Main

CF is the most common fatal autosomal recessive disease, affecting Caucasian populations, with an incidence between 1:2000 and 1:3000 live births(1). CF is caused by mutations in the gene encoding the CFTR(2). The protein product, CFTR, is an apical cAMP-regulated chloride channel considered important for proper hydration of epithelial secretions(3). Defective Cl- transport or insufficient hydration leads to dysfunction of multiple organs including sweat glands, respiratory tract, pancreas, and intestine(1). Ducts within certain exocrine glands become plugged with inspissated secretions, due to the precipitation of secreted macromolecules resulting from insufficient hydration(3,4).

The gastrointestinal tract is affected in CF beginning in utero. Meconium ileus, which manifests clinically at birth with inspissation of the terminal ileum, occurs in more than 15% of newborns with CF(5). Bowel obstruction after the newborn stage, termed "distal intestinal obstruction syndrome" or "meconium ileus equivalent," also involves acute and subacute obstruction of the terminal ileum and proximal colon with mucofeculent material(6,7). Although the pathogenesis of meconium ileus was previously attributed to poor digestion of intraluminal contents, due to a failure of exocrine pancreatic enzyme secretion, it is more likely due to accumulation of hyperviscous intestinal secretions within the crypts and in the intestinal lumen(1). The intestinal pathology in patients with CF occurs primarily in the ileum and proximal colon where the mucosal surface seems to be more susceptible to adherence of intestinal contents(1).

In CF, abnormal electrolyte transport in intestinal epithelia leads to changes in the composition of the luminal contents(1). In addition, there are biochemical alterations in the exocrine secretion of mucus(8). Both of these factors might alter surface properties of the intestinal epithelium, which, in turn, could influence adhesion patterns of macromolecules, bacteria, and microbial toxins to the intestinal surface. Hydrophobicity is a surface property that affects adhesion. Surface hydrophobicity can be determined through contact angle measurements(9). The principles of contact angles and surface hydrophobicity, as a measure of adhesion between macroscopic entities, is complex but not at all obscure. Underlying such efforts is the observation that adhesion of macroscopic entities, such as bacteria, is not exclusively governed (or cannot be readily described) in terms of molecular specificities. These relationships have been studied extensively in the physical sciences(9). For this reason, thermodynamic approaches seem to be appropriate for assessing biologic systems, such as adhesion patterns of mucosal surfaces. From the perspective of thermodynamics, the key quantity to describe adhesion is the free energy of adhesion. Free energy of adhesion can be expressed in terms of interfacial tensions between the interacting entities, which in turn, can be determined from contact angle data(9).

We hypothesize that altered surface hydrophobicity in CF intestinal epithelia may be a factor accounting for intestinal pathology in CF patients. We measured the contact angles of water droplets formed on the intestinal epithelia of a mouse model of CF(10). CF mouse models (Cftr-/-), show severe intestinal complications(10). A quarter of the Cftr-/- mice die at birth from complications closely resembling meconium ileus; most surviving mice die of intestinal obstruction shortly after being weaned to a solid diet(10). Recently, it was shown that introduction of a liquid diet to feed the Cftr-/- mice at weaning dramatically improves long term survival(11). The aim of this study was to use such a murine model of CF to compare the surface hydrophobicity of wild type and Cftr-/- gastrointestinal epithelia.

METHODS

Preparation of intestinal segments for contact angle measurements. A breeding colony of heterozygous CF mice was established using animals developed by Snouwaert et al.(10). This CF mouse model was produced by disrupting exon 10 of the murine Cftr gene. Cftr-/- mice and wild type mice (Cftr+/+) were bred in the animal facility at the Hospital for Sick Children, Toronto. Mice were weaned to a liquid diet (Liquidet, Bio-Serv, Frenchtown, NJ) to extend their lifespan(11). In brief, the liquid diet was fed to the animals in glass liquid mouse feeders suspended in microisolator cages. Corncob bedding was changed daily and both the feeders and the microisolators were sterilized. Fresh liquid diet and sterile water were provided daily. Mice were kept in a room with a 12-h light-dark cycle and ventilated at 20 air changes per hr with high efficiency particulate filtered air.

Mice were killed by cervical dislocation at 40 to 50 d of age. Preparation of the intestinal segments was adapted from a study of rabbits conducted by us previously(12). A midline abdominal incision was made to open the peritoneal cavity. The gastrointestinal tract, from the stomach to the distal colon, was removed intact, flushed with sterile saline at 4°C, and carefully trimmed of any adhering pancreatic and fat tissue to avoid irregularities in the tissues of interest. The tissue was then divided into seven segments (stomach, duodenum, upper jejunum, lower jejunum, ileum, proximal colon, and distal colon). Segments were opened along their longitudinal axis and spread onto dental was wrapped in aluminum foil, which was previously baked at 60°C between two glass plates to fix the foil onto the was. Because sessile water droplets fail to form on freshly isolated luminal surfaces, segments were air dried for approximately 30 min until the mucosa had a dull matted appearance, as described previously(12). Experimental protocols were approved by the Hospital for Sick Children Animal Care Committee.

Measurement of surface hydrophobicity with ADSA-CD. Surface hydrophobicity of mucosal epithelia was determined by measuring contact angles using the ADSA-CD technique(13,14). A water droplet applied to a hydrophobic surface has a relatively large contact angle, whereas a hydrophilic surface has a relatively small contact angle (Fig. 1). ADSA-CD uses the Laplace equation of capillarity to determine contact angles; liquid surface tension, drop volume, and contact diameter of a water droplet are input parameters. Sessile drops of distilled water were placed onto the mucosal surface. The precise drop volume was determined by using a micrometer syringe that delivers drops with an accuracy of 0.022 µL (Gilmont, Toronto). An image was taken from above the drop with the use of a stereomicroscope (M7S Zoom, Wild Heerbrugg, Germany). The contact angle of the drop was determined by computer digitization of the droplet periphery along with the input parameters. Surface tension of water droplets was determined by a different implementation of ADSA, called ADSA-P(9).

Figure 1
figure 1

Illustration of contact angles (θ). Contact angle of a water droplet on the surface in panel A is lower than that on the surface represented in panel B. The surface represented in panel A is less hydrophobic than the one in panel B.

Full size image

Statistics. Results were expressed as mean ± SEM. Comparisons between multiple groups were determined using one way analysis of variance (ANOVA) at 95% confidence intervals. Comparisons between two segments of intestine and between the same segments of Cftr+/+ and Cftr-/- intestine were performed using the unpaired two-tailed t test(15).

RESULTS

Eight Cftr+/+ and eight Cftr-/- mice were evaluated. Due to variability in the dimensions of each intestinal segment, a range of water droplets (11-70 drops) was formed on each section (Table 1). To evaluate consistency of contact angle values within each tissue segment of each animal, plots of contact angles from proximal to distal ends of the tissue were constructed. Contact angle measurements did not change within each segment (data not shown). Contact angles for each section were determined by the mean of the total droplets placed. As shown previously in rabbit(1) and ferret(16) gastrointestinal tracts, there was considerable variation in contact angle in different gastrointestinal segments of Cftr+/+ mice. In Cftr+/+ mice, no differences were observed in the contact angles between the duodenum (39.8° ± 1.2), upper jejunum (44.3° ± 2.4), lower jejunum (41.0° ± 1.7), and ileum (38.9° ± 2.3) [Fig. 2, p > 0.05]. Contact angles in the small intestine were higher than those obtained using the gastric mucosa (25.1° ± 1.7, p < 0.05). As in rabbit and ferret intestinal tissues, contact angles in both proximal (70.1° ± 2.2) and distal colon (53.8° ± 4.9) were higher than measurements obtained in the gastric mucosa (p < 0.05) and all small intestinal surfaces (p < 0.05). Contact angle of the proximal colon was greater than the distal colon (p < 0.05).

Table 1 Contact angles of water droplets formed on gastrointestinal segments of Cftr+/+ and Cftr-/- mice
Full size table
Figure 2
figure 2

Contact angles of water droplets formed on intestinal segments of wild type and cystic fibrosis mice. Results are expressed as mean ± SEM. For both Cftr+/+ and Cftr-/- mice, there were no differences in contact angles between the duodenum, upper jejunum, lower jejunum, and ileum (ANOVA, p > 0.05). Contact angles of the gastric mucosa in both Cftr+/+ and Cftr-/- mice were lower than those from the rest of the gastrointestinal tract (ANOVA, p < 0.05). Contact angles of the proximal colon and distal colon were both higher than those of the gastric mucosa (p < 0.05) and all regions of the small intestine (p < 0.05). The contact angles of the ileum and proximal colon from the Cftr-/- mice were higher than the corresponding sections of Cftr+/+ mice (*p < 0.05).

Full size image

Contact angles along the gastrointestinal tract of Cftr-/- mice followed the same general changes as those observed in each section from the Cftr+/+ mice (Fig. 2). Contact angles in the duodenum (40.0° ± 1.3), upper jejunum (39.7° ± 2.1), lower jejunum (38.0° ± 1.6), and ileum (44.4° ± 1.7) did not differ from each other (p > 0.05). As observed in the Cftr+/+ mice, mean values in the small intestinal sections were greater than that found in the gastric mucosa (30.7° ± 2.4) and lower than both the proximal colon (77.9° ± 1.9) and distal colon (57.5° ± 4.5) [p < 0.05]. Contact angle of the proximal colon was higher than that of the distal colon (p < 0.05), and both had significantly greater contact angles than those found in the gastric mucosa (p < 0.05) and all small intestinal segments (p < 0.05). Contact angles in each gastrointestinal section of the Cftr+/+ and Cftr-/- mice were compared. The ileum and proximal colon from the Cftr-/- mice each had higher contact angles than those determined in the corresponding sections of the Cftr+/+ animals (Fig. 2, p < 0.05).

DISCUSSION

In this study, we quantitated surface hydrophobicity of the intestinal epithelia using a CF mouse model by measuring contact angles of sessile water droplets placed onto mucosal surfaces. In the physical sciences, there is broad agreement about the principles of free energy of adhesion; however, several of the details are somewhat controversial. One of these concerns is the applicability to biologic measurements of one of the thermodynamic equilibrium conditions, the Young equation of capillarity. This equation interrelates contact angles and interfacial tensions and is therefore the point of departure of all such studies(9). Specifically, this relation assumes that the surface on which the contact angle is measured is smooth and homogeneous. Although it is well known from studies in the physical sciences that surfaces do not have to be completely smooth and of similar molecular homogeneity, concern remains that intestinal surfaces may not meet these assumptions completely. Surface roughness, which tends to increase contact angles, may lead to an overestimate of surface hydrophobicity(12). Because surface roughness was similar along the murine gastrointestinal tract, it cannot account for the differences observed in this study. At the present time, there are no means to test the applicability of the Young equation for biologic surfaces. For this reason, it was decided to use contact angles only as an empirical parameter. Thus, a difference in contact angle measurements between surfaces indicates differences in their adherence properties.

A specific limitation of the technique used in this study is that the need to dry the tissue may limit the biologic representation of gastrointestinal epithelia. Drying could affect cell viability, which could alter cellular secretions, cell surface components, and/or alter surface properties of the mucosa. Despite this, under identical preparatory conditions, we found regional differences in hydrophobicity of the gastrointestinal tract of both Cftr+/+ and Cftr-/- mice. Furthermore, the regional differences in surface hydrophobicity in the murine intestinal mucosa were similar to those observed in other species(12,16). The murine stomach was the least hydrophobic epithelial surface, whereas the colonic epithelium was significantly more hydrophobic than the rest of the gastrointestinal tract.

In terms of accuracy, ADSA-CD has been shown to outperform conventional techniques of contact angle determination(14,17). It has been successfully used to assess nonideal surfaces (i.e. rough and nonhomogeneous) including biologic epithelia. To improve accuracy, ADSA-CD requires an epithelial surface area large enough to allow the formation of a large number of water droplets with large volumes. Although drops on biologic surfaces are asymmetric, accuracy of measurement is possible by using the average contact diameter of the drop(13,14). The average contact diameter, which is used as input to the ADSA-CD routine, corresponds to a best-fit circle to the shape of the contact zone.

The main objective of this study was to compare gastrointestinal surface hydrophobicity between Cftr+/+ and Cftr-/- mice. The epithelial surfaces of the ileum and proximal colon of the Cftr-/- mice were more hydrophobic than the corresponding sections in Cftr+/+ mice. This observation, which reflects changes in adhesiveness, is of potential clinical significance because the ileum and proximal colon of Cftr+/+ mice(10) and the same regions of the CF intestinal tract in man(1), are prone to inspissation and obstruction with adherent mucofeculent material. Kent et al. demonstrated that Cftr-/- mice on the liquid diet had more severe intestinal pathology in the distal ileum and colon compared with the rest of the intestinal tract, even in the absence of obstruction. In Cftr-/- mice, the ileal and colonic crypts were deeper and stored more secreted mucin than those found in the Cftr+/+ mice. In addition, in the Cftr-/- mice, the accumulation of mucin in crypt lumina of the colon was even greater than that observed in the ileum(11).

Defective cAMP-dependent Cl- channel conductance in gastrointestinal epithelia(3) reduces fluid secretions(1) and may alter properties of the intestinal surface such as pH, surface charge(18), and mucus constituents, including changes in protein configuration. In CF, there are characteristic changes in mucus secretions(8,19,20). A study by Kent et al.(11) of Cftr-/- mice suggested a decrease in sialic acid and/or an increase in sulfation of their intestinal mucins. Wesley et al.(8) showed in humans, an increase in fucose, galactose, and N-acetyl glucosamine residues and a greater sulfation of CF intestinal mucins. In addition to greater mucin glycosylation, they showed lengthening of the polysaccharide chains in CF mucin and postulated that this would increase the hydrodynamic volume of mucin molecules thereby promoting greater viscosity and gelling. A study of mucus glycoproteins in the meconium from CF patients showed an increase in fucose and a decrease in sialic acid residues(19). Fucose has a hydrophobic C6 methyl group, and it was proposed that an increase in fucose would enhance hydrophobic interactions between mucin polysaccharide side chains, resulting in a greater gelling of mucin and possibly contributing to the pathogenesis of meconium ileus(19). An increase in mucin carbohydrate may contribute to the hyperviscosity observed in CF mucus secretions, leading to the mucus plugging of exocrine ducts(8). Mucin is the principal component of the mucus layer overlying gut surface epithelial cells(12). Therefore, differences in surface hydrophobicity between Cftr+/+ and Cftr-/- mice could be due to the observed changes in mucin characteristics associated with CF.

The colon was more hydrophobic than the rest of the gastrointestinal tract in both Cftr+/+ and Cftr-/- mice. It has been observed that mucins derived from the colon of a number of mammalian species demonstrate a greater content of hydrophobic amino acids compared with mucins of the small intestine(21). Although Wesley et al. found no change in amino acid profiles between normal and CF intestinal mucus compositions(8), Chace et al. reported that the amino acid composition of CF airway mucus was altered(22). Differences observed between the ileum and proximal colon of Cftr+/+ and Cftr-/- mice could be due to biochemical alterations of the mucins in these regions. Mucus hydrophobicity also seems to be related to its lipid constituents, particularly phospholipids, due to their amphipathic nature(23). The hydrophobic properties of lipids are postulated to form the basis of gastric mucosal and colonic mucosal hydrophobicities in many species. Lipid metabolism in CF intestinal mucosa may be altered as a consequence of the basic CF defect or as a result of a changed intestinal environment created by abnormalities in mucus secretion.

Bacterial adhesion is a central virulence factor in the initiation of infections(24). Bacterial binding has been described as a two-step process mediated first by nonspecific physicochemical properties (e.g. surface hydrophobicity) and then by specific stereochemical interactions (e.g. receptor-ligand)(25). In both Cftr+/+ and Cftr-/- mice, the colon was substantially more hydrophobic than the rest of the gastrointestinal tract, which could prove to be a factor contributing to the preference for certain enteric pathogens to colonize the colon(26). Furthermore, a change in epithelial cell surface hydrophobicity may either increase or decrease infectivity of certain pathogenic bacteria. For example, studies of Escherichia coli strain RDEC-1 (a rabbit enteropathogen) showed that binding of the organism to intestinal microvilli is dependent upon hydrophobic protein appendages, variously called pili or fimbriae, thereby promoting RDEC-1 adherence to hydrophobic surfaces(27).

Clinical studies have suggested, despite frequent and chronic use of antimicrobial agents for pulmonary diseases, there seems to be a reduced risk of antibiotic-associated diarrhea and colitis caused by an infection with Clostridium difficile(28). It has been suggested that several bacteria (e.g. Lactobacillus) with known inhibitory effects on C. difficile cytotoxin are found more frequently in the stool of CF patients(29). A change in surface hydrophobicity may influence the binding of bacterial cytotoxins such as C. difficile cytotoxins, and/or the colonization of certain bacteria that inhibit colonization of C. difficile. Additional factors, such as alterations of pH, protease (trypsin) activity, fat content, motility, etc., could affect the colonization, growth, and toxin production of various bacteria, including C. difficile(29).

In summary, we have demonstrated that the epithelial surfaces of the ileum and colon of Cftr-/- mice were more hydrophobic than the corresponding portions of the gastrointestinal tract in wild type mice. Intestinal pathology in CF is primarily in the ileum and colon. Thus, this observation is of potential clinical significance because the differences in hydrophobicity, which reflect differences in adhesion to macromolecules, were only found between CF and wild type ileal and colonic surfaces. Future studies are planned to confirm that alterations in the properties of intestinal mucus are responsible for the differences in surface hydrophobicity. A complementary technique, called ADSA-P, will be used to measure surface tension of intestinal mucins purified from various segments of Cftr+/+ and Cftr-/- murine intestines(9). Hydrophobicity can be characterized from surface tension since the two properties are inversely related. We will also compare the surface adherence properties of specific intestinal pathogens in CF and non-CF intestinal epithelia.