Previously, we demonstrated specific 3H-bradykinin receptor-binding sites in guinea pig ileum17. The potencies of peptides at these binding sites correlated with their potencies in contracting smooth muscle, suggesting that the binding sites might be localized to muscle. The results shown in Table 1 demonstrate that specific bradykinin receptor-binding sites occur in the mucosa as well as in the muscle area. The relative potencies of bradykinin and six analogues listed in Table 1 are quite similar in both mucosa and muscle and correspond well with those observed previously in the total ileal tissue17. The 25-fold-enhanced potencies of the peptides observed in this study, compared to our earlier work, are due to enhanced specific activity and diminished degradation of the 3H-bradykinin. The relative potencies of bradykinin and its analogues at the 3H-bradykinin receptor-binding site closely parallels their potency in stimulating active chloride secretion. Removal of the arginine in the 9 position, forming Des-Arg9-bradykinin, destroys receptor-binding affinity and eliminates biological activity. Replacement of proline in position 3 with a-aminoisobutyric acid (AIB), forming AIB3-bradykinin reduces receptor-binding potency markedly in parallel with a decrease in biological activity. In contrast to internal changes in the bradykinin structure, addition of lysine to the amino-terminal only reduces the receptor-binding and secretory activity slightly. While Met-Lys-bradykinin is 30-50% as potent as bradykinin in affecting short-circuit current, it is only 10% as potent at 3H-bradykinin receptor-binding sites. It is possible that in the short-circuit current assay. Met-Lys-bradykinin is metabolized to the more potent Lys-bradykinin while conditions of the binding assay retard such metabolism (see Table 1 legend). The localization of bradykinin receptors to both the mucosa and muscle layers of the ileum was confirmed by light microscopic autoradiographic studies (Fig. 1). Autoradiographic grains were apparent in highest density over the lamina propria surrounding both villus and crypt epithelial cells. Grains were also concentrated over the circular and longitudinal muscle layers. By contrast, grain density was sparse or absent over lymphatic tissue, as well as the myenteric ganglia. Inclusion of 1 jxM unlabelled bradykinin in incubations greatly reduced grain density with the remaining grains being distributed diffusely. The pharmacological specificity was also indicated by biochemical experiments using the six bradykinin analogues listed in Table 1 and slide-mounted tissue sections. After incubation and washing, the tissue section was wiped off the slide with a circle of filter paper and the bound radioactivity assessed by scintillation counting. In these experiments the affinity of bradykinin receptors was the same as in homogenate estimations. Also, 3H-bradykinin binding is displaced by peptides with the same potencies as in homogenate experiments.Fig. 1 Autoradiographic localization of 3H-bradykinin (BK) receptorbinding in guinea pig ileum. The dark-field photomicrograph (a) shows the autoradiographic grain distribution over the same tissue as the bright-field photomicrograph (b). Note that grains are concentrated over muscle layers () and also over the villi (between double arrows). Grains are almost entirely localized over the centre or lamina propria of the villi and are sparse or absent over the luminal surface of the epithelium. Grains are also absent inside the crypts of Leiberkuhn (single arrows) but are concentrated over the serosal side of the crypt epithelial cells. The dark-field photomicrograph (c) shows the absence of specific receptor binding in an adjacent 10-n,M section when 1 u,M unlabelled BK was included in the incubation buffer. The coverslip method of Young and Kuhar33 was used. Guinea pigs were perfused through the heart with 500ml of 0.1% formaldehyde in a mixture of equal parts of phosphate-buffered saline and 20% sucrose, pH 7.4. The ilea were then removed, embedded in homogenized calf cerebral cortex and flash-frozen in liquid nitrogen. Sections (10 jjim thick) were cut in a Cryostat at -12 C, thaw-mounted onto chrome alum gelatin subbed microscope slides and stored frozen at -20C. Slides were warmed to room temperature, then incubated for 120 min at 4 C in medium containing 25 mM trimethylaminoethanesulphonic acid (TES) pH 6.8, 1 jiM 1,10-phenanthroline, 0.2% bovine serum albumin (BSA) (protease-free), 1 mM dithiothreitol, 0.1 mM bacitracin, 1 jxM captopril (SQ 14,225), 300 mM sucrose, and 0.2 nM 3H-BK (52 Ci mmoF1). Blanks were incubated in the same medium with the addition of 1 jxM unlabelled BK. After incubation tissue sections were washed in 25 mM TES pH 6.8, and 1 mM 1,10-phenanthroline for 60 min (2 x 30 min) at 4 C. Sections were then dried rapidly in cold, dry air, NTB-2 emulsion-coated coverslips were apposed and the sections were exposed to the emulsion for 1 month. Table 1 Relative potenciesof bradykinin (BK) and several related peptides in receptor binding and short-circuit current assays Stimulation of Inhibition ofshort-circuit 3H-BK binding current Mucosa Muscle layerKinin Structure EC50(nM) IC50(nM) IC50(nM) 123 45678 9Bradykinin (BK) H-Arg-Pro-Pro-Gly-Phe-Ser-Pro-Phe-Arg-OH 1.00.2 0.2 0.2 Kallidin (Lys-BK) Lys-Arg-Pro-Pro-Gly-Phe-Ser-Pro-Phe-Arg-OH 2.80.5 0.4 0.2 Met-Lys-GK Met-Lys-Arg-Pro-Pro-Gly-Phe-Ser-Pro-Phe-Arg-OH 2.60.8 1.8 2.2 AIB3-BK 183.044 5.0 5.0 Hydroxyproline3-BK 3.10.3 0.4 0.6 0-(2-Thienyl)-Ala58-BK 5.81.4 0.5 0.4 Des-Arg9-BK No response >106 >10"63H-BK was prepared (by NEN) by reduction of 2,3-dehydroproline-BK with tritium gas, resulting in a specific activity of either 27 or 52 Ci mmol. The radiochemical purity of the ligand was determined by HPLC both before and after binding assay. The HPLC system used separates all partial sequence fragments of BK including BK2_9, BKi_8 and BKi_7. A jxBondapak Ci8 (Waters Ass.) column (4 mm i.d. x 30 cm) was used with a linear gradient solvent system (a, H2O + 0.2% H3PO4, pH 2.4; 6, CH3CN + 0.2% H3PO4; 12% B-24% B) over 20 min with a flow rate of 2.0 ml min"1. Fractions (30-s) were collected and the radioactivity measured by liquid scintillation counting. Using this HPLC system we were able to monitor the metabolism of the 3H-ligand after the binding assay and thus develop a mixture of peptidase inhibitors effective in stabilizing BK. This increase in BK stability and use of a ligand of 2-4 times higher specific activity accounts for the almost 10-fold greater detected affinity of BK for its receptor than published previously17. In all assays mucosa was separated from the muscle layers by scraping gently with a glass slide. Crude membranes were prepared by homogenizing tissues in 20 vol 25 mM TES, pH 6.8 +1 mM 1,10-phenanthroline with a Brinkmann Polytron PT10 (setting 6 for 30 s). The homogenates were centrifuged twice at 50,000g for 10 min with an intermediate rehomogenization in buffer. For routine studies the final pellets were resuspended in 300 vol of incubation buffer (25 mM TES pH 6.8, 1 mM 1,10 phenanthroline, 0.2% BSA (Sigma; crystallized, lyophylized), 0.1 mM bacitracin, 1 mM dithiothreitol, 1 jxM captopril). To triplicate polypropylene incubation tubes were added (in a total volume of 4 ml); 850 u-1 of freshly resuspended tissue, 25,000 d.p.m. of H-BK (final concentration 0.05 nM) and competing agents. After incubation for 90 min at 4 C the samples were rapidly filtered over GF/B filters previously coated with 0.1% aqueous polyethyleneimine solution, then the filters were washed three times with 3 ml ice-cold buffer. Filter-bound radioactivity was determined by liquid scintillation counting at 40% efficiency in 10 ml Formula 947 (NEN). Typical c.p.m. were 1,000 total and 35 blank with a tissue concentration of 1 ml tissue wet weight per ml. For metabolism studies, samples were centrifuged after incubation and the supernatant analysed by HPLC. Emax; maximal effect; EC50, concentration to elicit 50% maximal effect relative to bradykinin Emax in stimulating /sc in guinea pig ileum; IC50, concentration of nonradioactive BK or BK analogue required to displace 50% of specifically bound H-BK from its receptor EC50 values represent means.e.m. Bradykinin receptors in the muscle layers presumably mediate the contractile effects of bradykinin. To investigate the possible function of the mucosal receptors, we evaluated the influence of bradykinin on the transepithelial potential difference (p.d.) and short-circuit current (/sc) of the mucosal layer of the guinea pig ileum using techniques previously described18 (Fig. 2, Table 1). Tissue conductance showed no appreciable change and therefore the change in /sc was proportional to the change in p.d. There was no effect on p.d. or /sc when bradykinin in concentrations up to 10 |xM was added to the mucosal bathing solution. In striking contrast, addition of nanomolar concentrations of bradykinin to the serosal surface rapidly increased p.d. and 7SC, with maximal effects within 2 min (Fig. 2). The finding that bradykinin acts only when added to the serosal surface agrees with the localization of bradykinin receptors to the serosal surface of villus and crypt epithelium. These effects resemble those produced in this tissue by other peptides such as vasoactive intestinal peptide19, substance P20, neurotensin20 and bombesin21.Table 2 Effect of bradykinin on chloride fluxes m-s CF fluxs -" m Net Control Bradykinin (105M) 8.20.8 8.30.4 8.70.8 -0.51.6 12.11.1-3.81.3Values are the jx equiv. cm 2 h xs.e.m. for six paired experiments. Two 10-min samples were taken starting 15-20 min after the addition of 36C1" and 5 min after adding bradykinin.P<0.025. Fluxes were measured with tissues short-circuited as described previously18. Flux is from mucosa to serosa (m-"s) or serosa to mucosa (s-"m). After quickly becoming maximal, the /sc diminished with a rather slower time course similar to that observed for the 'fading' of the smooth muscle-contacting effects of bradykinin4. Whether it is related to tachyphylaxis or peptide metabolism is unclear. In the absence and presence of the kininase II inhibitor, captopril, 50% of the maximal response occurs at 10 and 1.0 nM bradykinin, respectively. With this 10-fold enhancement of potency, bradykinin has a similar potency in altering 7SC and binding to receptors (Table 1).Several experimental results suggest that the bradykinin-induced increase in Jsc is due to stimulation of anion secretion. Thus, the effects of bradykinin on p.d. and 7SC are reduced by 86% when Cl and HCOs are replaced by sulphate (data not shown). Other evidence derives from direct measurement of unidirectional Cl fluxes across the short-circuited ileal mucosa in the presence and absence of bradykinin (Table 2). In control tissues there is a mean net Cl secretion of 0.5 jxequiv. cnT2 h"1. In the presence of bradykinin, net Cl secretion increases to -1-3.8 nequiv. cm2 h'1. These experiments used high concentrations of bradykinin to obtain maximal and prolonged effects. The effects of bradykinin do not seem to be mediated by influences on neuronal conduction, as they are not affected by tetrodotoxin (Table 3); calcium movements may be involved, as verapamil, a Ca2^ channel blocker, markedly inhibits the influences of bradykinin on /sc (Table 3). Bradykinin stimulation of /sc is also completely blocked following the addition of the Ca2+ channel blocker CoCl2 (1 mM). A possible role of Ca2+ agrees with data that Ca2+ regulates intestinal secretion22 and bradykinin receptor binding17.It also seems that the effects of bradykinin on CP secretion may be mediated by arachidonic acid and its metabolites. We have found that the action of bradykinin can be inhibited by a variety of agents that act as inhibitors of cyclooxygenase (for example, indomethecin) and phospholipase A2 (for example, mepacrine). Moreover, exogenously added arachidonic acid or prostaglandin E2 (PGE2) are both powerful secretory stimuli and bradykinin can stimulate the in vitro release of PGE2 (ref. 23). Interestingly, phospholipase A2, the enzyme that releases arachidonic acid from membrane phospholipids, is dependent on calcium. Conceivably, bradykinin alters the availability of calcium to phospholipase A2 by opening calcium channels as indicated by the inhibitory effects of Ca2+ channel blockers. The present results suggest that bradykinin may have both a physiological and pathological role in the intestine. Our in vitro findings are consistent with the observations of Hardcastle et at.15 that bradykinin causes a transient increase in trans-colonic and transjejunal p.d. in the rat. Independently Cuthbert and Margolius24'25 have also found potent enhancement of gut chloride secretion by kallidin and have also suggested a role for prostaglandins. The bradykinin receptors in the crypts might regulate this anion secretion. Bradykinin may enhance intestinal absorption of amino acids and glucose25'26, effects that could involve bradykinin receptors in the tips of the intestinal villi where such absorption takes place. Autoradiography at the light microscopic level does not allow the precise determination of the cellular location of receptors. Bradykinin receptors in the lamina propria are probably located on the basolateral membranes of the epithelial cells, however, receptors may also occur on smooth muscle cells, blood vessels and/or sensory nerve fibres in the villi.Excessive kinin activity might account for some pathological conditions. In carcinoid syndrome, serotonin may be responsible for some of the symptoms5"7. However, although serotonin does elicit diarrhoea28, it causes vasoconstriction7 rather than the vasodilatation and the associated flushing, light-headedness and headache that typically occurs in carcinoid patients. Moreover, while some increase of blood level of serotonin occurs in certain patients with carcinoid syndrome, there is no correlation with the onset of their symptoms7. By contrast, increase blood levels of kinins have been detected in carcinoid patients in the course of their symptoms5"7. Fig. 2 Effect of BK on intestinal potential difference (p.d.). Ileal mucosa was obtained from female Hartley guinea pigs weighing 400-600 g. Segments of distal ileum, 10-15 cm long, were excised 5 cm above the ileo-caecal junction. The segments were cut along their mesenteric border and placed in oxygenated ice-cold Ringer's. After stripping off the serosa and underlying longitudinal muscle layer, each tissue was mounted between two Leucite half chambers, the exposed areas being 0.64cm2. The apparatus for measuring p.d. and /sc was as described previously 8. a, Effect of single dose of BK on transepithelial p.d. across guinea pig ileal mucosa. 6, Effect of increasing concentrations of BK analogue (ThiAla5'8) on p.d. Peptides were added at T- Tracings are from single experiments, but are characteristic of tracings from at least 10 other experiments.Kinins could also play a part in inflammatory bowel disease. In these conditions the plasma levels of kinins correlate very closely with the onset of vasomotor and gastrointestinal symptoms6'9. As inflammation is associated with the formation of kinins29, one would expect kinins to be released into the local or systemic circulation. Indeed, inflamed intestinal tissue from patients with ulcerative colitis contains abnormally high levels of active kallikrein, the kinin-releasing enzyme14. Plasma and tissue levels of peptidyl dipeptidase which degrades kinins, are depressed in patients with regional enteritis; this would elevate kinin levels30'31. Also, in regional enteritis plasma levels of a2-macroglobulin are reduced32. a2-Macroglobulin inhibits the bradykinin releasing enzyme kallikrein, thus reduced levels of the inhibitor should facilitate kinin synthesis. If indeed kinins are associated with symptoms of carcinoid syndrome or inflammatory bowel disease, then administration of the peptidyl dipeptidase inhibitor captopril to such patients may be contrain-dicated. Table 3 Effect of verapamil and tetrodotoxin on peak increments in 7SC elicited by bradykininResponse to 1.0 jjiM bradykinin Change in 7SC on drug (% of control addition (|xA cm )response) Tetrodotoxin Verapamil 4 4 -44.44.8 -24.65.2 119.417.1 29.910.3Values are means.e.m.; n, number of paired experiments. Negative values reflect decreases in 7SC.Plt;0.01 compared with control. We thank Lynda Hester for technical assistance and Dr John Stewart, University of Colorado, for the gift of the bradykinin analogues. This work was supported by USPHS grants DA 02121, AM 21345, DA 00266 and NS 16374, and grants from the National Foundation of Ileitis and Colitis, the McKnight Foundation and RSA award DA 00074 to S.H.S.