 |  |
| ||||
Copyright © 2006 The Authors. Journal compilation © 2006 The Physiological Society Epithelial carbonic anhydrases facilitate PCO2 and pH regulation in rat duodenal mucosa Corresponding author J. D. Kaunitz, Bldg 114, Suite 217, West Los Angeles VA Medical Center, 11301 Wilshire Blvd, Los Angeles, CA 90073, USA. Email jake/at/ucla.edu Revised February 14, 2006; Accepted March 21, 2006.  This article has been cited by other articles in PMC. | |||||||||||||||||||
Abstract The duodenum is the site of mixing of massive amounts of gastric H+ with secreted HCO3−, generating CO2 and H2O accompanied by the neutralization of H+. We examined the role of membrane-bound and soluble carbonic anhydrases (CA) by which H+ is neutralized, CO2 is absorbed, and HCO3− is secreted. Rat duodena were perfused with solutions of different pH and PCO2 with or without a cell-permeant CA inhibitor methazolamide (MTZ) or impermeant CA inhibitors. Flow-through pH and PCO2 electrodes simultaneously measured perfusate and effluent pH and PCO2. High CO2 (34.7 kPa) perfusion increased net CO2 loss from the perfusate compared with controls (pH 6.4 saline, PCO2 ≈ 0) accompanied by portal venous (PV) acidification and PCO2 increase. Impermeant CA inhibitors abolished net perfusate CO2 loss and increased net HCO3− gain, whereas all CA inhibitors inhibited PV acidification and PCO2 increase. The changes in luminal and PV pH and [CO2] were also inhibited by the Na+–H+ exchanger-1 (NHE1) inhibitor dimethylamiloride, but not by the NHE3 inhibitor S3226. Luminal acid decreased total CO2 output and increased H+ loss with PV acidification and PCO2 increase, all inhibited by all CA inhibitors. During perfusion of a 30% CO2 buffer, loss of CO2 from the lumen was CA dependent as was transepithelial transport of perfused 13CO2. H+ and CO2 loss from the perfusate were accompanied by increases of PV H+ and tracer CO2, but unchanged PV total CO2, consistent with CA-dependent transmucosal H+ and CO2 movement. Inhibition of membrane-bound CAs augments the apparent rate of net basal HCO3− secretion. Luminal H+ traverses the apical membrane as CO2, is converted back to cytosolic H+, which is extruded via NHE1. Membrane-bound and cytosolic CAs cooperatively facilitate secretion of HCO3− into the lumen and CO2 diffusion into duodenal mucosa, serving as important acid–base regulators. | |||||||||||||||||||
The duodenum must absorb ~450 mmol of H+ per 24 h in order to decrease [H+] of the luminal content by 6 log orders over its 15 cm length (Feldman & Colturi, 1984). HCl in the duodenal lumen is neutralized by HCO3− secreted by the pancreas and duodenal epithelium, generating extremely high luminal CO2 pressures (PCO2 > 30 kPa) which dissipate by the proximal jejunum (Rune & Henriksen, 1969; Winship & Robinson, 1974). Gastric mucosal CO2 and H+ permeability is low, since pyloric obstruction leads to severe metabolic alkalosis due to the inability of the stomach to absorb substantial quantities of H+ or CO2 (Gamble & Ross, 1925; Javaheri & Nardell 1981). Thus, the duodenum is the major site for intestinal H+ and CO2 absorption. Transmucosal bulk absorption of CO2 and H+ is likely to follow the sequence of transport across the apical cell membrane into the cytosol, transport across the basolateral membrane of the epithelium into the subepithelial interstitium, followed by transport into the portal vein. The mechanism for CO2 and H+ transport across cell membranes remains incompletely understood. In terms of transapical CO2 and H+ transport, one accepted model involves conversion of luminal H+ and HCO3− to CO2 and H2O, diffusion of CO2 across the apical plasma membrane, and hydration of CO2 to HCO3− and H+ in the cytoplasm. This process, sometimes termed the Jacobs-Stewart cycle after its original description in red blood cells (Jacobs & Stewart, 1942), requires the presence of intracellular and extracellular carbonic anhydrases (CAs) and a plasma membrane anion exchanger. Since the duodenal epithelium has abundant cytoplasmic and membrane-associated CA activity (Sugai et al. 1994; Lönnerholm et al. 1989; Parkkila et al. 1994; Saarnio et al. 1998; Purkerson & Schwartz, 2005; Leppilampi et al. 2005) and apical anion exchangers (Wang et al. 2002; Spiegel et al. 2003), we hypothesized that most of the excess duodenal luminal H+ was absorbed through neutralization of secreted HCO3−, yielding luminal CO2. CO2 is the molecular species that then traverses the apical membrane, which enters the cytoplasm, and is hydrated to H2CO3, which then dissociates in the cytoplasm to H+ and HCO3−. HCO3− is transported into the lumen whereas H+ is transported into the submucosal space by membrane transport proteins. In essence, luminal H+ is transported through the apical membrane as CO2, but HCO3− is simultaneously secreted in its anionic form. Several observations support this hypothesis. We, and others have found that luminal acidification or elevation of luminal PCO2 provokes several epithelial responses, such as acidification of the cytoplasm and subepithelial interstitial fluid, increased mucosal blood flow, increased HCO3− secretion, and increased mucus secretion (Flemström & Kivilaakso, 1983; Flemström, 1994; Seno et al. 1998; Paimela et al. 1990, 1992; Akiba & Kaunitz, 1999; Akiba et al. 2000, 2001a,b, 2006). Since these responses appear to be dependent on acidification of the subepithelial space, and since elevated luminal CO2 or H+ provoke similar responses, it appears that luminal CO2 must be converted to subepithelial H+, which then signals these protective mechanisms (Allen & Flemström, 2005). The further characterization of transmucosal H+ and CO2 movement thus has larger implications for the understanding of mucosal protective mechanisms and the signalling pathways coordinating mucosal responses to acid perfusion. In order to further test our hypothesis, we devised a system in which the movement of CO2 and H+ between lumen, mucosa and the portal vein was measured. In order to determine the contribution of cytoplasmic and extracellular CAs towards H+ and CO2 movement, we used cell-permeant and -impermeant CA inhibitors. Finally, transmucosal tracer carbon movement was measured using 13C. Our results support our hypothesis that luminal H+, neutralized by secreted HCO3−, is converted to CO2 prior to entry into the cytoplasm of the epithelial cells, and that cellular CO2 is reconverted to H+, which then is transported into the portal vein. | |||||||||||||||||||
Methods Chemicals and animals The sulphonamide compounds 1-[4-aminosulphonyl]phenyl]-2,4,6-trimethylpyridinium perchlorate (6a) and 1-[4-([5-(aminosulphonyl)-1,3,4-thiadiazol-2-yl]aminosulphonyl)phenyl]3,5-nonylene-2,6-dimethylpyridinium perchlorate (14v) are potent, cell-impermeant and selective inhibitors of extracellular CA (Ki for CA IV and IX ≈ 5–300 nm) (Casey et al. 2004; Pastorekova et al. 2004). 13C-labelled sodium bicarbonate (NaH13CO3) was purchased from Cambridge Isotope Laboratories, Inc. (Andover, MA, USA). S3226, a selective Na+–H+ exchanger 3 (NHE3) inhibitor (Schwark et al. 1998; Knutson et al. 1988; Wiemann et al. 1999; Vallon et al. 2000; Ledoussal et al. 2001; Furukawa et al. 2004) was a kind gift of Aventis Pharma Deutschland (Frankfurt am Main, Germany). Methazolamide (MTZ), acetazolamide (ACZ), 5-(N,N-dimethyl)-amiloride (DMA), Hepes, and all other chemicals were obtained from Sigma Chemical. Krebs solution contained (mm): 136 NaCl, 2.6 KCl, 1.8 CaCl2, and 10 Hepes at pH 7.0. For neutral pH–high CO2 perfusion, solutions were made as previously described (Furukawa et al. 2005) using a 50 mm NaHCO3–105 mm NaCl solution and 20 mm HCl–135 mm NaCl solution, prewarmed to 37°C, generating an isotonic (310 mosmol l−1) pH 6.4 solution with PCO2= 34.7 kPa, calculated using a CO2 aqueous solubility constant of 0.24 mm kPa−1 and the first pKa of carbonic acid = 6.1 at 37°C (Geers & Gros, 2000). Solutions were vigorously mixed 1 min prior to perfusion. We confirmed that the pH of the freshly mixed solution reached steady state within 10 s, whereas [CO2] reached its equilibrium of carbonic acid, and H+ and HCO3− by 1 min, and was stable for at least 10 min, as measured by pH and CO2 electrodes (Lazar Research Laboratories, Inc., Los Angeles, CA, USA). Each solution was prewarmed at 37°C using a water bath, and temperature was maintained with a heating pad during the experiment. For low pH–high CO2 perfusion, 10 mm NaHCO3–135 mm NaCl solution was titrated to pH 2.2 with 1 n HCl, generating a PCO2= 41.5 kPa at 37°C, as previously described (Akiba et al. 2001a). The composition of the perfusates used in this study is listed in Table 1. Stock solutions of MTZ, ACZ, S3226, DMA, compound 6a and compound 14v were prepared by dissolving in dimethyl sulfoxide (DMSO), stored at −20°C until use. DMSO (0.1%) in saline or Hepes buffer was used for vehicle perfusions. All studies were performed with approval of the Veterans Affairs Institutional Animal Care and Use Committee (VA IACUC). Male Sprague-Dawley rats weighing 200–250 g (Harlan, San Diego, CA, USA) were fasted overnight, but had free access to water.
Measurement of luminal pH and [CO2] Under continuous isoflurane anaesthesia (1.5–2.0%) using a rodent anaesthesia inhalation system (Summit Medical Systems, Bend, OR, USA), rats were placed supine on a recirculating heating block system (Summit Medical) to maintain body temperature at 36–37°C, as monitored by a rectal thermistor. Prewarmed saline was infused via the right femoral vein at 1.08 ml h−1 using a Harvard infusion pump (Harvard Apparatus, Holliston, MA, USA); blood pressure was monitored via a catheter placed in the left femoral artery using a pressure transducer (Kent Scientific, Torrington, CT, USA). Duodenal loops were prepared and perfused as previously described (Akiba et al. 2001a,b; Furukawa et al. 2005). Briefly, the abdomen was incised, both stomach and duodenum were exposed, and the forestomach wall was incised 0.5 cm using a thermal cautery (Geiger Medical Technologies, Inc., Monarch Beach, CA, USA). A polyethylene tube (diameter 5 mm) was inserted through the incision until it was 0.5 cm caudal from the pyloric ring, where it was secured with a nylon ligature. The distal duodenum was ligated proximal to the ligament of Treitz before the duodenal loop was filled with 1 ml saline prewarmed at 37°C. The distal duodenum was then incised, and another polyethylene tube was inserted through the incision and sutured into place. To prevent contamination of the perfusate from bile or pancreatic juice, the pancreaticobiliary duct was ligated just proximal to its insertion into the duodenal wall and cannulated with a PE-10 tube to drain the juice. The resultant closed proximal duodenal loop (perfused length 2 cm) was perfused with prewarmed saline by using a peristaltic pump (Fisher Scientific, Pittsburgh, PA, USA) at 1 ml min−1. In a modification of published methodology (Morgan et al. 1997; Dalenback et al. 1995), we continuously measured pH and PCO2 in the perfusate and in the effluent. Input (perfusate) and effluent pH and [CO2] were continuously measured with two sets of pH and CO2 electrodes, respectively, placed in series using flow-through cells (Micro Flow-Through pH and CO2 electrodes, Lazar Research Laboratories). These flow-through cells were immersed in a water bath maintained at 37°C with a thermo regulator (Fisher Scientific). For the mixed solutions, CO2 electrodes were calibrated every day with 0.1, 1.0 and 10 mm NaHCO3 in pH 5 citrate buffer, which generated 0.1, 1.0 and 10 mm total CO2 content ([CO2]tot), respectively. For the gas equilibrated solutions, we calibrated with 0.5, 5.5 and 20.5 mm NaHCO3 in pH 7.4 100 mm Hepes−105 mm NaCl, which generated 0.047, 0.473 and 1.388 mm CO2, respectively. This system enabled us to calculate the concentration changes in luminal CO2, total CO2, HCO3− and H+ by the following equations derived from the Henderson-Hasselbalch equation:
where [CO2] was calculated according to the calibration curve as mentioned above and pKa, the first dissociation constant of carbonic acid = 6.1 at 37°C. The net changes in concentration was determined as:
expressed so that positive quantities denote net secretion whereas negative quantities indicate net absorption from the lumen. For example,
were used to calculate the net secretion or absorption of [CO2] and [H+], respectively. For saline ([CO2]≈ 0) perfusion, the perfusate and effluent were circulated through a reservoir in which the perfusate was bubbled with 100% O2, and stirred and warmed at 37°C with a heating stirrer (Barnstead Int., Dubuque, IO, USA). The pH of the perfusate was kept constant at pH 7.0 with a pH stat (models PHM290 and ABU901; Radiometer Analytical, Lyon, France). For 5% CO2–HCO3− buffer perfusion, the perfusate, consisting of pH 7.4 Hepes 10 mm, NaHCO3 25 mm and NaCl 125 mm, was constantly bubbled with 5% CO2–95% O2, stirred and warmed at 37°C with a heating stirrer, and circulated with a peristaltic pump, whereas the effluent was discarded. Figure 1 shows a time course of input and effluent pH and PCO2 measurements during luminal perfusion with saline and with a high PCO2 solution.
Measurement of portal venous pH and PCO2 Before preparing the duodenal loop as described above, the gastroduodenal branch of the portal vein (PV) was cannulated with a 23-gauge metal cannula connected to a PE-50 tube as previously described (Wachter et al. 1998). The catheter was fixed with cyanoacrylate glue and the tube was filled with heparinized saline enabling repeated blood sampling. Portal blood samples were obtained as described below, and pH and PCO2 were measured with a blood gas analyser (ABL5, Radiometer, Copenhagen, Denmark).Experimental protocol Mixed high CO2 solution After stabilization of pH and [CO2] measurements with continuous perfusion of pH 7.0 saline for ~30 min, the time was set as t = 0. The duodenal loop was perfused with pH 7.0 saline from t = 0 min until t = 20 min (basal period). The perfusate was then changed to pH 6.4 saline ([CO2]≈ 0, [HCO3−]= 0, neutral pH–low CO2 solution), the high CO2 solution (pH 6.4, PCO2 34.7 kPa, neutral pH–high CO2 solution) prepared as described above, pH 2.2 acid saline ([CO2]≈ 0, [HCO3−]≈ 0, low pH–low CO2 solution), or pH 2.2 high CO2 solution (PCO2= 41.5 kPa, low pH–high CO2 solution) prepared as described above (Table 1), from t = 20 min until t = 30 min (challenge period), with or without inclusion of inhibitors (described below). At t = 20 min, the system was gently flushed so as to rapidly change the composition of the perfusate. During the challenge period, the solution was perfused with a syringe pump. At the end of the challenge period (t = 30 min, 10 min after CO2 or acid stress), an aliquot of portal blood was analysed. Abdominal aortic blood was analysed for comparison. To examine the effect of the inhibition of cytosolic CA on luminal and PV pH and [CO2], the duodenal loop was pretreated with MTZ (1 mm) dissolved in pH 7.0 Krebs buffer, for 5 min during the basal period, followed by perfusion of the high CO2 solution or pH 2.2 acid saline. Cell-permeant MTZ inhibits all cytosolic and extracellular CA; when used as a pretreatment, inhibition of extracellular CA is presumably selectively lessened due to a presumed faster washout from extracellular than from intracellular sites. We have reported that 1 mm MTZ pretreatment successfully inhibits CO2-induced duodenal bicarbonate secretion and epithelial acidification (Furukawa et al. 2005). Additionally, to confirm the effects of inhibition of cytosolic CA as well as subepithelial CAs including the basolateral CA and vascular CA, ACZ (10 mg kg−1) was given intravenously at t = 10 min, 10 min before the high CO2 challenge or pH 2.2 acid challenge. To inhibit the extracellular CA, cell-impermeant CA inhibitor compounds 6a and 14v (0.001–0.1 μm) were used. Those compounds have a positively charged residue with an in vitroKi for CA II, IV and IX = 0.1 μm with minimal cytoplasmic permeation (Pastorekova et al. 2004; Casey et al. 2004). The duodenal loop was perfused with 6a or 14v dissolved in the neutral pH–high CO2 solution. The effect of 6a (0.1 μm) dissolved in pH 2.2 acid saline was also examined. To further examine the role of NHE in CO2 and H+ movement during the high CO2 challenge, the non-selective NHE inhibitor DMA (0.1 mm), which inhibits NHE1 even when luminally perfused, on the basis of intracellular pH (pHi) studies and human absorption studies (Williams et al. 1987; Akiba & Kaunitz, 1999; Furukawa et al. 2005), or the selective NHE3 inhibitor S3226 (10 μm) (Furukawa et al. 2004) was perfused with the high CO2 solution. We confirmed that DMA, S3226, 6a, or 14v at these concentrations had no effect on solution pH or [CO2]. Equilibrated high CO2 solution As an alternative approach to generating solutions with high [CO2], we made solutions that were bubbled with CO2–O2 mixtures. A pH 7.4, 5% CO2–HCO3− buffer containing 10 mm Hepes, 25 mm NaHCO3 and 125 mm NaCl was prepared by constant bubbling with 5% CO2–95% O2. This solution was perfused until the pH and PCO2 measurements stabilized, after which the time was set as t = 0. The duodenal loop was then perfused with the 5% CO2–HCO3− buffer from t = 0 min until t = 20 min (basal period). The duodenum was superfused with either the same solution, or the solution was changed to a 30% CO2–HCO3− buffer, consisting of pH 6.5 10 mm Hepes, 25 mm NaHCO3 and 125 mm NaCl, constantly bubbled with 30% CO2−70% O2 (Praxair, Torrance, CA, USA), generating pH 6.5, PCO2 30.4 kPa (Table 1), from t = 20 min until t = 30 min (challenge period), with or without the inhibitors described below. At t = 20 min, the system was gently flushed in order to rapidly change the perfusate; otherwise, the system was perfused with a peristaltic pump (Fisher Scientific). MTZ (1 mm) dissolved in 5% CO2–HCO3− buffer was flushed into the system at t = 15 min and then perfused from t = 15 to t = 20 min, followed by the high CO2 challenge. Compound 6a (0.1 μm) or 14v (0.1 μm) was perfused with the high CO2 solution.Luminal 13CO2 diffusion into portal vein To elucidate whether luminal CO2 directly diffuses into PV and when luminal CO2 appears in PV, we perfused 13C-labelled CO2–HCO3− solution through the duodenal loop. The high CO2 solution was made from the equal mixture of 50 mm NaH13CO3−105 mm NaCl solution and 20 mm HCl–135 mm NaCl solution, both prewarmed at 37°C and was perfused during the challenge period with or without inhibitors 6a or 14v (0.1 μm) or pretreatment with MTZ (1 mm) as described above. In the experiments, four samples of PV blood (each 0.2 ml) were taken followed by flushing with heparinized saline (each 0.2 ml) at the end of the basal period (t = 20 min), during the challenge period (t = 21 min and t = 25 min) and at the end of the challenge period (t = 30 min). We confirmed that the four 0.2 ml blood withdrawals followed by saline flushing did not affect blood pressure or body temperature. A 0.1 ml aliquot was immediately transferred to a borosilicate microvial with Teflon–silicone septum (Corning, Acton, MA, USA) followed by freezing with dry-ice ethanol and storage at −80°C until analysis. Another 0.1 ml was used for blood gas analysis.PV 13CO2/12CO2 was analysed with gas chromatography-mass spectrometer (Finnegan MAT Delta IRMS coupled to a Hewlett-Packard 5890 gas chromatograph; Thermo-Finnigan, Waltham, MA, USA) as previously described (Kasho et al. 1996). Immediately after thawing the blood sample at 25°C, a 25-μl aliquot was transferred into another vial containing 50 μl 0.1 m NaHCO3, immediately followed by the addition of 50 μl 1 n HCl to convert all HCO3− to CO2 at which time the vial was sealed with a metal cap and rubber septum. After vortexing, the vial was set into an auto sampler. Relative enrichment of 13C was expressed as δ13C/12C relative to Pee Dee Belemnite limestone by the software program Isodat. Duplicate measurements of δ13C/12C was made using 5% CO2 gas as an internal control, with results expressed as the difference, positive or negative, of δ13C/12C from the reference. The sample obtained at t = 20 min, prior to 13CO2 solution perfusion, was defined as baseline, then the difference of δ13C/12C from the baseline, expressed as Δδ13C/12C, was measured at 1, 5, or 10 min after CO2 challenge (t = 21, 25 or 30 min). Statistics All data were derived from experimental groups with n = 6 and were expressed as means ±s.e.m. Comparisons between groups were made by one-way ANOVA followed by Fischer's least significant difference test. P < 0.05 was taken as significant. | |||||||||||||||||||
Results Net [CO2] and [HCO3−] movement and PV pH and PCO2 in response to luminal perfusion Premixed solution Baseline studies and the effect of MTZ or ACZ We initially performed measurements of perfusate and effluent [H+] and [CO2], from which we could calculate net gain or loss of either species, and also derive net changes of [HCO3−]. A time point of 10 min after the solution change was chosen to allow for equilibration of the perfusate with the mucosa, and also to coincide with portal blood measurements. Since there is no consensus regarding how best to generate a high PCO2 solution, we first perfused duodenal loops with a solution generated by the mixture of acid and a bicarbonate solution, as previously described (Holm et al. 1998; Furukawa et al. 2005). Furthermore, studies of transmucosal 13C movement necessitate the use of high CO2 solutions generated from the mixture of HCO3− and HCl. The stable, positive net changes of perfusate [CO2] and [HCO3−] (Δ[CO2] and Δ[HCO3−], respectively) were observed during the perfusion of saline (pH 7.0) during the basal period (t = 0–30 min), due to the basal HCO3− secretion (Fig. 1; data not shown). Figure 2A and B depicts Δ[CO2] and Δ[HCO3−] at t = 30 min, 10 min after perfusion of either pH 6.4 saline ([CO2]≈ 0), or perfusion with pH 6.4 high CO2 saline (PCO2= 34.7 kPa). Figure 2C–E depicts PV pH, PCO2 and [CO2]tot at t = 30 min. Positive Δ[CO2] and Δ[HCO3−] in the saline control group revealed net basal HCO3− secretion. Compared with saline controls, Δ[CO2] was decreased to negative values and Δ[HCO3−] was increased during exposure to high [CO2], consistent with simultaneous net movement of CO2 from perfusate to mucosa (absorption) and HCO3− secretion (Fig. 2A and B) or luminal conversion of CO2 to HCO3− in bulk solution. Furthermore, perfusion of high CO2 solutions lowered PV pH and elevated PV PCO2 (Fig. 2C and D), consistent with CO2 or H+ absorption into the portal venous blood, supporting luminal CO2 loss due to CO2 movement into the mucosa, and not due to luminal conversion of CO2 to HCO3−. Since PV [CO2]tot was unchanged during perfusion with the high CO2 solution (Fig. 2E), only [H+] was increased in the PV, indicating that luminal CO2 was absorbed and then converted to H+. We then examined the effect of MTZ pretreatment, in which the epithelial cells were preloaded with MTZ for 5 min, which was then washed out before the high CO2 challenge, in order to selectively inhibit cytosolic CA. MTZ (1 mm) pretreatment attenuated the CO2-induced CO2 loss and HCO3− increase (Fig. 2A and B), and also abolished the changes in PV pH and PCO2 (Fig. 2C and D), suggesting that cytosolic CA activity mediates CO2 absorption and CO2-induced HCO3− secretion, the latter finding consistent with our prior observations (Furukawa et al. 2005) and with recent observations made in CA II knockout mice (Leppilampi et al. 2005). As a further control for selective soluble CA inhibition, we infused ACZ (10 mg kg−1i.v.) 10 min before the high CO2 challenge, which produced an effect identical with Δ[CO2], but decreased Δ[HCO3−] to a negative value (Fig. 2A and B).
Effect of cell-impermeant CA inhibitors We next examined the effect of cell-impermeant CA inhibitors, in order to study the effect of extracellular CAs on H+ and CO2 movement in the absence of inhibition of cytosolic CA. Figure 3 depicts measurements made at t = 30 min, 10 min after the perfusion of either saline or the high CO2 solution. The cell-impermeant CA inhibitors, compounds 6a and 14v, dose-dependently reversed the effects of perfusion of the high CO2 solution, converting net CO2 loss into net CO2 secretion (Fig. 3A). Furthermore, 0.1 μm 6a and 14v augmented the increase of Δ[HCO3−] during high CO2 exposure (Fig. 3B), suggesting that the inhibition of extracellular CA stimulates HCO3− secretion during CO2 exposure, as also found recently in CA IX knockout mice (Leppilampi et al. 2005). CO2-induced PV acidification and PCO2 increase were inhibited by 6a and 14v (Fig. 3C and D), consistent with the observed inhibition of CO2 loss from the lumen. The challenge with high CO2 perfusate, with or without MTZ or impermeant CA inhibitors in the duodenal loop, had no effect on the systemic arterial blood pH, PCO2 and PO2 at the end of the challenge period (data not shown), confirming that high CO2 exposure or luminal CA inhibitors had no effect on the systemic acid–base balance.
Effect of NHE inhibitors Since NHE3, which is expressed in the epithelial cell apical membrane, facilitates H+ secretion and Na+ absorption (Vallon et al. 2000), and since basolateral NHE1 regulates intracellular pH by H+ extrusion into the subepithelial space (Praetorius et al. 2000; Pedersen & Cala, 2004), we also examined the effect of NHE inhibitors on CO2 and H+ movement during the high CO2 challenge. Figure 4A and B depicts Δ[CO2] and Δ[HCO3−] at t = 30 min, 10 min after perfusion of either saline or the high CO2 solution and Fig. 4C and D depicts PV pH and PCO2 at t = 30 min. DMA (0.1 mm) inhibited CO2 absorption (Fig. 4A) and HCO3− secretion (Fig. 4B) during perfusion of the high CO2 solution, accompanied by inhibition of PV acidification (Fig. 4C) and PV PCO2 increase (Fig. 4D). These observations are consistent with net H+ movement through the basolateral NHE1 from the epithelium into PV blood. Moreover the selective NHE3 inhibitor S3226 (10 μm) had no effect on CO2 loss, HCO3− secretion, or on PV pH or PCO2, suggesting that the apical NHE3 is not involved in CO2–H+ movement across the duodenal epithelium. These results suggest that loss of CO2 from the luminal solution corresponds to CO2 movement across the apical membrane, not to H+ movement via NHE3. Absorbed CO2 exits the cell via NHE1 as H+, which acidifies PV blood.
Effect of luminal acidity We then measured Δ[CO2]tot and Δ[H+] during perfusion with low CO2 solutions using perfusate [H+] as the variable. Under acidic conditions, [CO2]tot≈[CO2] and [HCO3−]≈ 0. As shown in Fig. 5A and B, net changes in total CO2 concentration (Δ[CO2]tot) and [H+] (Δ[H+]) at t = 30 min were assessed. Since pH 2.2 saline (low pH–low CO2 solution) contained ~0 [CO2], the positive Δ[CO2]tot measured during pH 2.2 saline was consistent with HCO3− secretion during exposure to luminal acid. Nevertheless, Δ[CO2]tot was diminished and Δ[H+] was negative during acid exposure compared with pH 7.0 saline (Fig. 5A), similar to our previous report (Akiba et al. 2001a), and consistent with partial CO2 loss, or diminution of HCO3− secretion, and H+ loss during acid perfusion. Perfusion with a low pH–low CO2 solution acidified the PV blood and increased portal PCO2 (Fig. 5C and D), corresponding to transmucosal CO2 or H+ absorption. MTZ (1 mm) pretreatment reversed these changes of Δ[CO2]tot and Δ[H+] in the perfusate and pH and PCO2 in the PV (Fig. 5A–D), suggesting that H+ loss from the lumen is also mediated by the cytosolic CA. No change of [CO2]tot was measured in PV (Fig. 5E), consistent with net transmucosal transfer of H+, and not CO2, under this condition. ACZ (10 mg kg−1, i.v.) pretreatment 10 min before pH 2.2 acid challenge also inhibited the changes of Δ[CO2]tot and Δ[H+] (Fig. 5A and B), confirming the involvement of cytosolic and subepithelial CAs in H+ absorption. However, ACZ had no effect on PV pH (Fig. 5C), probably due to the systemic acidosis induced by ACZ i.v. injection (Table 2), although pH 2.2 acid challenge itself had no effect on the systemic acid–base balance. Furthermore, the cell-impermeant CA inhibitor 6a (0.1 μm) perfused with pH 2.2 solution inhibited the changes in perfusate and PV CO2 and H+ (Fig. 5A–D), indicating that apical extracellular CAs also mediated net H+ absorption from the lumen.
To further investigate the movement of CO2 and H+ from the lumen, we perfused a low pH–high CO2 solution (pH 2.2, PCO2 41.5 kPa), comparable to the postprandial condition. This solution was alternately perfused with the pH 6.4–high CO2 solution in order to measure the effect of pH on Δ[CO2] (Fig. 6A), and with pH 2.2 saline in order to measure the effect of [CO2] on Δ[H+] (Fig. 6B). CO2 loss from the pH 2.2–high CO2 group was greater than from the pH 6.4–high CO2 solution (Fig. 6A). Conversely, H+ loss from the pH 2.2–high CO2 solution was less than from pH 2.2 saline (Fig. 6B), suggesting that luminal acidity enhances CO2 loss, whereas luminal CO2 reduces H+ loss. These data are also consistent with net conversion of H+ and HCO3− to CO2 in the lumen, with CO2 transported into the mucosa, since increasing luminal [H+] increases luminal [CO2], increasing the magnitude of the CO2 diffusion gradient. In Fig. 6B, net loss of luminal H+ presumably occurs through titration of basal HCO3− secretion. Increasing luminal [CO2] in this condition actually slows neutralization of luminal H+ by HCO3− to CO2 by mass action, driving the reactions in the direction of H+ and HCO3− and away from H2O and CO2, decreasing apparent H+ loss.
Studies with gas-equilibrated solutions To further confirm the effect of CA inhibition on CO2 absorption/loss from the duodenal lumen, we used pH-buffered, gas-equilibrated solutions, which represent an alternative approach to generating high PCO2 conditions. The duodenal loop was superfused with a 5% CO2–HCO3− buffer (pH 7.4) followed by 30% CO2–HCO3− buffer (pH 6.5; Table 1). Figure 7A depicts the time course of Δ[CO2] during the perfusion of 5% and 30% CO2 solutions. During the basal period, Δ[CO2] was ~0, probably due to the lack of a transmucosal PCO2 gradient under this condition. Perfusion of a 30% CO2 solution gradually decreased Δ[CO2], consistent with net CO2 loss from the perfusate. MTZ inhibited luminal CO2 loss and the decrease of PV pH and the increase of PV PCO2 effects consistent with those observed with the mixed high CO2 solutions (Fig. 7A–C).
Next, we recorded pH and PCO2 at 1 min intervals following the perfusion of the 30% CO2 solution, starting at t = 20 min. At t = 21 there was a rapid loss of CO2, followed by a brief recovery, followed by a more gradual and steady loss (Fig. 8A). The overall amount of CO2 loss was attenuated by co-perfusion of the membrane-impermeant CA inhibitors compound 6a and 14v (0.1 μm), suggesting that the extracellular CA also participates in the CO2 absorption from the lumen to the mucosa. These findings were supported by the inhibition of PV pH increase and PCO2 increase following perfusion of 30% CO2, effects also attenuated or abolished by co-perfusion of the cell-impermeant CA inhibitors (Fig. 8B and C).
Changes in pH, PCO2, total CO2 and Δδ13C/12C in PV blood after luminal CO2 exposure The time course of pH, PCO2, total CO2 and change in δ13C/12C (Δδ13C/12C) in the PV blood sample from just before luminal 13CO2 exposure (t = 20 min, at 0 min after CO2 challenge) to the end of the exposure (t = 30 min, at 10 min after CO2 challenge) are shown in Fig. 9A–D. PV pH decreased and PCO2 increased only after a 10 min exposure to elevated luminal CO2 solution, an effect blocked by MTZ (1 mm) pretreatment and by impermeant CA inhibitors (0.1 μm), whereas [CO2]tot was unchanged over time (Fig. 9A–C). PV Δδ13C/12C was sharply increased 1 min after CO2 challenge, remained at the same level 5 min thereafter, and significantly increased 10 min after CO2 challenge (Fig. 9D). These changes were inhibited by MTZ (1 mm) pretreatment, and by compound 6a and 14v (0.1 μm) co-perfusion. Nevertheless, PV total CO2 ([CO2]tot) did not change (Fig. 9C), again consistent with the lack of net bulk transmucosal CO2 transfer, even though Δδ13C/12C increased over time in PV. Note that Δδ13C/12C in systemic arterial blood of high CO2 group remained low at 10 min after the 13CO2 challenge, confirming that the increase of Δδ13C/12C was not due to the recirculation of the systemic elevated 13CO2.
| |||||||||||||||||||
Discussion We hypothesized that (1) luminal H+ secreted by the stomach is neutralized by secreted HCO3− in the duodenum, generating CO2; (2–3) CO2 traverses the epithelial cell apical membrane into the cytosol, where it is converted into H+ and HCO3− by cellular CA; (4–6) cellular H+ exits the cells via the basolateral NHE1 into the subepithelial interstitium, acidifying PV blood; (7–8) cellular HCO3− exits the cell via the apical anion exchanger into the lumen, where it is converted into CO2 by apical extracellular CA at moderate pH and by H+ at low pH (Fig. 10). Cytosolic CA participates in the cellular conversion of CO2 to H+ and HCO3−, whereas the apical extracellular CA helps convert HCO3− to CO2. Finally, cytosolic and extracellular CAs facilitate CO2 diffusion across the apical membrane, resulting in net H+ absorption from duodenal lumen to portal blood. Inhibition of cytosolic CA may inhibit CO2 diffusion by inhibiting conversion of CO2 to H+ whereas inhibition of apical extracellular CA may delay CO2 diffusion by reducing conversion of luminal HCO3− to CO2, with a resultant accumulation of secreted HCO3− in the lumen. This is the first study to show the role of extracellular and cytosolic CA in duodenal transmucosal absorption of CO2 and H+, and to demonstrate the facilitation of CO2 diffusion by epithelial CAs.
During perfusion with saline, HCO3− was secreted, corresponding to the basal HCO3− secretion observed previously (Furukawa et al. 2005). During perfusion with a high PCO2 solution, CO2 was lost from the lumen, although HCO3− was concurrently secreted, consistent with luminal CO2 diffusing to the cytoplasm simultaneously with transport of HCO3− from cytoplasm to lumen. During CO2 challenge, blood PCO2= 5.3 kPa, lumen PCO2= 34.7 kPa, blood [HCO3−]= 24 mm, lumen [HCO3−]= 16.7 mm, creating gradients for CO2 diffusing into the cell and HCO3− diffusing into the lumen down their respective concentration gradients. These data, which confirm prior observations of augmented HCO3− secretion during luminal perfusion with high PCO2 solutions (Furukawa et al. 2005; Holm et al. 1998), are consistent with the occurrence of two distinct processes: diffusion of CO2 into the mucosa, facilitated by CAs, and active HCO3− secretion, presumably facilitated by the SLC26Ax class of anion exchangers (Xu et al. 2003; Wang et al. 2002). These net movements were accompanied by PV acidification without change of PV [CO2]tot, consistent with net transfer of H+, and not CO2, from lumen to PV. These changes were reversed by pretreatment with the CA inhibitor MTZ, confirming the role of cytoplasmic CA in the loss of luminal CO2 and in the augmentation of HCO3− secretion. In the presence of selective inhibition of extracellular CAs by impermeant CA inhibitors, CO2 was secreted rather than being absorbed, and net HCO3− secretion was increased, in contrast to the results obtained with MTZ pretreatment or with i.v. ACZ. These results are best explained by the formation of a luminal ‘bicarbonate trap’ in which conversion of luminal HCO3− into CO2 and H2O is impeded, strongly supporting the role of apical extracellular CAs in luminal HCO3−↔ CO2 interconversion. These studies are indicative of a specialized role of extracellular CAs such as CA IV, IX, and possibly XII and XIV in duodenal CO2 dynamics, in which extracellular CA facilitates CO2 entry into the mucosa, and are supported by the findings made in mice with a deletion for the extracellular CA IX, in which basal and stimulated HCO3− secretion was augmented (Leppilampi et al. 2005), and in eel intestine, in which luminal ACZ inhibited the loss of luminal HCO3−, but amiloride had no effect (Maffia et al. 1996). One interpretational caveat is that CA IX is present on the basolateral membrane of duodenal epithelial cells (Saarnio et al. 1998; Hilvo et al. 2004) which is presumably inaccessible to luminally perfused 6a and 14v. Facilitation of transmembrane CO2 transport by extracellular CAs has been reported in a variety of organs, and is considered to be a major pathway for transmembrane movement of H+ equivalents (Heming et al. 1994; Henry et al. 1997; Geers & Gros, 2000). Luminal perfusion with acid affects CO2 and H+ movement differently than does perfusion with a high PCO2 solution. For example, luminal acid perfusion reduces baseline HCO3− secretion, an effect that we have previously observed (Akiba et al. 2001a) and have used to support our hypothesis that intracellular HCO3− is an important duodenal epithelial defense mechanism (Akiba et al. 2001a, 2005b; Hirokawa et al. 2004). Since pHi decreases during luminal perfusion with acidic solutions (Paimela et al. 1992; Akiba & Kaunitz, 1999), net HCO3− secretion might be inhibited due to decreased cystic fibrosis transmembrane conductance regulator (CFTR) and apical SLC26A anion exchanger function as a result of acidic pHi (Willumsen & Boucher, 1992; Reddy et al. 1998). Reversal of the changes observed during acid perfusion by MTZ pretreatment supports the role of cytoplasmic CA in net HCO3− secretion and in H+ absorption. The reversal of net H+ absorption from the lumen by MTZ or the impermeant CA inhibitor 6a, accompanied by inhibition of PV acidification, strongly supports our hypothesis that luminal acid enters the cell as CO2 with the aid of the apical extracellular CA and is then converted to H+ in the cytosol prior to transport out of the cell. Since membrane-bound CAs, such as CA IV, are localized on endothelial cell membranes (Fleming et al. 1995) and CA IX is expressed on the basolateral membrane of the epithelial cells in rat duodenum (Saarnio et al. 1998; Hilvo et al. 2004), these CAs may also affect transmucosal CO2 absorption and its submucosal conversion, as suggested by the inhibition of CO2 and H+ absorption and PV PCO2 increase by i.v. ACZ. Duodenal apical H+ uptake was formerly thought to be in exchange for cellular Na+, presumably reflecting apical NHE activity (Wormsley, 1968; Shaffer & Winship, 1971; Winship et al. 1972; Gurll et al. 1976), which would correspond to NHE3. The lack of effect of S3226 on CO2 absorption or on PV acidification does not support this hypothesis. Rather, the NHE1 inhibitor DMA inhibited luminal CO2 loss and PV acidification during high CO2 challenge, whereas NHE3 inhibitor S3226 had no effect, further supporting our hypothesis that CO2–H+ movement across the apical membrane is due to CO2 diffusion into the cell and H+ exit from the cells due to NHE1 activity. Although the duodenum fails to absorb all gastric acid in Zollinger-Ellison syndrome, leading to distal intestinal injury (Meko & Norton, 1995), its overall capacity to absorb fluid and electrolytes is normal (Rambaud et al. 1978). Luminal CO2 acidifies epithelial cells within ~1 min by an MTZ-inhibitable mechanism (Furukawa et al. 2005), demonstrating that the duodenal epithelial apical membrane is CO2 permeable. The imposition of a CO2 gradient across the mucosa rapidly increased CO2 loss from the lumen, followed by recovery, and then steadily increasing CO2 loss. The initial rapid increase of CO2 absorption probably reflected the large CO2 gradient applied across the apical membrane, followed by diminished CO2 absorption, probably reflecting a diminished lumen–cell CO2 gradient, due to cellular acidification, which converts cellular HCO3− to CO2. The gradually increasing CO2 loss represents steady-state CO2 loss. The phases of cellular acidification coincide with the mucosal hyperaemic response to CO2 stress (Akiba et al. 2006). Finally, CO2 loss from the lumen was mirrored by increased PV PCO2 and PV Δδ13C/12C, suggesting that luminal CO2 is transported across the mucosa, although there is no net [CO2]tot increase in PV. In conclusion, the duodenum is the major locus of intestinal CO2 and H+ disposal, that enables the completion of the carbon cycle that starts with CO2 uptake into parietal cells, followed by conversion to H+ and HCO3−, secretion of H+ into the gastric lumen, with secretion of HCO3− into the gastric venous circulation (alkaline tide), duodenal mixing of H+ and HCO3− with release of CO2, absorption of CO2, and transport of CO2 to the gastric parietal cell. Epithelial cytosolic and extracellular CAs cooperatively facilitate CO2 absorption, resulting in net H+ absorption by the rat duodenum. | |||||||||||||||||||
Acknowledgments We thank Rebecca Cho for her assistance with manuscript preparation and Ira Kurtz, MD, for his helpful comments. This work was supported by a Department of Veterans Affairs Merit Review Award, NIH-NIDDK RO1 DK54221 (J.K.), and the animal core of NIH-NIDDK P30 DK0413 (J.E.Rozengurt). | |||||||||||||||||||
References
| |||||||||||||||||||
