Are physiological bile acids toxic to the esophagus in an acid environment? Is there a difference with commercial bile as used in experimental studies?
J.W. Harmon (Baltimore), L.F. Johnson (Birmingham)
We know that reflux of gastric duodenal contents into the esophagus is associated with esophageal injury [1-3]. The precise injurious role of the various components of the gastric duodenal contents is still being evaluated. When an endoscopist visualizes green lumenal contents refluxing into the lower esophagus, it is not possible to know the damaging potential of that refluxate. The green color comes from the bilirubin breakdown products and these components of the refluxate have not been shown to cause injury. Therefore, the intensity of the pigment cannot be related to injurious potential. In contrast, acid, bile acid, pepsin, and trypsin can all be shown to be damaging in various circumstances [4-8].
The pH will be critical in determining the effect of the bile acids. It will determine whether trypsin and pepsin are active or inactive. It will also determine the activity of the bile acids present because unconjugated bile acids are precipitated and innocuous at acidic pH, while conjugated bile acids that remain in solution can be driven into the mucosa and become more damaging in the presence of acid.
In this report we will summarize our findings using a rabbit model of reflux esophagitis to characterize the role of bile and other noxious endogenous substances in causing mucosal injury.
Materials and methods
We utilize a continually perfused in vivo rabbit esophageal model to evaluate the injurious effects of the components of gastric duodenal contents (Figure 1) [9]. We perfused the esophagus for a one hour baseline, then for one hour with the test solution, and then for an additional test hour. We assessed hydrogen ion flux measured with a pH stat autoburette system (Radiometer, Copenhagen, Denmark). We titrated to pH 2 with 0.4 N HCl. The amount of acid added to keep the pH constant equalled the amount of acid disappearing from the system. At the completion of period 3 the test hour, the animal was sacrificed and the esophageal segment was removed and scored for gross pathologic injury by an un-informed observer without knowledge of the specific agent being tested. Gross esophagitis was scored by the following criteria: one plus - normal esophagus, two plus - minimal areas of superficial erosion, three plus - erosive epithelial loss over most of the surface, four plus - extensive erosions with intramural hemorrhage (Figure 2). The histologic pathology was also scored one to four by observers blinded to the protocol.
Figure 1. Experimental set-up.
Figure 2. Photomicrographs of perfused esophageal segments. A. Normal esophagus (+). B, C, D. ++, +++, and ++++ esophageal injury, respectively. (Original magnification X 100.)
In our experiments we used commercial bile acid as opposed to natural bile. Natural bile is a mixture of the four bile acids: deoxycholic, cholic, lithocholic, and chenodeoxycholic. The bile acids are conjugated with taurine or glycine when they are secreted into the bile (Figure 3). Free unconjugated bile acids are found in the intestines as a result of bacterial hydrolysis. Complex measurements are required to quantitate the precise mixture of bile acids and their state of conjugation in natural bile. Hemoglobin breakdown products are present as well. For simplicity and to control the exact components present, commercial bile acids are usually used in experiments. For our experiments with bile we chose to work with commercial taurodeoxycholate. Deoxycholate is the most lipid soluble and in most assays the most damaging bile acid. The taurine conjugate allows the bile acid to be soluble in acidic solution by lowering the pKa from about 6 to about 2. When the pH equals the pKa, the bile acid is half ionized and half protonated. The ionized moiety is soluble. When the pKa exceeds the pH, the acid is predominately protonated and is thereby insoluble. Taurine conjugated bile acids have the lowest pKa and are the most soluble in acid. All bile acids are soluble at pH 7.
Figure 3. Bile acid structure.
Results
Acid esophagitis
When the perfusion of the esophagus was carried out at pH 2 , pepsin caused injury typical of erosive esophagitis in a dose-dependent manner (Figure 4). There was no significant increase in gross or microscopic injury when taurodeoxycholate 5 mM or trypsin 100 unit per cc were perfused along with hydrochloric acid at pH 2. For the same experiments hydrogen ion flux levels were measured before and after exposure in periods one and three (Figure 5). In these experiments pepsin increased esophageal mucosal permeability in a dose-dependent manner. Trypsin had no effect. Remarkably, taurodeoxycholate increased hydrogen ion permeability dramatically even while it had not produced significant evidence of pathologic injury according to our assessment. This shows that even at an acid pH this bile acid had a significant effect on the mucosa.
Alkaline perfusion
The results of the perfusion with pH 7.5 test solution are shown on Figure 6. Here, in contrast to the results with acid perfusion, trypsin increased the pathologic injury score dose-dependently while pepsin was inactive. The hydrogen ion flux assessment showed that trypsin produced hemorhagic erosive esophagitis with only a modest increase in hydrogen ion flux (Figure 7). Again, taurodeoxycholate dramatically increased hydrogen ion flux despite its failure to produce pathologic injury according to our assessment.
Figure 4. The microscopic and gross pathologic scores for the acid perfusion experiments.
Discussion
These results show that the pH of the gastroduodenal contents is a critical determinant of the mechanism of injury in gastroesophageal reflux. In acid reflux, pepsin is the injurious substance with the capability to cause hemorrhagic ulcerative erosive esophagitis acutely.
Acid alone is relatively benign in this setting. We know from more detailed electron microscopic assessment that acid alone is damaging to the mucosa , but the damage is much less that seen with acid pepsin, and the permeability changes are minimal compared to those seen with bile acids [8].
Figure 5. The hydrogen ion back-diffusion results for the acid perfusion experiments.
Figure 6. The microscopic and gross pathologic scores for the alkaline perfusion experiments.
Figure 7. The hydrogen
ion back-diffusion results for the alkaline perfusion experiments.
The role of bile acids, in an acid refluxate, is more complex. In acidic conditions, the bile acid taurodeoxycholate increased acid flux dramatically [4]. The failure of our gross and light microscopic examination to identify injury associated with taurodeoxycholate exposure does not mean that no injury occurred. We know from the permeability assessment that injury did occur, but it was too subtle to be identified by our pathologic examination. In a chronic situation this increased permeability might prove to be important in producing pathologic conditions such as Barrett's esophagus. Unconjugated bile acids would be precipitated in acid conditions, and we have shown they have little effect in acid. But taurine conjugated bile acids with a pKa around 2 are approximately half solubilized and half precipitated at pH2. This explains the effectiveness of taurodeoxycholate at pH2.
When the reluxate is alkaline, trypsin becomes the primary injurious factor in causing acute erosive esophagitis [5]. Again, the bile acid taurodeoxycholate dramatically increased hydrogen ion permeability in this setting. The bile acid, despite its failure to produce erosive esophagitis, could be a determinant of more chronic manifestations of esophageal mucosal injury such as the development of Barrett's esophagus. There was no evidence that the pH 7.5 "alkaline" refluxate itself produced any injury.
The pathologic results we observed occurred within two hours of exposure to damaging agents. Clinicians must be concerned with these immediate signs of injury, but also with the mechanism producing Barrett's esophagus as well as those that produce dysplasia and finally cancer. The experimental protocol that we utilize here cannot assess the role of these reflux components in these more chronic forms of injury. It is possible that the more subtle permeability changes produced by the bile salts, while they do not produce acute hemorhagic esophagitis, may be contributing to the more chronic forms of pathology.
References