Abstract
Parietal cells undergo a differentiation process while they move from the isthmus toward the pits and the base region of the gastric gland. The aim of this work was to analyze the rat gastric glands by lectin histochemistry to show the glycans expressed by upper (young) and lower (old) parietal cells. We used lectins recognizing the most frequent sugar moieties in mammals. Each lectin was assayed alone and in combination with several deglycosylation pretreatments: (1) β-elimination, which removes O-linked oligosaccharides; (2) incubation with Peptide-N-glycosidase F, to remove N-linked glycans; (3) acid hydrolysis, which removes terminal sialic acid moieties; (4) methylation-saponification, to remove sulfate groups from sugar residues; and (5) glucose oxidase, a technique carried out with the lectin concanavalin A to convert glucose into gluconic acid. The lectins from Helix pomatia, Dolichos biflorus (DBA), Glycine max (soybean), Maclura pomifera, Arachis hypogaea (peanut), Bandeiraea simplicifolia (lectin I-B4), and Datura stramonium showed a different glycan expression in the parietal cells throughout the gastric gland. This difference supports that parietal cells undergo a maturation/degeneration process while the cells descend along the gland. The role of DBA as a marker of parietal cells previously reported should be taken with caution because these cells showed different reactivity for the lectin, ranging from negative to strong labeling.
Introduction
The gastric mucosa is divided into different regions. In some laboratory animals, like the rat and the mouse, the proximal third of the stomach, the forestomach, is lined by a stratified squamous epithelium without gastric glands. The second third forms the corpus, which contains oxyntic or fundic glands, and the remaining distal third is the pyloric antrum, with mucous antral glands. Corpus and antrum form the glandular stomach. 1 The glands are open at the bottom of surface depressions, called gastric pits. The gastric pit and the gastric gland together form a gastric unit. At the glandular stomach, the epithelium of the surface and gastric pits contains mainly mucous superficial cells (MSCs). The epithelium of fundic glands contains parietal cells, mucous neck cells (MNCs), zymogenic or chief cells (ZCs), and several types of enteroendocrine cells. Antral glands are formed by mucous antral gland cells and enteroendocrine cells.2–4 The complexity of fundic glands, with cells involved in many gastric functions (production of mucous barrier, HCl, and digestive enzymes), has attracted the attention of research.
The parietal cells secrete HCl. They are large cells scattered along the fundic glands, and it has been estimated that parietal cell number comprises 16–21% of all gastric epithelial cells in rats, and 12% in humans.5,6 These cells can be found in the gastric pit and in the three regions of the fundic gland, the isthmus, the neck, and the base. It is now established that all cells in the glandular epithelium arise from the same lineage, with a somatic stem cell, or SSC (not yet identified), at the isthmus of the gland, which produces precursor cells of MSCs, MNCs, and parietal cells. Eventually, MNCs differentiate into ZCs as they descend from the isthmus to the base.7–10
The study of the parietal cell differentiation is interesting both to understand the turnover of the gastric gland epithelium and to better understand the gastric pathological disorders.10,11 Cell differentiation involves the expression of new proteins or the expression of new oligosaccharides on glycoproteins and other glycosylated molecules (called glycoconjugates). Changes in the glycoconjugates of the gastric mucosa have been reported in precancerous intestinal metaplasia, 12 gastric carcinoma, 13 and infection by Helicobacter pylori. 14
Carbohydrate residues in glycan chains can be detected in situ by lectin histochemistry. 15 Lectins are carbohydrate-binding proteins which are not enzymes acting on the sugar ligand, carbohydrate-specific antibodies, or transport proteins for free mono- to oligosaccharides. 16 Lectin histochemistry uses this property to identify specific sugars in tissue sections. When combined with deglycosylation pretreatments, lectin histochemistry can also provide information about the nature of the glycan chain.
The aim of this work was to analyze the glycan composition of the parietal cells of rat fundic glands by means of lectin histochemistry, using for the first time a battery of lectins—which recognize the most frequent sugar moieties—in combination with deglycosylation pretreatments. The information provided by this analysis can help us to understand the process of parietal cell differentiation in a largely used animal model.
Materials and Methods
Reagents
Peroxidase-labeled agglutinin from Limax flavus (LFA), biotinylated Aleuria aurantia lectin (AAL), Galanthus nivalis agglutinin (GNA), Glycine max agglutinin (soybean agglutinin [SBA]), Ulex europaeus agglutinin-I (UEA-I), and Canavalia ensiformis agglutinin (concanavalin A [Con A]) were supplied by EY Laboratories (San Mateo, CA).
Type III glucose oxidase from Aspergillus niger, BSA, 3,3′-diaminobenzidine (DAB), peroxidase-labeled agglutinins from Helix pomatia (HPA) and Maclura pomifera (MPA/MPL), and biotinylated Arachis hypogaea (peanut agglutinin [PNA]) and Lotus tetragonolobus (LTA) agglutinins were purchased from Sigma-Aldrich (Madrid, Spain).
The enzyme Peptide-N-glycosidase F (PNGase F) from Flavobacterium meningosepticum and expressed in Escherichia coli, horseradish peroxidase–labeled anti-digoxigenin antibody (HRP-anti-DIG), digoxigenin-labeled Datura stramonium (DSA), and Sambucus nigra (SNA) agglutinins were obtained from Roche (Barcelona, Spain).
Avidin–biotin–peroxidase complex (VECTASTAIN ABC kit peroxidase standard, avidin–biotin blocking kit, and biotinylated Maackia amurensis hemagglutinin (MAH), Bandeiraea simplicifolia lectin I-B4 (BSI-B4), Ricinus communis agglutinin-I (RCA-I), Dolichos biflorus (DBA), and Triticum vulgaris (wheat germ agglutinin [WGA]) agglutinins were from Vector Laboratories (Burlingame, CA) and supplied by Atom (Barcelona, Spain).
Preparation of Tissue Samples
We used tissue samples from our archives that were obtained during 2002 and 2003. These samples were obtained as follows: eight adult male Sprague-Dawley rats, weighing 250–300 g, were supplied by the Animal Facility Service-SGIker of the University of the Basque Country UPV/EHU (Leioa, Vizcaya, Spain). To ameliorate suffering, animals were killed by CO2 inhalation before obtaining samples. All procedures involving animals were followed in accordance with the researchers’ institutional and local research guidelines.
Samples of gastric corpus were obtained and processed immediately: fixed in 10% formalin in PBS at 4C for 24 hr, washed in PBS, dehydrated, and embedded in paraffin. Paraffin-embedded samples were stored until use. To make lectin histochemistry, 4-µm sections were obtained.
Samples of the testis and intestine were also obtained to carry out control of deglycosylation techniques (see below), and processed in the same way.
Sections of human gall bladder from our archives, obtained in previous works, 17 were used as control of acid hydrolysis technique (see below).
Histochemical Procedure
After paraffin removal, sections were hydrated before lectin histochemistry. Seventeen lectins were used. The lectins, their origins, abbreviations, and binding specificities are listed in Table 1. Hydrated sections were immersed in 1% (v/v) H2O2 in TBS to block the endogenous peroxidase.
Lectins Used in This Work and Their Binding Specificity.
Decreasing specificity is indicated by one or more “greater than” symbols.18–22 Abbreviations: Fuc, fucose; Gal, galactose; GalNAc, N-acetylgalactosamine; NeuAc, sialic acid; GlcNAc, N-acetylglucosamine; Man, mannose; Glc, glucose.
AAL is sometimes abbreviated as AAA but then could be confused with the lectin from fresh water eel (Anguilla anguilla), which is usually abbreviated as AAA.
MAH is also abbreviated as MAA-II, MAA-2, and MAL-II. MAH should not be confused with MAL (Maackia amurensis leukoagglutinin, also abbreviated as MAM, MAL-I, and MAA-I) or MAA (an undefined mixture of MAL and MAH).
Histochemistry with peroxidase-labeled lectins was carried out as follows: After washing in TBS, sections were incubated with the lectin diluted in TBS. Lectins were diluted as follows: 6 µg/ml HPA, 20 µg/ml MPA/MPL, and 25 µg/ml LFA. Incubation was done at room temperature in a moist chamber for 1 hr 30 min. Then, sections were washed with TBS, and the peroxidase was developed as indicated below.
When histochemistry was carried out with digoxigenin-labeled lectins, the procedure differed as follows: after blocking the endogenous peroxidase, the sections were incubated in a moist chamber at room temperature for 10 min with 1% (w/v) BSA in TBS to block unspecific binding of the antibody. Then, the sections were incubated with the lectin diluted in 1% (w/v) BSA in TBS for 1 hr 30min. The lectins were diluted as follows: 10 µg/ml DSA and 30 µg/ml SNA. After washing with TBS, sections were incubated with 0.6 U/ml HRP-anti-DIG in 1% BSA in TBS.
The procedure carried out with biotinylated lectins differed because TBS was replaced by PBS, but blocking the endogenous peroxidase and unspecific binding sites for antibodies, and the incubation with the lectins was performed in a similar way. The lectins were used in the following dilutions: 5 µg/ml RCA-I, WGA, and MAH; 3 µg/ml BSI-B4, 50 µg/ml DBA, PNA, LTA, and SBA; 10 µg/ml AAL, UEA-I, and Con A; and 60 µg/ml GNA. After incubation with the lectin, sections were washed with PBS and incubated with the ABC kit for 1 hr.
All the histochemical procedures (peroxidase-labeled, digoxigenin-labeled, and biotinylated lectins) finished developing the peroxidase with 0.1% (v/v) H2O2 and 0.25 mg/ml DAB in TBS for 7 min. The sections were then counterstained with hematoxylin
Deglycosylation Pretreatments
The histochemical procedure with every lectin was carried out in three ways: (1) without previous treatment, (2) after previous chemical deglycosylation (β-elimination) procedure, and (3) after previous enzymatic deglycosylation with PNGase F.
The procedure with some lectins was also carried out after previous treatment of β-elimination and PNGase F in the same section.
Beta-elimination is a method to remove O-linked oligosaccharides described by Ono and coworkers. 23 The technique was carried out immersing the hydrated sections in 0.5-N NaOH in 70% ethanol at 4C. For every lectin, the technique was performed in different sections for 1 and 5 days, to discriminate both labile and resistant O-glycans. 24
The removal of N-linked oligosaccharides was carried out using the enzyme PNGase F. 25 The sections were incubated with 40 U/ml PNGase F in moist chamber at 37C. 26
Both deglycosylation pretreatments, β-elimination and incubation with PNGase F, enhance endogenous biotin labeling by the ABC kit in gastric gland cells (unpublished observations). Thus, endogenous biotin was blocked using the avidin–biotin blocking kit.
Some lectins (Con A, GNA, AAL, LTA, PNA, RCA-I, BSI-B4, MPA/MPL, HPA, DBA, SBA, and DSA) were also tested after removal of sialic acid (NeuAc) in terminal position from oligosaccharides. For this, acid hydrolysis was performed by section immersion in 0.1-N HCl at 82C–84C for 3 hr before the histochemical procedure with the lectin. 17 This technique allows us to know the subterminal sugars to terminal NeuAc. It was also used with WGA, a lectin which recognizes both N-acetylglucosamine (GlcNAc) and NeuAc (Table 1), to determine which carbohydrate was labeled by this lectin.
Some carbohydrates could be sulfated, and then they cannot be labeled by the lectins. To show the presence of sulfated terminal sugars in oligosaccharides, the procedure of methylation-saponification was used with Con A, GNA, AAL, LTA, RCA-I, BSI-B4, MPA/MPL, HPA, DBA, and DSA lectins. This technique is carried out by immersing the deparaffined sections in pure ethanol for 4 min and then in 1% (v/v) HCl at 60C in methanol for 5 hr. After rehydration, the sections were immersed in 1.8% (w/v) barium hydroxide at 4C for 1 hr. 27
Finally, to elucidate whether Con A is labeling glucose (Glc) or mannose (Man), histochemistry with this lectin was also carried out after incubation with 50 U/ml type III glucose oxidase at 37C overnight. This procedure converts Glc into gluconic acid, which is not recognized by Con A. 28
Controls
The following controls were used: (1) substitution of the lectins or the enzyme by the buffer alone, (2) preincubation of the lectins with the corresponding hapten sugar inhibitor—fucose (Fuc) for AAL, LTA, and UEA-I; N-acetylgalactosamine (GalNAc) for HPA, DBA, SBA, and MPA/MPL; galactose (Gal) for PNA, BSI-B4, and RCA-I; lactose for MAH and SNA; NeuAc for LFA, chitotriose for WGA and DSA 29 ; and Man for GNA and Con A, at a concentration of 0.2 M; and (3) staining of sections of other tissues of known altered binding pattern for each of the chemical and enzymatic pretreatments. As specific control of β-elimination, the procedure was carried out with sections of the rat testis combined with HPA histochemistry, because this tissue is not labeled by HPA after the procedure. 26 The specific control for incubation with PNGase F was the AAL histochemistry in the rat testis, because there is no labeling after removal of N-glycans. 26 To ensure the correct functioning of the acid hydrolysis, PNA histochemistry was carried out in human gall bladder sections to verify the labeling of secretory granules of principal cells. 17 PNA histochemistry after methylation-saponification in rat intestine labels microvilli and globet cells, 27 so this procedure was used as specific control for this technique.
Analysis of Results
Three histological sections from each sample were used for the histochemical procedure with each lectin and with each deglycosylation procedure. The staining intensity in the gastric cells was evaluated by three independent observers. It was classified into six categories: no labeling (0), very weak (1), weak (2), moderate (3), strong (4), and very strong (5). This classification let us evaluate the effects of the pretreatments. Sometimes, when there was a different staining intensity of the parietal cells in different sections, or in different fields of the same section, it was classified within a range indicated with the minimum and maximum values.
Results
Results are detailed in Table 2, which shows the intensity of staining of parietal cells for each lectin, including the pretreatments tried. To simplify the table, the results after some pretreatments are not shown: In detail, (1) methylation-saponification and acid hydrolysis have not been described because the obtained results were similar to the results with the lectin alone and (2) the double pretreatment with PNGase F and β-elimination in the same section have not been showed in the table, because the labeling was only modified for some lectins, and these differences are indicated at footnotes.
Evaluation of Lectin Labeling of Parietal Cells in the Gastric Glands.
The staining intensity was evaluated and classified into five categories: no labeling (0), very weak (1), weak (2), moderate (3), strong (4), and very strong (5). When there was a variation in staining of the same structure, the range of staining is indicated with the minimum and maximum values separated by a dash. When labeling of parietal cells differed in the upper and lower regions of the gland, it is indicated as u.c. (cells at the upper region of the gland) and l.c. (cells at the lower region of the gland). Results of lectin histochemistry after some pretreatments (methylation-saponification, acid hydrolysis) are not shown because the results did not differ from the lectin histochemistry without pretreatment. Abbreviations: PNGase F, Peptide-N-glycosidase F; AAL, Aleuria aurantia lectin; LTA, Lotus tetragonolobus agglutinin; UEA-I, Ulex europaeus agglutinin-I; HPA, Helix pomatia agglutinin; DBA, Dolichos biflorus agglutinin; SBA, Glycine max agglutinin; MPA/MPL, Maclura pomifera agglutinin; PNA, peanut (Arachis hypogaea) agglutinin; BSI-B4, Bandeiraea simplicifolia lectin I-B4; RCA-I, Ricinus communis agglutinin-I; MAH, Maackia amurensis hemagglutinin; SNA, Sambucus nigra agglutinin; LFA, Limax flavus agglutinins; WGA, wheat germ (Triticum vulgaris) agglutinin; DSA, Datura stramonium agglutinin; GNA, Galanthus nivalis agglutinin; Con A, Canavalia ensiformis agglutinin.
Staining with BSI-B4 was reduced when PNGase F and β-elimination for 1 day were done in the same section, and totally removed if the β-elimination was for 5 days.
WGA staining was not significantly modified by acid hydrolysis pretreatment.
After acid hydrolysis or methylation-saponification, both upper and lower parietal cells were stained with the same intensity.
After glucose oxidase pretreatment, the staining decreased to 1.
None of the Fuc-labeling lectins, that is, AAL, LTA, and UEA-I, stained parietal cells in any case (Fig. 1).
Fucose-binding lectins. The micrographs show the gastric gland bases stained with AAL (A), the pit and isthmus areas stained with LTA (B), and the neck region stained with UEA-I (C) of the gastric units. The parietal cells were negative for the three lectins. Lectins are abbreviated as in Table 1. Scale bars: A–C = 20 µm.
Parietal cells in the upper region of the gland (pits and isthmus) showed a stronger labeling with the lectins which recognize GalNAc (HPA, DBA, SBA, and MPA/MPL) than cells in the lower area (neck and base; Fig. 2). In fact, parietal cells in the lower region were negative for HPA. Labeling was irregularly observed in the cytoplasm. In some cases, as was for DBA, some labeling could be attributed to the intracellular secretory canaliculi surrounding the nucleus of these cells. The same pattern of labeling was seen in the parietal cells in the upper and lower areas, which only showed a different intensity of labeling. Labeling with these lectins remained after incubation with PNGase F (except for SBA) but diminished or disappeared after β-elimination pretreatment. For SBA, after PNGase F pretreatment upper parietal cell labeling diminished and lower parietal cells turned negative. Acid hydrolysis and methylation-saponification procedures did not modify the labeling pattern of parietal cells for GalNAc-binding lectins.
GalNAc-binding lectins. (A, B, C) Micrographs showing HPA histochemistry at low magnification of the gastric units (A), and higher magnification of the upper (B) and lower (C) regions. The staining of the parietal cells was stronger at the upper than the lower regions. (D) When HPA histochemistry was carried out after β1d, all the parietal cells were negative. (E) Gastric glands stained with DBA. The parietal cells at the upper region of the gastric units showed various intensities of labeling, ranging from zero to very strong, whereas most of the parietal cells at the lower region were negative, and only a few cells were moderately or strongly stained. (F) Isthmus region of gastric glands showing parietal cells with various intensities of labeling with DBA. The intensity of labeling is indicated by numbers which correspond to the scale used in Table 2. (G) After β5d, labeling with DBA showed a lower intensity. (H, I, J) Micrographs showing SBA histochemistry at low magnification of gastric units (H), and higher magnification of upper (I) and lower (J) regions. The staining of the parietal cells was stronger at the upper than the lower regions. (K) After incubation with PNGase F, the parietal cells at the lower region of the glands turned negative for SBA. (L) Histochemistry with MPA/MPL strongly labeled the parietal cells of the upper region but not those of the lower region. Some cells of the upper region are shown in the inset. (M) The parietal cells were negative for MPA/MPL after pretreatment of β5d. The inset shows parietal cells at the upper region of the gland at higher magnification. Arrows, parietal cells. Numbers (0, 2, 3, 4, and 5) indicate intensity of staining following the scale used in Table 2. Abbreviations: GalNAc, N-acetylgalactosamine; β1d, β-elimination for 1 day; β5d, β-elimination for 5 days; PNG-F, incubation with PNGase F; PNGase F, Peptide-N-glycosidase F. Lectins are abbreviated as in Table 1. Scale bars: A–M = 20 µm.
Concerning lectins recognizing Gal moieties (Fig. 3), PNA labeled only the upper parietal cells, but the others (BSI-B4 and RCA-I) labeled all parietal cells. When parietal cells were stained, the lectins labeled all the cytoplasm with different degree of intensity. PNA-labeling was unaltered by PNGase F pretreatment. Beta-elimination for 1 day caused a staining decrease, and the same pretreatment for 5 days turned parietal cells negative for the lectin. BSI-B4 differentially labeled upper and lower parietal cells after PNGase F incubation. Labeling with this lectin decreased or disappeared with β-elimination for 1 or 5 days, respectively, only when they were combined with PNGase F pretreatment. Parietal cells showed a similar staining pattern for RCA-I after each pretreatment procedure. Pretreatments of acid hydrolysis or methylation-saponification did not modify the staining pattern for any Gal-binding lectins.
Galactose-binding lectins. (A) Only parietal cells at the upper region of gastric units were very strongly positive for PNA. (B) At the lower region of gastric glands the parietal cells were negative for PNA. (C) Labeling with PNA of upper parietal cells was very weak or weak after β1d. (D) BSI-B4 strongly labeled all the parietal cells along the gastric units. The inset shows a higher magnification view of strongly labeled parietal cells. (E) BSI-B4 after PNGase F pretreatment showing a huge reduction of labeling of parietal cells at the lower zone of the glands, whereas the parietal cells at the upper zone remain strongly labeled. (F) RCA-I labeled all the parietal cells. Arrows, parietal cells. Abbreviations: β1d, β-elimination for 1 day; PNG-F, incubation with PNGase F; PNGase F, Peptide-N-glycosidase F. Lectins are abbreviated as in Table 1. Scale bars: A–F = 20 µm.
LFA was the only NeuAc-binding lectin which labeled the cytoplasm of parietal cells slightly, which were always negative for MAH and SNA (Fig. 4). LFA labeling remained after PNGase F or β-elimination procedures.
NeuAc-binding lectins. (A, B) The parietal cells were negative for MAH (A) and SNA (B). (C) The parietal cells showed a very weak or weak labeling to LFA. A negative control (substitution of the lectin by the buffer alone) is showed in the inset. Lectins are abbreviated as in Table 1. Arrows, parietal cells. Scale bars: A–C = 20 µm.
Both GlcNAc-binding lectins labeled parietal cells (Fig. 5), showing an irregularly and diffuse labeling of the cytoplasm. WGA also labeled the nuclear envelope. Labeling with WGA was not significantly modified after PNGase F, β-elimination, or acid hydrolysis. The upper parietal cells were slightly more intensely labeled by DSA than lower parietal cells. This labeling pattern remained after β-elimination. However, both upper and lower parietal cells were stained with the same intensity after PNGase F, acid hydrolysis, or methylation-saponification procedures.
GlcNAc-binding lectins. (A) The parietal cells were positive for WGA. (B) Labeling of the parietal cells after acid hydrolysis pretreatment, which removes sialic acid moieties. (C) The parietal cells of the upper region were more strongly labeled than the parietal cells of the lower region with DSA. Lectins are abbreviated as in Table 1. GlcNac, N-acetylglucosamine; ah, acid hydrolysis. Scale bars: A–C = 20 µm.
Finally, both Man-binding lectins, GNA and Con A, labeled the cytoplasm of parietal cells (Fig. 6). In addition, GNA labeled the nuclear envelope region. The cells were negative for both lectins after PNGase F pretreatment, while the staining remained or was slightly reduced, after β-elimination, acid hydrolysis, or methylation-saponification. As Con A lectin can also recognize Glc, pretreatment with glucose oxidase was also carried out. The parietal cells showed a very weak staining after this procedure.
Man-binding lectins. (A) GNA moderately to strongly labeled the parietal cells. (B) No GNA-labeling of the parietal cells was seen after incubation with PNGase F. (C) Moderate labeling of the parietal cells after Con A histochemistry was seen. (D) The pretreatment of PNGase F turned the parietal cells negative for Con A. (E) After incubation with glucose oxidase, which converts glucose residues into gluconic acid, the intensity of labeling of the parietal cells was reduced. Arrows, parietal cells. Abbreviations: Man, mannose; PNG-F, incubation with PNGase F; PNGase F, Peptide-N-glycosidase F; go, incubation with glucose oxidase. Lectins are abbreviated as in Table 1. Scale bars: A–E = 20 µm.
Discussion
Lectin histochemistry alone or combined with deglycosylation techniques is a useful tool which has been extensively used to analyze the in situ glycan composition of tissues and cells.15,30 The results obtained with lectin histochemistry could be interpreted taking into account that the specificity of a lectin is affected by (1) the linkage of the labeled sugar,15,17,18 (2) the protein backbone of the glycoprotein, 31 and (3) the neighboring glycans, so the removal of O-linked glycans can modify the affinity for N-glycans or viceversa 22 or can unmask carbohydrates which were previously inaccessible to the lectin. 21 In this work, we have used 17 selected lectins which recognize the most frequent carbohydrate moieties in mammalian glycoproteins, that is, Fuc, GalNAc, Gal, NeuAc, GlcNAc, and Man, to identify the glycan composition of parietal cells in rat gastric glands.
None of the Fuc-binding lectins used in this work labeled the parietal cells. Other cell types were positive to some lectins, indicating that lectin histochemistry worked correctly. The cells were also negative for the lectins after any of the deglycosylation pretreatments, suggesting that glycans with terminal Fuc were not masked to lectin by other N- or O-glycans. Acid hydrolysis was performed to remove terminal NeuAc and to allow subterminal sugar residues to be identified by the lectins. 17 However, no Fuc residues in subterminal position to NeuAc were located in parietal cells. Sulfated sugars could not be labeled by the lectins. Then, the pretreatment of methylation-saponification, which removes sulfate groups from sulfated sugars, was also carried out. 27 After this pretreatment, no Fuc residue was identified by any of the Fuc-binding lectins used. In conclusion, glycans of parietal cells lack terminal Fuc moieties. This agrees with the poor or no labeling of human parietal cells by UEA-I previously reported.32,33
All the GalNAc-binding lectins showed different labeling in upper and lower parietal cells. As a general rule, labeling in upper cells was stronger than in lower cells. This difference has not been reported in human parietal cells, which are labeled by DBA and SBA.32,33 Considering the binding specificity of every lectin (references in Table 1) and the labeling patterns observed after deglycosylation pretreatments, we can draw the following conclusions: Some parietal cells, mainly at the upper region of the gastric unit, showed GalNAc in resistant O-linked oligosaccharides labeled by DBA and MPA/MPL. All the parietal cells showed GalNAc in resistant O-linked glycans labeled by SBA. In addition, upper parietal cells showed GalNAc in labile O-linked oligosaccharides, labeled by HPA and SBA. Lower parietal cells showed GalNAc in N-glycans labeled by SBA and a minor expression of labile O-linked oligosaccharides labeled by MPA/MPL. Furthermore, the stronger labeling of upper parietal cells with DBA and MPA/MPL, suggests that they have a higher expression of GalNAc than lower cells. In particular, DBA labeling was complex: Parietal cells at the upper part of the gland showed various intensities of labeling, ranging from no labeling (0) to very strong (5), whereas only a few parietal cells were positive for the lectin at the base of the gland. A stronger DBA staining in human parietal cells after acid secretion stimulation has been reported. 34 Then, in our rat model, different intensity of DBA staining in the parietal cells of the higher part of the gastric glands could be indicative of different levels of acid secretion. The unlabeling of some parietal cells is surprising because DBA has been used as a marker for parietal cells, at least in mouse stomach. 35 Thus, the use of DBA as a marker for parietal cells should be taken with caution, even if different fixation protocols applied in different papers are to be considered, too.
A different labeling pattern of upper and lower parietal cells was also observed with two of the Gal-binding lectins, PNA and BSI-B4. A similar result was reported for PNA in a previous work, but no deglycosylation pretreatment was carried out and no interpretation was proposed. 36 Although some previous works have reported labeling of human parietal cells by PNA and BSI-B4, no difference depending on their location in the gland has been described so far.32,33 In the present work, Galβ(1,3)GalNAc, identified by PNA, was located in both resistant and labile O-linked glycans only in upper parietal cells. Galα(1,3) has been located in N-linked oligosaccharides by BSI-B4 in lower parietal cells. Staining of upper parietal cells with BSI-B4 remained almost unaltered after any of the pretreatments (incubation with PNGase F or β-elimination) but was partially or fully removed when both pretreatments were performed on the same section, suggesting that the lectin is labeling to Gal moieties in both N- and O-linked glycans. In contrast, N-acetyllactosamine, recognized by RCA-I, could be located in glycosphingolipids in all the parietal cells, because the staining remained when both pretreatments were combined on the same section.
The parietal cells were negative for two of the NeuAc-binding lectins, MAH, which recognizes NeuAcα(2,3) linked, and SNA, which labels NeuAcα(2,6) linked. These results could be unexpected, because the sialylated glycans recognized by both lectins are widely found in glycoproteins and glycolipids, 37 but absence of labeling was also reported for mouse parietal cells. 38 The parietal cells were weakly labeled by the other NeuAc-labeling lectin used, LFA, a lectin which labels NeuAc linked in anyway. Considering that the labeling remained after the removal of either N- or O-linked glycans and that LFA can label NeuAc in glycosphingolipids, 39 we could consider that LFA is labeling NeuAc in glycosphingolipids.
Labeling of parietal cells with WGA remained after acid hydrolysis, suggesting that the lectin is labeling GlcNAc but not NeuAc. As the labeling remained after removal of both N- and O-linked glycotopes, GlcNAc could be located in both types of glycans or in glycolipids. The same happened for DSA, the other GlcNAc-binding lectin, although a slightly stronger staining was seen in the upper parietal cells.
All the parietal cells showed labeling with Man-binding lectins. As expected, Man was located in N-linked oligosaccharides. In addition, pretreatment with glucose oxidase decreased Con A staining, which indicates that some of the Con A labeling could be due to Glc residues, according to Con A specificity showed in Table 1. However, lectin histochemistry has suggested a lack of mannosyl and glucosyl moieties in human parietal cells.32,33
Now, it is well established that stem cells in the isthmus give rise to three main progenitors: prepit, preneck, and preparietal cells. Parietal cells, originated mainly from the preparietal cells, migrate in two directions: toward the pit and toward the base 10 ; however, unlike the mouse, the parietal cells in rat are fewer in the pit. 40 This migration is associated with a loss in functional activity; thus, two distinct populations can be found: young parietal cells in the isthmus and neck and old parietal cells in the base regions (and in the pit region in the mouse). The existence of these two populations of parietal cells has been supported by six approaches: (1) young parietal cells have more developed canaliculi and rough endoplasmic reticulum than old parietal cells, which show signs of degeneration 41 ; (2) young parietal cells, but not base parietal cells, are activated by food 42 ; (3) H+/K+ ATPase activity found in young parietal cells is absent from old parietal cells 43 ; (4) upper parietal cells respond much more efficiently than basal parietal cells to acid inhibitors and secretagogues 44 ; (5) the level of H+/K+ ATPase β-subunit mRNA is high in parietal cells near the isthmus but is low in distant cells 44 ; and (6) there is a different production of the osteoblast activating peptide by parietal cells of the neck and base regions. 45 Here, we have shown a seventh approach to support the existence of two subpopulations of parietal cells. Our findings indicated that parietal cells express different glycoconjugates depending on their location in the upper or lower region of the gland: Cells in the upper region of the gland, near the isthmus, express different carbohydrates than cells in the base of the gland.
It can be assumed that the above explained differences must be acquired during migration of cells. The parietal cells synthesize the morphostat Sonic hedgehog (Shh), which forms a gradient from the isthmus toward the base of the gland, 4 and they also express Patched, the receptor of Shh.46,47 Then, the different glycan composition in each region of the gland could be due to an autocrine regulation. In summary, we have shown the labeling pattern of glycoconjugates of parietal cells by means of lectin histochemistry combined with deglycosylation pretreatments. The results showed two subpopulations of parietal cells expressing different glycans: one population comprising the cells near the isthmus, at the upper part of the gland, and the other comprising the cells migrating toward the base of the gland. This finding supports previous works which indicate that basal parietal cells degenerate or have roles other than acid secretion. Moreover, we have shown that some parietal cells of the upper region of the gland and most of the lower region were negative for DBA; thus, the use of DBA as a marker of parietal cells should be cautiously considered.
Footnotes
Acknowledgements
Mrs M.J. Fernández and C. Tobillas contributed to sample preparation. We thank Mrs M.J. Aldasoro for her support in the office work.
Competing Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Author Contributions
LG-S and EA have contributed to research design, acquisition of data, data analysis, and revision of the manuscript. JFM and LD-F have contributed to research concept and design, acquisition of data, data interpretation, and critical revision of the manuscript. FJS has contributed to research concept and design, data analysis and interpretation, and drafting and critical revision of the manuscript.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by the University of the Basque Country UPV/EHU (grant number EHUA13/15, EHUA15/26, and UFI 11/44), and Fundación Séneca from the Comunidad Autónoma de la Región de Murcia (grant number 04542/GERM/06).