ANALYST, MAY 1991, VOL. 116 463 pH Dependence of Hydrochloric Acid Diffusion Through Gastric Mucus: Correlation With Diffusion Through a Water Layer Using a Membrane-mounted Glass pH Electrode C. V. Nicholas, M. Desai and P. Vadgama Department of Medicine (Clinical Biochemistry), University of Manchester, Clinical Sciences Building, Hope Hospital, Salford M6 8HD, UK M. 6. McDonnell Department of Chemistry, University of Southampton, Southampton SO9 5HH, UK S. Lucas Co m pu ta tio na I Group, Computing Department, University of Man c h este r, Stop fo rd Building, Man c h este r M73 9PT, UK Solute diffusion coefficients (D) can indicate a dependence upon actual solute concentrations. Here a single compartment has been utilized, in which effective HCI diffusion t o a membrane-mounted glass pH electrode can be measured across the pH spectrum.The study has investigated HCI diffusion through both mucus and water layers as a function of HCI concentration. The observed dynamic responses of a liquid-film and mucus- coated electrodes over a range of HCI concentrations suggest that the speed at which equilibrium is attained is pH dependent; equilibrium was reached rapidly under more acidic and alkaline conditions. Estimated values of DHCl also indicate a strong pH dependence for both liquid film and mucus. In both instances, a >lo-fold reduction in DHcl at pH 7.5 as compared with that at pH 3.5 has been demonstrated. Furthermore, estimated values of DHCl are approximately 4-fold smaller through the mucus gel, as compared with a water layer. The findings indicate that the most powerful influence on diffusional resistance is pH itself, whereby a marked drop in H+ diffusion is likely t o occur towards neutral pH irrespective of the composition of the gel barrier.Possible implications of the findings are discussed in relation t o mucosal protection from acid. Keywords: Hydrochloric acid diffusion; gastric mucus; pH electrode Mucus forms a continuous, adherent visco-elastic gel layer over the gastrointestinal mucosa, which in man has an estimated depth of 50450 pm (Kress et a f . I ) . In the stomach and duodenum, mucus has been considered to play an important role in protecting the mucosa from damage by luminal acid (Allen and Garner’). Studies of its diffusional resistance to HCI have revealed a resistance that is 4-5-fold greater than that of an equivalent unstirred layer of water (Williams and Turnberg,3 Pfeiffer4 and Turner et a f .5 ) . Furthermore, steep pH gradients have been observed within mucus, across intact gastric mucosa, by means of implanted micro-pH electrodes (Williams and Turnberg,6 Bahari et a f . ,’ Takeuchi et a f . * and Flemstrom and Kivilaakso’). These have been explained on the basis of an intra-gel neutralization of HCI with hydrogen carbonate secreted by the surface epi- thelium (Allen and Garner2 and Flemstrom and Kivilaakso’). In this concept of the mucus-hydrogen carbonate barrier, the neutralizing action of HC03- is considered to be potentiated by neutralization occurring in the restricted volume of the mucus gel phase.A mathematical model developed by Engel et a f . 1 0 for intra-mucus neutralization, however, suggested that a relatively minor (5 mmol dm-3) drop in H+ concentra- tion was likely to be generated across the mucus layer based on existing, reported values for gel layer thickness, hydrogen carbonate flux and HCI diffusion. Solute diffusion coefficients can show a dependence upon actual solute concentrations (Crank”) and in the present study HCI diffusion through both mucus and water layers as a function of HCI concentration has been examined. A signifi- cant change from earlier reported values for HCI diffusion could provide further information concerning the resistive contribution made by mucus in its protective role over the gastroduodenal mucosa. A more detailed profiling of hydrogen ion diffusion is warranted, as its diffusion is unique in involving passage from one water molecule to the next, through the formation of a sequence of hydrogen bonds (Robinson and Stokes”).In a previous study, HCI diffusion through water was observed to be retarded by a factor of 100, as compared with earlier reported values, when conditions approached neutrality (Nicholas et af. 13)- The correlation between diffusion through water and mucus layers is reported here, and possible implications of these findings for mucosal protection from acid are discussed. Theory The two-compartment diffusion chamber is appropriate for the study of most solute species (IUPAC14). However, for H+ diffusion, measurement requires highly acidic (pH <2) con- ditions (Williams and Turnberg,6 Robinson and Stokes12 and Slomiany et af.15) if buffering is to be avoided.A one- compartment system based on the glass pH electrode, which allows measurements under acidic through to neutral and alkaline conditions, was used in the present study. The dynamic response of a pH electrode and its approach to an equilibrium may be modelled in terms of diffusion through a stagnant, unstirred layer over the sensor surface (Morf and Simonl6). Provided that the diffusion layer, and not the intrinsic electrode response, is rate limiting, the change in the electrode e.m.f. showed the following dynamic time depen- dence in its approach to an equilibrium response: Here, E, is the electrode e.m.f. at any given time t , E,, is the electrode equilibrium response, S is the slope of the pH calibration graph (mV per decade), [H+Io is the H+ concentra- tion at time zero, [H+],, is the H+ concentration at the final equilibrium response and ‘I: is the time constant for the system.The value of T is governed by both the thickness of the464 ANALYST, MAY 1991, VOL. 116 unstirred layer, d , and by the H+ diffusion coefficient, D , within that layer: The measurement of T permits the calculation of D provided d is known. Alternatively, mucus and liquid films have been created over the glass surface of a pH electrode (Nicholas et aZ.13), which provided a well-defined boundary layer in a stirred solution and which, furthermore? were of sufficient depth both to define the dynamic response of the pH electrode according to eqns.(1) and (2) and eliminate the effects of an external Nernst diffusion layer. Experimental The measuring glass pH electrode (Type CETL; Russell, Fife, UK) was used in conjunction with a saturated calomel reference electrode (Microelectrodes, Londonderry , NH, USA). Electrode e.m.f. was measured using a pH meter (PCMKI, Newcastle upon Tyne, UK) and output recorded at a strip-chart recorder (Linseis, Selb, Germany). A combi- nation pH electrode served as a follower electrode to monitor bulk solution pH during the addition of HCl in the pH jump experiments. All standard reagents were of AnalaR grade and purchased from BDH (Poole, Dorset, UK); bovine serum albumin (BSA) was obtained from Sigma (Poole, Dorset, UK). Native pig gastric mucus was removed as described previously (Williams and Turnberg3), from the stomachs of abattoir animals that had been killed recently.Mucus was applied to the tip of the measuring glass pH electrode which had a pre-mounted 135 pm nylon netting that acted as a spacer. A uniform gel or mucin layer was then created by stretching an external 10 pm Cuprophan dialysis membrane layer, using a Cuprophan from a haemodialysis cartridge (Gambro, Lund, Sweden). For measurements through aqueous films the nylon spacer and dialysis mem- brane were used alone. Measuring and follower electrodes were immersed in a chamber containing 175 ml of solution that was stirred rapidly (Vadgama and Albertil’), and 1 mol dm-3 HCI was injected via an automatic pipette over a period of 1-2 s in order to create a change in the pH of the bulk solution of about 1 pH unit; the temperature of the solution was 21 k 2 “C.Results The dynamic response of the uncovered glass pH electrode in stirred solution, as monitored at the strip-chart recorder, was complete within 2 s of the addition of HCI over the entire pH range used in these studies. The magnitude and dynamic response were unaffected by either previous contact with mucus gel or with the bulk solutions used. Dynamic response profiles were reproducible to within 5% with respect to e.m.f. The observed dynamic responses of a liquid-film and mucus- coated electrodes over a range of pH values in the presence of albumin, as a non-diffusible buffer, are shown in Fig. l(a) and ( b ) , respectively. These suggest that the speed at which equilibrium is attained depends upon pH, with equilibrium being reached more rapidly under more acidic and alkaline conditions; this would not be expected on the basis of eqn.(1) (Morf and Simonl6) , which predicts similar dynamic responses across the pH spectrum, provided the magnitude of the pH jump is uniform. However, when eqn. (1) was used to calculate the dynamic electrode response, it showed good agreement with observed e.m.f changes, both for liquid-film electrodes (Fig. 2) and the mucus-coated electrodes (Fig. 3) over a range of pH changes. This indicates a change in effective diffusion coefficients for HCI (DHCI) over a range of pH values. The value of DHCl was estimated using BSA as a non-diffusible buffer to assist in the pH stabilization of the t + E ui H-r G- EF- D- C / b’T Gf F/ E A 1 min Time -c Fig.1 Dynamic responses of a pH electrode mounted with a 135 pm nylon spacer and dialysis membrane. (a) Liquid film; pH change: A , 5.59-4.98; G, 4.98-3.07; and H, 3.07-2.51. (b) Mucus layer; pH change: A, 10.02-9.08; B, 9.08-7.80; C, 7.80-6.73; D , 6.73-5.62; E, 5.62-4.72; F, 4.724.00; and G, 4.00-2.88. Albumin (20 g 1-1) was used as a non-diffusible buffer 8.70-7.42; B, 7.42-7.07; C, 6.83-6.64; D, 6.46-4.29; E, 5.89-5.59; F, bulk solution. The results are shown in Fig. 4; these indicate a strong pH dependence for DHCl for both a liquid film and mucus. In both situations there is a >lO-fold reduction in DHCl at pH 7.5, as compared with pH 3.5. In addition, estimated diffusion coefficients are approximately 4-fold smaller through the mucus gel, as compared with a water layer.Interestingly, as conditions are made more alkaline, DHCl is seen to increase again. A reliable estimation of DHCl at pH <2 was precluded by the very rapid responses of the film-coated electrode. For the citrate buffer examined here, background ionic strength had no apparent effect on the trend in DHCl values observed over the pH range investigated, as shown in Fig. 5. However, results obtained using different concen- trations of glucosamine as a diffusible buffer in the presence of albumin (Fig. 6) suggest that there may be an effective increase in DHCI, particularly at more acidic and alkaline pH values when a high concentration of such a diffusible buffer is used. This possibility receives some support from the finding of a higher DHCl with a mucus-coated electrode at pH 10.5 in the presence of 30 mmol dm-3 salicylate and albumin (Fig.7) as compared with albumin alone. Discussion The concept of the mucus-hydrogen carbonate barrier (Allen and Garner2 and Flemstrom and Kivilaaksog) has received important supportive evidence (Williams and Turnberg6 and Flemstrom and Kivilaaksog), and continues to attract interest (Munster et al. 19). The effectiveness of this barrier would be critically affected by the rate at which protons approach the surface epithelium from the lumen. A mucus gel layer with a high diffusional resistance would appear to be ideal for such a system; however, the retardation of the diffusion of H+ in mucus would appear to be insufficient to explain the type of pH gradients observed (Engel et al.lo). Part of the explanation for the retardation of the diffusion of the H+ ion in mucus is the net negative charge of the constituent glycoprotein. This may operate by means of a Donnan exclusion mechanism, although comparison with uncharged gel suggest that the effect is minor (Lee and Nicholls20). Any specific ordering of water molecules around the glycoprotein structure is likely to be minimal (Soggett21) and, therefore, unlikely to result in significant additional diffusional resistance, particularly in view of the low ( ( 5 % ) concentration of the glycoprotein in the gel (Allen and Garner2). For native mucus, particulateANALYST, MAY 1991. VOL. 116 465 20 (a) 25 20 - 15 - 10 - - 5 - 0 - 1 1 I I 5 0 20 40 60 80 100 0 100 200 300 400 500 2 0 + 30 E d 25 20 15 10 5 0 14 12 10 8 6 4 2 0 I I I I I I 0 100 200 300 400 500 0 20 40 60 80 100 Time/s Spacer and dialysis membrane mounted pH electrode response profiles tor a liquid film at various pH jumps.(a) pH 10.99-9.99; (b) Fig. 2 9.00-7.4; (c) 6.97-5.91; and (d) 4.03-2.96. Calculated (0) and measured (a) e.m.f. values are compared 0) I I I I I I I 2 0 30 60 90 120 150 180 210 .;. 40 E d 30 20 10 200 400 600 800 1000 0 0 200 400 600 800 1000 0 60 120 180 240 300 360 420 Time/s Fig. 3 7.98-7.03; (c) 6.1 1-5.04; and (d) 3.98-2.93. Calculated (0) and measured (0) e.m.f. values are compared. Spacer and dialysis membrane mounted pH electrode response profiles for a mucus layer at various pH jumps.(a) pH 11.5-9.98; (b) c 80 70 7 60 50 .O 40 = 30 N z 0, .- g 20 2 10 .- v) i 0 2 4 6 8 1 0 1 2 Mid-point pH jump Fig. 4 Effective DHCl calculated from dynamic responses of a spacer and dialysis membrane mounted pH electrode in albumin buffer: 0, liquid film only; and 0, liquid film with mucus. pH values are the mid-points of pH jumps of magnitude about 1 material, protein and lipid, do appear to confer additional diffusional resistance (Slomiany et a1.22). The summation of all the above effects is undoubtedly important in reducing DHCl relative to that of a liquid film. The effect observed in this study is consistent with that reported previously by Williams and Turnberg,3 who used a classical two-compartment diffusion chamber. However, the present work would appear to indicate that the most powerful influence on diffusional resistance is pH itself, whereby a marked drop in H+ diffusibn is likely to occur towards neutral pH irrespective of the composition of a gel barrier (Fig.4). The DHCl has not previously been measured directly, largely because of the problems of achieving a stable pH under near neutral conditions without buffering. As a result, all previous studies of DHC] in mucus have been limited to using HCI at a pH of about 1. The technique reported here permits deter- mination of &-] in mucus over most of the pH spectrum. In the present study, the small unbuffered compartment, i.e., the466 F 2 60 a 0 5 40 .- 8 .- s 20- s In b- 0 - ANALYST, MAY 1991, VOL. 116 - - c I ," 30 E 2 % 20 5 s s 10 z n r- .- 0 .- In 0 2 4 6 8 Mid-point pH jump Fig.5 Effective DHCl calculated from dynamic responses of a liquid film mounted pH electrode for solutions containing 10 mmol dm-3 citrate in 0 , 1 0 mmol dm-3 NaCl; and @, 300 mmol dm-3 NaCl (the latter data are redrawn from reference 18 for comparison) 0 0 2 4 6 8 Mid-point pH jump Fig. 6 Effective DHCl calculated for a liquid film electrode from dynamic responses in glucosamine solutions at various concentra- tions: 0, 10 mmol dm-3 (redrawn from reference 18 for comparison); A , 30 mmol dm-3; and @, 60 mmol dm-3 in albumin buffer 0 .- In 0- 4 6 8 10 Mid-point pH jump Fig. 7 Effective DHCl for a mucus-coated electrode calculated from dynamic responses for 30 mmol dm-3 salicylate in albumin buffer for a range of pH changes liquid or mucus layer over the pH electrode, was more readily controlled with regard to pH by incorporation of a non- diffusible buffer in the much.larger external compartment, i.e., the bulk solution. The effective DHC] values at about neutral pH would appear to have major implications for H+ diffusion in biological systems generally. The mechanism for such pH dependence remains to be elucidated; however, one possibility is the unique mechanism for proton transfer through water, involving multiple hydrogen bonding. The effect of a diffusible buffer (B) in solution is to augment proton transfer by means of a buffer shuttle (Engasser and Horvath23) : HB * HB H+ i I I 6- 4 B- This system would be expected to operate in a concentration- dependent manner and also to have maximal effect at a pH close to the pK, of the buffer (Vadgama and Alberti24).Both phenomena have been observed previously to affect the dynamic response of a pH glass electrode mounted with immobilized protein (Deem et al.25), and also native gastric mucus (Vadgama and Albertil7). Over the buffer concentra- tions used in the present study, any possible effect was minor, except when a high concentration of buffer was used (Figs. 6 and 7). The over-all reduction of at low glucosamine concentration may have been the result of buffer shuttling, although this could have been affected by the binding of ghcosamine to albumin. Estimated DHC] values obtained in different low ionic strength buffers of low relative molecular mass18 demonstrated some differences between citrate, ascor- bate and glucosamine, but these were relatively minor as compared with the steep drop in as neutral pH was approached.Values of DHCl were also lower for the diffusible buffers as compared with the albumin system. This raises the possibility that in the total absence of a buffer effect, values of may be even lower than reported here, as neutrality is approached. The results augment considerably the postulated resistive property of the surface mucus layer; for a relatively small rise in pH, H+ diffusion may be reduced by a factor of =lo. A possible implication of the present findings for the mucus-hydrogen carbonate barrier is that HC03- secretion into mucus may be designed to adjust the pH of the mucus to a range were DHCl is reduced, rather than to effect complete neutralization.Indeed, at the equivalence point HC03- cannot neutralize HCI; for 100 mmol dm-3 HCI the final pH would be 3-5 (Vadgama and Alberti24). It is even conceivable that a high concentration of buffer within the mucus layer, including the HC03-/C02 buffer system, would actually accelerate proton fluxes to the surface epithelium by the operation of a buffer shuttle. Thus, while the high urease activity of Heliobacter pylori generates ammonia which can neutralize H+ within mucus (Thompson et a1.26), there may be significant associated shuttling of H+ along the pH gradient in mucus which might actually contribute to mucosal damage associated with this organism. The present studies reconcile the idea of mucus as a resistive barrier (Williams and Turnberg6) with that of mucus as simply an unstirred water layer (Morris27). In conclusion, the results obtained should be regarded as effective DHCl values, but ones which nevertheless reflect the diffusion behaviour of the physiological system.Further comparative studies are in progress for other ion and solute species using the coated electrode technique described. References 1 Kress, S., Allen, A., and Garner, A., Clin. Sci., 1982, 63, 187. 2 Allen, A., and Garner, A., Gut, 1980, 21, 249.ANALYST, MAY 1991, VOL. 116 467 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Williams. S. E . , and Turnberg, L. A . , Gastroenterology, 1980, 79, 299. Pfeiffer, C. J., Am. J. Physiol., 1981, 240, B176. Turner, N . C., Martin, G. P., and Marriott, C., J. Pharm. Pharmacol., 1985, 37, 776.Williams. S. E., and Turnberg, L. A., Gut, 1981, 22,94. Bahari, H. M. M.. Ross, I . N., andTurnberg, L. A., Gut, 1982, 23, 513. Takeuchi. K.. Magee, D., Critchlow, J., Matthew, J., and Gilen, W., Gastroenterology, 1982,84, 331. Flemstrom, G., and Kivilaakso, E., Gastroenterology. 1983,84, 787. Engel, E., Peskoff, A., Kauffman. G. L., and Grossman, M. I., Am. J. Physiol., 1984, 247, G321. Crank, J., The Mathematics of Diffusion, Oxford University Press, Oxford, 1959. Robinson. R. A., and Stokes, R. H., Electrolyte Solutions. Butterworth. London, 1955. Nicholas. C. V., McDonnell, M. B., and Vadgama, P., 1. Chem. SOC., Chern. Cornmun., 1990, 4. 320. IUPAC, Conditional Diffusion Coefficients of Ions and Mol- ecules in Solution. An Appraisal of the Conditions and Methods of Measurements, Pure Appl. Chem.. 1979.51, 1575. Slomiany, B. L., Laszewicz, W., and Slomiany. A., Digestion. 1986, 33. 146. Morf, W. F., and Simon, W., in Ion-Selective Electrodes in Analytical Chemistry. ed. Freiser, H . . Plenum, New York, 1978, vol. 1, p. 211. 17 18 19 20 21 22 23 24 25 26 27 Vadgama, P., and Alberti, K. G. M. M., Experientia, 1983,39, 573. Vadgama, P., Nicholas, C. V., McDonnell, M. B., Lucas, S., and Desai, M., J. Chem. SOC., Faraday Trans., 1991,87,293. Munster, D . J., Robertson, A. M., and Bagshaw, P. F., N. 2. Med. J . , 1989, 102, 607. Lee, S. P., and Nicholls, J. F., Biotechnology, 1987, 24, 565. Soggett, A., in Water: A Comprehensive Treatise, ed. Franks, F.. Plenum, New York. 1980, vol. 4, p. 519. Slomiany, B. L., Piasek, A., Sarasick, J., and Slomiany, A., Scand. J. Gastroenterol., 1985, 20, 1191. Engasser, J. M., and Horvath, C., Biochim. Biophys. Acta, 1974, 358, 178. Vadgama, P., and Alberti, K. G. M. M., Digestion, 1983, 27, 203. Deem, G. S., Zabusky, N. J., and Sternlicht, H., J. Mernbr. Sci., 1978, 4, 61. Thompson, L., Tasman-Jones, C., Morris, A., Wiggins, P., Lee, S., and Furlong, L., Scand. J. Gastroenterol., 1989, 24, 761. Morris, G . P., Gastroenterol. Clin. Biol., 1985, 9, 106. Paper 1/001321 Received January I 1 th, 1991 Accepted January 23rd, 1991