FEBS Letters 580 (2006) 2123–2128

Hepatic guanylate cyclase activity is decreased in a model of cirrhosis:A quantitative cytochemistry study

Nathan A. Daviesa, Stephen J. Hodgesa, Andrew A. Pitsillidesb, Rajeshwar P. Mookerjeea,Rajiv Jalana,*, Siroos Mehdizadehc

a The UCL Institute of Hepatology, Division of Medicine, University College London, 69–75 Chenies Mews, London, UKb Department of Veterinary Basic Sciences, Royal Veterinary College, London, UK

c Department of Clinical Chemistry, Imperial College School of Medicine, Charing Cross Campus, London, UK

Received 18 January 2006; accepted 23 February 2006

Available online 15 March 2006

Edited by Barry Halliwell

Abstract The production of nitric oxide (NO) in liver diseaseand its role in vascular control has been a subject of much inter-est in recent years. However, the activity of guanylate cyclase(GC), the enzyme activated by NO has received little attentionwith regard to liver disease. In this study we have utilised a quan-titative cytochemical technique to examine the activity of GC ona per cell basis in a rat model of cirrhosis. Our results show asignificant reduction in GC activity, indicating that vascular reg-ulation is likely to be substantially affected irrespective of NOgeneration in this disease model.� 2006 Federation of European Biochemical Societies. Publishedby Elsevier B.V. All rights reserved.

Keywords: Guanylate cyclase; Nitric oxide; Cirrhosis;Quantitative cytochemistry

1. Introduction

The role of nitric oxide (NO) as a vascular mediator is well

described; NO stimulates guanylate cyclase (GC) activity in

smooth muscle cells to produce cyclic guanosine monophos-

phate (cGMP), facilitating vascular relaxation [1,2]. Evidence

suggests that NO production is decreased in the cirrhotic liver

and that this contributes to a reduction in blood flow [3]. How-

ever, there is scant information addressing the role of GC as a

modulator of local blood flow in cirrhosis and indeed its regu-

latory role in normal liver function. Herein, we use the estab-

lished bile duct ligation (BDL) model [4] to examine whether

there is modulation of hepatic GC activity in cirrhosis. We sug-

gest that, in addition to NO production, the regulation of GC

activity is pivotal to the control of blood flow within the liver,

and that factors restricting GC activity would to be detrimen-

tal to hepatic function.

To investigate the NO-stimulated activity of GC at the cellu-

lar level we have utilised a quantitative cytochemical (QC)

technique [5–8]. This is an extremely sensitive method, which

allows examination of GC activity in individual cells within tis-

sues with complex architecture. In this method, tissues are

cryo-preserved, sectioned and then immediately reacted on

the microscope slide. The chosen enzyme reaction is optimised

*Corresponding author. Fax: +44 0 2073800405.E-mail address: r.jalan@ucl.ac.uk (R. Jalan).

0014-5793/$32.00 � 2006 Federation of European Biochemical Societies. Pu

doi:10.1016/j.febslet.2006.02.080

to produce a precipitated chromophore; the resultant optical

density is stoichiometrically related to the specific cellular

enzymatic activity. This methodology allows regional exami-

nation of tissues and the quantification of activity in individual

cell types [9,10].

The aim of this study was to utilise a QC method to evaluate

GC activity to determine the ability of the liver to respond to

hepatic production of NO. The specificity and sensitivity of

this technique provides the opportunity to investigate whether,

and to what extent, cGMP production may be altered in cir-

rhosis.

2. Materials and methods

2.1. Animal studiesAll studies were conducted according to Home Office guidelines un-

der the Animals in Scientific Procedures Act 1986. Male Sprague–Dawley rats (230–280 g, Charles Rivers Ltd.) were given free accessto normal rodent chow and water, with a light dark cycle of 12 h,19–23 �C, and humidity of approximately 50%.

BDL was achieved by making a midline abdominal incision underanaesthesia [4]. In the BDL group, the common bile duct was isolated,triply ligated with 3–0 silk, and sectioned between the ligatures. Thesham-operated (control) group were anesthetised, operated upon, butdid not have their bile duct ligated. After BDL, all animals continuedto gain weight and were comparable with sham-operated controls. Theoverall mortality in both groups was less than 10% and occurred within36 h of the operation.

The animals were killed by exsanguinations under anaesthesia, withblood and liver tissue collected for analysis. All studies were performedon liver tissue at 25–28 days post BDL or sham procedure, and 6 ani-mals were used in the final analysis in each group.

2.2. Tissue NOS activity determinationTissue for this assay was flash-frozen in liquid nitrogen and stored at�80 �C until use. The method used was a variation on the widely re-ported 3H-arginine to 3H-citrulline assay [1,2], modified for liver tissue.Briefly, liver tissue was homogenised in ice-cold HEPES buffer(20 mM, pH 7.4) containing EDTA (1 mM); sucrose (250 mM); valine(60 mM); protease inhibitor cocktail (1 ml/20 g tissue); and PMSF(3.4 mg/ml). This was then centrifuged (1000 · g, 10 min, 4 �C) andthe supernatant retained and measured for protein content using theBiuret method [11]. For NOS activity, 5 ll of the supernatant was incu-bated with a reaction medium (40 ll) containing: Tris–HCl buffer(30 mM, pH 7.4); NADPH (1.25 mM); 3H-arginine (10 lCi/ml, Amer-sham, UK); norvaline (5 mM); and either CaCl2 (400 lM) or EGTA(600 lM). The mixture was incubated at 30 �C for 30 min before stop-ping with ice-cold citrate buffer (1 ml, 50 mM, pH 5) containing EDTA(1 mM). The arginine:citrulline ratio was determined after thin layerchromatography separation on silica plates (Kieselgel 60, Merck,Darmstadt, Germany). Non-tritiated amino acids were added to aid

blished by Elsevier B.V. All rights reserved.

2124 N.A. Davies et al. / FEBS Letters 580 (2006) 2123–2128

the detection of spots and the components separated using a runningmixture of CHCl3:MeOH:NH4OH:H2O (ratio of 10:45:20:10), afterdrying the plates were developed using ninhydrin (0.5% in butanol).The individual spots were collected and the scintillation activity mea-sured (Ultima Gold scintillant, Perkin–Elmer, Boston, MA. Using aTri-Carb 2100TR, liquid scintillation analyser, Packard Biosciences,Berks, UK). Activity is expressed as lmol citrulline/mg protein/h.

2.3. Cytochemical assessment of GC activityFor QC analysis, small fresh liver segments (0.5 cm3) were coated in

a 5% aqueous solution of polyvinyl alcohol (PVA, Grade G04/140,Wacker chemicals Ltd., UK) and chilled to �70 �C by immersion inn-hexane (HPLC grade. Rathburn, Peebleshire, Scotland) in an outerbath of solid carbon dioxide in ethanol. The samples were then storedat �80 �C for less than 28 days before analysis [5–7].

14 lm tissue sections were cut in a cryostat (Bright’s InstrumentsHuntingdon, UK), using a microtome fitted with an automatic driveto ensure constant section thickness. The cabinet temperature wasmaintained at �25 �C, and the cutting blade further cooled by packingthe haft in solid CO2. Sections were flash dried onto uncoated glassslides before reaction for enzyme activity.

GC activity was disclosed according to the method of Mehdizadehet al. [6]. Briefly, fresh unfixed cryostat sections were reacted in med-ium containing: guanosine triphosphate (GTP, 10 mM); sodium azide(2.5 mM); sodium acetate (0.2 mM); manganese chloride (20 mM); so-dium fluoride (10 mM); LL-p-bromotetramisole (0.4 mM); in Tris buffer(pH 7.0, 0.2 M) containing 35% polypep 5115. Tissue sections were re-acted for 30 min at 37 �C before stopping by immersion in ice-coldpolypep 5115 (2%). Under these conditions, GC enzyme activity resultsin the generation of pyrophosphate which was trapped by the additionof freshly prepared lead ammonium citrate/acetate solution at a finalconcentration of 32 mg/ml [7]. The resultant lead phosphate precipitatewas subsequently disclosed by immersion in ammonium polysulfide(0.5%) for 1 min, to produce a coloured (brown) reaction product, be-fore repeated washing in distilled water then allowed to air dry. Sec-tions were mounted in absolute glycerine for measurement.

To measure the maximum extent to which endogenous hepatic NOSactivity contributes to the level of GC activity disclosed, serial sectionswere incubated in ‘full’ reaction medium (described above) that wasfurther supplemented with NO synthase substrates, namely: 1 mMNADPH, 5 mM arginine, and 3 mM norvaline (to inhibit endogenousarginase activity). The increased pyrophosphate production above thatevident in serial sections reacted for GC activity alone, is therefore ameasure of the level of GC activity attributable to endogenous NOSactivity.

Control studies included reacting sections incubated in medium sup-plemented with the NOS inhibitor LL-NAME [12] to demonstrate thatthe action of arginine and NADPH occurred via NOS stimulation.Complimentary use of the NO donor, 50 lM DEANO((CH3CH2)2N(NONO)Na) [13], in place of NOS substrates, was usedto evaluate the action of direct, non-enzymatic NO donation uponoverall GC activity in additional serial sections.

A known inhibitor of GC activity, ODQ [14], was also added to thereaction medium (1 mM), in both the presence and absence of GC acti-vators. This was included to demonstrate the specificity of the reactionfor GC activity and as a means of determining whether backgroundlevels of GC-independent phosphate generation were modulated byBDL treatment.

Unless otherwise stated, all chemicals and reagents were obtainedfrom the Sigma Chemical Company (Poole, UK).

GC activity was measured in individual histologically defined hepa-tocytes using a Vickers M85A scanning and integrating microdensi-tometer (Biorad Instruments, York, UK), fitted with a 40· 0.6 NAobjective, using spot size 1, a 20 lm mask and k = 585 nm [5,6]. Activ-ity was expressed as units of mean integrated extinction/cell multipliedby 100 (MIE · 100). Measurements of GC activity were made in 15individual cells in each of duplicate sections of tissue segments col-lected from the left medial lobe of the liver.

Fig. 1. NOS activity in liver tissue homogenates from BDL andcontrol animals as determined by the conversion of 3H-arginine to 3H-citrulline. Data are expressed as mean (±S.E.M.) scintillation counts/h/mg of protein, determined over a 30-min period (n = 6 per group). Nosignificant difference was found between these two groups.

2.4. Statistical analysisData are shown as mean ± standard error of the mean (S.E.M.).

Two-tailed unpaired t tests were used to define differences betweenmeans of normally distributed data of equal variance, using a commer-cially available package (GraphPad Prism, version 4.0; GraphPad

Software, Inc., SanDiego, CA, USA). For statistical evaluation of datathat was not normally distributed, a Mann–Whitney test was used.Dose concentration response curves were calculated using non-linearregression. Results were considered significant if P < 0.05.

3. Results

3.1. Animal studies

Animals in both groups continued to gain weight at similar

rates following surgery. After 28 days the BDL-treated animals

(305–345 g) were only slightly heavier than the control group

(270–355 g) and this difference, although not statistically sig-

nificant, was attributed to the presence of ascites.

3.2. Tissue NOS activity determination

The induction of liver cirrhosis was confirmed through the

appearance of steatosis [15] and large increases in connective

tissue that were apparent in the BDL group. An analysis of tis-

sue NOS activity using 3H-arginine, demonstrated that the

induction of cirrhosis following BDL fails to produce any sig-

nificant changes in the level of NOS activity in the soluble frac-

tion extracted from the tissue homogenates (Fig. 1). Though,

after centrifugation of the tissue homogenates, the protein con-

centration of the supernatants was found to be comparable in

the two groups (data not shown). 3H-Ornithine production

was also observed in both control and BDL treated groups,

indicating that despite the presence of norvaline some arginase

activity remained. 3H-Arginine was still present at the end of

the study period in all samples examined, demonstrating that

NOS activity was not limited by a lack of substrate availability

in this assay.

3.3. Cytochemical assessment of GC activity

Measuring GC activity by QC provides an alternative strat-

egy to disclose changes in the NOS-GC system and provides an

activity measurement on a per cell basis. This may be particu-

larly pertinent when measuring differences between normal

and diseased tissues in which pathological changes are associ-

ated with marked changes in histology and tissue architecture,

such as those addressed herein. It is apparent that the fibrotic

areas, evident in the H&E stained sections, showed little or no

GC activity (Fig. 2), and that although the general level of

activity appears lower in the livers of the BDL group, individ-

Fig. 2. Panels A and B show haematoxylin and eosin stained sections (16·) of liver tissue taken from a control and BDL animal, respectively. Insection B areas of fibrosis can be clearly seen. Panels C (control) and D (BDL) show corresponding liver sections following a quantitativecytochemistry study for GC activity, in which the brown colouration indicates the presence of phosphate.

N.A. Davies et al. / FEBS Letters 580 (2006) 2123–2128 2125

ual cells within these sections demonstrated high levels of

activity.

Microdensitometric assessment of GC activity in both the

sham and BDL groups disclosed a significant increase in mea-

surable GC activity/cell following the addition of NOS sub-

strates (P < 0.001 in each case, Fig. 3). This indicates that

endogenous NOS is capable of promoting increases in GC

activity in liver cells in both control and BDL rats. These mea-

surements also show that BDL treatment significantly reduces

both basal and NOS-activated GC activity to levels signifi-

cantly below those in sham-operated control animals

Fig. 3. Quantitative cytochemistry measurement for GC activityassessed by determination of pyrophosphate production in tissue slicesfrom control and BDL treated animals, in the presence and absence ofsubstrates for NOS (arginine and NADPH). Data are expressed asmean ± S.E.M., n = 6 in each case. *** signifies P < 0.001.

(P < 0.001). It is of note that the relative level of enhancement

in GC activity, induced by the supplementation of substrates

for endogenous NOS, was similar in both the BDL and sham

control groups (sham = 29.9%, BDL = 26.5%). This suggests

that BDL results in a reduction of GC activity per cell, but that

this is achieved without affecting the capacity for the GC pres-

ent to be activated by endogenous NOS.

In order to further address these issues, we used the NO do-

nor, DEANO, to stimulate GC activity in a manner that was

independent of endogenous NOS activity. Firstly we examined

the effect of differing concentrations of DEANO on GC activ-

ity in sections of liver from control rats. This showed that

DEANO induced increases in GC activity in control tissues

in a log dose-dependent manner (Fig. 4). Following such cali-

bration, we supplemented the GC reaction media with a con-

centration of DEANO that would produce an excess of NO.

Fig. 4. Quantitative cytochemistry measurement for GC activity,assessed by determination of pyrophosphate production in tissue slicesfrom control animals in the presence of increasing concentrations ofDEANO. Data are expressed as individual data points taken fromsections taken in triplicate from three animals. The line indicates a lognon-linear regression curve fit where r2 = 0.97.

2126 N.A. Davies et al. / FEBS Letters 580 (2006) 2123–2128

Such supplementation was found to stimulate similar increases

in GC activity/cell as those induced by NOS substrate supple-

mentation (Fig. 5A and B).

The effect of increasing concentrations of DEANO on GC

activity in serial sections of a control liver demonstrates that

the greater NO availability, engendered by inclusion of

increasing concentrations of DEANO, produced increased

density of reaction product (activity) which is particularly

apparent in cells at the edges of the sinusoids (Fig. 6). This sug-

gests that endogenous NO generation in control tissue is capa-

ble of inducing maximal increases in GC activity levels.

Furthermore, this NOS-stimulated increase in GC activity

was prevented in the presence of the NOS inhibitor LL-NAME

Fig. 5. Quantitative cytochemistry measurement for GC activityassessed by determination of pyrophosphate production in tissue slicesfrom (A) Control and (B) BDL treated animals, in the presence andabsence of substrates for NOS (arginine and NADPH), DEANO andNOS substrates plus the NOS inhibitor LL-NAME. Data are expressedas mean ± S.E.M., n = 6 in each case. ** signifies P < 0.01.

Fig. 6. Illustrative example of the staining density obtained withincreasing concentrations of DEANO in 14 lm sections of liverobtained from a control animal (at 10· magnification).

(Fig. 5). Indeed, the addition of LL-NAME decreased baseline

GC activity below that found in the absence of NOS substrates

(P < 0.01 in both sham and BDL groups), suggesting that a

proportion of this basal GC activity is attributable to the pres-

ence of endogenous NOS activity.

To confirm that the observed reaction product was depen-

dent upon the activity of GC, the effect of ODQ, an inhibitor

of GC, was examined. This showed that the level of GC activity

Fig. 7. Quantitative cytochemistry measurement for GC activityassessed by determination of pyrophosphate production in tissue slicesfrom control animals in the presence and absence of NOS substrates,expired (ex) DEANO, and the specific GC inhibitor ODQ. Data areexpressed as mean ± S.E.M., n = 3 in each case, *** indicates P < 0.001where indicated.

N.A. Davies et al. / FEBS Letters 580 (2006) 2123–2128 2127

per cell was significantly reduced by ODQ in control tissues

(Fig. 7). The remaining background activity is likely to be

due to the spontaneous GTP breakdown (data not shown).

GC activity, assessed in the presence of either NOS substrates

or DEANO, was also similarly inhibited by the inclusion of

ODQ (Fig. 7). This demonstrates that GC activity is required

for both the NOS-dependent and NO donor-dependent in-

crease in measured optical density. It was also found that a

solution of DEANO in which the NO donating capacity had

expired with prolonged storage, had no effect on GC activity.

4. Discussion

The results of his study have consistently shown that hepa-

tocellular GC activity in the BDL treated animals was signifi-

cantly lower than in controls. Stimulation of GC by NO

increased activity to a similar extent in both control and

BDL groups. Yet in all cases the total GC activity in the

BDL animals remained lower. Coupled with the formation

of large fibrotic areas in which GC active cells were almost

negligible, this indicates that overall hepatic GC activity is se-

verely restricted following BDL.

NOS activity in cirrhosis has been the subject of much inves-

tigation, and it has been suggested that decreased hepatic NO

generation may be responsible for reduced vascular control

and increased intra-hepatic resistance [16]. In this study we

found no difference in NOS activity between the two groups.

This similarity in NOS activity between the control and BDL

treated groups was found when assessed with an isolated max-

imal rate assay. A similar finding was made by Wei et al. [17],

although in this study it was found that though there was no

change in total NO production, the activity of constitutive

NOS was reduced and iNOS increased in BDL animals com-

pared to controls. It is also interesting to note that in a related

study by the same authors; the chronic administration of an

iNOS specific inhibitor (though not a non-specific NOS inhib-

itor) ameliorated liver injury in a BDL rat [18]. However, it is

important to consider that in both this study and that by Wei

et al. [17], only the soluble fraction of the tissue homogenate

was analysed after removal of the connective tissue, and that

the amount of NOS per gram of liver tissue may in fact be re-

duced in vivo. Furthermore, it is important to stress that this

assay precludes the effect of endogenous competitive NOS

inhibitors such as ADMA [19] which may have been present

prior to tissue processing.

A limitation of homogenate assays is that they cannot exam-

ine the specific activity in the different cell types present within

a tissue. QC is not restricted in this manner [20]. Furthermore,

though it is possible to add inhibitors/activators to homoge-

nate assays, again these may have diverse cell specific effects

that remain undetectable. Our investigations in the cirrhotic li-

ver, demonstrate the clear advantage in examining activity per

cell rather the common standardisation methods for homoge-

nate assays. These often utilise tissue weight or protein con-

tent, both of which would skew the results due to the

presence of fibrotic lesions.

It is clear from our findings that irrespective of the level of

NO generation in the liver, its efficacy would be reduced by

a decreased level of GC activity. This suggests that in situations

where the NO availability is reduced, the effect on vascular reg-

ulation would be magnified via an inability to produce cGMP.

Felipo and co-workers have previously shown that GC

activity is reduced by up to 50% in the brain of BDL rats fol-

lowing NO stimulation [21]. Similarly, brain GC activity was

also reduced in a portcaval anastamosis rat model of acute li-

ver failure, but the magnitude of this effect varied across differ-

ent regions [22,23]. Other studies by this group have linked the

effects on GC to hyperammonemia [23,24]. We found that the

plasma ammonia was slightly, but not significantly, elevated in

the BDL group compared to controls (23%, data not shown).

We are therefore unable to make any definitive conclusions

regarding ammonia affecting hepatic GC activity.

Previous studies have examined the activity of GC in

homogenates in various tissues in BDL models [25,26]. This

is the first study however, to directly examine GC activity in

the liver of BDL rats. It is also the first QC study to measure

GC activity in individual hepatocytes, and demonstrates

marked decreases associated with the development of cirrhosis.

This is an important point considering the large tracts of fibro-

tic tissue containing few GC active cells. Our study suggests

that it is likely that in this situation regulation of blood flow

would be severely compromised, and that any effects of NOS

substrates or NO donors acting via their influence on NOS

activity would be limited [27]. These considerations are also

relevant to NO produced independently from NOS, such as

that generated by the reaction between nitrite (NO2) and

deoxyhaemoglobin [28]. Gladwin and co-workers have sug-

gested that this mechanism may facilitate NO-stimulated vaso-

relaxation in areas of low oxygen tension [29–31]. This

mechanism however, presumably relies on the down-stream

stimulation of GC to affect a response, and is therefore likely

to remain limited in the cirrhotic liver.

The use of a phosphodiesterase (PDE) inhibitor was found to

improve memory in a hyperammonaemic rat model in which

GC activity was diminished [32]. It may be that the addition

of PDE inhibitors to prolong the effective lifetime of cGMP

may prove beneficial, particularly in the presence of an NO

donor. This combination may prove to be a useful therapeutic

option for the future, particularly if the effects could be targeted

to the liver. However, unfortunately it would not be possible to

examine the effects of PDE inhibitors with this technique as it is

the alternate reaction product (phosphate) which is measured.

In conclusion, we have demonstrated using quantitative

cytochemistry, that GC activity is significantly diminished in

hepatocytes of a cirrhotic liver, and that this is regardless of

the amount of NO generated. The implication of this study

is that the control of vasodilation in the cirrhotic liver is ulti-

mately dependent on the amount and activity of GC, and

not exclusively due to the availability of NO.

References

[1] Bredt, D.S. and Snyder, S.H. (1989) Nitric oxide mediatesglutamate-linked enhancement of cGMP levels in the cerebellum.Proc. Natl. Acad. Sci. USA 86, 9030–9033.

[2] Bredt, D.S. and Snyder, S.H. (1990) Isolation of nitric oxidesynthetase, a calmodulin-requiring enzyme. Proc. Natl. Acad. Sci.USA 87, 682–685.

[3] Mittal, M.K., Gupta, T.K., Lee, F.Y., Sieber, C.C. and Grosz-mann, R.J. (1994) Nitric oxide modulates hepatic vascular tone innormal rat liver. Am. J. Physiol. 267, G416–G422.

[4] Harry, D., Anand, R., Holt, S., Davies, S., Marley, R., Fernando,B. and Goodier, D., et al. (1999) Increased sensitivity toendotoxemia in the bile duct-ligated cirrhotic rat. Hepatology30, 1198–1205.

2128 N.A. Davies et al. / FEBS Letters 580 (2006) 2123–2128

[5] Mehdizadeh, S., O’Farrell, A., Bitensky, L., Weisz, J., Alagh-band-Zadeh, J. and Chayen, J. (1995) Histochemistry of guany-late cyclase activity. J. Histochem. Cytochem. 43, 1235–1240.

[6] Mehdizadeh, S., O’Farrell, A., Bitensky, L., Weisz, J., Alagh-band-Zadeh, J. and Chayen, J. (1995) Measurement of nitricoxide synthase activity in sections of rat liver. J. Histochem.Cytochem. 43, 1229–1233.

[7] Alaghband-Zadeh, J., Mehdizadeh, S., O’Farrell, A., Weisz, J.,Bitensky, L. and Chayen, J. (1999) An improved histochemicalmethod for measuring nitric oxide synthase activity. Cell Bio-chem. Funct. 17, 217–220.

[8] Alaghband-Zadeh, J., Mehdizadeh, S., Khan, N.S., O’Farrell, A.,Bitensky, L. and Chayen, J. (2001) The natural substrate for nitricoxide synthase activity. Cell Biochem. Funct. 19, 277–280.

[9] Bitensky, L. and Chayen, J. (1978) Quantitative cytochemistry:the basis of sensitive bioassays, for comparison of bio-andimmuno-reactive hormone values. Clin. Chem. 24, 1399–1407.

[10] Chayen, J., Chayen, R., Besser, G.M. and Bitensky, L. (1979)Quantitative cytochemistry in the detection and analysis ofsensitive rapid effects of hormones. Biochem. Soc. Trans. 7,857–860.

[11] Gornall, A.G. (1949) Determination of serum proteins by meansof the biuret reaction. J. Biol. Chem. 177, 751–766.

[12] Boer, R., Ulrich, W.R., Klein, T., Mirau, B., Haas, S. and Baur, I.(2000) The inhibitory potency and selectivity of arginine substratesite nitric-oxide synthase inhibitors is solely determined by theiraffinity toward the different isoenzymes. Mol. Pharmacol. 58,1026–1034.

[13] Prast, H., Tran, M.H., Fischer, H. and Philippu, A. (1998) Nitricoxide-induced release of acetylcholine in the nucleus accumbens:role of cyclic GMP, glutamate, and GABA. J. Neurochem. 71,266–273.

[14] Zhao, Y., Brandish, P.E., DiValentin, M., Schelvis, J.P., Babcock,G.T. and Marletta, M.A. (2000) Inhibition of soluble guanylatecyclase by ODQ. Biochemistry 39, 10848–10854.

[15] MacSween, R.N. and Burt, A.D. (1986) Histologic spectrum ofalcoholic liver disease. Semin. Liver Dis. 6, 221–232.

[16] Gupta, T.K., Toruner, M., Chung, M.K. and Groszmann, R.J.(1998) Endothelial dysfunction and decreased production of nitricoxide in the intrahepatic microcirculation of cirrhotic rats.Hepatology 28, 926–931.

[17] Wei, C.L., Khoo, H.E., Lee, K.H. and Hon, W.M. (2002)Differential expression and localization of nitric oxide synthases incirrhotic livers of bile duct-ligated rats. Nitric Oxide 7, 91–102.

[18] Wei, C.L., Hon, W.M., Lee, K.H. and Khoo, H.E. (2005) Chronicadministration of aminoguanidine reduces vascular nitric oxideproduction and attenuates liver damage in bile duct-ligated rats.Liver Int. 25, 647–656.

[19] Tran, C.T., Leiper, J.M. and Vallance, P. (2003) The DDAH/ADMA/NOS pathway. Atheroscler. Suppl. 4, 33–40.

[20] Altmann, F.P. and Chayen, J. (1965) Retention of nitrogenousmaterial in unfixed sections during incubation for histochemicaldemonstration of enzymes. Nature 207, 1205–1206.

[21] Rodrigo, R., Jover, R., Candela, A., Compan, A., Saez-Valero, J.,Erceg, S. and Felipo, V. (2005) Bile duct ligation plus hyperam-monemia in rats reproduces the alterations in the modulation ofsoluble guanylate cyclase by nitric oxide in brain of cirrhoticpatients. Neuroscience 130, 435–443.

[22] Rodrigo, R., Montoliu, C., Chatauret, N., Butterworth, R.,Behrends, S., Del Olmo, J.A. and Serra, M.A., et al. (2004)Alterations in soluble guanylate cyclase content and modulationby nitric oxide in liver disease. Neurochem. Int. 45, 947–953.

[23] Rodrigo, R., Erceg, S. and Felipo, V. (2005) Neurons exposed toammonia reproduce the differential alteration in nitric oxidemodulation of guanylate cyclase in the cerebellum and cortex ofpatients with liver cirrhosis. Neurobiol. Dis. 19, 150–161.

[24] Corbalan, R., Montoliu, C., Minana, M.D., Del Olmo, J.A.,Serra, M.A., Aparisi, L. and Rodrigo, J.M., et al. (2002) Alteredmodulation of soluble guanylate cyclase by nitric oxide in patientswith liver disease. Metab. Brain Dis. 17, 295–301.

[25] Liu, H., Song, D. and Lee, S.S. (2001) Role of heme oxygenase–carbon monoxide pathway in pathogenesis of cirrhotic cardio-myopathy in the rat. Am. J. Physiol. Gastrointest. Liver Physiol.280, G68–G74.

[26] Liu, H., Ma, Z. and Lee, S.S. (2000) Contribution of nitric oxideto the pathogenesis of cirrhotic cardiomyopathy in bile duct-ligated rats. Gastroenterology 118, 937–944.

[27] Brown, K.M., Brems, J.J., Moazzam, F.N., Hartman, G.G.,Gamelli, R.L. and Ding, J.W. (2003) The nitric oxide donormolsidomine improves survival and reduces hepatocyte apoptosisin cholestasis and endotoxemia. J. Am. Coll. Surg. 197, 261–269.

[28] Huang, K.T., Keszler, A., Patel, N., Patel, R.P., Gladwin, M.T.,Kim-Shapiro, D.B. and Hogg, N. (2005) The reaction betweennitrite and deoxyhemoglobin. Reassessment of reaction kineticsand stoichiometry. J. Biol. Chem. 280, 31126–31131.

[29] Crawford, J.H., Isbell, T.S., Huang, Z., Shiva, S., Chacko, B.K.,Schechter, A.N. and Darley-Usmar, V.M., et al. (2005) Hypoxia,red blood cells and nitrite regulate NO-dependent hypoxicvasodilatation. Blood.

[30] Cosby, K., Partovi, K.S., Crawford, J.H., Patel, R.P., Reiter,C.D., Martyr, S. and Yang, B.K., et al. (2003) Nitrite reductionto nitric oxide by deoxyhemoglobin vasodilates the humancirculation. Nat. Med. 9, 1498–1505.

[31] Gladwin, M.T., Crawford, J.H. and Patel, R.P. (2004) Thebiochemistry of nitric oxide, nitrite, and hemoglobin: role inblood flow regulation. Free Radic. Biol. Med. 36, 707–717.

[32] Erceg, S., Monfort, P., Hernandez-Viadel, M., Rodrigo, R.,Montoliu, C. and Felipo, V. (2005) Oral administration ofsildenafil restores learning ability in rats with hyperammonemiaand with portacaval shunts. Hepatology 41, 299–306.