Gastric motility in soluble guanylate cyclase α₁ knock-out mice

Abstract

The principal target of the relaxant neurotransmitter nitric oxide (NO) is soluble guanylate cyclase (sGC). As the α₁β₁-isoform of sGC is the predominant one in the gastrointestinal tract, the aim of this study was to investigate the role of sGC in nitrergic regulation of gastric motility in male and female sGCα₁ knock-out (KO) mice. In circular gastric fundus muscle strips, functional responses and cGMP levels were determined in response to nitrergic and non-nitrergic stimuli. sGC subunit mRNA expression in fundus was measured by real-time RT-PCR; in vivo gastric emptying of a phenol red meal was determined. No changes were observed in sGC subunit mRNA levels between wild-type (WT) and KO tissues. Nitrergic relaxations induced by short trains of electrical field stimulation (EFS) were abolished, while those by long trains of EFS were reduced in KO strips; the latter responses were abolished by 1H[1,2,4,]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ). The relaxations evoked by exogenous NO and the NO-independent sGC activator BAY 41-2272 were reduced in KO strips but still sensitive to ODQ. Relaxations induced by vasoactive intestinal peptide (VIP) and 8-bromo-cGMP were not influenced. Basal cGMP levels were decreased in KO strips but NO, long train EFS and BAY 41-2272 still induced a moderate ODQ-sensitive increase in cGMP levels. Gastric emptying, measured at 15 and 60 min, was increased at 15 min in male KO mice. sGCα₁β₁ plays an important role in gastric nitrergic relaxation in vitro, but some degree of nitrergic relaxation can occur via sGCα₂β₁ activation in sGCα₁ KO mice, which contributes to the moderate in vivo consequence on gastric emptying.

Nitric oxide (NO) is considered to be the most important non-adrenergic, non-cholinergic (NANC) neurotransmitter in the gastrointestinal tract. In the classic concept of nitrergic neurotransmission, NO synthesized from l-arginine by neuronal NO synthase (nNOS, NOS-1) in NANC neurons diffuses to the gastrointestinal smooth muscle cells where its principal target is soluble guanylate cyclase (sGC) generating cGMP (Rand & Li, 1995; Toda & Herman, 2005). cGMP activates cGMP-dependent protein kinase (PKG, cGK), which can lead to relaxation by several mechanisms such as the inhibition of InsP₃-dependent calcium release, the stimulation of the sarco(endo)plasmic reticulum Ca²⁺-ATPase (SERCA), which decreases intracellular calcium levels and the inhibition of phosphorylation of myosin light chains (Schlossmann & Hofmann, 2005). Recent evidence suggests that at some levels of the gastrointestinal tract, interstitial cells of Cajal are involved in nitrergic neurotransmission towards smooth muscle (Ward & Sanders, 2006).

sGC is a heterodimeric protein that contains an α and a β subunit, each containing a haeme-binding domain, a catalytic domain and a dimerization domain (Hobbs, 1997). NO stimulates sGC through the formation of a nitrosyl-haeme complex, leading to a 400-fold increase in catalytic activity (Stone & Marletta, 1996). Both the α and β subunit exist in two isoforms, α₁ and α₂ and β₁ and β₂. The α₁β₁ and the α₂β₁ heterodimers seem to be the physiologically active forms with higher basal and NO-stimulated activity for the α₁β₁ isoform (Harteneck et al. 1991) although more recent reports found no differences in kinetic properties and sensitivity towards NO between the two isoforms (Russwurm et al. 1998; Mergia et al. 2003). At the mRNA level, more α₁ than α₂ mRNA is found in the gastrointestinal tract and the α₁β₁ isoform is thus thought to be the predominant one in the gastrointestinal tract (Mergia et al. 2003).

The NOS-1–sGC-cGK signalling pathway plays an essential role in the regulation of gastrointestinal motility (Shah et al. 2004). In NOS-1 knock-out (KO) mice, the most important abnormality is an enlarged stomach with hypertrophy of the pyloric sphincter and the circular muscle layer (Huang et al. 1993). Smooth muscle relaxation induced by electrical field stimulation (EFS) in NANC conditions is reduced in gastric fundus and jejunum strips of these mice (Xue et al. 2000; Dick et al. 2002) and gastric emptying of liquids and solids is delayed (Mashimo et al. 2000). Of the two forms of cGK, cGKI is highly expressed in smooth muscle (Keilbach et al. 1992). cGKI KO mice show dilatation of the stomach and the caecum (Ny et al. 2000); electrically induced relaxation of gastric fundus strips is impaired, as well as gastric emptying and intestinal peristalsis of a barium suspension (Pfeifer et al. 1998). Still, NO has sGC-independent effects such as nitrosylation of proteins and sGC can be stimulated by other stimuli than NO such as carbon monoxide (CO; Schmidt, 1992; Jaffrey et al. 2001; Behrends, 2003). The aim of our study was therefore to investigate the role of sGCα₁ in gastric smooth muscle regulation by studying in vitro and in vivo gastric motility in sGCα₁ KO mice, lacking exon 6 of the sGCα₁ gene, coding for an essential part in the catalytic domain. The in vivo cardiovascular consequences of knocking out sGCα₁ have been reported preliminarily by Sips et al. (2005) and the influence on in vitro nitrergic vascular smooth muscle relaxation was recently published (Nimmegeers et al. 2007).

Methods

Animals

sGCα₁ KO mice were generated as described by Sips et al. (2005). Briefly, sGCα₁ KO mice were generated by targeting exon 6 of the sGCα₁ gene, which codes for an essential part of the catalytic domain. Chimeras were generated by aggregating R1 embryonic stem cells carrying the mutant sGCα₁ allele with Swiss morula-stage embryos. These sGCα₁ KO mice express a mutant α₁ protein, but when this was coexpressed with sGCβ₁ in insect cells, the resulting heterodimer was inactive (authors' unpublished results).

Wild-type (WT) and KO Swiss/129 mice of both sexes (WT male: 8–50 weeks, 20–53 g; KO male: 8–40 weeks, 27–56 g; WT female: 8–50 weeks, 23–44 g; KO female: 8–42 weeks, 22–44 g) had free access to water and commercially available chow except for the study of gastric emptying; food was then withheld for 16 h overnight with free access to water in a cage with a grid floor to prevent coprophagy. Animals were killed by cervical dislocation. All experiments were performed on mice of both sexes unless otherwise indicated. All experimental procedures were approved by the Ethical Committee for Animal Experiments from the Faculty of Medicine and Health Sciences at Ghent University.

Real-time quantitative RT-PCR

Gastric fundus tissue was harvested, rinsed in Krebs solution and transferred to RNAlater stabilization reagent (Qiagen) until RNA was prepared using the RNeasy® Protect Mini kit from Qiagen, according to the manufacturer's instructions. One microgram of RNA was used in the SuperScript™ First-Strand Synthesis System for RT-PCR from Invitrogen to produce cDNA using random hexameric primers. cDNA was subsequently used for relative expression quantification using real-time fluorescence detection. To this end, specific primer-probe sets were synthesized for sGCα₁ (detecting exon 5), α₂, β₁ and the household gene hypoxanthine-guanine phosphoribosyltransferase (HPRT) for normalization (Table 1). Quantitative real-time PCR was performed using the Taqman universal PCR master mix (Applied Biosystems) in an ABI Prism 7700 sequence detector. The cycling parameters were: 95°C for 10 min, followed by 40 cycles at 95°C for 15 s and 60°C for 1 min.

Table 1.

Sequence of primer and probe sets for real-time RT-PCR analysis

Primer		Sequence
sGCα1	forward	GCCCCAGTCCTCGCTAGTG
	reverse	GCATGAAATGGAACGGGAAA
	probe	FAM-TCCCCGCTTCGCTCTTCTGCAAG-TAMRA
sGCα2	forward	TGCATCTCTCAGACATCCCTATTC
	reverse	TGGGCCTTGGCTTGCTC
	probe	FAM-CCAACTAAAATTACATCTCGGGTGGCATCA-TAMRA
sGCβ1	forward	TCAGTGTGGCAATGCCATCTA
	reverse	AGGGCGGACCAGAGAGAAGA
	probe	FAM-CCAGCCTGGGAACTGCAGCCTTCT-TAMRA
HPRT	forward	TTATCAGACTGAAGAGCTACTGTAATGATC
	reverse	TTACCAGTGTCAATTATATCTTCAACAATC
	probe	FAM-TGAGAGATCATCTCCACCAATAACTTTTATGTCCC-TAMRA

Relative changes in gene expression were determined using the comparative C_T method described in Applied Biosystems User Bulletin no. 2 (1997) in which the amount of target, normalized to an endogenous reference and relative to a calibrator is given by . In brief, the C_T, or threshold cycle, represents the PCR cycle at which an increase in reporter fluorescence above a baseline signal can first be detected, ΔC_T refers to the difference between the threshold cycles of the target (sGC subunits) and the endogenous reference (HPRT); and ΔΔC_T indicates the difference between the ΔC_T values obtained in the KO animals to the mean ΔC_T value obtained in the calibrator group (WT mice).

Muscle tension experiments

Tissue preparation

Animals were killed by cervical dislocation; the gastrointestinal tract was removed and put in aerated Krebs solution (composition in mmol l⁻¹: NaCl 118.5, KCl 4.8, KH₂PO₄ 1.2, MgSO₄ 1.2, CaCl₂ 1.9, NaHCO₃ 25.0 and glucose 10.1). The gastric fundus and the pyloric region were separated from the rest of the stomach. Two full wall thickness fundus strips (width: 2 mm; length without tension: 11 mm) were prepared from the fundus by cutting in the direction of the circular muscle layer; 1 full wall thickness ring segment (width: 2 mm) was prepared from the pyloric region.

Isometric tension recording

After a silk thread (USP 4/0) was attached to both ends of the fundus strips and 2 L-shaped tissue hooks were inserted into the pyloric ring, strips and rings were mounted in 10 ml (pylorus) or 15 ml (fundus) organ baths between two platinum plate electrodes (6 mm apart). The organ baths contained aerated (5% CO₂ in O₂) Krebs solution, maintained at 37°C. Changes in isometric tension were measured using MLT 050/D force transducers (ADInstruments) and recorded on a Graphtec linearcorder F WR3701 (Graphtec, Yokohama, Japan; fundus) or on a PowerLab/8sp data recording system (ADInstruments) with Chart software (pylorus).

After an equilibration period of 30 min with flushing every 10 min at a load of 0.75 g (fundus) or 0.25 g (pylorus), the length–tension relationship was determined. Muscle tissues were stretched by load increments of 0.25 g and at each load level exposed to 0.1 μmol l⁻¹ (fundus) or 10 μmol l⁻¹ (pylorus) carbachol to determine the optimal load (L_o; the load at which maximal response to the contractile agent occurred). Tissues were then allowed to equilibrate for 60 min at L_o with flushing every 15 min in Krebs solution.

Protocols in fundic strips

Cumulative contractile responses to carbachol (1 nmol l⁻¹ to 30 μmol l⁻¹) or prostaglandin F2α (PGF2α; 1 nmol l⁻¹ to 3 μmol l⁻¹) were obtained in Krebs solution without atropine and guanethidine. All other experiments were performed after switching to Krebs solution containing 1 μmol l⁻¹ atropine and 4 μmol l⁻¹ guanethidine to block cholinergic and noradrenergic responses, respectively (NANC conditions). Ten minutes after adding 300 nmol l⁻¹ PGF2α to induce contraction, relaxations were induced with 5 min interval in between by application of electrical field stimulation (EFS; 40 V, 0.1 ms, 1–8 Hz for 10 s) via the platinum plate electrodes by means of a Grass S88 stimulator (Grass, W. Warwick, RI, USA), followed by the application of exogenous NO (1 μmol l⁻¹ to 100 μmol l⁻¹) and vasoactive intestinal peptide (VIP; 100 nmol l⁻¹). Strips were washed for 30 min, and interfering drugs were then incubated for 30 min. PGF2α was then applied again and the responses to EFS, NO and VIP were studied again in the presence of the NOS inhibitor l-NAME (300 μmol l⁻¹), the sGC inhibitor 1H[1,2,4,]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ; 10 μmol l⁻¹) or the small conductance Ca²⁺-dependent K⁺-channel (SK_Ca-channel) blocker apamin (500 nmol l⁻¹; only studied in male mice versus EFS and NO). The reproducibility of the responses to EFS, NO and VIP was evaluated by running time-control strips in parallel that did not receive the interfering drugs or that received the solvent of these drugs when it was not water (ethanol for ODQ). In a similar way, the influence of l-NAME and ODQ was also studied (in male mice only) versus the responses to 60 s trains of EFS (40 V, 0.1 ms, 1–16 Hz), and that of ODQ versus 8-bromoguanosine 3′,5′-cyclic monophosphate (8-Br-cGMP, 10 μmol l⁻¹) and [5-cyclopropyl-2-[1-(2-fluoro-benzyl)-1H-pyrazolo[3,4-b]pyridine-3-yl]pyrimidin-4-ylamine] (BAY 41-2272; 0.1 to 10 μmol l⁻¹ added cumulatively). As preliminary experiments had shown that DMSO, the solvent of BAY 41-2272, decreased the PGF2α-induced contraction to some extent when added cumulatively, DMSO was added cumulatively during a third contraction cycle and these responses were subtracted from these by BAY 41-2272.

Protocol in pyloric rings

All experiments were performed in NANC conditions. Relaxations were induced by application of EFS (40 V, 0.5 ms, 2–16 Hz for 60 s) via the platinum plate electrodes by means of a Grass S44 stimulator with a 5 min interval between stimulation trains, followed by the application of exogenous NO (10 to 100 μmol l⁻¹). Rings were washed for 15 min and the responses to EFS and NO were studied again twice, the last time after incubation for 30 min with ODQ (10 μmol l⁻¹); in female mice, an additional series was performed studying the influence of l-NAME (300 μmol l⁻¹).

Data analysis

The amplitude of contractile and relaxant responses was measured and expressed in grams per gram wet weight. In fundic strips, the relaxant responses were expressed as the percentage of contraction evoked by PGF2α. EC₅₀ values of the concentration–response curves were calculated by linear interpolation.

cGMP analysis

Circular gastric fundus muscle strips were prepared and weighed to obtain the tissue wet weight before they were mounted as described above, under the mean optimal load determined in the tension experiments (1.5 g). Pre-contracted fundus strips were snap-frozen in liquid nitrogen either at maximal relaxation by NO (10 μmol l⁻¹), upon EFS (40 V, 0.1 ms) at 4 Hz for 10 s or 16 Hz for 60 s (in males only), or 10 min after application of BAY 41-2272 (10 μmol l⁻¹, in males only); control strips received nothing (NO and EFS) or DMSO (solvent of BAY 41-2272). Experiments with EFS at 16 Hz for 60 s were done in the continuous presence of the phosphodiesterase type 5 (PDE-5) inhibitor zaprinast (10 μmol l⁻¹). In fundus strips of males, the response to EFS (16 Hz for 60 s) was also studied in the presence of ODQ (10 μmol l⁻¹) or l-NAME (300 μmol l⁻¹), and the response to NO and BAY 41-2272 was studied in the presence of ODQ (10 μmol l⁻¹). l-NAME and ODQ were incubated for 30 min.

Snap-frozen tissues were stored at −80°C until further processing. They were pulverized by a Mikro-dismembrator U (B-Braun Biotech International, Germany) and dissolved in cold 6% trichloroacetic acid to give a 10% (w/v) homogenate. The homogenate was centrifuged at 2000 g for 15 min at 4°C; the supernatant was recovered and washed four times with five volumes of water-saturated diethyl ether. The aqueous extract was then dried under a stream of nitrogen at 60°C and dissolved in a 10, 50 or 100 times volume of assay buffer depending on the cGMP content detected in preliminary experiments. cGMP concentrations were determined using an enzyme immunoassay kit (EIA Biotrak System; Amersham Biosciences, UK) after acetylation of the samples and according to the manufacturer's instructions. The optical density was measured with a 96-well plate reader (Biotrak II, Amersham Biosciences) at 450 nm. The tissue cGMP concentration was expressed as picomoles per gram tissue wet weight.

Gastric emptying

For gastric emptying measurements, we used a modification of the technique previously described by de Rosalmeida et al. (2003).

After food was withheld overnight, mice (age: 20 ± 3 weeks) were administered 250 μl of a phenol red meal (1 mg ml⁻¹ phenol red dissolved in water) by gavage with a feeding needle. Fifteen or sixty minutes later, mice were killed by cervical dislocation and the stomach and small bowel were clamped at both sides. Tissues were cut into small fragments and placed into 20 ml of 0.1 n NaOH in a 50 ml Falcon tube. This mixture was homogenized for approximately 30 s and allowed to stand for 20 min at room temperature. Ten millilitres of supernatant was placed into a 15 ml Falcon tube and centrifuged for 10 min at 2800 r.p.m. (1580g). Proteins in 5 ml supernatant were precipitated with 0.5 ml of 20% (w/v) trichloroacetic acid and the solution was centrifuged for 20 min at 2800 r.p.m. (1580g) 0.5 ml of supernatant was added to 0.667 ml of 0.5 n NaOH and the absorbance of 300 μl of this mixture was spectrophotometrically determined at 540 nm in a Biotrak II plate reader (Amersham Biosciences).

Gastric emptying was calculated as the amount of phenol red that already left the stomach as a percentage of the total amount of phenol red recovered and the phenol red recovery was determined as the amount of phenol red recovered, expressed as a percentage of the amount of phenol red administered.

Statistics

All results are expressed as means ±s.e.m.n refers to tissues obtained from different animals unless otherwise indicated. Comparison between KO and WT tissues or between parallel tissues of either WT or KO was done with Student's t test for unpaired data. Comparison within tissues of either WT or KO was done by a t test for paired data. When relaxant responses in time-control strips showed a significant decline when studied a second time, the change in relaxant response by an interfering substance in the parallel tissue was compared to the spontaneous change in the control strips by an unpaired student's t test. When more than two groups of tissues had to be compared, one-way ANOVA followed by a Bonferroni's corrected t test was applied. P ≤ 0.05 was considered to be statistically significant (Prism, Graphpad, San Diego, CA, USA).

Drugs used

The drugs used were as follows: apamin (Alomone Laboratories, Israel), atropine (Sigma-Aldrich), BAY 41-2272 (Alexis, Switzerland), 8-bromoguanosine 3′,5′-cyclic monophosphate sodium salt (8-Br-cGMP; Sigma-Aldrich), carbachol (Fluka AG, Switzerland), guanethidine (Sigma-Aldrich), N^ω-nitro-l-arginine methyl ester (l-NAME; Sigma-Aldrich), 1H[1,2,4,]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ; Tocris Cookson, UK), phenol red (Sigma-Aldrich), prostaglandin F2α (Sigma-Aldrich), vasoactive intestinal peptide (VIP; Sigma-Aldrich) and zaprinast (a gift of Rhone-Poulenc, UK). All drugs were dissolved in deionized water except ODQ and BAY 41-2272. ODQ was dissolved in 100% ethanol and the 1 and 10 mmol l⁻¹ stock solutions of BAY 41-2272 were dissolved in 100% DMSO. Further dilutions of BAY 41-2272 were made with deionized water. Saturated NO solution was prepared from gas (Air Liquide, Belgium) as described by Kelm & Schrader (1990).

Results

The majority of experiments were performed in mice of both sexes because sex-related differences in the cardiovascular phenotype were observed: male but not female sGCα₁ KO mice develop hypertension (Sips et al. 2005). No systematic differences between sexes were found in fundic strips, explaining why some experiments were carried out in male mice only for this tissue. Results are presented for male mice; when a result was different in female mice, this is reported. Muscle tension experiments in pyloric rings showed differences between sexes and are therefore presented for male and female mice. No gross morpho-logical gastrointestinal differences between WT and KO animals were observed.

Real-time quantitative RT-PCR

No significant differences were found in sGC subunit expression between WT and KO gastric fundus tissues; the mean levels of α₁, α₂ and β₁ subunit mRNA in KO tissues versus WT tissues were: 1.182 ± 0.222, 1.247 ± 0.553 and 1.218 ± 0.270 (n = 4), respectively.

Muscle tension experiments

General observations

The length of the fundus strips, measured after equilibration and determination of optimal load, and the tissue wet weight, measured immediately after the experiment, were not different between WT and KO mice. In the series where NO-, EFS- and VIP-induced responses were studied, length and weight of the strips were, respectively, 17 ± 1 mm and 7.99 ± 0.67 mg in WT (n = 18) and 16 ± 1 mm and 7.25 ± 0.90 mg in KO tissues (n = 17). The wet weight of the pyloric rings did not differ in male KO compared to male WT mice (16.33 ± 2.23 mg in WT and 16.51 ± 1.28 mg in KO, n = 6) but was increased in female KO compared to female WT mice (16.28 ± 0.85 mg in WT, n = 13 and 19.55 ± 1.29 mg in KO, n = 11, P < 0.05). When interfering agents were administered to study their effects on the responses to the different stimuli, their possible influence on basal tone during their incubation period was studied. None of the agents tested had a consistent influence on basal activity.

Contractile responses to carbachol and PGF2α in fundus

Contractile responses to carbachol and PGF2α did not significantly differ in gastric fundus of KO mice compared to WT mice. The EC₅₀ (nmol l⁻¹) and E_max (g per g wet weight) for carbachol were 202 ± 51 and 323 ± 43 in WT (n = 9), and 128 ± 16 and 369 ± 36 in KO (n = 8); for PGF2α, they were 138 ± 34 and 272 ± 52 in WT (n = 9) and 214 ± 47 and 299 ± 45 in KO (n = 8).

Inhibitory responses to electrical field stimulation, exogenously applied NO and BAY 41-2272 in fundus

In fundus, the contractile response to 300 nmol l⁻¹ PGF2α was not significantly different between WT and KO strips. Also, ODQ (10 μmol l⁻¹), apamin (500 nmol l⁻¹) and l-NAME (300 μmol l⁻¹) had no significant influence on the PGF2α-induced response. In WT strips, 10 s trains of EFS induced frequency-dependent relaxations that were reproducible; these responses were either abolished (Fig. 1) in KO strips or a small relaxation was retained (Fig. 2). In WT strips and also in KO strips, if some response to EFS occurred, the EFS-evoked responses were abolished by ODQ (n = 7–8, Fig. 2) and l-NAME (n = 7). Apamin had no influence on the electrically evoked responses in both WT and KO strips (n = 4).

Figure 1.

Smooth muscle responses to short train EFS and NO in gastric fundus strips of WT and KO mice Traces showing the inhibitory responses of precontracted (PGF2α; 300 nmol l⁻¹) circular muscle strips of male mouse gastric fundus to electrical field stimulation (EFS; 40 V, 0.1 ms, 1–8 Hz, 10 s trains) and exogenously applied NO (1–100 μmol l⁻¹). A, wild-type mouse (WT); B, knock-out mouse (KO). Arrows indicate the moment of EFS or administration of PGF2α or NO.

Figure 2.

Influence of ODQ on the responses to short train EFS, NO and BAY 41-2272 in gastric fundus strips of WT and KO mice Frequency–response curves of electrical field stimulation (EFS, 40 V, 0.1 ms, 1–8 Hz, 10 s trains, A), and concentration-response curves of NO (1–100 μmol l⁻¹, B) and of BAY 41-2272 (0.1–10 μmol l⁻¹, C) in gastric fundus circular muscle strips of male wild-type (WT) and knock-out (KO) mice before and after incubation with ODQ (10 μmol l⁻¹). Data are expressed as percentage relaxation of the contraction induced by PGF2α. Values are means ±s.e.m. of n = 7–9 strips. †P < 0.05, ††P < 0.01: unpaired student's t test (KO before ODQ versus WT before ODQ). *P < 0.05, **P < 0.01, ***P < 0.001: paired student's t test (after ODQ versus before ODQ). The decrease in response to 100 μmol l⁻¹ NO and to 1 and 3 μmol l⁻¹ BAY 41-2272 after ODQ in WT strips, was significant upon paired testing within the tissue as shown, but it was not significantly different from the decrease in response after ethanol in the parallel time control strips (unpaired student's t test).

Application of exogenous NO (1 to 100 μmol l⁻¹) in fundic strips induced concentration-dependent relaxations that tended to decline when repeated for a second time. In KO strips, the amplitude of the relaxant responses to NO was reduced in comparison to WT strips but pronounced relaxation was still systematically obtained (Figs 1 and 2); in 65% of the KO strips also the duration of the response to NO was diminished (Fig. 1). The responses to exogenously applied NO were not influenced by l-NAME (n = 9) and apamin (n = 4), but were reduced by ODQ in both WT and KO strips; the inhibition by ODQ was more pronounced in KO strips (Fig. 2). ODQ reduced the response by 100 μmol l⁻¹ NO by 20.5 ± 6.6% in WT (n = 9) and 68.9 ± 2.8% in KO strips (n = 8, P < 0.001).

The responses to BAY 41-2272 were significantly reduced in KO strips when compared to WT strips, but the EC₅₀ values were not significantly different (WT: 2.1 ± 0.2 μmol l⁻¹, n = 11 strips, 6 animals; KO: 2.2 ± 0.5 μmol l⁻¹, n = 12 strips, 6 animals). The concentration–response curve of BAY 41-2272 tended to decline when studied for a second time in control conditions. ODQ did not influence the EC₅₀ and, when compared to the control tissues, only reduced the response to 10 μmol l⁻¹ BAY 41-2272 in both WT and KO strips (n = 6).

In WT fundus strips, 60 s trains of EFS induced reproducible relaxations. They had a fast onset, tended to diminish somewhat during the stimulation train at low frequencies but were well maintained at high frequencies, and they were followed by an off-contraction. Both l-NAME (n = 6) and ODQ (n = 6) abolished these responses and on-contractions occurred during stimulation; with EFS at 16 Hz, a relaxation tended to develop during stimulation after a primary on-contraction (Fig. 3). In KO strips, no relaxation occurred with EFS at 1 Hz and the relaxant response with EFS at 2–16 Hz occurred at a slower rate; the amplitude was reduced and less well maintained in comparison to WT strips (Fig. 3); in one strip only contractions were observed. ODQ abolished the responses at 2–8 Hz and greatly reduced that at 16 Hz (n = 5, Fig. 3); l-NAME had a similar effect (n = 4).

Figure 3.

Smooth muscle responses to long train EFS in gastric fundus strips of WT and KO mice Traces showing the inhibitory responses of precontracted (PGF2α; 300 nmol l⁻¹) circular muscle strips of mouse gastric fundus to electrical field stimulation (EFS; 40 V, 0.1 ms, 1–16 Hz, 60 s trains) in tissues of male wild-type (WT; trace 1 and 2) and knock-out (KO; trace 3 and 4) mice before (trace 1 and 3) and after administration of ODQ (10 μmol l⁻¹; trace 2 and 4). Arrows indicate the moment of administration of PGF2α. Bars represent EFS.

Inhibitory responses to electrical field stimulation and exogenously applied NO in pylorus

In male WT pyloric rings, 60 s trains of EFS induced well-maintained relaxations that were followed by an off-contraction; application of exogenous NO (10 to 100 μmol l⁻¹) evoked a decline in tone that slowly recuperated (Fig. 4A). The frequency and concentration dependency was minimal. The relaxant responses in KO rings were not significantly lower than in WT rings (n = 6, Fig. 4A and B). In both WT and KO rings, the responses to EFS and NO were reproducible; ODQ abolished the responses to EFS in both WT and KO rings, while it reduced the responses to NO (e.g. 100 μmol l⁻¹ in WT from 4.43 ± 0.81 to 1.62 ± 0.55, P < 0.001 and in KO from 3.31 ± 0.54 to 0.18 ± 0.18 g (g wet wt)⁻¹, P < 0.001, n = 6, Fig. 4A).

Figure 4.

Smooth muscle responses to EFS and NO in pyloric rings of WT and KO mice A, traces showing the inhibitory responses of pyloric full thickness muscle rings to electrical field stimulation (EFS; 40 V, 0,5 ms, 8 Hz, 60 s train) and exogenously applied NO (100 μmol l⁻¹) in tissues of male wild-type (WT) and knock-out (KO) mice before and after administration of ODQ (10 μmol l⁻¹). Bars represent EFS. Arrows indicate moment of administration. B, frequency–response curves of electrical field stimulation (EFS, 40 V, 0.5 ms, 2–16 Hz, 60 s trains, left), and concentration–response curves of NO (10–100 μmol l⁻¹, right) in pyloric full thickness muscle rings of male and female WT and KO mice. Data are expressed as g (g wet weight)⁻¹. Values are means ±s.e.m. of n = 6 rings. *P < 0.05: unpaired student's t test (KO versus WT).

In female KO rings, the reduction of the EFS- and NO-induced relaxations compared to WT was much more pronounced than in male mice and significant (n = 6, Fig. 4B). In WT rings, these responses were reduced by ODQ (e.g. EFS at 2 Hz from 5.09 ± 1.64 to 1.26 ± 1.26, P < 0.001 and at 16 Hz from 5.73 ± 1.79 to 1.01 ± 1.01 g (g wet wt)⁻¹, P < 0.001; NO at 10 μmol l⁻¹ from 5.38 ± 1.39 to 2.65 ± 0.88, P < 0.001 and at 100 μmol l⁻¹ from 5.74 ± 1.38 to 3.50 ± 0.87 g (g wet weight)⁻¹, P < 0.01, n = 6); in KO rings, the small responses to EFS and NO were nearly abolished by ODQ (n = 6). l-NAME did not influence the responses to NO; it reduced the responses to EFS in WT rings (e.g. at 2 Hz from 3.30 ± 0.77 to 1.09 ± 0.80, P < 0.001 and at 16 Hz from 3.81 ± 0.94 to 0.84 ± 0.84 g (g wet wt)⁻¹, P < 0.001, n = 7) and nearly abolished the small EFS-induced responses in KO rings (e.g. at 2 Hz from 0.90 ± 0.56 to 0.00 ± 0.00 and at 16 Hz from 1.41 ± 0.60 to 0.28 ± 0.28 g (g wet wt)⁻¹, n = 5).

Inhibitory responses to exogenously applied VIP and 8-Br-cGMP in fundus

Exogenously applied VIP (100 nmol l⁻¹) completely reversed PGF2α-induced tone and tone even decreased somewhat below that prevailing before PGF2α administration (relaxation = 123 ± 5%) in WT strips (n = 36 strips, 18 animals). This was not changed in KO strips (115 ± 3%; n = 34 strips, 18 animals).

The relaxant response to 8-Br-cGMP (10 μmol l⁻¹) was similar in WT strips (77 ± 4%, n = 12 strips, 6 animals) and in KO strips (85 ± 3%, n = 12 strips, 6 animals); the response to 8-Br-cGMP was not reduced by ODQ in both WT and KO strips.

cGMP analysis

In WT strips, cGMP levels were significantly increased 34-fold by NO versus basal (7.3 ± 0.9 pmol (g tissue)⁻¹, n = 7) but were not changed by the 10 s train of EFS (40 V, 0.1 ms, 4 Hz; Fig. 5A). In KO strips, the basal cGMP level (4.1 ± 1.0 pmol (g tissue)⁻¹, n = 7) was lower than in WT strips; NO still induced a moderate but significant 3-fold increase in cGMP levels (Fig. 5A). In a separate series of experiments, the influence of ODQ on the increase in cGMP by NO was studied. ODQ largely prevented the increase in cGMP by NO in both WT and KO strips (Fig. 5B). In WT strips, the cGMP levels generated by NO in the presence of ODQ were 2.2 ± 0.6 pmol (g tissue)⁻¹ (n = 5) versus 1.5 ± 0.3 pmol (g tissue)⁻¹ in basal conditions (n = 3) while in KO strips, they were 0.6 ± 0.2 pmol (g tissue)⁻¹ (n = 4) versus 0.7 ± 0.1 pmol (g tissue)⁻¹ in basal conditions (n = 4).

Figure 5.

cGMP levels in response to short and long train EFS, NO and BAY 41-2272 in gastric fundus strips of WT and KO mice Basal cGMP levels and cGMP levels measured (A) upon electrical field stimulation (EFS, 40 V, 0.1 ms, 4 Hz, 10 s) or at maximal relaxation by NO (10 μmol l⁻¹), (B) at maximal relaxation by NO (10 μmol l⁻¹) in control conditions and after incubation with ODQ (10 μmol l⁻¹), (C) upon 60 s of EFS (40 V, 0.1 ms, 16 Hz) in control conditions and after incubation with l-NAME (300 μmol l⁻¹) or ODQ (10 μmol l⁻¹), and (D) 10 min after application of BAY 41-2272 (10 μmol l⁻¹) in control conditions and after incubation with ODQ (10 μmol l⁻¹) in gastric fundus circular muscle strips of male wild-type (WT; left) and knock-out (KO; right) mice. The experiments depicted in C were performed in the presence of zaprinast (10 μmol l⁻¹). Data are expressed as pmol (g tissue)⁻¹ and scales are different for WT and KO panels. Values are means ±s.e.m. of n = 6–8 strips except for panel B (WT and KO, n = 3–5) and panel D (KO, n = 4–6). *P < 0.05, **P < 0.01, ***P < 0.001: ANOVA followed by a Bonferroni's corrected t test (stimulus versus basal). †P < 0.05, ††P < 0.01, †††P < 0.001: unpaired student's t test (KO versus WT).

Preliminary experiments showed that also EFS with 60 s trains (40 V, 0.1 ms, 16 Hz) did not induce a rise in cGMP levels. Experiments with 60 s trains of EFS in WT and KO strips were therefore studied in the presence of the PDE-5 inhibitor zaprinast. Zaprinast did not influence basal cGMP levels but in its presence, the 60 s train of EFS elicited a significant 4.2-fold increase in cGMP level versus basal (7.6 ± 1.3 pmol (g tissue)⁻¹, n = 8) in WT gastric fundus; ODQ and l-NAME abolished the EFS-induced increase in cGMP level (Fig. 5C). In KO strips, the basal cGMP levels were 2.3 ± 0.3 pmol (g tissue)⁻¹ (n = 8); EFS induced a significant 1.8-fold increase in cGMP level that was also abolished by incubation with ODQ and l-NAME (Fig. 5C).

In WT strips, application of 10 μmol l⁻¹ BAY 41-2272 produced a significant 8.5-fold increase in cGMP levels versus basal; ODQ inhibited this effect (Fig. 5D). Administration of BAY 41-2272 in KO strips resulted in a significant 3.8-fold increase of the cGMP level that was reversed in the presence of ODQ (n = 6, Fig. 5D).

Gastric emptying

The phenol red recovery was 80.5 ± 1.4% and 77.5 ± 1.5% in, respectively, male WT and KO mice and 75.5 ± 2.0% and 73.3 ± 2.0% in, respectively, female WT and KO mice (n = 16 in each group).

Fifteen minutes after gavage, liquid gastric emptying was significantly higher in male KO mice but this difference disappeared 60 min after gavage (Fig. 6). In female mice, no significant difference in gastric emptying between WT and KO mice was observed (Fig. 6).

Figure 6.

Gastric emptying in WT and KO mice Gastric emptying 15 and 60 min after gavage of 250 μl phenol red meal (1 mg ml⁻¹ phenol red dissolved in water) in male and female wild-type (WT) and knock-out (KO) mice. Values are means ±s.e.m. of n = 8 animals. *P < 0.05: unpaired student's t test (KO versus WT).

Discussion

The aim of this study was to investigate the influence of sGCα₁ deficiency on the gastrointestinal tract concentrating on the proximal stomach, in view of the principal role of the NOS-1–sGC–cGK pathway in NANC relaxation of this tonic muscle region (Li & Rand, 1990; Lefebvre et al. 1992; Dick & Lefebvre, 1997; Pfeifer et al. 1998; Bayguinov & Sanders, 1998). When studying the cardiovascular consequences in the sGCα₁ KO mice with an exon 6 deletion, Sips et al. (2005) observed a sex-specific phenotype, with male KO mice showing testosterone-dependent hypertension. Both male and female KO mice still showed blood pressure decreases in response to NO donors but not to the direct sGC stimulator BAY 41-2272, so it was suggested that NO might modulate blood pressure via sGC-independent mechanisms. In the proximal stomach, nearly all differences between KO and WT mice were present in both sexes; the relaxant effect of both endogenous and exogenous NO and of BAY 41-2272 was decreased in KO mice but still present, sensitive to sGC inhibition and accompanied by a cGMP rise suggesting that NO is able to act via sGCα₂ in the stomach. Also at the level of the pylorus, responses to exogenous and endogenous NO were decreased in KO mice but still sensitive to sGC inhibition. However the decrease in NO-induced responses was more pronounced in female KO mice, with implications for gastric emptying (see below).

In contrast to NOS-1 (Huang et al. 1993) and cGKI KO mice (Ny et al. 2000), the sGCα₁ KO mice did not show an enlarged stomach. The influence of knocking out sGCα₁ is thus clearly less dramatic for gastric motility. This might be related to a partially maintained relaxant response to endogenous NO. We found no up-regulation of sGCα₂ in the proximal stomach but the normal sGCα₂ amount might be sufficient to obtain a rather well preserved nitrergic response in the stomach, as is suggested by the results with exogenous NO in the gastric fundus strips. The sGC inhibitor ODQ was tested versus exogenous NO in a concentration of 10 μmol l⁻¹. This concentration of ODQ has been shown to produce maximal inhibition of wild-type sGCα₁β₁ expressed in COS-7 cells (Schrammel et al. 1996) and to abolish the relaxant responses to the NO donor sodium nitroprusside in gastric fundus longitudinal muscle strips of Balb/C mice (Selemidis & Cocks, 2000). In the circular muscle strips of WT mice in this study, ODQ nearly abolished the relaxant response to 1 μmol l⁻¹ NO but only reduced the responses to 10 and 100 μmol l⁻¹ NO. ODQ is a competitive inhibitor of sGC (Schrammel et al. 1996) and 10 μmol l⁻¹ ODQ might not be sufficient to fully antagonize the sGC activation by higher concentrations of NO. However, while the relaxation by 10 μmol l⁻¹ NO was only halved by ODQ in WT strips, the increase in cGMP levels by this concentration of NO was nearly fully prevented by ODQ. This illustrates that the correlation between functional nitrergic relaxation on the one hand and measurable cGMP levels on the other hand is not linear. It has been shown before that nitrergic relaxations of equal size do not systematically induce a similar degree of cGMP increase (Garcia-Pascual & Triguero, 1994). Alternatively, cGMP-independent relaxation of gastrointestinal smooth muscle by NO via SK_Ca-channel activation has been reported (Martins et al. 1995; Watson et al. 1996) but the SK_Ca-channel blocker apamin did not influence the response to NO. In KO mice, the responses to NO were only moderately decreased and were even more sensitive to 10 μmol l⁻¹ ODQ. This suggests that NO can relax gastric fundus strips of KO mice through activation of sGCα₂, as corroborated by the NO-induced increase in cGMP levels. The observation that the moderate increase in cGMP levels by NO in KO mice induces nearly the same degree of relaxation as the more than 30-fold increase in WT mice suggests that the subcellular localization of cGMP produced by sGCα₂ in KO strips may differ from that by activation of sGCα₁, leading to more difficult recovery of cGMP when homogenizing the tissues for cGMP analysis. A different subcellular localization of the two heterodimers has indeed been reported with the α₁β₁ heterodimer being found in the cytosol and recruited to the membrane upon activation by an elevation of intracellular free calcium concentration while the α₂β₁ heterodimer is membrane-associated in a postsynaptic density 95 (PSD-95)-dependent manner in brain (Russwurm et al. 2001; Zabel et al. 2002). sGCα₂β₁ activation might also lead to cGMP generation in closer proximity of its relaxant effectors, which might also explain the same degree of relaxation via a smaller amount of measurable cGMP. Our results in sGCα₁ KO mice are similar to those of Mergia et al. (2006), who recently reported the cardiovascular phenotype in sGCα₁ KO mice with an exon 4 deletion. Because Mergia et al. started their numbering from the first coding exon while Sips et al. (2005) included the first 5′ noncoding exon in their numbering, exon 4 of Mergia et al. corresponds with exon 5 in the exon count of Sips et al. In the sGCα₁ KO mice of Mergia et al. the vasorelaxant response of the aorta to NO donors was well maintained and sensitive to ODQ, although sGCα₂, representing only 6% of total sGC content in WT aorta, was not up-regulated. No gastrointestinal data were reported for the exon 4-deleted sGCα₁ KO mice by Mergia et al. (2006). However, the same group has now reported that sGCβ₁ KO mice, which lack sGC activity and no longer show aortic relaxation to NO donors, had a severe gastrointestinal phenotype with enlarged caecum and increased whole-gut transit time (Friebe et al. 2007). In vitro data in gastrointestinal tissues of these sGCβ₁ KO mice were very recently reported preliminarily: in the gastric fundus of these mice, the relaxant response to the NO donor DEA-NO was abolished below a concentration of 30 μmol l⁻¹ (Groneberg et al. 2007). This corroborates that exogenous NO is only acting through sGC in the gastric fundus and that the partially maintained relaxations to exogenous NO in the sGCα₁ KO mice in our study must be related to sGCα₂ activity.

The results with EFS indicate that also endogenous NO can relax gastric fundus strips through sGCα₂ in sGCα₁ KO mice, at least when released for a sufficiently long time. Relaxations induced by EFS with 10 s trains were abolished by l-NAME and ODQ in WT mice indicating NOS-1 derived NO release and activation of sGC. Still, no cGMP rise could be detected upon EFS for 10 s suggesting that the subcellular localization of cGMP induced by release of endogenous NO along the nerve varicosities upon EFS differs from that induced by exogenous NO diffusing into the smooth muscle strip. Major differences between EFS- and NO-induced cGMP increases in gastrointestinal smooth muscle preparations have been reported before (Smits & Lefebvre, 1996). In sGCα₁ KO mice, relaxant responses to 10 s trains of EFS were very small or absent, suggesting that sGCα₁ is essential for a fast nitrergic response. As a non-nitrergic component in EFS-induced relaxation has been observed in previous studies in mice, especially at higher frequencies and longer stimulation trains (Ny et al. 2000; Dick et al. 2002), EFS was also performed with 60 s trains. In the WT mice of this study, the relaxations by EFS at 1–8 Hz were completely abolished by l-NAME and ODQ, indicating that they are fully nitrergic; this is also supported by the l-NAME- and ODQ-sensitive rise in cGMP measured for EFS at 16 Hz, although PDE inhibition was required to detect this. The contractions occurring during EFS in the presence of l-NAME or ODQ indicate the release of (a) non-cholinergic contractile neurotransmitter(s), most probably tachykinins (Pheng et al. 1997). With EFS at 16 Hz, relaxation still occurred in the presence of l-NAME or ODQ, which might correlate with the release of a non-nitrergic neurotransmitter such as VIP (Li & Rand, 1990; D'Amato et al. 1992). In sGCα₁ KO mice, relaxant responses still occurred with 60 s trains of EFS (2–16 Hz) and those at 2–8 Hz were abolished by l-NAME and ODQ similar to WT mice. This suggests that more prolonged release of endogenous NO can also lead to activation of sGCα₂ in sGCα₁ KO mice; this is supported by the modest but significant increase in cGMP upon EFS at 16 Hz. A final argument for a significant role of sGCα₂ in sGCα₁ KO mice was obtained with BAY 41-2272, a NO-independent activator of both sGCα₁β₂ and sGCα₂β₁ (Koglin et al. 2002). The relaxant response to BAY 41-2272, was clearly reduced but not abolished by ODQ. The cGMP-dependent component was illustrated by the rise in cGMP levels with 10 μmol l⁻¹ BAY 41-2272, which was abolished by ODQ. The non-cGMP-dependent component might be related to stimulation of sarcolemmal sodium pumps (Bawankule et al. 2005), inhibition of Ca²⁺ entry or stimulation of protein phosphatases (Teixeira et al. 2006), as proposed in vascular preparations. In sGCα₁ KO mice, the relaxant response to BAY 41-2272 was reduced, but the remaining response was partially ODQ-sensitive, and accompanied by a rise in cGMP, which was abolished by ODQ.

No adaptive changes occurred in the relaxant pathway downstream of sGC, in cAMP-induced relaxations or in contractile mechanisms. The relaxant response to the cell-permeant cGMP analogue 8-Br-cGMP was not different between WT and KO mice, which differs from observations in the sGCα₁ KO mice with exon 4 deletion, where increased sensitivity to a cGMP analogue in vascular tissue was reported (Mergia et al. 2006). Relaxation by VIP, which acts via activation of adenylate cyclase (AC; Simon & Kather, 1978; Laburthe & Couvineau, 2002), was well maintained in sGCα₁ KO mice. Also cGKI KO mice exhibit no compensatory up-regulation of the VIP–AC–cAMP relaxant pathway (Ny et al. 2000; Bonnevier et al. 2004). Contractions to both carbachol and PGF2α were not changed in sGCα₁ KO mice; this corresponds to the maintained contractile response to carbachol in small intestine and to endothelin in gastric fundus of cGKI KO mice (Ny et al. 2000; Bonnevier et al. 2004).

The ability of endogenous NO to act via sGCα₂ in sGCα₁ KO mice probably contributes to the moderate implication of knocking out sGCα₁ on in vivo gastric emptying. As we are particularly interested in the consequences of knocking out sGCα₁ on nitrergic relaxation in the gastric fundus and, at least in humans, gastric fundus activity is mainly important for liquid emptying (Minami & McCallum, 1984), gastric emptying of a phenol red solution was studied. Nitrergic relaxation in the fundus is important for gastric accommodation (Desai et al. 1991) and enhanced nitrergic proximal gastric relaxation will lead to a longer stay of liquids in the stomach, as has been shown in humans by administration of the NO-donor nitroglycerin or by inhibition of cGMP breakdown with the PDE-5 inhibitor sildenafil (Sun et al. 1998; Sarnelli et al. 2004). Conversely, reduction of the nitrergic storage capacity in the fundus is thus expected to accelerate liquid emptying as was observed in the male sGCα₁ KO mice 15 min after gavage. However, most studies investigating the influence of NOS inhibitors on gastric emptying report that NOS inhibition delays gastric liquid emptying (Plourde et al. 1994; Tanaka et al. 2005), which can be ascribed to increased pyloric tone by inhibition of nitrergic relaxation in the pylorus; this will counteract the stimulating effect on liquid emptying of decreased gastric accommodation capacity (Anvari et al. 1998). Also NOS-1 and cGKI KO mice show delayed gastric liquid emptying (Pfeifer et al. 1998; Mashimo et al. 2000) corresponding with pyloric hypertrophy (Huang et al. 1993) leading to gastric dilatation (Huang et al. 1993; Ny et al. 2000). Although an increased pyloric weight was measured in female sGCα₁ KO mice, the sGCα₁ KO mice did not show gastric enlargement pointing to less important consequences at the pyloric level of knocking out sGCα₁ than of NOS-1 and cGKI. Still, in vitro, nitrergic relaxation in the pylorus was decreased in sGCα₁ KO mice. The decrease was more pronounced in female mice, influencing the balance between the two opposite forces controlling liquid emptying (propulsion by enhanced fundic pressure and resistance by enhanced pyloric tone). In female mice, liquid gastric emptying will be more opposed at the level of the pylorus, which probably explains why gastric emptying at 15 min was not increased as in male mice. The observation that gastric emptying in male sGCα₁ KO mice is no longer significantly increased when measured at 60 min after gavage might be related to a less important influence of reduced fundic storage capacity on the much smaller amount that has to be emptied between 15 and 60 min than within the first 15 min.

In conclusion, sGCα₁β₁ plays an important role in gastric nitrergic relaxation in vitro, but some degree of nitrergic relaxation can occur via sGCα₂β₁ activation in sGCα₁ KO mice, which contributes to the moderate in vivo consequence on gastric emptying.