Materials and Methods

Chemicals.

Leupeptin, phenylmethylsulfonyl fluoride, and pepstatin A were obtained from Sigma Chemical Co. (St. Louis, MO). The specific antibody against the C terminus of rat mrp2 (Liu et al., 1999) and mrp2 cDNA (Madon et al., 1997) were generous gifts from Dr. Peter Meier (University Hospital, Zurich, Switzerland). The horseradish peroxidase-linked secondary antibody used in Western blot studies was from Amersham Pharmacia Biotech, Inc. (Piscataway, NJ), whereas a biotinylated secondary antibody (Vector Laboratories Inc., Burlingame, CA) was used for immunohistochemical studies. All other chemicals were of analytical grade purity and were used as supplied.

Animals.

Adult female Sprague-Dawley rats weighing 180 to 230 g were purchased from Harlan Laboratories (Indianapolis, IN). Adult female mutant Wistar rats (TR⁻) hereditarily deficient in mrp2 protein weighing 180 to 230 g were bred in our animal facility. All animals were maintained in an environmentally controlled facility with diurnal light cycling and free access to food and water for at least 1 week before use. All procedures involving animals were conducted in accordance with National Institutes of Health Guidelines for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee of the University of Kentucky.

Specimen Collection.

Animals were sacrificed by decapitation. The whole small intestine was divided into nine equal segments (11–12 cm each) and carefully rinsed with ice-cold saline. The segments were placed in saline at 4°C until use in mucosa tissue preparation. The most proximal segment, starting from the pylorus, was given the number 1, whereas the most distal segment close to the ileo-cecal valve was given the number 9. The intestinal segments were opened lengthwise, the mucus layer was carefully removed, and the mucosa obtained by scraping (Catania et al., 1998). The tissue thus obtained was used for total homogenate or brush border membrane (BBM) preparation.

To study the gradient of mrp2 expression along the villus, three different cell populations were prepared from the two most proximal segments (taken together) using a differential scraping technique (Hoensch et al., 1976). Briefly, the upper layer of tissue corresponding to villus tip cells was removed by light hand scraping. The second cell layer was obtained by applying slightly increased scraping pressure that yielded mainly lower villus cells. The last fraction, enriched in crypt cells, was harvested by abrasion of the remaining mucosa. The effectiveness of the extraction process was monitored by visual analysis of the remaining tissue using low magnification (100×) and confirmed by testing the activity of alkaline phosphatase, a marker of villus gradient, in the preparations. Liver samples from normal rats were also collected, and mixed plasma membranes were prepared as described (Meier and Boyer, 1990).

For intestinal total RNA extraction, the two most proximal (i.e., 1 and 2) and the two most distal (i.e., 8 and 9) segments were rinsed with ice-cold saline, and the whole tissue, or alternatively the three different cell populations obtained by differential scraping of the proximal region (segments 1 and 2), were immediately placed in liquid nitrogen.

Preparation of BBM.

Mucosa samples were homogenized in buffer [50 mM mannitol, 2 mM Tris (pH 7.1), 25 μg/ml leupeptin, 40 μg/ml phenylmethylsulfonyl fluoride, and 0.5 μg/ml pepstatin A], and BBMs were prepared from total homogenate by a divalent cation precipitation method followed by differential centrifugation (Kessler et al., 1978). The final pellet was resuspended in a 300 mM mannitol, 10 mM HEPES/Tris (pH 7.5), 25 μg/ml leupeptin, 40 μg/ml phenylmethylsulfonyl fluoride, and 0.5 μg/ml pepstatin A solution. Aliquots of the homogenate and BBM preparations were stored in liquid nitrogen and used within a week for Western blot analysis and alkaline phosphatase activity determination. Protein concentration in homogenate and BBM preparations was measured (Lowry et al., 1951) with BSA as standard. Alkaline phosphatase activity was determined usingp-nitrophenylphosphate as substrate (Kit DG1245-K; Sigma Chemical Co.). Apical membrane enrichment was estimated by calculation of the ratio of the alkaline phosphatase activity in BBM to the alkaline phosphatase activity in homogenate.

Western Blot Studies.

Western blotting for mrp2 was performed with homogenate and BBM using an amount of protein (15 μg) in the gels that was found to give a densitometric signal in the linear range of the response curve for the anti-mrp2 antibody (data not shown). Preparations were loaded onto 10% SDS-polyacrylamide gels (Laemmli, 1970) and subjected to electrophoresis. After electrotransfer onto nitrocellulose membranes (Protran; Schleicher and Schuell, Keene, NH), the blots were blocked overnight at 4°C with Tris-buffered saline containing 0.1% Tween 20 and 5% nonfat dry milk and then incubated for 1 h with the primary mrp2 antibody (1:2000). The immune complex was detected by incubation with the horseradish peroxidase-linked secondary antibody (1:2000) for 1 h. Immunoreactive bands were detected using a chemiluminescence kit (ECL+Plus; Amersham Pharmacia Biotech, Inc.), exposed to Bio-Max MR-2 film (Sigma Chemical Co.) for 5 min, and quantified by densitometry (Shimadzu CS-9000; Shimadzu Corporation, Kyoto, Japan).

Northern Blot Studies.

Total RNA was isolated from mucosa samples frozen in liquid nitrogen by a guanidinium thiocyanate extraction procedure (Chomczynski and Sacchi, 1987). Total RNA (15 μg) was denatured, electrophoresed through a 1.2% agarose/formaldehyde gel, transferred to a nylon membrane (Duralon-UV; Stratagene, La Jolla, CA) overnight by capillary blotting, and UV cross-linked (FB-UVXL-1000; Fisher Scientific, Westbury, NY). The membranes were prehybridized in 50% formamide, 1% SDS, 10% dextran sulfate, and 1 mM NaCl at 42°C for at least 1 h. Hybridization was performed at 42°C for 16 h after addition of 100 mg/ml salmon sperm DNA and the full-length mrp2 probe (Madon et al., 1997). The probe was labeled with [α-³²P]dCTP by random priming using the Prime-a-Gene kit (Promega, Madison, WI). To correct for the variance of RNA loading and transfer among the lanes, a 28S rRNA oligoprobe was end-labeled with [γ-³²P]ATP by polynucleotide kinase using the 5′ end labeling kit (Promega) and hybridized. The Northern blots were analyzed with a Molecular Dynamics Phosphorimager (Sunnyvale, CA) and quantitated using ImageQuant software (Molecular Dynamics). The relative optical densities of mrp2 mRNA were then expressed relative to that for the 28S rRNA. The blots were also exposed to Bio-Max MR-2 film for 1 to 3 days.

Immunohistochemistry.

For in situ immunodetection of mrp2, Sprague-Dawley rats were anesthetized with ethyl ether and perfused via transcardiac puncture with 0.1 M Dulbecco's PBS (Life Technologies, Grand Island, NY) (pH 7.4) followed by 4% paraformaldehyde in the same buffer. After fixation, the small intestine was removed, and segments 1, 2, and 9, corresponding approximately to duodenum, proximal jejunum, and distal ileum, respectively, were kept overnight at 4°C in the above fixative containing 30% sucrose. Forty-micrometer-thick freezing microtome sections were washed for 30 min in Tris-HCl buffer (0.05 M, pH 7.6) containing 10% normal horse serum, 0.1% sodium azide, and 0.2% Triton X-100 for 1 h and then incubated overnight at room temperature with the primary antibody (1:100). After rinsing with Tris-HCl buffer, sections were incubated for 1 h with the secondary biotinylated antibody (1:400), washed in Tris-HCl buffer, and exposed for 1 h to the avidin-biotin-peroxidase complex (Elite; Vector Laboratories Inc.). Sections were stained with a solution containing 50 mg of 3,3′-diaminobenzidine tetrahydrochloride dihydrate (Aldrich, Milwaukee, WI) and 5 μl of H₂O₂ in 100 ml of Tris-HCl buffer. To distinguish between specific and nonspecific staining, immunohistochemistry was also performed in intestine from TR⁻ rats, which are genetically deficient in mrp2, so that only nonspecific staining is expected.

Statistical Analysis.

Data on densitometric analysis of Western and Northern studies and on alkaline phosphatase activities are presented as the means ± S.D. Statistical analysis was performed using the Newman-Keuls multiple-range test (Tallarida and Murray, 1986), which includes ANOVA. Values of P < .05 were considered to be statistically significant.

Results

Expression of mrp2 Protein along the Small Intestine.

Western blot analysis of total homogenate and BBM preparations from the most proximal and distal segments is shown in Fig.1A. To compare the relative amounts of mrp2 protein in intestine and liver, a sample of total homogenate and mixed plasma membrane from a normal rat liver and homogenate and BBM from proximal and distal portions of the small intestine were examined (Fig. 1A). The relative content of mrp2 in small intestine was approximately 7 to 10 times higher for BBM than for homogenate; a similar relationship between plasma membrane and homogenate was found for the liver. The expression of the export pump was clearly higher in the proximal intestine, suggesting a gradient along the intestinal tract. To determine the specificity of mrp2 antibody reaction in intestinal membranes, we also examined homogenate and BBM samples from the proximal and distal intestine of TR⁻ rats, which lack expression of mrp2 protein in liver. Neither homogenate nor BBM from TR⁻ rats presented any detectable immunoprecipitation band, thus validating the use of the antibody in intestine.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Western analysis of mrp2 protein content in intestinal preparations. A, a typical gel of homogenate (lanes 1, 3, 5, 7, and 9) and plasma membrane (lanes 2, 4, 6, 8, and 10) from liver (lanes 1 and 2) and proximal (lanes 3 and 4) and distal intestine (lanes 5 and 6) of a normal rat and from proximal (lanes 7 and 8) and distal intestine (lanes 9 and 10) of a TR⁻ rat. B, a typical gel of apical membrane preparations from different segments of small intestine in a normal rat, starting from the most proximal (number 1). Bottom, densitometry of the immunoprecipitation bands is shown, where data represent means ± S.D. (n = 3). a, b, and c, significantly different from the precedent segments (P < .05).

The gradient of mrp2 expression along the intestinal tract was determined by examining the whole small intestine, which was divided into nine equal segments. Alkaline phosphatase activity was measured in BBM and homogenate, and the corresponding ratio was calculated to ensure comparable BBM enrichment in all segments. Although the absolute enzyme activity values decreased from the proximal to the distal intestine (data not shown), the ratio was similar for the different segments (4.5 ± 1.4, n = 27) and comparable to that reported by others using a similar BBM preparation methodology (Arrese et al., 1998). Figure 1B shows mrp2 protein expression throughout the intestine, starting from the pylorus as the most proximal segment. A gradient from the duodenum to the distal ileum was clearly observed, with the proximal jejunum showing the maximal mrp2 content. Densitometric analysis, shown at the bottom of Fig. 1B, confirmed this pattern of distribution.

Expression of mrp2 Protein along the Villus.

A simple method was used to obtain different cellular populations starting from the tip region of the villus and going down to the crypts (Hoensch et al., 1976). Figure 2 shows that mrp2 protein content decreased from the upper to the lower region of the villus and that mrp2 expression was very low for the crypt cells. A 7- to 9-fold enrichment in mrp2 expression in BBM versus homogenate was observed for each of the cellular populations and was similar to that shown previously. Densitometric analysis is shown at the bottom of the same figure. Figure 2 also shows that alkaline phosphatase activity exhibited a villus-crypt gradient as was previously described (Pinkus et al., 1977; Dawson and Bridges, 1981), validating the differential scraping methodology. The ratio of alkaline phosphatase activity in BBM to that in homogenate was similar for the three different cell preparations (5.1 ± 1.8, n = 9).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Western analysis of mrp2 protein content and alkaline phosphatase activity along the villus-crypt axis. Top, a typical gel of the mrp2 content in homogenate (lanes 1, 3, and 5) and in BBM (lanes 2, 4, and 6) prepared from the upper villus (lanes 1 and 2), lower villus (lanes 3 and 4), and crypt (lanes 5 and 6) regions of the proximal intestine. Bottom, densitometry of mrp2 bands and alkaline phosphatase activity determined in the different mucosa layers (n = 3). The data represent means ± S.D. a, significantly different from the upper villus (P < .05). b, significantly different from the upper and lower villus (P < .05).

In Situ Localization of mrp2.

Figure3 shows the results of immunohistochemical staining in different regions of the small intestine. Intense staining was detected at the external surface of the villi of the most proximal segments (Fig. 3, A and B), whereas the immunoreactivity was weaker in the ileum (Fig. 3C). No staining was found on the external surface of villi from the proximal intestine of the TR⁻ rat (Fig. 3E). The data obtained by immunohistochemistry also show that mrp2 protein was preferentially localized in the upper region of the villus (Fig. 3, A and B) with weaker immunoreactivity in the crypt regions (Fig. 3D).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Immunohistochemical analysis of mrp2 expression in the small intestine of normal and TR⁻ rats. A and B show a brown staining on the surface of the villus in the duodenum and jejunum of a normal rat. The staining was stronger at the tip of the villus (see black arrows). C shows a low expression of mrp2 on the surface of a villus from ileum (see black arrows) in a normal rat, whereas the immunostaining is very weak in the crypt region (asterisk) of jejunum of the same rat (D). A villus preparation from jejunum of a TR⁻ rat does not show any staining at the apical membrane (see black arrows) (E). The bars represent 50 μm; magnification was 100 to 150×.

Analysis of mrp2 mRNA Gradient.

Figure4A shows the content of mrp2 mRNA in proximal (segments 1 and 2) and distal (segments 8 and 9) regions of the intestine. Analysis of the data did not reveal statistically significant differences between the two regions. Figure 4B shows the content of mrp2 mRNA in the three different enterocyte populations. Statistical analysis of the densitometry values (see Fig. 4, bottom) indicated significantly higher mrp2 mRNA in the cells of the villus tip.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Northern analysis of mrp2 mRNA content in proximal and distal regions of the intestine (A) and on the three cell populations obtained by differential scraping (B). To correct for differences in total RNA loading and transfer among the lanes, the content of 28S rRNA was also estimated. The densitometric analysis is shown as the ratio of mrp2 to 28S rRNA (mean ± S.D.,n = 3). No significant difference was found between mrp2 expression in the proximal and distal intestine. a, significantly different from the lower villus and crypt regions (P < .05).

Discussion

The contribution of the small intestine to the overall phase II metabolism in the body is well established. The enterocytes not only form an effective and active barrier against xenobiotics present in the luminal chyme but also provide a mechanism to remove compounds from the blood and excrete them into the gut contents. Mulder and collaborators have calculated that the rat intestine extracts approximately 40% of 4-methylumbelliferone from the incoming arterial blood and that subsequent conjugation with glucuronic acid and sulfate can account for the whole intestinal clearance of this compound (Mulder et al., 1984).Koster and Noordhoek (1983) and Wollenberg and Rummel (1985) also demonstrated in rats that immediately after the intestinal conjugation reactions take place, significant amounts of the conjugated products are secreted into the intestinal lumen. Recently, Hirohashi et al. (2000) reported the presence of ATP-dependent transport of glucuronic acid and glutathione conjugates in BBM from human colon adenocarcinoma (Caco-2) cells in association with detection of mRNA encoding mrp2. Studies in Ussing chambers demonstrated greater serosal-to-mucosal versus mucosal-to-serosal flux of 2,4-dinitrophenyl-S-glutathione in Sprague-Dawley versus Eisai hyperbilirubinemic rats (EHBR) deficient in mrp2 (Gotoh et al., 2000). These studies provide additional, although indirect, evidence for the expression of MRP transporters at the apical level of the intestinal tract. We report here for the first time that mrp2 protein is present in the apical membrane of the rat small intestine and follows patterns of distribution along the digestive tract and the villus similar to those observed for phase II enzymes. It is well accepted that glucuronidation, glutathione conjugation, and sulfation are the major routes in intestinal phase II metabolism, affecting a wide variety of compounds including food contaminants, therapeutic agents, peroxides, and free radicals liberated during phase I biotransformation reactions (reviewed in Laitinen and Watkins, 1986). Potentially toxic endogenous compounds that can enter the intestinal cells from the blood or bile can also be secreted by this route. mrp2 expression is thus able to account for the secretory activity of the small intestine involving many of these compounds and is the most likely candidate to explain the active nature of transport of conjugated derivatives across the intestinal wall (Koster and Noordhoek, 1983; Wollenberg and Rummel, 1985; Gotoh et al., 2000). Although the relative content of mrp2 in BBM is clearly lower than that in liver plasma membranes (see Fig. 1), the participation of the small intestine in absorptive and secretory processes may be greatly enhanced by the presence of villi and microvilli, leading to an enhanced surface optimal for transport function. This is particularly relevant in the proximal intestine where villi exhibit maximal development.

Characterization of the distribution of mrp2 mRNA and protein revealed a dissociation between mrp2 protein and mRNA content along the small intestine. Thus, although mrp2 detected by Western analysis differed by 10-fold between the proximal and distal segments of the intestine, mRNA detected by Northern analysis differed only slightly. The presence of mrp2 mRNA in both the proximal and distal regions of the small intestine indicates that the low expression of the protein found in ileum is not a consequence of lack of gene transcription. mrp2 protein content in homogenate from the distal region of the intestine could not be detected, even on loading twice as much BBM protein (data not shown). The immunohistochemical data (Fig. 3) also confirmed in situ the proximal-distal gradient of mrp2 expression and found no staining inside distal enterocytes, so it is unlikely that the lack of detectable mrp2 is due to inability of the protein to migrate to the plasma membrane. Dissociation between the expression of the mrp2 protein and the mRNA content is most likely derived from factors affecting synthesis and/or degradation of this specific protein. Further studies are necessary to clarify this point. Regardless of the nature of the factors accounting for the dissociation between expression of mrp2 mRNA and protein, our data indicate that mRNA analysis of mrp2 expression along the length of the small intestine is not a valid index of protein expression and/or functional capacity.

We also analyzed mrp2 protein content in the different cell populations obtained by differential scraping of the proximal region of the intestine. The data revealed that the villus tip presents the highest protein expression, a finding that was confirmed by immunohistochemistry. A similar pattern of distribution was reported for phase II enzymes (Pinkus et al., 1977; Chowdhury et al., 1985). The distribution of mrp2 along the villus is also in accordance with the higher density of mitochondria found in the villus when compared with the crypt (Jeynes and Altmann, 1975). Thus, the export pump may be acting along the villus-crypt axis in concordance not only with phase II enzymes but also with ATP synthesis. The content of mrp2 mRNA was also found to be the highest at the villus tip region. However, it is not possible to establish whether preferential expression of mrp2 associated with increased gene transcription in the villus is a consequence of the inducing properties of compounds acting preferentially on the upper region of the villus as was suggested for metabolizing enzymes (reviewed in Hanninen, 1986) or merely a consequence of a more general phenomenon associated with cell migration and differentiation.

The MRP proteins include at least six members, all of which belong to the family of ABC transporters (Keppler et al., 1997; Kool et al., 1997). Proteins other than mrp2 may be involved in active transport of conjugated compounds across the apical membrane of the enterocyte. In fact, it was reported that intestinal secretion of 1-naphthol glucuronide in TR⁻ rats is not substantially different from that in the normal rat (deVries et al., 1989). Because the excretion of 1-naphthol glucuronide by the liver of TR⁻ rats was substantially impaired, the authors concluded that transport of the conjugated compound by the liver and intestine involves distinct organ-specific transport systems. In contrast to this finding, Gotoh et al. (2000) reported that after i.v. administration of 1-chloro-2,4-dinitrobenzene to EHBR, intestinal excretion of the corresponding glutathione derivative was substantially decreased with respect to that in normal rats. Based on studies in Ussing chambers, the authors demonstrated that the serosal-to-mucosal flux was the most decreased in EHBR. These data strongly suggest that mrp2 is involved in the secretion of organic anions in the small intestine.

Expression of other members of the MRP family in intestine from human and laboratory animals has been detected by mRNA analysis (Kool et al., 1997), but as we suggest above, this does not necessarily reflect protein level. Development of specific antibodies directed against the different MRPs will clarify the significance of the expression of each of the different export pumps in intestine and, more importantly, whether they are expressed at the apical level of the enterocyte. Thus, although mrp3 mRNA has been detected in small intestine and colon in the rat and human, it is thought by most investigators to be expressed on the basolateral domain of hepatocytes and enterocytes (König et al., 1999; Kool et al., 1999; Gotoh et al., 2000). Recent studies characterizing expression of mrp1 protein in mouse small intestine and colon found mrp1 expressed in the basolateral membranes of intestinal crypt cells (primarily in Paneth cells) but not in differentiated enterocytes (Peng et al., 1999). Furthermore, mrp1 was expressed in all colon cells along the entire crypt-villus axis. These data indicate distinct functions for intestinal mrp1 and mrp2 in spite of their similar substrate specificities (Keppler et al., 1997).

In humans, distribution of phase II enzymes along the intestine does not follow a consistent pattern, so the contribution of each region to the metabolism and subsequent excretion to either the intestinal lumen or the vascular compartment must be separately analyzed for each substrate. Studies analyzing the expression and localization of MRP proteins along the intestinal tract are lacking; however, it is known that mRNA encoding basolateral MRPs (i.e., MRP1 and MRP3) is present in both duodenum and colon, whereas mRNA encoding apical MRPs, such as MRP2, is only detected in duodenum (Kool et al., 1997). Although confirmation of a similar pattern of distribution for the corresponding proteins needs to be addressed, it is likely that compounds undergoing phase II metabolism in the distal intestine (e.g., steroid hormones) would be preferentially reabsorbed and probably excreted in urine, whereas compounds conjugated in the proximal intestine or all along the intestinal tract would be preferentially secreted to the lumen (e.g., bilirubin, 2-naphthol, bulky phenols).

It is now clear that decreases in oral bioavailability resulting from intestinal cytochrome P-450-mediated metabolism of therapeutic drugs, such as cyclosporine, are complemented by active secretion mediated by multidrug resistance 1 (MDR1) P-glycoprotein (Benet et al., 1996). MDR1 P-glycoprotein represents the first ATP-dependent transporter identified in the apical domain of the enterocyte (reviewed in Gatmaitan and Arias, 1993). P-glycoprotein is also located at the apical membrane of the mature enterocyte and is not located in crypt cells (Thiebaut et al. 1987). In contrast to the distribution of P-450, the phase II enzymes, and mrp2, MDR1 expression increases progressively from a low level in stomach, to intermediate levels in jejunum, and to high levels in the colon (Fricker et al., 1996). Clearly, the inducing properties of drugs and food contaminants acting preferentially on the proximal regions of the digestive tract is not the only factor involved in regulation of the preferential localization of the different members of the ABC transporter proteins. Additional factors may be the intestinal metabolism preceding secretion, and particularly for mrp2, the activity of hydrolytic enzymes (i.e., β-glucuronidase, sulfatase) associated with intestinal flora that can convert conjugated metabolites excreted in bile back to the parent compound.

In summary, we report localization of mrp2 in the intestinal BBM. Thus mrp2 represents the second export pump to be localized to the apical surface of the enterocyte. The data demonstrate that mrp2 expression is highest in jejunum and follows a pattern of distribution similar to that of phase II enzymes; it may therefore act coordinately to contribute to first-pass metabolism in the intestine. The actions of mrp2, coupled with the conjugation enzymes, may thus form a barrier to prevent absorption of food contaminants and drugs that enter the enterocytes via the digestive tract and may also decrease the enterohepatic circulation of compounds secreted in bile.