INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

CARBON MONOXIDE (CO) is a low molecular weight gas that is produced under physiological conditions (see Ref. 11 for review). There are at least two pathways by which CO is produced endogenously. A minor pathway is through NADPH-dependent enzymatic peroxidation of microsomal lipids, with the predominant source of endogenous CO being the NADPH-dependent oxidative heme destruction catalyzed by heme oxygenase (HO) isozymes (11). Cleavage of heme by HO results in the production of CO and biliverdin and release of iron. Two isoforms of HO have been identified and designated as HO-1 and HO-2 (14, 27). The two isoforms differ in their regulatory mechanisms. HO-1 is inducible by many agents and stimuli, including metal ions, heme, organic solvents, hormones, bacterial toxins, neoplasms, and alkylating agents (11, 14). The only inducers of HO-2 identified to date are adrenal glucocorticoids (30).

The potential role for CO as a cellular messenger has recently received considerable attention. CO is known to increase intracellular guanosine 3',5'-cyclic monophosphate (cGMP) levels in many cell types (1-3, 28). In the brain, CO has been postulated to play a role in long-term potentiation (LTP) and in the regulation of cGMP levels (13, 26, 28, 31). Inhibition of HO activity prevents the induction of LTP in guinea pig hippocampal slices, and perfusion of stimulated nerve fibers with CO enhances synaptic transmission (31). In mouse and rat hippocampal slices, inhibition of HO activity inhibits the induction of LTP and reverses established LTP (26). In olfactory neurons, inhibition of HO activity reduces endogenous levels of cGMP (28). In addition, in isolated rabbit corneal epithelial cells, CO increases the open probability of a non-Ca²⁺-dependent K⁺ channel and increases intracellular cGMP levels (22). Other systems in which CO function has been suggested include regulation of carotid body sensory activity in the rat and regulation of the release of corticotropin-releasing hormone and gonadotropin-releasing hormone (9, 18, 19).

There is also a growing body of evidence that suggests a role for CO as a messenger in the gastrointestinal tract. In the opossum, CO relaxes the internal anal sphincter and inhibition of HO suppresses nerve-evoked nonadrenergic, noncholinergic relaxation of the sphincter (21). In the feline lower esophageal sphincter, HO-2 is present in neuronal cell bodies and exogenous CO evokes a concentration-dependent relaxation of the sphincter (17). In human jejunal circular smooth muscle cells, exogenous CO stimulates an outward K⁺ current and a leak current, hyperpolarizes the membrane potential, and elicits transient membrane hyperpolarizations (7).

Central to the hypothesis that CO is a physiological messenger in smooth muscle is the ability of cells in the gastrointestinal tract to produce CO. CO may be produced from neurons, smooth muscle cells, or other cells normally found in the smooth muscle layers. CO produced from neurons may act as an autocrine messenger, regulating neuronal electrical activity and subsequently smooth muscle contractile activity, or as a paracrine messenger, acting on adjacent cells such as smooth muscle. Alternatively, CO may be produced by smooth muscle cells and may act as a local control mechanism for smooth muscle electrical and mechanical activity. Therefore, the aims of this study were to examine the distribution of HO in the canine jejunum and to determine the effects of exogenous CO on jejunal circular smooth muscle cells.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Immunostaining for HO. The use of canine jejunum was approved by the Institutional Animal Care and Use Committee. A piece of jejunum measuring ~1 × 1 cm, containing the entire thickness of the intestinal wall, was cut from the jejunum of four dogs just distal to the ligament of Treitz. The tissues were fixed by overnight immersion in 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4, at 4°C. The tissues were then rinsed several times in phosphate-buffered saline (PBS), immersed in 30% sucrose in PBS overnight at 4°C, and then quick-frozen at 40°C. From each tissue, 20 sections (20 µm thick) were cut on a cryostat, thawed, and mounted onto chrome alum-coated glass slides. In the sections, the longitudinal muscle layer was in cross section, and the circular muscle was cut tangential to its long axis. Alternate consecutive sections were immunostained for HO-1 or HO-2, using an indirect immunofluorescence procedure, as follows. All incubations were done at room temperature and in a moist chamber. Sections were preincubated with 10% normal goat serum (NGS) in PBS containing 0.3% Triton X-100 for 45 min, rinsed in PBS, and then incubated overnight in rabbit antisera to HO-1 or HO-2 diluted 1:500 to 1:1,000 in PBS containing 5% NGS and 0.3% Triton X-100. Sections were again rinsed in PBS and incubated with rhodamine-conjugated goat anti-rabbit immunoglobulin G diluted 1:80 in PBS containing 1.25% NGS and 0.3% Triton X-100 for 90 min. Sections were rinsed with PBS and then mounted and coverslipped in PBS-glycerol (1:1) containing an antifade reagent (Slow Fade; Molecular Probes, Eugene, OR) and examined with a Zeiss Axiophot fluorescence microscope equipped with a rhodamine filter set. As controls, nonimmune rabbit serum was used in place of the primary antiserum, and HO-2 antibody was preabsorbed with recombinant rat HO-2 protein (100 mg/ml). Immunolabeling was not observed in these controls. Preabsorption of HO-2 antibody with recombinant rat HO-1 protein did not diminish immunolabeling.

In situ hybridization. In situ hybridization for HO-1 and HO-2 was carried out as detailed by Ewing (4). Tissues were fixed and embedded in paraffin, and sections 5 µm thick were obtained. Tissue was cleared of paraffin with xylene and rehydrated by sequential equilibration with graded ethanol solutions (99-50%) and was then denatured by immersion in 0.2 N HCl for 20 min at 25°C. Tissue was enzymatically deproteinated by incubation (15 min, 37°C) with proteinase K and acetylated in 0.1 M triethanolamine, pH 8, containing 0.25% (vol/vol) acetic anhydride (5 min). Acetylated tissue was dehydrated with graded ethanol (50-90%) and then prehybridized in buffer containing 50% (vol/vol) deionized formamide, 4× SSC (1× SSC is 0.15 M NaCl and 0.015 M sodium citrate, pH 7.0), 1% Denhardt's solution, 0.25 mg/ml yeast tRNA, 10% dextran sulfate, and 0.5 mg/ml heat-denatured salmon sperm DNA. Tissue was hybridized with 2 ng/ml of the appropriate digoxigenin-labeled (sense and antisense) probe under parafilm coverslips in a humidified chamber for 16 h at 37°C.

Whole cell current measurements. Single isolated jejunal circular smooth muscle cells were obtained from adult mongrel dogs of either sex. Each dog was euthanized with an overdose of barbiturate (45 mg/kg), and a 10-cm piece of jejunum was removed just distal to the ligament of Treitz. The dissociation procedures used to obtain single relaxed circular smooth muscle cells were as previously described (6, 8). In brief, full-thickness strips of jejunum were pinned to the floor of a dissecting dish, and incisions were made parallel to the longitudinal muscle axis extending to, but not into, the circular muscle layer. The serosa and longitudinal muscle layer were removed, leaving the circular muscle and submucosa. Incisions were next made parallel to and through the circular muscle axis. Strips of circular muscle were gently peeled off of the submucosa, placed in modified Hanks' solution (Sigma H8389), and cut into 2-mm pieces. They were placed in 8 ml of Hanks' solution containing 15 mg of papain (Sigma P4762) and 3.1 mg of dithiothreitol (Sigma D0632) and gently stirred for 25 min at 37°C. After centrifugation to remove the enzyme solution, the tissue was transferred to fresh Hanks' solution and mechanically dissociated at 37°C to obtain single relaxed circular smooth muscle cells.

cGMP measurements. cGMP was measured using a radioimmunoassay technique as previously reported (22). Freshly isolated canine jejunal circular smooth muscle cells were counted, centrifuged for 7 min at 180 g, and resuspended in 1 ml of normal Ringer solution containing 0.1 mM 3-isobutyl-1-methylxanthine with or without 1% CO for 1 min. The cells were frozen in liquid nitrogen and homogenized, and 5% trichloroacetic acid was added to precipitate soluble protein. The samples were centrifuged at 1,500 g for 10 min, and the supernatant was then extracted with water-saturated ether four times to remove the trichloroacetic acid. Measurements were made in triplicate for each dilution.

Materials and solutions. CO was obtained from Scott Specialty Gases (Troy, MI). Quinidine, amphotericin B, oligo(dT) cellulose, deoxyribonuclease I, proteinase K, Triton X-100, yeast tRNA, dextran sulfate, 4-chloro-1-naphthol, and paraformaldehyde were obtained from Sigma (St. Louis, MO), 8-bromo-cGMP (8-BrcGMP) was from Boehringer Mannheim (Indianapolis, IN), and KT-5823 was from Calbiochem (San Diego, CA). Copper protoporphyrin IX (CuPP-IX) was obtained from Porphyrin Products (Logan, UT). HO-1 and HO-2 antisera and recombinant rat HO-1 and HO-2 protein were obtained from StressGen Biotechnology (Victoria, BC, Canada). Whole cell recordings were made using the following solution in the pipette (in mM): 25 KCl, 125 potassium methanesulfonate, 2.54 CaCl₂, and 5 N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid; pH was adjusted to 7.00. Smooth muscle cells were bathed in normal Krebs solution of the following composition (in mM): 137.4 Na⁺, 5.9 K⁺, 2.5 Ca²⁺, 1.2 Mg²⁺, 15.5 , 1.2 H₂PO₄, and 11.5 glucose, adjusted to pH 7.35.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Distribution of HO isozymes in the canine jejunum. Immunoreactivity for HO-2 was observed in a select population of nerve cell bodies of both myenteric and submucous plexuses in tracts interconnecting ganglia in a population of nerve fibers in the circular but not longitudinal muscle layers (Fig. 1). In the ganglia of the myenteric and submucosal plexuses some nerve cell bodies fluoresced intensely for HO-2, whereas others fluoresced more faintly or at background levels only. Because we did not use a marker to label all nerve cell bodies, we were unable to obtain an accurate percentage of nerve cell bodies containing HO-2. However, HO-2-containing nerve cell bodies were noted in all ganglia examined. Immunoreactivity for HO-2 was also observed in epithelial cells of the jejunal mucosa (data not shown). In contrast, no HO-1-immunoreactive structures were noted in the tissue preparations. HO-2 immunoreactivity was not observed in smooth muscle with the HO antibodies used.

View larger version (79K): [in this window] [in a new window]   Fig. 1.   Immunostaining for heme oxygenase-2 (HO-2) in full-thickness sections (except mucosa) of wall of canine jejunum. Longitudinal muscle (lm) layer was cut in cross section; circular muscle (cm) layer was cut tangentially. A: myenteric plexus. HO-2 immunoreactivity in a subpopulation of nerve cell bodies (arrows) and fibers interconnecting ganglia (arrowhead). B: HO-2 immunoreactivity in nerve fibers (arrowheads) in inner circular muscle layer. C: submucous plexus. HO-2 immunoreactivity in nerve cell bodies and nerve fiber strands interconnecting ganglia. Calibration bar (100 µm) applies to all panels.

View larger version (149K): [in this window] [in a new window]   Fig. 2.   In situ hybridization localization of HO-2 mRNA in canine jejunum. A and B: whole wall thickness showing positive staining for HO-2 mRNA using antisense probe (A) in circular smooth muscle layer and in outer region of longitudinal smooth muscle, and absence of staining using HO-2 sense probe (B). Magnification, ×4. C and D: higher magnification (×10) of circular smooth muscle layer shown in A and B. E and F: isolated circular smooth muscle cells (×100) showing positive staining using HO-2 antisense probe (E) and negative staining with HO-2 sense probe (F).

Effects of exogenous CO on isolated jejunal circular smooth muscle cells. CO (1%) increased outward current and hyperpolarized the membrane voltage. An increase in outward current was observed in 19 of 23 cells, with an increase of 285 ± 86% (n = 23, P < 0.05). The "resting" membrane potential was 37.2 ± 1.7 mV. This value is similar to that previously reported in isolated jejunal circular smooth muscle cells (8). The increase in outward current was accompanied by an 8.5 ± 2.9 mV membrane hyperpolarization. An example of the effect of 1% CO on an isolated canine jejunal circular smooth muscle cell is shown in Fig. 3. CO (1%) increased outward current by 255% in this cell (Fig. 3B). Figure 3C shows the current-voltage relationships of the whole cell currents. The membrane potential of the cell was 34 mV when bathed in normal Krebs solution. Changing the bath solution to CO (1%) resulted in an increase in outward current and a shift in the membrane potential from 34 to 50 mV. The difference current was obtained by subtracting the control current from the stimulated current and represents the portion of the current selectively activated by CO. The difference current reversed at 78 mV, close to the equilibrium potential for K⁺ (82 mV), suggesting a specific effect of CO on K⁺ current. The leak current, defined as an ohmic current that reversed at 0 mV, did not change in this cell. However, the leak current increased by 20 ± 18% in the 23 cells studied (P < 0.05).

View larger version (28K): [in this window] [in a new window]   Fig. 3.   Effect of carbon monoxide (CO) on an isolated canine jejunal circular smooth muscle cell. Whole cell current recorded in normal Krebs solution (A) increased by 255% when CO (1%) was present (B). C: current-voltage relationships show effects of CO on outward current and membrane potential. CO increased outward current and shifted membrane potential from 34 to 50 mV. Difference current was obtained by subtracting control current from CO-stimulated current and represents portion of current selectively activated by CO. Difference current reversed at 78 mV, near equilibrium potential for K+ (82 mV), suggesting an effect of CO on K+ current. No change in leak current, defined as an ohmic current that reversed at 0 mV, was noted in this cell.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Effect of the K+ channel blocker quinidine (Quin) on CO-stimulated outward current. A: whole cell current obtained in Krebs solution. B: effect of CO (1%) on outward current. C: effect of K+ channel blocker quinidine (50 µM) on CO-stimulated outward current. Note block of outward current by quinidine. D: current-voltage relationships obtained from whole cell currents. CO increased outward current and shifted reversal potential from 35 to 47 mV. Quinidine (50 µM) blocked outward current and shifted reversal potential from 47 mV to 2 mV. Results suggest that CO activated the quinidine-sensitive K+ current that determined the membrane potential.

View larger version (30K): [in this window] [in a new window]   Fig. 5.   Effect of CO on an isolated canine jejunal circular smooth muscle cell in presence of charybdotoxin. Whole cell current recorded in normal Krebs solution (A) increased by 136% when CO (1%) was applied in presence of charybdotoxin (100 nM) to block Ca2+-activated K+ channels (B), suggesting that Ca2+-activated K+ channels were not involved in the mechanism of action of CO. C: current-voltage relationships show effects of CO on outward current and membrane potential. CO increased outward current and shifted membrane potential from 30 to 51 mV. Difference current, obtained by subtracting control current from CO-stimulated current, reversed at 67 mV and represents portion of current selectively activated by CO.

Effects of long-term application of CO on outward current. In isolated human jejunal circular smooth muscle cells, long-term application (minutes) of CO evokes cyclic changes in outward current and in membrane potential (7). To determine whether the same effects were also present in isolated canine jejunal circular smooth muscle cells, whole cell outward currents were measured over a period of at least 15 min. In six of six cells tested, CO evoked cyclic changes in whole cell outward current. An example is shown in Fig. 6. The figure shows peak current measured at +40 mV and recorded at 1-s intervals. Marked oscillations in outward current were seen in the presence of CO, in contrast to the control tracing obtained in Krebs solution in the absence of CO. Because the oscillations were suggestive of spontaneous transient outward oscillations in current evoked by transient release of intracellular Ca²⁺, changes in intracellular Ca²⁺ were monitored in a separate series of experiments. Isolated canine jejunal circular smooth muscle cells were loaded with fura 2-acetoxymethyl ester, and the fluorescence emitted at excitation wavelengths of 340 and 380 nm was recorded. No effect of CO (1%) on intracellular Ca²⁺ levels was noted. The 340:380 nm ratio was 1.4 ± 0.4 before and 1.36 ± 0.5 after application of CO (P > 0.05, n = 6). Also, in suspensions of isolated canine jejunal circular smooth muscle cells loaded with indo 1, no changes in intracellular Ca²⁺ were seen using a fluorescence-activated cell sorter. These experiments, however, do not rule out localized increases in intracellular Ca²⁺ below the resolution of these techniques. To determine whether the increase in outward current observed with CO was dependent on intracellular Ca²⁺, traditional whole cell experiments were performed with 2 mM ethylene glycol-bis(-aminoethyl ether)-N,N,N',N'-tetraacetic acid in the recording pipette. After access to the cell was obtained, the cell interior was allowed to equilibrate with the pipette solution for 10 min. Application of CO (1%) increased outward current by 58 ± 40% (n = 6, P < 0.05) and hyperpolarized the membrane potential by 5.7 ± 3 mV (n = 6). As the increase in current in the standard whole cell configuration was less than that seen in amphotericin perforated patches, the data suggest that intracellular Ca²⁺ may be necessary for the full effects of CO, although washout of other intracellular messengers necessary for the effects of CO was also a possibility.

View larger version (21K): [in this window] [in a new window]   Fig. 6.   Cyclic changes in whole cell outward current evoked by application of CO (1%). Shown is peak current measured at +40 mV and recorded at 1-s intervals. Oscillation in outward current seen in presence of CO is in contrast to the absence of any oscillation when cell was treated with normal Krebs solution.

Effect of CO on intracellular cGMP. CO is among the endogenous factors known to activate soluble guanylyl cyclase (24) and therefore to stimulate the production of cGMP in many cell types (1-3, 28). The effect of CO (1%) on intracellular cGMP levels is shown in Fig. 7. CO (1%) increased cGMP levels from 86 ± 32 to 178 ± 70 pmol/10⁶ cells (n = 5, P < 0.05). To determine whether exogenous cGMP could mimic the effects of CO on outward current, 8-BrcGMP (membrane-permeable form of cGMP) was applied to the bathing solution. 8-BrcGMP (2 mM) increased whole cell current (94.2 ± 37%, n = 10, P < 0.05) and hyperpolarized (5.4 ± 2.7 mV, n = 10) the membrane potential (Fig. 8). If the effects of CO were mediated solely through cGMP, the further addition of CO would not be expected to further affect the outward current or membrane potential. Addition of 1% CO to the cGMP-stimulated current further increased the whole cell current, suggesting that in amphotericin perforated patch recordings the effects of CO may not be solely mediated through cGMP (Fig. 9). Cyclic increases in the whole cell current accompanied by membrane hyperpolarization were also noted with cGMP (data not shown).

View larger version (12K): [in this window] [in a new window]   Fig. 7.   Effect of CO on intracellular cGMP concentration measured by radioimmunoassay. Freshly dispersed cells in Krebs solution with 0.1 mM 3-isobutyl-1-methylxanthine, a phosphodiesterase inhibitor used to prevent cGMP breakdown, were divided into 2 samples (n = 5). 1% CO significantly increased cGMP levels from 86 ± 32 to 178 ± 70 pmol/106 cells (P < 0.05).

View larger version (31K): [in this window] [in a new window]   Fig. 8.   Effect of cGMP on whole cell outward current membrane potential. A: whole cell current recorded in Krebs solution. B: whole cell current recorded 8 min after application of 8-bromo-cGMP (8-BrcGMP; 2 mM). Current increased by 85%. C: current-voltage relationships for whole cell currents with a 17 mV hyperpolarization in membrane potential in presence of cGMP.

View larger version (26K): [in this window] [in a new window]   Fig. 9.   Effect of CO on cGMP-stimulated outward current. A: whole cell current recorded in Krebs solution. B: whole cell current recorded 7 min after application of 2 mM 8-BrcGMP. Current increased by 30%. C: effect of CO (1%) on cGMP-stimulated current. Outward current further increased by 65%. D: current-voltage relationships for whole cell currents. cGMP increased outward current and hyperpolarized membrane potential from 30 mV to 35 mV. CO further increased outward current and hyperpolarized membrane potential to 43 mV.

Effect of KT-5823. The effects of KT-5823, an inhibitor of cGMP-dependent protein kinase, on outward current and membrane potential were studied to further determine whether the hyperpolarization and increase in outward current seen on application of CO to the bath was due to cGMP or cGMP-dependent protein kinase. Cells were patch clamped in the standard whole cell configuration with KT-5823 (1 µM) in the recording pipette solution. After establishing access, control currents were observed for 10 min to allow KT-5823 to diffuse into the cell. CO (1%) was then added to the bath solution. The whole cell outward current increased by 7 ± 5%, and the membrane potential hyperpolarized by 0.3 ± 1.6 mV (n = 4, P > 0.05, data not shown). The data suggest that under standard whole cell recording conditions the increase in outward current and hyperpolarization evoked by CO was mediated primarily through a cGMP-dependent protein kinase.

Inhibition of HO. A major hurdle to the study of the role of CO as a physiological messenger is the absence of a specific blocker of HO. Zinc protoporphyrin IX has been used in several studies to inhibit HO; however, zinc protoporphyrin IX is also known to directly inhibit guanylyl cyclase (16) and to inhibit the actions of vasoactive intestinal polypeptide and atrial natriuretic peptide, both smooth muscle relaxants, through a non-HO pathway (16). Recently, CuPP-IX has been proposed as an inhibitor of HO with no effect on guanylyl cyclase (19, 29). Therefore, the effects of CuPP-IX on the whole cell current of isolated canine jejunal circular smooth muscle cells were determined. On addition to the bath, CuPP-IX (10 µM) evoked an immediate increase in outward current of 259 ± 94% (P < 0.05, n = 4). Prolonged exposure to CuPP-IX (30 min) resulted in a 25.3 ± 9% (P < 0.05) decrease in current over the maximal current recorded in the presence of CuPP-IX. However, the whole cell outward current did not return to baseline in all four cells. Washout of CuPP-IX (30 min) resulted in a 17 ± 6% increase in current (Fig. 10). The results may suggest a potential role for endogenous CO as an intracellular messenger in smooth muscle; however, due to the initial stimulatory effect of CuPP-IX, the role of CO as a messenger remains uncertain until more specific HO or CO inhibitors become available.

View larger version (10K): [in this window] [in a new window]   Fig. 10.   Effect of copper protoporphyrin IX (CuPP-IX; 10 µM) on whole cell current recorded from an isolated canine jejunal circular smooth muscle cell. Peak current (at +40 mV) is plotted against time. Prolonged exposure to CuPP-IX decreased outward current by 18% over the maximal outward current recorded in presence of CuPP-IX. Washout of CuPP-IX resulted in a 36% increase in outward current.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

We have shown in canine jejunum that HO-2, an enzyme whose activity leads to the endogenous production of CO, was present in neuronal cell bodies in a population of neurons in the myenteric and submucous plexuses, in nerve fibers innervating the circular smooth muscle layer, and in smooth muscle cells as well as in epithelial cells in the mucosa. We have also shown that exogenous CO stimulated intracellular cGMP production and activated a K⁺ conductance in circular smooth muscle cells, resulting in membrane hyperpolarization.

Endogenous production of CO as a byproduct of heme metabolism has been well known for many years (10). However, the potential role of CO generated by HO activity as a physiological messenger was only recently postulated (12, 15, 25). The data presented in this study indicate that the machinery for CO production (HO) was present in both neurons and smooth muscle, suggesting that CO may function as both a paracrine and an autocrine messenger in the gastrointestinal tract. Exogenous CO resulted in an increase in cGMP production in isolated jejunal circular smooth muscle cells and activated a K⁺ current. Both effects are expected to result in smooth muscle relaxation, suggesting that CO may act as an inhibitory messenger in the gastrointestinal tract.

The marked variation in the distribution of HO-2 among neuronal cell bodies in the canine jejunum, with some nerve cell bodies staining deeply for HO-2 while adjacent nerve cell bodies in the same ganglion show little or no staining, would argue in favor of a specific role for HO-2, perhaps in CO production, as opposed to reflecting a ubiquitous distribution of the enzyme. The two known constitutive isoforms of HO (HO-1 and HO-2) differ in their molecular weight, structure, and response to inducers. No evidence for specific immunostaining with HO-1 was seen in canine jejunal myenteric nerve cell bodies or in smooth muscle. HO-1, under normal conditions, is expressed at low levels in all tissues but the spleen. However, HO-1 activity in the rat can be increased 10- to 100-fold by a variety of agents, including heme and metal ions (10, 11). In the control state it appears that HO-2 is the predominate isoform in the canine jejunum; however, the presence of inducible HO-1 in canine jejunum at levels below the detection limits of our technique is also possible. This is in fact likely, as molecular biological techniques have demonstrated the presence of induced HO-1 in all tissues tested. For example, in the brain, levels of HO-1 are below the detection limit of Western immunoblotting, but HO-1 can be induced by hyperthermia and has been shown to be present by immunohistochemistry (5).

The effects of CO on peak whole cell current and membrane potential were similar to results previously reported in the human jejunum. In isolated human jejunal circular smooth muscle cells, 1% CO increases outward current by 175 ± 40% and hyperpolarizes the membrane potential by 15.6 ± 3.6 mV (7), similar to the 285 ± 86% increase in whole cell current and 8.5 ± 2.9 mV hyperpolarization of the membrane potential reported in this study on isolated canine jejunal circular smooth muscle cells. In both canine and human cells, long-term application of CO evoked cyclic changes in outward current and membrane potential, suggesting a similar mechanism of action for CO in both species.

In summary, in the canine jejunum, HO-2, a source for endogenous CO, was localized in a defined population of neuronal cell bodies, in nerve fibers, and in smooth muscle cells. Exogenous CO evoked an increase in outward K⁺ current accompanied by membrane hyperpolarization in isolated canine jejunal circular smooth muscle and increased cGMP levels. The data suggest that CO is endogenously produced in the canine jejunum in both nerves and smooth muscle, raising the possibility that CO may be an autocrine and paracrine messenger in the gastrointestinal tract.