio de Ciencia e Innovacion (SAF2005-06089, SAF2008-00062, CSD2008-00005) and from Consellería de Educación (ACOMP/2009/191; ACOMP-2009-025; PROMETEO/2009/027), and AP-092/09, AP-024/08 and A-01/08 from Conselleria de Sanitat of Generalitat Valenciana.

Ancillary