Expression of Neuronal Nitric Oxide Synthase in Several Cell Types of the Rat Gastric Epithelium

Abstract

The aim of this study was to identify which cell types of the rat gastric epithelium express neuronal nitric oxide synthase (nNOS) because the results of the previous studies have been very divergent regarding this point. By the combination of immunohistochemical (IHC) and in situ hybridization (ISH) techniques, we detected expression of nNOS in chief and mucosecretory cells of the gastric epithelium. Moreover, some gastric endocrine cells were immunoreactive for nNOS, although they could not be distinguished in sections treated with ISH techniques. The strongest signal for all antibodies in IHC techniques was obtained when microwave (MW) heating was performed before the IHC procedure. Our results indicate that in the gastric epithelium a variety of cell types are able to produce NO. The NO produced by the different cell types (chief, mucous, and endocrine) may form a complex network of paracrine communication with an important role in gastric physiology. (J Histochem Cytochem 48:1111–1119, 2000)

Nitric oxide (NO) is a multifunctional messenger that is involved in a wide range of physiological processes in many systems (Moncada 1992). NO is synthesized intracellularly from l-arginine by the catalytic action of enzymes collectively known as nitric oxide synthase (NOS) that require NADPH as a co-factor and are inhibited by some analogues of l-arginine. At least three different isoforms of NOS have been purified and cloned: a neuronal NOS (nNOS) and an endothelial NOS (eNOS), which are both calcium- and calmodulin-dependent, and an inducible NOS (iNOS), which is calcium-independent (Knowles and Moncada 1994). The nNOS was first isolated from cerebellum (Bredt et al. 1990). Since that time, nNOS has been identified in many neuronal and non-neuronal cell types such as skeletal muscle, lung and kidney, which are some of the more recently established sources of NOS (Knowles et al. 1989; Springall et al. 1994; Boissel et al. 1998). Current data suggest that regulation of NOS expression as well as NO synthesis and release might have an important role in the physiology and the pathology of many systems (Moncada 1992; Kerwin and Heller 1994; Springall 1995).

In the gastrointestinal tract, NO may influence muscle tone as well as endocrine and exocrine secretions (Whittle 1994). Furthermore, the dual role of NO as a cytoprotective or a cytotoxic free radical gas has been noted under physiological and pathological conditions (Konturek and Konturek 1995). In particular, in the gastric mucosa, different studies have agreed on the high content of a calcium-dependent constitutive NOS (Whittle et al. 1991) that produces NO, which is involved in maintaining mucosal integrity and regulating blood flow to the epithelium, but these studies have not agreed on the particular cell types responsible for this activity. These authors, using different techniques, have reported the production of NO by mucous (Brown et al. 1992b; Price et al. 1996; Byrne et al. 1997; Ichikawa et al. 1998; Price and Hanson 1998), chief (Fiorucci et al. 1995a), or endocrine (Akiba et al. 1995a; Burrell et al. 1996) epithelial cell types. Furthermore, the presence of NOS has also been reported in the epithelium of the forestomach (Schmidt et al. 1992) and the gastric brush cells (Kugler et al. 1994) of the rat.

Because the results of the previous studies have been very divergent regarding the localization of nNOS in the gastric mucosa, we tried to ascertain the possible reasons for such dissimilar results. Therefore, the aim of this study was to identify, by the combination of immunohistochemical and in situ hybridization techniques, which cell types of the rat gastric epithelium express nNOS.

Materials and Methods

Adult Wistar rats were sacrificed and pieces from the oxyntic and pyloric regions of the stomach and from the cerebellum were dissected. Part of the material was used for extraction of RNA. Other fragments were used for the histological study, some being fixed in Bouin's fluid or 10% neutral formalin (pH 7.4) for 12–24 hr, dehydrated, and embedded in paraffin. Finally, some fragments were fixed in 2% paraformal-dehyde/1% glutaraldehyde for 2 hr, washed in PBS 0.1 M, pH 7.4, dehydrated with ethanol, and embedded in Epon 812 for the ultrastructural study.

Reverse Transcriptase-Polymerase Chain Reaction (RT-PCR)

RNA was extracted from cerebellum, total gastric wall, or gastric mucosa using Ultraspec (Biotecx Labs; Houston, TX). The RNA (1 μg) was reverse-transcribed with a random hexamer and M-MLV reverse transcriptase (Gibco BRL; Paisley, UK). After an initial denaturation for 5 min at 95C, the resulting complementary DNA (cDNA) was amplified by PCR for 30 cycles of amplification (95C for 1 min, 56C for 1 min, and 68C for 1 min), followed by a 10-min extension at 72C, with a pair of primers (5′ CTACAAGGTCCGATTCAACAG 3′ sense and 5′ CCCACACAGAAGACATCACAG 3′ antisense) flanking a 315-bp fragment (2868–3182) of the rat neural NOS cDNA (Bredt et al. 1991), homologous to the human sequence and encompassing a 115-bp non-coding region of the human nNOS gene.

A 315-bp nNOS PCR product was purified by agarose electrophoresis, extracted from the gel, and ligated into a pGEM-T vector (Promega; Madison, WI). The identity of the fragment was confirmed by sequencing (DNA sequencing kit; PE Biosystems, Foster City, CA) in an automated sequencer (310 Genetic Analyzer; PE Biosystems).

Amplification of rat glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was performed simultaneously on the same samples to assess RNA integrity. The oligonucleotide primer sequences for GAPDH (Tso et al. 1985) were 5′ CCAAGGTCATCCATGACAA 3′ (sense) and 5′ TGTCATACCAGGAAATGAGC 3′ (antisense). All primers were synthesized by Genset SA (Paris, France). PCR products were analyzed by electrophoresis on a 1% (w/v) agarose gel.

Antisera

We used three antisera specific for neuronal NOS. One was a polyclonal antiserum (A1) against the synthetic peptide 520–540 of the cloned rat neural NOS, a kind gift from S. Moncada (present address: The Cruciform Project, University College London, London, UK). The amino acid sequence of this peptide was LPLLQANGNDPELFQIPPELC, as described by Riveros-Moreno et al. (1993). The second one (A2) was a polyclonal antiserum against the recombinant rat nNOS (Charles et al. 1993), courtesy of Wellcome Research Laboratories (Research Triangle Park, NC). The third (A3) was a monoclonal antiserum raised against the peptide 1095–1289 of the human nNOS from Transduction Laboratories (Lexington, KY) (code N31020). The three were used at a dilution of 1:500.

Antigen Retrieval Microwave (MW) Heating Technique

Before the immunohistochemical procedure, tissue sections were deparaffinized and rehydrated to water, and endogenous peroxidase was blocked with 3% H₂O₂ for 15 min. Slides were washed with distilled water for 5 min, placed in a citrate buffer 0.01 M (pH 6.0), and heated in the microwave oven (Balay W-2112, 1150–700 W; Madrid, Spain) for 15 min at maximal power and for 15 min at medium power. After rinsing in tapwater, the immunohistochemical procedure was performed as usual. For electron microscopy, thin sections were exposed to MW heating in the same manner but were heated for 50 min at the lowest power.

Immunohistochemistry

Light Microscopy. Paraffin sections 5 μm thick were mounted on slides coated with Vectabond (SP-1800; Vector Laboratories, Burlingame, CA). Single immunohistochemical staining was performed by the avidin-biotin-peroxidase complex method (ABC) (Hsu et al. 1981). The detailed protocol used was previously described by Martánez et al. (1993), the only difference being that the peroxidase activity was revealed in 0.03% 3,3′diaminobenzidine (DAB) in 0.1 M sodium acetate/acetic acid buffer, pH 6.0, containing 2.5% nickel ammonium sulfate, 0.2% β-d-glucose, 0.04% ammonium chloride, and 0.001% glucose oxidase (Shu et al. 1988).

Electron Microscopy. Thin sections from Epon-embedded material were mounted on nickel grids, MW-treated, and incubated for 1 hr in 1% normal goat serum in gold buffer (Tris-HCl-buffered saline 0.05 M, pH 7.2, 0.15 M ClNa, with 1% bovine serum albumin and 0.05% sodium azide), and then incubated in the polyclonal anti-nNOS, diluted 1:100, overnight at 4C. Subsequent passages, all at room temperature, included rinses in gold buffer, incubation for 1 hr with the gold-labeled secondary antiserum (goat anti-rabbit 20-nm gold conjugate, diluted 1:20; BioCell Research Laboratories, Cardiff, UK), and rinses in gold buffer again. Finally, sections were double stained in uranyl acetate and lead hydroxide.

Probes

The nNOS fragment-containing plasmid was linearized with PstI to create a template for antisense probe production or with SacII to create a template for a sense probe (Boehringer Mannheim; Mannheim, Germany) according to the manufacturer's instructions. Digoxigenin (DIG)-labeled antisense and sense probes were synthesized with T7 or SP6 RNA polymerases, respectively, using a DIG RNA (SP6/T7) labeling kit (Boehringer Mannheim).

Figure 1

RT-PCR analysis of neuronal NOS mRNA from the separate oxyntic and pyloric mucosa (Lanes 1 and 2), the total oxyntic and pyloric wall (Lanes 3 and 4), and the cerebellum (Lane 5) of the rat. M is the phiX174/Hae III molecular weight marker. The size of the predicted amplified product nNOS (315 bp) is indicated at right.

In Situ Hybridization

Sections 5 μm thick were mounted on coated slides ProbeON Plus (Fisher Biotech; Pittsburgh, PA), dewaxed, and prepared for hybridization with RNA probes as described by Gibson and Polak (1990). Peptidase pretreatment was carried out by incubation with proteinase K (Sigma; St Louis, MO), 20 μg/ml in 0.1 M Tris/0.05 M ethylene diamine tetra-acetic acid (EDTA), pH 8.0, for 30 min at 37C. For each section, hybridization was performed in 15 μl hybridization buffer [50% formamide, 5 × SSC, 1% dextran sulfate, 5 × Denhardt's solution, 2% sodium dodecyl sulfate (SDS), and diethyl pyrocarbonate-treated water], supplied with 20 ng/ml riboprobe, for 20 hr at 44C in a moist chamber. Hybridization was followed by four washes in 2 × SSC/0.1% SDS, two washes in 0.1 × SSC/0.1% SDS at 48C, brief rinses in 2 × SSC, incubation in 2 × SSC containing 25 μg/ml RNase (Boehringer Mannheim) at 37C for 15 min, and additional rinses in 2 × SSC. Visualization of DIG was performed by incubation with a monoclonal antibody coupled to alkaline phosphatase (anti-DIG-AP Fab fragments; Boehringer Mannheim), diluted 1:500, for 2 hr at room temperature. Nitroblue tetrazolium chloride and 5-bromo-4-chloro-3-indolyl-phosphate (Boehringer Mannheim) were used as substrates for the alkaline phosphatase. Controls include use of the sense probe, omission of the probe, and treatment of the section with RNase before hybridization.

Results

Expression of nNOS mRNA in Gastric Mucosa

The PCR products amplified using the nNOS-specific primers showed a clear band at the predicted size of 315 bp in control tissue (cerebellum) and in both the total gastric wall and the separate mucosa of the oxyntic and pyloric regions (Figure 1). As a positive control, the rat GADPH was assayed in all the specimens to verify the efficiency of cDNA synthesis from extracted RNA.

Figure 2

Pair of consecutive paraffin sections of formalin-fixed rat intestine immunostained for nNOS (A) without and (B) with MW treatment, using the A2 antiserum. The MW heating enhanced the intensity of immunoreaction. Note the ganglia of the myenteric plexus (arrowheads), as well as the number of immunoreactive elements (arrows), especially in the circular muscle layer (CM). LM, longitudinal muscular layer. Bars = 40 μm.

Immunohistochemical Results

The strongest immunostaining signal for all the antisera specific for nNOS used in this study was obtained by MW heating of tissue sections in citrate buffer solution (pH 6.0) before the immunohistochemical procedure. The immunostaining was particularly enhanced in formalin-fixed material, both in control tissues (brain and cerebellum) and in the neural plexi of the digestive organs (Figure 2).

In the gastric epithelium, immunostaining for nNOS appeared in chief cells (Figures 3 and 4) of the oxyntic region, and in mucosecretory (Figure 3) and endocrine (Figure 5) cells in both the oxyntic and pyloric regions. No immnunostaining was found in parietal cells. This immunolabeling pattern was achieved with the two polyclonal antisera for nNOS used in this study. The antiserum against the synthetic peptide (A1) stained all three cell types, and that against the recombinant rat nNOS (A2) only the chief and muco-secretory cells. The monoclonal antibody (A3) stained neural elements of the plexi but did not give positive results in the gastric mucosa (see Table 1).

Figure 3

Detection of nNOS in the oxyntic glands by immunohistochemistry (IHC) using the A2 antiserum (A,C,E) and in situ hybridization (ISH) (B,D,F). The distribution pattern of the labeling with the nNOS antiserum (A) is similar to that of the nNOS probe (B). The staining is stronger in the lower half of the oxyntic glands with both techniques. Bars = 100 μm. (C,D) Details of the bottom rectangles from A and B, respectively. In this region, the staining appeared in chief cells (asterisks). (E,F) Details of the top rectangles from A and B, respectively, showing the mucosecretory cells located at the neck portion of the oxyntic glands (arrows) positive for IHC and ISH, although less intensely than chief cells. Bars = 30 μm.

Figure 4

(A) High-magnification micrograph of two oxyntic glands showing the apical granular labeling of chief cells with IHC (A2 antiserum). Bar = 8 μm. (B) Electron micrograph of an oxyntic gland immunostained with anti-nNOS (A2 antiserum). Chief cells (C) are immunoreactive, whereas a mucosecretory cell (M) and an endocrine cell (E) are negative. Bar = 2 μm. (C) Detail of the nNOS-immunostained chief cells, showing that the labeling appeared in the pepsinogen granules. Bars = 1.5 μm.

Figure 5

Scattered cells similar to endocrine cells are immunoreactive for nNOS, using the A1 antiserum, in both oxyntic (A) and pyloric (B) glands. Bars = 20 μm.

Figure 6

(A) Panoramic view of the antral epithelium stained with ISH for nNOS. The labeling appeared in mucosecretory cells at the base of the glands. Bar = 40 μm. (B) A higher magnification of the positive cells. Bars = 20 μm.

Detection of nNOS mRNA by ISH

ISH studies were consistent with the immunohistochemical results. nNOS mRNA was detected in the chief and mucosecretory cells (Figures 3 and 6) of the gastric glands, both of which were also positive for the nNOS protein. Hybridization was also found in control tissues: brain, cerebellum, and neurons of the gastric plexi (Figure 7A). When the sense probe was used, no stain was detected either in neuronal elements (Figures 7A and 7B) or in the epithelial cells (Figures 7C and 7D). Endocrine cells were the only cell type positive with immunohistochemistry that could not be distinguished in sections treated with ISH techniques. As occurred with IHC, parietal cells were also negative with these techniques (Figures 3E and 3F).

Description of the NOS-expressing Gastric Cell Types

Chief Cells. Both immunohistochemical (Figures 3A and 3C) and ISH (Figures 3B and 3D) techniques stained chief cells, which in the rat are preferentially located in the lower region of the oxyntic glands but are absent in the pyloric glands. In these cells the immunolabeling had a granular aspect and was specifically concentrated in the apical region (Figures 3C and 4A), whereas the non-granular staining with ISH was widespread throughout the cell (Figure 3D).

Table 1

Detection of nNOS in rat gastric epithelium using immunohistochemistry (IHC) and in situ hybridization (ISH) techniques

	Epithelial cells
Techniques	Mucosecretory	Chief	Endocrine	Parietal	Control neurons
IHC-A1 a	+	+	+	–	+
IHC-A2 b	+	+	–	–	+
IHC-A3 c	–	–	–	–	+
ISH	+	+	ni d	–	+

^a

Antibody against the synthetic peptide no. 53 of the rat nNOS.

^b

Antibody against the recombinant rat nNOS.

^c

Antibody against human nNOS from Transduction Laboratories.

^d

ni, not identified.

Figure 7

Specificity controls for ISH. Pairs of consecutive paraffin sections treated with the anti-sense (A,C) and the sense probe (B,D). The stain disappeared in both neurons of the enteric plexi (A,B) and epithelial cells of the gastric mucosa (C,D). Bars: A,B = 100 μm; C,D = 30 μm.

Under the electron microscope, immunostaining for nNOS using the A1 antiserum was also found in chief cells, the signal appearing only in the secretory granules (Figures 4B and 4C).

Mucosecretory Cells. These cells were stained for nNOS both with IHC (Figures 3A and 3E) and ISH (Figures 3B and 3F) techniques. With both techniques, the intensity of the stain was lower than in the chief cells.

Mucous cells in the neck portion of the oxyntic glands were stained with both the A1 and the A2 anti-serum (Figure 3E), whereas immunolabeling throughout the pyloric glands was achieved only with the A2 antiserum. The pattern of immunoreactivity of the A2 antiserum was similar to that of the mRNA staining.

Endocrine Cells. Some cells displaying the typical features of gastric endocrine cells were detected with the antiserum A1 in both the oxyntic (Figure 5A) and the pyloric regions (Figure 5B), although they were more easily recognized in the pyloric antrum owing to the absence of chief cells in this region. In the oxyntic region, the stained endocrine cells were more visible when the antiserum was diluted more than the optimal dilution of 1:500, which decreased the staining of chief cells (Figure 5A).

Discussion

This study shows that nNOS mRNA and nNOS protein are present not only in a unique cell type but also in several cell types of the rat gastric epithelium, i.e., the chief, mucous, and endocrine cells. It is important to highlight the coincident pattern of stain we have obtained with immunohistochemical and in situ hybridization techniques, although the difficulty in identifying endocrine cells in ISH-stained sections could be due to their sparse distribution among other more abundant, strongly positive cells: chief cells in the oxyntic glands and mucosecretory cells in the pyloric glands.

The first evidence for the presence of a high content of a calcium-dependent constitutive NOS in the gastric mucosa came from biochemical studies (Brown et al. 1992b). This activity was due not only to eNOS from blood vessels but also to another NOS localized in the gastric epithelium. However, the studies that tried to identify the cell type responsible for the production of NO in the gastric epithelium of different mammals, using in situ techniques, have reported very contrasting results. The research carried out by Brown et al. 1992b) on dispersed gastric epithelial cells suggested that the mucosecretory cells are the NO-producing cells. With immunohistochemical methods, nNOS had been detected in the epithelium of the rat forestomach (Schmidt et al. 1992), brush cells (Kugler et al. 1994), chief cells (Fiorucci et al. 1995b), some endocrine cell types (Akiba et al. 1995a; Burrell et al. 1996), and mucosecretory cells (Brown et al. 1992b; Price et al. 1996; Byrne et al. 1997; Ichikawa et al. 1998; Price and Hanson 1998). In none of these studies were parietal cells reported to be positive for NOS. Our results encompass the different results obtained in previous research that has detected NOS independently in only one of these three cell types: mucous, chief, and endocrine cells.

Given the striking difference in results, we made an effort to optimize the conditions for the in situ techniques to clarify the differing results about the NO-producing cells in the gastric epithelium. We tested several antisera raised against different regions of nNOS on material processed with different fixatives, with or without MW pretreatment. The use of MW very much enhanced the immunoreactivity in control tissues, thus suggesting that it is an important factor in unmasking this antigen and obtaining a more complete pattern of immunoreactivity (Shi et al. 1991, 1993), at least with the nNOS antisera used in this study. In the gastric mucosa, the MW pretreatment also improved the immunostaining, both enhancing the intensity of the immunostaining of cells that were positive without pretreatment and unmasking the immunoreactivity in new positive cell types. As far as we know, this is the first time that MW heating has been used for the retrieval of NOS antigen. This technical aspect can explain, in part, the fact that our results encompass some of the different results of the previous immunohistochemical works in the gastric mucosa.

Our results indicate that in the stomach, as in other systems (Schmidt and Walter 1994), several cell types are able to synthesize NO. These nNOS-immunoreactive cell types in the gastric epithelium differ from each other in the pattern of immunoreactivity obtained with the different antisera used in this study, thus indicating possible dissimilarities between the nNOS forms present in each cell type. An interesting point is that the monoclonal antibody raised against human brain nNOS gave positive results in neurons of the rat stomach but not in the gastric epithelium. Because different studies have reported unexpected staining patterns with some commercially available NOS antibodies (Coers et al. 1998), in addition to IHC we also used the ISH technique for the detection of NOS mRNA. As far as we know, this is the first time that ISH has been used for study of NOS distribution in the gastric epithelium.

Chief Cells

These are the cell type that shows the most consistent labeling both with immunohistochemical and in situ hybridization techniques. Only in one of the previous studies was NOS activity reported in chief cells (Fiorucci et al. 1995a). These authors described the granular aspect of the NOS immunolabeling at the light microscopic level in isolated guinea pig chief cells. This granular immunostaining coincides with our results at both light and electron microscopy. Recent evidence suggests that the NO produced in chief cells appears to be involved in modulating pepsinogen-secreting cells (Fiorucci et al. 1995a).

Mucosecretory Cells

Several previous works reported that mucosecretory cells were the cell type responsible for the generation of NO in the gastric epithelium (Brown et al. 1992b; Price et al. 1996; Byrne et al. 1997; Ichikawa et al. 1998; Price and Hanson 1998). In our study, NOS protein and mRNA have also been detected in mucosecretory cells. These findings are in agreement with a possible role for NO in modulation of epithelial integrity and for secretion of mucus or bicarbonate, which are both important mechanisms of mucosal defense against luminal aggressors in the stomach (Whittle et al. 1990; Brown et al. 1992a).

Endocrine Cells

Only in our previous results (Burrell et al. 1996) and in those from another group (Akiba et al. 1995a, 1995b) have gastric endocrine cells been considered as a possible source of NO. The presence of NOS has been documented in several endocrine tissues in which NO may play an important role in regulation of hormone secretion.

We propose that, in the digestive system, NO could derive not only from different tissues (e.g., neurons, epithelium, muscle), as is proposed by Schmidt and Walter (1994) for many systems, but also from different cell types from the same tissue, as we have described for the gastric epithelium in the present study. The NO produced by the different cell types may form a complex network of paracrine communication with an important role in gastric physiology.

Note Added in Proof

While this manuscript was in revision, Cao et al. (J Histochem Cytochem 48:123–131, 2000) reported the use of MW for retrieval of NOS.