INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

GASTRIC UNITS IN THE zymogenic zone of the mouse and rat contain five principal epithelial lineages (11): mucous surface cells, parietal cells, zymogenic cells that contain intrinsic factor (IF) and pepsinogen, enteroendocrine cells, and brush or caveolated cells. The gastric gland is divided vertically into three regions: the pit, roughly corresponding with the upper third; the isthmus and neck, corresponding to the middle third; and the base or lower third of the gland (11). In the rat, the neck and isthmus regions contain four cell types: nondifferentiated cells, immature surface cells, mucous neck cells, and neck parietal cells (4, 15). The neck parietal cell appears to have features intermediate between the immature surface cell and the mature parietal cell found at the base of the gastric gland (4). The neck parietal (pre-parietal) cells do not appear to divide but acquire differentiated functions [e.g., H⁺-K⁺-adenosinetriphosphatase (H⁺-K⁺-ATPase), canaliculi] as they migrate from the isthmus to the neck and eventually to the basal region of the gastric gland (15). In the mouse, the cell dynamics appear to be similar, but the quantitation of the precursor cell types has been better delineated. When pre-neck precursors arrive at the upper portion of the base region of the gland, they differentiate into pre-neck (pre-zymogenic) cells with granules (12). Transformation of most (98%) pre-neck cells into zymogenic cells occurs as the cells continue to descend into the lower portion of the gastric gland. Murine pre-parietal cells, on the other hand, arise from all three major granule-free precursor cell types (pre-pit, pre-parietal, pre-neck), with an estimated 2% of parietal cells deriving from the pre-neck precursors (12).

IF production in the rat (3) and mouse (18) has been thought to occur exclusively in chief cells. This finding was confirmed originally in the rat, using parietal cell preparations >80% pure isolated by Percoll gradients (23). However, work from our laboratory (16), using antiserum raised against native rat IF, first suggested that some parietal cells expressed IF, particularly in the region of the isthmus and neck, although the reactivity was not strong. Moreover, analysis of the rat stomach by in situ hybridization revealed weak reactivity over some parietal cells (5). A more recent study reported no signal over rat parietal cells using radioactive probe RNA (19). In contrast, in the mouse IF is found only in differentiated zymogen cells, despite the fact that pepsinogen is also found in mucous neck cells in the mucoparietal section of the gastric mucosa (18). Finally, IF has been demonstrated in multiple cell types in human gastric mucosa (9), in which the parietal cell is the usual site for immunocytochemical localization (17). Subsets of positive chief and enteroendocrine cells contained immunoreactive IF, both at the light and electron microscopic level (9).

To resolve these discrepant findings and to test if IF could be used as a marker for a subset of parietal cells that might derive from chief cell precursors, we localized IF and another chief cell marker, pepsinogen, by single- and double-labeled immunocytochemistry in the rat glandular stomach and in isolated parietal cells from that tissue. We used monospecific, high-titer antisera raised against recombinant human IF, a peptide from rat H⁺-K⁺-ATPase, and human pepsinogen II that recognizes rat pepsinogen. Because IF has never been found in ectopic locations in gastric mucosa (18), the mouse was not chosen for these studies, despite the fact that cellular proliferation kinetics are better understood in that model.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Animals. Male rats (180 g) obtained from Sasco (Omaha, NE) were fasted overnight. After rats were killed by cervical dislocation, we removed the glandular stomach. Next, sections of the body (containing the zymogenic and parietal cells) and antrum to be used for tissue immunocytochemistry were fixed in 10% buffered Formalin for 2 h at room temperature before exposure to 70% ethanol overnight at 4°C. For gastric mucosal cell isolation, the rat mucosal cells were first released by enzymatic digestion (Pronase E). Then the cells were separated according to size and density by sequential use of counterflow elutriation and density gradient centrifugation to yield a highly enriched chief cell fraction containing <1% parietal cells and a parietal cell fraction enriched almost to purity (24). The purity of the parietal cell preparation was judged by the addition of nitro blue tetrazolium (Sigma Chemical, St. Louis, MO) at a concentration of 1 mg/ml to a fresh parietal cell suspension (10⁶ cells/ml) air dried on a slide. After incubation at 37°C for 30 min, the slides were washed with phosphate buffer. Under the light microscope the parietal cells appear dark blue to purple, and nonparietal cells are not stained. Cell pellets derived from both parietal and chief cell fractions were prepared by fixing an aliquot of the fraction in Bouin's solution for 2-4 h, followed by centrifugation at 6,000 g for 10 min and storage in 70% ethanol at 4°C until used. Three separate groups of isolated cells were examined.

Immunocytochemistry. Enriched chief and parietal cell pellets or gastric mucosal tissue blocks were embedded in paraffin, and 5-µm-thick sections were cut and mounted on slides. The sections were subsequently treated with xylene to remove the paraffin and dehydrated in graded ethanol. The standard avidin-biotin-peroxidase complex method was used (26). Sections were treated with 1% H₂O₂ in methanol for 20 min, followed by preincubation for 20 min at 37°C in 0.1 M phosphate-buffered saline (PBS) containing 5% bovine serum albumin and 10% normal goat serum. The primary antisera used were rabbit anti-H⁺-K⁺-ATPase -subunit, raised against amino acids 2-23 of the rat protein (1:500) [kindly supplied by Michael Caplan, Yale University (Ref. 18)], rabbit polyclonal anti-human IF (1:200), raised against recombinant human IF produced in baculovirus-infected Sf9 cells (9), and rabbit anti-human pepsinogen II (1:1,000) (the latter provided courtesy of Dr. Michael Samloff, University of California Los Angeles School of Medicine). Anti-rat IF had been shown previously to cross-react with human IF (16), but the availability of anti-human IF was now much greater, because of the markedly increased yield of human IF over rat IF in the baculovirus expression system. Human and rat IF share 80% amino acid identity (8), and as expected, rat and human IF both react strongly with anti-human IF (D. H. Alpers and M. M. Gordon, unpublished observations). The anti-pepsinogen II identifies one of the major pepsinogens in humans (22) and the only pepsinogen found in the rat (14).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Parietal cell distribution in the rat gastric body. Sections of the rat gastric body were examined for the distribution of parietal cells by using GS1 lectin, which binds to parietal cells but not to chief cells (13). Figure 1A shows that the stained cells occupied the lower two-thirds of the gastric glands, including the isthmus and neck (middle third) and base (lower third). Occasional GS1 lectin-stained cells were found in the lower portion of the upper third or pit region. This distribution fits well with the fact that parietal cells originate in the isthmus region from pre-parietal cells (10), but within the isthmus and neck region soon acquire H⁺-K⁺-ATPase-rich tubulovesicles, rudimentary canaliculi, and long microvilli characteristic of maturing parietal cells (12). When the rat gastric body was immunostained for H⁺-K⁺-ATPase (Fig. 1B), the distribution of positive cells was virtually identical with that seen using GS1 lectin (Fig. 1A). This result is consistent with the early appearance of tubulovesicles (the intracellular site of H⁺-K⁺-ATPase) in parietal cells found in the isthmus and neck region (4, 12, 15).

View larger version (131K): [in this window] [in a new window]   View larger version (116K): [in this window] [in a new window]   Fig. 1.   Histocytochemical localization of parietal cells in the rat glandular stomach. Tissue was stained as described in MATERIALS AND METHODS, using avidin-biotin peroxidase complex (ABC) techniques. A: biotinylated Griffonia simplicifolia lectin 1. B: antiserum against human H+-K+-ATPase. Note that parietal cells are distributed in lower 2/3 of the mucosa, comprising the isthmus and neck and basal regions of gastric glands. A and B: original magnification, ×100.

Immunostaining of the rat gastric body using chief cell markers. In contrast to the distribution of the total parietal cell population (Fig. 1), the distribution of IF and pepsinogen was confined largely to the basal third of the gastric glands (Fig. 2, A and C). However, occasional cells with the shape and size of parietal cells were found scattered rather evenly throughout the basal third of the gland; such cells are identified in Fig. 2, B and D (arrows). None of these cells had mucus secretory granules visible by light microscopy. Between 1.5% and 3.1% of the parietal cells in this region stained positively for IF and/or pepsinogen. Most chief (zymogenic) cells were positive using these markers. When double-labeled immunofluorescence was performed with these two antisera, only occasional cells outside the basal third of the gland stained positively with both antisera (Fig. 2E). These cells, similar to all the others, stained yellow in color prints (see Fig. 3), consistent with the presence of both antigens in the cells. Sections stained with normal rabbit serum were consistently negative (Fig. 2F).

View larger version (131K): [in this window] [in a new window]   View larger version (132K): [in this window] [in a new window]   View larger version (136K): [in this window] [in a new window]   View larger version (127K): [in this window] [in a new window]   View larger version (115K): [in this window] [in a new window]   View larger version (92K): [in this window] [in a new window]   Fig. 2.   Immunocytochemical localization of intrinsic factor (IF) and pepsinogen (Pep) in the rat glandular stomach. Immunocytochemistry was performed as described in MATERIALS AND METHODS, using antisera against human IF and pepsinogen II. A-D and F: ABC method utilized. E: double-labeled immunofluorescence used. A and B: anti-IF. C and D: antipepsinogen. E: double-labeled with second antisera labeled with Cy3 (red) for IF and with fluorescein isothiocyanate (FITC) (green) for pepsinogen. F: normal rabbit serum. Note that cell staining is largely confined to lower 1/3 of gastric glands. Most chief cells and some parietal cells (see arrows) are positive for IF and pepsinogen, and no cells are clearly positive for each protein independently. A, C, E, and F: original magnification, ×100. B and D: original magnification, ×400.

View larger version (121K): [in this window] [in a new window]   Fig. 3.   Double-labeled immunofluorescent localization of H+-K+-ATPase, IF, and pepsinogen in the rat glandular stomach. Immunofluorescence was performed as described in MATERIALS AND METHODS using same dilutions used for immunocytochemical studies. A-C: lower half of basal region (lowest third) of gastric glands. Note that parietal cells [PC; stained green for H+-K+-ATPase (H/K)] occur singly throughout this region, even up to the base of the glands. Chief cells (CC; stained red) containing IF but not H+-K+-ATPase are more concentrated at the base of these glands. Double-labeled parietal cells (stained yellow) are fairly frequent in this region. D-F: upper half of basal region of gastric glands. Here, IF-containing parietal cells are less frequent, and these cells also contain pepsinogen, confirming single-labeled study shown in Fig. 2. A-C: original magnification, ×200. D-F: original magnification, ×400.

Immunostaining of isolated parietal and chief cell preparations. To better characterize and document the subset of parietal cells expressing IF and pepsinogen, we prepared rat gastric mucosal cell fractions enriched for chief and parietal cells. In the chief cell fraction, 87-89% of the cells were positive for IF and pepsinogen (Fig. 4, A and B). In double-labeled fluorescence studies, no cells were identified that were positive for one but not the other protein (see Fig. 5). No freshly isolated cells stained positively with nitro blue tetrazolium as expected for parietal cells, but only a small number (<1%) of the fixed chief cell population stained positively using antibodies against H⁺-K⁺-ATPase or using GS1 lectin, both markers for parietal cells in the rat. When almost pure (>95%) parietal cell preparations were examined in the same way the results were quite different. In one preparation, 3.5% of the parietal cells that were identified by positive nitro blue tetrazolium staining were positive for IF and/or pepsinogen (Fig. 4, C and D). Under higher power magnification, the cells that were positive for IF did not appear to be morphologically different from the other cells (Fig. 4, E and F). In other preparations, 4% and 11% of the isolated parietal cells were positive. Normal rabbit serum applied to chief cells (Fig. 4G) and parietal cells (not shown) displayed no reactivity.

View larger version (143K): [in this window] [in a new window]   View larger version (146K): [in this window] [in a new window]   View larger version (125K): [in this window] [in a new window]   View larger version (126K): [in this window] [in a new window]   View larger version (110K): [in this window] [in a new window]   View larger version (112K): [in this window] [in a new window]   View larger version (80K): [in this window] [in a new window]   Fig. 4.   Immunocytochemical staining of isolated rat chief and parietal cells. A: chief cells (CC), anti-IF, no counterstain. B: chief cells, antipepsinogen (PEP), no counterstain. C: parietal cells (PC), anti-IF counterstained with hematoxylin and eosin. D: parietal cells, antipepsinogen counterstained with hematoxylin and eosin. E: parietal cells, anti-IF, no counterstain. F: parietal cells, anti-IF counterstained with hematoxylin and eosin. G: chief cells, normal rabbit serum. Most isolated chief cells, but only 3-4% of isolated parietal cells, are positive for both IF and pepsinogen. Arrow in A marks a chief cell that stained very positively for IF. Other arrows mark parietal cells staining positively for IF (C, E) and for pepsinogen (D). E and F: * indicates a parietal cell unstained (E) and stained (F) with hematoxylin and eosin. A-D and G: original magnification, ×200. E and F: original magnification, ×800.

View larger version (127K): [in this window] [in a new window]   Fig. 5.   Double-labeled immunofluorescence of isolated parietal cells. Immunofluorescent staining was carried out as described in MATERIALS AND METHODS. A and D: cells reactive with anti-IF first antibody that were detected with Cy3 (red)-labeled second antibody. B: cells reactive with antipepsinogen first antibody and detected with FITC (green)-labeled second antibody. C: use of both anti-IF and antipepsinogen first antibodies. Note that all positive parietal cells react with antibody against both chief cell markers. E: anti-H+-K+-ATPase with FITC (green)-labeled second antibody. F: anti-IF and anti-H+-K+-ATPase. Note that all IF positive cells are also positive for the parietal cell marker. A-C: arrows mark parietal cells positive for the chief cell markers IF and pepsinogen. D-F: upper arrows mark parietal cells positive for IF and for the parietal cell marker H+-K+-ATPase; lower arrows identify a cell that is H+-K+-ATPase positive but IF negative. A, D, and F: original magnification, ×400.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The findings in the present study confirm the original observations of Lee et al. (16) that some parietal cells in the rat express IF. We have extended this observation by demonstrating pepsinogen expression in the same cells and by estimating the percentage of parietal cells that express these chief cell markers in the glandular portion of rat stomach and in a purified preparation of parietal cells. Double-labeled experiments using the parietal cell marker H⁺-K⁺-ATPase confirmed that the cells expressing IF and pepsinogen were indeed parietal cells. Moreover, the isolated parietal cells expressing the chief cell markers were morphologically the same as the rest of the parietal cell preparation. These data are somewhat different from those reported by Maeda et al. (19), in which in situ hybridization with rat IF mRNA was used. Maeda et al. (19) report no signal over parietal cells, although some low level signal does appear to be present in their study. IF mRNA was seen at low levels in rat parietal cells in another study from our laboratory (5). Maeda et al. (19) conclude that "the primary transcription site of the rat IF gene is gastric chief cells," a conclusion with which the present data agree. It seems likely that expression of IF in a small percentage of parietal cells might not have been detected, especially if the protein concentration in parietal cells was less than in chief cells (16).

The percentage of parietal cells that express chief cell markers appears to range from 1.5% to 10%. This is the range of positive cells found in gastric glands both in whole tissue sections (1.5-3.1%), where the prevalence differs from one gland to another (3-10%), and in preparations of isolated parietal cells (4-11%). There may be several explanations for this variability. First, there could be biological variation from one set of animals to another. Second, there could be differences in sampling from different regions of the stomach. Third, local tissue factors could stimulate production of IF and pepsinogen in parietal cells. These latter two possibilities are supported by the variability in positive cells from gland to gland and the increased abundance of IF-expressing parietal cells toward the base of the glands. Fourth, it is possible that during cell isolation, a process that takes a number of hours, IF synthesis might remain active, while IF secretion is impaired. This discrepancy might increase the sensitivity of IF detection in parietal cells that do not appear to contain the protein in situ. Such a phenomenon has been seen with isolated enterocytes producing apolipoprotein B (apo B) (7). Although it was thought from this study (7) that enterocyte content of apo B increased after fat feeding, in fact the content fell in tissue (1). However, that the variability was seen both in tissue and in isolated cells somewhat opposes this methodological explanation. The important point is that there does exist a subset of parietal cells in the rat that express chief cell markers, and these cells are a small percentage of the total parietal cell population. There are other examples of subsets of epithelial cells in the gastrointestinal tract that express a product unique to that cell type. Such examples would include the cystic fibrosis transmembrane regulator (2).

A pyloric glandular region is present in fish, adult amphibians, reptiles, and mammals, and in amphibians the mucosal structure is less complex because it only includes three cell types (mucous, oxynticopeptic, and endocrine) instead of the four (mucous, parietal, zymogenic, and endocrine) found in mammals (6). The fact that oxyntic (parietal) and peptic (zymogenic) functions have been combined in a single cell during phylogeny makes the ectopic location of IF and pepsinogen seem more logical. Pepsinogen (but not IF) is produced in the oxynticopeptic cell. One can speculate that the ability to produce pepsinogen (and by inference IF) is retained in the parietal and zymogenic cells, which appear to derive from their ancestral cell in amphibians.

The significance of the IF and pepsinogen-containing parietal cell subset and its origin in the rat are not clear. It is intriguing to think that this subset may correspond with that proportion of parietal cells that derive from the pre-neck lineage. In the mouse this lineage supplies ~2% of parietal cells (12). The regulatory factors that would lead to expression of IF in this subset of cells are not known. Nucleotides from 1029 to +55 in the 5' upstream region of the mouse IF gene direct IF expression only in parietal cells (18). By Harr analysis, the first 400 base pairs of the 5' upstream sequence of the rat IF gene were highly conserved compared with the mouse sequence. Thus, in both mouse and rat, it seems that there are unidentified cis-acting elements that affect the promoter activity and help to direct IF expression in chief cells. Regulation of these elements may allow expression in adult parietal cells, and this regulation may occur in the subset of parietal cells that derive from the pre-neck lineage.

However, expression of IF and pepsinogen does not occur in the isthmus and neck region in which pre-neck cells are first identified, but only when the cells have migrated into the basal region. Thus IF and pepsinogen expression may be the result of maturation of the pre-neck lineage of parietal cells. In fact, a vertical expression gradient consistent with increasing cell maturation is apparent; IF-positive cells increase in frequency toward the base of the gastric glands. It seems possible that local factors may play a role in IF and pepsinogen expression in this subset of parietal cells. The continuous presence of glucocorticoids is necessary for active transcription of pepsinogen mRNA in the adult rat gastric mucosa (25). Many cytokines, especially interferon- and interleukin-1, are important in stimulating gene expression in epithelial cells. Local factors may account for the mosaicism seen in epithelial mucosal cells in the intestine (20, 21). It should be possible to examine the role of local factors in IF and pepsinogen expression in organ explants or in isolated preparations of parietal cells.