Villin as a structural marker to study the assembly of the brush border

Summary Brush borders which are localized at the apical face of enterocytes, are composed of thousands of stiff microvilli containing bundles of microfilaments made of actin. Their assembly occurs during terminal differentiation of the enterocytes when these cells migrate along the villus of the intestinal mucosa. The cell line HT 29 derived from a human colonic adenocarcinoma whose differentiation can be induced, can also be used as a model to study in culture the assembly of the intestinal brush border.Villin is one of the actin binding proteins found in microvilli which compose brush borders. Villin is expressed in the adult and in the embryo before the appearance of the brush border. Villin can be used as a tissue-specific marker for normal diffentiated and undifferentiated cells derived from gastrointestinal tractus in the adult as well as in the embryo. Since villin is a good marker for intestinal cells and plays a structural role in the assembly of the brush border we have analysed its expression and its localization in HT 29 cells. In HT 29 cells, as in the tissue, villin is synthesized at low levels before the appearance of the brush border. The high rate of synthesis and the recruitement of villin at the apical pole of the cells can be correlated with the existence of a well developed brush border.     

Cytoskeletal protein and mRNA accumulation during brush border formation in adult chicken enterocytes

We have explored the development of the brush border in adult chicken enterocytes by analyzing the cytoskeletal protein and mRNA levels as enterocytes arise from crypt stem cells and differentiate as they move toward the villus. At the base of the crypt, a small population of cells contain a rudimentary terminal web and a few short microvilli with long rootlets. These microvilli appear to arise from bundles of actin filaments which nucleate on the plasma membrane. The microvilli apparently elongate via the addition of membrane supplied by vesicles that fuse with the microvillus and extend the membrane around the actin core. Actin, villin, myosin, tropomyosin and spectrin, but not myosin I (previously called 110 kD; see Mooseker and Coleman, J. Cell Biol. 108, 2395-2400, 1989) are already concentrated in the luminal cytoplasm of crypt cells, as seen by immunofluorescence. Using quantitative densitometry of cDNA-hybridized RNA blots from cells isolated from crypts, villus middle (mid), or villus tip (tip), we found a 2- to 3-fold increase in villin, calmodulin and tropomyosin steady-state mRNA levels; an increase parallel to morphological brush border development. Actin, spectrin and myosin mRNA levels did not change significantly. ELISA of total crypt, mid and tip cell lysates show that there are no significant changes in actin, myosin, spectrin, tropomyosin, myosin I, villin or alpha-actinin protein levels as the brush border develops. The G-/F-actin ratio also did not change with brush border assembly. We conclude that, although the brush border is not fully assembled in immature enterocytes, the major cytoskeletal proteins are present in their full concentration and already localized within the apical cytoplasm. Therefore brush border formation may involve reorganization of a pool of existing cytoskeletal proteins mediated by the expression or regulation of an unidentified key protein(s).     

Posttranslational regulation of sucrase-isomaltase expression in intestinal crypt and villus cells

Expression and synthesis of sucrase-isomaltase (SI) were studied in human jejunum and in the colon tumor cell lines Caco-2 and HT-29. Twelve monoclonal antibodies produced against the adult human intestinal enzyme were shown to recognize specifically SI by immunoprecipitation of 14C-labeled membrane proteins, analysis of enzyme activities in the immunoprecipitates, and immunoblotting. These antibodies produced markedly different patterns of immunofluorescent staining of the intestinal mucosa. Three of them were specific for the absorptive villus cells, while the other nine also stained the luminal membrane of the proliferative crypt cells, with different intensities which paralleled their ability to recognize SI in immunoblots. Sequential immunoprecipitation of SI solubilized from purified brush borders or entire jejunum with four selected antibodies demonstrated the presence of different forms of the enzyme, expressed by either villus or crypt cells. Two immunologically distinct forms of high mannose precursor (hmP1 and hmP2) were also identified in both jejunal mucosa and colon tumor cells. They were present as monomers and their immunological differences were preserved under various ionic and pH conditions. Pulse-chase studies indicated that, in Caco-2 cells, hmP1 is converted into hmP2 within 30 min of chase, and hmP2 is then processed into the complex-glycosylated precursor destined for the brush border membrane. hmP1 was immunologically related to the mature SI present in crypt cells and lacked the epitopes specific for mature SI expressed by villus cells. These results demonstrated that sucrase-isomaltase is synthesized by both crypt and villus cells, but processing of the cotranslationally glycosylated high mannose precursor is dependent on the state of differentiation of the enterocytes. This may represent a general mechanism for the regulation of expression of differentiated cell products at the post-translational level.     

Alterations in the physical state and composition of brush border membrane lipids of rat enterocytes during differentiation

The physical state of the membrane lipid of brush border membranes, prepared from rat small intestinal villus and crypt cells, was examined by steady-state fluorescence polarization using three lipid-soluble fluorophors. Membranes prepared from crypt cells were found to possess a higher lipid fluidity than those of villus cells with each probe. Analysis of the composition of these membranes revealed that those from crypt cells had lower ratios of cholesterol/phospholipid (mol/mol), protein/lipid (w/w), and saturated fatty acyl chains/unsaturated chains (w/w). Alterations in the levels of stearic (18:0) and oleic (18:1) acids were responsible for differences in the latter ratio. The results, therefore, demonstrate that alterations in the lipid composition and fluidity of brush border membranes of enterocytes occur during the process of differentiation.     

The possible role of pancreatic proteases in the turnover of intestinal brush border proteins.

1. Intestinal brush border enzymes have heterogeneous rates of turnover, the largest proteins having the fastest turnover. Since the membrane faces the intestinal lumen, the effects of pancreatic factors were examined in mediating this turnover. Surgical subtotal pancreatectomy was used as an experimental model to study the turnover of brush border proteins in the absence of most pancreatic secretions. 2. Subtotal (95%) pancreatectomy of rats was found to cause elevations by about 50% of total activity and specific activities of certain brush border enzymes (maltase, sucrase, lactase), but not of others (alkaline phosphatase, trehalase). Rats were judged to be functionally deficient in pancreatic proteolytic enzymes (a) by demonstration of vitamin B-12 malabsorption, which was corrected by trypsin, and (b) by the finding of only about 20% of proteolytic activity appearing in the lumen after a test meal when compared to control. 3. To measure protein turnover in vivo the method of double labelling was used, where [3H]- and [14C]valine were administered intraduodenally in sequence 10 h apart. With this technique, a high 3H/14C ratio is correlated with rapid turnover. Proteins with apparent molecular weights of about 200 000-270 000 were found to turn over more rapidly than smaller proteins. 3H/14C ranged from 4.7 to 6.2 in animals without pancreatic insufficiency. In the face of decreased pancreatic proteolysis, the 3H/14C ratio was 2.3-3.1, similar to that of proteins with a slow half life. 4. Estimates of relative synthetic rates of large brush border proteins were lower than normal in pancreatectomized animals, but were constant over the period of the labelling experiment. The high enzyme levels in the face of lower synthetic rates confirms that, at the new steady rate, degradation rates must be slower for large brush border proteins in pancreatic insufficiency. 5. In vitro, using purified brush borders, unfractionated pancreatic enzymes were found to remove sucrase, maltase and lactase, but not alkaline phosphatase and trehalase. The enzyme most potent in this respect was the pancreatic protease, elastase. Non-proteolytic enzymes (amylase, lipase, phospholipase A) were inactive in removing enzyme from the brush border. The addition of elastase to pancreatectomized animals in vivo restored the rapid turnover rate of large brush border proteins. 6. A model is thus proposed for the normal catabolism of some large intestinal brush border proteins. It is suggested that the surface of intestinal absorptive cells is being constantly remodelled, and that certain surface enzymes are in part removed from the membrane by the action of pancreatic proteases. A possible special role for elastase is suggested.     

Intestinal calcium-binding protein

Summary The vitamin D-dependent calcium-binding protein (CaBP) was studied in relation to the age of the cell, in isolated epithelial cell populations removed from rat duodenum. Alkaline phosphatase and thymidine kinase activities were used as markers to characterize differentiated villus cells and undifferentiated (mitotically active) crypt cells, respectively. CaBP distribution along the length of the villus, as established by radioimmunoassay, appears as a gradient increasing from the crypt to the tip of the villus. CaBP concentration in cells is shown to be (i) negatively correlated with the thymidine kinase activity of cells, and (ii) positively correlated with the alkaline phosphatase activity of cells. This indicates that CaBP is absent in crypt cells and appears in differentiated cells with the development of the brush border. Thus CaBP, like alkaline phosphatase, can be considered as an indicator of enterocyte maturation. These data were also confirmed by studying the cellular localization of the protein. In addition both indirect immunofluorescence and immunoperoxidase staining methods reveal that antibody against CaBP decorates the terminal web, but not the microvilli of the brush border of mature absorptive cells. The results suggest that CaBP may act as a modulator of some Ca²⁺-mediated biochemical processes at the level of the enterocyte brush border.Portions of this work were presented at the Fourth International Workshop on Calcified Tissues, Israel (March 1980)     

Keratin expression in rat intestinal crypt and villus cells. Analysis with a panel of monoclonal antibodies.

Seven monoclonal antibodies were prepared against cytoskeletal components of rat intestinal brush borders. In the following paper (Chandler, J. S., Calnek, D., and Quaroni, A., J. Biol. Chem. 266, 11932-11938), three of them were shown to be specific for, respectively, keratin 8 (RK4), keratin 19 (RK7), and a newly identified type I keratin (keratin 21) (RK5). With these antibodies we have investigated the changes in keratin gene expression accompanying intestinal cell differentiation. Keratin 21 was detected exclusively in differentiated villus cells and in goblet, enteroendocrine, and Paneth cells in the crypts; in the proliferative crypt cells keratin 19 was predominant. Analysis of keratins expressed by cultured rat crypt cells (IEC cells) confirmed the absence of keratin 21 in undifferentiated intestinal cells. Changes in keratin's expression similar to those observed with cell differentiation in the adult intestinal mucosa were also demonstrated during early fetal intestinal development: the stratified epithelium present at 15-16 days of gestation contained predominantly keratin 19 with only a small amount of keratin 8; keratin 21 was first detected at 18-19 days of gestation, concomitant with the appearance of a well formed brush border and an apical cytoplasmic terminal web. These results suggest that keratin tonofilaments may play a role in the morphological and structural alterations accompanying intestinal cell differentiation in vivo.     

Mechanism of inhibition of proton: Dipeptide co-transport during chronic enteritis in the mammalian small intestine

Amino acids, a critical energy source for the intestinal epithelial cells, are more efficiently assimilated in the normal intestine via peptide co-transporters such as proton:dipeptide co-transport (such as PepT1). Active uptake of a non-hydrolyzable dipeptide (glycosarcosine) was used as a substrate and PepT1 was found to be present in normal villus, but not crypt cells. The mRNA for this transporter was also found in villus, but not crypt cells from the normal rabbit intestine. PepT1 was significantly reduced in villus cells also diminished in villus cell brush border membrane vesicles both from the chronically inflamed intestine. Kinetic studies demonstrated that the mechanism of inhibition of PepT1 during chronic enteritis was secondary to a decrease in the affinity of the co-transporter for the dipeptide without an alteration in the maximal rate of uptake (V_max). Northern blot studies also demonstrated unaltered steady state mRNA levels of this transporter in the chronically inflamed intestine. Proton dipeptide transport is found in normal intestinal villus cells and is inhibited during chronic intestinal inflammation. The mechanism of inhibition is secondary to altered affinity of the co-transporter for the dipeptide.     

Does heme oxygenase-1 have a role in Caco-2 cell cycle progression?

Intestinal epithelium undergoes a rapid self-renewal process characterized by the proliferation of the crypt cells, their differentiation into mature enterocytes as they migrate up to the villi, followed by their shedding as they become senescent villus enterocytes. The exact mechanism that regulates the intestinal epithelium renewal process is not well understood, but the differential expression of regulatory genes along the crypt-villus axis may have a role. Heme oxygenase-1 (HO-1) is involved in endothelial cell cycle progression, but its role in the intestinal epithelial cell turnover has not been explored. With its effects on cell proliferation and its differential expression along the crypt-villus axis, HO-1 may play a role in the intestinal epithelial cell renewal process. In this study, we examined the role of HO-1 in the proliferation and differentiation of Caco-2 cells, a well-established in vitro model for human enterocytes. After confluence, Caco-2 cells undergo spontaneous differentiation and mimic the crypt to villus maturation observed in vivo. In preconfluent and confluent Caco-2 cells, HO-1 protein expression was determined with the immunoblot. HO-1 activity was determined by the ability of the enzyme to generate bilirubin from hemin. The effect of a HO-1 enzyme activity inhibitor, tin protoporphyrin (SnPP), on Caco-2 cell proliferation and differentiation was examined. In preconfluent cells, cell number was determined periodically as a marker of proliferation. Cell viability was measured with MTT assay. Cell differentiation was assessed by the expression of a brush border enzyme, alkaline phophatase (ALP). HO-1 was expressed in subconfluent Caco-2 cells and remained detectable until 2 days postconfluency. This timing was consistent with cells starting their differentiation and taking the features of normal intestinal epithelial cells. HO-1 was inducible in confluent Caco-2 cells by the enzyme substrate, hemin in a dose- and time-dependent manner. SnPP decreased the cell number and viability of preconfluent cells and delayed the ALP enzyme activity of confluent cells. HO-1 may be involved in intestinal cell cycle progression.     

Down-regulation of reduced folate carrier may result in folate malabsorption across intestinal brush border membrane during experimental alcoholism

Folate plays a critical role in maintaining normal metabolic, energy, differentiation and growth status of all mammalian cells. The intestinal folate uptake is tightly and diversely regulated, and disturbances in folate homeostasis are observed in alcoholism, attributable, in part, to intestinal malabsorption of folate. The aim of this study was to delineate the regulatory mechanisms of folate transport in intestinal absorptive epithelia in order to obtain insights into folate malabsorption in a rat model of alcoholism. The rats were fed 1 g.kg(-1) body weight of ethanol daily for 3 months. A reduced uptake of [(3)H]folic acid in intestinal brush border membrane was observed over the course of ethanol administration for 3 months. Folate transport exhibited saturable kinetics and the decreased intestinal brush border membrane folate transport in chronic alcoholism was associated with an increased K(m) value and a low V(max) value. Importantly, the lower intestinal [(3)H]folic acid uptake in ethanol-fed rats was observed in all cell fractions corresponding to villus tip, mid-villus and crypt base. RT-PCR analysis for reduced folate carrier, the major folate transporter, revealed that reduced folate carrier mRNA levels were decreased in jejunal tissue derived from ethanol-fed rats. Parallel changes were observed in reduced folate carrier protein levels in brush border membrane along the entire crypt-villus axis. In addition, immunohistochemical staining for reduced folate carrier protein showed that, in alcoholic conditions, deranged reduced folate carrier localization was observed along the entire crypt-villus axis, with a more prominent effect in differentiating crypt base stem cells. These changes in functional activity of the membrane transport system were not caused by a general loss of intestinal architecture, and hence can be attributed to the specific effect of ethanol ingestion on the folate transport system. The low folate uptake activity observed in ethanol-fed rats was found to be associated with decreased serum and red blood cell folate levels, which might explain the observed jejunal genomic hypomethylation. These findings offer possible mechanistic insights into folate malabsorption during alcoholism.     

Changes in fatty acid composition during cell differentiation in the small intestine of suckling piglets.

Alterations of phospholipid fatty acid composition in the renewing intestine were studied in the infant piglet. Newborn piglets were fed from birth to 2 weeks of age a concentrated cow's milk which defined a standard supply of dietary fatty acids. Phospholipids were isolated from the whole mucosa, isolated intestinal cells and purified brush border membranes. Intestinal cells were isolated according to their position along the crypt-villus axis and cell phospholipids were extracted at each step of differentiation. Changes in fatty acid composition of cell phospholipids were related to those of lactase activity in the corresponding cell homogenates. In cell phospholipids, the relative content of linoleic and linoleic acids increased about 2-fold from crypt base to villus tip. Substantial contents of alkenylacyl glycerophospholipids (plasmalogens) were found in crypt cell phospholipids and in purified brush border membrane phosphatidylethanolamine (11 and 14% of alkenyl groups by weight of total fatty acids, respectively). The proportion of alkenylacyl glycerophospholipids decreased as cells ascended the villus column and became more differentiated. The results show that fatty acid compositional changes in differentiating cell phospholipids occurred in the immature intestine (before weaning) and suggest that these alterations might be related to the appearance of specific functions.     

An in vitro system to study interactions between bacteria and epithelial cells at the molecular level

This paper describes an experimental system to study interactions between porcine enterotoxigenic Escherichia coli (ETEC) and porcine intestinal epithelial cells in vitro at the molecular level. Radiolabelled bacteria or bacterial membrane fractions were incubated with brush borders prepared from purified epithelial cells, which were then washed repeatedly. The bacterial components removed by washing or retained by the brush borders were analysed to determine their composition and source. For this it was necessary to develop a minimal medium in which attachment factors of porcine ETEC could be radiolabelled. Furthermore, an improved method for the isolation of porcine intestinal epithelial cells was developed, since other procedures did not yield sufficiently pure preparations. The resulting method was rapid and yielded large quantities of viable epithelial cells, free from crypt cells and contaminating intestinal contents. Finally, we adapted existing procedures to isolate brush borders from these epithelial cells with special emphasis on the removal of nuclear and cytosolic material and on the isolation of morphologically intact brush borders. Using this system, mixtures of bacterial cytoplasmic and outer membranes were incubated with brush borders. Cytoplasmic membranes were easily removed by washing, while the outer membranes were not.     

Changes in the synthesis of ribosomal ribonucleic acid and of poly(A)-containing ribonucleic acid during the differentiation of intestinal epithelial cells in the rat and in the chick.

Epithelial cells were isolated from rat and chick small intestine by techniques which separated subpopulations of differentiating villus and upper crypt cells from each other and from populations of mitotically dividing lower crypt cells. Incorporation of precursors into epithelial-cell DNA, cytoplasmic rRNA and cytoplasmic poly(A)-containing RNA occurred in the lower crypt cells in vivo when precursor was supplied from the vascular system of the intestine. Incorporation of precursor into 28S and 18S rRNA continued in the upper crypt cells, but occurred to only a very slight extent (if at all) in villus cells, whereas incorporation into poly(A)-containing RNA continued (at a diminishing rate) as the differentiating cells migrated along the villi. When precursor was supplied from the intestinal lumen, its incorporation into DNA and into rRNA of crypt cells was not very different from that observed with the other mode of precursor administration, but incorporation into villus-cell poly(A)-containing RNA then occurred at essentially the same rate in all intestinal epithelial cells in vivo. Cytoplasmic poly(A)-containing RNA appeared to turn over in rat crypt cells with a half-life not exceeding 24 h; crypt-cell rRNA showed no turnover and no evidence could be found for the presence of 'metabolic DNA'.     

Immunolocalization of transferrin and transferrin receptor in mouse small intestinal absorptive cells

The mechanisms by which the duodenal mucosa absorbs iron are unknown. Insorption into absorptive cells of luminal iron bound to transferrin via receptor-mediated endocytosis has been hypothesized, but transferrin and transferrin receptor are absent in apical microvillous brush borders of small bowel biopsies taken from fasted patients and normal volunteers. We hypothesized that a normal iron-containing diet might induce the transient appearance of transferrin and transferrin receptor in apical brush borders of small intestinal absorptive cells in a normal mouse that was provided iron-containing chow until the moment of sacrifice. Light and electron microscopic immunolocalization of transferrin and transferrin receptor in proximal small intestinal absorptive cells was limited to basolateral membranes and coated pits of cells predominantly in the crypts and basal regions of the villi. Transferrin and transferrin receptor were not detected in apical microvillous brush border membranes of these enterocytes. In parallel immunolocalization protocols designed to show the ability to immunodetect other antigens at these locations, maltase and proteoglycan were demonstrated in apical microvillous brush border membranes and in basolateral membranes, respectively, in absorptive cells of small intestinal villous tip, base, and crypt regions. Furthermore, transferrin and transferrin receptor were immunolocalized in hepatocyte sinusoidal microvillus membranes. We conclude that food does not induce the appearance of immunodetectable transferrin and transferrin receptor in the apical microvilli of small intestinal absorptive cells and, therefore, that these iron transport proteins are not involved in the apical microvillous membrane transport of luminal dietary iron.     

Studies on the brush border membrane of the mouse duodenum

Summary Brush border membranes have been isolated from villus epithelial cells of the adult Swiss mouse duodenum. Preparations of these membranes are not contaminated by other organelles as judged from electron-micrographs of sectioned pellets of brush borders. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of proteins from brush borders solubilized in Tris-sodium dodecyl sulfate buffer reveals a reproducible Coomassie Brilliant Blue pattern of 17 bands. By comparing the brush border protein band positions with those of standard proteins run concurrently on sodium dodecyl sulfate-polyacrylamide gel slabs it is estimated that the 17 brush border proteins and subunits have molecular weights ranging from over 250,000 to around 16,000. Periodate-fuchsin sulfite staining shows that the five more slowly migrating, high molecular weight proteins are glycoproteins. The two proteins of smallest molecular size react positively with Oil Red O but have very small amounts of lipophilic amino acid residues, which indicates that the lipid extractable from the gels in these areas is a contaminant and is not bound to the proteins.     

Studies on the brush border membrane of the mouse duodenum. I. Membrane isolation and analysis of protein components.

Brush border membranes have been isolated from villus epithelial cells of the adult Swiss mouse duodenum. Preparations of these membranes are not contaminated by other organelles as judged from electron-micrographs of sectioned pellets of brush borders. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of proteins from brush borders solubilized in Tris-sodium dodecyl sulfate buffer reveals a reproducible Coomassie Brilliant Blue pattern of 17 bands. By comparing the brush border protein band positions with those of standard proteins run concurrently on sodium dodecyl sulfate-polyacrylamide gel slabs it is estimated that the 17 brush border proteins and subunits have molecular weights ranging from over 250,000 to around 16,000. Periodate-fuchsin sulfite staining shows that the five more slowly migrating, high molecular weight proteins are glycoproteins. The two proteins of smallest molecular size react positively with Oil Red O but have very small amounts of lipophilic amino acid residues, which indicates that the lipid extractable from the gels in these areas is a contaminant and is not bound to the proteins.     

Fetal gut mesenchyme induces differentiation of cultured intestinal endodermal and crypt cells

An experimental model was designed to analyze the effect of fetal gut mesenchyme on the cytodifferentiation of crypt cells and of embryonic progenitor cells. The cells used were the rat intestinal crypt cell line, IEC-17, and primary cell cultures prepared form isolated 14-day-old fetal intestinal endoderm (EC). Both cultures prepared from isolated 14-day-old fetal rat intestinal endoderm (EC). Both types of cells were associated with 14-day-old fetal rat gut mesenchyme (Rm) and grafted under the kidney capsule of adult rats. Seventy percent of the Rm/EC and ten percent of the Rm/IEC recombinants, recovered after 9 days, exhibited well-vascularized structures in which the mesenchyme had induced morphogenesis of the cells into a villus epithelium. The four main intestinal epithelial cell types, absorptive, goblet, endocrine, and Paneth cells, were identified using electron microscopy. Biochemical determinations of enzyme activities associated with brush border membranes revealed that alkaline phosphatase, lactase, sucrase, and maltase were expressed in both types of associations. These results were confirmed by immunofluorescence staining using monoclonal antibodies to brush border enzymes. Both enzyme assays and immunocytochemistry showed that the amount of enzymes present in the brush border membrane of Rm/IEC grafts was in general lower than that of the Rm/EC recombinants. The results indicate that fetal rat gut mesenchyme enables morphogenesis and cytodifferentiation of both crypt and embryonic progenitor cells.     

Na-glutamine co-transporters B0AT1 in villus and SN2 in crypts are differentially altered in chronically inflamed rabbit intestine

Glutamine is a major nutrient utilized by the intestinal epithelium and is primarily assimilated via Na-glutamine co-transport (NGcT) on the brush border membrane (BBM) of enterocytes. Recently we reported that B⁰AT1 (SLC6A19) mediates glutamine absorption in villus while SN2 (SLC38A5) does the same in crypt cells. However, how B⁰AT1 and SN2 are affected during intestinal inflammation is unknown. In the present study it was shown that during chronic enteritis NGcT was inhibited in villus cells, however, it was stimulated in crypt cells. Our studies also demonstrated that the mechanism of inhibition of NGcT during chronic enteritis was secondary to a reduction in the number of B⁰AT1 co-transporters in the villus cell BBM without a change in the affinity of the co-transporter. In contrast, stimulation of NGcT in crypt cells was secondary to an increase in the affinity of SN2 for glutamine without an alteration in the number of co-transporters. Thus, glutamine assimilation which occurs via distinct transporters in crypt and villus cells is altered in the chronically inflamed intestine.