The Polymorphism at Codon 54 of the FABP2 Gene Increases Fat Absorption in Human Intestinal Explants*

Based on titration microcalorimetry and Caco-2 cell line transfection studies, it has been suggested that the A54T of the FABP2 gene plays a significant role in the assimilation of dietary fatty acids. However, reports were divergent with regard to the in vivo interaction between this polymorphism and postprandial lipemia. We therefore determined the influence of this intestinal fatty acid-binding protein polymorphism on intestinal fat transport using the human jejunal organ culture model, thus avoiding the interference of various circu-lating factors capable of metabolizing in vivo postprandial lipids. Analysis of DNA samples from 32 fetal intes-tines revealed 22 homozygotes for the wild-type Ala-54/ Ala-54 genotype (0.83) and 10 heterozygotes for the polymorphic Thr-54/Ala-54 genotype (0.17). The Thr-en-coding allele was associated with increased secretion of newly esterified triglycerides, augmented de novo apolipoprotein B synthesis, and elevated chylomicron out-put. On the other hand, no alterations were found in very low density lipoprotein and high density lipoprotein production,

The intestine is the site essential for the transport of alimentary fat in the form of lipoproteins. After the digestive phase, the lipolytic products are absorbed by the enterocyte, in which a complex series of sequential events result in their packaging as chylomicrons (1)(2)(3). The formation and secretion of chylomicrons are key steps in the transport of dietary fats and fatsoluble vitamins (4,5). The assembly of these triglyceride-rich lipoproteins within the enterocyte is a multistep pathway including the uptake and translocation of lipolytic products from the brush border membrane to the endoplasmic reticulum by fatty acid-binding proteins (FABPs), 1 lipid esterification, synthesis and posttranslational modification of different apolipoproteins, as well as the packaging of lipid and apolipoprotein components into lipoprotein particles (1)(2)(3)(4)(5)(6).
Intestinal FABP (I-FABP) is a small cytosolic protein involved in intracellular fatty acid (FA) transfer and metabolism (7). However, its specific function remains unclear. It is encoded by the FABP2 gene, located in the long arm of chromosome 4 (8,9). The G to A polymorphism of codon 54 of the FABP2 gene results in the substitution of a Thr (mutated-type) in I-FABP (10). This Thr variant exhibits a 2-fold greater affinity for long-chain FA compared with Ala-containing I-FABP (wild-type) in vitro (10). In vivo studies reveal that the Thr-encoding allele is associated with insulin resistance and enhanced fat oxidation rates among Pima Indians, a group known to have the highest prevalence of type 2 diabetes mellitus (10,11). A similar association observed among aboriginal Canadians between the Thr-54 I-FABP allele and elevated plasma triglycerides supports the hypothesis that this polymorphism is related to an increased affinity to dietary longchain FA and enhanced intestinal lipid uptake (12). This same polymorphism was further associated with increased fat transport in Caco-2 cells (13) and postprandial lipemic response (14). Nevertheless, other reports did not confirm the role of the Thr-54 allele in lipid and glucose metabolism. The same mutation in a Finnish population had no impact on the FA composition of plasma lipids, the basal metabolic rate, or the insulin, glucose, and lipid levels (15)(16)(17). Moreover, the increased lipid oxidation and postprandial response associated with the Thr-54 allele may be due to various confounding in vivo factors, including insulin resistance, fasting triglyceride concentrations, and various polymorphisms of apo-E, apo-C-II, lipoprotein lipase, and cholesteryl ester transfer protein (16, 18 -20). Additionally, the Ala-to-Thr shift in I-FABP may alter the amount of absorbed FA transported by the portal route to the liver (21). Finally, recent experimental studies in mice and studies using Caco-2 cells have suggested that I-FABP is not essential for dietary fat absorption (22,23).
The aim of our study was to investigate the relationship between polymorphisms of I-FABP and intracellular lipid transport in the human intestine. We also sought to determine how the Thr-encoding allele influences lipid esterification, de novo synthesis of apolipoproteins A-I and B, and the secretion of lipoproteins, as well as glucose uptake and metabolism. To address these issues in vitro, we utilized an organ culture of human intestine with fetal explants. This represents a unique experimental model to further our understanding of the complex biosynthetic molecular events essential to the formation and secretion of lipoproteins in the human intestine (24 -28). Thus, the intestine organ culture model affords the opportunity to investigate the functional role of polymorphisms of I-FABP while minimizing the influence of the many confounding factors seen in the in vivo situation, thus facilitating the interpretation of data.

MATERIALS AND METHODS
Specimens-Intestinal tissue was obtained from 32 normal fetuses at 17-20 weeks of gestation after elective pregnancy terminations. No specimens were collected from cases associated with any congenital abnormalities or fetal deaths. Studies were approved by the institutional Human Subject Review Board. The entire small intestine was immersed in Leibovitz L-15 medium containing garamycin (40 g/ml) and transported immediately to the laboratory at room temperature. The proximal half of the intestine excluding the first 3 cm was used and defined as jejunum.
Intestinal Organ Cultures-The jejunum was cleansed of mesentery, split longitudinally, washed in culture medium, and cut into explants (3 ϫ 7 mm). An average of 10 -15 explants were randomly transferred onto a lens paper and placed mucosal side up into each organ culture dish (Falcon Plastics, Los Angeles, CA). Six dishes were used for each experimental condition. 0.8 ml of Leibovitz L-15 medium was added to the central well of each culture dish, a volume sufficient to dampen the tissue without immersing it (29). Explants were cultured in serum-free L-15 medium as described previously (25,30,31). After an initial 3-h stabilization period, the medium was refreshed with a final concentration of 1.0 mol/ml unlabeled oleic acid with 0.3 Ci of [ 14 C]oleic acid (specific activity, 53.9 mCi/mmol; Amersham Pharmacia Biotech) (25)(26)(27)(28). Explants originating from each fetus were genotyped for the FABP2 A54T polymorphism (10). Lipid analyses were carried out on both the tissue and culture media. The morphological integrity of the jejunal explants was confirmed at the onset and after 5 days of culture, as described previously (31).
Isolation of Lipoproteins-After incubation, the explants were homogenized in isotonic saline containing antibacterial and antiprotease agents (0.01% sodium azide, 0.1% EDTA, and 10,000 (kallitrein inactivating units (kiu)/ml Trasylol). The medium was mixed with the plasma lipid carrier (2:1) (v/v), and the lipoproteins were isolated by sequential ultracentrifugation using a TL-100 ultracentrifuge (Beckman Instruments Inc., Montreal, Canada), as described previously (25,32). Briefly, after the removal of the 0.97 g/ml density fraction by centrifugation using a TLS 55 rotor (20,000 rpm for 20 min), lipoproteins with density Ͻ 1.006 g/ml were separated by spinning at 100,000 ϫ g for 2.26 h with a TL 100.3 rotor at 5°C. The high density fraction was obtained by adjusting the infranatant to a density of 1.21 g/ml and by centrifugation for 6.5 h at 100,000 rpm. Each lipoprotein fraction was exhaustively dialyzed against 0.15 M NaCl and 0.001 M EDTA, pH 7.0, at 4°C for 24 h.
De Novo Apolipoprotein Synthesis-After the incubation period in the presence of unlabeled oleic acid to stimulate the synthesis of lipids and apoproteins, jejunal explants were washed twice with methioninefree Leibovitz medium. They were then incubated in the same medium containing unlabeled oleic acid for 60 min at 37°C in the presence of [ 35 S]methionine (300 Ci/ml; specific activity, 1062 Ci/mmol) (27,28,33). At the end of the labeling period, explants were washed three times and homogenized in phosphate-buffered saline (20 mM sodium phosphate and 145 mM NaCl, pH 7.4) containing 1% (w/v) Triton X-100, 2 mM methionine, 1 mM phenylmethylsulfonyl fluoride, and 1 mM benzemidine. Aliquots of tissue homogenates were then treated with 20% trichloroacetic acid, and the protein precipitates were washed three times with 5% trichloroacetic acid. The radioactivity was then determined using a liquid scintillation spectrometer (Beckman Instruments Inc.). Another aliquot of the homogenate was also centrifuged (4°C) at 105,000 ϫ g for 60 min in a 50-Ti rotor (Beckman Instruments Inc.) to determine apolipoprotein levels. The immunoprecipitation of apo-A-I and apo-B was then carried out on supernatants with excess specific monoclonal antibodies (Roche Molecular Biochemicals). Pansorbin (Calbiochem, San Diego, CA) was then added, and the mixture was reincubated at 20°C for 60 min. The immunoprecipitate was washed extensively and analyzed by a linear 4 -20% acrylamide gradient preceded by a 3% stacking gel, as described previously (27,28,33). Gels were sectioned into 4-mm slices and counted after an overnight incubation at 20°C with 1 ml of BTS-450 (Beckman Instruments Inc.) in 10 ml of liquid scintillation fluid (Ready solvent; Beckman Instruments Inc.).
FABP2 Polymorphisms-DNA from intestinal explants was isolated (34), and genotyping was performed using the solid-phase minisequencing method (35). Three different primers were used. First, exon 2 of FABP2 was amplified by PCR using forward primer 5Ј-ACAGGTGTTA-ATATAGTGAAAAGG-3Ј and reverse biotinylated primer 5Ј-TACCCT-GAGTTCAGTTCCGTCTGC-3Ј. The product was analyzed using detection primer 5Ј-TCACAGTCAAAGAATCAAGC-3Ј in a single-step extension reaction in avidine-coated wells with a radiolabeled nucleotide (either G or A). The PCR conditions were as follows: denaturing at 95°C, annealing at 60°C, and extension at 72°C, each for 30 s for a total of 30 cycles. The PCR reaction volume was 40 l. The reaction temperature for detection was 50°C.
Reverse Transcription-PCR Analysis-After incubating explants in the presence of unlabeled oleic acid, RNA was extracted and amplified by PCR to determine apo-B and glyceraldehyde-3-phosphate dehydrogenase mRNA. All these techniques were described in detail previously (28).
Lipid Analyses-Aliquots of explant homogenates and their respective incubation media were lipid-extracted with chloroform:methanol (2:1) (v/v) (36). Small amounts of lipid standards were added to the samples before separation of individual lipid classes by one-dimensional silica gel TLC (Eastman Kodak Co.), as described previously (25,37). The nonpolar solvent system was hexane:diethylether:glacial acetic acid (80:20:3) (v/v/v). The radioactivity of the separated fractions was measured using a liquid scintillation spectrometer. Quenching was corrected using computerized curves generated with external standards. An aliquot of the tissue homogenate was used for protein determinations (38). The hydrolytic activities of sucrase, glucoamylase, lactase, and alkaline phosphatase of cultured explants were assayed as described by Ménard et al. (31). Glucose oxidation was measured by incubating jejunal explants with D-[U-14 C]glucose (specific activity) and collecting 14 CO 2 by soaking paper rolls in hyamine hydroxyde (39). The activity and mass of intestinal microsomal triglyceride transfer protein (MTP) were evaluated as described previously (40).

RESULTS AND DISCUSSION
Analysis of I-FABP polymorphisms among the 32 intestinal specimens revealed 22 homozygotes for the FABP2 Ala-54/ Ala-54 genotype, 1 homozygote for the Thr/Thr genotype, and 9 heterozygotes for the FABP2 Thr-54/Ala-54 genotype. Thus, the frequencies of the alanine-encoding (Ala-54) and threonineencoding (Thr-54) FABP2 alleles were determined to be 0.83 and 0.17, respectively. Similar prevalences were observed in Caucasians and among Pima Indians (10, 14, 15). Because only one subject was found to be homozygous for the Thr-54 allele, his data were pooled with those of the Thr-54/Ala-54 heterozygotes. Subsequent experiments were carried out to determine whether the alanine to threonine substitution in the FABP2 gene product results in altered intestinal lipid transport.
The incorporation of labeled oleic acid by intestinal explants was used to quantify the rate of synthesis of total lipids as well as that of the specific lipid classes. As shown in Fig. 1, lipid synthesis was significantly greater in explants expressing the Thr-54 I-FABP variant. Analysis of the different lipid classes by TLC revealed that triglyceride and phospholipid esterification was significantly augmented (ϳ16% and 35%, respectively) in jejunal explants with the Thr-54-containing I-FABP compared with wild-type jejunal explants ( Fig. 2A). Similarly, the secretion of labeled triglycerides was higher (ϳ41%) in the medium of the heterozygous I-FABP tissue (Fig. 2B). No significant variations were noted in phospholipid and cholesteryl ester export. When lipid composition was assessed as a percentage of total lipid distribution, no significant differences were noted between the two groups (Table I). Overall, our findings suggest that the expression of Thr/Ala I-FABP is characterized primarily by an increased synthesis and secretion of triglycerides, the major class of dietary lipids.
Experiments were also conducted to examine the impact of the I-FABP polymorphism on the transport of newly synthe-  sized lipids by lipoproteins. Culture supernatants were collected and immediately ultracentrifuged to separate chylomicrons, very low density lipoproteins, and high density lipoproteins. As expected from the aforementioned lipid data, alterations in lipoprotein exocytosis were found in association with the Thr-encoding allele (Fig. 3). The secretion of chylomicrons, the predominant lipoprotein fraction incorporating [ 14 C]oleic acid, was greater in jejunal explants with Thr-containing I-FABP than in those with Ala-containing I-FABP. Very low density lipoprotein and high density lipoprotein production was not significantly different between tissues with the Thr-encoding allele of the FABP2 gene and tissues with the Ala-encoding allele of the FABP2 gene. Our data thus suggest that chylomicrons are the predominant lipoprotein particles, which are influenced by the I-FABP polymorphism. The next set of experiments was designed to examine the effect of I-FABP polymorphism on the process of apolipoprotein biogenesis. Their synthesis was estimated by the incorporation of [ 35 S]methionine, with subsequent immunoprecipitation and separation by SDS-polyacrylamide gel electrophoresis. Incubation of jejunal I-FABP polymorphic explants with [ 35 S]methionine resulted in increased apolipoprotein synthesis and secretion of apo-B-48 and apo-B-100, whereas the magnitude of apo-A-I production was not modified by the Ala to Thr shift in I-FABP (Table II). To determine whether the increase in apo-B synthesis associated with FABP2 gene polymorphism was related to augmented apo-B mRNA expression, reverse transcription-PCR analysis was performed. No significant effect of the I-FABP polymorphism was noted on the abundance of apo-B mRNA (Fig. 4) because the relative apo-B mRNA levels (expressed as average ratio values of apo mRNA:glyceraldehyde-3-phosphate dehydrogenase mRNA) in three experiments were similar between Thr-54 carriers and Ala-54 homozygotes (0.96 Ϯ 0.11 versus 1.07 Ϯ 0.13, respectively). This suggests that co-translational or posttranslational mechanisms may be responsible for the regulation of apo-B formation in the presence of Thr-containing I-FABP. An extensive body of literature suggests that apo-B mRNA levels are relatively stable and generally do not change under conditions that alter extracellular apo-B secretion (27,28,(41)(42)(43). On the other hand, there is evidence that intracellular degradation of apo-B is the major regulatory event for TG-rich lipoprotein secretion. The stabilization phase necessary for the correct folding of translocating apo-B and its assembly with TG-rich lipoproteins requires a bulk transfer of lipid to apo-B (44 -46). Otherwise, apo-B is misfolded and is probably degraded via the ubiquitin-dependent proteasomal pathway (47)(48)(49). It is therefore conceivable that the changes in protein properties caused by the I-FABP polymorphism result in more efficient conveyance of fatty acids to the endoplasmic reticulum, which preserves apo-B form degradation. Accordingly, the threonine-containing I-FABP has been shown to possess higher affinity for long-chain fatty acids (10). Therefore, we can speculate that the alanine to threonine substitution at residue 54 facilitates the cytoplasmic attachment of fatty acids and their fast release to the endoplasmic reticulum, where they undergo esterification to TG, resulting in both apo-B protection from proteolytic degradation and enhanced chylomicron formation, as suggested by our data.
The small intestinal enterocyte contains two distinct FABPs, liver FABP and I-FABP, which are 29% homologous (1). To determine whether their abundance is altered by the Ala-to-Thr shift in FABP, two intestinal specimens were submitted to Western analysis. Although I-FABP polymorphic explants produced more triglycerides (38%) and chylomicrons (31%) than wild-type explants, no marked differences were noted in I-FABP and L-FABP levels (Fig. 5), thus suggesting the involvement of I-FABP functional properties in the magnitude of fat

Synthesis of apolipoproteins A-I and B by cultured jejunal explants
Fetal jejunal explants were incubated with [ 35 S]methionine in the presence of unlabeled oleic acid for 60 min to stimulate the biogenesis of apolipoproteins. At the end of the labeling period, explants were washed, homogenized, and centrifuged. Supernatants from the cell homogenates and concentrated media were then reacted with excess antibodies for 18 h at 4°C to precipitate specific apolipoproteins. Immune complexes were washed and analyzed by linear 4 -20% SDSpolyacrylamide gel electrophoresis. After electrophoresis, gels were sliced and counted for radioactivity. Data represent the means Ϯ SEM of five specimens in each group and are expressed as the percentage of trichloroacetic acid-precipitable proteins/mg tissue protein. absorption. This is consistent with the findings of Baier et al. (10), whose observations indicate that Thr-54 substitution in I-FABP occurs in a region of the molecule involved in fatty acid binding that could alter its overall stability, ligand affinity, or the kinetics of fatty acid acquisition/release. As mentioned before, the extended affinity could be expected to enhance intestinal fat absorption by increasing intracellular fatty acid trafficking and processing by the secretory pathway. However, other mechanisms are possible because additional functions have been suggested for FABP, including targeting fatty acids to intracellular organelles and different metabolic pathways, stimulating the activities of specific protein enzymes associated with FA metabolism (microsomal glycerol 3-phosphate acyltransferase and lysophosphatidyl acyltransferase), protecting intracellular polyunsaturated fatty acids against peroxidation, and regulating signal transduction pathways (50). Nevertheless, the use of a very limited number of specimens in our studies emphasizes the need for careful characterization of cellular I-FABP and L-FABP in relation to FABP2 genotype to validate our data and draw definitive conclusions about the mechanisms involved in intestinal fat transport. MTP resides in the endoplasmic reticulum in tissues secreting apo-B-containing lipoproteins. Absence of the functional MTP causes abetalipoproteinemia, a recessive disease characterized by a deficiency in the assembly process and secretion of TG-rich lipoproteins into plasma (51). Given its critical role in the formation of apo-B-containing lipoproteins, we determined the protein mass and activity of MTP in intestinal tissues with Thr-containing I-FABP and Ala-containing I-FABP (Fig. 6). However, the Ala to Thr shift did not alter MTP mass. MTP activity in the intestinal explants was then estimated by monitoring the transfer of triglyceride from donor to acceptor vesicles. No marked changes were noted in MTP activity between I-FABP wild-type and polymorphic tissues. It seems therefore that the ability of the Thr-encoding allele to enhance enterocyte fat transport is due to the stimulation of triglyceride synthesis, a mechanism likely involving apo-B protection.
The ability of I-FABP to influence intestinal carbohydrate metabolism was also investigated in the present studies. The evaluation of brush border disaccharidase hydrolytic activities and glucose metabolism did not reveal any significant differences between mutant and wild-type I-FABP intestinal specimens (Table III). The uptake and oxidation of glucose by these tissues was comparable, as shown by quantification of CO 2 production from [U-14 C]glucose and was not influenced by I-FABP polymorphism (Table III).
Our observations stress the potential impact of I-FABP polymorphisms on the biosynthetic events involved in the formation and secretion of lipoproteins. Individuals endowed with the Thr-54 genotype are presumed to express augmented in-testinal absorption and elevation of plasma postprandial chylomicron concentrations in response to dietary long-chain fatty acids. However, such an association has not been validated in a certain number of in vivo studies (15)(16)(17). One may thus suspect that other mechanisms involved in plasma lipid clearance may obscure the relationship between I-FABP variants and intestinal fat contribution. Common genetic variants influence the lipolytic process and receptor removal pathway of TG-rich lipoproteins (18 -20, 52, 53). These determinants include lipoprotein lipase, apo-E, apo-C-II, apo-B, and apo-E receptor that can modulate chylomicron degradation and remnant uptake, thereby altering their clearance from the circulation. The increased availability of free fatty acids via chylomicron degradation in the peripheral circulation of subjects with I-FABP polymorphism may affect intermediary metabolism. It has been reported (54) that high plasma concentrations of free fatty acids decrease the rate of insulin-stimulated glucose uptake in skeletal muscle, resulting in a compensatory increase in plasma insulin concentrations. Chronically, this situation leads to insulin resistance, as documented in the Pima Indians, a native population of the American Southwest with an extraordinarily high frequency of type 2 diabetes (10). The same I-FABP polymorphism was associated with a defect in postprandial lipemia (14) and other dyslipidemias (55). Recently, I-FABP has also been suggested to function physiologically as a FIG. 5. Effect of I-FABP polymorphism on protein mass of I-FABP and L-FABP. Jejunal explants were immunoprecipitated with specific anti-I-FABP and anti-L-FABP antibodies. The immunoprecipitates were run on SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membrane, which was blotted with I-FABP and L-FABP antibodies. There were no consistent differences in the expression of I-FABP and L-FABP between wild-type (lanes 1 and 3) and polymorphic (lanes 2 and 4) specimens.  lipid-sensing component of energy homeostasis (22). Our studies here provide further evidence that polymorphisms in I-FABP are not silent genetic markers; rather, they likely play a key role in lipid production and metabolism.