An immunohistochemical study of the distribution of nitric oxide synthase-immunoreactive neurons and fibers in the reticular groove of suckling lambs

Abstract

The reticular groove (RG) is a specialized region of ruminant forestomach which, in suckling animals, via a vagovagal reflex, transforms itself into a tube to ensure the direct transport of milk from the esophagus to the abomasum. The nervous mechanism controlling the RG movement is not fully understood; however, at this level, the enteric nervous system (ENS) shows the highest neuronal density when compared with other forestomach compartments. Because nitric oxide is considered the putative major mediator of non-adrenergic non-cholinergic smooth muscle relaxation, the aim of the present study was to investigate the ENS of the RG of suckling lambs, both in the floor and in the lip, with particular regard to nitric oxide synthase (NOS)-immunoreactivity (-IR), by means of double immunohistochemical staining. NOS antiserum was used in association with some neurochemical markers which have been utilized by many authors in ENS. A rich innervation of fibers extended along the entire length of the RG. Proceeding distally, the number of neurons stained with a pan-neuronal marker increased; they were more numerous in the lips and lip–floor junction than in the floor itself. However, the percentage of NOS-IR neurons was the same in the proximal and distal parts. Many NOS-IR neurons often co-expressed galanin and dopamine β-hydroxylase. Neurochemical markers, such as calbindin, calcitonin gene-related peptide, IB4 and neurofilament 200 kDa, usually used to identify primary sensory neurons were not expressed in RG neurons, and the co-localization of NOS with tyrosine hydroxylase and substance P was rarely found. When compared with other districts, the RG showed some peculiar aspects, such as the lack of both neurons in the submucosal plexus and the lack of typical sensory neurons.

Introduction

The ruminant forestomach consists of specialized regions, each having a distinct function and motility. In particular, the reticular groove (RG) of the suckling lamb functions as a bypass between the esophagus and the abomasum to ensure the direct transport of milk via a vagovagal reflex integrated in the gastric centers of the medulla oblongata (Titchen, 1976). Following vagal stimulation of the afferent nerve endings situated in the mouth and pharynx, the lips of the RG, extending from the cardia to the reticulo-omasal orifice, strongly contract coming into close apposition; a twisting movement of the right lip then draws the epithelium of the reticulum adjacent to it and, in this way, a tube for milk transport is formed (Phillipson, 1970). The responsiveness of the RG reflex normally diminishes with age; in the adult, drinking water seldom evokes this reflex; the RG reflex has therefore received attention for its potential use in the oral administration of drugs (Sharifi et al. 2009). The nervous mechanism controlling the RG movement is not fully understood; however, when compared with other forestomach compartments, the RG shows the greatest amount of vagal extrinsic innervation (Morrison & Habel, 1964; Chiocchetti et al. 2003). Moreover, the RG shows the highest densities of ganglia and nerve fibers in the intramural plexuses with respect to other forestomach compartments (Kitamura et al. 1986; Teixeira et al. 1998). Although forestomach motility is strongly regulated by sympathetic and parasympathetic extrinsic innervation, enteric neurons influence muscle activity by releasing excitatory and inhibitory neurotransmitters, such as serotonin (Vassileva, 1977), vasoactive intestinal peptide (Denac et al. 1990), acetylcholine (Vassileva et al. 1978) and nitric oxide (NO; Schneider & Eades, 1998). Intrinsic nerve pathways might be capable of coordinating the muscle activity of all the forestomach compartments in response to the vagal impulses received in the RG; in fact, although motility is dramatically impaired after a vagotomy, it is, however, still present (Gregory, 1982). The demonstration that forestomach motility is blocked following surgical ablation or cauterization of the RG (Borgatti et al. 1959) further supports the hypothesis that the RG may have an important role as the coordinating center for the control of reticulo-ruminal motility (Kitamura et al. 1986).

To study the enteric nervous system (ENS) of mammals, physiological, pharmacological and immunohistochemical studies have generally been carried out on the small or the large intestine. In the ENS, two main ganglionated plexuses are present, both in small and large mammals: the myenteric plexus (MP), lying between the longitudinal and circular muscle layers and mainly regulating muscle activity, and the submucosal plexus (SMP), embedded in the submucosal connective tissue and mainly regulating mucosal functions. Sensory neurons, motor neurons and interneurons are organized into functional reflex circuits, which make the ENS capable of integrated reflex behavior, without the involvement of the central nervous system (CNS) (Wood et al. 1999). There are only a few immunohistochemical studies referring to the ENS of the RG of bovines (Kitamura et al. 1986, 1987; Barahona et al. 1997; Teixeira et al. 1998) and sheep (Pfannkuche et al. 2003). Pharmaco-physiological studies on the RG of calves have demonstrated that both nor-adrenaline (Denac et al. 1991) and acetylcholine (Barahona et al. 1997) play an essential role in the modulation of the activity and maintenance of muscle tone; in addition, a functional study involving esophageal groove strips revealed that NO plays a role in non-adrenergic non-cholinergic relaxation, which is significantly inhibited by incubation with L-NOARG, an inhibitor of NO synthesis (Barahona et al. 1998).

The aim of the present study was to investigate the presence and distribution of NO-producing neurons and their phenotype by means of histochemical and double-immunofluorescence stains in cryostat sections in suckling lambs. Neuronal nitric oxide synthase (NOS) antiserum was used in association with antisera raised against neurochemical markers that have been utilized by other authors to identify adrenergic, sensory, inhibitory and excitatory neurons and fibers, such as: tyrosine hydroxylase (TH), dopamine β-hydroxylase (DBH), substance P (SP), galanin (GAL), calcitonin gene-related peptide (CGRP), calbindin (CALB), isolectin B₄ (IB4) derived from the Griffonia simplicifolia plant, and neurofilament 200 kDa (NF).

Materials and methods

The experiments described in this report were carried out in agreement with Italian legislation on experimental animals. The RG was dissected along its entire length (from the esophagus to the reticulo-omasal orifice) in three suckling lambs, 1–3 days old, which were slaughtered at the public slaughterhouse, pinned tautly on balsa wood and separated into proximal (the half next to the cardia) and distal (the half next to the reticulum-omasum orificium) parts. The specimens were fixed in 4% paraformaldehyde in phosphate buffer (0.1 m, pH 7.2) at 4 °C for 24 h, then rinsed in 0.01 m phosphate-buffered saline (PBS) and stored at 4 °C in PBS containing 30% sucrose and sodium azide (0.1%). The samples were frozen in isopentane cooled in liquid nitrogen. Transverse (i.e. perpendicular to the floor of the groove) cryosections having a thickness of 20 μm were obtained along the entire length and mounted on poly-l-lysine- (Sigma Aldrich Chemie) coated slides, and stored at −20 °C until histochemical and immunohistochemical staining were carried out. To best define the histological aspect, some sections were stained with Masson's trichrome (McManus & Mowry, 1960).

Immunohistochemistry and NADPH-d staining

Cryostat sections were processed by double-labeling indirect immunofluorescence. NOS was detected both in combination with neuron-specific enolase (NSE; to describe the overall distribution of innervation of the RG and to evaluate the percentage of NOS neurons with respect to the total number of neurons) and with TH, DBH, SP, GAL, CGRP, CALB, IB4 and NF. Table 1 summarizes the antibodies used and their dilutions. The sections were fixed in acetone at 4 °C for 10 min, washed in PBS and then incubated in 2% of the appropriate normal serum in PBS containing 1% Triton X-100 at room temperature (RT) for 30 min to permeabilize the tissue to the antisera. The sections were then incubated in a humid chamber at 4 °C for 2 nights in the mixture of primary antibodies diluted in PBS. After washing for 3 × 5 min in PBS, the sections were incubated for 1 h at RT with a mixture of fluorescein isothiocyanate (FITC)-conjugated and tetramethyl rhodamine isothiocyanate (TRITC)-conjugated secondary antibodies (or the secondary TRITC-conjugated antibody plus streptavidin FITC-conjugated if biotinylated IB4 was used) diluted in PBS. The sections were then washed in PBS and coverslipped with buffered glycerol, pH 8.6.

Table 1.

List of antibodies used in this study

Neuronal markers	Species	Code	Dilution	Supplier
NSE	Rabbit	N1516	1 : 5	DAKO
NOS	Mouse	Sc-5302	1 : 200	Santa Cruz
NOS	Rabbit	SA-227	1 : 300	Biomol
GAL	Rabbit	IHC 7153	1 : 400	Peninsula
DBH	Mouse	MAB308	1 : 200	Chemicon/Millipore
TH	Mouse	MONX10786	1 : 60	Monosan
SP	Rat	10-S15A	1 : 400	Fitzgerald
CGRP	Rabbit	C8198	1 : 800	Sigma-Aldrich
CALB	Rabbit	CB-38	1 : 800	Swant
NF200 kDa	Rabbit	N 4142	1 : 1000	Sigma-Aldrich
IB4	–	L2140	1 μg mL	Sigma-Aldrich
Secondary antisera
FITC-conjugated donkey anti-mouse IgG			1 : 100	Jackson
TRITC-conjugated donkey anti-rabbit IgG			1 : 500	Jackson
Alexa 594-conjugated donkey anti-rat IgG			1 : 400	Mol. Probes
FITC-conjugated Streptavidin			1 : 100	Vector

CALB, calbindin; CGRP, calcitonin gene-related peptide; DBH, dopamine β-hydroxylase; GAL, galanin; IB4, isolectin B₄; NF200 kDa, neurofilament 200 kDa; NOS, nitric oxide synthase; NSE, neuron-specific enolase; SP, substance P; TH, tyrosine hydroxylase.

For NADPH-d staining, additional sections were left to dry at RT before being incubated at 37 °C for 45 min in a 0.1 m Tris–HCl buffer (pH 8) plus 0.3% Triton containing 0.6 mm nitroblue tetrazolium and 2.2 mmβ-NADPH-d, according to Grozdanovic et al. (1992). The sections were then washed in PBS and coverslipped as described above.

Quantitative and statistical analysis

The specimens were examined by a Zeiss Axioplan microscope equipped with appropriate filter cubes for distinguishing between FITC and TRITC. Double-labeling was evaluated by switching the filter cube. The percentages of neurochemically identified NOS neurons were calculated in non-consecutive sections, using NOS/NSE stain, and at least 100 cells were counted both in the lips and in the floor of each subject. For double-stained NOS/GAL and NOS/DBH, at least 50 neurons were counted for each subject both in the lips and in the floor. The images were recorded with a Polaroid DMC digital photocamera (Polaroid, Cambridge, MA, USA), and the dmcv 2 software images were further processed using Corel Photo Paint and Corel Draw software programs (Corel, Milan, Italy). The values of the lip–floor junctions were included in the lip count. The percentages obtained using each immunohistochemical stain were first tested for normality using the Shapiro–Wilk's W-test. As the data showed a normal distribution, comparisons were made using the anovat-test, with a conventional 1% level being used to define statistical significance. In detail, for both the lip and floor, the percentages (Table 2) were first compared between the cranial and caudal parts, then the percentages of both lips and floors of the three subjects were grouped together irrespective of cranial and caudal considerations, and the lips and floors were compared. Values are expressed as mean ± standard deviation (SD).

Table 2.

Percentages of NOS-IR neurons co-localized with other markers

	NOS/NSE	NOS/GAL	GAL/NOS	NOS/DBH
Proximal lips	24.2 ± 9.5 (302/1250)	47 ± 7.5 (150/319)	28.4 ± 8.9 (150/529)	39.8 ± 25 (153/384)
Distal lips	27.6 ± 11.8 (303/1099)	38.5 ± 8.3 (152/394)	42 ± 22.7 (152/362)	37.5 ± 17.4 (152/405)
Proximal floor	48.1 ± 19.4 (303/630)	75.4 ± 12.2 (150/199)	82 ± 22.2 (150/183)	58.1 ± 7.3 (150/258)
Distal floor	48.2 ± 2.9 (302/626)	84.7 ± 5.4 (155/183)	81.2 ± 3.7 (155/191)	32.3 ± 18 (291/900)

The number of cells, counted in three subjects, are in parentheses. Values are expressed as mean ± SD.

DBH, dopamine β-hydroxylase; GAL, galanin; NOS, nitric oxide synthase; NSE, neuron-specific enolase.

Specificities of the antibodies

Wood et al. (2005) tested the specificity of the mouse NOS antibody on sheep by Western blot analysis; in past unpublished experiments we found total co-localization of this antibody with both antibodies used in this study. In addition, in past studies (Lalatta-Costerbosa et al. 2007) and in the present study we found a total overlapping between histochemical and immunohistochemical stained enteric neurons. For CGRP, the specificity was tested by Chiocchetti et al. (2006), for SP by Pfannkuche et al. (2002), for TH and DBH by Russo et al. (2009), and for IB4 by Mazzoni et al. (2010). The specificity of the CALB and GAL antibodies has previously been published (Dunning et al. 1986; Reiche et al. 1999). We believe that these antibodies recognize the appropriate antigens also in ruminants. However, they have not been directly tested in this study. We did not test the specificity of NF and, as we cannot exclude cross-reaction with other related proteins, the term like immunoreactivity (-LI) will be used hereafter for these neurotransmitters. To test the specificity of the secondary antibodies, they were applied without the use of primary antibodies. No staining was seen after omitting the primary antibodies.

Results

Histological aspect and NADPH-d histochemical staining

Two thick lips, right and left, delimited the floor. The entire RG was limited by a stratified squamous epithelium. The muscularis mucosae was not continuous and, in the lips, thick fascicles were immersed in the loose connective tissue embedded in the epithelium. Mucous glands (Fig. 1a) were found in the loose connective tissue of the sub-mucosa at the level of lip–floor junctions; they extended along the entire length of the RG and became most numerous in the distal part, where they also extended between the internal and external muscle layers. In the floor, the tunica muscularis was composed of a thin uncompleted external longitudinal layer (Fig. 1b) and a thick inner circular layer; in the most cranial part of the floor, the muscular layer was formed by striated fibers (Fig. 1c), and elsewhere by smooth muscle cells. Some circular fascicles were bent at the level of the lip–floor junctions and joined the lips as small muscular fascicles surrounded by very developed longitudinal fascicles. This latter was very developed in the lips in contrast to the thin uncompleted longitudinal layer of the floor. Both in the floor and in the lips, many NADPH-d fibers were observed along the entire length of the RG, parallel to the longitudinal axis of the muscle fascicles. No obvious gradient of innervation was detected. Only a few NADPH-d fibers were located in the external longitudinal muscle layer of the floor and in the submucosa, and few fibers were detected in close apposition to the epithelium. NADPH-d neurons were located in the MP; they were more numerous in the floor than in the lips, and we rarely observed neurons located singly or grouped into small clusters of two–three cells among the muscle fascicles of the lips.

Fig. 1.

(A) Transverse section of the right half of the reticular groove of lambs after NADPH-d staining at the distal level. Mucous glands (arrowheads) are present in the submucosa of the lip-floor junction, and between the circular (1) and longitudinal (2) muscle layers of the lips. (B) A higher magnification of the floor after Masson's trichrome staining to show the internal (1) and the external interrupted muscle layer (2). (C) Internal muscle layer of the floor after NADPH-d staining to show the nervous fibers parallel to the striated muscle fibers. Bars: A = 800 μm; B = 100 μm; C = 50 μm.

NSE immunoreactivity

NSE-immunoreactive (-IR) nerve fibers were observed along the entire length of the RG, there was no obvious gradient of innervation between the proximal and the distal parts. Numerous nerve trunks were detected in the MP and they branched through the fascicles of the muscular layer to form a dense network around smooth muscle fascicles both of the tunica muscularis and the muscularis mucosae (Fig. 2a). In the connective tissue of the submucosa thin fascicles or isolated fibers formed a fine plexus. Fibers were sometimes found in close apposition to the basal layers of the epithelium (Fig. 2a) or between the adenomers of the mucous glands (Fig. 2b), and nerve fibers could frequently be seen around the blood vessels forming the perivascular plexuses. The neurons were unevenly distributed; they were most numerous in the lips and their number increased proceeding distally. In one subject in which all NSE neurons were counted in a section of both the proximal and the distal parts, in the proximal part, we counted 80 neurons in a lip (a mean value of 8.5 cells per ganglion) and 54 neurons in the floor (a mean value of 9 cells per ganglion); in the distal part, we counted 271 neurons in a lip (a mean value of 24.6 cells per ganglion) and 143 in the floor (a mean value of 13 cells per ganglion). The neurons were generally clustered in the ganglia varying in size; large ganglia (up to 50 neurons) were located in the MP, and small ganglia (two–four neurons) or isolated neurons were observed intermingled with the longitudinal muscular fascicles of the lips. No neurons were found in the SMP.

Fig. 2.

Immunohistochemical staining with NSE (A-B), SP (C) and CGRP (D). (A) Note the immunoreactive fibers (arrowheads) in close apposition to the stratified epithelium (asterisk) and in the lamina propria (arrow); (B) immunoreactive fibers around the adenomers of the mucous glands (asterisks). (C) SP-IR fibers form perivascular plexuses around the blood vessels of the floor (arrowhead) and, (D) there are CGRP-IR varicose fibers (arrowhead) around the CGRP negative neurons (asterisks) of the myenteric plexus of the floor. Bars: A, D = 100 μm; B = 50 μm; C = 200 μm.

NOS immunoreactivity

Immunohistochemical staining confirmed the NADPH-d observations; however, with the immunohistochemical staining, the morphology of the neurons was better defined. NOS-IR neurons showed great variability both in size and morphology; they exhibited an irregular outline with short, well-stained processes or an ovoid shape. In these neurons, a long process arising from the side opposite the nucleus could sometimes be seen, without showing a given polarity (Fig. 3a,b). The neurons were more numerous in the floor than in the lips, and their number increased proceeding distally; however, their percentage was similar in the proximal and distal parts (Table 2). Almost all ganglia in the floor contained NOS-IR neurons, which were often arranged around the periphery of the ganglia, and some ganglia showed only NOS-IR neurons. NOS-IR neurons were not found in the submucosa. About 80% of the NOS-IR neurons in the floor co-expressed GAL (Table 2; Fig. 4a,b); this percentage decreased in the lips (Table 2). Many neurons co-expressed DBH (Table 2; Fig. 4c,d), but rarely (< 0.5%) TH and SP. We never observed NOS-IR neurons expressing CGRP, CALB, IB4 and NF. We also never observed statistically significant differences between the proximal and distal parts in the percentages of NOS-IR neurons and their co-localization in either the lips or floors. On the contrary, a statistically significant difference (P≤ 0.01) was observed between the lips and floors when we grouped the percentage of the three animals with the exception of the colocalization NOS/DBH.

Fig. 3.

NOS immunohistochemical staining of the floor. NOS-IR neurons show variability in size and morphology. Small neurons are intercalated to large ones. Arrowheads indicate neurons with irregular outlines; arrows indicate ovoid neurons with a long process starting from the side opposite the nucleus without showing a given polarity. Bars: 100 μm.

Fig. 4.

Double immunofluorescence staining with NOS (A, C, E, G) and GAL (B), DBH (D), TH (F), SP (H). (B) Many NOS-IR neurons varying in size and morphology co-express GAL (arrowheads); (D) all DBH-IR neurons co-express NOS and show a generally ovoid shape; (F) TH-IR neurons (arrowheads) are intermingled with NOS-IR neurons and no co-localization is present; (H) an SP-IR single-stained neuron (asterisk) and SP-IR fibers forming a basket around the NOS-IR neurons showing both irregular (arrowhead) or ovoid (arrow) profiles. Bars: A, B, E, F = 100 μm; C, D, G, H = 50 μm.

GAL immunoreactivity

Many GAL-LI fibers were observed along the entire length of the RG; as NOS-IR fibers they were most numerous in the circular layer of the floor and in the longitudinal layer of the lips. In the floor, about 80% of the GAL-LI neurons, varying in size and morphology, co-expressed NOS-IR; this percentage decreased in the lips (Table 2; Fig. 4a,b). We also did not observe any statistically significant difference between the proximal and distal parts of the lips and floors, regarding the percentage of GAL/NOS. On the contrary, a statistically significant difference (P≤ 0.01) was observed between the lips and floors when we grouped the percentages of the three animals.

DBH-IR

Many DBH-IR fibers were detected around the blood vessels, and among the muscle cells of both the floor and the lips. All DBH-IR neurons co-expressed NOS, only two DBH neurons did not co-express NOS. Many, but not all of them, showed an ovoid shape (Fig. 4c,d). No statistically significant differences were observed.

TH-IR

TH-IR fibers were found in the myenteric ganglia and around the blood vessels; few fibers were observed among the muscle fascicles, both in the floor and in the lips. TH-IR neurons were intermingled with NOS-IR neurons (Fig. 4e,f); only in one subject were three TH-IR neurons co-expressing NOS-IR detected.

SP-IR

SP-IR fibers were observed along the entire length of the RG and were unevenly distributed. Many SP-IR fibers were detected in the longitudinal muscle layer of the lips and in the muscularis mucosa. In the floor, only a few SP-IR fibers were found among the fascicles of the internal muscle layers; on the contrary, they were numerous in the myenteric ganglia where they showed numerous varicosities and formed a basket around the NOS-IR neurons (Fig. 4g,h). In the circular muscle layer of the lips and in the loose connective tissue of the submucosa fibers, SP-IR fibers were not numerous; perivascular plexuses (Fig. 2c), fibers in close apposition to the epithelium and a few fibers between the adenomers of the mucous glands were also detected.

CGRP-IR

Both in the floor and in the lips, only a few CGRP-IR fibers were present among the smooth muscle cells and in the submucosa, sometimes in close apposition to the epithelium. In the myenteric ganglia, varicose CGRP-IR fibers formed a basket around the CGRP-negative neurons (both NOS positive and negative; Fig. 2d). No CGRP-IR neurons were found.

CALB-IR

Only a few CALB-LI fibers were present in the circular muscle layer of the floor, more numerous in the proximal than in the distal part of the RG. The fibers were sometimes arranged around the NOS-IR neurons. CALB-LI neurons were rarely found.

NF-IR

Some NF-LI fibers were found in the MP, sometimes in close apposition to the NOS-IR neurons. Few NF-LI fibers were located in the muscularis mucosa and in close apposition to the epithelium. NF-LI neurons were rarely observed.

IB4 stain

No stained neurons were found with this neurochemical marker.

Discussion

The present study describes the ENS organization in the RG of lambs with particular attention to NOS-IR neurons and fibers. ENS neurons, stained with NSE, were unevenly distributed; their number increased in the distal part and they were most abundant in the lip and floor–lip junctions. In contrast, there was no obvious gradient of innervation between the proximal and the distal parts as concerned the distribution of fibers. This might indicate that the proximal part of the RG is more influenced by extrinsic innervations with respect to the distal part, which might be more involved in the regulation of food processing, regulating the passage of large food particles into the omasum by mean of their unguiculiform papillae (Teixeira et al. 2009). In contrast, NOS-IR fibers did not show an obvious gradient of innervations, and the percentage of NOS-IR neurons was the same in the proximal and distal parts. Contrary to what has been observed in the intestinal plexuses of both small and large mammals, including sheep (Lalatta-Costerbosa et al. 2007), no neurons were detected in the SMP. Similarly, the SMP neurons, which are generally involved in secretion and vasodilation, are very few or absent in the esophagus and the stomach of many mammals (Furness, 2006), including the sheep forestomachs (Yamamoto et al. 1995). In particular, Kitamura et al. (1986), in the RG of suckling, weaning and adult cattle, did not find any neurons, in contrast to Barahona et al. (1998), who in adult cattle found many NOS-IR neurons in both the MP and SMP. Therefore, it is plausible to think that the myenteric ganglia in the lamb RG control the mucosal functions.

However, in many cases, NO does not appear to be a primary transmitter and is often co-localized with other putative inhibitory neurotransmitters (Chiocchetti et al. 2006). In the present study, many NOS-IR neurons co-expressed GAL. This neuropeptide is widely expressed in the CNS as well as in the ENS; several studies have indicated that it is involved in many physiological processes including cognitive performance, feeding behavior, nociception and nerve regeneration. In the gastrointestinal tract (GIT), the best known effect of GAL is its ability to regulate GIT motility, either acting as a neurotransmitter by binding to three different G-protein-coupled receptors, all three expressed in the GIT smooth muscle cells, or acting as neuromodulator by both increasing and decreasing the release of neurohumoral substances (Gross & Pothoulakis, 2007). In sheep myenteric forestomach neurons, the lack of the GAL R1 receptor and only a few neurons with both R2 or R3 have recently been observed (Arciszewski et al. 2008), which supports the minor role of GAL in the indirect mediation of forestomach motility. The extensive co-localization of NOS and GAL in the floor suggests that both neurochemical markers can act as inhibitory neuromodulators/neurotransmitters for both smooth and striated cells, both of which are present in the lamb RG. Widespread co-localization of NOS and GAL has been evidenced in many neurons and fibers in many districts, including the enteric neurons co-innervating the esophageal endplates. Recently, in rats, Boudaka et al. (2009)in vitro vagus nerve–esophagus preparations demonstrated the indirect inhibitory role of GAL on vagally mediated striated muscle contraction. The authors hypothesized that GAL may be co-released with NO from enteric neurons as part of a local neuronal reflex arc thus modulating cholinergic vagal contraction. The inhibitory role of NO on the muscle cells of the RG of bovines has been clearly demonstrated by Barahona et al. (1998), and may be supported also by the lack of the co-expression of SP in nitrergic neurons in the present study. In fact, analogously to what has been observed in the GITs of many other mammals, it is plausible to hypothesize that SP has an excitatory role (Brookes, 2001); furthermore, both acetylcholine and SP evoke contractions of the isolated ruminal smooth muscle (Vassileva et al. 1978; Veenendaal et al. 1982). The excitatory effect of SP is also hypothesized by Pfannkuche et al. (2003), who found extensive co-localization of SP and acetylcholine in the RG of sheep and whose contractile effects on the reticulum have been reported by Wong & McLeay (1988). The prevalent distribution of this neuropeptide in the lips and in the lip–floor junction supports the great importance of excitatory stimuli to contract the lips and form the tube for milk transport following vagal stimulation. However, the presence of SP-IR fibers in close apposition to the epithelium or among the adenomers of the mucous glands indicates a possible different role of SP, other than having an excitatory effect on the smooth muscle cells, it may also regulate secretion as already suggested by Wathuta (1986), in the sheep digestive system.

In the RG lips, the percentage of NOS-IR neurons decreased as did the colocalization of NOS/GAL, and many single-stained NOS or GAL neurons were detectable; it is then possible to hypothesize a different role for both neurotransmitters, such as involvement in the control of mucosal function as occurs in other districts, and which could explain the presence of NOS-IR and GAL-LI fibers in the submucosa. Moreover, for many neurotransmitters, a range of activities and variations between regions and species have been demonstrated. For example, in the small and large intestine of guinea pigs, CALB, which in many cases co-expressed IB4 (Hind et al. 2005), is the putative neurotransmitter of intrinsic primary neurons (Chiocchetti et al. 2006) but, in the sheep ileal plexus, CGRP is a more selective marker of the sensory neurons. In the present study, we rarely found neurons expressing CALB, CGRP or IB4. Consequently, we can hypothesize that the sensory fibers of the RG may have an extrinsic origin coming from the spinal ganglia, in which both CGRP and NF neurons have been shown to project into the sheep duodenum and ileum (Mazzoni et al. 2010), or from the distal vagal ganglia or from primary sensory neurons located in other districts of the forestomach. Alternatively, it is tempting to speculate that, in the RG, intrinsic primary neurons, through which the enteric reflex is initiated (Furness et al. 2004), express unknown specific neurotransmitters, different with respect to what has been observed in other districts of the same species. However, such hypotheses need to be further explored by morphological and electrophysiological studies.

In the RG, as previously observed in the lamb ileum (Chiocchetti et al. 2004), TH-IR neurons were present in the ganglia intermingled with NOS-IR neurons. In contrast to our previous observations in the sheep ileum, however, no co-localization of TH/NOS was observed, whereas many NOS-IR neurons co-expressed DBH. The presence of catecholaminergic neurons in the intramural cells is unusual, although intramural noradrenergic neurons have been described in the porcine urinary bladder (Pidsudko, 2004); we therefore think that, as has been observed in mammalian fetal and neonatal organs (Tharakan et al. 1995; Dixon et al. 1998), DBH in the RG of lambs is transiently expressed as the expression of crest-derived cells, and DBH cells may be the progenitors of NOS neurons, such as has been demonstrated by Young et al. (2002) in the ENS of mice for TH-expressing neurons.

In conclusion, this study shows a rich innervation in the RG of lambs having a peculiar aspect with respect to the other districts of the same species, probably related to its particular function in suckling and adult animals. In particular, the general organization of intramural plexuses and the lack of typical primary sensory neurons should induce some caution when extrapolating data from other species or from other districts of the same species.