THE ANATOMICAL RECORD 229:51-60 (1991) Ontogenic Appearance of Three Fatty Acid Binding Proteins in the Rat Stomach SHOICHI ISEKI, TATSUO KANDA, MASAHIRO HITOMI, A N D TERUO O N 0 Department of Anatomy, Kanazawa University School of Medicine, Kanazawa (S.I.1, and Department of Biochemistry, Niigata Universit.y School o f Medicine, Niigata (T.K., M.H., T.O.),Japan ABSTRACT With the use of specific antibodies against three structurally different fatty acid binding proteins (FABPs),viz, liver FABP (L-FABP),heart FABP (H-FABP), and intestinal FABP (I-FABP),the localization and relative amount of the immunoreactive proteins were determined by immunoblotting and immunocytochemistry in the gastric epithelium of rats during prenatal and postnatal development. H-FABP immunoreactivity was first detected at embryonic day 20 (E201, with predominant localization in the parietal cells, whereas I-FABP immunoreactivity was detected at the day of birth in the surface mucous cells. Both immunoreactivities were continuously localized in the same cell types with increasing intensity into adulthood. In contrast, the immunoreactivity for L-FABP showed remarkable changes in intensity and localization during development of t h e rat stomach. It was first detected in the surface mucous cells of E19. In the first 2 weeks of postnatal life, i.e., the suckling period, L-FABP immunoreactivity reached a peak in intensity and was localized not only in the surface mucous cells, but also in some of the parietal cells, brush cells, and endocrine D cells. In the following few weeks of weaning, the reactivity of surface mucous cells and parietal cells disappeared, leaving only a small amount of total L-FABP immunoreactivity in the adult stomach, which was localized exclusively in the brush cells and D cells. These results revealed t h a t the appearance of the three types of FABPs in the r a t stomach is specific to cell types and developmental stages. Fatty acid binding proteins (FABPs) represent a class of low-molecular-weight cytosolic proteins present in a variety of mammalian tissues and having high binding affinity for long-chain fatty acids (Ockner e t al., 1972; Glatz and Veerkamp, 1985; Sweetser e t al., 1987). In the rat, three structurally different FABPs have been identified, isolated, and named liver FABP (L-FABP; Ockner e t al., 1982; Takahashi e t al., 1983), heart FABP (H-FABP; Fournier e t al., 1978; Said and Schultz, 1984), and intestinal FABP (I-FABP; Alpers e t al., 1984). These proteins may function in t h e uptake, intracellular transport, and metabolism of free fatty acids and their acyl-CoA esters. The occurrence of the three FABPs in rat tissues has been studied biochemically in terms of the tissue-specific expression of the mRNAs (Gordon et al., 1985; Heuckeroth e t al., 1987) and of the proteins (Bass and Manning, 1986; Paulussen e t al., 1989). According to those studies, L-FABP is abundantly expressed in the liver and intestine, HFABP in the heart and muscle, and I-FABP in the intestine. It should be noticed that the stomach is one of the few organs which express all three types of FABPs to some extent, and therefore is suitable to investigate the specific roles of different FABPs in different cell populations of the same organ. However, immunohistochemical evidence for the localization of FABPs in particular cell types of the gastric epithelium is limited. Studies using adult rat stomachs have described the localization of L-FABP in brush cells and D cells c 1991 WILEY-LISS. INC (Iseki and Kondo, 1989, 1990) and of H-FABP in parietal cells (Kanda e t al., 1989). In the present study, we have utilized specific antibodies to L-, H-, and I-FABPs to elucidate the ontogenic changes in the relative amount and localization of the three FABPs in the rat stomach. MATERIALS AND METHODS Antibodies L-, H-, and I-FABPs were purified from rat materials, antisera against them were raised in rabbits, and the IgG fractions were purified as described previously (Takahashi e t al., 1983; Iseki e t al., 1989; Kanda e t al., 1989, 1990). Anti-L-FABP antibody was further puritied by affinity chromatography using a n antigen column prepared with Sephadex 4B and L-FABP. For negative control experiments, antibodies were absorbed overnight by 100 Lgiml of corresponding FABPs prior to use. For anti-L-FABP antibody, the pass-through fraction of the affinity chromatography was also used as a control. Received February 6. 1990: accepted May 17, 1990. Address reprint requests to Dr. S. Iseki, Dept. Anatomy, Kanazawa University School of Medicine, Kanazawa. Ishikawa 920, J a p a n . 52 S. ISEKI ET AL. Animals Male Wistar rats were used in the present study except for the homogenization of fetal rat stomachs, which was performed without identification of the sex of animals. Rats of the following ages were used: embryonic day 18 (E18), E19, E20, E21, postnatal day 0 (newborn, about 8 hours after birth), 3,7,10,14,17,21, 24,28,35,42, and 56 (adult).The day of the appearance of spermatozoa in the vaginal smear was counted a s EO. The animals were grown with free access to mother’s milk and chow diet in the first 3 postnatal weeks. At 21 days of age, they were separated from their mothers and subsequently grown on standard chow and water. The animals were sacrificed in the morning either by decapitation (prenatal animals) or by transcardial perfusion with cold physiological saline under Nembutal anesthesia (postnatal animals). lrnrnunochernical Studies Immediately after the sacrifice of animals, whole stomachs were removed, cleared of contents by cold physiological saline, and homogenized in ice-cold 10 mM potassium phosphate, pH 7.4, containing 150 mM KCl. At least ten stomachs were combined to obtain the fetal materials. The homogenates were centrifuged for 105,OOOgto yield a cytosolic protein preparation. In the adult rats, cytosols of the liver, jejunum, and heart were also prepared. The cytosolic protein samples (5 pg) were separated by electrophoresis in 15% (W/V) acrylamide slab-gels at pH 8.8 in the presence of 0.1% (WIV) SDS (Laemmli, 1970). The proteins were then transferred electrophoretically to nitrocellulose membranes (Schleicher & Schuell) according to Towbin et al. (1979). After blotting, the membranes were incubated for 2 hours in phosphate-buffered saline (PBS) (10 mM sodium phosphate, pH 7.4, 150 mM NaCl) containing 5% (WIV) bovine serum albumin (BSA). Blots were then incubated for 2 hours at room temperature with the antibodies against FABPs diluted in buffer A (0.05% VIV Tween 20 and 1% BSA in PBS). After washing with buffer B (0.05% Tween 20 in PBS) for 30 minutes, the blots were treated for 1hour with biotinylated antirabbit IgG (Dako) diluted at 1:200 in buffer A. They were then washed again with buffer B and subsequently incubated for 1hour with streptavidin-conjugated horseradish peroxidase (Dako) diluted at 1:600 in buffer A. After washing with buffer B, the immunoreaction was visualized by incubation of the membranes with 0.1 mg/ml of diaminobenzidine tetrahydrochloride and 0.01% (VIV) H 2 0 2in 50 mM Tris-HC1, pH 7.6. lrnrnunocytochernical Studies For preparing the histological specimens, postnatal animals were perfused transcardially first with physiological saline and then with 4% (WIV) paraformaldehyde in 0.1 M phosphate buffer, pH 7.4. The stomach specimens were excised and immersed in the same fixative for a n additional 4 hours. The stomach specimens from the decapitated prenatal animals were also fixed by immersion in the same fixative. Three animals of each age were used for histology. After a rinse with phosphate buffer, all tissue blocks were immersed overnight in 30% (WIV) buffered sucrose at 4°C. For light microscopic immunohistochemistry, thick (15 pm) frozen sections were cut by a cryostat and mounted on gelatin-coated glass slides. They were incubated with the antibodies against FABPs diluted with PBS overnight a t room temperature. The sites of the immunoreactions were visualized by the peroxidase-antiperoxidase (PAP) method (Sternberger, 1974). Details of the immunohistochemical procedures have been described previously (Iseki et al., 1989). For electron microscopic immunocytochemistry, the cryostat sections immunostained by the PAP method were postfixed in 0.5% (WIV) Os04 in 0.1 M phosphate buffer for 15 minutes and stained e n bloc with 1% (WI V) uranyl acetate for 20 minutes. They were then dehydrated in a n alcohol series and embedded in Epon 812 (Serva) to obtain ultrathin (0.1 pm) sections for electron microscopy. RESULTS lrnrnunochernical Analysis of Cytosols The specificity of the antibodies raised against the three FABPs was examined by Western blot analysis of cytosols from various tissues of adult (56 day) rats. As shown in Figure 1, reactivity with anti-L-FABP antibody was prominent in the liver and jejunum in the 14 kD region, whereas it was negligible in the heart and stomach. Anti-HI-FABP antibody formed a clear protein band of 15 kD in the stomach a s well as in the heart. Reactivity with anti-I-FABP antibody in the 15 kD region was observed primarily in the jejunum and also in the stomach at a much lower intensity. These results confirmed the FABP-type specificity of the three antibodies used in the present study and also were consistent with the previous biochemical result that the major type of FABP in the adult rat stomach is H-FABP (Kanda et al., 1989). Next, whole stomachs of rats from various stages of development were homogenized and the cytosolic fraction was analyzed by Western blots using the three antibodies (Fig. 2). The reactivity of the 14 kD protein band to anti-L-FABP antibody, which was scarcely detectable a t E18, showed a marked increase in intensity after birth, reaching a peak at 7 days, and remained high in the first 2 weeks of postnatal life. It then declined substantively in intensity at 21-28 days and became scarcely detectable at 42-56 days after birth. It was noted that the reactivity was transiently stronger a t 28 days than a t 21 days. In contrast, reactivity to anti-H-FABP in the 15 kD region was continuously present from birth to adulthood and increased substantially in intensity between 14 and 21 days. I n the immunoblot using anti-I-FABP antibody, the intensity of reaction in the 15 kD region increased gradually until 21 days after birth and remained constant thereafter. lrnrnunocytochernical Localization of Three FABPs in the Developing Stomach The immunoreactivity for L-FABP was first detected in the gastric mucosa of E19, with localization to the superficial epithelium (Fig. 3a). The immunopositive epithelial cells were arranged with the negative cells in a mosaic manner. Immunoelectron microscopy of the L-FABP-reactive cells identified them a s the surface mucous cells by their apical cytoplasm containing mucous globules (Fig. 3b). Although the immunoreactive - H-FABP L FABP Mr (kD) 53 FABPs I N DEVELOPING RAT STOMACH I =FABP 4329- L J H S L Fig. 1. Western blots of cytosolic proteins from various organs shown by immunostaining for FABPs. Samples of total cytosolic proteins (5 pgilane) from the liver (L), jejunum (J),heart (HI, and stomach (S)of adult rats were subjected to SDS/polyacrylamide-gel elec- Ageldaysl El8 PO 3 7 J H S L J H S trophoresis followed by electrophoretic transfer to nitrocellulose. The blots were immunostained with the antibodies against L-, H-, and I-FABPs. 10 14 21 28 42 56 L- FABP - H FABP I -FABP Fig. 2. Western blots of cytosolic proteins from the stomachs in various developmental stages shown by immunostaining for FABPs. Samples of total stomach cytosolic proteins (5 pgilane) from rats a t ages of E l 8 to 56 days were electrophoresed, blotted, and immunostained for L-, H-, and I-FABPs. The immunopositive protein bands of 14 kD (L-FABP) and 15 kD (H- and I-FABPs) are shown. material was distributed throughout the cytoplasm, a circumscribed intense reaction was noted around the empty spaces representing the glycogen areas, which are abundant in this cell type a t prenatal stages (Helander, 1969b). No immunoreaction was observed in the parietal cells, mucous neckichief cells. Approximately 8 hours after birth, L-FABP-immunoreactive cells still constituted only a part of the total surface mucous cell population in both the fundic and pyloric regions (Fig. 4a). The glycogen areas were no longer present, but instead large lipid droplets were noted in the cytoplasm (Fig. 4b). No circumscribed enrichment of the immunoreaction was observed around the lipid droplets of reactive cells. Occasionally, solitary L-FABP-immunoreactive cells with a quite different figure from that of surface mucous cells were encountered in the superficial epithelium of the newborn gastric mucosa (Fig. 5a). These cells were smaller than the surface mucous cells and had a barrel-like shape with a constricted apical portion protruding to the gastric lumen. In appropriate sections, a thin cytoplasmic process was seen extending from the basal portion of the cell toward the basement membrane. In immunoelectron micrographs, no mucous globules were found in the apical cytoplasm of these cells (Fig. 5b). Instead, they had closely packed large microvilli, approximately 0.7 pm long, straight microfilaments extending from the cores of the microvilli, and numerous round vesicles 50-300 nm in diameter in the supranuclear cytoplasm (Fig. 5c). By such apical features, these cells were identified as the brush cells (Isomaki, 1973; Nabeyama and Leblond, 1974; Iseki and Kondo, 1989). The immunoreactive material appeared diffusely in the cytoplasmic matrix of the brush cells without association with subcellular structures. A careful observation of the newborn gastric epithelium failed to reveal any brush cells that were immunonegative for LFABP. 54 S.ISEKI ET AL FABPs I N DEVELOPING RAT STOMACH Through 3 to 14 days after birth, i.e., in the suckling period, the immunoreactivity for L-FABP was observed in nearly all superficial epithelial cells and also in the cells located in the gastric pits and upper portions of the gastric and pyloric glands (Fig. 6a). In immunoelectron microscopy, all surface mucous cells, including relatively immature ones in the gastric pits, a s well a s the parietal cells located in the pits and upper portions of the gastric glands were immunoreactive (Fig. 6b). The mucous neck cells and chief cells, which were indistinguishable in the early postnatal days, were immunonegative for L-FABP. The reactive brush cells were found only occasionally throughout the first 2 postnatal weeks. In addition, starting around 7 days after birth, another type of immunoreactive cell became evident in the lower portions of the gastric and pyloric glands (Fig. 7a). By immunoelectron microscopy, they were identified a s the endocrine D cells by their characteristic basal granules, 200-300 nm in diameter, with limiting membranes very closely applied to moderately electron-dense cores (Figs. 7b, c; Alumets et al., 1977). Empty spaces seemingly representing lipid droplets were often found in this cell type during the suckling period. By a careful survey of the gastric epithelium by immunoelectron microscopy, the D cells without immunoreaction for L-FABP were also found. The immunoreaction product appeared evenly in the entire cytoplasmic matrix of reactive D cells without association with subcellular structures, although it could not be determined conclusively if the cores of the endocrine granules were free of the reaction. From 17 to 28 days after birth, i.e., in the weaning period, there was a progressive decrease in the intensity of L-FABP-immunoreactivity in the surface mucous cells, leading eventually to no immunostaining of 55 this cell type by 28 days (Figs. 8-11). The reactivity of the parietal cells located in the gastric pits also decreased progressively in this period. However, the parietal cells located in the neck region of the gastric glands showed a transient rise in the reactivity to antiL-FABP antibody a t 24-28 days (Fig. lo), probably accounting for the quantitative result of the immunoblot analysis (Fig. 2). A remarkable rise in the number of immunoreactive brush cells was noted after 17 days, reaching the adult value (approximately a cell in every five gastric pits; Iseki and Kondo, 1989) by 28 days after birth. A group of immunopositive brush cells in the distal wall of the groove between the forestomach and glandular stomach (gastric groove) became apparent around 24 days (Fig. 9). The number of immunoreactive D cells also increased progressively in this period. After 28 days, the immunoreactivity in the neck parietal cells disappeared rapidly. As a consequence, in the mature stomach of 42-56 days postpartum, the reactivity to anti-L-FABP-antibody was found exclusively in the brush cells located in the superficial epithelium and gastric pits and in the endocrine D cells located in the gastric glands (Fig. 11). Reactivity to anti-H-FABP antibody was first detected in the stomach of E20 (Fig. 12a) and localized exclusively in the parietal cells, which were so recognized by the presence of intracellular secretory canaliculi (Fig. 12b; Helander, 1969a). In this stage, there was a marked heterogeneity in the parietal cell population in terms of the intensity of immunoreaction. Within the first postnatal week, all parietal cells in the stomach became intensely immunoreactive for H-FABP and remained so to adulthood (Fig. 13). No reactivity to anti-H-FABP antibody was detected in any cell types other than the parietal cells. The immunoreaction product was distributed evenly throughout the cytoplasmic matrix of the parietal cells, whereas interior of the intracellular canaliculi, mitochondria, endoplasmic reticulum, and Golgi apparatus was free of the reaction Fig. 3.a: Light micrograph of the gastric mucosa of E l 9 treated by (Fig. 12c). Furthermore, no regional enrichment of the immunohistochemistry for L-FABP. Immunoreactive epithelial cells immunoreactivity was discerned in the cytoplasmic are seen in a mosaic pattern. x 300. b: Immunoelectron micrograph of portion adjacent to any of those subcellular structures. the El9 gastric mucosa. A surface mucous cell intensely immunoreThe reactivity to anti-I-FABP antibody was first deactive for L-FABP is shown. The cytoplasmic immunoreactivity is enriched in intensity around the glycogen area (Gy). Arrows indicate tected in the newborn gastric mucosa (Fig. 14a) and localized exclusively in the surface mucous cells (Fig. the mucous granules. x 13,700. 14b). The immunopositive cells were intermingled with Fig. 4.a: Light micrograph of the newborn gastric mucosa immu- the negative cells a s in the case of L-FABP. This mosaic nostained for L-FABP. Immunoreactive cells occur in the surface arrangement of I-FABP-immunoreactive cells was reepithelium in a mosaic pattern. M, muscularis mucosae. ~ 3 0 0b. Immunoelectron micrograph of the newborn gastric mucosa. L-FABP- placed after 3 days by eventually all the surface muimmunoreactive surface mucous cells are shown. The reaction product cous cells lining the luminal surface of fundic and pyis distributed evenly in the entire cytoplasm. L, lipid droplets. loric mucosa reactive to the antibody. These cells were x 5,500. continuously immunoreactive into adulthood (Fig. 15). Fig. 5. a: Light micrograph of the newborn gastric mucosa immu- However, the relatively immature cells of the same nostained for L-FABP showing an immunoreactive brush cell (arrow). type, located in the deep regions of the gastric pits, The basal cell process of the brush cell is not contained in this parremained immunonegative for I-FABP. The immunoticular section. x 600. b: An immunoelectron micrograph showing a L-FABP-immunoreactivebrush cell found in the newborn gastric mu- reactive material was distributed evenly throughout cosa. Note large microvilli (Mv)protruding to the gastric lumen and the cytoplasmic matrix of the reactive surface mucous the absence in apical cytoplasm of the mucous granules characteristic cells. of the surface mucous cells. An empty space (asterisk) presumably Incubation of the sections with the three control anrepresents a lipid droplet. In this particular section, the immunoreaction appears stronger in the apical and lower lateral cytoplasm, tibodies preabsorbed by corresponding types of FABPs brought about no immunostaining of any portion of the probably due to uneven penetration of the antibody. ~ 9 , 8 0 0 c: . Higher magnification of the apical portion of the brush cell shown in stomach, confirming that the observed immunoreacb. Note large microvilli (Mv), extension of the core microfilaments tions were specific to the three FABPs. Table 1 sum(Mf) deep into the cytoplasm, bundles of intermediate filaments (IF), and numerous vesicles (V). D, desmosomes; Mi, mitochondrion. marized the occurrence of the three FABPs in various cell types in the developing stomach described above. x 21,000. Scale bar = 500 nm. 56 S. ISEKI ET AL. Fig. 6. a: Light micrograph of the 7 day gastric mucosa immunostained for L-FABP. Immunoreactivity covers the entire surface epithelium and also extends to the upper portions of the gastric glands. M, muscularis mucosa. x 300. b Immunoelectron micrograph of the 7 day gastric mucosa. Not only surface mucous cells (SM) but also the parietal cells located in the upper portions of gastric glands (arrows), are immunoreactive for L-FABP. X 1,700. Fig. 7. a: A light micrograph of the 7 day gastric mucosa immunostained for L-FABP, showing an immunoreactive D cell (arrow)in the DISCUSSION In the suckling rat stomach, considerable hydrolysis of milk fat by lingual lipase is known to take place (Hamosh, 1979) at a pH optimum around 5.0, followed by absorption of fatty acids by the surface mucous cells (Helander and Olivecrona, 1970). The absorbed fatty acids are stored in large lipid droplets, which are not associated with limiting membranes derived from endoplasmic reticulum or Golgi apparatus, unlike the absorptive epithelial cells of the small intestine. The present immunochemical and immunocytochemical evidence for the occurrence of L-FABP in the surface mucous cells during the suckling period and its disappear- lower portion of gastric gland. x 600. b: An immunoelectron micrograph of the 7 day gastric mucosa. A L-FABP-immunoreactive D cell found in the lower portion of gastric gland is shown. Note numerous round endocrine granules in the cytoplasm apposing the basement membrane (BM).Empty spaces (asterisks)presumably represent lipid droplets. The immunoreaction product occurs evenly throughout the cytoplasm. Chief cells ( C ) are immunonegative. x 8,300. c: Higher magnification of the basal portion of the D cell shown in b. Note the endocrine granules with limiting membrane very closely applied to the core. x 22,000. Scale bar = 500 nm. ance after weaning suggests that L-FABP is somehow involved in the absorption of milk fat by the gastric epithelium of suckling rats. On the other hand, I-FABP is also demonstrated immunohistochemically in the surface mucous cells throughout the postnatal development of rats, without relation to suckling and weaning. The amount of I-FABP in the gastric epithelium is probably very low, judging from the relative intensity of immunoreactivity in the immunoblot. The physiological significance of I-FABP in the surface mucous cells remains unclear. The parietal cells are first recognized in the gastric mucosa of the rat at E19-E20 both morphologically FABPs I N DEVELOPING RAT STOMACH 57 Fig. 8. Light micrograph of the 21 day gastric mucosa immunostained for L-FABP. Both the number of immunoreactive surface mucous cells and the intensity of the reactivity are decreased compared with the 7 day gastric mucosa. Note the strongly immunoreactive brush cells (arrowheads). x 300. Fig. 10.A light micrograph of the 24 day gastric mucosa immunostained for L-FABP. Intense immunoreactivity is no longer present in the surface mucous cells. Brush cells (B) in the gastric pits and D cells (D) in the gastric glands are immunoreactive. Also note that many parietal cells (P) located in the neck region of gastric glands are intensely to moderately reactive for L-FABP. x 300. Fig. 9. A light micrograph of the 24 day gastric mucosa immunostained for L-FABP. The junctional region between the forestomach (F) and the glandular stomach is shown. Note a cluster of immunoreactive brush cells in the distal wall of the gastric groove ( G ) . x 300. Fig. 11. Light micrograph of the 56 day gastric mucosa immunostained for L-FABP. Immunoreactivity is no longer present in the surface mucous cells or neck parietal cells and is exclusively localized to the brush cells (B) and D cells (D). x 300. and by their carbonic anhydrase activity (Helander, 1969a). They differentiate progressively after birth a s shown by the continuous decline of the pH of gastric contents. The present results demonstrate that HFABP occurs in the parietal cells from very early stages of their development and is continuously expressed in this cell type. Thus, H-FABP is a good histochemical marker of parietal cells. The relative amount of H-FABP, a s shown by the immunoblot analysis, appears to increase in coincidence with the weaning of animals. In the context of the functional role of H-FABP, past studies have demonstrated that H-FABP of the rat is widely expressed in the energy-consuming organs such as the heart, muscle and kidney (Bass and Manning, 1986; Paulussen et al., 1989). It is generally believed that H-FABP acts in these tissues as a n intracellular carrier of fatty acids to mitochondria for 6-oxidization (Fournier et al., 1978; Glatz and Veerkamp, 1985). Because the parietal cells of the stomach probably undergo active energy consumption, a s suggested by the presence of numerous mitochondria in their cytoplasm, it is plausible that H-FABP is involved in this cell type in the utilization of plasma-free fatty acids as energy sources. On the other hand, the present study also reveals L-FABP to be expressed in a subpopulation of the 58 S.ISEKI ET AL. Figs. 12-1 5 59 FABPs IN DEVELOPING RAT STOMACH TABLE 1. Ontogenic appearance of the immunoreactivity for three FABPs in the rat stomach FABP type L H I Cell type Surface mucous cell Parietal cell Brush cell D cell Parietal cell Surface mucous cell El8 19 20 PO 7 -' +- + ++ + + + ++ + ~ - - - - ++- - - + + - - - + - ~ Age (days) 14 21 24 28 42 56 - - ++ + + + - - ++ + + ++ + ++ + ++ + ++ + ++ + ++ + + + + 'The intensity of the immunoreactivity and the number of immunoreactive cells are synthetically expressed a s parietal cells throughout the suckling period followed by its transient increase after weaning. The functional significance of L-FABP in the parietal cells remains unclear. In addition to the two cell types described previously, the present study reveals two minor cell types in the gastric epithelium to express L-FABP. The D cells, identified by their characteristic endocrine granules that contain somatostatin, are known to appear in a limited number in the rat gastric epithelium a few days after birth and to increase in number linearly from 10 days to 3 weeks (Alumets et al., 1977). The present results suggest that D cells express L-FABP from the early stages of their ontogenic development. However, it should also be noted that only a subpopulation, about two thirds in the adult case (Iseki and Kondo, 19901, of the total D cell population in the stomach are immunoreactive for L-FABP throughout development. The brush cells, also termed tuft cells, fibrillovesicular cells, or caveolated cells, represent a rarer but distinct cell population found in the endoderm-derived mucosal epithelia including the gastric epithelium (Isomaki, 1973; Nabeyama and Leblond, 1974; Iseki and Kondo, ~ - - ++ ++ + ++ + ++ + + to + t 1989). The nature of this cell type is largely unknown, and very little information is available on its ontogeny. The present work reveals for the first time that the brush cells appear in the rat gastric mucosa as early a s the day of birth, with intense immunoreactivity for LFABP. Unlike D cells, all brush cells in the stomach seem to be immunoreactive. Based on this selective reactivity, it is concluded that the frequency of brush cells remains low in the rat gastric epithelium during the first 2 weeks of postnatal development but increases remarkably in the following 2 weeks. This is coincident with the end of the suckling period and the beginning of the weaning period. In the adult rat stomach, L-FABP, although low in amount in the whole stomach homogenate, is exclusively expressed in D cells and brush cells, as reported previously (Iseki and Kondo, 1990). The occurrence of L-FABP in these minor cell types, which are generally considered not to be involved in active fat absorption, implies further aspects of the role played by L-FABP. The present study indicates that structurally different FABPs are expressed in the rat stomach, depending on the cell types and developmental stages. ACKNOWLEDGMENTS Fig. 12 a: Light micrograph of the E20 gastric mucosa immunostained for H-FABP. Immunoreactive cells are mainly found in the developing gland epithelium. M, muscularis mucosa. x 300. b Immunoelectron micrograph of the E20 gastric mucosa. Two parietal cells identified by t h e intracellular secretory canaliculi (asterisks) a r e shown, one with intense and the other with weaker immunoreactivity for H-FABP. A surface mucous cell (SM) is immunonegative. x 6,100. c: Higher magnification of the intensely H-FABP-immunoreactive parietal cell shown in b. Note that the immunoreaction product is distributed evenly in the cytoplasmic matrix, but the interior of the membranous organelles, such a s intracellular canaliculi (IC), mitochondria (Mi), endoplasmic reticulum (ER) and Golgi apparatus (G), is free of the reaction. x 17,000. Fig. 13. A light micrograph of the 56 day gastric mucosa immunostained for H-FABP. Parietal cells in the entire thickness of the mucosa a r e intensely immunoreactive. M, muscularis mucosae. x 150. Fig. 14. a: A light micrograph of the newborn gastric mucosa immunostained for I-FABP. Immunoreactive cells a r e found in the surface epithelium. x 300. b Immunoelectron micrograph of the newborn gastric mucosa. Surface mucous cells with various intensities of I-FABP-immunoreactivity are shown. The reaction product occurs evenly throughout t h e cytoplasm. x 5,500. Fig. 15. Light micrograph of the 56 day gastric mucosa immunostained for I-FABP. Immunoreactivity is localized exclusively to the surface mucous cells. x 300. The authors are indebted to Mr. S. Yamazaki for photographic work. This work was supported in part by Grant-in-Aid 01570006 from the Ministry of Education, Science and Culture, Japan. LITERATURE CITED Alpers, D.H., A.W. Strauss, R.K. Ockner, N.M. Bass, and J.I. Gordon 1984 Cloning of a cDNA encoding rat intestinal fatty acid binding protein. Proc. Natl. Acad. Sci. U.S.A., 81t313-317. Alumets, J., F. Sundler, and R. Hakanson 1977 Distribution, ontogeny and ultrastructure of somatostatin immunoreactive cells in the pancreas and gut. Cell Tissue Res., 185:465-479. Bass, N.M., and J.A. Manning 1986 Tissue expression of three structurally different fatty acid binding proteins from rat heart muscle, liver, and intestine. Riochem. Biophys. Res. Commun., 137: 929-935. Fournier, N.C., M. Geoffroy, and J. Deshusses 1978 Purification and characterization of a long-chain fatty acid-binding protein supplying the mitochondria1 B-oxidative system in the heart. Biochim. Biophys. Acta, 533t457-464. Glatz, J.F.C., and J.H. Veerkamp 1985 Intracellular fatty acidbinding proteins. Int. J. Biochem., 17t13-22. Gordon, J.L., N. Elshourbaby, J.B. Lowe, W.S. Liao, D.H. Alpers, and J.M. Taylor 1985 Tissue specific expression and developmental regulation of two genes coding for r a t fatty acid binding proteins. J. Biol. Chem., 260:1995-1998. Hamosh, M. 1979 A review. Fat digestion in the newborn: role of 60 S. ISEKI ET AI, lingual lipase and preduodenal digestion. Pediatr. Res., 13t615622. Helander, H.F. 1969a Ultrastructure and function of gastric parietal cells in the rat during development. Gastroenterology, 56t35-52. Helander, H.F. 1969b Ultrastructure and function of gastric mucoid and zymogen cells in the rat during development. Gastroenterology, 56.53-70. Helander, H.G., and T. Olivecrona 1970 Lipolysis and lipid absorption in the stomach of the suckling rat. Gastroenterology, 59.22-35. Heuckeroth, R.O., E.H. Birkenmeier, M.S. Levin, and J.I. Gordon 1987 Analysis of the tissue-specific expression, developmental regulation, and linkage relationships of a rodent gene encoding heart fatty acid binding protein. J. Biol. Chem., 262.9709-9717. Iseki, S., and H. Kondo 1989 Specific localization of hepatic fatty acid-binding protein in the gastric brush cells of rats. Cell. Tissue Res., 257t545-548. Iseki, S., and H. Kondo 1990 An immunocytochemical study on the occurrence of liver fatty-acid-binding protein in the digestive organs of rats: specific localization in the D cells and brush cells. Acta Anat., I38t15-23. Iseki, S., M. Hitomi, T. Ono, and H. Kondo 1989 Immunocytochemical localization of hepatic fatty acid binding protein in the rat intestine: effect of fasting. Anat. Rec., 223.283-291. Isomaki, A.M. 1973 A new cell type (tuft cell) in the gastrointestinal mucosa of the rat. Acta Pathol. Microbiol. Immunol. Scand. ISuppl. A], 240r1-35. Kanda, T., S. Iseki, M. Hitomi, H. Kimura, S. Odani, H. Kondo, Y. Matsubara. T. Muto. and T. Ono 1989 Purification and characterization of a fattv-acid-bindine urotein from the gastric mucosa of rats Eur J Blochem , 185 27-33 Kanda, T ,T Ono, Y Matsubara, and T Muto 1990 Possible role of rat fatty acid-binding proteins in the intestine as carriers of phenol and phthalate derivatives. Biochem. Biophys. Res. Commun., 168:1053-1058. - Laemmli, U.K. 1970 Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227:680-685. Nabeyama, A., and C.P. Leblond 1974 Caveolated cells characterized by deep surface invaginations and abundant filaments in mouse gastro-intestinal epithelia. Am. J. Anat., 140:147-166. Ockner, R.K., J.A. Manning, and J.P. Kane 1982 Fatty acid binding protein. Isolation from rat liver, characterization, and immunochemical quantification. J. Biol. Chem., 257t7872-7878. Ockner, R.K., J.A. Manning, R.B. Poppenhausen, and W.K.L. Ho 1972 A binding protein for fatty acids in cytosol of intestinal mucosa, myocardium, and other tissues. Science, 177t56-58. Paulussen, R.J.A., M.J.H. Geelen, A.C. Beynen, and J.H. Veerkamp 1989 Immunochemical quantitation of fatty-acid-binding proteins. I. Tissue and intracellular distribution, postnatal development and influence of physiological conditions on rat heart and liver FABP. Biochim. Biophys. Acta, 1001t201-209. Said, B., and H. Schultz 1984 Fatty acid binding protein from rat heart. The fatty acid binding proteins from rat heart and liver are different proteins. J. Biol. Chem., 259t1155-1159. Sternberger, L.A. 1974 Immunocytochemistry. Prentice Hall Inc., Englewood Cliffs, New Jersey. Sweetser, D.A., R.O. Heuckeroth, and J.I. Gordon 1987 The metabolic significance of mammalian fatty-acid-binding proteins: Abundant proteins in search of a function. Annu. Rev. Nutr., 7t337357. Takahashi, K., S. Odani, and T. Ono 1983 Isolation and characterization of the three fractions (DE-I, DE-I1 and DE-111) of rat-liver Z-protein and the complete primary structure of DE-11. Eur. J. Biochem., 136:589-601. Towbin, H., T. Staehelin, and J. Gordon 1979 Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: Procedure and some applications. Proc. Natl. Acad. Sci. USA, 76: 4350-4354.