1Basal electrogenic Cl− secretion, measured as the short-circuit current (Isc), was variable in ileum removed from tetrahydrobiopterin (BH4)-deficient hph-1 mice and wild-type controls in vitro, although values were not significantly different.
2The basal nitrite release and mucosal cyclic guanosine 3′,5′-monophosphate (cyclic GMP) production were similar in control and BH4-deficient ileum.
3Mucosally added Escherichia coli heat-stable toxin (STa, 55 ng ml−1) increased the nitrite release, cyclic GMP levels and the Isc in control ileum, but its secretory actions were reduced in BH4-deficient ileum.
4L-Arginine (1 mM) increased the nitrite release, cyclic GMP production and the Isc in control ileum, but the actions were reduced in BH4-deficient ileum.
5Serosal carbachol (1 mM) stimulated maximum short-circuit currents of similar magnitude in both control and BH4-deficient ileum, whilst nitrite release and cyclic GMP production were minimal.
6E. coli STa and L-arginine increased electrogenic Cl− secretion across intact mouse ileum in vitro by releasing nitric oxide and elevating mucosal cyclic GMP. The inhibition of these processes in the hph-1 mouse ileum suggests that BH4 may be a target for the modulation of electrogenic transport, and highlight the complexity of the interactions between nitric oxide and cyclic GMP in the gut.
Because increasing evidence implicates NO as a regulator of intestinal ion transport, we have studied its putative role by using the hph-1 mouse. The metabolism of GTP-cyclo-hydrolase is impaired in the hph-1 mouse, which results in a deficiency of the NO synthase cofactor tetrahydrobiopterin (BH4; McDonald & Bode, 1988). BH4 is an essential cofactor for all isoforms of NO synthase and regulates the synthesis of NO from L-arginine (Griscavage, Fukuto, Komori & Ignarro, 1994). The secretory actions of E coli STa, the NO substrate L-arginine and the cholinergic agonist carbachol, were assessed in hph-1 mice and wild-type controls. Preliminary data have been presented to The Physiological Society (Rolfe, Brand, Heales, Lindley & Milla, 1995).
Measurement of electrogenic ion secretion in vitro
Albino male control (CD1) mice, BH4-deficient (hph-1) mice and wild-type controls (C57BLxCBA) were used (aged 30 days old). Mice were housed at constant temperature (23 ± 1 °C) and humidity (72%) with a 12 h light and 12h dark cycle, and had free access to food (Special Diet Services, Witham, Essex, UK) and water. On the day of use, mice were anaesthetized with Sagatal (60 mg (kg body weight)−1, I.P.), and a mid-line incision made along the abdomen. The ileum was removed at a distance of 4 cm proximal to the ileo-caecal valve and flushed with 0.9% saline. Tissue was mounted as intact flat sheets in Ussing-type chambers over an aperture of 0.33 cm2. The chambers were filled with Krebs bicarbonate saline (KBS, pH 7.4) containing the following concentrations of ions (mM): 143.0 Na+, 125.7 Cl−, 24.9 HCO3−, 5.9 K+, 2.5 Ca2+, 1.2 H2SO4, 1.2 SO42- and 1.2 Mg2+ and gassed with 95% 02 and 5% CO2 at 37 °C. The serosal bathing solution contained 10 mM glucose for metabolic requirements and the mucosal solution contained 10 mM mannitol as an osmotic balance. The short-circuit current (Isc) across the tissue was monitored by an automatic voltage clamp (DVC 1000, WPI Instruments, Stevenage, UK). The maximum change in Isc (▵Isc,max= maximum Isc– basal Isc) was used as an index of electrogenic ion secretion. Mice were killed by anaesthetic overdose and incision of the heart, after the removal of intestine for the in vitro experiments.
Measurement of nitrite release
The release of nitrite into the serosal and mucosal bathing solutions was measured (as an index of NO synthesized) by the Greiss reaction (1 ml sample + 0.5 ml 1.0% (w/v) sulphanilamide in 1.0 M HCl + 0.5ml 0.1% (w/v) napthylethylamine diethylamine in distilled H2O after 5 min; from Archer, 1993). The nitrite content was measured by spectrophotometer (Cecil CE2040) at a wavelength of 540 nm, and concentrations were calibrated from standard NaN02 curves. Mucosal nitrite release was unchanged throughout the experiments, so all data will describe changes in serosal nitrite release only.
Measurement of cyclic GMP levels
Sections of ileum approximately 2 cm long were incubated in KBS containing either STa, L-arginine or carbachol, and gassed at 37 °C for 10 min. The mucosa was removed by scraping the surface with a glass microscope slide. The mucosa was homogenized in perchloric acid (0.1 mM) containing 3-isobutyl-1-methylxanthine (1 mM), and centrifuged for 2 min at 14000 g. The cyclic nucleotide content in the supernatant was measured by scintillation proximity assay (125I-labelled cGMP; Amersham, UK).
Ileum tetrahydrobiopterin content
Sections of mouse ileum (2 cm) were homogenized in perchloric acid (0.1 mM) and centrifuged at 14000g for 2 min. The BH4 content was measured in the supernatant by high pressure liquid chromatography, and BH4 concentrations calibrated from standard curves.
The results are expressed as the means ±s.e.m., followed by the numbers of mice used. Statistical significance was tested using Student's unpaired ? test, and multiple comparisons were made using Bonferroni's correction. Significance was assumed at P < 0.05. It should be noted that the mean Isc data plotted as time course profiles or concentration–response curves are obtained from single experiments whose maximum responses occur at different times, so the mean maximum Isc obtained from these data may vary from the maximum values shown in the text.
Sagatal was obtained from May & Baker (Dagenham, Essex, UK) whilst all other chemicals were obtained from Sigma (Poole, Dorset, UK).
Effects of Cl− removal on STa, l-arginine and carbachol action in control (CD1) ileum
The effects on Isc of adding cumulative concentrations of STa to the mucosal solution, and l-arginine and carbachol to the serosal incubation solution are shown in Fig. 1. The effects on Igc of substituting gluconate for Cl− was assessed. The substitution of gluconate for Cl− in the Krebs bicarbonate buffer inhibited the maximum Isc generated by STa by 94% (Fig. 1A, n= 4, P < 0.002), that by l-arginine by 72% (Fig. 1B, n= 6, P < 0.05) and that by carbachol by 66% (Fig. 1C, n=6, P < 0.02). The basal ISC was also significantly inhibited by the removal of Cl− (control 38 ± 6 μA cm−2versus Cl−-free 12 ± 2 μA cm−2, n= 10, P < 0.001). This confirms that the majority of the basal Iscand stimulated secretory Isc is carried by the Cl− ion.
Basal activity of control (C57BLxCBA) and BH4− deficient (hph-1) ilea
The basal Isc was variable between control and BH4- deficient ileum used for the transport studies, but the mean basal values were similar in both and not significantly different (control 23 ± 5 μA cm−2 (n= 24) versus BH4 deficient 28 ± 3 μA cm−2 (n= 21), P < 0.05). The release of nitrite into the serosal bathing solution and mucosal cyclic GMP levels were also not significantly different between control and BH4-deficient ileum (nitrite release: control 0.13 ± 0.02 μM (n= 24) versus BH4 deficient 0.09 ± 0.02 μM (n= 21), P > 0.05; cGMP: control 98.1 ± 13.4 fmol(μg protein)−1 (n= 6) versus BH4 deficient 99.7 ± 18.2 fmol (μg protein)−1(n= 6), P > 0.05).
Effects of STa, l-arginine and carbachol on the Iscand nitrite release in control and BH4-deficient ilea STa.
The effects of STa (55 ng ml−1 EC50 value) placed in the mucosal bathing solution of control and BH4-deficient ilea are shown in Fig. 2A. The value of ▵Isc,max induced by STa in control ilea was reduced significantly (P < 0.05) by 74% in BH4-deficient ilea (control 13 ± 2 μA cm−2 (n= 8) versus BH4 deficient 3 ± 1 μA cm−2 (n= 5)). The release of nitrite into the serosal bathing solution 5 min after the addition of STa was significantly reduced (P < 0.05) in BH4-deficient intestine (control 0.26 ± 0.01 μm (n= 8) versus BH4 deficient 0.05 ± 0.02 μm (n= 5)).
The values of ▵Isc,max generated by l-arginine (1 mM) placed in the serosal solutions were reduced significantly (P < 0.01) in BH4-deficient ilea (Fig. 2B, control 16 ± 3 μA cm−2 (n= 16) versus BH4 deficient 2 ± 1 μA cm−2 (n= 8)), and the serosal nitrite release was suppressed (control 0.50 ± 0.07 μM (n= 8) versus BH4 deficient 0.02 ± 0.01 μM (n= 8), P < 0.01).
The effects of carbachol (1 mM) placed in the serosal bathing solution of control and BH4-deficient ilea are shown in Pig. 2C. Carbachol induced similar values of ▵Isc,max (control 35 ± 2 μA cm−2 (n= 13) versus BH4 deficient 32 ± 6 μA cm−2 (n= 9), P > 0.05) and minimal nitrite release (control 0.01 ± 0.001 μM (n= 8) versus BH4 deficient 0.04 ± 0.03 μM (n= 8)), which were not significantly different (P > 0.05).
Efiects of STa, l-arginine and carbachol on mucosal cyclic GMP levels
Since electrogenic Cl− secretion activated by STa and l-arginine was dramatically reduced in BH4-deficient rat ileum, and cyclic GMP is an intracellular mediator of Cl− secretion induced by E. coli STa, the influence of BH4 deficiency on mucosal cyclic GMP levels was tested. The effects of STa, l-arginine and carbachol placed in the bicarbonate buffer used to incubate the ilea, on mucosal cyclic GMP, is shown in Fig. 3. The STa (55 ng ml−1) elicited an increase in mucosal cyclic GMP levels in control gut mucosa, but levels were reduced significantly in BH4-deficient ileum by 46% (n= 6, P < 0.05). Similarly, l-arginine (1 mM) elevated cyclic GMP levels in control ilea, but levels were reduced significantly in BH4-deficient ilea by 66% (n= 6, P < 0.05). In contrast, the cyclic GMP levels in control and BH4-deficient ileum exposed to carbachol were minimal (n= 6, P > 0.05).
Ileal BH4 content
The ileal BH4 content measured by HPLC was significantly reduced in hph-1 mice compared with wild-type controls (C57BLxCBA 0.64 ± 0.04 pmol (μg protein)−1 (n= 4) versus hph-1 0.20 ± 0.02 pmol (μg protein)−1 (n= 4), P < 0.001).
Since the mechanisms by which nitric oxide (NO) regulates intestinal transport are unclear, the secretory actions of NO in the hph-1 mutant mouse, which is deficient in the nitric oxide synthase (NOS) cofactor tetrahydrobiopterin (BH4), has been studied. Ileal concentrations of BH4 were reduced significantly in hph-1 mice (by 69%) compared with wild-type controls, whilst a 50% reduction in BH4 was observed in brain tissue from hph-1 mice (Brand, Heales, Land & Clark, 1995). The localization of BH4 in the gut is not known and may include a variety of cells within the enteric nervous system, smooth musculature, lamina propria and epithelium, and this diversity of cell types may account for the variability of the BH4 levels between the gut and the brain. The basal Isc values of ileum from wild-type control mice and BH4-deficient mice fluctuated during the course of experimentation. Although the mean basal Isc values of the two groups were not significantly different, it is not clear whether the drift may be a result of variation within the strains of mice, or a result of the BH4 deficiency. Hence no conclusions regarding the nature of the basal current can be made. The basal cyclic GMP content of the mucosa was similar in BH4-deficient and control ileum. Therefore, BH4 is not involved in the generation of basal electrogenic ion secretion or the regulation of cyclic GMP metabolism in mouse ileum.
The experiments of this study with STa and l-arginine show that BH4 is an important regulator of intestinal electrogenic transport in vitro. Enterotoxin STa added to the mucosal surface and L-arginine added to the serosal surface induces electrogenic Cl− secretion in control mouse ileum, but it is dramatically reduced in both the magnitude and duration of action in BH4-deficient ileum. The reduction in secretory activity is accompanied by a decrease in nitrite release, indicating that NO synthesis was suppressed, presumably as a result of the inactivation of NOS. Furthermore, both STa and l-arginine increased cyclic GMP levels in control ileum, whilst in BH4-deficient ileum the cyclic GMP level in the mucosa was minimal. The depression of the action of STa and l-arginine in BH4-deficient ileum suggests that electrogenic Cl− secretion is NO dependent, and the data of this study show that the secretory action involves a cyclic GMP-signalling pathway. The actions of carbachol on eliciting electrogenic secretion were similar in ilea from control and BH4-deficient mice, and both nitrite release and cyclic GMP levels were minimal. This highlights the specificity of BH4 for NO-dependent secretory pathways.
Our data suggest that STa and L-arginine stimulate the release of NO, which in turn increases cyclic GMP production. However, the cellular sources of BH4 that are altered in hph-1 mice are not known, but are presumably colocalized with NOS. NOS isotypes have been identified in the enteric nerves (constitutive, NOS I), in epithelial and inflammatory cells (inducible synthase, NOS II) and vascular endothelium (constitutive, NOS III) (Schini-Kerth & Vanhoutte, 1995; Alican & Kubes, 1996). In our experiments we cannot conclude whether the NO generated by l-arginine or STa was of epithelial or neural origin. However, in the rat, STa elicts an indirect secretory action via local nitrinergic reflexes (Rolfe & Levin, 1994), thus if the NO-dependent element of this neural reflex is sensitive to BH4 deficiency, this would account for the reduction in electrogenic Cl− secretion observed in the mouse in vitro. Furthermore, if NO is released as an efferent neurotransmitter from the myenteric plexus, it would trigger cyclic GMP-dependent secretion by acting on the enterocytes. This would account for why in the BH4-deficient mouse, the cyclic GMP levels were not altered by exposure to STa. It is surprising that STa did not activate guanylate cyclase and raise intra-cellular cyclic GMP, which is a well-established primary mechanism of action in the rabbit (Field et al. 1978). This may be due to impaired guanylate cyclase–cyclic GMP signalling in the BH4-deficient mouse, or species variation.
The receptor targets for NO in the intestine are purely speculative, although soluble guanylate cyclase (GC) is a primary target for NO in many mammalian cells. In the intestine, GC is predominantly particulate (membrane bound; De Jonge, 1975), but the nature of the interactions between NO and particulate GC is not known. The NO-donating compound sodium nitroprusside stimulated electrogenic secretion in rat ileum, which was sensitive to cystamine (Rolfe & Levin, 1994), an inhibitor of particulate guanylate cyclase (Brandwein, Lewicki & Murad, 1981). This suggests that NO may indeed interact with guanylate cyclase in the gut. These interactions may involve alteration of the redox state, in order for NO to activate guanylate cyclase, as demonstrated in a variety of cell types (Murad, Mittal, Arnold, Katsuki & Kimura, 1978). We suggest, therefore, that NO may activate particulate guanylate cyclases for the induction of cGMP-dependent secretion in the mouse intestine.
Although BH4 is a cofactor for the synthesis of biogenic amines such as 5-hydroxytryptamine (5-HT), it is unlikely that the transport changes observed in our current experiments were a result of a reduction in 5-HT release. Peterson & Whipp (1995) found that STa did not stimulate 5-HT release in pig small intestine, whereas cholera toxin and STb were effective. In vivo studies of fluid secretion in the rat small intestine showed that the 5-HT receptor antagonist granisetron was ineffective in reducing STa-induced secretion, but its effectiveness was tested by its inhibition of cholera toxin (Mourad et al. 1995). In tissue desensitized to 5-HT in vitro, STa induced ion transport independently of 5-HT in the rat (Rolfe, Levin & Young, 1991). Beubler, Badhri & Schirgi-Degen (1992), however, speculated that 5-HT released by STa may result in fluid secretion in the small intestine of the rat in vivo.
In conclusion, this study has been shown that STa and l-arginine stimulate cyclic GMP-dependent electrogenic Cl− secretion across intact mouse ileum in vitro which involves release of nitric oxide. These actions are suppressed in ileum deficient in the NOS cofactor BH4, suggesting that BH4 may provide a target for the modulation of intestinal ion transport function.
V.R. was in receipt of sponsorship from the National Association of Crohn's and Colitis, and this work was partly funded by the University Central Funding Scheme, University College London.