Introduction

Top Summary Introduction Materials & Methods Results Discussion References

An important process of detoxification in human catabolic metabolism is performed by the UGTs localized in the endoplasmic reticulum. UGTs catalyze the formation of hydrophilic glucuronides, which facilitates the elimination of substrates from the body via urine or feces (Dutton et al., 1980; Bock et al., 1987; Mackenzie et al., 1997). Two families of UGT proteins have been defined and termed UGT1 and UGT2 (Burchell et al., 1991). The UGT2 structural genes are located on chromosome 4 (Monaghan et al., 1992). In contrast, the human UGT1A locus is located on chromosome 2, spans >160 kilobase pairs of DNA, with at least 12 individual first exon cassette sequences followed by exons 2-5 at the 3' end of the locus (Moghrabi et al., 1992; Ritter et al., 1992; Mackenzie et al., 1997). The 5' flanking region of each first exon cassette contains appropriate promoter elements. Transcription of each individual first exon leads to a strategy of exon sharing, combining the first exon sequences with common exons 2-5, a mechanism that can generate up to nine functional transferases (Mackenzie et al., 1997). UGT1A proteins eliminate by glucuronidation a broad array of endobiotic and exobiotic substrates, including reactive oxygen products of bioactivation, carcinogens, therapeutic drugs, complex phenolic compounds, and the heme synthetic byproduct bilirubin (Ebner et al., 1993; Ritter et al., 1991; Green and Tephly, 1996; Mojarrabi et al., 1996; Kim et al., 1997; Strassburg et al., 1998).

Investigation of UGT1A expression has focused primarily on the liver, which is considered to be the most important location of human glucuronidation. From liver RNA, five UGT1A cDNAs have been identified and cloned (Harding et al., 1988; Ritter et al., 1991; Wooster et al., 1991; Mojarrabi et al., 1996). It has been demonstrated that hepatocellular tissue expresses without interindividual variation UGT1A1, UGT1A3, UGT1A4, UGT1A6, and UGT1A9 gene transcripts (Strassburg et al., 1997b). We have recently characterized the tissue specific regulation of the human UGT1A locus, which has led to the identification of three additional UGT1A proteins, selectively expressed in extrahepatic tissues (Strassburg et al., 1997b; Strassburg et al., 1998). The identification of UGT1A7 expression in gastric epithelium, UGT1A8 expression in colonic epithelium, and UGT1A10 expression in gastric, biliary, and colonic epithelium but not in hepatocellular epithelium indicates a complex control of this gene locus and emphasizes a unique physiological role of glucuronidation in nonhepatocellular epithelial tissues.

Immediate and prolonged contact with ingested matter requires the stomach to exercise an effective defense against chemical and biological influences. The expression of individual UGT1A proteins in gastric mucosa may determine the stomachs epithelial glucuronidation capacity and may constitute an important factor for extrahepatic metabolism and first pass effects (Bock et al., 1987; Bock and Lilienblum, 1994; Strassburg et al., 1997a; Strassburg et al., 1998). Therefore, variations in drug efficacy and of carcinogenic risk between individuals may be the result of interindividual differences of UGT1A regulation in extrahepatic surface tissue such as gastric epithelium. Because data on the regulation of the UGT1A locus in human gastric epithelium is scarce, we examined UGT1A mRNA regulation by quantitative DRT-PCR combined with UGT activity analysis.

	Materials and Methods

Top Summary Introduction Materials & Methods Results Discussion References

Tissue samples. Gastric tissue samples were obtained from six female (mean age 58.67 ± 16.79 years) and eight male German patients (mean age 56.75 ± 11.68 years) with histologically confirmed gastric adenocarcinoma undergoing gastrectomy at the University of Hannover Medical Center, Hannover, Germany. The panel included 14 tissue sample pairs, each consisting of a sample of tumor tissue and a sample of healthy gastric mucosa from the same resection specimen. In each gastric tissue sample the mucosa was dissected from the underlying muscle layer and only mucosal tissue was used for the analyses. The normal gastric mucosa samples were microanatomically free of any detectable concurrent disease such as gastritis, Helicobacter pylori infection, or dysplasia and were all sampled from the area of the corpus/antrum junction. The patients did not receive chemotherapeutic compounds before resection and no drug therapy of any significance was administered. Patient records indicated absence of tobacco consumption for at least 6 months before sampling. Normal liver tissue samples were collected from 16 patients undergoing hemihepatectomy (n = 14) or liver transplantation (n = 2) for hepatocellular carcinoma (n = 12, 11 men, one woman, mean age 58.08 ± 10.44 years) or focal nodular hyperplasia (n = 4, four women, mean age 37.25 ± 14.48 years) (Strassburg et al., 1997a). Tumor tissue was histologically confirmed and showed no signs of necrosis. In all instances, normal sampling was performed at the distal resection margin of the specimen and exhibited no signs of macroscopic deterioration such as necrosis or any histopathological abnormalities. All tissue was immediately frozen in liquid nitrogen and stored at 80° until use.

RNA Isolation, cDNA synthesis, and RNA purity analysis. Tissue was pulverized under liquid nitrogen and immediately lysed in acidic phenol/guanidinium-isothiocyanate solution (Tripure; Boehringer Mannheim, Mannheim, Germany) as described previously (Strassburg et al., 1997b). RNA Concentrations were determined by spectrophotometry at 260 and 280 nm. Samples were stored in water at 80° until further analysis.

DRT-PCR for UGT1A transcripts. DRT-PCR of a 487-bp fragment of the conserved UGT1A exons 2-5 has been described previously in detail (Strassburg et al., 1997b). Briefly, coamplification was carried out for six cycles of PCR synthesis with UGT1A primers at a concentration of 2 mM. After the addition of -actin primers to a concentration of 0.4 mM, cycling was continued for a total of 32 cycles at 94° for 1 min, 59° for 1 min, and 72° for 1 min. PCR was preceded by a 3-min incubation of the reaction mixture at 94° and was followed by a 7-min elongation at 72°. DRT-PCR products were separated in a 2% agarose gel stained with ethidium bromide. Polaroid (Cambridge, MA) type 665 positive/negative film was used to quantify bands separated in 2% agarose using laser densitometry on a LKB 2222-020 UltroScan XL densitometer (LKB, Bromma, Sweden). Arbitrary units were calculated relative to -actin products according to the following formula: (mean peak area for UGT/mean peak area for -actin) × 100 = relative arbitrary units. Independent and combined linear kinetics for both products during the amplification process were established as previously described in detail (Strassburg et al., 1997b). Statistical analysis was performed using Student's t test from the GraphPad Prism software (GraphPad, San Diego, CA).

Exon-1-specific DRT-PCR. The UGT1A locus predicts the existence of UGT1A1 and UGT1A3-1A10. UGT1A2, -1A11, and -1A12 do not have an uninterrupted open reading frame and are therefore considered to be pseudogenes (Mackenzie et al., 1997). DRT-PCR detection of all nine UGT1A transcripts predicted by the human UGT1A locus was performed using nine exon-1-specific sense primers and two antisense primers located within exons 2-5 or within a common portion of the 3' end of the first exons. As already described elsewhere (Strassburg et al., 1997b), the primers lead to RT-PCR products of distinct molecular sizes: UGT1A1, 644 bp; UGT1A3, 483 bp; UGT1A4, 572 bp; UGT1A5, 659 bp; UGT1A6, 562 bp; UGT1A7, 754 bp; UGT1A8, 514 bp; UGT1A9, 392 bp; and UGT1A10, 478 bp. Briefly, coamplification of UGT1A first-exon and -actin sequences was performed using three cycling protocols: UGT1A1 and UGT1A6: 94° (1 min), 59° (1 min), 72° (1 min); UGT1A3, UGT1A4, UGT1A5: 94° (1 min), 56° (1 min), 72° (1 min); UGT1A7, UGT1A8, UGT1A9, UGT1A10: 94° (1 min), 64° (1 min), 72° (1 min). Each protocol was preceded by a 3-min incubation of the reaction mixture at 94° and was followed by a 7-min elongation at 72°. The specificity and kinetics of this assay have previously been reported (Strassburg et al., 1997b). Experiments were performed in duplicate, and controls were performed without cDNA, primers, or thermophilic polymerase included. Quantification of products by laser densitometry was performed as described above.

Isolation of microsomal protein from tumor and normal gastric tissue. Approximately 200 mg of tissue was pulverized under liquid nitrogen, resuspended in 1 ml of buffer (50 mM Tris·HCl, pH7.4, 10 mM MgCl₂) and homogenized with a Potter-Elvehjem tissue grinder. The tissue homogenate was centrifuged at 10,000 × g for 5 min at 4° in a microcentrifuge and the supernatant was collected. The pellet was resuspended in 0.5 ml of buffer, centrifuged, and the supernatant was collected. The combined supernatants were centrifuged at 150,000 × g for 60 min at 4° in a TL100 ultracentrifuge (Beckman, Palo Alto, CA) and the pellet was resuspended in 0.2 ml of buffer. Protein concentration was determined by the method of Bradford et al. (1976). Microsomal protein was stored at 80°.

UGT enzymatic activity assay. 1-naphthol, 4-nitrophenol, 4-methylumbelliferone, 4-isopropylphenol, octylgallate, estriol, estrone, naringenin, hyodeoxycholic acid (all from Sigma, St. Louis, MO), and 8-hydroxy-benzo[]pyrene were suspended in methanol, 7-hydroxy-benzo[]pyrene was suspended in acetone (-pyrenes from National Cancer Institute Chemical Carcinogen Repository, Midwest Research Institute, Kansas City, MO). Microsomal protein (25 µg) in reaction buffer (50 mM Tris-Cl, pH 7.6, 10 mM MgCl₂) were incubated in the presence of 0.1 mM UDP-glucuronic acid, 0.08 mg/ml phosphatidylcholine, 0.04 µCi of ¹⁴C-labeled glucuronic acid, and 0.1 mM of test substrate for 60 min at 37°. Protein was precipitated by the addition of ethanol and subsequent centrifugation. Lyophilized supernatants were resuspended in 50 µl of methanol and separated by thin layer liquid chromatography with n-butanol/acetone/acetic acid/water (35:35:10:20%) as the running solvent. The production of ¹⁴C-labeled glucuronides was detected by autoradiography. To determine specific catalytic activities the ¹⁴C-labeled glucuronides were quantified by liquid scintillation counting and expressed as picomoles of glucuronide formed per minute per milligram of microsomal protein.

Western blot analysis. Hepatic and gastric microsomal protein (25 µg) were boiled for 90 sec in loading buffer (2% sodium dodecyl sulfate, 62.5 mmol/liter Tris·HCl, pH 6.8, 10% glycerol, and 0.001% bromphenol blue) with 2% -mercaptoethanol and separated by 8% sodium dodecyl sulfate-polyacrylamide gel electrophoresis before electrotransfer onto nitrocellulose membrane. As a control, 15 µg of baculovirus-expressed UGT1A1 protein (Strassburg et al., 1996) was included. Immunodetection was performed according to previously published protocols (Strassburg et al., 1996). UGT1A protein was detected using a rabbit anti-UGT1A antiserum raised against the peptide SSLHKDRPVEPLDLA located between amino acids 441 and 455 of exon 5 of the constant UGT carboxyl-terminal portion (Strassburg et al., 1998). Visualization was achieved with an alkaline phosphatase conjugated goat anti-human IgG diluted at 1:1000.

	Results

Top Summary Introduction Materials & Methods Results Discussion References

Quantitative variability of gastric UGT1A mRNA regulation. Total RNA from gastric carcinoma and corresponding normal tissue was analyzed for the expression of UGT1A transcripts by quantitative DRT-PCR (Fig. 1). A 487-bp fragment from the UGT1A constant exons 2-5 was amplified by RT-PCR and quantified relative to a coamplified 202-bp fragment of -actin, as shown in Fig. 1. Overall UGT1A expression was lower in gastric tissue compared with hepatic tissue. Between the 14 samples of gastric adenocarcinoma and normal gastric epithelium, UGT1A expression was found to be significantly (p = 0.013) down-regulated, indicating that a quantitative differential regulation of UGT1A mRNA exists between hepatic and gastric tissue, but also between gastric adenocarcinoma tissue and healthy gastric epithelium (Fig. 2).

View larger version (36K): [in this window] [in a new window]   Fig. 1.   UGT1A mRNA expression in human gastric epithelium and adenocarcinoma. UGT1A mRNA was detected in normal gastric epithelium and corresponding adenocarcinoma by DRT-PCR. The photograph of an ethidium bromide-stained gel demonstrates an example of UGT1A mRNA expression in two gastric mucosa/adenocarcinoma sample pairs and normal human liver. Quantification was performed by laser densitometry to assess UGT1A mRNA levels in all studied samples. standard, HaeIII digest of X174 DNA.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Graphic representation of the quantitative interindividual UGT1A regulation in hepatic, normal gastric, and gastric adenocarcinoma tissue. UGT1A transcript levels were determined by quantitative DRT-PCR. In 16 samples of hepatic tissue, UGT1A mRNA expression levels were homogeneous among subjects. In normal gastric epithelium and gastric adenocarcinoma tissue, differential quantitative regulation of UGT1A mRNA was evident among individual samples and differential down-regulation was observed in gastric adenocarcinoma. Interindividual differences of UGT1A mRNA regulation in human gastric tissue are preserved despite neoplastic transformation to adenocarcinoma.

Polymorphic regulation of UGT1A isoform expression in human gastric tissue. First-exon sequences encoded by the UGT1A locus share extensive sequence similarity, with 94% similarity among UGT1A7, UGT1A8, UGT1A9, and UGT1A10. Therefore, a transcript specific DRT-PCR assay was employed, capable of detecting all nine predicted UGT1A transcripts (Fig. 3 and Table 1). Gastric mucosa and carcinoma samples were characterized by a unique UGT1A expression pattern. Analysis of the healthy gastric tissue samples revealed UGT1A7 expression in 12 of the 14 samples, whereas UGT1A10 was expressed in all 14 normal tissue samples (Table 1). In contrast, the expression of UGT1A1 was identified in four samples, UGT1A3 in only three samples, and UGT1A6 in five samples. Sequences of the PCR products were confirmed by dideoxy sequencing. The expression of UGT1A4, UGT1A5, UGT1A8, and UGT1A9 was not detected in any of the tissues. The patterns identified in the healthy gastric tissue specimens were confirmed in the corresponding adenocarcinoma tissue samples. In these samples, UGT1A7 and UGT1A10 were expressed in 13 of 14 samples, UGT1A1 in three of 14 samples, UGT1A3 in one of three samples and UGT1A6 in four of 14 samples, whereas UGT1A4, UGT1A5, UGT1A8, and UGT1A9 were also not detected. These findings indicate a polymorphic expression of UGT1A isoforms in human gastric tissue that is present in both normal epithelium and adenocarcinoma tissue from the same subject.

View larger version (66K): [in this window] [in a new window]   Fig. 3.   Polymorphic regulation of UGT1A mRNA in human gastric tissue. DRT-PCR specifically detected the expression of UGT1A transcripts in gastric carcinoma (T, tumor sample) and corresponding normal mucosa (N, normal mucosa) and yielded the PCR products of expected sizes (as described in Materials and Methods). Assays for UGT1A5 and UGT1A8 were negative in all samples and are not shown. A-D, Each panel demonstrates the DRT-PCR results obtained from a single patient tumor/normal tissue sample pair and is an example of different UGT1A isoform expression patterns found between the studied individuals. A, major UGT1A expression pattern found in five of 14 sample pairs. B-D, major UGT1A expression patterns found in two of 14 sample pairs. Up-regulation of UGT1A7 and down-regulation of all other UGT1A transcripts is demonstrated. The same subjects demonstrated in this figure were analyzed for specific gastric microsomal UGT activity and are reported in Table 2. Products were separated in a 2% agarose gel stained with ethidium bromide. LS, molecular length standard.

View this table: [in this window] [in a new window]   TABLE 1 Expression pattern of the UGT1A locus in gastric carcinoma and normal gastric mucosa Studied samples were collected as pairs of tumor tissue and surrounding healthy mucosa from the same organ. Total denotes presence of specific UGT1A transcripts in either tumor or normal sample of the individual patient. Polymorphic regulation of UGT1A1, UGT1A3, and UGT1A6 are noted in human gastric epithelium. In contrast, the analysis of a panel of 16 hepatic RNAs [as reported in Strassburg et al. (1997b)] displays no polymorphic UGT1A expression in human liver.

View larger version (48K): [in this window] [in a new window]   Fig. 4.   Differential down-regulation of the UGT1A transcripts expressed in gastric carcinoma. UGT1A transcripts were quantified by DRT-PCR as outlined in Fig. 2. All comparisons except UGT1A7 were characterized by down-regulation of UGT1A transcripts in gastric carcinoma () versus normal surrounding gastric mucosa ). *, Down-regulation of UGT1A3 (p = 0.049) and UGT1A10 (p = 0.02) was statistically significant (Student's t test). n, number of samples out of the 14 studied that expressed each individual UGT1A isoform (also summarized in Table 1).

Functional polymorphism of UGT catalytic activity in gastric microsomal protein. To assess the functional impact of the observed polymorphic expression of UGT1A gene products in gastric epithelium, four microsomal protein preparations of normal gastric tissue samples expressing different UGT1A transcript patterns were analyzed (Table 2). In Table 2, sample 1 expressed Fig. 3, pattern C, sample 2 expressed pattern B, sample 3 expressed pattern D, and sample 4 expressed pattern A. Microsomal glucuronidation activities were examined with a panel of nine substrates ranging from planar and complex phenols to steroid hormones, flavones, bile acid metabolites, and hydroxylated benzo[]pyrene. Interindividual glucuronidation activity levels varied up to 4-fold among the four tissue protein samples (1-naphthol, 4-nitrophenol, 4-methylumbelliferone), which is in agreement with the quantitative differences found at the transcript level (Fig. 2). Sample 3 displayed the highest level of 1-naphthol, 4-nitrophenol, 4-methylumbelliferone, and 4-isopropylphenol glucuronidation. On the mRNA level, this sample is characterized by an abundant expression of UGT1A7 (compare DRT-PCR analysis of this sample in Fig. 4D). This result is in agreement with the recently described specific activities of UGT1A7 that favor phenolic substrates (Strassburg et al., 1998). Interestingly, hyodeoxycholic acid, which has been described as a substrate of UGT2B4 (Jackson et al., 1987; Fournel-Gigleux et al., 1991) and has not been found to be glucuronidated by UGT1 proteins was glucuronidated by one of the four samples (Table 2, sample 1). Taken together, these data provide evidence for the polymorphic regulation of the human UGT1A locus at the RNA transcript and functional levels, in addition to demonstrating that UGT2B proteins are potentially regulated in a polymorphic fashion in human gastric tissue.

Detection of UGT1A protein in gastric microsomal fractions. Overall UGT1A protein was detected with an antibody directed against a 15-mer epitope within exon 5 of the common UGT1A carboxyl-terminal portion. Normal liver microsomal protein was compared with a pair of gastric adenocarcinoma/normal gastric mucosa microsomes. In agreement with the observed differences of UGT1A isoform mRNA expression detected by DRT-PCR between the two tissues, band mobility was unique to each tissue type (Fig. 5, lanes 3-5). In addition, the analysis of pairs of adenocarcinoma and normal gastric mucosa tissue demonstrated lower levels of overall UGT1A protein in the adenocarcinoma samples. This finding is in agreement with the results obtained by DRT-PCR on the mRNA level (Fig. 1 and Fig. 2) indicating that down-regulation of UGT1A mRNA leads to down-regulation of UGT1A protein in these samples.

View larger version (39K): [in this window] [in a new window]   Fig. 5.   Western blot analysis of UGT1A protein in hepatic and gastric microsomal protein. Normal microsomal protein from liver and gastric microsomal protein show a tissue-specific banding pattern for each tissue, which is consistent with the differential expression of UGT1A isoform mRNA detected by DRT-PCR. The pattern is consistent among gastric adenocarcinoma and normal mucosa samples. However, differential quantitative expression of UGT1A mRNA detected by DRT-PCR (Fig. 1 and 2) results in down-regulation of UGT1A protein in gastric adenocarcinoma compared with the surrounding normal tissue (two right lanes). Recombinant UGT1A1 (Strassburg et al., 1996) was included as positive control (rec. UGT1A1).

	Discussion

Top Summary Introduction Materials & Methods Results Discussion References

Human UGT1A genes are regulated in a tissue-specific fashion in hepatic and extrahepatic tissue (Strassburg et al., 1997a; Strassburg et al., 1998). This distribution of individual UGT1A isoforms is believed to determine tissue specific metabolic and detoxification requirements. Although most UGT1A cDNAs have been cloned and characterized using hepatic tissue (Harding et al., 1988; Ritter et al., 1991; Wooster et al., 1991; Mojarrabi et al., 1996), the regulation and function of extrahepatic UGTs is less well established. As an organ of the digestive tract characterized by prolonged and immediate contact to dietary constituents, including xenobiotics, carcinogens, and therapeutic drugs, human gastric epithelium is likely to play a significant role in extrahepatic detoxification (Peters et al., 1989; McDonnell et al., 1996). Previous studies have identified UGT1A1 and UGT1A4 expression in the extrahepatic gastrointestinal tract (McDonnell et al., 1996) in addition to significant bilirubin UGT activity in small intestine (Peters et al., 1989).

In the present study, we provide evidence for a polymorphic regulation of the UGT1A locus in gastric mucosa and carcinoma. In a panel of 14 normal gastric epithelium/adenocarcinoma tissue sample pairs that were examined, UGT1A7 was expressed in 13 and UGT1A10 was expressed in all samples, confirming UGT1A7 and UGT1A10 expression as the predominant extrahepatic UGT1A gene products in human gastric tissue (Strassburg et al., 1997b). However, in contrast to findings obtained with a panel of hepatic tissue samples, gastric tissue was not characterized by a uniform presence of UGT1A gene products consistent among individuals. Analysis of the overall UGT1A mRNA in gastric epithelium displayed significant interindividual differences. In addition, the expression of UGT1A1 mRNA was observed in four of 14 samples, UGT1A3 mRNA in three of 14 samples, and UGT1A6 mRNA in five of 14 samples. This finding documents a polymorphic expression of the UGT1A locus within a single tissue type. Because the history of the sampled subjects indicates no relevant drug or nicotine use, and all samples were obtained from the area of the corpus/antrum junction, these findings are not likely to be the result of sampling or external factors. In addition to examining normal gastric epithelium, corresponding specimens of adenocarcinoma tissue from the same organ were analyzed for UGT1A mRNA expression. The polymorphic expression of UGT1A mRNAs was confirmed. Moreover, differential down-regulation of UGT1A1, UGT1A3, UGT1A6, and UGT1A10 was observed in the tumor specimens. Interestingly, UGT1A7 mRNA was up-regulated in gastric adenocarcinoma.

The results also demonstrate that in normal gastric epithelium and paired adenocarcinoma samples, there occurs significant differential regulation of UGT1A gene products. This regulation includes the consistent expression of predominant and tissue-specific UGT1A gene products such as UGT1A7 and UGT1A10, the variable coexpression of individual UGT1A genes such as UGT1A1, UGT1A3, and UGT1A6 in selected individuals, as well as down-regulation and up-regulation of UGT1A gene expression in the course of neoplastic transformation to gastric adenocarcinoma. Combined, the data provides compelling evidence for individual regulation of the human UGT1A gene products in gastric tissue.

Metabolism by glucuronidation in human gastric tissue may therefore be determined by a regulatory polymorphism that may potentially represent the molecular biological basis of interindividual differences in gastric UGT activity. These differences could impact on interindividual rates of glucuronidation of drugs that could potentially be absorbed through the gastrointestinal lining. With few exceptions, all drugs can be classified as uncharged drugs, organic acids, or organic amines. In an acidic environment such as that of the stomach, amines would exist predominately as ionized and protonated molecules that are not readily accessible for transport into the gastric epithelium. It is interesting to note that UGT1A4, which has specificity toward the metabolism of tertiary amines to form quaternary ammonium glucuronides (Green and Tephly, 1996), is not expressed in gastric epithelium. However, other agents, such as weak acids, would be protonated, uncharged, and would have a high oil/water partition coefficient, allowing for efficient absorption, thus serving as excellent candidates for glucuronidation in the stomach. The interindividual variation observed with the expression of UGT1A1, UGT1A3, UGT1A6, UGT1A7, and UGT1A10 could potentially impact on drug disposition as a result of variations in first-pass metabolism through the gastric mucosa. The variations observed in stomach are in contrast to the consistent UGT1A mRNA regulation in human liver (Strassburg et al., 1997a), which further highlights the differences between hepatic and extrahepatic regulation of the UGT1A locus.

The polymorphic regulation of a group of xenobiotic and drug metabolizing enzymes is an interesting finding in view of the stomach's physiological role as a first site of prolonged immediate contact with exobiotic compounds. These differences could potentially modify the cellular defense potential of the stomach against diet-borne cytotoxic and mutagenic compounds, thereby affecting gastric cancer predisposition (Kim et al., 1997; Strassburg et al., 1997a). The analysis of the catalytic ability of gastric microsomal protein to glucuronidate different substrates was exploited to further elucidate a functional impact of the discovered polymorphic regulation. In normal gastric microsomal samples, a 2- to 4-fold difference in specific activities was noted. In particular, phenolic 4-nitrophenol, 4-methylumbelliferone, and 1-naphthol UGT activities catalyzed by UGT1A proteins (Ebner et al., 1993; Ebner and Burchell et al., 1993; Green and Tephly, 1996; Burchell and Brierley 1998; Strassburg et al., 1998), were highly variable. The considerable overlap of specific substrate activities of the majority of UGT1A proteins prevents an exact assignment of specific activity to an individual UGT1A isoform when several UGT1A proteins are simultaneously expressed in a tissue. This is exemplified by the substrate estrone, which has been found to be glucuronidated by UGT1A3 (Mojarrabi et al., 1996) and UGT1A10 (Strassburg et al., 1998). Because expression of both isoforms was demonstrated in gastric tissue in this study, their individual contribution to microsomal estrone activity cannot be estimated by catalytic activity assays using microsomal protein preparations. However, the activity of UGT1A7, which has been found to be uniquely expressed in stomach, was found to be associated with higher specific activities for phenolics such as 1-naphthol, 4-methylumbelliferone, and 4-nitrophenol (Fig. 4D; Table 2, sample 3), all of which have recently been identified as preferred substrates of UGT1A7 (Strassburg et al., 1998). The panel included the putative tobacco carcinogen metabolite 7-hydroxy-benzo[]pyrene (Strassburg et al., 1998), which was also glucuronidated with specific activities that varied 4-fold. This may contribute to interindividual differences in cytoprotection and genoprotection. Interestingly, a qualitative differential activity was demonstrated for the glucuronidation of hyodeoxycholic acid. Because hyodeoxycholic acid has not been identified as a substrate for the UGT1A gene products but has been found to be conjugated by UGT2B4 (Fournel-Gigleux et al., 1991), the polymorphic patterns of catalytic activities detected here are very likely to include UGT2B proteins.

The findings presented in this study differ from the classical genetic polymorphisms of drug metabolizing enzymes, which have been documented for such cytochrome P450 enzymes as CYP2D6, as well as glutathione S-transferase, N-acetyltransferases, and methyltransferases (Daly, 1995; Meyer and Zanger, 1997). Most genetic polymorphisms have been discovered as a result of bifunctional distribution of drug metabolizing enzyme function, which becomes clinically evident in adverse drug reactions. The genetic basis of this feature is the presence of monogenic traits in the normal population, which is represented as two phenotypes. As a result, drug metabolizing enzyme function can be substantially altered by homozygous combination of "loss of function alleles" or enhanced by duplication or amplification events. Other than in the investigation of Crigler Najjar's and Gilbert's diseases (Mackenzie et al., 1997) genetic polymorphisms have not been documented for the majority of UGT proteins. We report here a differing observation leading to polymorphic expression of UGT1A gene product levels and isoform expression, not following a bimodal distribution. This seems to be the result of tissue-specific polymorphic regulation of the human UGT1A gene locus in gastric tumor as well as in normal gastric epithelium. In contrast to the homogeneous expression of UGT1A gene products in human hepatic tissue, this also implies that interindividual regulatory mechanisms may affect the function of drug metabolizing enzymes in the stomach and thus determine interindividual variations in metabolism.

In summary, a quantitative and qualitative polymorphism of UGT1A gene regulation in human gastric epithelium has been demonstrated. Interindividual differences in UGT activity included a 4-fold quantitative variation and the qualitative variations of specific substrate activities. The regulation of UGT1A transcripts in gastric tissue follows a complex mode different from that in liver, which may indicate a unique control of human extrahepatic glucuronidation.