Gastric ulcers involving the NG mucosa of horses are recognized worldwide, with prevalence estimated from 40% to 90%, depending on populations surveyed and type of athletic activity.1–11 The high prevalence of gastric ulcers in horses has been attributed to the anatomy of the equine stomach, diet, and environmental stress.12 The exact cause of gastric ulcers in the NG region of horses has not been determined but may be the result of exposure to acids (hydrochloric, volatile fatty, and bile acids) and the lack of adequate barrier defenses, including a thick mucus layer and bicarbonate.13–18
Performance horses are typically fed grain containing high amounts of soluble carbohydrates, and these carbohydrates may be fermented by resident stomach bacteria to end products such as VFAs, which are found in gastric contents of horses fed hay and grain diets.13,19 Furthermore, HCl and VFAs have been shown to cause disruption in bioelectric properties and barrier function of the NG mucosa of horses in an in vitro Ussing chamber system.14–16
In a previous study13 conducted by our laboratory, we constructed a gastric ulcer model in horses fed bromegrass hay and an alfalfa hay-grain diet. In that study, horses fed a diet high in fermentable carbohydrates that had gastric ulcers had higher stomach propionic, butyric, and valeric acid concentrations and a lower stomach pH than horses without gastric ulcers. Additionally, the high-calcium and -protein diet (alfalfa hay-grain) reduced the severity of gastric ulcers, leading us to conclude that calcium, protein, or both may have provided a protective effect on the NG mucosa. Also, in more recent studies,15,16 we have shown that 60mM acetic, propionic, butyric, and valeric acids at a pH of ≤ 4 decreased Isc and PD, which are indicators of damage to the cellular Na+-K+ ATPase pump and barrier disruption, respectively. Furthermore, cellular swelling was evident on histologic examination. Results of these 2 studies confirm that high VFA concentrations within an acid environment cause injury to the stomach NG mucosa.
Although 60mM VFAs have been found in gastric contents of horses, concentrations are typically much lower (1.1 to 20.3mM) in horses fed hay and grain feeds containing highly fermentable carbohydrates.13 Thus, we hypothesize that lower concentrations of VFAs (< 60mM) within a low-pH (≤ 4.0) environment cause acid injury and cellular damage to the NG mucosa of the equine stomach and that this damage occurs in a dose-dependent manner. Furthermore, in a previous study,13 we found that a diet high in calcium and protein exerted a protective effect upon the NG mucosa. The concentration of calcium in that diet was 17.7 mg/g, so we hypothesized that calcium carbonate (20 mg/mL) would protect and reverse the damage occurring in the NG mucosa from a result of HCl and VFA exposure. We chose to evaluate the protective effects of calcium rather than protein because calcium is the main component in antacid preparations and commonly added to equine diets.
The purpose of the study reported here was to determine effects of pH and VFAs (ie, acetic, butyric, propionic acids, and valeric acids) at various concentrations (5, 10, 20, and 40mM) on sodium transport (Isc) and barrier function (PD and tissue R) of the NG stratified squamous mucosa of the equine stomach and to determine whether the addition of calcium carbonate (20 mg/mL) was effective in reversing that injury. Our objective was to elucidate the role of HCl and VFAs in the pathogenesis of gastric ulcers in horses and the role of calcium carbonate in reversing that injury.
Materials and Methods
Animals—Gastric tissues obtained from 48 adult horses (27 geldings, 1 stallion, 19 mares, and 1 hermaphrodite) that were euthanatized for reasons of debilitation or donation. Information regarding sex, breed, and age of each horse was obtained.
Tissue acquisition—Horses were euthanatized by lethal injection of an overdose of barbiturate.a The stomach of each horse was removed within 1 hour of euthanasia and dissected along the greater curvature to expose the NG mucosa. The stomach, including NG mucosa, was cleaned of ingesta by use of deionized water and an NRS (142mM sodium, 124mM chloride, 25mM bicarbonate, 10mM glucose, 5mM potassium, 1.65mM HPO4−2, 1.25mM calcium, 1.1mM magnesium, and 0.3mM H2 PO4− [pH, 7.4]) prepared in our laboratory. The mucosa was examined for gastric ulcers, which were graded by use of a scoring system for horses.20 The NG mucosa was obtained by sharply dissecting the submucosa from the underlying muscular tissue. Specimens were cut into disks (approx 3.5 cm2 in diameter). A specimen of NG mucosa from each horse was immediately placed in neutral-buffered 10% formalin solution and underwent histologic examination.
Study design—Specimens of NG gastric mucosa were manipulated in Ussing chambers, as previously described.15
The NG mucosa was pinned, mucosal side down, to a paraffin tray and bathed in oxygenated NRS maintained at 20° to 22°C (ie, room temperature.). Each Ussing chamber had an aperture area of 7.07 cm2. Tissues were bathed in NRS on the mucosal and submucosal surfaces. The NRS was oxygenated by circulating the solutions in water-jacket reservoirs with a gas lift and by use of 95% O2 and 5% CO2, at 37°C.
Each experiment lasted 330 minutes. Tissues were allowed to equilibrate in the NRS tissue baths for 30 minutes. The mucosal surface of tissues then were bathed for 240 minutes with 10 mL of NRS (control solution) or 10 mL of NRS to which a VFA had been added and that was adjusted to attain a specific pH (1.5, 4.0, or 7.0); the submucosal surface of tissues was bathed with 10 mL of NRS at a pH of 7.0. Luminal reservoirs were then drained and refilled with NRS or NRS containing the respective VFAs, to which was added calcium carbonate (20 mg/mL), and the pH was adjusted to 7.0. Values were recorded for an additional 60 minutes. This sequence was used for all the experiments in which tissues were exposed to VFAs and then allowed to recover with added calcium carbonate.
When acetic, propionic, butyric, or valeric acids were added to the bath solution, the sodium salt of each VFA was substituted for an equivalent amount of sodium chloride in the NRS so that the final concentration was 5, 10, 20, and 40mM, resulting in ARS, PRS, BRS, and VRS. The pH (1.5, 4.0, or 7.0) of the solution used to bathe the mucosal surface was adjusted by titration with 1.0N HCl while measuring with a pH electrode.b These VFA concentrations and pH values were selected because they are representative of the environment found within the equine stomach after feeding a typical hay and grain diet13 and because VFAs are fully dissociated (ionized) and not lipophilic at a pH of 7.0, partially dissociated and partially lipophilic at a pH of 4.0, and undissociated (nonionic) and highly lipophilic at a pH of 1.5. For each tissue bathed in NRS (control solution) and VFA solution, chambers were set up with pH of 1.5, 4.0, or 7.0 on the mucosal side for a total of 3 chambers of each VFA and 3 chambers of NRS in each experiment.
Because dissimilar solutions were used in the mucosal and submucosal baths, liquid junction potentials (caused by net diffusion of sodium ions from solution to bridge to electrode, resulting in PD changes) were measured to determine their magnitude and orientation. External circuits were created by use of 2 experimental solutions as previously described.21 Junction potentials for all experimental bathing solutions were < 1.0 mV; therefore, no corrections were made.
Spontaneous PD was measured by use of Ringer-agar bridges, the composition of which was identical to that of the submucosal solution used to bathe the tissues. The PD was nullified by use of an automatic voltage clamp through Ag-AgCl2 electrodes. Tissue PD and Isc were recorded every 15 minutes throughout the 330-minute duration of each experiment. The pH of each bath solution was measured hourly and adjusted to maintain stability. Tissue R and G were calculated from the open-circuit PD and from the current necessary to nullify the PD (ie, Isc) by use of Ohm's law (ie, voltage = current × resistance) as reported elsewhere.15
Immediately after the stomach was incised, a specimen of NG mucosa was collected from the margo plicatus and placed in neutral-buffered 10% formalin. This tissue was used to determine whether the NG mucosa was histologically normal prior to VFA or control treatment. After completion of the incubation period in the Ussing chambers, mucosa was cut into 2 equal pieces; 1 piece was placed in neutral-buffered 10% formalin, and the other piece was placed in NRS and frozen at −70°C for subsequent analysis of VFA concentrations. Fixed tissues were prepared for routine histologic examination and stained with H&E as described elsewhere.15 Slides were examined by use of light microscopy to determine cellular swelling and necrosis.
Analysis of VFA concentrations in tissues—To determine uptake of VFAs into the NG mucosa, specimens of frozen tissue were thawed and analyzed by use of gas chromatography in accordance with the following procedure. Tissue was removed from the NRS, weighed, minced, and placed in 5 mL of deionized water. Specimens were then homogenized for 5 minutes by use of a homogenizer.c Specimens were transferred into 15-mL polypropylene centrifuge tubes and centrifuged at 2,000 × g for 20 minutes. An aliquot (2 mL) of supernatant was placed into a 5-mL polypropylene test tube. Another aliquot (1 mL) of supernatant was placed in a 1.5-mL microcentrifuge tube and cooled on ice. A volume (0.2 mL) of an internal standard solution (25% phosphoric acid) was added to each aliquot, and tubes were then vortexed at 13,000 rpmd for 20 minutes. The supernatant was removed, placed into a gas chromatography vial (approx 1 mL), and clamped with an aluminum vial closure. Vials were frozen at −70°C for subsequent analysis. Tissue concentrations of each VFA (ie, acetic, propionic, butyric, and valeric acids) were measured in the homogenized tissue with a gas chromatographe with a method described in a previous study.15 Tissue VFA concentrations were measured to determine whether uptake occurred into tissues, which would add credence to the theory of cell acidification and decreased sodium transport as the cause of the histopathologic changes.
Statistical analysis—Data were analyzed by use of a computer software program.f The model used for tissue type and variable was a randomized block design repeated-measures ANOVA with each horse as a block. Treatment factors of solution condition (VFA), concentration, and pH along with all interactions were in the main plot, and time and interactions of time with main plot factors were in the subplot. Time was the repeated-measure factor. Least squares means and SEM were calculated, and mean separation by use of the Fisher protected least significant difference test was performed. Values of P < 0.05 were considered significant.
Results
Animals and gastric tissues—Horses (n = 48) ranged from 3 to 32 years of age (mean, 13.9 years), and 25 of 45 horses (56%) had gastric ulcers in the NG mucosa. The presence or absence of gastric ulcers was not recorded for 3 horses. Reasons for euthanasia included chronic laminitis and lameness (n = 12), weight loss (8), neurologic disease (5), blindness (4), recurrent airway obstruction (3), urinary incontinence (2), congestive heart failure (1), fistulous withers (1), chronic diarrhea (1), peritonitis (1), suspensory desmitis (1), ringbone (1), and hermaphrodite (1). None of the horses were treated at our facility prior to being euthanatized, so treatments administered prior to arrival were not recorded. Medical conditions and prior treatment may have altered the responses seen in the tissues of these horses because we did not control for these factors. However, the study was designed (use of control specimens in NRS at each pH [1.5, 4.0, and 7.0] along with each VFA treatment) to account for variations between and within horses.
Ulcer score was determined for the 25 horses with gastric ulcers. Ulcers were scored from 1 (small focal or multifocal ulcers) to 3 (extensive coalescing ulcers). Mean (range) ulcer score was 1.4 (1 to 3). Of the 25 horses, 16 horses (64%) had an ulcer score of 1, 6 horses (24%) had an ulcer score of 2, and 3 horses (12%) had an ulcer score of 3. Sixteen tissue specimens were collected from grossly normal NG mucosa at or adjacent to the margo plicatus in the stomach of each horse, which is the region of the equine stomach where most ulcers develop.9 Of the 768 tissue specimens collected from this region (48 horses; 16 tissue specimens/horse), 48 (1 from each horse) were immediately placed in neutral-buffered 10% formalin for subsequent histologic evaluation. None of these tissue specimens had evidence of gross or histopathologic changes. Of the remaining 720 specimens, 690 were analyzed in Ussing chambers. Thirty tissues from 2 horses were not analyzed as a result of malfunction of the Ussing chambers. The variables measured did not differ significantly between tissues from horses with ulcers and tissues from horses without ulcers.
HCl exposure and recovered mucosa—Mean Isc in tissues perfused with NRS at a pH of 1.5 decreased significantly by 195 minutes after onset of exposure, compared with values at the same time points for tissues exposed to NRS at a pH of 7.0 and 4.0 (Figure 1). Mean PD across the tissue decreased significantly by 30 minutes after onset of exposure, compared with mean values at the same time points for tissues exposed to NRS at a pH of 4.0 and 7.0. Mean Isc was initially lower in the tissues exposed to NRS at a pH of 4.0 but did not differ significantly, compared with mean Isc value for tissues exposed at a pH of 7.0 and 1.5. Mean Isc in tissues perfused with NRS at a pH of 7.0, 4.0, and 1.5 decreased 36%, 45%, and 58% from their initial values, respectively, during the 270-minute exposure. Mean PD across the tissues decreased by 57% in tissues exposed to NRS at a pH of 1.5, compared with 9% from their initial values in tissues exposed to NRS at a pH of 7.0 and 4.0. Mean R also decreased significantly by 150 minutes after onset of exposure, compared with the same time points in tissues exposed to pH 7.0 and 4.0. Mean R significantly decreased by 12% from initial values in the tissues exposed to NRS at a pH of 1.5 during the 240-minute incubation period.
After adding calcium carbonate (recovery) with an adjusted pH of 7.0, mean Isc immediately increased to near control values in tissues exposed to NRS at a pH of 1.5 (Figure 1). Although mean PD across the tissues remained significantly decreased in tissues exposed to NRS at a pH of 1.5, mean tissue R immediately increased to near control values after addition of calcium carbonate. Analysis of these data suggests that sodium transport in the NG tissue and barrier function recovered with addition of calcium carbonate. Thus, the effect of HCl on sodium transport and barrier function may be reversible; however, prolonged exposure may lead to irreversible tissue damage and ulceration.
Acetic acid exposure and recovered mucosa— Mean Isc in tissues perfused with all concentrations of ARS at a pH of 4.0 and 7.0 did not significantly change from control tissues. However, mean Isc significantly decreased within 30 minutes after the pH change in tissues perfused with 5, 10, 20, and 40mM of ARS at a pH of 1.5, compared with tissues exposed to NRS at the same time points (Figure 2). Mean Isc remained significantly decreased for 60 minutes in tissues exposed to 5, 10, and 20mM of ARS and then stabilized and was not significantly different from controls during the remainder of the 240-minute exposure period. In contrast, mean Isc in tissues exposed to 40mM ARS remained significantly decreased for the 240-minute exposure period, compared with NRS-bathed tissues at the same time points. Mean Isc decreased 116.5% from initial values in tissues exposed to 40mM ARS at a pH of 1.5, whereas tissues exposed to ARS (5, 10, and 20mM) at a pH of 1.5 decreased 55% from initial values by the end of the 240-minute exposure period, which did not differ from controls for the same period. Mean PD across the tissues also significantly decreased 30 minutes after the pH was decreased to 1.5 in tissues exposed to 40mM ARS and remained significantly decreased throughout the 240-minute experimental period. Mean R and G in tissues perfused with the ARS did not significantly change over the experimental period. However, during the recovery phase of the experiment, when calcium carbonate was added, Isc and PD returned to near control values in tissues perfused with ARS at all concentrations, compared with tissues perfused with NRS at the same time points. Concentrations of ARS > 20mM lowered Isc and PD significantly, compared with lower concentrations, and in a time-dependent manner. The pH threshold for affecting the Isc was < 4.0, a value equating to 30mM of undissociated acetic acid.
Propionic, butyric, and valeric acid exposure and recovered mucosa—Mean Isc and PD in tissues perfused with PRS, BRS, and VRS at a pH of 7.0 did not differ significantly from control tissues; however, mean Isc and PD in tissues perfused with these VFAs at a pH of 4.0 and 1.5 had a concentration-dependent decrease, compared with control tissues at the same pH (Figures 3 and 4). Mean Isc and PD in tissues perfused with 20 and 40mM BRS and VRS at a pH of 4.0 had a significant decrease within 30 minutes after the pH was decreased in mucosal solutions and remained significantly decreased throughout the 240-minute exposure period, compared with tissues exposed to NRS and the lower concentrations (5 and 10mM) at the same pH. Mean R increased in all tissues perfused with the VFAs at all concentrations at a pH of 4.0, but this increase was not significant. However, mean R increased significantly in tissues exposed to 40mM BRS and mean R increased and then decreased in tissues exposed to 40mM VRS at a pH of 1.5 over the 240-minute exposure period (Figures 5 and 6). Mean tissue R decreased to a value that was 40% of baseline values over the 240-minute exposure period. Mean tissue G decreased slightly over the 240-minute exposure time in tissues exposed to the VFAs at a pH of 1.5, but was not significantly different than that of tissues exposed to NRS at the same pH.
When CaCO3 was added during the recover period, Isc and PD increased to near pretreatment values in the PRS and BRS exposed tissues at a pH of ≤ 4.0, but remained significantly decreased in tissues exposed to 20 and 40mM VRS at a pH of ≤ 4.0 (Figures 3 and 4). Mean R increased slightly in tissues exposed to the lower concentrations of VFAs at a pH of 1.5 and 4.0; however, mean R increased in tissues exposed to BRS and VRS at the higher concentrations (20 and 40mM) at a pH of 1.5 and 4.0, compared with tissues exposed to NRS at the same pH (Figures 5 and 6).
Mucosal concentrations of VFAs—Concentrations of VFAs were determined for NG mucosa exposed to each of the bath solutions. Mean tissue acetic acid concentration was highest in all tissues regardless of the VFA exposure. Mean tissue acetic acid concentration in control tissues was 0.41 ± 0.53 mmol/g of tissue and consisted of 91.6% of the total tissue VFAs measured. Acetic acid concentration was significantly increased in tissues exposed to 20mM (0.54 ± 0.11 mmol/g of tissue) and 40mM (1.29 ± 0.84 mmol/g of tissue) ARS at a pH of 1.5. Mucosal tissues exposed to BRS, PRS, or VRS contained higher concentrations of the respective VFAs in tissues, but these values were low, compared with tissue concentrations of acetic acid.
HCl and VFA mucosal histopathologic changes— Histologic examinations of specimens of NG mucosa exposed to NRS and the various VFAs at a pH of 7.0 were generally normal (Figure 7). On the other hand, mucosa exposed to NRS at a pH of 1.5 had cellular swelling and a mottled appearance in the superficial stratum corneum and stratum transitionale. Mucosa specimens exposed to ARS had the most dramatic cell swelling in tissues exposed to the higher concentrations (Figure 8). Furthermore, cell swelling was seen in the stratum transitionale and stratum spinosum in tissues exposed to PRS and BRS at a pH of 1.5 or 4.0. Cellular swelling in the stratum spinosum cells was most dramatic in specimens exposed to BRS at a pH of 4.0 and 1.5 and VRS at all pH values and was less apparent in tissues exposed to ARS and PRS at a pH of 4.0 and 1.5 (Figure 9).
Discussion
Prevalence of gastric ulcers in horses of our study was 56% (25/45 horses). Mean ulcer score for horses with ulcers was 1.4, with 64% of horses having an ulcer score of 1, 24% of horses having an ulcer score of 2, and 12% of horses having an ulcer score of 3. These findings were similar to those of 2 previous reports15,16 in which older horses were donated because of underlying medical disease and musculoskeletal lameness. In those reports, prevalence of gastric ulcers was 53% and 46%, respectively, and mean gastric ulcer score was 1.9 and 1.8, respectively. Because most of horses donated to our study were affected by medical conditions, we believe the higher prevalence of gastric ulcers in older horses could be a result of stress caused by systemic illness.
Our results indicate that concentration-dependent alterations in tissue bioelectric properties occur in the NG mucosa of the equine stomach when it is exposed to acetic acid at a pH of < 4.0, propionic and butyric acid at a pH of ≤ 4.0, and valeric acid at a pH of ≤ 7.0. This functional change in the tissues was more than that induced by HCl alone. Functional mucosal damage was manifested as a decrease in sodium transport (ie, Isc), with an increase in tissue R and decrease in tissue G. These changes were associated with histologic evidence of cellular swelling in the NG (squamous) mucosa.
Bioelectric measurements of the NG mucosa in horses are based on results of previous studies22–28 on tissues of other organs and species. Values for Isc and PD in our study were similar to those previously reported for equine NG mucosa exposed to HCl and VFAs.15,16 However, tissues in those studies were exposed to a higher concentration (60mM) of VFAs. The Isc is a direct indicator of active sodium ion transport and, therefore, a measure of epithelial function and tissue viability. Experiments with frog skin,24,25 rumen epithelium,26 and rabbit esophagus27 revealed that Isc is equal to the net sodium transport across tissues. In rumen epithelium and rabbit esophageal epithelium, it was found that the viable layers immediately beneath the stratum corneum are involved in sodium transport because of the high density of intra-cellular sodium pumps (Na+-K+ ATPase) in these tissues.26,27 Also, results of studies by Goldstein et al29 and Snow et al30 indicate that rabbit esophageal cells of the stratum spinosum have pH-sensitive basolateral potassium channels. Acidification of rabbit esophageal tissues resulted in inhibition of the potassium channels and abolished Isc, which prevented the regulatory volume decrease usually seen after hypotonic medium–induced cell swelling. A more recent study by Widenhouse et al14 revealed that basolateral potassium channels are present in equine NG mucosal cells and that these channels are important for active sodium absorption across the apical surface of the epithelium, maintenance of cell volume, and control of intracellular pH. Inhibition of these channels by acidification (pH, 1.7) abolished Isc and PD and decreased R, which made cells more susceptible to HCl damage. Thus, cellular Na+-K+ ATPase and basolateral potassium channels in the equine NG mucosa may play a major role in maintaining cellular fluid volume, pH, and barrier function. Furthermore, Na+-K+ ATPase have been localized in the stratum spinosum layer by histochemical methods in equine NG mucosa.31
In our study, Isc in tissues perfused with HCl in NRS at a pH of 1.5 (and to a lesser extent at a pH of 4.0) significantly decreased after 195 minutes of exposure, whereas PD and R significantly decreased in the same tissues before changes in Isc (30 and 150 minutes, respectively). The significant decrease in PD and R, followed by the decrease in Isc, was similar to previously reported findings for equine NG mucosa.15,16 These data suggest that hydrogen ions initially cause an increase in outer barrier permeability (decrease in tissue PD and R), allowing diffusion into the deeper sodium-transporting cell layers (ie, stratum transitionale and stratum spinosum), which are then acidified, causing a subsequent decrease in sodium transport (decrease in Isc) and eventually cell swelling. This was evident by histologic examination of tissues exposed to HCl alone at a pH of 1.5.
In our study, tissues required 150 minutes of HCl (pH 1.5) exposure before tissue R was significantly decreased, whereas porcine gastroesophageal mucosa required only 75 minutes of HCl (pH 1.5) exposure before tissue R was significantly decreased.22 Equine NG mucosa may therefore be more resistant to acid damage than gastroesophageal tissues collected from pigs. Further evidence of this difference in tissue R between porcine and equine NG mucosa was seen when tissues were exposed to acetic acid at a pH of 1.5. In a study by Argenzio and Eisemann,22 porcine gastroesophageal mucosa exposed to 20mM ARS at a pH of 2.5 had a significant decease in Isc (tissue sodium transport) after 15 minutes of exposure, compared with tissues exposed to 5mM ARS, whereas equine NG mucosa exposed to 20mM ARS at a pH of 1.5 in our study did not have a significant decrease in sodium transport after 240 minutes of exposure, compared with tissues exposed with 5mM ARS. Furthermore, in the same study of pigs, gastroesophageal tissue had mucosal vesicle formation and separation when exposed to 20mM acetic acid for 75 minutes, whereas mucosal separation was not observed in our study or in previous in vitro studies of equine NG mucosa.15,16 The difference in tissue R to acid damage may explain why horses have milder clinical signs,32 compared with the clinical signs described in pigs.33
The effect of hydrogen ions on sodium transport and barrier function appears to be reversible in pigs.22 This was evident by Isc and tissue R returning to near control values during the 1-hour recovery period. The reversible nature of HCl damage has been reported previously when equine NG mucosa was exposed to pH of 1.5 and then changed to pH of 7.0.15,16 However, the effects of adding calcium carbonate to mucosal bath solutions have not been examined previously. In our study, tissues recovered from a pH of 1.5 when calcium carbonate was added to the mucosal solutions, which is similar to the recovery of tissues observed when pH alone is increased.15,16 The addition of calcium carbonate increased pH in the mucosal bathing solutions in our study. Addition of calcium carbonate would be expected to increase stomach pH to > 4.0 when fed to horses, as it is the primary ingredient of antacid tablets.34 Thus, the change in pH to > 4.0 in the stomach may reverse the severity of acid damage, whereas prolonged exposure (> 240 minutes) of the NG mucosa to acid conditions (pH ≤ 4) may lead to mucosal injury and gastric ulcers. Because horses deprived of feed for 24 hours have a low stomach pH, this finding may explain why horses deprived of feed for prolonged periods develop gastric ulcers.35
In contrast to the mild decrease in tissue R (19%) and Isc (58%) detected in tissues exposed to HCl, tissues exposed to various concentrations of VFAs in the presence of HCl had an immediate concentration-dependent decrease in Isc (ie, sodium transport) and PD during the 240-minute exposure period. Along with this alteration in Isc and PD, a milder increase in R and decrease in G were found during the same period. This decrease in sodium transport appears to be dependent on VFA concentration, pH of the bathing solution, and chain length of the VFA. Exposure of tissues to acetic acid (2 carbons) had the least effect on the mucosal tissue, whereas propionic (3 carbons), butyric (4 carbons), and valeric (5 carbons) acids completely abolished Isc (150% to 180% from initial values) at higher concentrations (20 and 40mM). The Isc in ARS-exposed tissue at a pH of 4.0 and 7.0 had no significant changes, compared with tissues exposed to NRS at the same time points. However, Isc significantly decreased immediately after the pH was lowered to 1.5, in tissues exposed to all concentrations of ARS, but Isc recovered by 60 minutes in the tissues exposed to lower concentrations (5, 10, and 20mM) of acetic acid, compared with control tissues at the same pH. In contrast, Isc of tissues exposed to 40mM ARS at a pH of 1.5 was abolished (116% of initial values) 30 minutes after exposure and remained significantly decreased for the remainder of the 240-minute exposure period. Therefore, the pH and concentration thresholds for long-term decreases of Isc (sodium transport) in NG mucosa exposed to acetic acid were 4.0 and > 20mM, respectively. Although acetic acid had the least effect on NG gastric mucosa in our study, compared with other VFAs, gastric contents in horses fed typical alfalfa hay and highly fermentable grain diets have been reported to contain 16 to 20mM acetic acid.12 In the aforementioned study,12 horses with a mean body weight of 411 kg were fed a diet consisting of alfalfa hay (5.5 kg) and grain (2.6 kg; 0.6 kg/100 kg BW). Acetic acid concentrations at or close to the 20mM threshold value were present in an environment with a pH of < 4.0 for 5 hours after feeding. A 5-hour exposure to acetic acid at a pH of < 4.0 could potentially damage sodium transport in the NG mucosa in horses and lead to the formation of gastric ulcers. Gastric ulcers were detected in 3 of 5 horses in that study.12 On the basis of these data, it appears that feeding grain at amounts of < 0.6 kg/100 kg BW results in intragastric acetic acid concentrations below the 20mM threshold. This is consistent with a recommendation of grain feeding of 0.5 kg/100 kg BW made in a previous report.36 We therefore conclude that horses fed grain at an amount of > 0.5 kg/100 kg BW every 6 to 8 hours are at higher risk for the development of NG mucosal ulcers. Furthermore, in our study and others,13,15,16 tissue acetic acid concentrations were highest in tissues exposed to gastric contents (in vivo) and to ARS (in vitro), compared with other VFAs. Thus, acetic acid may permeate NG tissue more readily than other VFAs, resulting in changes in bioelectric properties. However, further studies are required to determine the in vivo effects of acetic acid and the exact quantity of grain needed in horses to meet nutritional needs, while minimizing the effect of acetic acid on NG mucosal bio-electrical properties.
Exposure of tissues to propionic, butyric, and valeric acids had a more dramatic effect on bioelectric properties of the equine NG mucosa than those of acetic acid and HCl alone. The Isc significantly decreased in tissues exposed to propionic, butyric, and valeric acids in a dose-dependent manner and remained low throughout the 240-minute exposure period. Mean Isc decreased (reductions of > 100% of baseline values) in tissues exposed to 5 to 40mM propionic or butyric acids at a pH of 1.5. Furthermore, mean Isc in valeric acid–exposed tissues decreased by 73% at a pH of 7.0 and was completely abolished (> 100% of baseline values) at a pH of 1.5 and 4.0. Also, tissue R increased in a dose-dependent manner in tissues exposed to propionic and butyric acid and decreased in tissues exposed to valeric acid. These data are consistent with those of previous reports15,16 examining equine NG mucosa exposed to the same VFAs. These data suggest that propionic, butyric, and valeric acids have significant and immediate effects on sodium transport in these tissue as evident by the decrease in Isc within 30 minutes of exposure, in contrast with exposure to HCl alone, which resulted in a much later (195 minutes) decrease in Isc. This suggests that VFAs rapidly permeate the equine NG mucosal cells because of their lipophilic properties at a low pH (ie, pH ≤ 4.0). These data also suggest that VFAs can penetrate the outer barrier of the stomach mucosa and affect the sodium transport tissues underneath without altering the permeability of the outer barrier (ie, no change in R observed). Thus, the proposed sequence of events that lead to tissue injury associated with VFAs in an acidic environment can be hypothesized. Undissociated (nonionic) VFAs (at a pH of ≤ 4.0) penetrate the outer barrier layers and enter cells below and adjacent to the stratum corneum, potentially as a result of cotransport with sodium. The intracellular pH is higher than the pH of the extracellular fluid, which results in dissociation of these weak acids and acidification of the intracellular environment. This disrupts sodium transport (as measured by Isc) and the regulation of cell volume. Uptake of water and sodium across the apical membrane cannot be alleviated by sodium pumping or potassium leakage through the basolateral membrane, which results in cellular swelling and necrosis. Cellular swelling in the stratum spinosum results in an initial increase in tissue R and, eventually, a decrease in R and sloughing of the outer barrier. Initial damage may also occur as a result of hydrogen ions crossing tissues because an initial decrease in PD is observed, despite the increase in tissue R. This initial loss of barrier function may enable sodium and water to enter underlying layers and hasten the injury to tissues.
Absorption of VFAs was evident in exposed tissues. Control tissues contained acetic acid in highest concentration, which constituted 91% of the total VFAs in these tissues. Also, tissues exposed to ARS had the highest concentration of acetic acid at a pH of 1.5. Tissues contained VFAs other than acetic acid, but the concentrations were much lower, compared with acetic acid concentration. The fact that tissue concentrations of acetic acid significantly increased in tissues exposed to the higher concentrations of acetic acid may indicate that tissue absorption of acetic acid, especially when acetic acid concentration in gastric contents exceed threshold concentration, may lead to acid injury. Also, acetic acid may permeate tissues more readily than other VFAs, thus leading to changes in bioelectric properties. However, tissue concentrations of other VFAs may have been higher than that measured in our study because in a previous study,37 propionic acid was transported into, metabolized, and released from the serosal side of the mucosal cells at a rate that was half the mucosal uptake. Thus, acetic acid concentrations as well as other VFA concentrations may have been higher in tissues, which may account for the histopathologic changes observed.
Although propionic, butyric, and valeric acids had a more dramatic effect on sodium transport and barrier function than acetic acid in our study, gastric contents collected from horses generally contain < 5mM concentrations of these VFAs, which are below the lowest concentration (5mM) evaluated here.13,15 However, it should be mentioned that valeric acid caused acid injury at all pH values and had the most inhibitory effect on NG mucosa sodium transport. Valeric acid was the most important predictor of gastric ulcer severity in research horses fed various diets.13 Also, gastric contents contain multiple VFAs (acetic, propionic, butyric, and valeric acids) and lactic acid, which was not evaluated here, but may act synergistically to cause damage in vivo. To our knowledge, no studies exist examining the effects of various VFA combinations on the bioelectric properties of the equine NG mucosa.
Addition of calcium carbonate to mucosal bath solutions resulted in a return to baseline Isc and recovery in all tissues exposed to acetic acid at a pH of 1.5. The recovery of these tissues was attributed to a change in pH or stimulation of Na+-K+ ATPase by calcium carbonate. Calcium carbonate has been shown to stimulate Na+-K+ ATPase via calmodulin in other species.38 This recovery in Isc is greater than that observed in tissues when only the pH was increased.15,16 However, in the aforementioned studies,15,16 tissues were exposed to a higher concentration of acetic acid (60mM) and did not recover within an hour after the pH was adjusted to 7.0. On the basis of these findings, it appears that the addition of calcium carbonate to gastric contents in the form of calcium supplements or calcium-containing feed (alfalfa hay) might hasten the recovery of cellular sodium transport systems in NG mucosa of horses. Although 20 mg/mL was used in our in vitro study, further studies are needed to determine the amount of calcium or calcium carbonate required to initiate recovery of sodium transport systems in vivo and whether its use will help to prevent gastric ulcers from developing in horses. Furthermore, studies in people have shown that ingestion of calcium carbonate following a meal results in transient buffering of gastric acid followed by hypersecretion of acid or a rebound effect secondary to stimulation of gastrin or by other mechanisms.39 However, that rebound effect was not seen until 6 hours after feeding of a calcium-rich (alfalfa hay and grain) diet in horses.13
The Isc and PD returned to near control values after addition of calcium carbonate in tissues exposed to propionic and butyric acids at all concentrations, but did not return to control values in tissues exposed to valeric acid at a pH of 4.0 and 1.5. Recovery seen in our study was greater than observed in 2 previous studies, when equine NG mucosa was exposed in vitro to propionic and butyric acids and pH adjusted to 7.0.15,16 However, in those studies, a higher concentration (60mM) of propionic or butyric acid was used. Thus, addition of calcium carbonate may help improve the bioelectric properties and barrier function of NG mucosa exposed to VFAs. However, bioelectric properties of tissues exposed to valeric acid may not recover with addition of calcium carbonate.
In conclusion, our results indicate that HCl and VFAs cause functional damage to the equine NG mucosa in vitro. A threshold level of acetic acid was determined, which is relevant to the development of gastric ulcers because higher concentrations of this VFA have been detected in the gastric contents of horses.13,19 Hydrochloric acid and VFAs damage NG mucosal barrier function and sodium transport mechanisms in the living layers immediately deep to the stratum corneum of the NG mucosa, which undermines the superficial mucosa and leads to eventual sloughing and ulcers. Diets high in calcium carbonate may increase the pH of gastric contents and stimulate sodium transport in tissues, which could reverse acid injury caused by VFAs and HCl. Hydrochloric acid and VFAs cause acid injury in a pH- and concentration-dependent manner, and this may provide a reason for why diets high in fermentable carbohydrates have been implicated in the development of gastric ulcers in horses. Further research is needed to determine the in vivo effects of individual VFAs and to ascertain whether VFAs act synergistically to alter mucosal bioelectric properties. Our results also indicate that threshold VFA concentrations may be exceeded if excessive amounts of concentrate are fed to horses; further studies are required to determine the quantity and types of grain that can be safely fed to horses without exerting adverse effects on the stomach mucosa.
ABBREVIATIONS
NG | Nonglandular |
VFA | Volatile fatty acid |
Isc | Short-circuit current |
PD | Potential difference |
Na+-K+ ATPase | Sodium and potassium adenosinetriphosphatase |
R | Resistance |
NRS | Normal Ringer's solution |
ARS | Acetate Ringer's solution |
PRS | Propionic Ringer's solution |
BRS | Butyrate Ringer's solution |
VRS | Valerate Ringer's solution |
G | Conductance |
BW | Of body weight |
Beuthanasia, Schering-Plough, Kenilworth, NJ.
Accumet A-10 portable pH meter, Fisher Scientific, Pittsburgh, Pa.
Polytron, Brinkmann Instruments, Westbury, NY.
Vortex-Genie, Scientific Industries Inc, Bohemia, NY.
Hewlett Packard model 5890 gas chromatograph, Hewlett Packard, Avondale, Pa.
SAS, version 8.02, SAS Institute Inc, Cary, NC.
References
- 1.
Hammand CJ, Mason DK, Watkins KL. Gastric ulceration in mature Thoroughbred horses. Equine Vet J 1986;18:284–287.
- 2.
Vatistas NJ, Snyder JR & Carlson G, et al. Epidemiological study of gastric ulceration in the Thoroughbred racehorse: 202 horses (1992–1993), in Proceedings. Annu Meet Am Assoc Equine Pract 1994;4:125–126.
- 3.
Murray MJ, Schusser GF & Pipers FS, et al. Factors associated with gastric lesions in thoroughbred racehorses. Equine Vet J 1996;28:368–374.
- 4.
Murray MJ, Gordinsky C & Anderson CW, et al. Gastric ulcers in horses: a comparison of endoscopic findings in horses with and without clinical signs. Equine Vet J Suppl 1989;7:68–72.
- 5.
McClure SR, Glickman LT, Glickman NW. Prevalence of gastric ulcers in show horses. J Am Vet Med Assoc 1999;215:1130–1133.
- 6.
Bertone JJ. Prevalence of gastric ulcers in elite, heavy use western performance horses, in Proceedings. 46th Annu Conv Am Assoc Equine Pract 2000;46:256–259.
- 7.
Johnson JH, Vastistas N & Castro L, et al. Field survey of the prevalence of gastric ulcers in thoroughbred racehorses and on response to treatment of affected hores with omeprazole paste. Equine Vet Educ 2001;13:274–279.
- 8.
Mitchell RD. Prevalence of gastric ulcers in hunter/jumper and dressage horses evaluated for poor performance, in Proceedings. 21st Annu Meet Assoc Equine Sports Med 2001;74–77.
- 9.↑
Begg LM, O'Sullivan CB. The prevalence and distribution of gastric ulceration in 345 racehorses. Aust Vet J 2003;81:199–201.
- 10.
Dionne RM, Vrins AA, Doucet MY. Gastric ulcers in standardbred racehorses: prevalence, lesion description, and risk factors. J Vet Intern Med 2003;17:218–222.
- 11.
Nieto JE, Snyder JR & Beldomenico P, et al. Prevalence of gastric ulcers in endurance horses—a preliminary report. Vet J 2004;167:33–37.
- 12.↑
Andrews FM, Buchanan BR & Elliott SB, et al. Gastric ulcers in horses. J Anim Sci 2005;83 (suppl E):E18–E21.
- 13.↑
Nadeau JA, Andrews FM & Mathew AG, et al. Evaluation of diet as a cause of gastric ulcers in horses. Am J Vet Res 2000;61:784–790.
- 14.↑
Widenhouse TV, Lester GD, Merritt AM. Effect of hydrochloric acid, pepsin, or taurocholate on bioelectric properties of gastric squamous mucosa in horses. Am J Vet Res 2002;63:744–749.
- 15.↑
Nadeau JA, Andrews FM & Patton CS, et al. Effects of hydrochloric, acetic, butyric, and propionic acids on pathogenesis of ulcers in the nonglandular portion of the stomach of horses. Am J Vet Res 2003;64:404–412.
- 16.
Nadeau JA, Andrews FM & Patton CS, et al. Effects of hydrochloric, valeric, and other volatile fatty acids on pathogenesis of ulcers in the nonglandular portion of the stomach of horses. Am J Vet Res 2003;64:413–417.
- 17.
Berschneider HM, Blikslager AT, Roberts MC. Role of duodenal reflux in nonglandular gastric ulcer disease of the mature horse. Equine Vet J Suppl 1999;29:24–29.
- 18.
Bullimore SR, Corfield AP & Hicks SJ, et al. Surface mucus in the non-glandular region of the equine stomach. Res Vet Sci 2001;70:149–155.
- 19.
Stillions MC, Teeter SM, Nelson WE. Equine digestive volatile fatty acid concentration, in Proceedings. 2nd Equine Colic Res Symp 1970;21–22.
- 20.↑
Andrews FM, Sifferman RL & Bernard W, et al. Efficacy of omeprazole paste in the treatment and prevention of gastric ulcers in horses. Equine Vet J Suppl 1999;29:81–86.
- 21.↑
Orlando RC, Bryson JC, Powell DW. Mechanisms of H+ injury in rabbit esophageal epithelium. Am J Physiol 1984;246:G718–G724.
- 22.↑
Argenzio RA, Eisemann J. Mechanisms of acid injury in porcine gastroesophageal mucosa. Am J Vet Res 1996;57:564–573.
- 23.
Argenzio RA, Meuten DJ. Short-chain fatty acids induce reversible injury of porcine colon. Dig Dis Sci 1991;36:1459–1468.
- 24.
Ussing HH. Overview: the development of the concept of active transport. In: Skou JC, Norby JG, Maunsbach AB, et al, eds. The Na+-K+ pump part B: cellular aspects: proceedings of the 5th International Conference on Na+, K+ ATPase. Fuglso, Denmark. New York: Alan R Liss Inc, 1988;3–16.
- 25.
Koefoed Johnsen V, Ussing HH. The nature of the frog skin potential. Acta Physiol Scand 1958;42:298–310.
- 26.↑
Chien WJ, Stevens CE. Coupled active transport of Na and Cl across forestomach epithelium. Am J Physiol 1972;223:997–1003.
- 27.↑
Powell DW, Morris SM, Boyd DD. Water and electrolyte transport by rabbit esophagus. Am J Physiol 1975;229:438–443.
- 28.
Korner B, Moller M & Mertens H, et al. Pars proventricularis of pig stomach—electro and transport physiological studies, in Proceedings. 6th Int Symp Dig Physiol Pigs 1994;1:12–18.
- 29.↑
Goldstein JL, Folgelson BG & Snow JC, et al. Rabbit esophageal cells possess K+ channels: effect of hyposmotic stress on channel activity. Gastroenterology 1993;104:417–426.
- 30.↑
Snow JC, Goldstein JL & Schmidt LN, et al. Rabbit esophageal cells show regulatory volume decrease: ionic basis and effect of pH. Gastroenterology 1993;105:102–110.
- 31.↑
Schnorr B, Wille KH. Histochemical and ultrastructural study of the acid phosphatase, alkaline phosphatase, and adenosine triphosphatase in the pars proventricularis epithelium of the horse stomach [in German]. Z Zellforsch Mikrosk Anat 1971;114:482–492.
- 32.↑
Andrews FM, Nadeau JA. Clinical syndromes of gastric ulceration in foals and mature horses. Equine Vet J Suppl 1999;29:30–33.
- 33.↑
Argenzio RA. Comparative pathophyiology of nonglandular ulcer disease: a review of experimental studies. Equine Vet J Suppl 1999;29:19–23.
- 34.↑
Gilman AG. Antacids. In: Gilman AG, Rall TW, Nies AS, et al, eds. Goodman & Gilman's the pharmacological basis of therapeutics. 8th ed. New York: Pergamon Press, 1990;904–909.
- 35.↑
Murray MJ. Equine model of inducing ulceration in alimentary squamous epithelial mucosa. Dig Dis Sci 1994;39:2530–2535.
- 36.↑
Coenen M. The occurrence of feed-induced stomach ulcers in horses. Schweiz Arch Tierheilkd 1990;132:121–126.
- 37.↑
Breves G, Shroder B, Stein J. Transepithelial transport of short-chain fatty acids and their metabolism in pig hindgut. Dtsch Tierarztl Wochenschr 1995;102:159–160.
- 38.↑
Grinstein S, Wieczorek H. Cation antiports of animal plasma membranes. J Exp Biol 1994;196:307–318.