METHODS

Male Sprague-Dawley rats weighing 250–350 g were maintained on a standard diet with free access to water. Under pentobarbital sodium anesthesia (5 mg/100 g body wt), the proximal or distal 10 cm of colon were removed and rinsed with 0.9% saline. The serosa was stripped while the tissue was mounted on a glass rod.

Ion flux measurements.

Details of the method were previously described (9, 10). Briefly, tissue pairs were mounted in modified Ussing half-chambers exposing 1.12 cm² surface area. The transepithelial potential difference (PD) was referenced to the mucosal side. Tissues were studied under short-circuit conditions except for 1-s intervals every 100 s, during which bipolar pulses of 0.5 mV yielded electrical current values that were used to calculate tissue conductance (G). Tissues were paired for ion flux studies on the basis of differences inG no greater than 25%. The short-circuit current (I_sc) divided by G yielded the active transport PD.

The fluxes of SCFA were measured by adding 1 μCi¹⁴C-SCFA (10–20 mCi/mM specific activity; NEN, Boston, MA) to the mucosal side of one member of each tissue pair and to the serosal side of the other. Mucosal-to-serosal (J_m→s) and serosal-to-mucosal (J_s→m) fluxes were measured over a 16-min period after an initial 30-min equilibration period. Twelve minutes were allowed for each new steady state, and 32 min were allowed for the effect of ouabain. Net flux was calculated as (J_m→s−J_s→m).

Solutions and acid-base conditions.

The composition of the solutions is shown in Table1. TheN-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES)-Ringer solutions (A andF) were gassed with 100% O₂, and pH was titrated using 2 M H₂SO₄or 1 M NaOH. The ${HCO}_3^{-}$ solutions (B-E) were gassed with 1% CO₂($P{\text{CO}}_2$ = 7 mmHg), 3% CO₂($P{\text{CO}}_2$ = 21 mmHg), 5% CO₂($P{\text{CO}}_2$ = 35 mmHg), 11% CO₂($P{\text{CO}}_2$ = 75 mmHg), or 14% CO₂($P{\text{CO}}_2$ = 95 mmHg) (balance O₂) to obtain various pH values. All solutions were maintained at 37°C. The solutions were so designed that after the addition of the salt of an SCFA or gluconic acid, similar final osmolality and, where appropriate, Na⁺ concentration (always <150 mM) were achieved.

In the ion replacement experiments, choline and isethionate replaced Ringer Na⁺ and Cl⁻, respectively. SCFA concentration was 25 mM bilaterally in all experiments except in the gradient and transport kinetics studies where 1, 10, 25, 50, or 100 mM Na⁺ butyrate or Na⁺ isobutyrate were used on one side and Na⁺ gluconate on the other. In certain experiments, ouabain (1 mM; Sigma Chemical, St. Louis, MO) was added to either the mucosal or serosal bathing solution.

In several experiments in which butyrate flux was measured, the tops of the fluid reservoirs were sealed and vented through an ethanolamine trap (20).¹⁴CO₂was quantitated by liquid scintillation counting to determine the degree to which butyrate was metabolized by the colonic epithelium. We found that ∼7% of the butyrate that entered cells during transmucosal passage was metabolized to CO₂.

pH_i measurements.

The method is described in detail elsewhere (9-11, 13). Briefly, a segment of stripped distal colon (described previously) was mounted as a flat sheet over a 1-cm² hollow ring assembly. It was first incubated in Ringer containing 2 mMdl-dithiothreitol (Sigma), a mucolytic agent, for 10 min. It was then bathed in Ringer containing 9.68 μM 2′,7′-bis(2-carboxyethyl)-5(6)-carboxyfluorescein-acetoxymethyl ester (BCECF-AM; Molecular Probes, Eugene, OR) for 40 min. Intracellular cleavage of BCECF-AM by endogenous esterases produces the poorly permeable pH-sensitive dye BCECF. The mounted tissue was then washed three times in fresh Ringer solution (as designated by the experimental protocol) to remove extracellular dye. Fluorescence microscopy localized the dye primarily to surface epithelial cells (9).

A Perkin-Elmer LS-5B spectrofluorometer (South Plainfield, NJ) equipped with a thermoregulated cuvette holder was used. The mounted tissue was placed in a fixed position at the bottom of a 4-ml cuvette (Markson Science, Phoenix, AZ) with the mucosal surface facing the excitation beam at a 45° angle. In the experiments with a pH gradient a modified tissue holder was used (10). It consisted of a plastic divider that spanned two opposite inside corners of the cuvette. Stripped colonic tissue was placed over a plastic ring designed to snugly fit a 9-mm hole in the divider. The divider was snapped over the ring and tissue. Silicone grease was used to provide a water-tight seal between the divider and the cuvette.

When acid-base conditions were altered, all readings were performed after pH_i reached a plateau, but no less than 12 min after the acid-base condition was changed. During this time the solution in the cuvette was exchanged rapidly with fresh aliquots every 2 min. When ${HCO}_3^{-}$buffers were used the cuvette was tightly closed with a plastic cap to prevent CO₂ leakage between solution exchanges. These measurements represent steady-state values because in preliminary experiments they were not found to change for periods up to 40 min. Ratiometric fluorescence measurements were performed in triplicate using excitation wavelengths of 440 and 500 nm in sequence. The emission wavelength was 530 nm. The ratio was computed by dividing the fluorescence intensity at 500 nm by that at 440 nm. Only tissues that maintained total fluorescence above 2× autofluorescence for the duration of the experimental protocol were used. Autofluorescence was automatically subtracted.

pH_i calibration was done by the high-K, nigericin method (31). The calibration solution contained (in mmol/l) 21 HEPES, 140 KCl, 1 MgCl₂, 1 CaCl₂, 10 glucose, and 10 μg/ml nigericin (Sigma). A logarithmic regression line for the standard curve was used to accommodate the nonlinearity of fluorescence ratios at very low pH values.

Intracellular ${HCO}_3^{-}$ concentration ([${HCO}_3^{-}$]_i) was computed using the Henderson-Hasselbalch equation and the measured pH_i. Intracellular$P{\text{CO}}_2$ was assumed to be equal to the medium $P{\text{CO}}_2$, and the negative log of dissociation (pK ′) and CO₂ solubility were 6.115 and 0.0306, respectively. Bathing solution pH and$P{\text{CO}}_2$ were measured with a Radiometer BMS 3 Mk 2 system with a PHM 73 acid-base analyzer (London Company, Cleveland, OH).Extracellular [${HCO}_3^{-}$] was computed using the Henderson-Hasselbalch equation as described.

All data are expressed as means ± SE and were compared by paired Student’s t-test or analysis of variance (ANOVA). Two-tailed P values < 0.05 were considered significant.

RESULTS

Effect of pH on butyrate flux in distal colon.

Initially the effect of unilateral and bilateral changes in bathing solution pH on butyrate flux in HEPES Ringer were examined. In these experiments butyrate was present at 25 mM on both sides of the tissue. As shown in Fig. 1, at pH 7.38 the net flux of butyrate was −0.1 ± 0.3 μeq ⋅ cm⁻² ⋅ h⁻¹. As mucosal solution pH was decreased in steps from 7.38 to 5.47,J_m→sincreased from 3.4 to 7.2 μeq ⋅ cm⁻² ⋅ h⁻¹and net absorption was observed (3.5 ± 0.3 μeq ⋅ cm⁻² ⋅ h⁻¹). As shown when the luminal pH was then increased to 7.38, the increase in J_m→swas completely reversible. Luminal pH changes had no effect on theJ_s→m of butyrate.

As shown in Fig. 2, as serosal solution pH was decreased in steps from 7.38 to 5.57J_s→m of butyrate increased from 3.4 to 5.4 μeq ⋅ cm⁻² ⋅ h⁻¹. This change caused net butyrate secretion (−2.4 ± 0.5 μeq ⋅ cm⁻² ⋅ h⁻¹). When pH was then increased to 7.38 the increase inJ_s→m was completely reversible. Serosal pH changes had no effect on theJ_m→s of butyrate.

The effect of changing the pH of both bathing solutions on butyrate flux is shown in Fig. 3. As solution pH was decreased in steps from 7.39 to 5.53, bothJ_m→s andJ_s→m of butyrate increased from ∼3 to 7 μeq ⋅ cm⁻² ⋅ h⁻¹. Net butyrate flux was minimal at bilateral pH 7.39 (−0.8 ± 0.4 μeq ⋅ cm⁻² ⋅ h⁻¹) and remained minimal at pH 5.53 (0.1 ± 0.5 μeq ⋅ cm⁻² ⋅ h⁻¹). When pH was then increased to 7.39, the increases in the unidirectional fluxes were completely reversible.

Reductions in pH of one or both bathing solutions had no effect onG but decreasedI_sc. For example, when mucosal solution pH was reduced from 7.38 to 5.47,I_sc decreased from 0.7 ± 0.2 to 0.2 ± 0.3 μeq ⋅ cm⁻² ⋅ h⁻¹,P < 0.05. When pH was then increased to 7.38I_sc increased to 0.9 ± 0.4 μeq ⋅ cm⁻² ⋅ h⁻¹.

Effect of pH on propionate flux in distal colon.

Similar effects of pH changes on propionate fluxes in HEPES Ringer were observed. Propionate was present at 25 mM on both sides of the tissue. Decreases in luminal pH in steps from 7.36 to 5.46 selectively increasedJ_m→s of propionate from 2.7 ± 0.2 to 6.3 ± 0.4 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 4,P < 0.001. As serosal solution pH was decreased in steps from 7.36 to 5.46,J_s→mselectively increased from 4.1 ± 0.2 to 5.4 ± 0.5 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 4,P < 0.05. Minimal net propionate secretion was observed at bilateral pH 7.36 (−1.4 ± 0.3 μeq ⋅ cm⁻² ⋅ h⁻¹,P < 0.02), and zero net transport was found at pH 5.46 (0.9 ± 0.4 μeq ⋅ cm⁻² ⋅ h⁻¹,P < 0.01 compared with flux at pH 7.36).

Effect of butyrate gradient on butyrate flux.

We then examined the effects of unilateral pH changes in HEPES Ringer on butyrate fluxes in the presence of a 25 mM luminal to 0 mM serosal butyrate concentration gradient. As mucosal solution pH was decreased in steps from 7.39 to 5.58,J_m→sincreased from 1.0 ± 0.1 to 2.4 ± 0.2 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 4,P < 0.005. When pH was then increased to 7.39, the increase inJ_m→s was reversed to 1.1 ± 0.1 μeq ⋅ cm⁻² ⋅ h⁻¹,P < 0.005. When serosal pH changes were studied in the presence of a 25 mM serosal to 0 mM luminal butyrate concentration gradient, similar results were obtained. As serosal solution pH was decreased in steps from 7.39 to 5.56,J_s→m of butyrate increased from 2.0 ± 0.1 to 4.1 ± 0.1 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 4,P < 0.007. When pH was then increased to 7.39, the increase inJ_s→m was reversed to 2.3 ± 0.1 μeq ⋅ cm⁻² ⋅ h⁻¹,P < 0.007.

Effect of pH on butyrate flux in${HCO}_3^{-}$ Ringer.

The effect of bilateral changes in bathing solution pH on butyrate flux across distal colon also was examined in${HCO}_3^{-}$ Ringer where pH changes were induced by changing $P{\text{CO}}_2$. In these experiments butyrate was present at 25 mM on both sides of the tissue. In 1 mM ${HCO}_3^{-}$ Ringer at$P{\text{CO}}_2$ = 7 mmHg, pH 6.79, the net flux of butyrate was 0.1 ± 0.2 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 6. As solution pH was decreased in steps to 6.08 by increasing $P{\text{CO}}_2$ to 95 mmHg, net flux was little changed: −0.5 ± 0.2 μeq ⋅ cm⁻² ⋅ h⁻¹.J_m→s increased from 4.2 ± 0.2 to 4.8 ± 0.3 μeq ⋅ cm⁻² ⋅ h⁻¹and J_s→mincreased from 4.1 ± 0.3 to 5.3 ± 0.5 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 6,P < 0.05.

In 5 mM ${HCO}_3^{-}$ Ringer at$P{\text{CO}}_2$ = 7 mmHg, pH 7.24, minimal net butyrate secretion was observed (−1.0 ± 0.2 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 5). As solution pH was decreased to 6.44 in steps by increasing $P{\text{CO}}_2$to 95 mmHg, net flux was little changed: −0.4 ± 0.5 μeq ⋅ cm⁻² ⋅ h⁻¹.J_m→s increased from 2.6 ± 0.1 to 4.0 ± 0.3 μeq ⋅ cm⁻² ⋅ h⁻¹and J_s→mincreased from 3.6 ± 0.2 to 4.4 ± 0.2 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 5,P < 0.02. These flux changes in 1 and 5 mM ${HCO}_3^{-}$ Ringer were completely reversible and were similar in magnitude to flux changes in HEPES Ringer (∼1 μeq ⋅ cm⁻² ⋅ h⁻¹per pH unit). In addition, in both 1 and 5 mM${HCO}_3^{-}$ Ringer, reductions in pH did not affect G but reducedI_sc from 0.4 μeq ⋅ cm⁻² ⋅ h⁻¹to near zero.

Effect of butyrate metabolism on butyrate flux.

We then examined whether the metabolism of SCFA influenced the pattern of their transepithelial transport. We studied the effect of pH in HEPES Ringer on the flux across distal colon of isobutyrate, a weakly metabolized SCFA (22). Isobutyrate was present at 25 mM on both sides of the tissue. Decreases in luminal pH in steps from 7.35 to 5.45 increasedJ_m→s of isobutyrate from 2.1 ± 0.1 to 4.9 ± 0.1 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 4,P < 0.0001.J_s→m flux increased slightly from 2.2 ± 0.1 to 2.8 ± 0.1 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 4,P < 0.04. In a separate experiment, when pH was reduced in both bathing solutions in steps, net isobutyrate flux remained unchanged: −1.0 μeq ⋅ cm⁻² ⋅ h⁻¹at pH 7.34, 0.0 μeq ⋅ cm⁻² ⋅ h⁻¹at pH 6.68, −0.8 μeq ⋅ cm⁻² ⋅ h⁻¹at pH 6.04, and −0.5 μeq ⋅ cm⁻² ⋅ h⁻¹at pH 5.55. These changes in unidirectional fluxes were completely reversible.

Effect of luminal Na⁺ removal and ouabain.

To determine whether apical Na⁺/H⁺exchange activity was necessary for SCFA transport, the exchanger was inhibited by substituting choline for Na⁺ in the mucosal bathing solution. In HEPES Ringer, with butyrate present at 25 mM on both sides of the tissue, bilateral reductions in pH in steps stimulatedJ_m→s andJ_s→m of butyrate equivalently. At pH 7.41 and 5.68, net flux was unchanged and near zero, andJ_m→s was 1.3 ± 0.1 and 5.1 ± 0.1 μeq ⋅ cm⁻² ⋅ h⁻¹and J_s→mwas 2.1 ± 0.2 and 4.1 ± 0.1 μeq ⋅ cm⁻² ⋅ h⁻¹, respectively, n = 2,P < 0.05. In 5 mM${HCO}_3^{-}$ Ringer, similar results were obtained. At pH 7.32 and 6.51, net fluxes were unchanged and near zero, and J_m→swas 1.5 ± 0.1 and 2.1 ± 0.1 μeq ⋅ cm⁻² ⋅ h⁻¹and J_s→mwas 2.7 ± 0.1 and 3.4 ± 0.3 μeq ⋅ cm⁻² ⋅ h⁻¹, respectively, n = 3,P < 0.05. The effects of pH in both HEPES and ${HCO}_3^{-}$ Ringer were completely reversible.

The effect of luminal ouabain was tested to determine whether an apical membrane H⁺-K⁺-adenosinetriphosphatase (ATPase) participated in the action of luminal pH on SCFA absorption. The experiments were carried out in 5 mM${HCO}_3^{-}$ Ringer with butyrate at 25 mM on both sides of the tissue. Ouabain (1 mM) did not affectJ_m→s at pH 7.30 (2.7 ± 0.3 vs. 2.6 ± 0.3 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 6). Luminal ouabain also did not affect the stimulatory action of a luminal pH reduction to 6.36 on J_m→s(3.2 ± 0.5 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 6,P < 0.02).

The effect of serosal ouabain was tested to determine if any active transport process was involved in the SCFA response to pH. Fluxes were not measured for 32 min and/or until theI_sc was reduced to near zero. The experiments were carried out in 5 mM${HCO}_3^{-}$ Ringer with butyrate at 25 mM on both sides of the tissue. As shown in Fig.4, the addition of 1 mM ouabain to the serosal solution did not alter butyrate fluxes. When pH was reduced on both sides of the tissue from 7.29 to 6.45, increases in bothJ_m→s (3.0 ± 0.2 vs. 4.4 ± 0.2 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 4,P < 0.01) andJ_s→m (3.3 ± 0.2 vs. 4.0 ± 0.1 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 4,P < 0.05) were noted.

Effect of luminal Cl⁻ removal on butyrate flux.

To evaluate the role of apical membrane Cl⁻/SCFA⁻exchange, we examined the effect of substituting isethionate for Cl⁻ in the mucosal bathing solution. In 5 mM ${HCO}_3^{-}$ Ringer, the absence of luminal Cl⁻ at pH 7.21 did not significantly increaseJ_m→s of butyrate [2.6 ± 0.1 vs. 3.2 ± 0.3 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 5, not significant (NS)] but did decreaseJ_s→m 22% (3.6 ± 0.2 vs. 2.4 ± 0.2 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 5,P < 0.05). At pH 6.50, the removal of luminal Cl⁻ did not increaseJ_m→s (4.0 ± 0.3 vs. 4.7 ± 0.5 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 5, NS) and again decreasedJ_s→m∼32% (4.4 ± 0.2 vs. 3.0 ± 0.3 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 5,P < 0.005). The absence of luminal Cl⁻ also altered the effect of pH on butyrate flux. A reduction in pH from 7.21 to 6.50 stimulated net flux from 0.8 ± 0.2 to 1.7 ± 0.4 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 5,P < 0.01, as a consequence of a greater increase inJ_m→s (3.2 ± 0.3 vs. 4.7 ± 0.5 μeq ⋅ cm⁻² ⋅ h⁻¹,P < 0.01) than inJ_s→m (2.4 ± 0.2 vs. 3.0 ± 0.3 μeq ⋅ cm⁻² ⋅ h⁻¹,P < 0.02). Cl⁻ removal had similar effects on fluxes in HEPES Ringer, and the effects of pH in both${HCO}_3^{-}$ and HEPES Ringer were completely reversible.

Effect of SCFA concentration.

Carrier-mediated transport processes, unlike nonionic diffusion, exhibit saturation as evidenced by flattening of the curve describing the relationship between substrate concentration and flux. We examined for saturation by measuringJ_m→s of butyrate in HEPES or 5 mM ${HCO}_3^{-}$ Ringer at pH 6.40 or 7.40 in the presence of a mucosal-to-serosal butyrate gradient. As shown in Fig. 5, when the mucosal butyrate concentration was progressively increased from 1 to 100 mM, regardless of the Ringer or pH,J_m→sincreased in a linear fashion. Furthermore, the weakly metabolized SCFA isobutyrate exhibited similar transport behavior as its concentration was increased in ${HCO}_3^{-}$ Ringer at pH 6.40.

Effect of pH_i and [${HCO}_3^{-}$]_i.

To determine whether the relation of SCFA flux to solution pH changes could be accounted for by changes in intracellular acid-base conditions, we measured pH_i and calculated [${HCO}_3^{-}$]_iin distal colon during the various experimental conditions. As shown in Table 2, in HEPES Ringer containing 25 mM butyrate and gassed with 100% O₂, CO₂ tension and therefore [${HCO}_3^{-}$]_iwere zero. pH_i mirrored pH_e whether the pH change was unilateral or bilateral. However, when the pH_e decrease was unilateral, the decrease in pH_i was less than when the pH_e change was bilateral. Furthermore, mucosal changes in pH_e seemed to have a somewhat greater effect on pH_i than serosal changes.

In 5 mM ${HCO}_3^{-}$ Ringer containing 25 mM butyrate, qualitatively similar effects of bilateral changes in pH were observed. Increasing bathing solution$P{\text{CO}}_2$ from 7 mmHg (pH_e 7.43 ± 0.02) to 95 mmHg (pH_e 6.48 ± 0.01) decreased pH_i from 7.39 ± 0.02 to 6.46 ± 0.03, n = 6,P < 0.001, and increased [${HCO}_3^{-}$]_ifrom 4.8 ± 0.5 to 5.8 ± 0.4 mM,n = 6,P < 0.05. In 5 mM${HCO}_3^{-}$ Ringer containing 25 mM isobutyrate, similar effects of $P{\text{CO}}_2$on pH_i and [${HCO}_3^{-}$]_iwere observed. When bilateral pH_ewas decreased from 7.44 to 6.48, pH_i decreased from 7.20 ± 0.02 to 6.53 ± 0.02 and [${HCO}_3^{-}$]_iincreased from 3.0 ± 0.1 to 7.1 ± 0.3 mM,n = 5,P < 0.001. In both HEPES and${HCO}_3^{-}$ Ringer, the presence of unilateral or bilateral SCFA did not affect the steady-state value of pH_i.

Effect of pH on butyrate flux in proximal colon.

We then examined whether the pattern of pH effects on SCFA transport was similar in the proximal colon. Butyrate flux was measured when present at 25 mM on both sides of the tissue in HEPES Ringer. Bilateral decreases in pH from 7.39 to 5.69 in steps increasedJ_m→s from 2.9 ± 0.2 to 4.5 ± 0.3 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 4,P < 0.01, andJ_s→m from 2.2 ± 0.4 to 3.3 ± 0.4 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 4,P < 0.002. Net butyrate fluxes were minimal at bilateral pH 7.39 (0.8 ± 0.5 μeq ⋅ cm⁻² ⋅ h⁻¹) and pH 5.69 (1.2 ± 0.5 μeq ⋅ cm⁻² ⋅ h⁻¹,n = 4, NS). Reductions in pH also had no effect on G but decreasedI_sc. All of these changes were qualitatively and quantitatively similar to those observed in distal colon.

DISCUSSION

The importance of SCFA in colonic energy metabolism (5, 26), ion transport (1, 2, 4, 12, 17, 26, 28, 32), pH_i regulation (7, 14, 15), and systemic acid-base balance (5) has been recognized for some time. SCFA absorption precedes and is required for these functions. For many years the mechanism of SCFA absorption by the colon was believed to be by nonionic diffusion. The basis for this were the findings that chain length, luminal pH, and the concentration gradient from lumen to blood (or tissue or serosa) affected the absorption rate (5, 29, and see Ref.30 for a review of these considerations). However, such characteristics are compatible with SCFA absorption through apical and basolateral membrane SCFA⁻/${HCO}_3^{-}$exchange processes. Indeed, these process have recently been identified in colonic membrane vesicles (20, 24, 27).

Our experiments were designed to determine the relative importance of nonionic diffusion and anion exchange as mechanisms by which SCFA cross the colonic mucosa. We considered theK_m of the apical anion exchanger (27 mM for butyrate in rat colon) (24), the pKa of the SCFA under consideration (4.8–4.9), metabolism of SCFA by the colonic mucosa, the cell-to-lumen flux of SCFA by apical Cl⁻/SCFA⁻exchange (25), and the possibility that various SCFA or segments of the colon had different transport characteristics (30). Thus experiments were performed at unilateral or bilateral butyrate concentrations of 25 mM; with butyrate, propionate, and weakly metabolized isobutyrate (22); over a pH range of 7.4 to 5.3; and in the presence and absence of luminal Na⁺ or Cl⁻, or mucosal or serosal ouabain. Our findings strongly suggest that a passive transport pathway, presumably nonionic diffusion, accounts for SCFA transport across rat colon.

The most compelling evidence for nonionic diffusion includes the findings that unilateral reductions in pH_e had selective and quantitatively equivalent effects onJ_m→s andJ_s→m, that the effects of pH_e were similar in HEPES and ${HCO}_3^{-}$ Ringer, and that transepithelial transport was unsaturable at SCFA concentrations up to 100 mM. None of these findings would be expected or accounted for by a carrier-mediated epithelial transport process and by apical membrane SCFA⁻/${HCO}_3^{-}$exchange in particular. It is otherwise difficult to explain how SCFA could move at equivalent rates in both directions across colonic tissue if the transport process were not passive, how SCFA could traverse the tissue via a SCFA⁻/${HCO}_3^{-}$exchange process in the apparent absence of intracellular bicarbonate (in HEPES Ringer), and why transport saturation would not be observed at substrate concentrations almost four times greater than theK_m (observed in brush-border membrane vesicles).

The measurements of pH_i also shed light on the mechanism of SCFA transport. In HEPES Ringer, the effects of pH_e on unidirectional fluxes were equivalent whether the changes in pH_e were unilateral or bilateral. Moreover, the absolute values for the unidirectional fluxes were equivalent at similar values for unilateral pH_e and bilateral pH_e. However, in HEPES Ringer (Table 2), the effects of unilateral and bilateral changes in pH_e on pH_i were not equivalent. This suggests that the effect of pH_e on SCFA flux was primarily localized to the mucosal or serosal compartment in which it occurred rather than through the effects of pH_e on pH_i. As discussed previously, such an effect would be more compatible with the transport process of nonionic diffusion than anion exchange.

We also found that luminal removal of Na⁺ to inhibit apical Na⁺/H⁺exchange did not alter SCFA transport in rat colon or the effects of pH_e. Furthermore, neither luminal ouabain, which may inhibit colonic apical membrane H⁺-K⁺-ATPase (8, 16, 19, 23), nor serosal ouabain, which inhibits all active transport processes, affected SCFA transport or the effects of pH_e. Epinephrine stimulation of apical Na⁺/H⁺exchange has been shown to stimulate propionate absorption in rabbit proximal but not distal colon (29, 30). A pH gradient (presumably luminal microclimate pH_e < pH_i) was suggested as the mechanism of these effects (29). Because we could not confirm a role for apical Na⁺/H⁺exchange in SCFA absorption, we believe that the reported requirement for this exchanger may be species specific. The presence of a microclimate pH, of course, is consistent with both SCFA absorption by nonionic diffusion and anion exchange.

Metabolism of SCFA by the colonic mucosa certainly occurs, and in preliminary experiments we found that ∼7% of the butyrate that entered cells was metabolized to CO₂. In the rabbit proximal colon in vitro, from 4 to 7% of absorbed propionate was metabolized to CO₂ under similar experimental conditions (29). We do not believe that SCFA metabolism affected our flux measurements or their interpretation. First, in our studies the absolute fluxes and the effects of pH and substrate concentration were similar for butyrate, propionate, and weakly metabolized isobutyrate. Second, the presence of glucose in our studies reduced the fraction of colonic energy derived from SCFA (5). Third, our unidirectional flux calculations were based on the appearance of radiolabeled butyrate in the unlabeled “cold” bathing solution rather than on the disappearance of butyrate from the labeled “hot” side. Thus it is irrelevant to the calculation of unidirectional flux that ∼7% more butyrate entered cells than exited into the opposite bathing solution.

In addition to transport across the cell, SCFA may be recycled across the apical membrane. Recycling presumably occurs via the Cl⁻/SCFA⁻exchange process described by Rajendran and Binder (25). In our studies of rat colon, the fraction recycled was estimated by comparing theJ_m→s of butyrate in the presence and absence of luminal Cl⁻. We found thatJ_m→s was not significantly affected by the absence of this anion.J_s→m, however, was decreased 22–32%, depending on pH_e. Apparently, under the experimental conditions described (5 mM${HCO}_3^{-}$ Ringer at pH_e between 6.50 and 7.21) apical membrane Cl⁻/SCFA⁻exchange has a greater role in the transcellular secretory flux of SCFA than in apical membrane recycling of absorbed SCFA.

Our findings do not rule out a contribution of anion exchange at the apical and basolateral membranes to SCFA absorption in situ. The effects on SCFA transport of intact tissue layers, blood flow, cell membrane potential, competing substrates, and varying energy stores and demands are unknown. Moreover, the complexity of the in situ environment suggests that the relative importance of active and passive transport of SCFA may not be fixed. Nevertheless, the experimental conditions examined here do mirror in situ conditions where the [${HCO}_3^{-}$]_iis very low and a lumen-to-blood pH gradient and SCFA concentration gradient exist. Indeed, such conditions favor net absorption of SCFA by nonionic diffusion. Our studies suggest that at least in the rat colon nonionic diffusion is the most important if not the only mechanism of SCFA absorption.