The human intestinal epithelial cell line Caco-2; pharmacological and pharmacokinetic applications

Lipid peroxide-induced redox imbalance differentially mediates CaCo-2 cell proliferation and growth arrest

Dietary oxidants like lipid hydroperoxides (LOOH) can perturb cellular glutathione/glutathione disulphide (GSH/GSSG) status and disrupt mucosal turnover. This study examines the effect of LOOH on GSH/GSSG balance and phase transitions in the human colon cancer CaCo-2 cell. LOOH at 1 or 5 micro m were noncytotoxic, but disrupted cellular GSH/GSSG and stimulated proliferative activity at 6 h that paralleled increases in ornithine decarboxylase activity, thymidine incorporation, expression of cyclin D1/cyclin-dependent kinase 4, phosphorylation of retinoblastoma protein, and cell progression from G0/G1 to S. At 24 h, LOOH-induced sustained GSH/GSSG imbalance mediated growth arrest at G0/G1 that correlated with suppression of proliferative activity and enhanced oxidative DNA damage. LOOH-induced cell transitions were effectively blocked by N-acetylcysteine. Collectively, the study shows that subtoxic LOOH levels induce CaCo-2 GSH/GSSG imbalance that elicits time-dependent cell proliferation followed by growth arrest. These results provide insights into the mechanism of hydroperoxide-induced disruption of mucosal turnover with implications for understanding oxidant-mediated genesis of gut pathology.

The beta 1 isoform of protein kinase C mediates the protective effects of epidermal growth factor on the dynamic assembly of F-actin cytoskeleton and normalization of calcium homeostasis in human colonic cells

Using intestinal monolayers, we showed that F-actin cytoskeletal stabilization and Ca(2+) normalization contribute to epidermal growth factor (EGF)-mediated protection against oxidant injury. However, the intracellular mediator responsible for these protective effects remains unknown. Since the protein kinase C-beta1 (PKC-beta1) isoform is abundant in our naive (N) cells, we hypothesized that PKC-beta1 is essential to EGF protection. Monolayers of N Caco-2 cells were exposed to H(2)O(2) +/- EGF, PKC, or Ca(2+) modulators. Other cells were transfected to over-express PKC-beta1 or to inhibit its expression and then pretreated with low or high doses of EGF or a PKC activator, OAG (1-oleoyl-2-acetyl-sn-glycerol), before H(2)O(2). In N monolayers exposed to oxidant, pretreatment with EGF or PKC activators activated PKC-beta1, enhanced (45)Ca(2+) efflux, normalized Ca(2+), decreased monomeric G-actin, increased stable F-actin, and protected the cytoarchitecture of the actin. PKC inhibitors prevented these protective effects. Transfected cells stably over-expressing PKC-beta1 (+3.1-fold) but not N cell monolayers were protected from injury by even lower doses of EGF or OAG. EGF or OAG rapidly activated the over-expressed PKC-beta1. Antisense inhibition of PKC-beta1 expression (-90%) prevented all measures of EGF protection. Inhibitors of Ca(2+)-ATPase prevented EGF protection in N cells as well as protective synergism in transfected cells. EGF protects the assembly of the F-actin cytoskeleton in intestinal monolayers against oxidants in large part through the activation of PKC-beta1. EGF normalizes Ca(2+) by enhancing Ca(2+) efflux through PKC-beta1. We have identified novel biologic functions, protection of actin and Ca(2+) homeostasis, among the classical isoforms of PKC.

PKC-zeta is required in EGF protection of microtubules and intestinal barrier integrity against oxidant injury

Using monolayers of human intestinal (Caco-2) cells, we showed that epidermal growth factor (EGF) protects intestinal barrier integrity against oxidant injury by protecting the microtubules and that protein kinase C (PKC) is required. Because atypical PKC-zeta isoform is abundant in wild-type (WT) Caco-2 cells, we hypothesized that PKC-zeta mediates, at least in part, EGF protection. Intestinal cells (Caco-2 or HT-29) were transfected to stably over- or underexpress PKC-zeta. These clones were preincubated with low or high doses of EGF or a PKC activator [1-oleoyl-2-acetyl-sn-glycerol (OAG)] before oxidant (0.5 mM H(2)O(2)). Relative to WT cells exposed to oxidant, only monolayers of transfected cells overexpressing PKC-zeta (2.9-fold) were protected against oxidant injury as indicated by increases in polymerized tubulin and decreases in monomeric tubulin, enhancement of architectural stability of the microtubule cytoskeleton, and increases in monolayer barrier integrity toward control levels (62% less leakiness). Overexpression-induced protection was OAG independent and even EGF independent, but EGF significantly potentiated PKC-zeta protection. Most overexpressed PKC-zeta (92%) resided in membrane and cytoskeletal fractions, indicating constitutive activation of PKC-zeta. Stably inhibiting PKC-zeta expression (95%) with antisense transfection substantially attenuated EGF protection as demonstrated by reduced tubulin assembly and increased microtubule disassembly, disruption of the microtubule cytoskeleton, and loss of monolayer barrier integrity. We conclude that 1) activation of PKC-zeta is necessary for EGF-induced protection, 2) PKC-zeta appears to be an endogenous stabilizer of the microtubule cytoskeleton and of intestinal barrier function against oxidative injury, and 3) we have identified a novel biological function (protection) among the atypical isoforms of PKC.

The L-3,4-dihydroxyphenylalanine transporter in human and rat epithelial intestinal cells is a type 2 hetero amino acid exchanger

Information on the intestinal transport of L-3,4-dihydroxyphenylalanine (L-DOPA) is scarce. We present here the functional characteristics and regulation of the apical inward L-DOPA transport in two intestinal epithelial cell lines (human Caco-2 and rat IEC-6). The inward transfer of L-DOPA and L-leucine was promoted through an energy-driven system but with different sensitivity to extracellular Na(+) concentration: a minor component of L-leucine uptake (approximately 25%) was found to require extracellular Na(+) in comparison with L-DOPA transport which was Na(+)-independent. L-DOPA and L-leucine uptake was insensitive to N-(methylamino)-isobutyric acid, but competitively inhibited by 2-aminobicyclo(2,2,1)-heptane-2-carboxylic acid (BCH). L- and D-neutral amino acids, but not acidic and basic amino acids, markedly inhibited L-DOPA and [(14)C]L-leucine accumulation in both cell lines. The [(14)C]L-DOPA and [14C]L-leucine outward were markedly increased by L-leucine and BCH present in extracellular medium, but not by L-arginine. In both cell lines, L-DOPA transport was stimulated by acidic pH in comparison with [(14)C]L-leucine inward which was pH-independent. In conclusion, it is likely that system B(0) might be responsible for the Na(+)-dependent uptake of L-leucine in Caco-2 and IEC-6 cells, whereas sodium-independent uptake of L-leucine and L-DOPA may include system type 1 and type 2 L-amino acid transporter (LAT1 and LAT2), the activation of which results in trans-stimulation of substrates outward transfer.

Identification of a novel route of extraction of sirolimus in human small intestine: roles of metabolism and secretion

Using Caco-2 cell monolayers expressing CYP3A4, we investigated the interplay between metabolism and transport on the first-pass intestinal extraction of the immunosuppressant sirolimus, a CYP3A4/P-glycoprotein (P-gp) substrate. Modified Caco-2 cells metabolized [(14)C]sirolimus to the predicted amounts of CYP3A4-mediated products based on CYP3A4 content, which was approximately 20% of that measured in human small intestinal mucosal homogenate. [(14)C]Sirolimus also degraded to the known ring-opened product, seco-rapamycin. Unexpectedly, a ring-opened dihydro metabolite (M2) was the major product detected in cells at all sirolimus concentrations examined (2-100 microM). Greater M2 formation after apical versus basolateral dosing (1.6-fold) was explained by higher intracellular content of sirolimus after apical dosing. M2 was not detected in incubations with human liver and intestinal microsomes but was readily detected with corresponding homogenates. M2 formation was NADPH-dependent but unaffected by the CYP3A4 inhibitors ketoconazole and troleandomycin. Although M2 was formed from purified seco-rapamycin (20 microM) in the homogenates, it was not detected in cells when seco-rapamycin was added to the apical compartment, because seco-rapamycin was essentially impermeable to the apical membrane. Sirolimus, seco-rapamycin (basolaterally dosed), and M2 were all actively secreted across the apical membrane, and secretion of each was inhibited by the P-gp inhibitor LY335979 [(2R)-anti-5-[3-[4-(10,11-difluoromethanodibenzo-suber-5-yl)piperazin-1-yl]-2-hydroxypropoxy]quinoline trihydrochloride]. Along with CYP3A4-mediated metabolism and P-gp-mediated secretion, we conclude that the following novel pathway, which occurs at least in the intestine, may contribute significantly to the first-pass extraction of sirolimus in humans: intracellular degradation of sirolimus to seco-rapamycin, metabolism of seco-rapamycin to M2 by an unidentified non-microsomal enzyme, and P-gp-mediated secretion of M2 and seco-rapamycin.

L-type amino acid transporters in two intestinal epithelial cell lines function as exchangers with neutral amino acids

The present study examined the functional characteristics of the inward [(14)C]-L-leucine transporter in two intestinal epithelial cell lines (human Caco-2 and rat IEC-6). The uptake of [(14)C]-L-leucine was largely promoted through an energy-dependent and sodium-insensitive transporter, although a minor component of [(14)C]-L-leucine uptake ( approximately 15%) required extracellular sodium. [(14)C] -L-leucine uptake was insensitive to N-(methylamino)-isobutyric acid, but competitively inhibited by 2-aminobicyclo(2,2,1)-heptane-2-carboxylic acid (BCH). Both L- and D-neutral amino acids, but not acidic and basic amino acids, markedly inhibited [(14)C]-L-leucine accumulation. The efflux of [(14)C]-L-leucine was markedly increased (P < 0.05) by L-leucine and BCH, but not by L-arginine. In IEC-6 cells, but not in Caco-2 cells, the uptake of [(14)C]-L-leucine at acidic pH (5.0 and 5.4) was greater (P < 0.05) than at pH 7.4. In conclusion, it is likely that system B(0) might be responsible for the sodium-dependent uptake of L-leucine in Caco-2 and IEC-6 cells, whereas sodium-independent uptake of L-leucine may include system LAT1, whose activation results in transstimulation of L-leucine outward transfer.

Receptor-dependent transcriptional activation of cytochrome P4503A genes: induction mechanisms, species differences and interindividual variation in man

1. The importance of CYP3A enzymes in drug metabolism and toxicology has yielded a wealth of information on the structure, function and regulation of this subfamily and recent research emphasis has been placed on the human forms, namely CYP3A4, CYP3A5, CYP3A7 and CYP3A43. 2. The current review will focus on the receptor-dependency of CYP3A regulation and includes consideration of the regulatory roles of the glucocorticoid (GR), pregnane X (PXR) and constitutive androstane (CAR) receptors. 3. Emphasis has been placed on the topics of expression and substrate specificity, assessment of induction, species differences in induction, CYP3A promoter sequences and regulation of gene expression, structural and functional aspects of receptor-mediated, CYP3A gene activation, receptor variants and interindividual variation in human CYP3A expression, the latter encompassing environmental, physiological and genetic aspects. 4. An outline of future research needs will be discussed in the context of receptor-mediated molecular mechanisms of CYP3A gene regulation and the impact on interindividual variations in CYP3A expression. 5. Taken collectively, this review highlights the importance of understanding the molecular mechanisms of CYP3A induction as a means of rationalizing human responses to many clinically used drugs, in addition to providing a mechanistically coherent platform to understand and predict interindividual variations in response and drug-drug interactions.

Relevance of multidrug resistance proteins for intestinal drug absorption in vitro and in vivo

Multidrug resistance proteins (p-glycoprotein and mrps) are becoming increasingly important to explain the pharmacokinetics and action of drugs. Located in epithelial and endothelial cells of the gastrointestinal tract, liver, kidney, blood brain barrier, choroid plexus and other organs, they are critical determinants for the movement of a large number of commonly prescribed drugs across cellular barriers. Here we provide a brief overview of the role of multidrug resistance proteins in drug absorption from the gastrointestinal tract. We address the different types of multidrug resistance proteins involved, describe experimental models to study the influence of these proteins on transcellular transport and discuss the impact of multidrug resistance proteins on overall drug bioavailability in vivo.

Enhanced expression of the human multidrug resistance protein 3 by bile salt in human enterocytes. A transcriptional control of a plausible bile acid transporter

The enterohepatic circulation is essential for the maintenance of bile acids and cholesterol homeostasis. The ileal bile acid transporter on the apical membrane of enterocytes mediates the intestinal uptake of bile salts, but little is known about the bile salt secretion from the basolateral membrane of enterocytes into blood. In the basolateral membrane of enterocytes, an ATP-binding cassette transporter, multidrug resistance protein 3 (MRP3), is expressed, which has the ability to transport bile salts. We hypothesized that MRP3 might play a role in the enterohepatic circulation of bile salts by transporting them from enterocytes into circulating blood through the up-regulation of MRP3 expression, so we investigated the transcriptional control of MRP3 in response to bile salts. MRP3 mRNA levels were increased about 3-fold in human colon cells by chenodeoxycholic acid (CDCA), in a dose- and time-dependent manner. In the promoter assay, the promoter activity of MRP3 was increased about 3-fold over the basal promoter activity when treated with CDCA, and the putative bile salt-responsive elements exist in the region -229/-138 including two alpha-1 fetoprotein transcription factor (FTF)-like elements. Constructs with a specific mutation in the consensus sequence of FTF elements showed no increase in basal transcriptional activity following CDCA treatment. In electrophoretic mobility shift assay with nuclear extracts, specific binding of FTF to FTF-like elements was observed when treated with CDCA. The expression of FTF mRNA levels were also markedly enhanced in response to CDCA, and overexpression of FTF specifically activated the MRP3 promoter activity about 4-fold over the basal promoter activity. FTF thus might play a key role not only in the bile salt synthetic pathway in hepatocytes but also in the bile salt excretion pathway in enterocytes through the regulation of MRP3 expression. MRP3 may contribute as a plausible bile salt-exporting transporter to the enterohepatic circulation of bile salts.

Role of multidrug resistance protein 2 (MRP2) in glutathione-bimane efflux from Caco-2 and rat renal proximal tubule cells

1. The multidrug resistance protein 2 (MRP2) has been shown to play an important role in the transport of glutathione conjugates in the liver. Its importance in renal excretion, however, is still uncertain and other organic anion transporters may be involved. The objective of the present study was to characterize glutathione conjugate efflux from rat kidney proximal tubule cells (PTC), and to determine the contribution of Mrp2. 2. We used isolated PTC in suspension, as well as grown to monolayer density. For comparison, transport characteristics were also determined in the human intestinal epithelial cell line Caco-2, an established model to study MRP2-mediated transport. The cells were loaded with monochlorobimane (MCB) at 10 degrees C. MCB enters the cells by simple diffusion and is conjugated with glutathione to form the fluorescent glutathione-bimane (GS-B). 3. In primary cultures of rat PTC, no indications for a transporter-mediated mechanism were found. The efflux of GS-B from Caco-2 cells and freshly isolated PTC was time- and temperature-dependent. Furthermore, GS-B transport in both models was inhibited by chlorodinitrobenzene (CDNB), with an inhibitory constant of 46.8+/-0.9 microM in freshly isolated PTC. In Caco-2 cells, the inhibitory potency of CDNB was approximately 20 fold higher. Finally, efflux of GS-B from freshly isolated PTC from Mrp2-deficient (TR(-)) rats was studied. As compared to normal rat PTC, transport characteristics were not different. 4. We conclude that in freshly isolated rat PTC glutathione conjugate excretion is mediated by other organic anion transporters rather than by Mrp2.

PKC-beta1 mediates EGF protection of microtubules and barrier of intestinal monolayers against oxidants

Using monolayers of human intestinal (Caco-2) cells, we found that oxidants and ethanol damage the cytoskeleton and disrupt barrier integrity; epidermal growth factor (EGF) prevents damage by enhancement of protein kinase C (PKC) activity and translocation of the PKC-beta1 isoform. To see if PKC-beta1 mediates EGF protection, cells were transfected to stably over- or underexpress PKC-beta1. Transfected monolayers were preincubated with low or high doses of EGF (1 or 10 ng/ml) or 1-oleoyl-2-acetyl-sn-glycerol [OAG; a PKC activator (0.01 or 50 microM)] before treatment with oxidant (0.5 mM H(2)O(2)). Only in monolayers overexpressing PKC-beta1 (3.1-fold) did low doses of EGF or OAG initiate protection, increase tubulin polymerization (assessed by quantitative immunoblotting) and microtubule architectural integrity (laser scanning confocal microscopy), maintain normal barrier permeability (fluorescein sulfonic acid clearance), and cause redistribution of PKC-beta1 from cytosolic pools into membrane and/or cytoskeletal fractions (assessed by immunoblotting), thus indicating PKC-beta1 activation. Antisense inhibition of PKC-beta1 expression (-90%) prevented these changes and abolished EGF protection. We conclude that EGF protection against oxidants requires PKC-beta1 isoform activation. This mechanism may be useful for development of novel therapies for the treatment of inflammatory gastrointestinal disorders including inflammatory bowel disease.

Phospholipase C-gamma inhibition prevents EGF protection of intestinal cytoskeleton and barrier against oxidants

Loss of intestinal barrier integrity is associated with oxidative inflammatory GI disorders including inflammatory bowel disease. Using monolayers of human intestinal epithelial (Caco-2) cells, we recently reported that epidermal growth factor (EGF) protects barrier integrity against oxidants by stabilizing the microtubule cytoskeleton, but the mechanism downstream of the EGF receptor (EGFR) is not established. We hypothesized that phospholipase C (PLC)-gamma is required. Caco-2 monolayers were exposed to oxidant (H2O2) with or without pretreatment with EGF or specific inhibitors of EGFR tyrosine kinase (AG-1478, tyrphostin 25) or of PLC (L-108, U-73122). Other Caco-2 cells were stably transfected with a dominant negative fragment for PLC-gamma (PLCz) to inhibit PLC-gamma activation. Doses of EGF that enhanced PLC activity also protected monolayers against oxidant-induced tubulin disassembly, disruption of the microtubule cytoskeleton, and barrier leakiness as assessed by radioimmunoassay, quantitative Western blots, high-resolution laser confocal microscopy, and fluorometry, respectively. Pretreatment with either type of inhibitor abolished EGF protection. Transfected cells also lost EGF protection and showed reduced PLC-gamma phosphorylation and activity. We conclude that EGF protection requires PLC-gamma signaling and that PLC-gamma may be a useful therapeutic target.

Key role of PKC and Ca2+ in EGF protection of microtubules and intestinal barrier against oxidants

Using monolayers of human intestinal (Caco-2) cells, we showed that growth factors (GFs) protect microtubules and barrier integrity against oxidative injury. Studies in nongastrointestinal cell models suggest that protein kinase C (PKC) signaling is key in GF-induced effects and that cytosolic calcium concentration ([Ca2+](i)) is essential in cell integrity. We hypothesized that GF protection involves activating PKC and maintaining normal ([Ca2+](i)) Monolayers were pretreated with epidermal growth factor (EGF) or PKC or Ca2+ modulators before exposure to oxidants (H2O2 or HOCl). Oxidants disrupted microtubules and barrier integrity, and EGF protected from this damage. EGF caused rapid distribution of PKC-alpha, PKC-betaI, and PKC-zeta isoforms to cell membranes, enhancing PKC activity of membrane fractions while reducing PKC activity of cytosolic fractions. EGF enhanced (45)Ca2+ efflux and prevented oxidant-induced (sustained) rises in ([Ca2+](i)). PKC inhibitors abolished and PKC activators mimicked EGF protection. Oxidant damage was mimicked by and potentiated by a Ca2+ ionophore (A-23187), exacerbated by high-Ca2+ media, and prevented by calcium removal or chelation or by Ca2+ channel antagonists. PKC activators mimicked EGF on both (45)Ca2+ efflux and ([Ca2+](i)). Membrane Ca2+-ATPase pump inhibitors prevented protection by EGF or PKC activators. In conclusion, EGF protection of microtubules and the intestinal epithelial barrier requires activation of PKC signal transduction and normalization of ([Ca2+](i)).

Role of pre-existing redox profile of human macrophages on lipid synthesis and cholesteryl ester cycle in presence of native, acetylated and oxidised low density lipoprotein

The importance of the interactions of modified lipids and macrophages in foam cell generation is clear; however, little attention has been paid to the role of intra-macrophagic redox potential as a modulator of their lipid synthesis and metabolism. In this study, the effects of previously induced non-toxic manipulations of intracellular redox balance on lipid synthesis in human monocyte-derived macrophages (HMDM) was evaluated. Cells, pre-treated with 2.5 microM of the pro-oxidising agent CuSO(4) or with 5 mM of the antioxidant and thiol supplier N-acetylcysteine (NAC), were exposed to radiolabelled oleic acid alone or in combination with native low density lipoprotein (LDL) or modified LDL to evaluate the incorporation of radioactivity into cholesteryl ester, triacylglycerols and phospholipids. CuSO(4)-treated macrophages synthesised more lipids than NAC-treated cells in absence of exogenous lipid, and, generally, in the presence of native or acetylated, but oxidised LDL. In addition, the activities of the enzymes involved in cholesteryl ester storage were also influenced by the pro-oxidant condition. The ratio values between acyl-coenzyme A:cholesterol acyl transferase and cholesteryl ester hydrolase activity suggest that in CuSO(4)-treated macrophages the hydrolysis of cholesteryl ester is favoured with respect to esterification. The interaction of HMDM with oxidised LDL showed a significant different pattern in term of lipid synthesis with respect to those induced by native or acetylated LDL, disrespectful of the initial redox profile of the cells. On the whole, these results suggest that the pre-existing internal redox condition is a further parameter able to modulate the effects of native or acetylated LDL-cell interaction, influencing both HMDM lipid synthesis profile and cholesterol storage. Moreover, oxidised LDL represent a carrier of additional factor(s) able per se to introduce perturbation in the synthetic pathway of lipids, which is not influenced by the redox potential of the macrophage.

Charge-dependent interaction of self-emulsifying oil formulations with Caco-2 cells monolayers: binding, effects on barrier function and cytotoxicity

A positively charged self-emulsifying oil formulation (SEOF), aimed to enhance oral bioavailability of drugs poorly soluble in water, was recently developed. In the present study the Caco-2 cell model was used for the investigation of the charge-dependent interactions of this SEOF with human intestinal epithelial cells. The positively charged emulsions affected the barrier properties of the cell monolayer at high concentrations and reduced the cell viability. However, at the dilution with aqueous phase used in the present study (1:2000), the positively charged SEOF did not induce any detectable cytotoxic effect. The binding of the fluorescent dye DiIC(18)(3) was much higher from the positively charged SEOF, compared to the negatively charged formulation, suggesting an increased closer adhesion of the droplets to the cell surface due to the electrostatic attraction. No transepithelial transport of this compound across Caco-2 cell monolayers was observed with any SEOF formulation.

Prediction of drug absorption using multivariate statistics

Literature data on compounds both well- and poorly-absorbed in humans were used to build a statistical pattern recognition model of passive intestinal absorption. Robust outlier detection was utilized to analyze the well-absorbed compounds, some of which were intermingled with the poorly-absorbed compounds in the model space. Outliers were identified as being actively transported. The descriptors chosen for inclusion in the model were PSA and AlogP98, based on consideration of the physical processes involved in membrane permeability and the interrelationships and redundancies between available descriptors. These descriptors are quite straightforward for a medicinal chemist to interpret, enhancing the utility of the model. Molecular weight, while often used in passive absorption models, was shown to be superfluous, as it is already a component of both PSA and AlogP98. Extensive validation of the model on hundreds of known orally delivered drugs, "drug-like" molecules, and Pharmacopeia, Inc. compounds, which had been assayed for Caco-2 cell permeability, demonstrated a good rate of successful predictions (74-92%, depending on the dataset and exact criterion used).

Mucosal barrier injury: biology, pathology, clinical counterparts and consequences of intensive treatment for haematological malignancy: an overview

Mucositis is an inevitable side-effect of the conditioning regimens used for haematopoietic stem cell transplantation. The condition is better referred to as mucosal barrier injury (MBI) since it is primarily the result of toxicity and is a complex and dynamic pathobiological process manifested not only in the mouth but also throughout the entire digestive tract. A model has been proposed for oral MBI and consists of four phases, namely inflammatory, epithelial, ulcerative and healing phases. A variety of factors are involved in causing and modulating MBI including the nature of the conditioning regimen, the elaboration of pro-inflammatory and other cytokines, translocation of the resident microflora and their products, for example, endotoxins across the mucosal barrier, exposure to antimicrobial agents and whether or not the haematopoietic stem cell graft is from a donor. Neutropenic typhlitis is the most severe gastrointestinal manifestation of MBI, but it also influences the occurrence of other major transplant-related complications including acute GVHD, veno-occlusive disease and systemic infections. The pathobiology, clinical counterparts and the means of measuring MBI are discussed together with potential approaches for prevention, amelioration and, perhaps, even cure. Bone Marrow Transplantation (2000) 25, 1269-1278.

Induction of UDP-glucuronosyltransferase by the flavonoids chrysin and quercetin in Caco-2 cells

PURPOSE

Dietary flavonoids have been reported to be potent inhibitors of drug metabolizing enzymes. In the present study we examined the inducing effect of three of these compounds, chrysin, quercetin and genistein, on UDP-glucuronosyltransferase (UGT) in the human intestinal cell line Caco-2.

METHODS

The induction of UGT by flavonoid pretreatment was studied both in the intact cells and cell homogenates, measured as the glucuronidation of chrysin, and by immunoblot analysis of the UGT 1A protein.

RESULTS

Exposure of Caco-2 cells to 50 microM chrysin resulted in a 3.8-fold increase in chrysin glucuronidation in intact cells (p < 0.0001) with a 38% decrease in sulfation (p < 0.01). In the cell homogenate the induction was much larger, 14-fold. The induction was slow to develop with maximum induction after 3-4 days. Interestingly, the isoflavonoid genistein was without effect. Immunoblot analysis of Caco-2 cell microsomes with a UGT1A subfamily-selective antibody showed a markedly increased band at about 59 kDa, consistent with induction of one or more UGT1A isoforms. A 5-week exposure of Caco-2 cells to low concentrations (10 microM) of chrysin or quercetin also showed markedly increased glucuronidation activity.

CONCLUSIONS

Diet-mediated induction of intestinal UGT may be important for the bioavailability of carcinogens and other toxic chemicals as well as therapeutic drugs.

Caco-2 versus Caco-2/HT29-MTX co-cultured cell lines: permeabilities via diffusion, inside- and outside-directed carrier-mediated transport

PURPOSE

The objective of this study was a systematic characterization and evaluation of cell culture models based on mixtures of Caco-2/HT29-MTX co-cultures for their use in screening for drug absorption and intestinal permeability in comparison to the properties of the respective mono-cultures.

METHODS

Co-cultures of Caco-2 cells (absorptive-type) and HT29-MTX cells (goblet-type) were set up. Three different co-cultures (initial seeding ratios Caco-2/HT29-MTX: 90/10, 70/30, and 50/50) were grown on permeable filter supports, and monolayers were used for permeability studies with model compounds for paracellular absorption (atenolol, furosemide, H334/75, mannitol, terbutaline), transcellular absorption (antipyrine, ketoprofen, metoprolol, piroxicam), carrier-mediated absorption (D-glucose, Gly-Pro, and L-phenylalanine) as well as substrates for carrier-mediated secretion via P-glycoprotein (cimetidine and talinolol). Electrophysiological and microscopic controls were performed to characterize the cell cultures.

RESULTS

For compounds undergoing passive intestinal absorption permeabilities were generally higher in co-cultures than in Caco-2 monolayers, yielding highest values in pure HT29-MTX monolayers. This difference was most obvious for compounds transported via the paracellular pathway, where HT29-MTX cells may be up to 30 times more permeable than Caco-2 cells, whereas for lipophilic and highly permeable compounds, the difference in permeability values was less obvious. For drugs undergoing intestinal secretion mediated by P-glycoprotein, co-cultivation of Caco-2 cells with HT29-MTX cells led to increased apical to basolateral permeability which was decreased in the opposite direction, consistent with the fact that HT29-MTX cells do not express P-glycoprotein. When a carrier-mediated absorption mechanism is involved, the permeabilities observed were lower than the values reported for human small intestine and co-cultivation of HT29-MTX cells with Caco-2 cells resulted in even lower values as compared to the plain Caco-2 cultures.

CONCLUSIONS

Co-cultures of HT29-MTX and Caco-2 cells offer the opportunity of modifying the permeability barrier of the cell monolayers both with respect to paracellular resistance and secretory transport via P-gp. Thus, in special cases, they allow more flexibility in adapting the in vitro system to the in vivo situation as compared to the monocultures. Another advantage is the obvious robustness of the method with respect to the reproducibility of the results. A problem remaining, however, is the quantitative expression of carriers involved in intestinal uptake of many nutrients and drugs.

Extensive metabolism of the flavonoid chrysin by human Caco-2 and Hep G2 cells

1. Chrysin is one of many bioflavonoids with chemopreventive properties in cardiovascular disease and cancer. In an effort better to understand factors that may affect the oral bioavailability of the bioflavonoids from dietary sources, the metabolism of chrysin by cultured intestinal Caco-2 cells and hepatic Hep G2 cells was studied, together modelling human presystemic metabolism. 2. At concentrations that may be achieved in the diet, chrysin was extensively metabolized to two conjugated metabolites, M1 and M2, with no CYP-mediated oxidation. M1 was identified as a glucuronide, and M2 as a sulphate conjugate by LC/MS and other spectroscopic and biochemical techniques. Sulphate conjugation occurred at a rate twice that of glucuronic acid conjugation in both cell types. 3. M1 was catalyzed by UGT1A6 with a Km = 12 microM. M2 was catalyzed both by M- and P-form phenolsulphotransferases (SULT 1A3 and SULT 1A1) with very low Km of 3.1 and 0.05 microM respectively. 4. Pretreatment with 3-methylcholanthrene, interestingly, did not result in oxidation of chrysin but rather in increased glucuronidation. 5. Also, M1 and M2 were the only metabolites formed from chrysin in fresh rat hepatocytes. The metabolism of another flavonoid, apigenin, was very similar to that of chrysin. 6. These observations suggest that both sulphation and glucuronidation are critical determinants of the oral bioavailability of bioflavonoids in humans, although a contribution from CYP-mediated oxidation can not be excluded.

Transport of the cooked-food mutagen 2-amino-1-methyl-6-phenylimidazo-[4,5-b]pyridine (PhIP) across the human intestinal Caco-2 cell monolayer: role of efflux pumps

Cooked-food mutagens formed when frying meat have been suggested to contribute to the etiology of colon, breast and prostate cancer. The most prevalent of these mutagens is 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP), which after absorption is bioactivated by both phase I and phase II enzymes. Although available data suggest absorption of PhIP in humans, the extent and mechanism of absorption are unknown. In the present study we examined the transport of [(3)H]PhIP through the human Caco-2 intestinal epithelial cell monolayer, a well-accepted model of human intestinal absorption. The influx, or absorption, was extensive and linear for 2 h and up to a PhIP concentration of 5 microM. Still, the basolateral to apical efflux [apparent permeability coefficient (P(app)) 54.2 +/- 0.7x10(-6) cm/s, mean +/- SEM, n = 24] was 3.6 times greater than the apical to basolateral influx (P(app) 15.1 +/- 0.6x10(-6) cm/s, n = 21, P < 0.0001). Equilibrium exchange experiments demonstrated the efflux to be a true active process. Preincubations with verapamil, an inhibitor of P-glycoprotein-mediated transport, or MK-571, an inhibitor of multidrug resistance-associated protein-mediated transport, stimulated influx and reduced efflux of PhIP, suggesting that PhIP is a substrate for both of these transporters. These findings should be considered when determining exposure to the cooked food mutagens.

The influence of plasma binding on absorption/exsorption in the Caco-2 model of human intestinal absorption

The Caco-2 cell monolayer has become an increasingly useful in-vitro model of human intestinal absorption. In this study we have determined the effect of plasma on the basolateral side on the absorption as well as exsorption of several drugs that are highly bound to plasma proteins. The drugs used included propranolol and quercetin, which both use the transcellular route of absorption, and taxol and oestradiol 17 beta-D-glucuronide, which are thought to undergo efflux by P-glycoprotein and the multidrug resistance protein MRP, respectively. All experiments were carried out under sink conditions to mimic normal absorption. It was necessary to use heparin anticoagulation for generation of the plasma, as EDTA was found to make the monolayers very leaky. The apparent permeability (P(app)) values for absorption were 1.54 x 10(-6) cm s(-1) for oestradiol 17 beta-D-glucuronide, 3.33 x 10(-6) cm s(-1) for taxol, 20.8 x 10(-6) cm s (-1) for quercetin, and 35.3 x 10(-6) cm s(-1) for propranolol. For these four compounds, plasma on the basolateral side had no influence on absorption. However, plasma on the basolateral side significantly reduced the efflux of oestradiol 17 beta-D-glucuronide by 66%, taxol by 75%, propranolol by 82%, and quercetin by 94%. Failure to consider the effect of plasma binding can result in an overestimate of basolateral to apical efflux and result in misleading net flux calculations.

Transport of the flavonoid chrysin and its conjugated metabolites by the human intestinal cell line Caco-2

Chrysin (5,7-dihydroxyflavone), a natural product present in our daily diet, is a potent inhibitor of drug-metabolizing enzymes. However, its oral bioavailability is not known. This study examined the intestinal epithelial transport of chrysin (20 microM), using the human colonic cell line Caco-2 as a model of human intestinal absorption. The apical to basolateral flux of chrysin, with an apparent permeability coefficient (P(app)) during the first hour of 6.9 +/- 1.6 x 10(-6) cm x sec(-1) (mean +/- SEM), was more than 10-fold higher than for the paracellular transport marker mannitol, 0.42 +/- 0.12 x 10(-6) cm x sec(-1). Interestingly, the reverse, basolateral to apical flux of chrysin, P(app) = 14.1 +/- 1.6 x 10(-6) cm x sec(-1), was about 2-fold higher than the apical to basolateral flux (P < 0.01). In transport studies beyond 1 hr, there was a rapid decline in P(app). This correlated with the appearance of two metabolites, M1 (chrysin glucuronide) and M2 (chrysin sulfate), identified by enzymatic hydrolysis procedures and HPLC. Following apical loading of chrysin, as much as 90% of M1 + M2 appeared on the apical side, thus indicating clear efflux of the chrysin metabolites. The addition of the anion transport inhibitor MK-571 (50 microM) on the apical side produced a 71% (P < 0.0001) and 20% (P < 0.05) inhibition of the efflux of M1 and M2, respectively, suggesting the involvement of the multidrug resistance protein MRP2 pump. Indeed, using specific antibodies, MRP2 was in fact detected by western blotting in Caco-2 plasma membranes, whereas MRP1 was not. These observations suggest that chrysin has favorable membrane transport properties but that its intestinal absorption may be seriously limited by surprisingly efficient glucuronidation and sulfation by the enterocytes and almost quantitative efflux by MRP2 of the metabolites formed.

Bioavailability of genistein, daidzein, and their glycosides in intestinal epithelial Caco-2 cells

In this study information was obtained on bioavailability of genistein, daidzein and their glycosides in human intestinal epithelial Caco-2 cells grown on semi-permeable filters. The integrity of Caco-2 monolayers was confirmed by transepithelial electrical resistance measurements and by determination of the permeability of the radioactive marker polyethylene glycol (PEG4000). After 6 h approximately 30-40% of genistein and daidzein added at the apical side was transported to the basolateral side and this level was maintained for 24 h, The glycosides were barely transported through the Caco-2 cells. No significant metabolism of genistein and daidzein in the Caco-2 cells occurred, whereas the glycosides were mainly metabolised to their respective aglycones. Obviously, our data indicates that Caco-2 cells contain an endogenous glycosidase activity.

A contribution to safety assessment of veterinary drug residues: in vitro/ex vivo studies on the intestinal toxicity and transport of covalently bound residues

1. The gastrointestinal fate of protein-bound residues of the model compound furazolidone (FZD) was investigated in vitro and ex vivo. Protein-bound residues were generated in rat liver microsomes, isolated by solvent extraction and digested with 0.5% hydrochloric acid and Pronase E. 2. During digestion, 3-amino-2-oxazolidinone (AOZ), the side chain of furazolidone, was partly released from bound residues. 3. The absorption of free AOZ and digested protein-bound residues was tested in isolated perfused rat gut segments (IPGS) and in the intestinal cell line Caco-2. Free AOZ was transfered both in the IPGS model and in Caco-2 monolayer cultures, while no indications for passage of bound residues were obtained. 4. No acute toxicity of AOZ or digested food residues respectively was observed in gut segments and Caco-2 cells at concentrations that were substantially above maximum residue levels to be expected in food of animal origin after administration of therapeutic doses. 5. The results demonstrate that digestive processes can alter the chemical nature of drug residues and yield degradation products that may be bioavailable for the consumer. Thus, the covalent binding of xenobiotics to macromolecular tissue constituents cannot necessarily be regarded as an irreversible endpoint of residue bioavailability and toxicity.

The effects of nitrogen-containing bisphosphonates on human epithelial (Caco-2) cells, an in vitro model for intestinal epithelium

Nitrogen-containing bisphosphonates (N-PCP) are bisphosphonates with an increased antiresorptive potency. Aminobisphosphonates, N-PCPs with an amino group, can cause nonspecific gastrointestinal complaints. It is not known whether these side effects are specific for these bisphosphonates or for the whole class of N-PCPs. In this study, we investigated the effects of two aminobisphosphonates (pamidronate and alendronate) and a structurally similar N-PCP (olpadronate) and their three respective calcium complexes on the viability and the intracellular calcium concentration ([Ca2+]i) of cultured Caco-2 cells a model for intestinal epithelium. These cells were also examined for apoptosis or necrosis. In the presence of calcium, pamidronate and alendronate were toxic to the cells, with pamidronate being more toxic than alendronate. Olpadronate induced toxicity only at concentrations more than ten times higher than the toxic concentrations of pamidronate. In the absence of calcium definite signs of toxicity were observed only with pamidronate at clinically relevant concentrations. The complexes of pamidronate and alendronate with calcium were considerably less soluble than the olpadronate calcium complex. There were no signs of apoptosis. [Ca2+]i was transiently raised after treatment with the N-PCPs. Doses at which responses were seen were, respectively, 0.02 mM (pamidronate), 0.3 mM (alendronate), and 2 mM (olpadronate). The peak of response was slightly greater after pamidronate treatment than after alendronate or olpadronate, respectively. In conclusion pamidronate, either as an ion or as a calcium complex, is the most toxic of the bisphosphonates tested for Caco-2 cells. Alendronate was less toxic while olpadronate was the least toxic in presence of calcium. The solubility of the bisphosphonate complexes with calcium may account for these differences in toxicity.

Primary cultures of fully differentiated and pure human intestinal epithelial cells

The epithelium of the small intestinal mucosa is a highly dynamic system particularly well suited for analyzing key biological phenomena such as cell differentiation and migration, cell-matrix interactions, and apoptosis. However, in vitro models of fully differentiated normal human enterocytes are still lacking. The objective of the present study was to investigate the possibility of generating such differentiated intestinal cell cultures from the fetal small intestine. For this purpose, various dissociation methods were tested in order to obtain pure, viable, and functional enterocytes. One of these methods, based on the procedure to recover epithelial cells grown on Matrigel and involving the use of Matrisperse, a nonenzymatic dissociating solution, was found to allow the isolation of the integral epithelial lining from the mesenchyme. In culture, these epithelial fractions plated on collagen I spread rapidly and reached confluence after 3-4 days. When tested after 5-7 days, these primary cultures of differentiated enterocytes (PCDE) remained well preserved. Both goblet and absorptive cells exhibit all the main characteristics of intact villus intestinal cells as assessed by electron microscopy. Indirect immunofluorescence and Western blot analyses confirmed the purity of PCDE. The functional status of these cells was demonstrated by the presence of uniformly distributed tight junction, zonula adherens, and desmosomal components at the region of cell to cell contact as well as expression of various brush border enzymes, namely sucrase-isomaltase and lactase, and goblet cell mucins. As expected, cell proliferation was found to be negligible as assessed by DNA synthesis. Taken together, these data show that primary cultures of pure and viable differentiated enterocytes can be generated from the human fetal small intestine.

Lectin-mediated bioadhesion: binding characteristics of plant lectins on the enterocyte-like cell lines Caco-2, HT-29 and HCT-8

In order to take advantage of the biorecognition between lectins and specific carbohydrates for targeted drug delivery, fluorescein-labelled lectins of different carbohydrate specificities were screened for binding to human colorectal carcinoma cell lines by flow cytometry and confocal microscopy. The lectin-binding rate increased as follows: Dolichos biflorus agglutinin, DBA<peanut agglutinin, PNA<Lens culinaris agglutinin, LCA<Solanum tuberosum lectin, STL<Ulex europaeus isoagglutinin I, UEA-I<wheat germ agglutinin, WGA (Caco-2); PNA<UEA-I<WGA (HT-29); DBA<UEA-I<WGA (HCT-8), thus reflecting the glycosylation pattern of the cells. Compared to the BSA-binding capacity of the cells, the extent of nonspecific binding was strongly dependent on the type of cell line and lectin under investigation being lower than 2% in the case of WGA, STL and UEA-I/Caco-2 and HT-29 cells. Whereas 50% of DBA was bound nonspecifically to Caco-2 cells, the interactions DBA/HCT-8 and PNA/HT-29 were due to nonspecific binding. By competitive inhibition of lectin-adhesion to the cells upon addition of the complementary carbohydrate, specificity of lectin-binding was confirmed except for the interaction of DBA/HCT-8 and PNA/HT-29. Following on from this work, we consider WGA-, STL- and UEA-I-mediated drug delivery to be a promising approach for peroral bioadhesive formulations adhering to absorptive enterocytes.

Effects of vegetables-fruit extracts and indole-3-carbinol on stearic acid-modulated intercellular communication and cytochrome P450-IA activity

Modulatory effects were investigated of extracts of a vegetables-fruit mixture and indole-3-carbinol (I3C) on stearic acid-modulated gap junctional intercellular communication (GJIC) and cytochrome P450-IA activity (EROD). In V79 cells, pure water and hexane extracts of a vegetables-fruit mixture and 25 μg/ml I3C significantly protected against decreased GJIC caused by 10 μM stearic acid. Furthermore, pure, 10× and 100× diluted vegetables-fruit extracts significantly maintained their capacity to induce EROD activity in Caco-2 cells, but only when these extracts were added to the cells in media already containing 500 μM stearic acid for 48 h. Stearic acid itself did not induce EROD activity. I3C (10, 25, and 50 μg/ml) clearly induced EROD activity in Caco-2 cells, irrespective of the order at which I3C and stearic acid were added to the cells. In conclusion, the present in vitro study showed that vegetables-fruit extracts and I3C modulate effects of stearic acid on intercellular communication and cytochrome P450-IA activity.

Transport of quercetin and its glucosides across human intestinal epithelial Caco-2 cells

There is mounting evidence from human epidemiological, animal in vivo, and in vitro studies to suggest beneficial effects related to the consumption of quercetin and its glucosides. However, there is limited knowledge on the oral bioavailability of these natural products. This study examined the intestinal epithelial membrane transport of quercetin, quercetin 4'-glucoside, and quercetin 3,4'-diglucoside, using the Caco-2 human colonic cell line, a model of human intestinal absorption. The apparent permeability (Papp) of each agent was measured in both apical to basal and basal to apical directions. The apical to basolateral flux of quercetin, Papp 5.8 +/- 1.1 x 10(-6) cm x sec(-1) (mean +/- SEM), was more than 10-fold higher than for the paracellular transport marker mannitol, 0.48 +/- 0.09 x 10(-6) cm x sec(-1) (P < 0.01). Under identical conditions, the Papp for the transcellular marker propranolol was about 5-fold higher than for quercetin (P < 0.001). Interestingly, the reverse, basolateral to apical, flux of quercetin (Papp 11.1 +/- 1.2 x 10(-6) cm x sec(-1)) was almost 2-fold higher than the apical to basolateral flux (P < 0.001). In similar experiments, quercetin 4'-glucoside demonstrated no absorption, Papp < 0.02 x 10(-6) cm x sec(-1) in the apical to basal direction, but did demonstrate basal to apical flux, Papp 1.6 +/- 0.2 x 10(-6) cm x sec(-1). Quercetin 3,4'-diglucoside showed a low apical to basolateral transport (Papp 0.09 +/- 0.03 x 10(-6) cm x sec(-1)); its reverse, basolateral to apical, transport was, however, 4-fold higher (P < 0.05). In these cells, glucose was actively transported with an apical to basolateral Papp of 36.8 +/- 1.1 x 10(-6) cm x sec(-1). These observations suggest facile absorption of quercetin through the human intestinal epithelium, but contrary to a previous proposal, they do not support an active transport process for quercetin glucosides.

Caco-2 cells express a combination of colonocyte and enterocyte phenotypes

Caco-2 cells are derived from a human colonic adenocarcinoma, but differentiate into small intestinal-like cells after confluence. While this enterocytic differentiation has been well studied, the presumed parallel loss of colonocyte function has not been as thoroughly examined. To follow the phenotype for both tissues, Western blots were performed using antisera recognizing liver/bone/kidney alkaline phosphatase and surfactant-like particle proteins found in normal human colon, along with antisera against the small bowel representatives of the same proteins. Antisera against proteins enriched in either enterocytes (alpha1-antitrypsin) or colonocytes (surfactant protein A) were also evaluated. Alkaline phosphatase activity increased from 3 to 18 days post-confluence. Activity at 3 days post-confluence derived substantially from both isomers. Thereafter, the colonic (liver/bone/kidney) isomer declined to low levels as the content of the enterocytic isomer rose. A similar pattern was found with colonic (decreasing expression) and enterocytic (increasing expression) surfactant-like particle proteins. In particular, the content of larger enterocytic particle proteins (97 and 116 kDa) increased with time in culture. Expression of alpha1-antitrypsin increased early and remained high, whereas surfactant protein A generally declined after the third day post-confluency. In summary, undifferentiated Caco-2 cells express very low levels of proteins characteristic of either colonocytes or enterocytes. Immediately after confluence, they expressed proteins characteristic of both cell types. Thereafter, the content of colonocyte-specific proteins decreased, whereas those specific for the enterocyte increased. The timing and degree of this phenotypic switch have implications for the interpretation of experiments using Caco-2 cells as a model of small intestinal function.

Poly(ADP-ribose) synthetase activation mediates increased permeability induced by peroxynitrite in Caco-2BBe cells

BACKGROUND & AIMS

Peroxynitrite induces cytotoxicity by generating DNA single-strand breaks and activating poly(ADP-ribose) synthetase (PARS), a nuclear enzyme that consumes oxidized nicotinamide adenine dinucleotide (NAD+) and depletes cellular adenosine triphosphate (ATP). The aim of this study was to examine this mechanism of injury in an intestinal epithelial cell model after exposure to exogenous peroxynitrite (ONOO-) and nitric oxide (NO).

METHODS

Caco-2BBe cell monolayers exposed to donors of peroxynitrite (3-morpholino-sydnonimine [SIN-1], 3 mmol/L) or NO (S-nitroso-N-acetyl penicillamine [SNAP]; 3 mmol/L) were analyzed for DNA strand breaks, [NAD+], [ATP], and transepithelial flux of fluorescein sulfonic acid.

RESULTS

SIN-1 but not SNAP induced DNA single-strand breakage. Both SIN-1 and SNAP reduced [ATP], but only SIN-1 reduced [NAD+]. Inhibition of PARS activity by the PARS inhibitors 5-iodo-6-amino 1,2-benzopyrone or 3-aminobenzamide prevented the SIN-1-induced reduction in [NAD+] and [ATP] but had no effect on the SNAP-induced reduction in [ATP]. PARS inhibition reduced SIN-1-but not SNAP-induced hyperpermeability.

CONCLUSIONS

Peroxynitrite but not NO increases transepithelial permeability by inducing DNA strand breaks that activate the PARS pathway and cause the depletion of intracellular energy stores. Inhibition of PARS activity may represent a novel strategy in ameliorating peroxynitrite-mediated epithelial injury during intestinal inflammation.

The drug transporter P-glycoprotein limits oral absorption and brain entry of HIV-1 protease inhibitors

Currently available HIV-1 protease inhibitors are potent agents in the therapy of HIV-1 infection. However, limited oral absorption and variable tissue distribution, both of which are largely unexplained, complicate their use. We tested the hypothesis that P-glycoprotein is an important transporter for these agents. We studied the vectorial transport characteristics of indinavir, nelfinavir, and saquinavir in vitro using the model P-glycoprotein expressing cell lines L-MDR1 and Caco-2 cells, and in vivo after intravenous and oral administration of these agents to mice with a disrupted mdr1a gene. All three compounds were found to be transported by P-glycoprotein in vitro. After oral administration, plasma concentrations were elevated 2-5-fold in mdr1a (-/-) mice and with intravenous administration, brain concentrations were elevated 7-36-fold. These data demonstrate that P-glycoprotein limits the oral bioavailability and penetration of these agents into the brain. This raises the possibility that higher HIV-1 protease inhibitor concentrations may be obtained by targeted pharmacologic inhibition of P-glycoprotein transport activity.

The efflux of lysine from the basolateral membrane of human cultured intestinal cells (Caco-2) occurs by different mechanisms depending on the extracellular availability of amino acids

The efflux of the nutritionally essential amino acid, L-lysine from the basolateral (BL) membrane was characterized in human cultured intestinal cells (Caco-2) grown and differentiated on permeable filter supports. Cells were loaded by incubating with 3H-lysine from the apical (AP) side in the absence of sodium (substituted with choline) in the BL medium; under these conditions, cells accumulated lysine in the intracellular soluble pool to 10- to 20-fold the extracellular concentration. L-Lysine efflux in the BL medium was then followed, and initial rates of efflux were calculated under different experimental conditions. L-Lysine efflux exhibited a strong energy dependence. The presence of an inwardly directed gradient of sodium or lithium stimulated lysine efflux; ouabain reduced efflux in both sodium- and lithium-containing medium. When zwitterionic or cationic amino acids were added to the BL medium, L-lysine efflux was strongly stimulated. The most efficient trans-stimulating amino acids were L-leucine > L-methionine = L-ornithine = L-arginine. In the presence of trans-stimulating amino acids in the BL medium, L-lysine efflux exhibited energy independence and was not affected by the presence of a sodium gradient. In addition, the sensitivity, of efflux to N-ethylmaleimide was different in the absence or in the presence of amino acids in the BL medium. These results suggest that different mechanisms may operate in the BL efflux of L-lysine from human intestinal epithelial cells, depending on the extracellular availability of other amino acids, to guarantee optimal bioavailability of this essential amino acid both in the postprandial absorptive period and between meals.