Cellular bicarbonate protects rat duodenal mucosa from acid-induced injury

Secretion of bicarbonate from epithelial cells is considered to be the primary mechanism by which the duodenal mucosa is protected from acid-related injury. Against this view is the finding that patients with cystic fibrosis, who have impaired duodenal bicarbonate secretion, are paradoxically protected from developing duodenal ulcers. Therefore, we hypothesized that epithelial cell intracellular pH regulation, rather than secreted extracellular bicarbonate, was the principal means by which duodenal epithelial cells are protected from acidification and injury. Using a novel in vivo microscopic method, we have measured bicarbonate secretion and epithelial cell intracellular pH (pH_i), and we have followed cell injury in the presence of the anion transport inhibitor DIDS and the Cl^– channel inhibitor, 5-nitro-2-(3-phenylpropylamino) benzoic acid (NPPB). DIDS and NPPB abolished the increase of duodenal bicarbonate secretion following luminal acid perfusion. DIDS decreased basal pH_i, whereas NPPB increased pH_i; DIDS further decreased pH_i during acid challenge and abolished the pH_i overshoot over baseline observed after acid challenge, whereas NPPB attenuated the fall of pH_i and exaggerated the overshoot. Finally, acid-induced epithelial injury was enhanced by DIDS and decreased by NPPB. The results support the role of intracellular bicarbonate in the protection of duodenal epithelial cells from luminal gastric acid.

The duodenal mucosa is regularly exposed to intermittent pulses of gastric acid, with luminal acidity varying rapidly between pH 2 and 7 (1). Without protective mechanisms in place, duodenal cells, as is the case with other cells in the upper gastrointestinal tract, are believed to irreversibly acidify in the presence of acidic luminal contents, injuring the epithelium (2). With the measurement of a neutral pH in the juxtamucosal mucus gel in the presence of luminal acid, the bicarbonate secretion hypothesis was developed, wherein bicarbonate, secreted by the epithelial cells, completely neutralized luminal acid diffusing through the mucus gel toward the epithelium (3). Correlation of bicarbonate secretion with mucosal protection from acid-related injury further bolstered this hypothesis (4–6). Recently, with the cloning of transport proteins and more sophisticated understanding of the bicarbonate secretory process, emphasis has been placed on cellular bicarbonate uptake and exit in the context of duodenal mucosal defense (7, 8). Furthermore, luminal acid upregulates other putative defense mechanisms, such as mucosal blood flow and mucus gel secretion, suggesting that bicarbonate secretion is part of a multicomponent defensive system (9–11). Nevertheless, the currently accepted primary duodenal mucosal defense factor is indeed secreted bicarbonate (12).

In our studies of duodenal epithelial pH_i measured in vivo, we made several observations, coupled with the analysis of others, that suggested that this hypothesis is not sufficient to fully account for duodenal cell protection. The first observation was that duodenal epithelial cells readily acidified when the mucosa was superfused with mildly acidic (e.g., pH 4.5) solutions (13), which indicated that acid was not completely neutralized in the mucus gel and could enter the epithelial cells. Second, after observing an increase of duodenal cell buffer capacity in response to perfused acid, we hypothesized that intracellular, rather than secreted extracellular bicarbonate, played an important role in protecting duodenal cells from acid injury (14). Third, we found that bicarbonate secretion did not increase when luminal acid was present, casting further doubt on the role of extracellular bicarbonate in acute mucosal protection from acid (13). Fourth, Wormsley has argued that complete neutralization of luminal acid is possible only with a vast molar excess of bicarbonate, as is the case in general when a strong mineral acid is titrated with a volatile base (15). Fifth, on the basis of mathematical modeling of pH within the mucus gel exposed to secreted bicarbonate and luminal acid, Engel and colleagues have suggested that the dimensions of a neutral layer within the gel would be too thin to afford substantial mucosal protection (16). Finally, neutralization of acid with secreted bicarbonate generates high levels of carbon dioxide, leaving the duodenal cells with an additional volatile acid stress. This phenomenon also argues against the importance of extracellular acid neutralization per se as the predominant protective mechanism (17, 18).

To determine the relative importance of the bicarbonate secretion hypothesis, we first devised experiments that were designed to “uncouple” bicarbonate secretion from mucosal protection. We chose to use two inhibitors, 4,4′diisothiocyanostilbene-2,2′-disulfonic acid (DIDS) and 5-nitro-2-(3-phenylpropylamino) benzoic acid (NPPB), both known to inhibit epithelial bicarbonate secretion (7, 19, 20), which were predicted to have divergent effects on pH_i regulation and mucosal protection. DIDS, by enhancing acidification and abolishing the pH_i overshoot following a pulse of perfused acid (14) behaves as if it were inhibiting a bicarbonate uptake mechanism, which was most likely the Na⁺:HCO₃^– cotransporter (NBC1), thought to be present in duodenal epithelial cells (7). NPPB, an anion channel inhibitor, was chosen because it was predicted to inhibit bicarbonate exit across the epithelial cell apical membrane, as is the case when the cystic fibrosis transmembrane conductance regulator (CFTR) chloride channel is dysfunctional (21, 22). Inhibition of apical bicarbonate exit may help preserve cellular buffering power by increasing intracellular bicarbonate concentration and hence buffering power, protecting the cells from acidification and consequent injury. If neutralization of gastric acid in the duodenal mucus gel is the primary mucosal defense mechanism, both inhibitors, by inhibiting bicarbonate secretion, should enhance mucosal injury equally. In contrast, if prevention of a marked reduction of intracellular bicarbonate/pH_i regulation is the prime protective mechanism, NPPB, which inhibits apical bicarbonate secretion, and DIDS, which inhibits basolateral bicarbonate uptake, should have opposite effects on duodenal cell pH_i and viability. To distinguish between these two mechanisms, we studied the effects of DIDS and NPPB on pH_i, bicarbonate secretion, and mucosal injury in the duodenum of living rats.

Effect of DIDS and NPPB on acid-induced bicarbonate secretion.

We first confirmed the effect of DIDS and NPPB on duodenal bicarbonate secretion by titration of the effluent of a perfused duodenal loop with a pH-stat, a measure of bicarbonate secretion. Figure 1 depicts that, following a 10-minute acid challenge, luminal alkalization gradually increased, consistent with increased bicarbonate secretion. DIDS and NPPB abolished the increase of bicarbonate secretion after acid exposure, confirming the importance of DIDS-and NPPB-sensitive transport processes in rat duodenal bicarbonate secretion.

Figure 1

Effect of DIDS and NPPB on acid-induced bicarbonate secretion. Titratable alkalinity (bicarbonate secretion) was measured by the pH-stat method in duodenal loop perfusion experiments. Both DIDS and NPPB inhibit the increased bicarbonate secretion following acid exposure. All data are expressed as mean ± SEM from six rats. *P < 0.05 vs. pH 7.0 saline perfusion, ^†P < 0.05 vs. pH 2.2 saline perfusion.

Effect of DIDS and NPPB on pH_i with or without luminal acid.

We then measured the effects of DIDS and NPPB on duodenal pH_i. Figure 2 depicts the effects of DIDS and NPPB on pH_i in the presence or absence of perfused luminal acid. Without luminal acid (Figure 2a), DIDS acidified duodenal epithelial cells, consistent with inhibition of cellular bicarbonate influx, whereas NPPB alkalinized the cells, consistent with inhibition bicarbonate exit. Divergent effects of DIDS and NPPB were also observed in the presence of perfused acid (Figure 2b). Perfused acid rapidly acidified the duodenal epithelial cells, followed by recovery of pH_i to baseline (t = 25). Following removal of the perfused acid, pH_i increased above the baseline value (overshoot) (t = 30, 35) as described previously (14). DIDS enhanced the intracellular acidification during the luminal acid challenge and abolished the overshoot following acid removal. These effects of DIDS are consistent with the inhibition of a bicarbonate influx process. In contrast, NPPB attenuated the decrease of pH_i during the luminal acid challenge and enhanced the overshoot of pH_i following acid removal, consistent with inhibition of cellular bicarbonate exit.

Figure 2

Effects of DIDS and NPPB on baseline pH_i and on pH_i during acid challenge of duodenal epithelial cells. (a) During perfusion with pH 7.0 solution, DIDS acidifies, whereas NPPB alkalinizes cells. (b) DIDS enhances, whereas NPPB attenuates acid-induced intracellular acidification during acid exposure. Note that DIDS abolishes, but NPPB enhances pH_i overshoot following acid challenge. *P < 0.05 vs. pH 7.0 Krebs group, ^†P < 0.05 vs. pH 2.2 Krebs group. All data are expressed as mean ± SEM from six rats.

Immunolocalization of the electrogenic sodium bicarbonate cotransporter pNBC1.

Since bicarbonate influx into duodenal cells by electrogenic sodium bicarbonate cotransport is thought to play an important role both in pH_i regulation following acute acid loading and in transepithelial bicarbonate secretion, we immunolocalized NBC1 protein variants in rat duodenum. For this purpose, rabbit affinity-purified polyclonal Ab’s were raised against the specific N-terminus of pNBC1 and kNBC1 variants of NBC1 (27). The pNBC1 Ab recognized a major band at approximately 150 kDa on immunoblots using membrane fractions from rat duodenum (Figure 3). The specificity of the labeling was confirmed using the pNBC1 Ab preabsorbed with the immunizing peptide. Moreover, data from pancreatic membranes, which strongly express pNBC1 (approximately 150-kDa band), is shown for comparison. The size of the recognized protein in all tissues is larger than the predicted size of approximately 120 kDa, suggesting that pNBC1 is glycosylated or posttranslationally modified in other ways. In duodenum, a minor band of approximately 90 kDa was also detected. Unlike pNBC1, kNBC1 was not detected in the rat duodenum. The results of the immunocalization studies are shown in Figure 4. pNBC1 was highly expressed on duodenal epithelial cells on the basolateral membrane. In contrast, kNBC1 was not detectable (not shown). These results demonstrate that pNBC1 is the basolateral electrogenic sodium bicarbonate cotransporter in rat duodenum.

Figure 3

Western blot of rat duodenum membrane proteins probed with pNBC1-specific (a) and kNBC1-specific (b) Ab’s. Rat duodenum membrane proteins (80 μg) were loaded in lanes 1 and 2 in both a and b. (a) Rat pancreas membrane proteins (25 μg)were loaded in lanes 3 and 4. (b) Rat kidney membrane proteins (100 μg) were loaded in lanes 3 and 4. The pNBC1-specific Ab and the kNBC1-specific Ab were used at dilution 1:1,000. For lanes 2 and 4, pNBC1 (a) and kNBC1 (b) Ab’s preincubated with the corresponding specific peptide (1:100) were used.

Figure 4

Immunolocalization of pNBC1 in rat duodenum. (a) pNBC1 is localized to the basolateral membranes of duodenal villus epithelial cells appearing as a cup-shaped structure (arrows). (b) When the pNBC1 Ab was preincubated with a specific peptide (1:100), staining was suppressed. (c) An image taken with Nomarski optics is included to provide orientation. The white rectangle encloses the same area shown in a. Calibration bar, 50 μm.

Detection of acid-induced epithelial injury with PI staining in vivo.

Initial studies were performed to address the feasibility of using PI staining to measure duodenal epithelial injury in response to a brief pulse of luminal acid. Figure 5 depicts images of PI-stained mucosa generated in vivo (Figure 5, a–c) and images of corresponding frozen sections (Figure 5, d–f). Figure 5, a and d, depicts mucosae exposed to pH 7.0; Figure 5, b and e, depicts mucosae 15 minutes after a 5-minute exposure to pH 1.8; and Figure 5, c and f, depicts mucosae 5 minutes after a 5-minute exposure to pH 1.0. During perfusion with neutral solution, only rare PI staining was visible at t = 30 minutes (Figure 5a), though submucosal PI-positive structures were visible in sections (Figure 5d). The red appearance of the villi in (Figure 5a) was due to BCECF fluorescence observed through a 590-nm filter. Fifteen minutes after a 5-minute exposure to a pH 1.8 solution, brightly fluorescent PI-positive dots, corresponding to PI-stained nuclei, were observed over the villi (t = 30 minutes; Figure 5b). Five minutes after a 5-minute pH 1.0 challenge, numerous PI-positive nuclei were observed, producing diffuse PI-positive staining over the villous tips (t = 20 minutes; Figure 5c). Frozen sections of acid-exposed mucosa (Figure 5, e and f) revealed that PI-related fluorescence was localized to the nuclei of the villous tip epithelial cells and the nuclei of detached epithelial cells above the villi, but not the middle portion of the villus or the crypts (not shown). Villous ballooning was additionally observed after exposure to pH 1.0 (t = 20 minutes; Figure 5f).

Figure 5

PI-positive staining in vivo and effect of acid. (a and d) A low number of visible PI-positive nuclei were visible 30 minutes after perfusion with pH 7.0 solution (t = 30 minutes). The red color of the villi is BCECF fluorescence observed through a 590-nm filter (thick arrows). (b and e) Fifteen minutes after a 5-minute pulse of pH 1.8 perfusate (t = 30 minutes), PI-positive dots are seen readily, localized to the villous tips (thin arrows). (c and f) Five minutes after a 5-minute pulse of pH 1.0 perfusate (t = 20 minutes), the villous tips have ballooned, and some are detached (yellow arrows). Note that all of the epithelial cells in the ballooned and detached villi are PI positive. Calibration bar, 100 μm. a, d, e, and f were printed so as to emphasize the faint staining of the villi.

Effect of DIDS and NPPB on acid-induced epithelial injury

In vivo PI staining. Having devised a means of reproducibly measuring acid-related injury by in vivo microscopy, we then measured injury in the presence of DIDS and NPPB to determine if bicarbonate secretion could be uncoupled from acid-related mucosal injury. Figure 6 depicts the time course of in vivo PI staining before and after a 5-minute pH 1.8 exposure. PI-positive cells gradually increased in number in the villous cells. In control studies, pH 7.0 Krebs perfusion slightly increased the number of PI-positive cells during a 30-minute experiment. To determine the effect of each inhibitor on cell viability, we used preperfusions of slightly different pH to best demonstrate their respective effects. A 5-minute exposure to pH 2.0 or pH 1.8 gradually increased the number of PI-positive cells (Figure 6, a and b). DIDS further increased the number of PI-positive cells, consistent with enhancement of the mucosal susceptibility to acid injury (Figure 6a). Conversely, NPPB reduced the number of PI-positive cells observed after a 5-minute perfusion with a pH 1.8 solution (Figure 6b).

Figure 6

Effect of DIDS and NPPB on acid-induced duodenal injury (PI study). (a) Effect of DIDS. A 5-minute pulse of pH 2.0 injures the mucosa, as measured by a progressive increase of PI-positive nuclei. DIDS further increased the number of PI-positive cells induced by a pH 2.0 pulse. *P < 0.05 vs. pH 7.0 Krebs group, ^†P < 0.05 vs. pH 2.0 Krebs group. (b) Effect of NPPB. NPPB lessens mucosal injury caused by a 5-minute perfusion with pH 1.8 solution, as assessed by decreased number of PI-positive cells. *P < 0.05 vs. pH 7.0 Krebs group, ^†P < 0.05 vs. pH 1.8 Krebs group. All data are expressed as mean ± SEM from six rats.

Conventional histology. To help confirm the injury studies, routine histology was performed on sections obtained from duodenal mucosa exposed to the same conditions used for the PI experiments. As seen in Figure 7, exposure to pH 7.0 (Figure 7a), pH 2.2 (Figure 7b), or pH 2.0 (Figure 7c) perfusate did not produce microscopic damage to the villi. Focal necrosis, ballooning, and sloughing of villi represent increasing severity of injury with increasing [H⁺] (26, 28). Exposure to pH 2.0 perfusate (Figure 7c) produced no visible damage; the addition of 0.5 mM DIDS to the pH 2.0 perfusate was associated with focal necrosis of the villus tips (Figure 7e). More severe damage, with frank sloughing of the villous tips, was observed with pH 1.8 perfusion (Figure 7d). The injury was attenuated to ballooning by the presence of NPPB (Figure 7f). Thus, routine histologic examination confirmed that DIDS augmented acid-induced injury, whereas NPPB diminished injury. The percentage of the damaged villi is shown in Table 1. As seen in Table 1, DIDS alone did not increase the injury score, but increased the injury score in the presence of acid. NPPB in the presence of acid decreased the injury score, consistent with the data obtained with PI.

Figure 7

Effect of DIDS and NPPB on acid-induced duodenal injury (H&E study). Injury was assessed in control and acid-exposed duodena with the use of conventional histologic sections. (a) Duodenal villi perfused with pH 7.0 solution, (b) pH 2.2 solution, (c) pH 2.0 solution, (d) pH 1.8 solution, (e) pH 2.0 + DIDS, and (f) pH 1.8 + NPPB. Note that DIDS increased mucosal injury, as manifest by focal necrosis seen in e (thin arrows), compared with no injury seen with pH 2.0 alone (c). The sloughing caused by pH 1.8 perfusion (thick arrows), seen in d, was not observed in the presence of NPPB (f), where ballooning (arrowheads), a more mild form of injury, was observed (H&E, ×400; calibration bar, 100 μm).

These results, which are consistent with the pH_i studies, further indicate that attenuation of a marked fall of pH_i/bicarbonate during an acid challenge is imperative for protection from mucosal injury. Furthermore, the divergent effects of DIDS and NPPB on mucosal injury, despite both abolishing bicarbonate secretion, would be predicted only if bicarbonate secretion was not a primary mucosal protective process.

We studied the importance of intracellular bicarbonate/pH_i regulation in the protection of duodenal epithelial cells from luminal acid challenge. Both DIDS and NPPB abolished bicarbonate secretion after a luminal acid challenge. DIDS and NPPB, however, had divergent effects on pH_i at baseline (neutral perfusion) and during acid exposure. In complementary studies, DIDS exaggerated, whereas NPPB lessened, acid-induced epithelial injury. Immunolocalization studies identified pNBC1 in the duodenal epithelial cell basolateral membrane as a potentially important DIDS-sensitive bicarbonate influx pathway that plays a role in the prevention of duodenal epithelial injury. This is the first study showing that two inhibitors of bicarbonate secretion have opposite effects on pH_i and mucosal protection, suggesting that intracellular bicarbonate/pH_i regulation plays an important role in the protection of duodenal cells from a luminal acid load.

DIDS inhibits bicarbonate secretion in bicarbonate-secreting epithelia (19, 20). The acidifying effect of DIDS on pH_i, however, is less easily understood. DIDS inhibits several anion transporters thought to be present on duodenal epithelial cells. Although early studies identified an anion exchange process in membrane vesicles derived from the intestinal brush border (29), the AE2 anion exchanger isoform has been immunolocalized to the basolateral membrane of intestinal, including duodenal, epithelial cells (30). Given the known gradients for chloride and bicarbonate, AE2 would be predicted to mediate basolateral bicarbonate efflux in the absence of luminal acid. Moreover, inhibition of AE2 with DIDS would be predicted to alkalinize the cells, as was found in isolated murine duodenocytes (8), in contrast to the findings we report in the current study. Nevertheless, our study confirms the original measurement of epithelial cell pH_i in isolated Necturus duodenum, in which Kivilaakso and coworkers observed that 4-acetamido-4′-isothiocyano-2,2′-disulfonic acid stilbene (SITS), an inhibitor closely related to DIDS, added to the serosal bath–acidified duodenal epithelial cells (2). An additional anion exchanger, downregulated in adenoma gene product (DRA), is present in the apical membrane of intestinal and colonic epithelial cells (31) and is believed to mediate intestinal and colonic fluid absorption. DRA, however, may only be variably sensitive to DIDS (32). The presence of NBC1 in duodenal epithelial cells has been demonstrated functionally previously and by PCR (7). Recently, NBC1 and a related electroneutral cotransporter NBCn1 were identified in the basolateral membrane of villus epithelial cells of murine duodenum (8). Using a pNBC1-specific Ab, our studies demonstrate that rat duodenal cells express the pNBC1 protein variant of the NBC1 gene (33, 34) on their basolateral membrane. Furthermore, we were unable to detect the kNBC1 N-terminal variant in duodenal cells either by Western blot analysis or immunocytochemistry using a kNBC1-specific Ab. Finally, it is also possible that DIDS inhibited outwardly rectifying basolateral Cl^– channels (35, 36), with the resultant hyperpolarization inhibiting bicarbonate uptake via pNBC1. Therefore, the effect of DIDS on pH_i and on duodenal cell–mediated bicarbonate secretion is in keeping with its inhibition of basolateral bicarbonate influx via pNBC1 or a related, DIDS-sensitive bicarbonate uptake mechanism. Although it was not possible to reliably add DIDS to the basolateral side directly in the in vivo preparation used in this study, our results, taken in the context of previous studies of isolated duodenal preparations (2), are best explained by postulating that DIDS permeated through intercellular junctions and inhibited basolateral pNBC1. This explanation is further supported by the observation that the duodenum is a low-resistance, leaky epithelium that can be penetrated by numerous small organic compounds such as capsaicin and bradykinin (11, 37). Moreover enhancement of intracellular acidification during acid challenge and abolition of the overshoot after acid removal by luminal DIDS application are most consistent inhibition of bicarbonate uptake. We are unable, however, at the present time to exclude additional effects of DIDS on apical transport mechanisms in these experiments. We also note that even though DIDS is fluorescent, its fluorescence at the concentrations and wavelengths used is unlikely to interfere with BCECF-based pH_i measurements (38).

We developed a method in which cell injury could be assessed dynamically in vivo by using fluorescence microscopic methodology to detect PI staining. As shown previously, acid-induced mucosal injury in rat duodenum is a function of perfusate pH and exposure time (28). Although only a few previous studies have used PI staining to assess injury in vivo (24, 39, 40), we chose this technique since it provides a useful method for dynamically quantifying the luminal response to acid perfusion in vivo. Our results compare favorably with the data of Livingston et al., where a conventional histological approach was used to demonstrate duodenal cell injury during luminal perfusion with solutions of pH less than 2.0 (28). The gradual increase in the number of PI-positive cells during perfusion with neutral solutions reflects, in part, the spontaneous turnover of villous tip cells (41). Acidic solutions of pH 2.0 and 1.8 were chosen for the current duodenal cell injury studies since these pH values are close to the physiological range and were based on the ability to produce a reproducible increase of nuclear staining without engendering immediate or severe cell necrosis. As seen in the conventional histologic sections, and in agreement with previous studies (26, 28), early villus injury progresses from blebbing of the apical membrane to ballooning of the villus tips through sloughing. The overall injury scores confirmed the data obtained with the PI technique that DIDS enhanced injury, whereas NPPB decreased injury. Conventional injury studies hence confirmed that DIDS enhanced, whereas NPPB lessened, the amount of visible acid-related damage. The demonstration that the observed damage is preceded by irreversible cellular acidification would further support the hypothesis concerning the etiology of duodenal epithelial injury. We did not measure pH_i in damaged cells in our system since pH_i measurements in the presence of injury are unreliable due to dye leakage and since the lower limit of reliability with BCECF is approximately pH 6.2. In the absence of acid perfusion, DIDS mildly and reversibly acidified the epithelial cells, similar to the effects of perfusion with pH 4.5 solution, which is not injurious (13, 28). The lack of injury with DIDS at neutral pH, which was predicted from the pH_i data, lends further confidence to the correlation between pH_i and injury measured with PI and conventional histology.

Recent studies of CFTR indicate that this channel plays a major role in bicarbonate secretion (21, 42, 43). This channel, present in pancreatic ducts, duodenal epithelial cells, and Brunner’s glands (44), is closely associated with epithelial HCO₃^– secretion. NPPB, a blocker of the CFTR Cl^– channel, inhibits bicarbonate secretion stimulated by cGMP agonists in rats (45) and by cAMP in mice (42), suggesting that bicarbonate exits from the cells through the CFTR itself or through CFTR-associated transporters. The mechanism by which bicarbonate is secreted across epithelia is controversial. CFTR may serve as a HCO₃^– conductance in human airway epithelia (46); alternatively, it may alter the electrical gradient for electrogenic bicarbonate entry, consistent with data obtained in studies of mouse pancreatic ducts (47–49). The former mechanism is compatible with the results of the current study. Specifically, NPPB, by inhibiting CFTR function, diminishes bicarbonate secretion by inhibiting exit across the apical membrane. The absence of bicarbonate exit might reset intracellular buffering at a higher level, protecting the cells from acidification due to acid loads. This contention is substantiated by the observation that duodenal epithelial cells obtained from patients with cystic fibrosis (CF) have impaired pH_i recovery from alkaline loads in chloride-free medium (50). Furthermore, our results are also consistent with the presence of a CFTR-like, NPPB-sensitive bicarbonate-selective anion channel in the apical membrane, as has been described recently for guinea pig gallbladder (51).

One of the most interesting aspects of this study was that DIDS, by presumably inhibiting bicarbonate uptake across the basolateral membrane, exaggerates perfused acid-induced epithelial cell injury measured by PI staining, whereas NPPB, which presumably inhibits bicarbonate exit across the apical membrane, protects the cells from acid-induced injury. The importance of this observation does not lie in the mechanism of action of each inhibitor, but rather in that a compound such as NPPB can, on the one hand, inhibit bicarbonate secretion while, on the other hand, can protect duodenal cells from acidification and acid-induced injury. This decoupling of bicarbonate secretion and mucosal protection contradicts previous animal and human studies (4, 6, 52), although selective inhibition of apical bicarbonate exit may be an unusual means by which transepithelial bicarbonate secretion can be altered.

The lack of duodenal ulcer disease among patients with CF may be explained by similar logic. Although gastroduodenal pH is low and pancreatic and duodenal bicarbonate secretion is subnormal in patients with CF (53–55), these patients have a paradoxical low incidence of peptic ulcer disease (56–58). This low incidence of peptic ulcer disease is not ascribable to an absence of upper gastrointestinal acid-peptic disease, since severe acid-peptic esophageal injury is common (59, 60). Furthermore, CF patients, who are frequently endoscoped for upper gastrointestinal symptoms, have characteristic duodenal mucosal abnormalities (61), and can be infected with Helicobacter pylori (62). These observations further support our “intracellular bicarbonate” hypothesis: diminished apical bicarbonate exit with intact basolateral bicarbonate influx increases the steady state intracellular bicarbonate/pH_i in the vulnerable duodenal epithelial cells in these patients, protecting them from injury due to luminal acid.

In conclusion, we have presented data that support our hypothesis that intracellular pH regulatory mechanisms are an important means by which duodenal epithelial cells are protected from luminal acid, part of an intricate defense system that includes secretion of HCO₃^– from the pancreas and possibly Brunner’s glands, and mucus from the epithelial cells and Brunner’s glands. As a potent bicarbonate influx pathway, basolateral pNBC1 is likely to be a component of the protective mechanism. The protective effect of NPPB during luminal acid loading may serve as a useful model for studying the cellular processes responsible for preventing ulcers in CF patients.

This work was supported by VA Merit Review Funding and National Institutes of Health grants R01-DK54221, DK07789, and DK85863. We would like to thank Dipty Shah for her technical assistance.

See the related Commentary beginning on page 1743.

Yasutada Akiba’s present address is: Department of Internal Medicine, School of Medicine, Keio University, Tokyo, Japan.

1. Rhodes, J, Apsimon, HT, Lawrie, JH. pH of the contents of the duodenal bulb in relation to duodenal ulcer. Gut 1966. 7:502-508.
   View this article via: PubMed CrossRef Google Scholar
2. Paimela, H, Kiviluoto, T, Mustonen, H, Kivilaakso, E. Intracellular pH of isolated Necturus duodenal mucosa exposed to luminal acid. Gastroenterology 1992. 102:862-867.
   View this article via: PubMed Google Scholar
3. Garner, A, Flemström, G, Allen, A, Heylings, JR, McQueen, S. Gastric mucosal protective mechanisms: roles of epithelial bicarbonate and mucus secretions. Scand J Gastroenterol Suppl 1984. 101:79-86.
   View this article via: PubMed Google Scholar
4. Isenberg, JI, Selling, JA, Hogan, DL, Koss, MA. Impaired proximal duodenal mucosal bicarbonate secretion in patients with duodenal ulcer. N Engl J Med 1987. 316:374-379.
   View this article via: PubMed Google Scholar
5. Hogan, DL, et al. Duodenal bicarbonate secretion: eradication of Helicobacter pylori and duodenal structure and function in humans. Gastroenterology 1996. 110:705-716.
   View this article via: PubMed CrossRef Google Scholar
6. Wenzl, E, Feil, W, Starlinger, M, Schiessel, R. Alkaline secretion: a protective mechanism against acid injury in rabbit duodenum. Gastroenterology 1987. 92:709-715.
   View this article via: PubMed Google Scholar
7. Jacob, P, et al. Role of Na⁺HCO₃^– cotransporter NBC1, Na^+/H⁺ exchanger NHE1, and carbonic anhydrase in rabbit duodenal bicarbonate secretion. Gastroenterology 2000. 119:406-419.
   View this article via: PubMed CrossRef Google Scholar
8. Praetorius, J, et al. Molecular and functional evidence for electrogenic and electroneutral Na⁺HCO₃^– cotransporters in murine duodenum. Am J Physiol Gastrointest Liver Physiol 2001. 280:G332-G343.
   View this article via: PubMed Google Scholar
9. Allen, A, Flemström, G, Garner, A, Kivilaakso, E. Gastroduodenal mucosal protection. Physiol Rev 1993. 73:823-857.
   View this article via: PubMed Google Scholar
10. Akiba, Y, Guth, PH, Engel, E, Nastaskin, I, Kaunitz, JD. Dynamic regulation of mucus gel thickness in rat duodenum. Am J Physiol Gastrointest Liver Physiol 2000. 279:G437-G447.
    View this article via: PubMed Google Scholar
11. Akiba, Y, Guth, PH, Engel, E, Nastaskin, I, Kaunitz, JD. Acid-sensing pathways of rat duodenum. Am J Physiol Gastrointest Liver Physiol 1999. 277:G268-G274.
12. Flemström, G, Isenberg, JI. Gastroduodenal mucosal alkaline secretion and mucosal protection. News Physiol Sci 2001. 16:23-28.
    View this article via: PubMed Google Scholar
13. Akiba, Y, et al. Acute adaptive cellular base uptake in rat duodenal epithelium. Am J Physiol Gastrointest Liver Physiol 2001. 280:G1083-G1092.
    View this article via: PubMed Google Scholar
14. Akiba, Y, Kaunitz, JD. Regulation of intracellular pH and blood flow in rat duodenal epithelium in vivo. Am J Physiol Gastrointest Liver Physiol 1999. 276:G293-G302.
15. Wormsley, KG. What happens to acid in the duodenum? Gastroenterology 1968. 55:441-443.
    View this article via: PubMed Google Scholar
16. Engel, E, Guth, PH, Nishizaki, Y, Kaunitz, JD. Barrier function of the gastric mucus gel. Am J Physiol Gastrointest Liver Physiol 1995. 269:G994-G999.
17. Rune, SJ, Henriksen, FW. Carbon dioxide tensions in the proximal part of the canine gastrointestinal tract. Gastroenterology 1969. 56:758-762.
18. Holm, M, Johansson, B, Pettersson, A, Fändriks, L. Carbon dioxide mediates duodenal mucosal alkaline secretion in response to luminal acidity in the anesthetized rat. Gastroenterology 1998. 115:680-685.
    View this article via: PubMed CrossRef Google Scholar
19. Nyberg, L, et al. Human proximal duodenal alkaline secretion is mediated by Cl^–/HCO₃^– exchange and HCO₃^– conductance. Dig Dis Sci 1998. 43:1205-1210.
    View this article via: PubMed CrossRef Google Scholar
20. Feldman, GM. HCO₃^– secretion by rat distal colon: effects of inhibitors and extracellular Na⁺. Gastroenterology 1994. 107:329-338.
    View this article via: PubMed Google Scholar
21. Hogan, DL, et al. Acid-stimulated duodenal bicarbonate secretion involves a CFTR-mediated transport pathway in mice. Gastroenterology 1997. 113:533-541.
    View this article via: PubMed CrossRef Google Scholar
22. Ulrich, CD. Bicarbonate secretion and CFTR: continuing the paradigm shift. Gastroenterology 2000. 118:1258-1261.
    View this article via: PubMed CrossRef Google Scholar
23. Weinlich, M, Heydasch, U, Starlinger, M, Kinne, RK. Intracellular pH-measurements in rat duodenal mucosa in vitro using confocal laserscan microscopy. Z Gastroenterol 1997. 35:263-270.
    View this article via: PubMed Google Scholar
24. Takase, S, Lerond, L, Bergan, JJ, Schmid-Schonbein, GW. The inflammatory reaction during venous hypertension in the rat. Microcirculation 2000. 7:41-52.
    View this article via: PubMed CrossRef Google Scholar
25. Lawlor, DK, Brock, RW, Harris, KA, Potter, RF. Cytokines contribute to early hepatic parenchymal injury and microvascular dysfunction after bilateral hindlimb ischemia. J Vasc Surg 1999. 30:533-541.
    View this article via: PubMed CrossRef Google Scholar
26. Leung, FW, Miller, JC, Reedy, TJ, Guth, PH. Exogenous prostaglandin protects against acid-induced deep mucosal injury by stimulating alkaline secretion in rat duodenum. Dig Dis Sci 1989. 34:1686-1691.
    View this article via: PubMed CrossRef Google Scholar
27. Bok, D, et al. Immunolocalization of electrogenic sodium-bicarbonate cotransporters pNBC1 and kNBC1 in the rat eye. Am J Physiol 2001. 281:F920-F935.
28. Livingston, EH, Passaro, EP, Miller, J, Guth, PH. Spectrum of injury produced in the duodenum by perfusion with luminal acid in the rat. Gastroenterology 1992. 103:481-489.
    View this article via: PubMed Google Scholar
29. Knickelbein, RG, Aronson, PS, Dobbins, JW. Membrane distribution of sodium-hydrogen and chloride-bicarbonate exchangers in crypt and villus cell membranes from rabbit ileum. J Clin Invest 1988. 82:2158-2163.
    View this article via: JCI PubMed CrossRef Google Scholar
30. Alper, SL, et al. Expression of AE2 anion exchanger in mouse intestine. Am J Physiol Gastrointest Liver Physiol 1999. 277:G321-G332.
31. Byeon, MK, Frankel, A, Papas, TS, Henderson, KW, Schweinfest, CW. Human DRA functions as a sulfate transporter in Sf9 insect cells. Protein Expr Purif 1998. 12:67-74.
    View this article via: PubMed CrossRef Google Scholar
32. Melvin, JE, Park, K, Richardson, L, Schultheis, PJ, Shull, GE. Mouse down-regulated in adenoma (DRA) is an intestinal Cl^–/HCO₃^– exchanger and is up-regulated in colon of mice lacking the NHE3 Na⁺/H⁺ exchanger. J Biol Chem 1999. 274:22855-22861.
    View this article via: PubMed CrossRef Google Scholar
33. Abuladze, N, et al. Structural organization of the human NBC1 gene: kNBC1 is transcribed from an alternative promoter in intron 3. Gene 2000. 251:109-122.
    View this article via: PubMed CrossRef Google Scholar
34. Abuladze, N, et al. Molecular cloning, chromosomal localization, tissue distribution, and functional expression of the human pancreatic sodium bicarbonate cotransporter. J Biol Chem 1998. 273:17689-17695.
    View this article via: PubMed CrossRef Google Scholar
35. Monaghan, AS, Mintenig, GM, Sepulveda, FV. Outwardly rectifying Cl^– channel in guinea pig small intestinal villus enterocytes: effect of inhibitors. Am J Physiol Gastrointest Liver Physiol 1997. 273:G1141-G1152.
36. Mignen, O, Egee, S, Liberge, M, Harvey, BJ. Basolateral outward rectifier chloride channel in isolated crypts of mouse colon. Am J Physiol Gastrointest Liver Physiol 2000. 279:G277-G287.
    View this article via: PubMed Google Scholar
37. Akiba, Y, et al. Sensory pathways and cyclooxygenase regulate mucus gel thickness in rat duodenum. Am J Physiol Gastrointest Liver Physiol 2001. 280:G470-G474.
    View this article via: PubMed Google Scholar
38. Peral, MJ, Calonge, ML, Ilundain, AA. Na⁺-HCO₃^– cotransporter and intracellular pH regulation in chicken enterocytes. Pflügers Arch 1995. 430:612-616.
    View this article via: CrossRef Google Scholar
39. Narayan, P, Mentzer (Jr), RM, Lasley, RD. Annexin V staining during reperfusion detects cardiomyocytes with unique properties. Am J Physiol Heart Circ Physiol 2001. 281:H1931-H1937.
    View this article via: PubMed Google Scholar
40. Suematsu, M, et al. Topographic dissociation between mitochondrial dysfunction and cell death during low-flow hypoxia in perfused rat liver. Lab Invest 1992. 67:434-442.
    View this article via: PubMed Google Scholar
41. Iwanaga, T, Han, H, Adachi, K, Fujita, T. A novel mechanism for disposing of effete epithelial cells in the small intestine of guinea pigs. Gastroenterology 1993. 105:1089-1097.
    View this article via: PubMed Google Scholar
42. Clarke, LL, Harline, MC. Dual role of CFTR in cAMP-stimulated HCO₃^– secretion across murine duodenum. Am J Physiol Gastrointest Liver Physiol 1998. 274:G718-G726.
43. Seidler, U, et al. A functional CFTR protein is required for mouse intestinal cAMP-, cGMP-, and Ca₂⁺-dependent HCO₃^– secretion. J Physiol (Lond) 1997. 505:411-423.
    View this article via: PubMed CrossRef Google Scholar
44. Strong, TV, Boehm, K, Collins, FS. Localization of cystic fibrosis transmembrane conductance regulator mRNA in the human gastrointestinal tract by in situ hybridization. J Clin Invest 1994. 93:347-354.
    View this article via: JCI PubMed CrossRef Google Scholar
45. Guba, M, et al. Guanylin strongly stimulates rat duodenal HCO₃^– secretion: proposed mechanism and comparison with other secretagogues. Gastroenterology 1996. 111:1558-1568.
    View this article via: PubMed CrossRef Google Scholar
46. Illek, B, Yankaskas, JR, Machen, TE. cAMP and genistein stimulate HCO₃^– conductance through CFTR in human airway epithelia. Am J Physiol 1997. 272:L752-L761.
    View this article via: PubMed Google Scholar
47. Shumaker, H, Amlal, H, Frizzell, R, Ulrich, CD, Soleimani, M. CFTR drives Na⁺-nHCO₃^– cotransport in pancreatic duct cells: a basis for defective HCO₃^– secretion in CF. Am J Physiol Gastrointest Liver Physiol 1999. 276:C16-C25.
48. Lee, MG, et al. Regulation of Cl^–/HCO₃^– exchange by cystic fibrosis transmembrane conductance regulator expressed in NIH 3T3 and HEK 293 cells. J Biol Chem 1999. 274:3414-3421.
    View this article via: PubMed CrossRef Google Scholar
49. Lee, MG, et al. Cystic fibrosis transmembrane conductance regulator regulates luminal Cl^–/HCO3^– exchange in mouse submandibular and pancreatic ducts. J Biol Chem 1999. 274:14670-14677.
    View this article via: PubMed CrossRef Google Scholar
50. Pratha, VS, et al. Identification of transport abnormalities in duodenal mucosa and duodenal enterocytes from patients with cystic fibrosis. Gastroenterology 2000. 118:1051-1060.
    View this article via: PubMed CrossRef Google Scholar
51. Meyer, G, Garavaglia, ML, Bazzini, C, Botta, G. An anion channel in guinea pig gallbladder epithelial cells is highly permeable to HCO₃^–. Biochem Biophys Res Commun 2000. 276:312-320.
    View this article via: PubMed CrossRef Google Scholar
52. Vattay, P, et al. Role of passive HCO₃—diffusion in duodenal acid-stimulated alkaline secretion. Acta Physiol Hung 1989. 73:251-255.
    View this article via: PubMed Google Scholar
53. Knauff, RE, Adams, JA. Duodenal fluid pH in cystic fibrosis. Clin Chem 1968. 14:477-479.
    View this article via: PubMed Google Scholar
54. Cox, KL, Isenberg, JN, Ament, ME. Gastric acid hypersecretion in cystic fibrosis. J Pediatr Gastroenterol Nutr 1982. 1:559-565.
    View this article via: PubMed Google Scholar
55. Youngberg, CA, et al. Comparison of gastrointestinal pH in cystic fibrosis and healthy subjects. Dig Dis Sci 1987. 32:472-480.
    View this article via: PubMed CrossRef Google Scholar
56. Lepore, MJ. Cystic fibrosis of the pancreas in the adult. Gastroenterology 1963. 44:696-699.
    View this article via: PubMed Google Scholar
57. Aterman, K. Duodenal ulceration and fibrocystic pancreas disease. Am J Dis Child 1961. 101:210-215.
    View this article via: PubMed Google Scholar
58. Rosenstein, BJ, Perman, JA, Kramer, SS. Peptic ulcer disease in cystic fibrosis: an unusual occurrence in black adolescents. Am J Dis Child 1986. 140:966-969.
    View this article via: PubMed Google Scholar
59. Thomas, D, Rothberg, RM, Lester, LA. Cystic fibrosis and gastroesophageal reflux in infancy. Am J Dis Child 1985. 139:66-67.
    View this article via: PubMed Google Scholar
60. Stringer, DA, et al. The association of cystic fibrosis, gastroesophageal reflux, and reduced pulmonary function. Can Assoc Radiol J 1988. 39:100-102.
    View this article via: PubMed Google Scholar
61. Phelan, MS, et al. Radiographic abnormalities of the duodenum in cystic fibrosis. Clin Radiol 1983. 34:573-577.
    View this article via: PubMed CrossRef Google Scholar
62. Israel, NR, et al. Seroprevalence of Helicobacter pylori infection in cystic fibrosis and its cross-reactivity with anti-pseudomonas antibodies. J Pediatr Gastroenterol Nutr 2000. 30:426-431.
    View this article via: PubMed CrossRef Google Scholar