Biomacromolecules 2003, 4, 922-927
922
The Gel Matrix of Gastric Mucus Is Maintained by a Complex Interplay of Transient and Nontransient Associations Catherine Taylor,*,† Adrian Allen,† Peter W. Dettmar,‡ and Jeffrey P. Pearson† Department of Physiological Sciences, The Medical School, University of Newcastle upon Tyne, Framlington Place, Newcastle upon Tyne, NE2 4HH U.K., and Reckitt Benckiser Healthcare (U.K.) Ltd., Dansom Lane, Hull HU8 7DS U.K. Received December 20, 2002; Revised Manuscript Received April 8, 2003
The gel nature of mucus is fundamental to its physiological functions; however, the structure of the mucus gel matrix is unclear. Here, small and large deformation rheology has been used to investigate the physical nature of the gel matrix and the forces that maintain this matrix in pig gastric mucus. The gelation process in mucus has been shown to be comparable with that of other polymer gel systems. Nongelling portions of mucin have been identified within the gel network, and the importance of transient, relaxable interactions to the maintenance of the mucus gel matrix has been demonstrated. The structure of the mucus gel matrix is considered in relation to the functional properties of mucus gels. Introduction
Experimental Section
Mucus secretions have many important biological functions in the digestive,1 respiratory,2 and reproductive3 tracts. Central to many of the functions of mucus is the gel forming ability of its major constituent molecules, the mucins. Gelforming mucins are a complex group of high molecular weight, polymeric glycoproteins based on four protein cores4 (coded for by the genes MUC2, MUC5AC, MUC5B, and MUC6) and a wide variety of carbohydrate side chain structures.5 However, the rheological similarities between mucus gels originating from different regions of the GI tract suggest that they have a common gel structure.6 A number of theories have been put forward to explain the process of gelation in mucus based on molecular level studies of mucin-mucin interactions;7-9 however, there is still no universally accepted model for the structure of the mucus gel matrix. In addition to polymeric mucin molecules the mucus gel also contains proteins, lipids and nucleic acids. The role of these additional components within the mucus gel is not fully understood, and although some studies suggest an integral role for other components within the mucus gel matrix,10 others suggest that the gel properties may result from mucins alone.11 Here we take a rheological approach to investigate the structure of a model mucus gel, pig gastric mucus, in the native, nonpurified state to reflect the complex nature of the mucus gel in vivo. We show that the rheological behavior of the mucus suggests it exists as an equilibrium gel and demonstrate the importance of both transient, relaxable, and nontransient interactions to the maintenance of the mucus gel matrix.
Sample Preparation. Pig stomachs obtained from a local abattoir were emptied immediately after slaughter then transported on ice to the laboratory. The mucosal surface was gently washed with running tap water to remove debris adhering to the mucus. Mucus gel was scraped from the mucosal surface of the gastric fundus using a glass microscope slide, particulate debris was removed, and it was stored at 4 °C and subjected to rheological testing within 48 h. Hyaluronic acid (potassium salt from human umbilical cord) was made up to 30 mg/mL in deionized water and allowed to hydrate overnight at 4 °C before rheological testing. Gelatin solutions were made up in deionized water (30 mg/mL) then heated to dissolve. The solution was then layered onto Petri dishes to a depth of 1 mm and allowed to cure for 1 h at 5 °C as this was found to give gels that were weak enough to breakdown without slippage on the rheometer plates. 25 mm diameter disks were punched from the gelatin sample for rheological testing. Oscillatory rheological measurements were carried out using a Bohlin CVO50 rheometer fitted with serrated 25 mm diameter parallel plates to eliminate gel slippage with the gap width set at 500 µm. The sample temperature was maintained at 25 °C by a thermostatically controlled circulating water bath. At all times, the gel sample was surrounded by a solvent trap filled with deionized water to minimize dehydration. Rheological Measurements. Frequency sweeps in the range of 0.01-3 Hz were carried out on samples of gastric mucus gel at target stress and strain values which lay well within the linear viscoelastic region, where the moduli values remained independent of the applied stress and induced strain. Typical target stress and strain values were set so that the induced strain was approximately 0.05 (5%). The gel samples were subjected to a repeated increase/ decrease amplitude (stress) sweep with log sampling and the
* To whom correspondence may be addressed. E-mail: Catherine.Taylor@ nt.ntnu.no. Current address: Institutt for Bioteknologi, Sem Sælands Vei 6/8, NTNU, Trondheim 7049, Norway. † University of Newcastle upon Tyne. ‡ Reckitt Benckiser Healthcare (UK) Ltd..
10.1021/bm025767t CCC: $25.00 © 2003 American Chemical Society Published on Web 05/30/2003
The Mucus Gel Matrix
Biomacromolecules, Vol. 4, No. 4, 2003 923
Figure 1. Frequency sweep behavior of a typical fundic gastric mucus gel at a target strain of 0.045 in the frequency range 3-0.01 Hz. The phase angle δ (b) remains essentially constant between 5 and 10 °. Both G′ and G′′ show a loss in magnitude with decreasing frequency. G′, 2; G′′, 9.
maximum applied stress being at least twice the breakdown stress. Breakdown was considered as the point at which G′′ became dominant over G′ (i.e., viscous behavior became dominant over elastic behavior). Breakdown and recovery stresses were quoted as the calculated stress at G′′ ) G′ for the upward and downward portion of the sweep, respectively. This point was calculated from the equation breakdown or recovery stress ) σa + {(σb - σa)[(45 - δa)/(δb - δa)]} where σa ) shear stress at first sample point below δ ) 45°, σb ) shear stress at first sample point above δ ) 45°, δa ) phase angle at first sample point below δ ) 45°, and δb ) phase angle at first sample point above δ ) 45° Results and Discussion Mucus Gel Rheology. All of the gastric mucus gels tested (n ) 10) showed highly reproducible behavior during the frequency sweep test with the phase angle (δ) lying between 5 and 10° (Figure 1) across the entire accessed frequency range. The magnitude of the elastic modulus G′ and the viscous modulus G′′ varied between gels with the range in G′ at 3 Hz being 189-648 Pa. In all cases, both moduli showed slight frequency dependence, dropping in value as the frequency was lowered (45% drop between 3 and 0.01 Hz; Figure 1). Unlike an entanglement network, the gel showed no increased tendency to flow at lower frequencies as demonstrated by the steady values for the phase angle δ (Figure 1). During the increase-decrease stress cycle, the gel flowed (G′′ became greater than G′ as a progressive change) rather than ruptured (which would produce gel fragments) in response to excess shear stress. The gastric mucus recovered gel properties (G′ > G′′) with a phase angle of 5-10° when the shear stress was reduced. The initial increase-decrease stress cycle caused a loss in the value of the moduli G′ and G′′ at minimum stress (Figure 2), but immediately repeating the cycle caused no further loss in the moduli at minimum stress. Rheological behavior did not change over 10 successive repeats of the increased-decrease stress cycle (all results not shown). The initial values of the moduli G′ and G′′ varied markedly
Figure 2. Behavior of the gastric mucus gel in response to repeated stress-induced breakdown at a frequency of 1 Hz (n ) 10). G′, solid line; G′′ lower dashed line; shear stress, upper dashed line.
between gels (G′ range 15-157 Pa at 1 Hz), but after the first breakdown and recovery stress cycle, there was good agreement between gels in the values of the moduli (G′ range 9-17 Pa at 1 Hz). As the shear stress is increased above that necessary for gel breakdown, the mucus flows as a viscous liquid. As the shear stress is reduced, the gel regains its solid dominant behavior G′ > G′′. After a large initial loss in the values of the moduli G′ and G′′ with the first stress induced breakdown, there is no further loss with subsequent breakdown/ recovery cycles. The initial high and varied values of G′ and G′′ may be an artifact of the gel loading procedure. It is normal practice to allow the test polymer to gel between the rheometer plates or to load a disk of gel of defined thickness for testing. This is not possible with mucus, and in practice, a lump of gel must be loaded which is then squeezed to the required thickness by lowering the rheometer plates. Reloading the mucus gel as a lump immediately after testing caused an increase in the moduli values. There was no significant difference between the mean stress at the point of breakdown (187 ( 14 Pa, mean (SEM) and the mean stress at the point of recovery (180 ( 19 Pa) (n ) 10) or the mean strain at the point of breakdown (15.5 ( 1.0) and the mean strain at the point of recovery (14.7 ( 0.7) (n ) 10). This behavior is compatible with the classification of the mucus gel as a physical gel maintained by noncovalent associations between the mucin chains as has been previously described.6,12 The category of physical gels has been subdivided into “weak” gels and “strong” gels based on the response of the gel to strain and its rheological reversibility.13 If the behavior of the gel is unaffected at strains of up to 25%, then it can be considered a strong gel, and in addition, strong gels would be expected to rupture rather than flow when exposed to excess strain, whereas weak gels demonstrate rheological reversibility. Although the mucus gel can maintain high strains without rupture, it demonstrates rheological reversibility and thus does not fit into either the weak or strong gel category and could therefore be considered an atypical biopolymer gel.
924
Biomacromolecules, Vol. 4, No. 4, 2003
Figure 3. Behavior of G′ and G′′ during the recovery (decrease stress) portion of the increase-decrease stress sweep at a frequency of 1 Hz shown for a single sample. G′, 2; G′′, 9; G′ increases steadily reaching a maximum at minimum stress. G′′ increases initially, passes through a maximum and then decreases as the stress is reduced.
When the shear stress applied to the gel was greater than the yield stress, its behavior was essentially that of mucus in the fluid state (G′′ > G′; Figure 3). As the applied shear stress was reduced, there was an increase in both G′ and G′′ (Figure 3). As the shear stress was reduced below the yield stress (the stress at which the gel collapsed), G′ becomes dominant over G′′ (Figure 3), which can be associated with gelation having occurred (moving into a regime of solid dominant behavior). This pattern is typical of a gelation event occurring and has been described for gels as diverse as heat set bovine serum albumin14 and synthetic polymethylsiloxane networks;15 however, in the case of mucus, the gel is “set” by a reduction in applied stress rather than alterations in temperature. The molecular events responsible for this characteristic pattern have been described based on spectroscopic studies of polymethylsiloxane networks. The behavior of G′ can be simply explained; it increases as more chains become involved in the gel network until it reaches a maximum plateau value when gelation is complete. G′′ represents the flow character of the gel and can be attributed to noninteracting or “dangling” chain ends in within the gel matrix. Each mucin molecule will have multiple binding sites involving a variety of interactions between protein and carbohydrate. These dangling ends represent noninteracting portions of the molecule within the gel at a given time. These portions of the polymer are free to move within the network at one end and hence do not exhibit the same elastic recoil as fully interacting molecules within the gel. When stress is applied to the gel, they move toward an entropically more favored position, i.e., one in which the molecule adopts a random rather than extended conformation. As the stress is reversed in the oscillatory test, the interacting, elastic portions of the chain will restore the energy spent on their deformation, but the free ends will only restore a portion of that energy as some has been used in the molecular rearrangement. These noninteracting chain ends will effectively “flow” in response to shear stress. Therefore, the rise and fall observed in the value of G′′ can be attributed to the presence of these dangling chain ends within the mucus matrix. G′′ initially rises as chains begin to interact with the network leading to dangling chain ends. As these chain ends form intermolecular interactions in the gelling system, they are no longer free to move and G′′ decreases. It has been
Taylor et al.
Figure 4. Behavior of G′ and G′′ during the breakdown (increasing stress) portion of the increase-decrease stress sweep at a frequency of 1 Hz shown for a single gel sample. G′, 2; G′′, 9. After an initial plateau, the value of G′ drops gradually with increasing stress. After a longer plateau, the value of G′′ rises past that of the decreasing G′.
suggested that the value of G′′ should become essentially zero when gelation is complete, as this would represent the maximum number of polymer interactions within the gel and therefore the maximum elastic response. This is not the case in the mucus gel where G′′ remains significantly nonzero. This is compatible with the equilibrium gel model in which in the native state the adherent gastric mucus gel matrix includes noninteracting, dangling chain ends, which retain the ability to interact. In an equilibrium gel at any given time, an equilibrium exists between molecules or regions of molecules involved in gelation and molecules or regions which are not involved in gelation but which have the potential to be involved. The presence of these noninteracting regions may allow the mucus to flow by sequential adjustment of the interacting and noninteracting chains in the direction of flow. These noninteracting regions could also impart the ability of the gel to reanneal, as there will be noninteracting regions on either side of the section line able to interact with each other and thus rejoin the gel. The gel shows no signs of rupture or catastrophic failure as the shear stress increases, with the gradual decrease in G′ (Figure 4) being indicative of structural interactions breaking consecutively rather than concurrently. This argues against a single interaction type maintaining the gel matrix. Effect of Frequency of Oscillation on Mucus Gel Behavior. Data obtained from the cyclical changes in stress magnitude were examined to determine the effect of altering the frequency of oscillation on the behavior of the elastic modulus G′. The behavior at a frequency of 1 Hz has previously been described, consisting of a plateau followed by a single gradual decrease in the value of the modulus (Figure 2 and 4). On decreasing the frequency of oscillation to 0.5 Hz, a second feature became apparent in the breakdown curve during increasing stress (Figure 5). After a short initial plateau, a slight decrease in magnitude occurs and then G′ increases, passes through a maximum, and then decreases rapidly as gel breakdown occurs (Figure 5). This demonstrates that as the shear stress is increased there is an increase in solid like behavior within the gel. This behavior is not seen at a frequency of 1 Hz but is seen at lower frequency values, e.g., 0.5 Hz and below. The effect of frequency of oscillation on gel rheology demonstrates that there are interactions within the gel matrix that are frequency
The Mucus Gel Matrix
Figure 5. Breakdown behavior of G′ at 0.5 Hz of gastric mucus. The behavior of the elastic modulus G′ over successive breakdown sweeps is plotted. G′, 2; shear stress, s. As the shear stress is increased, there is a slight decrease in G′ (A) before a rise to a maximum (B) followed by a rapid drop as gel breakdown occurs (C). The pattern is consistent over repeat sweeps.
Figure 6. Effect of frequency of oscillation on breakdown stress and strain in gastric mucus. Breakdown stress showed a general decrease with decreasing frequency with a pronounced change between 1 and 0.5 Hz. Breakdown strain showed a linear decrease with decreasing frequency across the frequency range (n ) 5).
sensitive, and so by examining the gel at different frequencies different regimes of behavior can be probed. The two-stage gel change in G′ seen at lower frequencies is direct evidence of at least two subsets of interactions within the mucus gel matrix with the second measured set of interactions (A-B, Figure 5) having greater elastic character than the initially measured interactions hence the increase in G′. The increase-decrease stress studies (five cycles per sweep) were repeated over a range of frequencies in the range of 1-0.1 Hz. At each frequency, the breakdown stress and strain were determined and plotted as a function of frequency. The strain at breakdown showed a gradual, linear decrease with decreasing frequency, with an excellent correlation between frequency and breakdown strain with a Pearson correlation value of 0.995 (Figure 6). The gel showed approximately a 50% decrease in breakdown strain over a decrease from 1 to 0.2 Hz (Figure 6). The stress at breakdown
Biomacromolecules, Vol. 4, No. 4, 2003 925
also decreased with decreasing frequency but not in a linear manner (Figure 6). The change in the stress/strain ratio at breakdown indicates there was a change in the dominant interactions within the gel at the point of gel breakdown as the frequency was lowered. The lowering of the breakdown stress with frequency is indicative of the initially dominant interactions within the gel being able to rearrange or relax at lower frequencies when the time scale of the oscillation is sufficiently slow. At a frequency of 1 Hz, breakdown is a onestage event as these initially measured interactions are the strongest subset of interactions within the gel; however, these interactions become weaker at lower frequencies and are broken at lower stresses exposing interactions within the gel that were previously masked, such as the more elastic interaction that leads to a peak in G′. To assess the contributions of these “transient” relaxable interactions to the mucus gel and to investigate the presence of any interactions which are nontransient under this experimental regime, the behavior of the gel at “zero frequency” was extrapolated. Plotting log[breakdown stress] against frequency allows the breakdown stress to be extrapolated to zero frequency (Figure 7). The frequency is the reciprocal of the relaxation time; therefore, at zero frequency, the relaxation time is infinite, and hence, the interactions remaining can be considered nontransient or “permanent” interactions within the gel under these test conditions. Extrapolating to zero frequency gives a breakdown stress of 10.9 Pa resulting from the “nontransient” interactions within the mucus gel. The analysis was validated by considering the behavior of a purely entangled system and a system that is essentially permanently cross-linked. Hyaluronate16 was chosen as the model entangled system and was subjected to an increasedecrease stress cycle (2 repeats) between 1 and 250 Pa at frequencies of 1, 0.5, 0.2, and 0.1 Hz. The breakdown stress of the hyaluronate was plotted against frequency and then linearized by plotting breakdown stress against log[frequency] to allow analysis of the behavior at zero frequency or infinite relaxation (Figure 7). The breakdown stress of the hyaluronate was reduced to zero even at a finite frequency of 0.1 Hz (Figure 7) indicating this frequency was sufficiently low to allow complete relaxation of all entanglements within the gel system. Gelatin17 was chosen as the permanently cross-linked system as the junction zones are formed by cooperative bonding and as such have a bond energy in excess of the available thermal energy and there is no evidence for entanglement in the gelatin network. As gelatin is not a rheologically reversible gel system, each gel sample was subjected to a single breakdown sweep in the stress range 1-250 Pa, with 4 gel samples being tested at frequencies of 1, 0.5, 0.2, and 0.1 Hz. The point of gel breakdown remained essentially constant across the frequency range (Figure 7), so no mathematical manipulation was needed to extrapolate behavior at zero frequency (Figure 7). This demonstrates that under this experimental regime the gelatin system is formed of essentially nontransient or permanent interactions. Considering the mucus gel, we see that the extrapolated breakdown stress for the mucus gel at zero frequency is 10.9
926
Biomacromolecules, Vol. 4, No. 4, 2003
Taylor et al.
regime show good agreement with previous studies that lead to the classification of mucus as a physical gel rather than an entangled system.6,12,18 The expansion of these studies to include large deformation regimes has, for the first time, allowed the rheological monitoring of the gel formation process in mucus in real time. In addition these studies have shown that the interactions within the mucus gel matrix can be subdivided based on their response to changing frequency, clearly demonstrating the presence of more than one interaction type within the gastric mucus gel matrix. The data presented here also indicate that gastric mucus exists as an equilibrium gel with at any given time an equilibrium existing between molecules or regions of molecules involved in gelation and molecules or regions which are not involved in gelation but which have the potential to be involved, which exist as “dangling chain ends” within the matrix. This gel model helps to explain the flow and reannealing properties of mucus that have previously been documented.19 Studies of mucus gel formation have focused on the molecular nature of mucin-mucin interactions in purified systems,7-9 and this study represents a novel approach to investigating these interactions in situ in the native mucus gel. The data presented here are concerned with the structure of the mucus gel matrix and the physical nature and time scale of the interactions that maintain it, in contrast to other studies that focus on the biochemical nature of the intermolecular associations within the mucus gel.7,8 Second, these data are representative of the multicomponent mucus gel and unlike the studies in purified systems do not distinguish which molecules are responsible for the interactions. Studying the mucus gel in the native state allows an understanding of the properties of the mucus gel as it exists in vivo and will allow further investigations as to which molecular components of the gel are necessary to fully replicate this behavior in purified systems. Combining such studies with investigations of interactions between these components on a molecular level and the effects of changes in the external environment such as pH should lead to a more complete understanding of the mucus gel.
Figure 7. Extrapolation of the breakdown stress of gastric mucus to zero frequency or infinite relaxation time allows analysis of the “nonrelaxable” interactions within the gel. These nonrelaxable interactions have a breakdown stress of 10.9Pa. Breakdown behavior with frequency of a model entangled gel system (hyaluronate) and a model “permanent” gel system (gelatin).
Pa compared to a breakdown stress of 203 Pa at 1 Hz. This is indicative of the mucus gel being a system where behavior is usually governed by transient, relaxable interactions; however, there are other interactions present within the gel matrix that do not relax under these experimental conditions and hence can be considered nontransient. Conclusions The study presented here is novel in that it uses both small and large deformation rheology to examine the architecture of a mucus gel. Results obtained in the small deformation
Acknowledgment. We thank Reckitt Benckiser Healthcare (UK) Ltd. and FMC Biopolymer a.s. for financial support for this study and Prof. Olav Smidsrød, NTNU Trondheim for helpful advice and discussion. References and Notes (1) Allen, A.; Pearson, J. P. Eur. J. Gastroenterol. Hepatol. 1993, 5, 193-199. (2) Puchelle, E.; Zahm, J. M. Biorheology 1986, 24, 146-150. (3) Chantler, E.; Elstein, M. Semin. Reprod. Endocrinol. 1986, 4, 333342. (4) Desseyn, J. L.; Aubert, J. P.; Porchet, N.; Laine, A. Mol. Biol. EVol. 2000, 17, 1175-1184. (5) Hounsell, E. F.; Feizi, T. Med. Biol. 1982, 60, 227-236. (6) Sellers, L. A.; Allen, A.; Morris, E. R.; Ross-Murphy, S. B. Biorheology 1987, 24, 615-623. (7) Cao, X.; Bansil, R.; Bhaskar, K. R.; Turner, B. S.; LaMont, J. T.; Niu, N.; Afdhal, N. H. Biophys. J. 1999, 76, 1250-1258. (8) Bromberg, L. E.; Barr, D. P. Biomacromolecules 2000, 1, 325334. (9) Berry, M.; McMaster, T. J.; Corfield, A. P.; Miles, M. J. Biomacromolecules 2001, 2, 498-503. (10) Thim, L.; Madsen, F.; Poulsen, S. S. Eur. J. Clin. InVest. 2002, 32, 519-527.
The Mucus Gel Matrix (11) Sellers, L. A.; Allen, A.; Morris, E. R.; Ross-Murphy, S. B. Biochim. Biophys. Acta 1991, 6, 174-179. (12) Bell, A. E.; Allen, A.; Morris, E.; Rees, D. A. AdV. Exp. Med. Biol. 1982, 144, 97-99. (13) Ross-Murphy, S. B.; Shatwell, K. P. Biorheology 1993, 30, 217225. (14) Kavanagh, G. M.; Ross-Murphy, S. B. Prog. Polym. Sci. 1998, 23, 533-562. (15) Bibbo, M. A.; Valles, E. M. Macromolecules 1984, 17, 360365.
Biomacromolecules, Vol. 4, No. 4, 2003 927 (16) Krause, W. E.; Bellomo, E. G.; Colby, R. H. Biomacromolecules 2001, 2, 65-69. (17) Lisetski, L. N.; Makarovskaya, Y. N.; Panikarsaya, V. D. Colloid Polym. Sci. 2001, 279, 283-285. (18) Sellers, L. A.; Allen, A.; Morris, E. R.; Ross-Murphy, S. B. Carbohydr. Res. 1988, 178, 93-110. (19) Bell, A. E.; Sellers, L. A.; Allen, A.; Cunliffe, W. J.; Morris, E. R.; Ross-Murphy, S. B. Gastroenterology 1985, 88, 269-280.
BM025767T
