Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

The H⁺/oligopeptide cotransporter PepT1 is a 12 transmembrane domain protein located in the brush-border membrane of enterocytes that is specific for di- and tripeptides arising from digestion of dietary proteins (for review, see Leibach and Ganapathy, 1996). Besides its key role in nutrient absorption, PepT1 is also pharmacologically relevant because its activity is responsible for absorption of peptidomimetic drugs such as -lactam antibiotics.

Although several studies have been carried out on the functional and molecular characteristics of PepT1 (Liang et al., 1995; Mackenzie et al., 1996), little information is available on the regulation of its activity (Brandsch et al., 1994; Muller et al., 1996; Fujita et al., 1997, 1999; Thamotharan et al., 1999b). A recent study demonstrated transcriptional enhancement of PepT1 expression in rats fed with protein-rich diet. However, in a previous work with a single-pass jejunal perfusion technique in anaesthetized rats, we showed that intestinal absorption of the -lactam antibiotic cephalexin (CFX), which is carried by PepT1, was influenced by the nervous system (Berlioz et al., 1999). In our experiments, stimulation of PepT1 occurred very rapidly, excluding a transcriptional control, and depended on the activity of intramural and/or extramural neuron networks, including nicotinic synapses, intestinal sensory neurons, and sympathetic noradrenergic fibers. Among the agents acting on neurotransmitter receptors that were tested, administration of the ₂-agonist clonidine induced a 2-fold increase of the intestinal absorption of CFX. In the small intestine, ₂-adrenergic receptors are present on enteric neurons, on extrinsic sympathetic postganglionic neurons (Cooke and Reddix, 1994), and also on intestinal epithelial cells (Laburthe et al., 1982; Nakaki et al., 1983). Thus, the precise location of the receptor responsible for the stimulatory effect of clonidine in vivo is unclear. The purpose of this study was to investigate whether ₂-adrenergic receptors can directly affect the transport of peptidomimetic drugs in cultured epithelial cells devoid of innervation.

To fulfill this objective, we needed to use a polarized intestinal cell line possessing PepT1 and ₂-adrenergic receptors. However, a cell model spontaneously expressing both proteins is not currently available. The HT29 colonic cells express ₂-adrenergic receptors but not PepT1 (Langin et al., 1989, 1995). Conversely, Caco-2 cells, frequently used to study intestinal drug absorption (Zweibaum et al., 1991) show enterocytic differentiation and express the PepT1 transporter (Liang et al., 1995) but not the ₂-adrenergic receptor (Devedjian et al., 1991). The absence of suitable model was recently circumvented by the generation of a clone of Caco-2 cells stably transfected with the ₂C10 adrenergic receptor gene. This clone, referred to as Caco-2 3B (Schaak et al., 2000), expresses ₂-adrenergic receptors at a density similar to that found on enterocytes and colonocytes of different species, including humans (Paris et al., 1990; Senard et al., 1990; Valet et al., 1993). As in normal intestinal cells, the receptor is coupled to G_i2 and G_i3 and its stimulation inhibits forskolin-stimulated cAMP production. We thus decided to use Caco-2 3B cells to study a possible direct involvement of ₂-adrenergic receptors in the activation of CFX transport.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. Caco-2 cells were purchased from the American Type Culture Collection (Rockville, MD). HT-29 clone 19A cells were a generous gift from C. Laboisse (CJF 9404, Nantes, France). The clone of Caco-2 cells (Caco-2 3B) expressing ₂-adrenergic receptors was obtained by transfection of the parental cell-line with the bicistronic plasmid p₂C10ENeo containing the coding region of the human _2A-adrenergic receptor subtype (Schaak et al., 1999). Dulbecco's modified Eagle's medium (DMEM), trypsin solution, and fetal calf serum (FCS) were purchased from Gibco-BRL (Cergy Pontoise, France). Cephalexin, clonidine, yohimbine, glycyl-sarcosine, amiloride, and colchicine were obtained from Sigma (St. Louis, MO). UK14304 and RX821002 were donated by Pfizer (Sandwich, UK) and Reckitt and Colman Laboratories (Kingston-upon-Hull, UK), respectively. [¹⁴C]Mannitol (specific radioactivity, 57 Ci/mmol) and [³H]RX821002 (specific radioactivity, 53 Ci/mmol) were purchased from Amersham (Amersham, UK). [³H]Clonidine (specific radioactivity, 66 mCi/mmol) was from New England Nuclear (Boston, MA).

Cell Culture. Caco-2 3B (passages 18-27) and Caco-2 (passages 35-37) cells were propagated in 25-cm² flasks at 37°C in a humidified 5% CO₂ incubator in DMEM supplemented with 20% FCS and 1% nonessential amino acids (Zweibaum et al., 1991). When reaching confluency, cells were trypsinized and plated (starting density, 5 × 10⁴ cells/cm²) on Transwell Clear polyester membranes, 1 cm² in surface and 0.4 µm in pore size (Costar, Dutscher, France). Culture medium was changed every day and, except where noted, monolayers at day 16 to 17 postseeding were used for transport experiments. HT-29 clone 19A cells (passage 154) were subcultured and plated as Caco-2, except they were grown in DMEM supplemented with 10% FCS and 1% nonessential amino acids.

Adrenergic Receptor Quantification. The expression of ₂-adrenergic receptors in the different cell types was assessed by binding studies with [³H]RX821002 (₂-antagonist) and [³H]clonidine (₂-agonist) as specific radioligands. Binding experiments were performed on crude membranes prepared from frozen cells as described previously (Paris et al., 1990). Briefly, total binding was measured by incubating 100 µl of membranes with the radioligand in a total volume of 400 µl of binding buffer (50 mM Tris-HCl, 0.5 mM MgCl₂, pH 7.5). After a 45-min incubation at 25°C, bound radioactivity was separated from free by filtration through GF/C Whatman filters with a Millipore manifold sampling unit. Filters were rapidly washed with ice-cold buffer and bound radioactivity was determined by liquid spectrometry. Specific binding was defined as the difference between total and nonspecific binding measured as described above but in the presence of 10 µM phentolamine. Final concentrations of radioligand ranged from 0.1 to 10 nM for [³H]RX821002 and from 0.05 to 8 nM for [³H]clonidine. Saturation isotherms were analyzed with the EBDA-LIGAND computer programs (McPherson, 1985) and protein concentration was determined with the Coomassie blue method (Bradford, 1976).

Reverse Transcription-Polymerase Chain Reaction (RT-PCR). Total RNAs were extracted from Caco-2, Caco-2 3B, and HT-29 19A cells with RNAXEL (Eurobio, Les Ulis, France) according to the manufacturer's instructions. Ten micrograms of total RNA was reverse transcribed with Moloney murine leukemia virus RNase at 37°C for 45 min and then heated at 80°C for 5 min. The synthesized cDNA was used for subsequent PCR with two sets of primers allowing us to amplify either PepT1 or GAPDH, taken as a control for housekeeping gene. The primers for PepT1 were identical with those used in Liang et al. (1995). The sense 5'-TCCACCGCCATCTACCATAC-3' and antisense 5'-GGACAAACACAATCAGGGCT-3' primers allow amplification of a 479-base pair fragment corresponding to nucleotides 210 to 708 of the human PepT1 cDNA. Primers for GADPH were as follows: 5'-TGAAGGTCGGAGTCAACGGATTTGGT-3' (sense) and 5'CATGTGGGCCATGAGGTCCACCAC-3' (antisense). Reaction mixtures were subjected to 35 cycles consisting of 30 s at 92°C, 30 s at 58°C, and 1 min at 72°C. The PCR products were electrophoresed on a 1% agarose gel and stained with ethidium bromide.

Transport Studies. The day of the experiment, the transepithelial resistance of each cell layer was measured with a Millicel-ERS ohmmeter (Millipore, St. Quentin en Yvelines, France). Monolayers exhibiting a transepithelial resistance above 150 /cm² were considered as satisfactory and were further used for the transport experiments. Before transport studies, the culture medium was removed and the apical and basolateral compartments were washed three times with Krebs' solutions [137 mM NaCl, 5.4 mM KCl, 2.8 mM CaCl₂, 1 mM MgSO₄, 0.3 mM NaHPO₄, 0.3 mM KH₂PO₄) buffered with either 10 mM HEPES/Tris, pH 7.4 (basolateral compartment) or 10 mM Mes/Tris, pH 6.0 (apical compartment)]. Cell monolayers were incubated for 15 min at 37°C under continuous circular shaking in Krebs' modified buffer (pH 6.0 in the apical compartment and 7.4 in the basolateral compartment). Unless otherwise specified, at zero time of the experiment, CFX (final concentration 1 mM) was added to the apical compartment. The rate of CFX transport was estimated over a 30 min-period by measuring CFX concentration (see below) in 50-µl aliquots taken from the basolateral compartment every 5 min. For competition experiments, glycyl-sarcosine (50 mM) was added to the apical compartment before the addition of CFX. The ₂-adrenergic drugs (clonidine, UK14304, yohimbine, or RX821002) were added to the basolateral compartment 15 min after CFX. The effects of these pharmacological agents were evaluated on the same cell monolayer by comparing basal CFX flux (calculated from the three samples taken during the 0- to 15-min incubation period) with the flux measured after their addition (calculated from the three samples taken during the 15- to 30-min incubation period). In some experiments, cells were pretreated with colchicine (20 µM) or amiloride (10 µM); these inhibitors were added to the apical compartment 25 and 10 min, respectively, before the addition of CFX.

Determination of CFX Concentration. CFX concentration was measured by HPLC on a Supelcosil LC18 column (250 × 4.6 mm, 5-µm particle size; Supelco, Touzart et Matignon, France). The HPLC apparatus comprised a WISP 712 automatic sampler (Waters, Paris, France), an SPD-6AV pump, and an SPD-6AV UV detector (Shimadzu, Paris, France) set at 260 nm to monitor the CFX peak that came off at 14 min. The mobile phase was a mixture of sodium acetate buffer (0.01 M, pH 5.2) and acetonitrile (94.5:4.5, v/v). The flow rate was 1.5 ml/min.

Data Analysis. The experiments were performed in triplicate and repeated twice. The data are presented as mean ± S.E.M. The means were compared by paired or unpaired Student's t test, as appropriate; P < .05 was considered significant. The kinetic constants of CFX transport were determined by applying a nonlinear regression method to fit the Michaelis-Menten kinetic equation with the GRAFIT program. V_max is the maximal CFX flux and K_m is the concentration of CFX that yielded one-half of V_max. To determine the number of systems involved in CFX transport, the CFX fluxes were transformed according to the Eadie-Hofstee method, by plotting V against V/S were V is the CFX flux in nanograms per square centimeter per minute and S the apical CFX concentration in millimolar.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Junctional Integrity and Expression of hPepT1 in Caco-2 3B Cell Line. Caco-2 is a human colon cancer cell line that has retained the remarkable property to spontaneously differentiate and to behave as a functional epithelium. Thus, this cell line has been extensively used as a model to study transepithelial transport of a large panel of molecules, including dipeptides. Emergence of the differentiated phenotype is growth-related and it can be monitored by measuring changes in transepithelial electrical resistance (TEER) and in paracellular diffusion, which both reflect appearance of tight junctions. The Caco-2 3B clone was recently generated by stable transfection and this model was never used before for transport study. Preliminary experiments were therefore designed to assess its epithelial properties. TEER and [¹⁴C]mannitol flux were measured in Caco-2 3B monolayers from day 3 to day 24 postseeding. During the time interval between days 3 and 15, TEER increased 4-fold from 45 ± 6 to 180 ± 5 /cm². Conversely, the percentage of passive diffusion of mannitol decreased 17-fold from 0.50 ± 0.02 to 0.030 ± 0.002%/cm²/min. The two parameters remained stable between days 15 and 24. All further experiments were carried out on monolayers at days 16 to 17 postseeding.

View larger version (25K): [in this window] [in a new window]   Fig. 1.   A, detection by PCR amplification of hPepT1 mRNA in Caco-2 3B, Caco-2, and HT29 clone 19A cell lines. Ten micrograms of total RNA from the indicated cell line was reverse transcribed and first-strand cDNA synthesized was amplified with a set of specific primers described in Experimental Procedures. The PCR products were separated by electrophoresis through a 1% agarose gel and stained with ethidium bromide. GADPH was taken as control for a housekeeping protein. B, effect of glycyl-sarcosine on CFX transport in Caco-2 and Caco-2 3B cell lines. Transport of CFX (1 mM) was measured in confluent monolayers of Caco-2 and Caco-2 3B cells for 30 min at 37°C under continuous circular shaking. pH was 6.0 in the apical compartment and 7.4 in the basolateral compartment. After 15 min of equilibration, CFX ± glycyl-sarcosine (GS; 50 mM) was added to the apical compartment. CFX alone fluxes were measured between 0 and 30 min (CFX alone) and compared with CFX + GS 50 mM fluxes between 0 and 30 min (CFX + GS). Data are mean ± S.E.M. of triplicate measurements repeated twice. *P < .05 versus corresponding CFX alone (unpaired t test). C, 2-adrenergic receptor quantification in Caco-2 3B, Caco-2, and HT29 cl 19A cell lines with Scatchard plots. The expression of 2-adrenergic receptors was assessed by binding studies with [3H]RX821002 (2-antagonist) as specific radioligand. Binding experiments were performed on crude membranes prepared from frozen cells as described previously (Paris et al., 1990). Specific binding was defined as the difference between total and nonspecific binding measured as described above but in the presence of 10 µM phentolamine.

Effect of ₂-Adrenergic Receptor Stimulation on CFX Absorption. To investigate whether activation of ₂-adrenergic receptors had any effect on CFX transport, clonidine was added to the basolateral compartment, 15 min after adding CFX to the apical compartment. As depicted in the Fig. 2, addition of the ₂-agonist caused a rapid increase of the flux of CFX through the Caco-2 3B monolayer. Calculation of the slope of the absolute values of CFX fluxes against time under basal and stimulated conditions (Fig. 2) revealed a doubling of the mean rate of CFX transport from 5.9 ± 0.5 to 12.0 ± 0.6 ng/cm² · min. The stimulatory effect of clonidine was not observed in the presence of glycyl-sarcosine, indicating that it reflects an increase of PepT1 activity (Fig. 3). The effect of clonidine was mimicked by UK14304 (Fig. 3). However, the effects of the two ₂-agonists were totally blocked by two specific ₂-antagonists with different chemical structures, yohimbine and RX821002, strongly suggesting the involvement of ₂-adrenergic receptors in the observed effect (Fig. 3).

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Mean concentrations of CFX versus time in the basolateral compartment and effect of clonidine (clo) 105 M. Columns represent CFX fluxes measured between 0 and 15 min (CFX alone) compared with fluxes between 15 and 30 min (CFX + clo 105 M). Transport of CFX (1 mM) was measured in confluent monolayers of Caco-2 3B cells for 30 min at 37°C under continuous circular shaking. pH was 6.0 in the apical compartment and 7.4 in the basolateral compartment. After 15 min of equilibration, CFX was added to the apical compartment. clo (105 M) was added to the basolateral compartment 15 min after CFX. Data are mean ± S.E.M. of triplicate measurements repeated twice. *P < .05 versus corresponding CFX alone (paired t test).

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Effect of 2-adrenergic receptor agonists clonidine (clo), UK14304 (UK) alone, or associated with 2-adrenergic receptor antagonists yohimbine (yo), RX821002 (RX), or with glycyl-sarcosine (GS) on CFX transport. Transport of CFX (1 mM) was measured in confluent monolayers of Caco-2 3B cells for 30 min at 37°C under continuous circular shaking. pH was 6.0 in the apical compartment and 7.4 in the basolateral compartment. After 15 min of equilibration, CFX ± GS (50 mM) was added to the apical compartment. GS (50 mM) was added to the apical compartment just before CFX. clo (105 and 106 M) and UK (105 M) were added to the basolateral compartment 15 min after CFX. The 2-adrenergic receptor antagonists yo (105 M) and RX (105 M) were added to the basolateral compartment 15 min after CFX, just before adding clo or UK. Results are expressed as the percentage of CFX fluxes measured between 15 and 30 min in the presence of the various agents compared with CFX fluxes between 0 and 15 min (CFX alone). The effect of the 2-adrenergic receptor agonists was suppressed by the antagonists. Data are mean ± S.E.M. of triplicate measurements repeated twice. *P < .05 versus corresponding CFX alone (paired t test).

Mechanism of PepT1 Stimulation. The stimulation of CFX uptake caused by the ₂-agonist may be the consequence of an increase in the affinity of the oligopeptide transporter for its substrate, and/or of an augmentation of the number of transporter molecules functionally available at the apical membrane of Caco-2 3B cells. The effect of clonidine on the kinetics of CFX transport was thus investigated to clarify this point. The transformation of the kinetic data yielded linear Eadie-Hofstee plots (Fig. 4), indicating the presence of a single class of transporters in both control and clonidine-treated cells. Furthermore, the treatment with the ₂-agonist resulted in a significant (P < .05) increase of the V_max value from 54.4 ± 9.1 to 92.8 ± 13.4 ng/cm² · min without any modification of the K_m value (2.64 ± 0.79 versus 2.78 ± 0.69 mM). These data indicate that another transporter system is not recruited after ₂-agonist exposure. They also eliminate the possibility that a change in the intrinsic properties of the oligopeptide transporter occurred.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Effect of the 2-adrenergic receptor agonist clonidine (clo) on kinetic parameters of CFX flux. Transport conditions as described in Fig. 3. After 15 min of equilibration, CFX was added to the apical compartment. clo (105 M) was added to the basolateral compartment 15 min after CFX. Fluxes were measured between 0 and 15 min (CFX alone, ) and between 15 and 30 min (CFX + clo, ). A, CFX flux versus CFX concentration. B, CFX flux versus CFX flux/CFX concentration (Eadie-Hofstee plot). Data are mean ± S.E.M. of triplicate measurements repeated twice.

View larger version (23K): [in this window] [in a new window]   Fig. 5.   Effect of amiloride and of colchicine on CFX transport. Transport conditions as described in Fig. 3. Caco-2 3B cells were pretreated with amiloride (10 µM) in the apical compartment, 10 min before CFX, or with colchicine (20 µM) in the apical and basolateral compartments, 25 min before CFX. CFX was then added to the apical compartment. Clonidine (clo; 105 M) was added to the basolateral compartment 15 min after CFX. Fluxes were measured between 0 and 15 min (CFX alone) and between 15 and 30 min (CFX alone in a group and CFX + clo at 105 M in another). Amiloride did not affect neither basal transport of CFX nor clo stimulation of CFX transport. Colchicine did not affect basal transport of CFX but suppressed clo stimulation of CFX transport. Data are mean ± S.E.M. of triplicate measurements repeated twice. *P < .05 versus CFX alone.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Previous experiments conducted in anaesthetized rats have demonstrated that administration of clonidine resulted in a rapid enhancement of intestinal absorption of CFX through increased activity of the oligopeptide transporter PepT1. This effect might either be the consequence of the stimulation of ₂-adrenergic receptors located on neurons from the central nervous system, sympathetic system, or intramural network or be due to the direct activation of ₂-adrenergic receptors located on enterocytes. A clone of Caco-2 stably expressing ₂-adrenergic receptors (Caco-2 3B; Schaak et al., 2000) was therefore used as a model to determine whether ₂-agonists can directly modulate the activity of the dipeptide carrier PepT1 in the absence of nerve endings.

Preliminary experiments were designed to verify that the transfection and the selection of the Caco-2 3B clone had no incidence on epithelial properties and PepT1 expression. The measurement of TEER and the estimation of mannitol diffusion demonstrated that tightness of Caco-2 3B monolayers was identical with that of the parental cell line. The expression of PepT1 was assessed by RT-PCR and the functionality of the oligopeptide transporter was evidenced by the large inhibition of CFX transport by glycyl-sarcosine. Finally, the K_m of PepT1 for CFX in Caco-2 3B (2.64 ± 0.79 mM) was fairly similar to that previously reported in Caco-2 cells (Gochoco et al., 1994).

Exposure of Caco-2 3B cells to clonidine resulted in a doubling of CFX transport. The involvement of ₂-adrenergic receptors in this response is obvious because the effect of clonidine was mimicked by UK14304 (another ₂-agonist), blocked by RX821002 or yohimbine (₂-antagonists), and absent in Caco-2 cells, which express PepT1 but not the receptor. However, the implication of PepT1 is proved because the clonidine-induced increase of CFX transport was abolished in the presence of an excess of glycyl-sarcosine and was not observed in HT29 clone 19A, which expresses the ₂-receptor but not PepT1.

Changes in PepT1 activity have been reported in the rat intestine as a consequence of starvation (Ogihara et al., 1999; Thamotharan et al., 1999a), dietary protein content (Erickson et al., 1995; Shiraga et al., 1999), or epithelium injury by 5-fluorouracil (Tanaka et al., 1998). In Caco-2, an augmentation in the amount of PepT1 also was found after complementation of the culture medium with glycyl-glutamine (Walker et al., 1998) and after cell exposure to pentazocine, a selective ligand of -receptor (Fujita et al., 1999). In all these cases, up-regulation of PepT1 correlated with increased level of its mRNA. Although mRNA levels were not examined in this study, a modification of PepT1 gene expression is unlikely because the increase of CFX transport occurred within minutes after exposure to the ₂-agonist.

Clonidine also was previously shown as able to raise the intracellular pH of intestinal epithelial cells by increasing the activity of the Na⁺/H⁺ exchanger (Sundaram, 1995). An alkalinization of the intracellular compartment might indirectly affect CFX flux because oligopeptide transport is electrogenic and PepT1 activity is highly dependent on the H⁺-gradient (Fei et al., 1994; Mackenzie et al., 1996). In our experiments, amiloride, however, did not alter the extent of clonidine effect, suggesting that the Na⁺/H⁺ exchanger is not involved.

The determination of PepT1 kinetic parameters showed that the stimulation of CFX flux by clonidine solely resulted from an increase in the population of functionally active PepT1 (increase in V_max but no change in K_m). Furthermore, this effect was abolished after microtubule disruption by colchicine. It is thus possible that ₂-agonists act by increasing the insertion of transporter molecules, recruited from a preformed cytoplasmic pool, into the apical membrane of Caco-2 3B. A similar mechanism of translocation was demonstrated to account for the effect of insulin on PepT1 activity in Caco-2 cells (Thamotharan et al., 1999b). Further study with anti-PepT1 antibody on brush-border membrane purified from Caco-2 3B is necessary to verify this hypothesis.

The intracellular messenger responsible for the redistribution of PepT1 molecules was not determined in this study. cAMP and protein kinase C may appear as possible candidates. Indeed, a previous study with Caco-2 has shown that elevation of cAMP levels induced by cholera toxin or heat-labile enterotoxin inhibits PepT1 activity (Muller et al., 1996). Because PepT1 is devoid of site for phosphorylation by protein kinase A but does possess two putative sites for phoshorylation by protein kinase C, and because phorbol esters inhibit the transporter activity (Brandsch et al., 1994), it is thought that the effect of cAMP may be indirectly mediated by protein kinase C (Muller et al., 1996). With the ₂-adrenergic receptor being negatively coupled to adenylate cyclase (Remaury et al., 1993), one could expect that decreased intracellular level of cAMP may conversely result in enhanced PepT1 activity. However, two arguments make such a mechanism of action unlikely. First, regardless the cellular system examined, ₂-agonists were never found to inhibit protein kinase C. Second, if clonidine and other ₂-agonists are able to inhibit forskolin-induced cAMP production in Caco-2 3B cells (Schaak et al., 2000), they are unable to lower the level of this intracellular messenger in basal conditions such as those under which the effects on PepT1 activity are observed. In contrast, ₂-agonists per se were found to activate mitogen-activated protein kinase in Caco-2 3B via a cascade of events comprising recruitment of Gi-proteins, G-subunit-mediated formation of Shc-Grb2-SOS complex, and subsequent activation of mitogen-activated protein kinase kinase 1. Furthermore, stimulation of ₂-adrenergic receptors was proved to evoke focal adhesion kinase phosphorylation and a rapid rearrangement of actin cytoskeleton in smooth muscle cells (Richman and Regan, 1998) and preadipocytes (Betuing et al., 1996). In this latter case, the effect of ₂-agonists was correlated with stimulation of RhoA via a G-subunit-independent mechanism. Activation of the RhoA/focal adhesion kinase pathway was not demonstrated yet in Caco-2 3B, but the possibility exists that the effect of ₂-agonists on PepT1 is triggered by one of these two cAMP-independent mechanisms. In support to this view, it is worth mentioning that substantial amounts of Gi-proteins were found intracellularly in Caco-2 (Lacombe et al., 1996) and that these proteins were demonstrated to regulate trafficking of a number of transporters, including cystic fibrosis transmembrane conductance regulator (Schwiebert et al., 1994) and aquaporin (Valenti et al., 1998).

Finally, another point that needs to be addressed is whether our results obtained on a cell line can be extrapolated to the in vivo situation. Indeed, previous studies have shown that ₂-adrenergic receptors are abundant in crypt cells, whereas they are scarce in villus cells (Paris et al., 1990; Valet et al., 1993). Conversely, PepT1 amount is high in villus cells but low in crypt cells (Ogihara et al., 1996). It is therefore questionable whether the density of receptors in villus cells is sufficient to affect PepT1 function. The measurement of CFX transport on cells isolated from the villi is presently impossible; but experiments on other clones of Caco-2 expressing a lower amount of receptor may provide an answer to this issue. It is however noteworthy that the density of ₂-adrenergic receptors in the villi is high enough to stimulate activity of the Na⁺/H⁺ exchanger (Sundaram, 1995). One could hypothesize that the same is true for PepT1.

In conclusion, our data demonstrate that activation of an epithelial Gi-protein-coupled neurotransmitter receptor can increase the intestinal absorption of peptides and peptidomimetic drugs. Although the mechanisms involved in this effect remain to be elucidated and the relative participation of epithelial and neural receptors remains to be precisely delineated in vivo, the amplitude of the ₂-agonist effect is sufficient to be of interest to peptidomimetic drugs poorly absorbed by the intestine and to potentially represent a new field for therapeutic application of ₂-agonists.