Osmotic flow of water frog gastric mucosa
in isolated
RICHARD P. DURBIN Cardiovascular Research Institute UCSF, San Francisco, California
and Department 94143
DURBIN, RICHARD P. Osmotic flow of water in isolated frog gastric mucosa. Am. J. Physiol. 236(l): E63-E69, 1979 or Am. J. Physiol.: Endocrinol. Metab. Gastrointest. Physiol. 5(l): E63-E69, 1979. -A gravimetric procedure was used to measure net volume flow across bullfrog gastric mucosa mounted between cKambers. A portion of the net volume flow towards the lumen was coupled to acid production. With an isotonic solution instilled on the luminal surface, the secreted acidity (ratio of increase in acid output to increase in volume flow) was hypertonic, in agreement with previous reports in mammalian stomach. Dilution of the secretory solution to 10% of normal nearly abolished the net volume flow coupled to acid production so that the mean secreted acidity rose to 1.87 M. Other experiments in which gastric juice was collected from this preparation showed that secretion into an initially empty lumen was only slightly hypertonic, as in mammalian stomach. The results indicate that instillation of secretory solution dilutes the endogenous osmotic gradient due to secreted HCl. This gradient is probably just outside the apical surface of the oxyntic cells of stomach.
gastric
acidity;
osmotic
permeability;
instillate
effect
collected from the mammalian stomach that is initially empty has nearly the same osmolarity as plasma. Several groups have reported, however, that prior instillation of a solution isotonic with plasma leads to an increase in secreted acidity (13, 18, 22, 23). Such results suggest that the instillate has interfered in some manner with an osmotic mechanism that determines the flow of water in gastric secretion. Relatively little is known about the secretion of water by the isolated amphibian mucosa in spite of the ease with which the environment of this preparation can be controlled. The present study shows that net volume flow in this preparation is proportional to acid secretion. The secreted acidity is affected by the luminal contents to the extent that instillation of a hypotonic (11 mM NaCl) secretory solution virtually abolished water flow associated with HCl production. GASTRIC JUICE
METHODS
A direct gravimetric method was used to measure net volume flow. The plastic chambers used for this purpose, modified from an earlier design (7), are shown in cross-section in Fig. 1. A bullfrog of moderate size (about 5 in.) was doubly pithed, the stomach removed, 0363-6100/79/0000-0000$01.25
Copyright
0
1979 the American
Physiological
of Physiology,
and the mucosa freed from outer smooth muscle by blunt dissection. The gastric epithelium was mounted so that its serosal surface faced the volume chamber, which contained nutrient solution: the luminal surface was bathed with secretory solution, in the larger chamber. This orientation was chosen because quantitative recovery of solution was more reliable from the serosal surface (7). It was not practical to oxygenate the solution in the volume chamber during an experimental period; hence bicarbonate-CO2 buffer was avoided. The nutrient solu30 Ntion contained (in mM): 68 NaCl, tris(hydroxymethyl)methyl-2-aminoethane sulfonic acid (TES), 4 KCl, 1.8 CaCl,, 0.8 MgS04, and 10 Na phydroxybutyrate. The solution was titrated to pH 6.9 with NaOH, and saturated with oxygen just before using. The secretory solution was 110 mM NaCl, bubbled continuously with 100% oxygen. In some experiments, hypotonic secretory solution (11 mM NaCl) was used instead. To begin an experimental period, usually 1 h in length, the stopcock (Fig. 1) was opened to drain the volume chamber. The residue in the stopcock was expelled by fitting an empty syringe to the upper filling tube and gently ejecting a few milliliters of air. The volume chamber was then filled from above with a weighed plastic syringe containing nutrient solution, and the syringe reweighed. A few minutes before the end of a period, the secretory solution was titrated to pH 7 with 15 mM NaOH from an autoburet (Radiometer). The more conventional procedure of keeping pH constant (pH stat) was avoided to prevent change in the osmolarity of the secretory contents during the hour. The period was terminated by collecting the nutrient solution in a preweighed vial, refilling the volume chamber, and instilling fresh secretory solution in the large chamber. The net volume flow was calculated from the differences in weight of filled and empty vial and syringe, taking the density of the solution to be unity. Weights were determined to 0.1 mg on an analytical balance (Sartorius). As a test of the procedure, a Parafilm membrane was mounted between the chambers and recovery of volume chamber contents measured. In ten trials in rapid succession, the change in volume was -0.7 t 8.9 ~1 (mean t SD). In another series of nine trials with a Parafilm membrane, recovery of nutrient contents was Society
E63
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (128.059.222.107) on January 14, 2019.
R. P. DURBIN RESULTS
tested after the solution was left in the chamber for 50 min. A loss was noted, presumably due to evaporation, which averaged -12.1 t 6.8 ~1 (mean t SD), or about -15 pi/h. In some experiments gastric juice was collected in the absence of a secretory instillate. The mucosa was mounted so that the luminal surface faced the volume chamber. Both chambers were filled for the first few hours of the experiment to allow recovery from dissection and to establish a steady rate of stimulated acid secretion. The volume chamber was then drained carefully and secretion collected every hour thereafter in a preweighed vial, using the air-filled syringe to empty the chamber. After weighing the vial, the secretion was diluted with exactly 1 ml of distilled water. Two 0.1.ml aliquots were frozen for later determination of Cl (by chloridimeter) and Na and K (by flame photometer), and the remainder was promptly titrated to pH 7. The water content of the tissue was measured at the end of certain experiments. To accomplish this the exposed area of tissue was promptly cut out with fine scissors and placed serosal side down on filter paper. The upper or mucosal surface was lightly scraped with a glass slide to remove the adherent mucus coat and the wet weight determined; dry weight was obtained after desiccating overnight at 9OOC.
A primary goal of this study was to test for the presence (or absence) of a relationship between net volume flow and acid production. In a group of experiments, gastric mucosae was brought to a near-resting (i.e., nonsecreting) state by either prolonged incubation or treatment with 10m3M metiamide. The latter procedure was quicker and more convenient, yielding a small or negligible secretory rate; otherwise, the two procedures gave similar results. Figure 2 shows a representative experiment in which a mucosa was treated with metiamide (lo-" M) in the nutrient solution for 1 h, washed for 3 h, and then maximally stimulated with histamine (10v4M) and theophylline (5 x 10e3 M) in the nutrient solution. Spontaneous acid secretion was promptly depressed by metiamide and net volume flow was also reduced after a lag period. In 3 h, acid secretion appeared to plateau at zero and net volume flow at about -30 pi/h (the minus sign indicating flow from the volume chamber towards the lumen). Stimulation of acid secretion was accompanied by a small but definite increase in volume flow towards the secretory solution. The relatively constant loss in volume of nutrient solution when the mucosa is resting is due to evaporation and other factors: it is examined in more detail in a later section. Because treatment with secretagogues increases both acid secretion (peq/h) and rate of volume loss from the nutrient solution (pi/h), it seems reasonable to define the secreted acidity as the quotient of these changes (peqlpl). The change is calculated between the mean of the two periods just before addition of secretagogues and the mean of the final two periods, omitting the first period of stimulation as a transient. On this basis, the experiment of Fig. 2 yields an acidity of 0.31 eqlliter, well above isotonicity. The peak rate of acid secretion in the experiment of Fig. 2 was 12.3 peq/h for a chamber area of 5.1 cm* or 2.4 peq/cm** h, which is somewhat less than values previously reported for the stimulated preparation (1, 9). The large area used in the present study could include some nonoxyntic glands or the drop in secretory pH during the hour (to about pH 2.5) could lead to backdiffusion or reduction in H+ output. The latter is not rate-limiting because instillation of a hypotonic
TEFLON
FIG. 1. Cross-section of chambers used to measure volume flow, roughly to scale. Approximate volume of right-hand chamber (volume chamber) 3.5 ml; left-hand chamber, 7 ml. Area of central aperture 5.1 cm*.
1-1 Acid
secretion
(yEq /Id
( W-B Al$tiomide 1 1
Volume
1
flow
(yl/h) -I20
-80
FIG. 2. Effect of secretagogues on mucosa brought to resting state with metiamide (10d3 M). Isotonic secretory solution (110 mM NaCl) was instilled in the lumen. Acid secretion, solid line; net volume flow, dashed line; minus sign for latter axis denotes flow out of volume chamber towards the larger chamber (lumen).
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (128.059.222.107) on January 14, 2019.
OSMOTIC
FLOW
OF
WATER
IN
GASTRIC
E65
MUCOSA
t-1 Acid
secretion
Volume (yl/ h)
flow
-120 FIG. 3. Effect of SCNAxes as in Fig. 2.
I --a--l
----.
(lo-*
M) on stimulated
mucosa .
-80
-P I loo
I
I
200
300
1 400
’ 500
1 O
Minutes
secretory solution improved the secretory performance (as shown below). It is more probable that oxygen supply to the tissue is the major factor limiting acid output; the experimental design did not permit stirring or oxygenating the nutrient solution. In other experiments mucosae were first stimulated after recovery from dissection and then inhibited. Either 10 mM NaCl in the nutrient solution was replaced by 10 mM NaSCN, leaving secretagogues in the solution; or 10 -3 M metiamide was included in the nutrient solution and secretagogues removed. Figure 3 shows an experiment with SCN. Whereas acid secretion reached a plateau within an hour of stimulation, net volume flow took longer to reach a steady state. The mean of the last two periods before SCN addition is taken to represent the secreting state, and the mean of the final two periods, the inhibited state. Calculation of the secreted acidity from the changes between these states yielded 0.27 eq/liter. Experiments with metiamide gave similar results, except that the final secretory rate was near zero in spite of the long prior exposure to histamine and theophylline. Each experiment was analyzed in the above manner to yield two ‘sets of data, at high and low secretory rates. The results of 20 experiments with either stimulation followed by inhibition, or vice versa, are combined in Fig. 4. Fitting the points by least squares (5) to the line, y = a + bx, yields net volume flow = -51.5 - 3.50 (acid secretion) Here volume flow is in microliters per hour and acid secretion in microequivalents per hour. The mean acidity from these experiments is the reciprocal of the slope, l/b = 0.286 eq/liter. The standard deviation of the slope, that the slope is zero sb = 0.668; hence the possibility can be rejected (P C 0.001). These results show that net volume flow is correlated with acid secretion, and the acidity of the secretion is well above isotonicity. If the instillate has diminished the endogenous gradient due to HCl, thereby reducing
Volume (yl /h)
flow
v -40
l .
l
. 1
IA
0’
R9sfhg
- stweting
A Secrefhg -mfihg fmetion7i&/ l Secreting - est..ng
I
I
J
4
8
I2
Acid
secretion
fSCN/
, 16
(y Eq /h )
FIG. 4. Cumulative plot of data from 20 mucosae. Filled circles denote resting and stimulated states from experiments like that of Fig. 2, including 6 with metiamide and 4 with prolonged incubation to lower spontaneous rates. Filled triangles denote stimulated and inhibited states for 5 experiments with metiamide, and filled squares, stimulated and inhibited states for 5 experiments with SCN- (as in Fig. 3).
water flow, it should be even more effective when diluted. Accordingly, a series of experiments were performed with 11 mM NaCl (10% of normal, or 0.1 S) as secretory solution. This solution was first instilled well before stimulating the mucosa, as illustrated in Fig. 5. The drop i .n volume flow towards the 1umen in the first of acid secretion few hours reflects both the inhibition by metiamide and the establishment of a passive osmotic flow out of the lumen. Stimulation of acid secretion led to a very large increase in this parameter and a modest change in net volume flow. The difference in means of the final two resting periods and the final two secreting periods gave an acidity of 1.02eqlliter in this experiment. Resting and secreting data were obtained in this manner for nine mucosae, plotted in Fig. 6. For comparison the results for stimulation with normal secretory
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (128.059.222.107) on January 14, 2019.
R. P. DURBIN
E66 S ( -1 171 Acid secretion (pEq/h)
Hyptomii Metio fni& I I
secretory 10.LSI
1 Histomhe + theophyhne I 1
I Volume flow (yllh) -120
1
FIG. 5. Effect of secretagogues on mucosa brought to resting state with metiamide, in the presence of hypotonic secretory solution (11 mM NaCl) on the lumen before and during stimulation. Axes as in Fig. 2.
-80
1
I -w-w
Minutes Volume flow solution from Fig. 4 are repeated here. Fitting both sets OJVhh) of points to the line, y = a + bx, by least squares yields -r20the parameters listed in Table 1. Comparison of the two slopes by standard methods (5) shows that their difference is highly significant (P < 0.001). Instillation of the hypotonic secretory solution has increased the mean Secreted acidity to ilO. = 1.87 eqlliter. Of interest here is the effect of hypotonic secretory solution to increase the maximal rate of acid production. The mean of this parameter (t SE) for the experiments 0 00 of Fig. 6 was 16.3 -+ 0.95 *q/h with 0.1 S, compared to 40I I I I I 11.3 2 0.59 peq/h with 1.0 S. The difference in means is 0 4 8 I2 I6 20 significant (P < 0.001). A clue that might account for Acid secretion (yEq/h) this result is provided by unpublished electron microFIG. 6. Cumulative plots for the resting to secreting transition. graphs of H. F. Helander. He found that the lateral Open circles denote resting and stimulated states from 9 experispace between oxyntic cells appeared dilated in tissues ments with hypotonic secretory solution, 4 like that of Fig. 5 using that had been fixed after exposure to 0.1 S, in comparimetiamide, and 5 using prolonged incubation to lower spontaneous son to control tissues. Osmotic expansion of such spaces rates. Filled circles are data repeated from Fig. 4 for the resting to could provide additional oxygen by diffusion from the secreting transition with isotonic secretory solution (10 expts). well-stirred secretory solution to the lateral surfaces of TABLE 1. Results of stimulation: least-squares the oxyntic cells. =a+bx The large variation in acidity reported here raises the jittoy possibility that the frog is deficient in some manner in Secretory SoluNo. of Expts Intercept (a), &h Slope (b 2 St)), pllpeq the osmotic mechanism that yields isotonic gastric juice tion in mammalian stomach. For this reason experiments 1.0s 10 -41.8 -4.24 + 0.67 were done in which gastric juice was collected from the 0.1s 9 1.8 -0.534 AI 0.65 empty lumen. The need for quantitative recovery required that the mucosa be mounted so that the luminal surface faced the volume chamber. Figure 7 shows a nutrient solution. In this respect frog gastric juice representative experiment in which secretion was col- conforms to mammalian gastric juice. It should be lected for 6 h from a stimulated mucosa. In agreement noted, however, that NaCl is probably absent from pure with a previous report (19), the secretion was a mixture oxyntic cell secretion in the mammal (17). of HCl, NaCl, and KCl. The Cl- concentration over the The intercepts on the y-axis in Figs. 4 and 6 reveal a six periods was 124 t 3 meq/liter (mean t SE), consistsubstantial loss of volume from the nutrient solution in ent with the sum of the three cations, 126 t 4 meq/liter. the absence of acid production. About 15 pi/h is evapoIt appears that these four ions account for the secretion, rative loss, provided this is the same with a mucosa which is thus slightly hypertonic to frog plasma or mounted between chambers as with Parafilm (cf. METH-
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (128.059.222.107) on January 14, 2019.
OSMOTIC
FLOW
OF
WATER
IN
GASTRIC
I50
E67
MUCOSA
Vohme
TABLE 2. Effects of inhibitors on volume flow in unstimulated mucosae
f/o w
Al/h Agent
50
Metiamide Ouabain mM
Dinitrophenol 50
No. of Periods
12 8 8
Values are means 2 SE for net volume was metiamide, 3; ou .abain 9 2; dinitrophenol,
Net Volume Flow
-28.7 -28.8
2 3.5 + 5.1
-24.4
+ 6.6
flow. Number 2.
of mucosae
0 t mM
50 0t IO
mM
-04-O
l/-•
0t
[K’l
[c/-l
-a-
+--
too mM t 0'
I
I 2
I
1 4
I
I 6
Hours 7. Consecutive juice collections from stimulated mucosa. Mucosa was dissected at -2 h and stimulated with histamine and theophylline at -1 h. Volume flow is positive here because mucosa is reversed, with lumen facing volume chamber. The large volume flow in the 1st h represents drainage of residual secretory instillate, in part; the steady state thereafter represents secretion. FIG.
ODS). An appreciable loss must still be accounted for. Sources-to be considered include a) secretion of NaCl to the lumen, a possibility raised by the appearance of NaCl in the collected gastric juice; b) secretion of NaHCO, to the lumen (11); c) a steady swelling of the mucosa or hydration of mucus by incorporation of fluid from the nutrient solution. One approach to the problem was to obtain a control group of mucosae in which acid secretion was inhibited by metiamide to compare with groups in which either 2,4dinitrophenol (DNP) was used as an inhibitor of oxidative phosphorylation or ouabain as an inhibitor of active Na+ transport. Metiamide was used at lo-” M for the 1st h and at 10e4 M for 6 h thereafter. A level of 2 x 10M4 M in the nutrient solution was adopted for DNP and ouabain, sufficient to inhibit acid secretion, electrogenie Cl- transport, and HCO,- secretion (9, 11). With all three inhibitors, acid secretion became negligible and net volume flow reached a plateau after 3 h. The results for the ensuing 4 h are summarized in Table 2. The net volume flow is seen to be the same for all inhibitors, within experimental error. The net volume flow in the presence of metiamide (Table 2) is probably a better estimate of the resting flow than the intercepts of Figs. 4 and 6. Thus only 2915 (evaporative loss) or 14 pi/h remain to be accounted for. Very little of this could be due to active secretion, as shown by the negative results with DNP and ouabain. It appears that the third possibility listed above is the most likely, that part of the resting volume flow is due to hydration of mucus or swelling of the tissue by absorption of water from the nutrient solution. It therefore seemed worthwhile to make a direct test
of tissue swelling. In a series of experiments, water content of mucosae exposed to metiamide for 2 h was compared to that of mucosae similarly treated for 9 h. Acid production was small or negligible. In the 2-h studies, water content was 6.47 t 0.28 ml H,O per g dry wt (mean t SE; n = 7) compared to 7.51 -t 0.37 (n = 7) in the 9-h experiments. The difference is barely significant (P < 0.05). The mean dry weight here was 64 mg; hence the average increase in tissue water in 7 h was 67 mg, i.e., slightly less than 10 pi/h. Thus part of the resting volume flow from the nutrient solution may reflect tissue swelling. Note, however, that the calculation of secreted acidity from the change in volume flow due to stimulation or inhibition is not affected by steady-state losses due to evaporation and tissue swelling. DISCUSSION
The present study shows that a net volume flow is associated with acid production by the isolated frog stomach. The finding is not unexpected because net water flow coupled to active ion transport has been described in a number of epithelia (16). To demonstrate the relationship here, it was necessary to use a large mucosal area and maximal changes in secretory rate, factors neglected in previous studies (8, 25). A significant component of net volume flow is independent of HCl or NaCl secretion, which may account for the negligible effect of Cl- withdrawal on volume flow seen by Villegas and Sananes (25) in unstimulated mucosa. The present chambers permit the use of bullfrogs of moderate size and hence are more convenient than those employed in a preliminary study (7). The apparatus is not suited to the measurement of small or rapidly changing flows, however; tissue swelling or shrinking can affect the volume of nutrient solution recovered, and time must be allowed for such changes to reach a steady state. Acid secretion measured in the presence of an isotonic instillate in the lumen was hypertonic, as found also for low rates of acid production in mammalian stomach. The effect has been shown to arise primarily from a decrease in water flow, rather than additional HCl secretion (22). The effect was observed when isotonic NaCl, sucrose, or mannitol solutions were instilled, but not with isotonic HCl (22). The latter is consistent with the usual finding that gastric juice collected from an otherwise empty lumen is isotonic. The instillate effect is not unique to stomach; it has been described also in gallbladder. This organ normally transfers an isotonic NaCl solution in the direction,
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (128.059.222.107) on January 14, 2019.
R. P. DURBIN
E68
osmotic permeability is increased by stimulating acid lumen-to-blood. The lateral and basal surfaces, bathed secretion (24). For simplicity consider P,,%{of the resting by the serosal solution in vitro, are the site of active Na+ transport (21). Isosmotic replacement of part of the mucosa. The two intercepts on the y-axis of Fig. 6 differ serosal NaCl by sucrose significantly increased the by 44 pllh. Hence tonicity of the absorbate, from 161 to 211 mM (26). Posh1(mucosa) In both stomach and gallbladder, the solution in the = 44/&l ~l/cm*~h per 0.9 x 110 mM of NaCl compartment towards which active transport is directed = 0.09 $/cm* h per 1 mM difference in [NaCl] has been altered. The effect is due to a change in solute: for example, HCI is replaced by NaCl in the case of Net flow has been divided by chamber area and the stomach, and NaCl by sucrose in gallbladder. In either solute (NaCl) assumed to have a reflection coefficient of instance, the replacing substance is larger and diffuses less readily into the region in which osmotic flow is unity. The osmotic permeability thus calculated is close to determined, near the transporting surface. The outcome that of frog skin in vitro. House (15) found for the latter suggests that the replacement has diminished the norPosM of 4 x 10B7 cm/s per atmosphere pressure difference. mal osmotic gradient. Whitlock and Wheeler (26) disIn the present units his figure yields: cussed their results quantitatively on the basis of the double-membrane model of osmotic water flow (2, 6). h Diamond (4) has criticized their treatment, but he also I& (skin) = 0.07 $/cm2. per 1 mM difference in [NaCl] invoked the difference in diffusion rates of NaCl and sucrose to explain their findings. It is of interest to compare these results for intact Instillation of hypotonic secretory solution had an epithelia with the corresponding figure for a synthetic especially striking effect in the present study, practilipid bilayer. Using such a membrane, Hanai and cally abolishing the component of water flow coupled to Haydon (12) found that an osmotic difference of 0.157 M acid secretion. This observation is a key finding because NaCl gave rise to a net flow of lo-” cm/s. In the units it shows clearly that the endogenous osmotic gradient used here in frog stomach is accessible to the external solution. An additional factor determining the amount of os- PosM(bilayer) = 0.23 pl/cm2 h motic water flow is the conformation of the transporting per 1 mM difference in [NaCl] surface. Dainty and House (3) considered the extreme case of active transport of solute into a flat, unstirred This is a value about three times that of intact frog skin layer parallel and adjacent to an epithelium. They point or gastric mucosa. Naively one might have expected the out that the magnitude of the osmotic gradient is contrary result. Thus folding of the gastric epithelium limited by diffusion of solute away from the unstirred presents many square centimeters of apical membrane layer. The resulting osmotic flow is small and the in parallel per square centimeter of chamber area (1, transported fluid hypertonic. 14). An increase in area might have been expected to In reality transporting epithelia are not flat, and the increase PosMin proportion. Either the osmotic permearelevant morphology needs to be examined in each bility of the transporting surface is much less than that instance. Infolding of the epithelium shields the endog- of a lipid bilayer or the full strength of the applied enous osmotic gradient, increasing the velocity of bulk osmotic gradient does not reach the appropriate surface. flow of secreted or absorbed fluid and hampering access The latter explanation seems more likely. The applied of external solute to that surface. gradient is flat at the macroscopic level of the two The apical surface of the oxyntic cell in stomach is solutions bathing the isolated epithelium. Clearly it almost surely the site of active transport, and therefore would be impossible for the gradient thus defined to the locus of the endogenous osmotic gradient. This reach all the transporting surface of frog skin, which is surface changes remarkably on stimulation of secretion, dispersed over a number of cell layers (20). forming numerous leaflets that result in an extensive The isolated gastric mucosa, on the other hand, proliferation of the apical surface (1, 14). If active presents a single cell layer that is highly folded. It transport and osmotic flow are distributed uniformly seems paradoxical that the instillate effects have shown over this surface, a modest osmotic gradient should that the transporting surface is accessible to the lusuffice to drive the observed water flow (10). It is this minal solution; yet, the low osmotic permeability suggradient that is affected by the secretory instillate, as gests that only a fraction of the applied osmotic gradient shown in this study. reaches this surface. A resolution of the problem will The difference in net volume flow observed for iso- require an exact analysis of the distribution of the tonic (1.0 S) and hypotonic (0.1 S) secretory solution can osmotic gradient over the complex Sdimensional strucbe used to estimate an apparent osmotic permeability, ture of the transporting surface. P oskl,for the intact mucosa. At best PosMcan only provide I thank Drs. Venetia France and Margareta Ekblad for helpful a crude estimate, due to lack of stirring of the nutrient criticism. solution and other special conditions in the present This work was supported by National Institutes of Health Grant study. AM-18395 and National Science Foundation Grant PCM 76-04148. The vertical separation of the two lines in Fig. 6 Received 16 May 1978; accepted in final form 28 August 1978. increases with acid secretion. This may indicate that l
l
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (128.059.222.107) on January 14, 2019.
OSMOTIC
FLOW
OF WATER
IN GASTRIC
E69
MUCOSA
REFERENCES 1. CARLISLE, K. S., C. S. CHEW, AND S. J. HER~EY. Ultrastructural changes and cyclic AMP in frog oxyntic cells. J. Cell Biol. 76: 31-42,1978. 2. Cuw, P. F. Na, Cl, and water transport by rat ileum in vitro. J. Gen. Physbl. 43: 1137-1148,196O. 3. DANIT, J., AND C. R. HOUSE. “Unstirred layers” in frog skin. J. Physiol. London 182: 6678,196s. 4. DIAMOND, J. M. Transport mechanisms in the gallbladder. In: Handbook of Physiology. Alimentary Canal. Washington, D.C.: Am. Physiol. Sot., 1968 . v% 6, vol. v, p. 2451-2482. 5. DIXON, W. J., AND F. J. MASSEY, JR. Introduction to Statistical Analysis. New York: McGraw-Hill, 1969, chapt. 11. 6. DURBIN, R. P. Osmotic flow of water across permeable cellulose membranes. J. Gen. Physiol. 44: 31%326,196O. 7. DURBIN, R. P. Coupling of water to solute movement in isolated gastric mucosa. Ciba Found. Symp. 38: 161-177,1976. 8. DURBIN, R. P., H. FRANK, AND A. K. SOLOMON. Water flow through frog gastric mucosa. J. Gen. Physiol. 39: 535-551, 1956. 9. DURBIN, R. P., F. MICHELANGELI, AND A. NICKEL. Active transport and ATP in frog gastric mucosa. B&him. Biophys. Acta 367: 177-189, 10. DURBIN, R.
11.
12. 13. 14.
1974.
P., AND H. F. HELANDER. Distribution of osmotic flow in stomach and gallbladder. B&him. Biophys. Acta. In press. FLEMSTR~M, G. Active alkalinization by amphibian gastric f’undic mucosa in vitro. Am. J. Physiol. 233: El-E12, 1977 or Am. J. PhysioZ.: EndocrinoZ. Metab. Gastrointest. PhySioZ. 2: El-E12,1977. HANAI, T., AND D. A. HAYDEN. The permeability to water of bimolecular lipid membranes. J. Theor. BioZ. 11: 370-382, 1966. HEINZ, E. -r die primare Aziditit der Magensaure. B&him. Biophys. Acta 6: 434-444,195l. HELANDER, H. F., AND R. P. DURBIN. Secretory surface area and phosphatase activity of frog gastric mucosa. Am. J. Physiol. 232:
15. 16.
17.
18.
E48-52, 1977 or Am. J. Physiol.: Endocrinol. Metab. Gastrointest. Physiol. 1: E48-E52, 1977. HOUSE, C. R. The nature of water transport across frog skin. Biophys. J. 4: 401-416,1964. HOUSE, C. R. Water Transport in CeZZs and Tissues. London: Arnold, 1974. HUNT, J. N., AND B. WAN. Electrolytes of mammalian gastric juice. In: Handbook of Physiology. Alimentary CanaZ. Washington, D.C.: Am. Physiol. Sot., 1967, sect. 6, vol. II, part 2, p. 781804. LINDE, S., T. TEORELL, AND K. J. OBRINK. Experiments on the primary acidity of gastric juice. Acta Physiol. Scund. 14: 220-
232,1947. 19. MAKHLOUF,
G. M., AND G. R. DUCKWORTH. Secretion and electrical activity of a unilateral in vitro gastric mucosa. Gastroenterology 65: 907-911, 1973. 20. MILLS, J. W., S. A. ERNST, AND D. R. DIBONA. Localization of Na+-pump sites in frog skin. J. CeZZBioZ. 73: 88-110, 1977. 21. MILLS, J. W., AND D. R. DIBONA. Distribution of Na+ pump sites in the frog gallbladder. Nature 271: 273-275, 1978. 22. MOOI)Y, F. G., AND R. P. DURBIN. Effects of glycine and other instillates on concentration of gastric acid. Am. J. Physiol. 209: 122-126,1965. 23. TEORELL, T.
On the primary acidity of the gastric juice. J, Physiol. London 97: 308-315, 1940. 24. VILLEGAS, L. Action of histamine on the permeability of the frog gastric mucosa to potassium and water. Biochim. Biophys. Actu 75: 377-386,1963. 25. VILLEGAS, L., AND
L. SANANES. Independence between ionic transport and net water flux in frog gastric mucosa. Am. J. Physiol. 214: 997-1000,1968. 26. WHITLOCK, R. T., AND H. 0. WHEELER. Coupled transport of solute and water across rabbit gallbladder epithelium. J. CZin. Invest. 43: 2249-2265, 1964.
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (128.059.222.107) on January 14, 2019.
in isolated
RICHARD P. DURBIN Cardiovascular Research Institute UCSF, San Francisco, California
and Department 94143
DURBIN, RICHARD P. Osmotic flow of water in isolated frog gastric mucosa. Am. J. Physiol. 236(l): E63-E69, 1979 or Am. J. Physiol.: Endocrinol. Metab. Gastrointest. Physiol. 5(l): E63-E69, 1979. -A gravimetric procedure was used to measure net volume flow across bullfrog gastric mucosa mounted between cKambers. A portion of the net volume flow towards the lumen was coupled to acid production. With an isotonic solution instilled on the luminal surface, the secreted acidity (ratio of increase in acid output to increase in volume flow) was hypertonic, in agreement with previous reports in mammalian stomach. Dilution of the secretory solution to 10% of normal nearly abolished the net volume flow coupled to acid production so that the mean secreted acidity rose to 1.87 M. Other experiments in which gastric juice was collected from this preparation showed that secretion into an initially empty lumen was only slightly hypertonic, as in mammalian stomach. The results indicate that instillation of secretory solution dilutes the endogenous osmotic gradient due to secreted HCl. This gradient is probably just outside the apical surface of the oxyntic cells of stomach.
gastric
acidity;
osmotic
permeability;
instillate
effect
collected from the mammalian stomach that is initially empty has nearly the same osmolarity as plasma. Several groups have reported, however, that prior instillation of a solution isotonic with plasma leads to an increase in secreted acidity (13, 18, 22, 23). Such results suggest that the instillate has interfered in some manner with an osmotic mechanism that determines the flow of water in gastric secretion. Relatively little is known about the secretion of water by the isolated amphibian mucosa in spite of the ease with which the environment of this preparation can be controlled. The present study shows that net volume flow in this preparation is proportional to acid secretion. The secreted acidity is affected by the luminal contents to the extent that instillation of a hypotonic (11 mM NaCl) secretory solution virtually abolished water flow associated with HCl production. GASTRIC JUICE
METHODS
A direct gravimetric method was used to measure net volume flow. The plastic chambers used for this purpose, modified from an earlier design (7), are shown in cross-section in Fig. 1. A bullfrog of moderate size (about 5 in.) was doubly pithed, the stomach removed, 0363-6100/79/0000-0000$01.25
Copyright
0
1979 the American
Physiological
of Physiology,
and the mucosa freed from outer smooth muscle by blunt dissection. The gastric epithelium was mounted so that its serosal surface faced the volume chamber, which contained nutrient solution: the luminal surface was bathed with secretory solution, in the larger chamber. This orientation was chosen because quantitative recovery of solution was more reliable from the serosal surface (7). It was not practical to oxygenate the solution in the volume chamber during an experimental period; hence bicarbonate-CO2 buffer was avoided. The nutrient solu30 Ntion contained (in mM): 68 NaCl, tris(hydroxymethyl)methyl-2-aminoethane sulfonic acid (TES), 4 KCl, 1.8 CaCl,, 0.8 MgS04, and 10 Na phydroxybutyrate. The solution was titrated to pH 6.9 with NaOH, and saturated with oxygen just before using. The secretory solution was 110 mM NaCl, bubbled continuously with 100% oxygen. In some experiments, hypotonic secretory solution (11 mM NaCl) was used instead. To begin an experimental period, usually 1 h in length, the stopcock (Fig. 1) was opened to drain the volume chamber. The residue in the stopcock was expelled by fitting an empty syringe to the upper filling tube and gently ejecting a few milliliters of air. The volume chamber was then filled from above with a weighed plastic syringe containing nutrient solution, and the syringe reweighed. A few minutes before the end of a period, the secretory solution was titrated to pH 7 with 15 mM NaOH from an autoburet (Radiometer). The more conventional procedure of keeping pH constant (pH stat) was avoided to prevent change in the osmolarity of the secretory contents during the hour. The period was terminated by collecting the nutrient solution in a preweighed vial, refilling the volume chamber, and instilling fresh secretory solution in the large chamber. The net volume flow was calculated from the differences in weight of filled and empty vial and syringe, taking the density of the solution to be unity. Weights were determined to 0.1 mg on an analytical balance (Sartorius). As a test of the procedure, a Parafilm membrane was mounted between the chambers and recovery of volume chamber contents measured. In ten trials in rapid succession, the change in volume was -0.7 t 8.9 ~1 (mean t SD). In another series of nine trials with a Parafilm membrane, recovery of nutrient contents was Society
E63
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (128.059.222.107) on January 14, 2019.
R. P. DURBIN RESULTS
tested after the solution was left in the chamber for 50 min. A loss was noted, presumably due to evaporation, which averaged -12.1 t 6.8 ~1 (mean t SD), or about -15 pi/h. In some experiments gastric juice was collected in the absence of a secretory instillate. The mucosa was mounted so that the luminal surface faced the volume chamber. Both chambers were filled for the first few hours of the experiment to allow recovery from dissection and to establish a steady rate of stimulated acid secretion. The volume chamber was then drained carefully and secretion collected every hour thereafter in a preweighed vial, using the air-filled syringe to empty the chamber. After weighing the vial, the secretion was diluted with exactly 1 ml of distilled water. Two 0.1.ml aliquots were frozen for later determination of Cl (by chloridimeter) and Na and K (by flame photometer), and the remainder was promptly titrated to pH 7. The water content of the tissue was measured at the end of certain experiments. To accomplish this the exposed area of tissue was promptly cut out with fine scissors and placed serosal side down on filter paper. The upper or mucosal surface was lightly scraped with a glass slide to remove the adherent mucus coat and the wet weight determined; dry weight was obtained after desiccating overnight at 9OOC.
A primary goal of this study was to test for the presence (or absence) of a relationship between net volume flow and acid production. In a group of experiments, gastric mucosae was brought to a near-resting (i.e., nonsecreting) state by either prolonged incubation or treatment with 10m3M metiamide. The latter procedure was quicker and more convenient, yielding a small or negligible secretory rate; otherwise, the two procedures gave similar results. Figure 2 shows a representative experiment in which a mucosa was treated with metiamide (lo-" M) in the nutrient solution for 1 h, washed for 3 h, and then maximally stimulated with histamine (10v4M) and theophylline (5 x 10e3 M) in the nutrient solution. Spontaneous acid secretion was promptly depressed by metiamide and net volume flow was also reduced after a lag period. In 3 h, acid secretion appeared to plateau at zero and net volume flow at about -30 pi/h (the minus sign indicating flow from the volume chamber towards the lumen). Stimulation of acid secretion was accompanied by a small but definite increase in volume flow towards the secretory solution. The relatively constant loss in volume of nutrient solution when the mucosa is resting is due to evaporation and other factors: it is examined in more detail in a later section. Because treatment with secretagogues increases both acid secretion (peq/h) and rate of volume loss from the nutrient solution (pi/h), it seems reasonable to define the secreted acidity as the quotient of these changes (peqlpl). The change is calculated between the mean of the two periods just before addition of secretagogues and the mean of the final two periods, omitting the first period of stimulation as a transient. On this basis, the experiment of Fig. 2 yields an acidity of 0.31 eqlliter, well above isotonicity. The peak rate of acid secretion in the experiment of Fig. 2 was 12.3 peq/h for a chamber area of 5.1 cm* or 2.4 peq/cm** h, which is somewhat less than values previously reported for the stimulated preparation (1, 9). The large area used in the present study could include some nonoxyntic glands or the drop in secretory pH during the hour (to about pH 2.5) could lead to backdiffusion or reduction in H+ output. The latter is not rate-limiting because instillation of a hypotonic
TEFLON
FIG. 1. Cross-section of chambers used to measure volume flow, roughly to scale. Approximate volume of right-hand chamber (volume chamber) 3.5 ml; left-hand chamber, 7 ml. Area of central aperture 5.1 cm*.
1-1 Acid
secretion
(yEq /Id
( W-B Al$tiomide 1 1
Volume
1
flow
(yl/h) -I20
-80
FIG. 2. Effect of secretagogues on mucosa brought to resting state with metiamide (10d3 M). Isotonic secretory solution (110 mM NaCl) was instilled in the lumen. Acid secretion, solid line; net volume flow, dashed line; minus sign for latter axis denotes flow out of volume chamber towards the larger chamber (lumen).
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (128.059.222.107) on January 14, 2019.
OSMOTIC
FLOW
OF
WATER
IN
GASTRIC
E65
MUCOSA
t-1 Acid
secretion
Volume (yl/ h)
flow
-120 FIG. 3. Effect of SCNAxes as in Fig. 2.
I --a--l
----.
(lo-*
M) on stimulated
mucosa .
-80
-P I loo
I
I
200
300
1 400
’ 500
1 O
Minutes
secretory solution improved the secretory performance (as shown below). It is more probable that oxygen supply to the tissue is the major factor limiting acid output; the experimental design did not permit stirring or oxygenating the nutrient solution. In other experiments mucosae were first stimulated after recovery from dissection and then inhibited. Either 10 mM NaCl in the nutrient solution was replaced by 10 mM NaSCN, leaving secretagogues in the solution; or 10 -3 M metiamide was included in the nutrient solution and secretagogues removed. Figure 3 shows an experiment with SCN. Whereas acid secretion reached a plateau within an hour of stimulation, net volume flow took longer to reach a steady state. The mean of the last two periods before SCN addition is taken to represent the secreting state, and the mean of the final two periods, the inhibited state. Calculation of the secreted acidity from the changes between these states yielded 0.27 eq/liter. Experiments with metiamide gave similar results, except that the final secretory rate was near zero in spite of the long prior exposure to histamine and theophylline. Each experiment was analyzed in the above manner to yield two ‘sets of data, at high and low secretory rates. The results of 20 experiments with either stimulation followed by inhibition, or vice versa, are combined in Fig. 4. Fitting the points by least squares (5) to the line, y = a + bx, yields net volume flow = -51.5 - 3.50 (acid secretion) Here volume flow is in microliters per hour and acid secretion in microequivalents per hour. The mean acidity from these experiments is the reciprocal of the slope, l/b = 0.286 eq/liter. The standard deviation of the slope, that the slope is zero sb = 0.668; hence the possibility can be rejected (P C 0.001). These results show that net volume flow is correlated with acid secretion, and the acidity of the secretion is well above isotonicity. If the instillate has diminished the endogenous gradient due to HCl, thereby reducing
Volume (yl /h)
flow
v -40
l .
l
. 1
IA
0’
R9sfhg
- stweting
A Secrefhg -mfihg fmetion7i&/ l Secreting - est..ng
I
I
J
4
8
I2
Acid
secretion
fSCN/
, 16
(y Eq /h )
FIG. 4. Cumulative plot of data from 20 mucosae. Filled circles denote resting and stimulated states from experiments like that of Fig. 2, including 6 with metiamide and 4 with prolonged incubation to lower spontaneous rates. Filled triangles denote stimulated and inhibited states for 5 experiments with metiamide, and filled squares, stimulated and inhibited states for 5 experiments with SCN- (as in Fig. 3).
water flow, it should be even more effective when diluted. Accordingly, a series of experiments were performed with 11 mM NaCl (10% of normal, or 0.1 S) as secretory solution. This solution was first instilled well before stimulating the mucosa, as illustrated in Fig. 5. The drop i .n volume flow towards the 1umen in the first of acid secretion few hours reflects both the inhibition by metiamide and the establishment of a passive osmotic flow out of the lumen. Stimulation of acid secretion led to a very large increase in this parameter and a modest change in net volume flow. The difference in means of the final two resting periods and the final two secreting periods gave an acidity of 1.02eqlliter in this experiment. Resting and secreting data were obtained in this manner for nine mucosae, plotted in Fig. 6. For comparison the results for stimulation with normal secretory
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (128.059.222.107) on January 14, 2019.
R. P. DURBIN
E66 S ( -1 171 Acid secretion (pEq/h)
Hyptomii Metio fni& I I
secretory 10.LSI
1 Histomhe + theophyhne I 1
I Volume flow (yllh) -120
1
FIG. 5. Effect of secretagogues on mucosa brought to resting state with metiamide, in the presence of hypotonic secretory solution (11 mM NaCl) on the lumen before and during stimulation. Axes as in Fig. 2.
-80
1
I -w-w
Minutes Volume flow solution from Fig. 4 are repeated here. Fitting both sets OJVhh) of points to the line, y = a + bx, by least squares yields -r20the parameters listed in Table 1. Comparison of the two slopes by standard methods (5) shows that their difference is highly significant (P < 0.001). Instillation of the hypotonic secretory solution has increased the mean Secreted acidity to ilO. = 1.87 eqlliter. Of interest here is the effect of hypotonic secretory solution to increase the maximal rate of acid production. The mean of this parameter (t SE) for the experiments 0 00 of Fig. 6 was 16.3 -+ 0.95 *q/h with 0.1 S, compared to 40I I I I I 11.3 2 0.59 peq/h with 1.0 S. The difference in means is 0 4 8 I2 I6 20 significant (P < 0.001). A clue that might account for Acid secretion (yEq/h) this result is provided by unpublished electron microFIG. 6. Cumulative plots for the resting to secreting transition. graphs of H. F. Helander. He found that the lateral Open circles denote resting and stimulated states from 9 experispace between oxyntic cells appeared dilated in tissues ments with hypotonic secretory solution, 4 like that of Fig. 5 using that had been fixed after exposure to 0.1 S, in comparimetiamide, and 5 using prolonged incubation to lower spontaneous son to control tissues. Osmotic expansion of such spaces rates. Filled circles are data repeated from Fig. 4 for the resting to could provide additional oxygen by diffusion from the secreting transition with isotonic secretory solution (10 expts). well-stirred secretory solution to the lateral surfaces of TABLE 1. Results of stimulation: least-squares the oxyntic cells. =a+bx The large variation in acidity reported here raises the jittoy possibility that the frog is deficient in some manner in Secretory SoluNo. of Expts Intercept (a), &h Slope (b 2 St)), pllpeq the osmotic mechanism that yields isotonic gastric juice tion in mammalian stomach. For this reason experiments 1.0s 10 -41.8 -4.24 + 0.67 were done in which gastric juice was collected from the 0.1s 9 1.8 -0.534 AI 0.65 empty lumen. The need for quantitative recovery required that the mucosa be mounted so that the luminal surface faced the volume chamber. Figure 7 shows a nutrient solution. In this respect frog gastric juice representative experiment in which secretion was col- conforms to mammalian gastric juice. It should be lected for 6 h from a stimulated mucosa. In agreement noted, however, that NaCl is probably absent from pure with a previous report (19), the secretion was a mixture oxyntic cell secretion in the mammal (17). of HCl, NaCl, and KCl. The Cl- concentration over the The intercepts on the y-axis in Figs. 4 and 6 reveal a six periods was 124 t 3 meq/liter (mean t SE), consistsubstantial loss of volume from the nutrient solution in ent with the sum of the three cations, 126 t 4 meq/liter. the absence of acid production. About 15 pi/h is evapoIt appears that these four ions account for the secretion, rative loss, provided this is the same with a mucosa which is thus slightly hypertonic to frog plasma or mounted between chambers as with Parafilm (cf. METH-
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (128.059.222.107) on January 14, 2019.
OSMOTIC
FLOW
OF
WATER
IN
GASTRIC
I50
E67
MUCOSA
Vohme
TABLE 2. Effects of inhibitors on volume flow in unstimulated mucosae
f/o w
Al/h Agent
50
Metiamide Ouabain mM
Dinitrophenol 50
No. of Periods
12 8 8
Values are means 2 SE for net volume was metiamide, 3; ou .abain 9 2; dinitrophenol,
Net Volume Flow
-28.7 -28.8
2 3.5 + 5.1
-24.4
+ 6.6
flow. Number 2.
of mucosae
0 t mM
50 0t IO
mM
-04-O
l/-•
0t
[K’l
[c/-l
-a-
+--
too mM t 0'
I
I 2
I
1 4
I
I 6
Hours 7. Consecutive juice collections from stimulated mucosa. Mucosa was dissected at -2 h and stimulated with histamine and theophylline at -1 h. Volume flow is positive here because mucosa is reversed, with lumen facing volume chamber. The large volume flow in the 1st h represents drainage of residual secretory instillate, in part; the steady state thereafter represents secretion. FIG.
ODS). An appreciable loss must still be accounted for. Sources-to be considered include a) secretion of NaCl to the lumen, a possibility raised by the appearance of NaCl in the collected gastric juice; b) secretion of NaHCO, to the lumen (11); c) a steady swelling of the mucosa or hydration of mucus by incorporation of fluid from the nutrient solution. One approach to the problem was to obtain a control group of mucosae in which acid secretion was inhibited by metiamide to compare with groups in which either 2,4dinitrophenol (DNP) was used as an inhibitor of oxidative phosphorylation or ouabain as an inhibitor of active Na+ transport. Metiamide was used at lo-” M for the 1st h and at 10e4 M for 6 h thereafter. A level of 2 x 10M4 M in the nutrient solution was adopted for DNP and ouabain, sufficient to inhibit acid secretion, electrogenie Cl- transport, and HCO,- secretion (9, 11). With all three inhibitors, acid secretion became negligible and net volume flow reached a plateau after 3 h. The results for the ensuing 4 h are summarized in Table 2. The net volume flow is seen to be the same for all inhibitors, within experimental error. The net volume flow in the presence of metiamide (Table 2) is probably a better estimate of the resting flow than the intercepts of Figs. 4 and 6. Thus only 2915 (evaporative loss) or 14 pi/h remain to be accounted for. Very little of this could be due to active secretion, as shown by the negative results with DNP and ouabain. It appears that the third possibility listed above is the most likely, that part of the resting volume flow is due to hydration of mucus or swelling of the tissue by absorption of water from the nutrient solution. It therefore seemed worthwhile to make a direct test
of tissue swelling. In a series of experiments, water content of mucosae exposed to metiamide for 2 h was compared to that of mucosae similarly treated for 9 h. Acid production was small or negligible. In the 2-h studies, water content was 6.47 t 0.28 ml H,O per g dry wt (mean t SE; n = 7) compared to 7.51 -t 0.37 (n = 7) in the 9-h experiments. The difference is barely significant (P < 0.05). The mean dry weight here was 64 mg; hence the average increase in tissue water in 7 h was 67 mg, i.e., slightly less than 10 pi/h. Thus part of the resting volume flow from the nutrient solution may reflect tissue swelling. Note, however, that the calculation of secreted acidity from the change in volume flow due to stimulation or inhibition is not affected by steady-state losses due to evaporation and tissue swelling. DISCUSSION
The present study shows that a net volume flow is associated with acid production by the isolated frog stomach. The finding is not unexpected because net water flow coupled to active ion transport has been described in a number of epithelia (16). To demonstrate the relationship here, it was necessary to use a large mucosal area and maximal changes in secretory rate, factors neglected in previous studies (8, 25). A significant component of net volume flow is independent of HCl or NaCl secretion, which may account for the negligible effect of Cl- withdrawal on volume flow seen by Villegas and Sananes (25) in unstimulated mucosa. The present chambers permit the use of bullfrogs of moderate size and hence are more convenient than those employed in a preliminary study (7). The apparatus is not suited to the measurement of small or rapidly changing flows, however; tissue swelling or shrinking can affect the volume of nutrient solution recovered, and time must be allowed for such changes to reach a steady state. Acid secretion measured in the presence of an isotonic instillate in the lumen was hypertonic, as found also for low rates of acid production in mammalian stomach. The effect has been shown to arise primarily from a decrease in water flow, rather than additional HCl secretion (22). The effect was observed when isotonic NaCl, sucrose, or mannitol solutions were instilled, but not with isotonic HCl (22). The latter is consistent with the usual finding that gastric juice collected from an otherwise empty lumen is isotonic. The instillate effect is not unique to stomach; it has been described also in gallbladder. This organ normally transfers an isotonic NaCl solution in the direction,
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (128.059.222.107) on January 14, 2019.
R. P. DURBIN
E68
osmotic permeability is increased by stimulating acid lumen-to-blood. The lateral and basal surfaces, bathed secretion (24). For simplicity consider P,,%{of the resting by the serosal solution in vitro, are the site of active Na+ transport (21). Isosmotic replacement of part of the mucosa. The two intercepts on the y-axis of Fig. 6 differ serosal NaCl by sucrose significantly increased the by 44 pllh. Hence tonicity of the absorbate, from 161 to 211 mM (26). Posh1(mucosa) In both stomach and gallbladder, the solution in the = 44/&l ~l/cm*~h per 0.9 x 110 mM of NaCl compartment towards which active transport is directed = 0.09 $/cm* h per 1 mM difference in [NaCl] has been altered. The effect is due to a change in solute: for example, HCI is replaced by NaCl in the case of Net flow has been divided by chamber area and the stomach, and NaCl by sucrose in gallbladder. In either solute (NaCl) assumed to have a reflection coefficient of instance, the replacing substance is larger and diffuses less readily into the region in which osmotic flow is unity. The osmotic permeability thus calculated is close to determined, near the transporting surface. The outcome that of frog skin in vitro. House (15) found for the latter suggests that the replacement has diminished the norPosM of 4 x 10B7 cm/s per atmosphere pressure difference. mal osmotic gradient. Whitlock and Wheeler (26) disIn the present units his figure yields: cussed their results quantitatively on the basis of the double-membrane model of osmotic water flow (2, 6). h Diamond (4) has criticized their treatment, but he also I& (skin) = 0.07 $/cm2. per 1 mM difference in [NaCl] invoked the difference in diffusion rates of NaCl and sucrose to explain their findings. It is of interest to compare these results for intact Instillation of hypotonic secretory solution had an epithelia with the corresponding figure for a synthetic especially striking effect in the present study, practilipid bilayer. Using such a membrane, Hanai and cally abolishing the component of water flow coupled to Haydon (12) found that an osmotic difference of 0.157 M acid secretion. This observation is a key finding because NaCl gave rise to a net flow of lo-” cm/s. In the units it shows clearly that the endogenous osmotic gradient used here in frog stomach is accessible to the external solution. An additional factor determining the amount of os- PosM(bilayer) = 0.23 pl/cm2 h motic water flow is the conformation of the transporting per 1 mM difference in [NaCl] surface. Dainty and House (3) considered the extreme case of active transport of solute into a flat, unstirred This is a value about three times that of intact frog skin layer parallel and adjacent to an epithelium. They point or gastric mucosa. Naively one might have expected the out that the magnitude of the osmotic gradient is contrary result. Thus folding of the gastric epithelium limited by diffusion of solute away from the unstirred presents many square centimeters of apical membrane layer. The resulting osmotic flow is small and the in parallel per square centimeter of chamber area (1, transported fluid hypertonic. 14). An increase in area might have been expected to In reality transporting epithelia are not flat, and the increase PosMin proportion. Either the osmotic permearelevant morphology needs to be examined in each bility of the transporting surface is much less than that instance. Infolding of the epithelium shields the endog- of a lipid bilayer or the full strength of the applied enous osmotic gradient, increasing the velocity of bulk osmotic gradient does not reach the appropriate surface. flow of secreted or absorbed fluid and hampering access The latter explanation seems more likely. The applied of external solute to that surface. gradient is flat at the macroscopic level of the two The apical surface of the oxyntic cell in stomach is solutions bathing the isolated epithelium. Clearly it almost surely the site of active transport, and therefore would be impossible for the gradient thus defined to the locus of the endogenous osmotic gradient. This reach all the transporting surface of frog skin, which is surface changes remarkably on stimulation of secretion, dispersed over a number of cell layers (20). forming numerous leaflets that result in an extensive The isolated gastric mucosa, on the other hand, proliferation of the apical surface (1, 14). If active presents a single cell layer that is highly folded. It transport and osmotic flow are distributed uniformly seems paradoxical that the instillate effects have shown over this surface, a modest osmotic gradient should that the transporting surface is accessible to the lusuffice to drive the observed water flow (10). It is this minal solution; yet, the low osmotic permeability suggradient that is affected by the secretory instillate, as gests that only a fraction of the applied osmotic gradient shown in this study. reaches this surface. A resolution of the problem will The difference in net volume flow observed for iso- require an exact analysis of the distribution of the tonic (1.0 S) and hypotonic (0.1 S) secretory solution can osmotic gradient over the complex Sdimensional strucbe used to estimate an apparent osmotic permeability, ture of the transporting surface. P oskl,for the intact mucosa. At best PosMcan only provide I thank Drs. Venetia France and Margareta Ekblad for helpful a crude estimate, due to lack of stirring of the nutrient criticism. solution and other special conditions in the present This work was supported by National Institutes of Health Grant study. AM-18395 and National Science Foundation Grant PCM 76-04148. The vertical separation of the two lines in Fig. 6 Received 16 May 1978; accepted in final form 28 August 1978. increases with acid secretion. This may indicate that l
l
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (128.059.222.107) on January 14, 2019.
OSMOTIC
FLOW
OF WATER
IN GASTRIC
E69
MUCOSA
REFERENCES 1. CARLISLE, K. S., C. S. CHEW, AND S. J. HER~EY. Ultrastructural changes and cyclic AMP in frog oxyntic cells. J. Cell Biol. 76: 31-42,1978. 2. Cuw, P. F. Na, Cl, and water transport by rat ileum in vitro. J. Gen. Physbl. 43: 1137-1148,196O. 3. DANIT, J., AND C. R. HOUSE. “Unstirred layers” in frog skin. J. Physiol. London 182: 6678,196s. 4. DIAMOND, J. M. Transport mechanisms in the gallbladder. In: Handbook of Physiology. Alimentary Canal. Washington, D.C.: Am. Physiol. Sot., 1968 . v% 6, vol. v, p. 2451-2482. 5. DIXON, W. J., AND F. J. MASSEY, JR. Introduction to Statistical Analysis. New York: McGraw-Hill, 1969, chapt. 11. 6. DURBIN, R. P. Osmotic flow of water across permeable cellulose membranes. J. Gen. Physiol. 44: 31%326,196O. 7. DURBIN, R. P. Coupling of water to solute movement in isolated gastric mucosa. Ciba Found. Symp. 38: 161-177,1976. 8. DURBIN, R. P., H. FRANK, AND A. K. SOLOMON. Water flow through frog gastric mucosa. J. Gen. Physiol. 39: 535-551, 1956. 9. DURBIN, R. P., F. MICHELANGELI, AND A. NICKEL. Active transport and ATP in frog gastric mucosa. B&him. Biophys. Acta 367: 177-189, 10. DURBIN, R.
11.
12. 13. 14.
1974.
P., AND H. F. HELANDER. Distribution of osmotic flow in stomach and gallbladder. B&him. Biophys. Acta. In press. FLEMSTR~M, G. Active alkalinization by amphibian gastric f’undic mucosa in vitro. Am. J. Physiol. 233: El-E12, 1977 or Am. J. PhysioZ.: EndocrinoZ. Metab. Gastrointest. PhySioZ. 2: El-E12,1977. HANAI, T., AND D. A. HAYDEN. The permeability to water of bimolecular lipid membranes. J. Theor. BioZ. 11: 370-382, 1966. HEINZ, E. -r die primare Aziditit der Magensaure. B&him. Biophys. Acta 6: 434-444,195l. HELANDER, H. F., AND R. P. DURBIN. Secretory surface area and phosphatase activity of frog gastric mucosa. Am. J. Physiol. 232:
15. 16.
17.
18.
E48-52, 1977 or Am. J. Physiol.: Endocrinol. Metab. Gastrointest. Physiol. 1: E48-E52, 1977. HOUSE, C. R. The nature of water transport across frog skin. Biophys. J. 4: 401-416,1964. HOUSE, C. R. Water Transport in CeZZs and Tissues. London: Arnold, 1974. HUNT, J. N., AND B. WAN. Electrolytes of mammalian gastric juice. In: Handbook of Physiology. Alimentary CanaZ. Washington, D.C.: Am. Physiol. Sot., 1967, sect. 6, vol. II, part 2, p. 781804. LINDE, S., T. TEORELL, AND K. J. OBRINK. Experiments on the primary acidity of gastric juice. Acta Physiol. Scund. 14: 220-
232,1947. 19. MAKHLOUF,
G. M., AND G. R. DUCKWORTH. Secretion and electrical activity of a unilateral in vitro gastric mucosa. Gastroenterology 65: 907-911, 1973. 20. MILLS, J. W., S. A. ERNST, AND D. R. DIBONA. Localization of Na+-pump sites in frog skin. J. CeZZBioZ. 73: 88-110, 1977. 21. MILLS, J. W., AND D. R. DIBONA. Distribution of Na+ pump sites in the frog gallbladder. Nature 271: 273-275, 1978. 22. MOOI)Y, F. G., AND R. P. DURBIN. Effects of glycine and other instillates on concentration of gastric acid. Am. J. Physiol. 209: 122-126,1965. 23. TEORELL, T.
On the primary acidity of the gastric juice. J, Physiol. London 97: 308-315, 1940. 24. VILLEGAS, L. Action of histamine on the permeability of the frog gastric mucosa to potassium and water. Biochim. Biophys. Actu 75: 377-386,1963. 25. VILLEGAS, L., AND
L. SANANES. Independence between ionic transport and net water flux in frog gastric mucosa. Am. J. Physiol. 214: 997-1000,1968. 26. WHITLOCK, R. T., AND H. 0. WHEELER. Coupled transport of solute and water across rabbit gallbladder epithelium. J. CZin. Invest. 43: 2249-2265, 1964.
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (128.059.222.107) on January 14, 2019.
