Regional variations in ABC transporter expression along the mouse intestinal tract

David M. Mutch1,2, Pascale Anderle3, Muriel Fiaux1, Robert Mansourian1, Karine Vidal1, Walter Wahli2, Gary Williamson1 and Matthew-Alan Roberts4

1 Nestlé Research Center, CH-1000 Lausanne 26
2 Center for Integrative Genomics, Université de Lausanne, CH-1015 Lausanne
3 Swiss Institute for Experimental Cancer Research, 1066 Epalinges s/Lausanne, Switzerland
4 Nestle Purina Pet Care, St. Louis, Missouri 63164


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIAL AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The ATP-binding cassette (ABC) family of proteins comprise a group of membrane transporters involved in the transport of a wide variety of compounds, such as xenobiotics, vitamins, lipids, amino acids, and carbohydrates. Determining their regional expression patterns along the intestinal tract will further characterize their transport functions in the gut. The mRNA expression levels of murine ABC transporters in the duodenum, jejunum, ileum, and colon were examined using the Affymetrix MuU74v2 GeneChip set. Eight ABC transporters (Abcb2, Abcb3, Abcb9, Abcc3, Abcc6, Abcd1, Abcg5, and Abcg8) displayed significant differential gene expression along the intestinal tract, as determined by two statistical models (a global error assessment model and a classic ANOVA, both with a P < 0.01). Concordance with semiquantitative real-time PCR was high. Analyzing the promoters of the differentially expressed ABC transporters did not identify common transcriptional motifs between family members or with other genes; however, the expression profile for Abcb9 was highly correlated with fibulin-1, and both genes share a common complex promoter model involving the NF{kappa}B, zinc binding protein factor (ZBPF), GC-box factors SP1/GC (SP1F), and early growth response factor (EGRF) transcription binding motifs. The cellular location of another of the differentially expressed ABC transporters, Abcc3, was examined by immunohistochemistry. Staining revealed that the protein is consistently expressed in the basolateral compartment of enterocytes along the anterior-posterior axis of the intestine. Furthermore, the intensity of the staining pattern is concordant with the expression profile. This agrees with previous findings in which the mRNA, protein, and transport function of Abcc3 were increased in the rat distal intestine. These data reveal regional differences in gene expression profiles along the intestinal tract and demonstrate that a complete understanding of intestinal ABC transporter function can only be achieved by examining the physiologically distinct regions of the gut.

ATP-binding cassette transporters; microarray; intestine; anterior-posterior axis


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIAL AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
THE INTESTINAL BARRIER PRESENTS a formidable obstacle to the absorption of both medicinal and nutritional compounds alike. Whereas small hydrophobic molecules and ions may pass freely from the lumen into the gastrointestinal enterocytes based on concentration gradients, hydrophilic and larger molecules require the assistance of transporters for their uptake into intestinal cells. To date, the ATP-binding cassette (ABC) proteins have been identified as the largest known family of transmembrane transporters (6).

ABC transporters are multispan membrane proteins that are involved in the directional transport of a wide variety of substrates, including sugars, amino acids, glycans, sterols, phospholipids, peptides, proteins, toxins, antibiotics, and xenobiotics across biological membranes (11). Family members have been identified and grouped based on conserved sequence motifs within their ATP-binding domains (also referred to as nucleotide binding domains), which lie between the characteristic Walker A and B motifs (31). Based on structural similarities and the sequence homology of these nucleotide-binding domains, the ABC superfamily has been further subdivided into seven subfamilies, ABCA to ABCG (7). Furthermore, ABCs are highly conserved between species, reinforcing their functional importance in the transport of molecular compounds. A thorough review by Dean (6) was recently published and described the structure, molecular organization both within and between species, and the known functions of ABC transporters and how they may contribute to the onset of disease. Within this review (6), ABC family homology was compared and contrasted between the mouse and human. To date, 49 ABC family members have been identified in the mouse, and most of these show high concordance with their corresponding human ortholog. The differences consist of: 1) an additional ABCG subfamily member in the mouse (Abcg3), 2) duplication of ABCB1 and ABCA8 in the mouse (Abcb1a/b and Abca8a/b, respectively), and 3) loss of Abca10 and Abcc12. Additionally, a cluster of three murine ABCA subfamily members has been identified in the mouse genome, but these have not been assigned human orthologs because of the incomplete characterization of this region of the human genome (6).

As a first step in elucidating their functions in eukaryotes, it is critical to ascertain where in the body these transporters are expressed. A recently published study by Langmann and colleagues (17) described a whole body gene transcript characterization of all currently known human ABC transporters using real-time PCR (RT-PCR). The authors analyzed and revealed the expression profiles of these genes in 20 different tissues and concluded that tissues involved in secretory function (adrenal gland), metabolic function (liver), barrier function (small intestine), and development (uterus, testis) had high levels of ABC transporter transcripts. The authors did not subdivide the small intestine into its functionally different sections: duodenum, jejunum, and ileum. These various sections have been previously shown to have different rates for the transport-mediated uptake of orally delivered compounds in the rat intestine (21, 27). Furthermore, the bioavailability of both medicinal and nutritional compounds is dependent on the intestinal environment, where factors such as disease, pH, motility, and microflora can vary along the intestinal tract (9). As ABC transporters are highly involved in the transport of nutrients and drug compounds alike, we propose that expression patterns are a prerequisite to deciphering their functions in the gastrointestinal tract (GIT). The mouse was used to generate statistically significant and biologically relevant information for this family of genes because of the high degree of homology and conservation of ABC transporters between eukaryotes and their potential to generate specific knockout models for further analysis.

Recently, Bates and colleagues (1) conducted a study examining the global differential gene expression along the anterior-posterior (A-P) axis of the adult mouse GIT. The authors used the mouse GEM1 cDNA microarray from Incyte Genomics, which contains ~8,000 sequence-verified expression sequence tags (ESTs), to identify novel genes and functional relationships in the GIT (1). Several ABC transporters (16 of 49) could be identified in their data set and indicated that some of these genes are differentially expressed along the intestinal tract; however, the majority of ABC transporters are either not annotated or not present on the mouse GEM1 array.

A more complete picture of murine intestinal mRNA expression patterns is obtained using high-density oligonucleotide microarray technology. The present work extracts all ABC transporters currently annotated and identifiable in the MuU74v2 GeneChip set (43 of 49 transporters currently annotated) for a focused discussion on this important molecular family. Despite both the differences in mouse strains (Hsd:ICR vs. C57BL/6) and array technology (Affymetrix vs. cDNA) used in this study vs. the Bates study (1), concordance was high for the 16 ABC transporters found in common, reinforcing the highly conserved nature of this transporter family. Our findings indicate that 8 of the 43 ABC transporters examined are differentially regulated along the murine intestinal tract. Furthermore, examining the promoter regions and the cellular location of differentially expressed ABC transporters revealed complementary information required for elucidating the functions of these transporters in the intestinal tract.


    MATERIAL AND METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIAL AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Animals and tissue handling.
Eight-week-old male Hsd:ICR(CD-1) mice (Harlan, Netherlands) were provided by AMS Biotechnology (Lugano, Switzerland). Mice were feed ad libitum a standard diet (Harlan Teklad diet 2018S) and housed in groups of five. Animals were then divided into three pools (n = 10) and euthanized. The small intestine was extracted and divided into three sections, where the first 2–3 cm after the stomach comprised the duodenum and the middle third the jejunum, and the section before the ileo-ceco-colic junction comprised the ileum. Colon was not divided into proximal and distal sections and was treated as a single intestinal section. Tissues destined for immunohistochemistry were rinsed in physiological buffer, cut longitudinally, embedded in Tissue-Tek (Sakura, Netherlands), frozen in liquid nitrogen, and stored at -80°C. The handling and killing of animals obtained by AMS Biotechnology was compliant with federal, state, and local laws and regulations, and in accordance with the Institute for Laboratory Animal Research (ILAR) "Guide for the Care and Use of Laboratory Animals."

Nucleic acid preparation.
Total RNA extracts were provided by AMS Biotechnology (Lugano, Switzerland), and RNA extraction was performed identically for each pool of mice. Within a pool, the duodenum, jejunum, ileum, and colon were harvested from each animal, pooled, and homogenized, and the RNA was extracted by a modified guanidium thiocyanate method (RNWAY Laboratories, Seoul, Korea) as previously described (28). Although pooling mRNA samples is commonly performed because of low quantities of mRNA and/or cost restrictions, recent evidence indicates that pooling RNA also provides equivalent statistical power compared with individual analyses (13, 24). To ascertain the condition of the RNA pools, RNA quality was tested by formaldehyde gel electrophoresis and was stored and transported on dry ice in an ethanol suspension containing 0.1 M sodium acetate. RNA was then repurified, according to manufacturer’s instructions, using the Nucleospin kit, and contaminating genomic DNA was removed with by DNase I treatment (Macherey-Nagel, Oensingen, Switzerland). All the samples were monitored by agarose gel and with the Agilent 2100 Bioanalyser (Agilent Biotechnologies, Germany) and consistently demonstrated high-quality RNA (28S/18S ratio approximately 2, but always less than 3).

cRNA preparation.
For each gut tissue section, 5 µg total RNA was used as the starting material for all individual samples. cRNA was prepared as previously described for all experimental samples except two (R. Mansourian, unpublished observation). Because of low yields in cRNA synthesis with two of the three jejunum samples, it was necessary to pool the cRNA from multiple synthesis steps. Corresponding data analysis [i.e., counting the number of outliers identified using both the global error assessment (GEA) model and the rudimentary 2- and 3-fold cutoff rules] indicated that this did not have any significant effects on the variability in gene expression measurements between jejunum samples compared with the other intestinal regions (data not shown).

Array hybridization and scanning.
Samples were hybridized to the Affymetrix MuU74v2 set (Affymetrix, High Wycombe, UK), which consists of three GeneChips (A, B, and C) containing ~33,000 unique genetic elements. For all experimental replicates, the same cRNA sample was hybridized to each of the three GeneChips. No detrimental effects on the quality of cRNA were observed after three hybridizations, as determined by comparing the variability in intensity signals to the intensity signals of all genes present on each GeneChip (data not shown). Scanning was performed as previously described (23). Readings from quantitative scanning were analyzed with Affymetrix Gene Expression Analysis Software (MAS 5.0). The complete data set is publicly available in the NCBI Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) through the accession number GSE849.

Data analysis.
Differential gene selection was determined using two statistical approaches: 1) a classic ANOVA and 2) a GEA method of analysis. ABC genes considered differentially regulated along the intestinal tract were statistically significant ({alpha} <= 0.01) by both the classic ANOVA and the GEA model.

The GEA approach takes advantages of several inherent characteristics of microarrays that result in an increased statistical power for the data analyst. When examining genomic data sets, most genes are found to be stably expressed across all conditions (14, 23). Therefore, rather than treating each gene on the microarray as a unique and unrelated element, neighboring genes are binned into groups of 200 based on similar intensity signals, and the mean squared error is calculated for each bin. The binning of genes greatly increases the statistical power of the GEA approach and yields results that are statistically significant and highly concordant with an alternate gene expression analytical platform (i.e., RT-PCR). For a detailed description and general protocol of the GEA model, we encourage readers to contact us.

Real-time polymerase chain reaction.
All mouse primer and probe sets were synthesized by Applied Biosystems (Foster City, CA) through their Assays on Demand (AoD) and Assays by Design (AbD) services. For those primer/probe sets created using the AbD service, complete sequences were obtained from the Ensembl database (http://www.ensembl.org/), and target sequences were selected over exon splicing sites to minimize any potential signal stemming from contaminating genomic DNA (Table 1). AbD and AoD primer/probe solutions were received at a premixed, ready-to-use concentration of 18 µM for each primer and 5 µM for the probe.


View this table:
[in this window]
[in a new window]
 
Table 1. Sequences of TaqMan primer/probe sets

 
RNA corresponding to the samples hybridized to microarrays was used, as RT-PCR measurements with pooled mRNA samples have been demonstrated to have a reduced variability and yield results comparable to those produced via individual measurements (13). Reverse transcription was performed with 2 µg of total RNA using the first-strand cDNA synthesis kit for RT-PCR (AMV; Roche Biomedical, Basel, Switzerland) using an oligo-d(T)15 primer. RT-PCR amplification was performed using an ABI 5700 machine (Applied Biosystems) with the following thermal cycling conditions: 2 min at 50°C, 10 min at 95°C, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min for denaturation, annealing, and elongation. All samples were performed in duplicate. Cycle-to-cycle fluorescence emission was monitored and quantified using the GeneAmp software provided by Applied Biosystems. Data was normalized to Gapdh, which both GEA and the classic ANOVA did not identify as differentially expressed (Table 2). A two-tailed, homoscedastic Student’s t-test ({alpha} = 0.01) was used to confirm differences in gene expression in pair-wise analysis (i.e., all four intestinal sections compared).


View this table:
[in this window]
[in a new window]
 
Table 2. ABC transporter gene expression along the intestinal tract

 
Comparing high-density oligonucleotide and cDNA microarray analysis of ABC transporters.
The data set by Bates and colleagues (1), herein referred to as the "cDNA" data set, was treated in the following manner: 1) results corresponding to the proximal and distal colon sections were averaged to produce a mean expression value for the various transporters in the colon, and 2) ratios were averaged for each individual technical replicate in instances where multiple cDNA probes existed for a single ABC transporter (as was the case for 5 of the 16 transporters). Trends along the gut were of primary importance, and statistical significance was used as a means to improve one’s confidence in the observed trends. The cDNA data set had an n value of 2, which is not optimal to establish statistical significance using classic methods; however, values were considered significantly different if these had a P value of <=0.05 using a Student’s t-test.

Promoter analysis.
Genes with expression profiles highly correlated (r = 0.90–0.99) with one of the eight differentially expressed ABC transporters were selected for promoter analysis. The putative human homolog corresponding to each gene was identified according to the Ensembl annotation. Using the publicly available promoter analysis tools Chip2Promoter, MatInspector, and FrameWorker by Genomatix (http://www.genomatix.de), known or putative promoter regions were identified and checked for both single promoter elements (MatInspector) and complex models (FrameWorker) reflecting both the promoter element composition and the functional organization of individual elements (15, 29). In all circumstances, promoter regions were defined by Genomatix software as the 500 base pairs (bp) upstream and the 100 bp downstream of the predicted transcript start site. Complex models were identified by the FastM method (15), which combines a search algorithm for individual transcription factor binding sites with a distance correlation. Only single elements or complex models were selected if they were common to both the murine and human sequence, as sequence conservation in noncoding, upstream regions of orthologous genes from man and mouse is likely to reflect common regulatory DNA sites (8). The distance between elements was required to be in the range of 10 to 100 base pairs. All remaining parameters used Genomatix defaults.

Immunohistochemistry.
Serial tissue sections of 5 µm were obtained using a cryostat (model HM 500 OM; Microm Laborgeräte, Walldorf, Germany), air dried, and fixed in acetone for 10 min. Slides were air dried and placed in 1x PBS until use. All future rinses were performed using 1x PBS. Endogenous peroxidase was blocked with 0.3% H2O2 (Merck Eurolab, Switzerland) and rinsed for 5 min. Biotin and avidin blocking was performed according to manufacturer’s instructions (DAKO, Carpinteria, CA). Slides were then blocked with 10% normal rabbit serum (Sigma, Buchs, Switzerland). Abcc3 protein was detected by incubating 1 µg/ml goat anti-mouse Abcc3 antibody (Santa Cruz, Heidelberg, Germany) in 10% normal rabbit serum for 1 h at room temperature, which was then rinsed with gentle agitation. Biotin-conjugated rabbit anti-goat antibody (Zymed Laboratories, S. San Francisco, CA) was then incubated at 2 µg/ml in 10% normal rabbit serum for 1 h at room temperature. After rinsing with gentle agitation, slides were treated with the TSA fluorescence system (PerkinElmer, Boston, MA) according to the manufacturer’s instructions. Controls included omission of the primary antibody or inclusion of an isotype control antibody (goat anti-mouse IgG, Santa Cruz).

Tissues were mounted with fluorescent mounting medium (DAKO), sealed, and stored at 4°C in the dark. All images were obtained using a Hamamatsu camera (Hamamatsu Photonics) and a Zeiss Axioplan II microscope (Zeiss, Feldbach, Switzerland).


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIAL AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
ABC transporter mRNA expression along the intestinal tract.
To elucidate the expression profiles of ABC transporters along the A-P axis of the GIT, murine RNA samples corresponding to the duodenum, jejunum, ileum, and colon were hybridized to Affymetrix MuU74v2 GeneChips. To date, 43 of 49 murine ABC transporters have been annotated by Affymetrix and are located across the three GeneChips. Those ABC transporters not yet annotated or present on the GeneChips are Abca8, Abca12, Abca13, Abcb5, Abcb8, Abcc4, and Abcc11.

Differential gene expression was determined primarily using the GEA method, as classic statistical methods (ANOVA, Student’s t-test) often do not prove to be robust using three replicates (R. Mansourian, unpublished observation). It has previously been demonstrated that the GEA approach can accurately identify differentially expressed genes at both low and high numbers of replication. To reinforce this point, GEA selected seven additional transporters (Abca2, Abca7, Abcb1a, Abcb4, Abcb10, Abcg1, and Abcg2) as differentially expressed in the gut, whereas the classic ANOVA attributed much less significance to these results (Table 2).

The natural logarithmic values of normalized expression results for all ABC transporters are listed in Table 2. It is interesting to note that, at the mRNA level, most of the ABC transporters are not differentially expressed along the intestinal tract. Indeed, only eight transporters were identified as differentially expressed in the gut by both the classic ANOVA and GEA ({alpha} < 0.01 for both statistical tests). Differentially expressed transporters were Abcb2, Abcb3, Abcb9, Abcc3, Abcc6, Abcd1, Abcg5, and Abcg8. When visualizing all ABC transporters on a bivariate plot comparing the small intestine (duodenum, jejunum, and ileum values averaged to obtain a single value) to the colon, where the mean of the natural logarithm of the fold change (M) is plotted against the mean expression value (A), one can immediately observe that the great majority of these transporters (identified by green dots) lie within the data cloud (Fig. 1). Additionally, estimation of a Lowess (also known as "Loess") regression function predicting the local mean standard deviation demonstrates that the majority of these transporters are not differentially expressed. Furthermore, as the x-axis is representative of absolute expression levels, it is clear that ABC transporters are present in the intestinal tract at various levels, from lowly or not at all expressed up to highly expressed. The aforementioned eight differentially expressed ABC transporters are indicated and place their degree of differential expression in context with all genetic elements present on the GeneChips.



View larger version (31K):
[in this window]
[in a new window]
 
Fig. 1. Comparing the variation in expression levels of ABC transporters in the small intestine and colon. Blue dots represent all genes on the MuU74A GeneChip set. Green dots represent the 43 ABC transporters annotated on the GeneChips. Overlaying red dots identify those ABC transporters identified as differentially expressed. Green and red lines show the estimation of a Lowess regression function predicting the local mean standard deviation (green line = estimated 3 x standard deviation; red = estimated 2 x standard deviation).

 
Real-time PCR validation of microarray results.
A set of eight ABC transporters was further validated by TaqMan RT-PCR, as is customary when building confidence in a transcriptomic data set (Table 3). The transporters selected for validation displayed one of the following trends: 1) no change along the intestinal tract (Abca1, Abcc1, Abcc6, and Abcd3), 2) an increase along the intestinal tract (Abcb1a), or 3) a decrease along the intestinal tract (Abcd1, Abcg5, and Abcg8). Overall, the concordance between the two techniques indicated that trends seen in microarray data could also be seen with RT-PCR, e.g., an increase in gene expression along the A-P axis could be seen with both techniques. Because of the greater dynamic range attainable with RT-PCR, this technique is more often used as a means to confirm trends in microarray data rather than duplicate the fold changes seen with GeneChip experiments (13, 23, 32). Therefore, it was not surprising to see that discrepancies in the two data sets arose when examining the statistical significance of fold changes (for both RT-PCR and microarray data) in gene expression levels of the jejunum, ileum, and colon in relation to the duodenum. Microarray data indicated that many of the gene changes observed were not found to be statistically different ({alpha} < 0.01), even in circumstances where a fold change of 4.7 (Abcc6 in the jejunum) is seen. In contrast, RT-PCR was able to identify additional statistically significant differences in situations where the microarray was not, such as the 3.8-fold change observed for Abcc6 in the jejunum. Findings such as these were expected because of the inherent differences in sensitivities between the two methods (33). Indeed, identifying 100% of the truly differentially regulated genes in a microarray experiment is still complicated by lowly expressed genes (i.e., corresponding to many transcription factors and receptors), which may be highly variable within and/or between biological treatments (19). The development of robust statistical methods, such as GEA, aim to dissociate those lowly expressed genes that are variable within a treatment, i.e., technical variability, from those that are variable between experimental conditions, i.e., biological variability (R. Mansourian, unpublished observation; 19). As demonstrated in the current study, GEA methodology proved to have increased sensitivity compared with the classic ANOVA for gene selection. Whereas GEA identified Abcb1a and Abcd1 as differentially expressed, the classic ANOVA failed to assign statistical significance to the changes in expression (Table 2). RT-PCR confirmed that these genes were indeed differentially expressed in the gut, reinforcing the improved sensitivity of the GEA method for the detection of low-abundance genes. Finally, in agreement with the decreased sensitivity of microarrays, the results demonstrate that microarray platforms tend to underestimate the relative changes in mRNA expression between samples (33).


View this table:
[in this window]
[in a new window]
 
Table 3. Comparison between RT-PCR and microarray gene expression data for eight ABC transporters along the intestinal tract

 
Comparing high-density oligonucleotide and cDNA microarray analysis of ABC transporters.
Comparisons with a publicly available cDNA data set produced by Bates and colleagues (1) provided an in silico means to further compare and validate our Affymetrix data set. As previously described in the literature, the hybridization specificity of spotted cDNA vs. oligonucleotide probe sets for target transcripts is still being explored (33); however, sufficient similarities remain to make the comparison useful. Despite the different strain of mouse used to obtain the cDNA data set, mice were killed at a similar age and consisted of only male animals.

The cDNA data set contained expression information for 16 of the 49 ABC transporters. Those ABC transporters found in common between the two data sets are identified with an asterisk in Table 2. Of the eight ABC transporters identified in our data set as differentially regulated along the gut, only two were present in the cDNA data set: Abcd1 and Abcg5. Figure 2 demonstrates that despite the experimental differences, both cDNA and Affymetrix platforms indicate that Abcg5 is expressed at significantly lower levels in the colon and similarly expressed in the duodenum, jejunum, and ileum. This finding is in agreement with RT-PCR data (Table 3). Expression-profiling trends revealed that Abcd1 is most highly expressed in the jejunum and most lowly expressed in the colon (Fig. 2), as confirmed by RT-PCR. The remaining 14 ABC transporters found in the cDNA data set were not differentially expressed along the intestinal tract. This agrees with our high-density oligonucleotide results and further specifies that the members of this family are expressed at various levels in the intestine and that most of these transporters are not differentially expressed.



View larger version (11K):
[in this window]
[in a new window]
 
Fig. 2. Comparison of differentially expressed ABC transporters between high-density oligonucleotide and cDNA microarray platforms. The two differentially expressed genes in common between data sets were examined. Solid lines represent Affymetrix expression profiles, and dotted lines represent the expression profiles in the cDNA data set. The trend for Abcg5 expression indicates low expression levels in the colon and similar levels in the small intestine (both open and solid triangles on black lines). The trend for Abcd1 is identical between the two data sets (both open and solid squares on gray lines). ln ADI, natural logarithm of the average difference intensity (represents the absolute expression of a gene).

 
Promoter analysis of differentially expressed ABC transporters.
Clustering analysis of differentially regulated genes can reveal numerous genes known to be regulated by specific transcription factors, thereby enabling transcription factor signatures to be established and potential common regulatory elements to be identified. In the present study, we searched for common promoter elements within the ABC superfamily members and between a given ABC transporter and all other genes present on the MuU74v2 GeneChips with a highly correlated expression profile. The promoter regions of the 43 ABC transporters have been characterized to varying degrees; therefore, the various promoters were analyzed according to Genomatix guidelines. As expected because of the widely different expression profiles of the 43 ABC transporters, no common promoter elements were identified among this protein family. Furthermore, comparing the promoters of the eight differentially expressed ABC transporters identified only the ETS1 factor as common among these murine genes (Fig. 3); however, the score assigned by FrameWorker was 0.43, which suggests a promoter model that is likely to match often to random DNA sequences (as defined by Genomatix, an "FW score" above 0.5 can be considered significant). In accordance with earlier comparative genome analyses, we found that both the TEF1 (TEAF) and GATA-binding factor 1 (GATA) regulatory elements are common between Abcg5 and Abcg8, as illustrated in Fig. 3 (20, 25). Furthermore, we have confirmed the findings of Remaley and colleagues (25) in that neither the LXR nor RXR elements are present in this bi-directional promoter.



View larger version (40K):
[in this window]
[in a new window]
 
Fig. 3. Identification of single elements within the promoter regions of the differentially expressed ABC transporters. Promoter elements were identified with FrameWorker for murine Abcb2, Abcb3, Abcb9, Abcc3, Abcc6, Abcd1, Abcg5, and Abcg8. For all transcripts, the putative promoter region was identified as the 500 bp upstream and 100 bp downstream of the predicted transcript start site. Only elements with an "FW score" >=0.5 and a core similarity of 1.00, with the exception of GATA and ETSF, are depicted. These two elements are included because GATA was found to be significant in the Abcg5/Abcg8 promoter comparison and ETSF is common to all eight transporters. Numbers directly above or below the identified binding sites indicate their start position (base pair) within the predicted promoter region. For reasons of clarity, only those elements present in at least three of the eight putative promote regions are indicated in color: FKHD, Forkhead domain factors; GATA, GATA binding factors; NFKB, nuclear factor kappa B/c-rel; ETSF, human and murine ETS1 factors; XSEC, Xenopus seleno cysteine tRNA activating factor; AP4R, AP4 and related proteins; SORY, SOx/sRY-sex/testis determining and related HMG box factors; MOKF, mouse Kruppel-like factor; PAX6, PAX-4/PAX-6 paired domain binding sites.

 
Hereafter, the putative promoter regions of each of the eight differentially expressed transporters and all genes with similar expression profiles (correlation = 0.90–0.99) were examined. Using our aforementioned criteria, Abcb2, Abcb3, Abcc3, Abcc6, Abcd1, Abcg5, and Abcg8 were not found to share common complex elements with other genes present on the Affymetrix GeneChips. Only one transporter, Abcb9, shared both common promoter elements and a common complex model with fibulin-1, an extracellular matrix protein. The transcript start site for both human and mouse fibulin-1 and mouse Abcb9 have been experimentally verified; therefore, Genomatix software considers this a "gold" promoter. In contrast, the transcript start site of human Abcb9 was confirmed by PromoterInspector prediction only; therefore, Genomatix software considers this a "silver" promoter. Nevertheless, both Abcb9 and fibulin-1 were found to contain the following common single elements with an FW score above 0.5: NF{kappa}B (NFKB; 0.57) and the NKX/DLX-homeodomain site (NKXH; 0.67). Other single elements were found in common, such as EGR/nerve growth factor-induced protein C (EGRF) and GC-box factors SP1/GC (SP1F), but their individual FW scores were below 0.5. However, transcription factors often associate into complexes to correctly activate the transcription of a given gene; therefore, those single elements that have an FW score below 0.5 may be found in complex models with significant scores. Indeed, we find that NF{kappa}B, ZBPF, SP1F, and EGRF associate to form a complex model with FW scores of 0.24/1.0, as illustrated in Fig. 4. Although additional complex models exist, this model consisted of the greatest number of single elements (i.e., 4) with the highest FW scores. Furthermore, the importance of Sp1 binding sites in both Abcb9 and fibulin-1 have been experimentally demonstrated by Castoldi and Chu (4) and Kobayashi et al. (16), respectively.



View larger version (23K):
[in this window]
[in a new window]
 
Fig. 4. Identification of a common complex model in both Abcb9 and fibulin-1. A complex model consisting of NF{kappa}B (blue), ZBPF (purple), SP1F (light green), and EGRF (yellow) was identified by FrameWorker as common to both the murine and human homologs for Abcb9 and fibulin-1 (FW score = 0.24/1.00). Colored nodes lying above the sequence are found on the forward strand, and nodes below the sequence are on the reverse strand. All possible combinations of the four elements are shown in each of the sequences.

 
Despite the different degree to which these gene promoters have been characterized, both the murine and human homologs contain these single and complex promoter elements and suggest a possibly correlated function in the intestine; however, the roles of both Abcb9, which has been located to lysosomes (34), and fibulin-1, located in the gut subepithelium (22), in the intestine are currently unknown.

Abcc3 protein distribution along the intestinal tract.
Abcc3 (MRP3) was selected for further analysis at the protein level because it was one of the eight differentially expressed ABC transporters. Immunohistochemical staining for Abcc3 revealed a similar cellular pattern of expression along the intestinal tract (Fig. 5). Protein location is restricted to the villi of the intestine, as no staining was present in the crypts of Lieberkühn (data not shown). Furthermore, staining appeared to increase in intensity from crypt to villi, suggesting that Abcc3 protein is predominantly expressed in fully differentiated enterocytes. In all segments of the intestine, Abcc3 is expressed basolaterally in enterocytes (Fig. 5, AD). Abcc3 is most highly expressed in the colon and to a lesser extent in the various regions of the small intestine. Although not quantitative, the high degree of protein staining in the colon visually concords with the mRNA expression profile, which also indicated a higher level of Abcc3 transcript in the colon. This qualitative statement is based on the lower exposure time required to capture protein staining in the colon (0.48 s) compared with the small intestine (between 1.04 and 1.44 s).



View larger version (88K):
[in this window]
[in a new window]
 
Fig. 5. Abcc3 protein localization along the murine intestinal tract. Expression of Abcc3 in the duodenum (A), jejunum (B), ileum (C), and colon (D). Arrows indicate the basolateral staining of Abcc3 in all sections of the intestine. Immunofluorescence was performed on frozen tissues (0.5 µm sections) with a TSA amplification kit (PerkinElmer). Scale bar = 30 µm.

 

    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIAL AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Microarray technology enables the simultaneous study of thousands of unique genetic elements and provides insight into the coordinated control of many genes (23). When considering the intestine, the complexity of this organ provides a formidable obstacle to understanding gene expression profiles. Therefore, a mandatory first step in understanding regional specificity is to examine the gene expression profiles occurring along the GIT tissue, which can then reveal important information concerning the genes involved in drug and nutrient metabolism. Bates and colleagues (1) provided an elegant first glimpse of the global gene expression profile along the murine GIT; however, the Incyte GEM1 microarray used by these authors is estimated to cover only 20–25% of the mouse genome (1). The present study is the first to assess the entire murine genome along the mouse intestinal tract; however, interpreting the expression information for all genetic elements is beyond the scope of this manuscript. Rather, we have preferred to focus our attention on the highly conserved, and functionally important, ABC transporter super family.

Few studies have examined the expression profiles of ABC transporters in the intestine to elucidate their region-specific functions. Studies by Rost et al. (26) and Chianale et al. (5) demonstrate the importance of treating the small intestine as functionally distinct regions rather than as a single entity. For example, Rost and colleagues (26) found that Abcc3 (MRP3) was predominantly expressed (mRNA and protein) in the ileum and colon of the rat intestine, and Chianale et al. (5) found that Abcb1a (corresponding to ABCB4 in the human) mRNA, protein, and transport rate were all approximately sixfold higher in the mouse ileum vs. the duodenum. Because of the high degree of homology for ABC transporters between eukaryotes, similar findings in humans would be expected. A good example supporting homology between species was demonstrated by Stephens et al. (30), in which the authors found that P-glycoprotein (human ABCB1)-mediated drug efflux increased along the rat and human intestine in a similar manner. Taken together, these examples demonstrate the importance of distinguishing between the various regions of the small intestine, and, furthermore, indicate that differences in mRNA levels can be indicative of functional differences across eukaryotes.

Approximately half of the 48 human ABC transporters have been characterized to various extents, with the remaining having currently unknown functions; therefore, interpreting the intestinal profiles of all differentially expressed ABC transporters is not possible at this time. Of noticeable interest are the overall low expression levels of ABC transporters in the gut. The association of these proteins in the transport of a multitude of nutritional and chemical compounds would imply that they might be highly expressed. This result may be explained in several ways. First, the functions/actions of many of the ABC transporters may stem from posttranscriptional regulation, i.e., low mRNA levels but a high degree of protein synthesis and/or stability. Second, ABC transporters may have highly sensitive promoter elements that permit a rapid response (i.e., increased transcription) when exposed to a specific substrate. Alternatively, this result may arise as a result of the experimental design rather than a biological phenomenon. The use of intestinal tissue rather than a specific intestinal cell type (e.g., absorptive cells) may have diluted their highly localized expression patterns.

Of those ABC transporters identified as differentially expressed in the gut, Abcb9, Abcc6, and Abcd1 have not been previously examined in the intestine, and their functions in this organ are currently unknown. Nevertheless, current information concerning their functions permits some speculation regarding their expression profiles in the intestine. For example, Abcb9 ("transporter associated with antigen processing-like," or TAPL) is closely related to the TAP2 gene and has a intestinal expression pattern similar to the TAP1/TAP2 complex (described in more detail below). This lysosomal protein may play a role in the translocation of peptides from the cytosol into the lysosome for degradation (16, 34). Abcc6 is a confirmed member of the MRP family of drug efflux pumps and has been demonstrated to transport glutathione conjugates, thereby having a potential role in the regulation of xenobiotic bioavailability in the intestine (2, 6). Finally, Abcd1 is a peroxisomal half transporter that is mutated in adrenoleukodystrophy, which is characterized by a reduced peroxisomal very-long-chain fatty acid (VLCFA) ß-oxidation (6). As the primary site for dietary fatty acid absorption occurs in the upper intestine, the higher Abcd1 expression levels in the duodenum and jejunum suggest that this peroxisomal transporter is actively involved in the metabolism of dietary VLCFA. Although the definitive functions of intestinal Abcb9, Abcc6, and Abcd1 have yet to be ascribed, two pairs of half transporters, Abcb2/Abcb3 (TAP1/TAP2) and Abcg5/Abcg8, have previously been studied in the intestine and have been implicated in immune responses and sterol transport, respectively.

TAP1 and TAP2 ("transporters associated with antigen presentation/processing") have been found to preferentially transport 9–12 amino acid peptides into the lumen of the endoplasmic reticulum and load these peptides onto major histocompatibility complex class 1 molecules, which are critical for an immune response (18). One possible explanation for their higher levels in the small intestine may stem from the reduced number of microorganisms. The much higher levels of microflora in the large intestine form an additional obstacle for passage of small molecules to the mucosal layer, as bacteria can metabolize and affect the bioavailability of various toxic chemicals (10, 12). Therefore the requirements for the TAP1/TAP2 complex, and the closely related Abcb9, could be diminished. With time, the function of these transporters and the biological purpose for this differential expression in the gut will undoubtedly be revealed.

Abcg5 and Abcg8, which are associated with sitosterolemia and the selective transport of sterol compounds (3), are highly expressed in the small intestine and found at much lower levels in the colon. This suggests that the selection process for the efflux of plant sterols vs. cholesterol from enterocytes back to the intestinal lumen is restricted to the small intestine and would not occur to a significant extent in the large intestine. The relative stability of mRNA expression in the duodenum, jejunum, and ileum would suggest that this active selection process could occur equivalently along the entire length of the small intestine; however, this will need to be examined via functional transport studies.

Abcc3 protein analysis provided additional information that is important in understanding the role of this differentially expressed transporter in the GIT. As indicated above, Rost and colleagues (26) found that Abcc3 (MRP3) was most highly expressed in the colon of the rat intestine. Our findings indicate that MRP3 is also highly expressed in the murine colon at both the mRNA and protein levels. Furthermore, the similar cellular location (basolateral in enterocytes) found between the mouse and rat further supports the notion of a high degree of conservation for ABC transporters amongst eukaryotes. This suggests that MRP3 may have a similar role in the ATP-dependent transport of 17ß-glucuronosyl estradiol, glucuronosyl bilirubin, monovalent bile salts (taurocholate and glycocholate), and sulfated bile salts (i.e., taurochenodeoxycholate-3-sulfate, taurolithocholate-3-sulfate) from the enterocyte to the blood in all higher mammals.

In conclusion, regional differences in the expression of eight ABC transporters along the mouse intestinal tract were identified and confirmed by both RT-PCR and an in silico comparison with a publicly available cDNA data set. Examining the expression profiles of differentially expressed ABC transporters within the context of all genetic elements present on the microarrays revealed that common promoter elements can be identified and yield important information for deciphering the functions and coordinate regulation of ABC transporters. Furthermore, linking mRNA expression analysis with immunohistochemistry yields complementary information required for the complete characterization of gene function. These data indicate that mRNA analysis is a necessary first step in understanding tissue- and region-specific gene function, yet herald the necessity to establish higher-throughput tools for the study of ABC transporters at the protein and functional levels, which in combination with gene expression data, will provide the means to fully characterize ABC transporter function in the intestine.


    ACKNOWLEDGMENTS
 
We are grateful for the communication and assistance of Drs. Michael Bates and Bruce Aronow (University of Cincinnati, OH) with the interpretation of their Incyte GEM1 cDNA data set. We thank Dr. Anne Donnet-Hughes and Catherine Schwartz (Nestlé Research Center) for assistance and advice regarding the euthanasia of animals and preparation of mouse tissue for immunohistochemistry. We also thank Prof. Jean-Pierre Kraehenbuhl and Dr. Grigorios Fotopoulos for critical review of this manuscript. Finally, we thank Alex Sim (AMS Biotechnology) for time and dedication to this study.


    FOOTNOTES
 
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).

Address for reprint requests and other correspondence: G. Williamson, Head of Nutrient Bioavailability, Nestlé Research Center, Vers-Chez-Les-Blanc, PO Box 44, CH-1000 Lausanne 26, Switzerland (E-mail: gary.williamson{at}rdls.nestle.com).

10.1152/ physiolgenomics.00150.2003.


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIAL AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Bates MD, Erwin CR, Sanford LP, Wiginton D, Bezerra JA, Schatzman LC, Jegga AG, Ley-Ebert C, Williams SS, Steinbrecher KA, Warner BW, Cohen MB, and Aronow BJ. Novel genes and functional relationships in the adult mouse gastrointestinal tract identified by microarray analysis. Gastroenterology 122: 1467–1482, 2002 (Supplemental information: http://trinity.chmcc.org/register/gastro/index.php)[ISI][Medline]
  2. Belinsky MG, Chen ZS, Shchaveleva I, Zeng H, and Kruh GD. Characterization of the drug resistance and transport properties of multidrug resistance protein 6 (MRP6, ABCC6). Cancer Res 62: 6172–6177, 2002.[Abstract/Free Full Text]
  3. Berge KE, Tian H, Graf GA, Yu L, Grishin NV, Schultz J, Kwiterovich P, Shan B, Barnes R, and Hobbs HH. Accumulation of dietary cholesterol in sitosterolemia caused by mutations in adjacent ABC transporters. Science 290: 1771–1775, 2000.[Abstract/Free Full Text]
  4. Castoldi M and Chu ML. Structural and functional characterization of the human and mouse fibulin-1 gene promoters: role of Sp1 and Sp3. Biochem J 362: 41–50, 2002.[CrossRef][ISI][Medline]
  5. Chianale J, Vollrath V, Wielandt AM, Miranda S, Gonzalez R, Fresno AM, Quintana C, Gonzalez S, Andrade L, and Guzman S. Differences between nuclear run-off and mRNA levels for multidrug resistance gene expression in the cephalocaudal axis of the mouse intestine. Biochim Biophys Acta 1264: 369–376, 1995.[ISI][Medline]
  6. Dean M. The Human ATP-Binding Cassette (ABC) Transporter Superfamily. Bethesda, MD: NCBI, Natl Library Medicine, 2002, p. 1–45. (http://www.ncbi.nlm.nih.gov/books/bv.fcgi?call=bv.View..ShowTOC&rid=mono_001.TOC&depth=2)
  7. Dean M, Rzhetsky A, and Allikmets R. The human ATP-binding cassette (ABC) transporter superfamily. Genome Res 11: 1156–1166, 2001.[Abstract/Free Full Text]
  8. Dieterich C, Wang H, Rateitschak K, Luz H, and Vingron M. CORG: a database for COmparative Regulatory Genomics. Nucleic Acids Res 31: 55–57, 2003.[Abstract/Free Full Text]
  9. Dressman JB, Bass P, Ritschel WA, Friend DR, Rubinstein A, and Ziv E. Gastrointestinal parameters that influence oral medications. J Pharm Sci 82: 857–872, 1993.[ISI][Medline]
  10. Femia AP, Luceri C, Dolara P, Giannini A, Biggeri A, Salvadori M, Clune Y, Collins KJ, Paglierani M, and Caderni G. Antitumorigenic activity of the prebiotic inulin enriched with oligofructose in combination with the probiotics Lactobacillus rhamnosus and Bifidobacterium lactis on azoxymethane-induced colon carcinogenesis in rats. Carcinogenesis 23: 1953–1960, 2002.[Abstract/Free Full Text]
  11. Gottesman MM and Ambudkar SV. Overview: ABC transporters and human disease. J Bioenerg Biomembr 33: 453–458, 2001.[CrossRef][ISI][Medline]
  12. Grangette C, Muller-Alouf H, Goudercourt D, Geoffroy MC, Turneer M, and Mercenier A. Mucosal immune responses and protection against tetanus toxin after intranasal immunization with recombinant Lactobacillus plantarum. Infect Immun 69: 1547–1553, 2001.[Abstract/Free Full Text]
  13. Kendziorski CM, Zhang Y, Lan H, and Attie AD. The efficiency of pooling mRNA in microarray experiments. Biostatistics 4: 465–477, 2003.[Abstract/Free Full Text]
  14. Kepler TB, Crosby L, and Morgan KT. Normalization and analysis of DNA microarray data by self-consistency and local regression. Genome Biol 3: RESEARCH0037, 2002.[Medline]
  15. Klingenhoff A, Frech K, Quandt K, and Werner T. Functional promoter modules can be detected by formal models independent of overall nucleotide sequence similarity. Bioinformatics 15: 180–186, 1999.[Abstract/Free Full Text]
  16. Kobayashi A, Hori S, Suita N, and Maeda M. Gene organization of human transporter associated with antigen processing-like (TAPL, ABCB9): analysis of alternative splicing variants and promoter activity. Biochem Biophys Res Commun 309: 815–822, 2003.[CrossRef][ISI][Medline]
  17. Langmann T, Mauerer R, Zahn A, Moehle C, Probst M, Stremmel W, and Schmitz G. Real-time reverse transcription-PCR expression profiling of the complete human ATP-binding cassette transporter superfamily in various tissues. Clin Chem 49: 230–238, 2003.[Abstract/Free Full Text]
  18. Lankat-Buttgereit B and Tampe R. The transporter associated with antigen processing: function and implications in human diseases. Physiol Rev 82: 187–204, 2002.[Abstract/Free Full Text]
  19. Lin H, Stoehr JP, Nadler ST, Schueler KM, Yandell BS, and Attie AD. Adaptive gene picking with microarray data: detecting important low abundance signals. In: The Analysis of Gene Expression Data: Methods and Software, edited by Parmigiani G, Garrett ES, Irizarry RA, and Zeger SL. New York: Springer, 2003.
  20. Lu K, Lee MH, Yu H, Zhou Y, Sandell SA, Salen G, and Patel SB. Molecular cloning, genomic organization, genetic variations, and characterization of murine sterolin genes Abcg5 and Abcg8. J Lipid Res 43: 565–578, 2002.[Abstract/Free Full Text]
  21. Makhey VD, Guo A, Norris DA, Hu P, Yan J, and Sinko PJ. Characterization of the regional intestinal kinetics of drug efflux in rat and human intestine and in Caco-2 cells. Pharm Res 15: 1160–1167, 1998.[CrossRef][ISI][Medline]
  22. Miosge N, Gotz W, Sasaki T, Chu ML, Timpl R, and Herken R. The extracellular matrix proteins fibulin-1 and fibulin-2 in the early human embryo. Histochem J 28: 109–116, 1996.[ISI][Medline]
  23. Mutch DM, Berger A, Mansourian R, Rytz A, and Roberts MA. The limit fold change model: a practical approach for selecting differentially expressed genes from microarray data. BMC Bioinformatics 3: 17, 2002.[CrossRef][Medline]
  24. Peng X, Wood CL, Blalock EM, Chen KC, Landfield PW, and Stromberg AJ. Statistical implications of pooling RNA samples for microarray experiments. BMC Bioinformatics 4: 26, 2003.[CrossRef][Medline]
  25. Remaley AT, Bark S, Walts AD, Freeman L, Shulenin S, Annilo T, Elgin E, Rhodes HE, Joyce C, Dean M, Santamarina-Fojo S, and Brewer HB Jr. Comparative genome analysis of potential regulatory elements in the ABCG5-ABCG8 gene cluster. Biochem Biophys Res Commun 295: 276–282, 2002.[CrossRef][ISI][Medline]
  26. Rost D, Mahner S, Sugiyama Y, and Stremmel W. Expression and localization of the multidrug resistance-associated protein 3 in rat small and large intestine. Am J Physiol Gastrointest Liver Physiol 282: G720–G726, 2002. First published October 10, 2001; 10.1152/ajpgi.00318.2001.[Abstract/Free Full Text]
  27. Saitoh H and Aungst BJ. Possible involvement of multiple P-glycoprotein-mediated efflux systems in the transport of verapamil and other organic cations across rat intestine. Pharm Res 12: 1304–1310, 1995.[CrossRef][ISI][Medline]
  28. Sambrook J and Russell D. Molecular Cloning: A Laboratory Manual (3rd ed.). New York: Cold Spring Harbor Laboratory Press, 2001, p. 2100.
  29. Scherf M, Klingenhoff A, and Werner T. Highly specific localization of promoter regions in large genomic sequences by PromoterInspector: a novel context analysis approach. J Mol Biol 297: 599–606, 2000.[CrossRef][ISI][Medline]
  30. Stephens RH, O’Neill CA, Warhurst A, Carlson GL, Rowland M, and Warhurst G. Kinetic profiling of P-glycoprotein-mediated drug efflux in rat and human intestinal epithelia. J Pharmacol Exp Ther 296: 584–591, 2001.[Abstract/Free Full Text]
  31. Walker JE, Saraste M, Runswick MJ, and Gay NJ. Distantly related sequences in the alpha- and beta-subunits of ATP synthase, myosin, kinases and other ATP-requiring enzymes and a common nucleotide binding fold. EMBO J 1: 945–951, 1982.[ISI][Medline]
  32. Wurmbach E, Yuen T, Ebersole BJ, and Sealfon SC. Gonadotropin releasing hormone receptor-coupled gene network organization. J Biol Chem 276: 47195–47201, 2001.[Abstract/Free Full Text]
  33. Yuen T, Wurmbach E, Pfeffer RL, Ebersole BJ, and Sealfon SC. Accuracy and calibration of commercial oligonucleotide and custom cDNA microarrays. Nucleic Acids Res 30: e48, 2002.[Abstract/Free Full Text]
  34. Zhang F, Zhang W, Liu L, Fisher CL, Hui D, Childs S, Dorovini-Zis K, and Ling V. Characterization of ABCB9, an ATP binding cassette protein associated with lysosomes. J Biol Chem 275: 23287–23294, 2000.[Abstract/Free Full Text]