The peritoneum is a large serous membrane laying in the abdominal wall and over the visceral organs. The part that lines the abdominal wall is named the “parietal peritoneum,” whereas the rest that reflects over the visceral organs is named the “visceral peritoneum.” It is well known that sensory nerves innervate both the visceral peritoneum (Bentley et al., 1981; Gebhart, 2000; Ward et al., 2003) and the parietal peritoneum (Bahns et al., 1986). The parietal peritoneum is innervated by the neurons of the dorsal root ganglia (DRG) that supply the muscles and the skin of the abdominal wall, whereas the visceral peritoneum is innervated by not only the neurons of the vagal ganglia but also the DRG (Warwick and Williams, 1973; Tanaka et al., 2002). There are several reports about the innervation of sensory afferent fibers in the serous membrane covering the gut (Iggo, 1986; Ward et al., 2003; Song et al., 2009; Zagorodnyuk et al., 2010). Iggo (1986) reported that there are distension-sensitive C-mechanoreceptor units in the serous membrane covering the gut or in the omentum. Song et al. (2009) described that mesenteric afferent mechanosensitive nerve endings correspond to varicose branching axons on mesenteric blood vessels. However, there have been few studies about the distribution and the ultrastructure of the sensory nerves supplying the parietal peritoneum.
In our previous study (Tanaka et al., 2002), the distribution of the cell bodies of sensory neurons innervating the peritoneum was studied using the retrograde tracer Fluorogold. When Fluorogold was applied on the parietal peritoneum, retrogradely labeled neurons were found only in the ipsilateral DRG and were distributed in a segmental fashion. When Fluorogold was applied to the visceral peritoneum, labeled neurons were found in the nodose ganglia in addition to the DRG on both sides. This study clarified the origin of the sensory neurons innervating both the visceral and the parietal peritoneum, and the different patterns of innervation in the visceral and the parietal peritoneum. However, the distribution of the nerve fibers and their terminal structure in the parietal peritoneum were not studied in detail.
This study was designed to reveal the morphology of sensory nerve fibers in the parietal peritoneum. We used immunohistochemistry with an antiserum against the protein gene product 9.5 (PGP9.5), a marker for nerve fibers (Doran et al., 1983), to identify the sensory nerve fibers. Furthermore, to distinguish the sensory afferent fibers from the sympathetic postganglionic fibers, double immunohistochemical staining for tyrosine hydroxylase (TH) with calcitonin gene-related peptide (CGRP) was performed in the parietal peritoneum.
MATERIALS AND METHODS
Twelve male Sprague-Dawley rats weighing 200–220 g were used. The rats were treated according to the animal care guidelines of the Hyogo College of Medicine. The Animal Care and Use Committee of the Hyogo College of Medicine approved all surgical procedures. The rats were anesthetized with sodium pentobarbital (40 mg/kg intraperitoneally). A midline abdominal incision was made. Both sides of abdominal wall including the parietal peritoneum and the muscles, caudal to the last rib and costal cartilage, lateral to the erector spinae muscle, and cranial to the inguinal ligament, were removed and pinned rapidly onto a flat silicone board with the peritoneal surface above. The abdominal wall was fixed by immersion with a 4% paraformaldehyde and 0.1% glutaraldehyde solution in 0.1 M phosphate buffer (PB) at pH 7.4 at 4°C for 60 min. We cut a 1-cm square of the abdominal wall, which was pinned onto another silicone board with the peritoneal surface on the bottom. The muscle fibers overlying the peritoneum were then carefully removed. The piece of peritoneum was pinned again onto small pieces of silicone board with peritoneal surface above, and processed for immunohistochemistry.
After a rinse with PB, the specimens were incubated with 1% bovine serum albumin in phosphate buffered saline (PBS) containing 0.9% NaCl and 0.1% Triton X-100 for 1 h at room temperature. They were then incubated with a rabbit anti-PGP9.5 serum (UltraClone, Isle of Wight, England, 1:400 dilution) for 1 day at 4°C. After rinsing with PBS, they were incubated with biotinylated goat anti-rabbit IgG (Vector Laboratories, Burlingame, CA) diluted 1:100 in PBS containing 0.2% bovine serum albumin for 60 min at room temperature. After rinsing with PBS, the pieces of peritoneum were incubated with the avidin-biotin complex (Vector) at a 1:100 dilution in PBS for 60 min at room temperature. They were then transferred to 0.05-M Tris HCl buffer (pH 7.2) and incubated with 0.02% 3,3′- diaminobenzidine tetrahydrochloride (DAB), 0.002% hydrogen peroxide in 0.05M Tris HCl buffer for 10 min. These whole-mount preparations of the peritoneum were postfixed with 1% osmium tetraoxide in PB for 1 h, dehydrated with ethanol and embedded flatly between the Aclar sheets in Spurr's resin.
We observed and photographed the flatly embedded specimens with a light microscope equipped with a digital camera system (DP70 Olympus). We made camera lucida drawings of the PGP9.5 immunoreactive nerve fibers on the monitor. The number of nerve endings directed toward the peritoneal surface and branches of the immunopositive fibers were counted in the photographs of 30 randomly selected regions of the peritoneum from three rats.
The serial ultrathin sections of labeled nerve fibers and endings were cut with an ultramicrotome, and observed with a JEOL 1200EX electron microscope after staining with uranyl acetate and lead citrate.
To distinguish primary afferent fibers from sympathetic axons, additional immunohistochemical experiments were performed. The whole-mount specimens of the parietal peritoneum and the peal of muscle layers of the duodenum fixed with 4% paraformaldehyde were incubated with 1% bovine serum albumin and 0.1% Triton X-100 in PBS for 1 h at room temperature. They were then incubated with mouse anti-TH serum (Chemicon, Temecula, CA; 1:500 dilution) and rabbit anti-CGRP serum (Yanaihara, Fujinomiya, Japan; 1:3000 dilution) for 1 day at 4°C. After rinsing with PBS, they were incubated with FITC conjugated donkey anti-mouse IgG (Jackson Laboratories, West Grove, PA; 1:200 dilution) and Cy3 conjugated goat anti-rabbit IgG (Jackson Laboratories, 1:200 dilution). The specimens were viewed with an Olympus AX-80 fluorescence microscope.
Immunohistochemistry with an antiserum against PGP9.5 revealed that many nerve fibers were labeled in the whole-mount preparations of the parietal peritoneum (Fig. 1A). The immunoreactive fibers were wavy, but generally ran straight. The labeled nerve fibers tended to run parallel to the direction of the intercostal nerves running in the abdominal wall (Fig. 1B). Branching of the fibers was sometimes observed. The number of branches were 2.14 ± 1.75 mm−2 (n = 30 regions from three rats). The immunoreactive fibers ran through the connective tissue between the layer of peritoneal cells and the muscle layer. The distal end of the labeled fibers did not branch or form characteristic morphological structure such as an end-net form or flower-like endings. The terminals of the fibers formed club-like endings (arrow in Fig. 2A). Thus, it was easy to distinguish the cut ends (arrowhead in Fig. 2A) and the terminals of the fibers. The density of the endings was 3.25 ± 1.66 mm−2 (n = 30 regions from three rats).
Double stainings for TH combined with CGRP were performed. Network of nerve fibers in the subperitoneal layer were labeled with CGRP (Fig. 3A), but they were not labeled with TH (Fig. 3B). To confirm the procedure of immunohistochemistry using these antibodies, wholemount preparations of the muscle layers of duodenum were also processed with TH and CGRP immunohistochemistry. Cy3-labeled CGRP immunoreactive fibers were found in the myenteric plexus (Fig. 3C), and FITC-labeled TH-immunoreactive fibers were found in the muscle layers in the same specimen (Fig. 3D).
The terminals of the imuunoreactive nerve fibers that contained DAB reaction product were trimmed, and serial ultrathin sections were made to observe the ultrastructure with the electron microscope. Electron microscopic observations revealed that the nerve bundle ran beneath the peritoneum, and through the subperitoneal space between the peritoneal cell layer and the muscle layer (Fig. 4). This nerve bundle was composed of a few thin myelinated nerve fibers. The nuclei of Schwann cells were often observed in the bundle. A layer of peritoneal cells, forming a simple squamous epithelium, divided the peritoneal cavity and the subperitoneal space. Collagen fibers were packed in both the subperitoneal space and the surface of the peritoneal cell layer. Myelinated nerve fibers diverged from the nerve bundle. Finally, the nerve fibers left the myelin sheath and penetrated the peritoneal cell layer to reach the peritoneal cavity (Figs. 5 and 6). The nerve endings were covered with collagen fibers in the peritoneal cavity. The fiber endings were thin (about 0.5–1.0 μm in width) with a club-like shape, and contained a few mitochondria, many neurofilaments but no synaptic vesicles (Fig. 5).
Unmyelinated nerve fibers were also observed in the parietal peritoneum. They were thin fibers, less than 1 μm in diameter, and had endings with slight swellings just inside the peritoneal cell layer (Fig. 2B). The nerve endings contained many neurofilaments, a few mitochondria, but no synaptic vesicles, just like the endings of the myelinated nerve fibers.
This article has described the morphology of the afferent fibers innervating the parietal peritoneum with immunohistochemistry using an antiserum against the neuronal marker PGP9.5. As PGP 9.5 is a general neuronal marker, the nature of the nerve fibers cannot be unequivocally identified. There could be efferent fibers supplying the abdominal muscles. Because, the efferent fibers leave the intercostal nerves on their way through the muscle layers which they supply, we considered that nerve fibers shown in this study running in the connective tissue between the muscle layers and the peritoneum were not efferent fibers. Postganglionic sympathetic nerve fibers might be found among unmyelinated fibers. Double stainings for TH combined with CGRP were revealed that the TH positive postganglionic sympathetic axons were observed only on the muscle fiber beneath the peritoneum but not in the subperitoneal layer. Thus, the PGP9.5 immunoreactive nerve fibers in the subperitoneal space shown in this study were thought to be afferent fibers coming from the DRG.
The labeled fibers tended to run parallel to the direction of the intercostal nerves running in the abdominal muscles underlying the peritoneum. The branches (2.14 ± 1.75 mm−2) and terminals (3.25 ± 1.66 mm−2) of the nerve fibers were observed. The labeled fibers were sometimes wavy or spiral. This may be due to the shrinkage of the tissue caused by fixation and dehydration. The densities of the nerve fibers and sensory end organs have been reported in the epidermis in the human hand (Kelly et al., 2005). The number of intraepidermal nerve fibers ***per mm of epidermis was 9.0 to 12.3 in the sections of the skin of the fingers. The density of intraepidermal nerve fibers was 12.4 fibers per mm in the sections of skin of human leg (Goransson et al., 2006), and the density of sensory corpuscles per millimeter square in the fingertip of digit III in the human finger was 33.02 (Nolano et al., 2003). In this study, the number of terminals of nerve fibers at the parietal peritoneum was 3.25 mm−2. Thus, the density of nerve terminals in the peritoneum is about one-tenth of those of sensory corpuscles in the human fingertip, which is one of the most sensitive regions in the human skin.
Electron microscopic observations of the immunoreactive nerve fibers revealed that there were both myelinated and unmyelinated afferent fibers in the parietal peritoneum. The unmyelinated fibers were less than 1 μm in diameter. We could not find unmyelinated nerve terminals that reached the peritoneal cavity across the boundaries of the peritoneal cell layers. They ended in the subperitoneal space.
We found myelinated fibers that formed nerve bundles beneath the peritoneal cell layer. Each myelinated fiber left the bundle and ran toward the peritoneal cavity, while keeping in contact with a Shwann cell. The most striking finding in this study is that the nerve fiber terminals penetrate through the peritoneal lining into the peritoneal cavity. The myelinated fiber made contact with a peritoneal cell, and became free from the myelin sheath. Finally, the neurits, which were covered with collagen fibers, penetrated the peritoneal cell layer to reach the peritoneal cavity without forming any special terminal structures.
In Figures 4 and 5, osmicated DAB reaction product are not recognized within the axons, although they are whole-mount specimens previously incubated for PGP 9.5. We could find out these nerve bundles in the specimen using light microscopy because of the brown DAB precipitate resulting the immunoreaction. The discrepancy between observations using electron and light microscopy may be the result of penetration problems of the immunoglobulin.
Cheng et al. (1997) reported that vagal sensory fibers originating from the nodose ganglion were distributed in the epicardium. They described that afferent fibers to the epicardium formed complexes associated with the cardiac ganglia. However, they did not mention about the fiber endings in the serous membrane. Pintelon et al. (2007) reported about the sensory receptors in the visceral pleura that covered the lung. Afferent fibers originating from the DRG formed slender bundles with a few nerve fibers. Separate nerve fibers generally gave rise to characteristic laminar endings associated with the elastic fibers of the visceral pleura.
In the skin, there are many kinds of receptor organs, such as Meissner, Merkel, Ruffini, and Pacinian corpuscles. These corpuscules are mechanosensors mediating various forms of touch or vibration sensation (Fawcett, 1986).
In addition to these receptor organs, free nerve endings are considered to be sensory terminals. Using the criteria established by Kruger et al. (1981), free nerve endings of both A delta and C fibers can be identified on the basis of their ultrastructural characteristics, that is, an intimate relationship between axons and the associated epithelium, the lack of a complete Schwann cell investment, and a 1:1 relationship between the axon and the investing Schwann cell for A delta terminals (Munger and Ide, 1988). The myelinated axons shown in Figures 4 and 5 in this study are considered to be A delta fibers on the basis of the size and the thickness of the myelin sheath. An intimate relationship between the axons and the peritoneal cells, and the lack of a complete Schwann cell investment were also observed. Thus, we considered the nerve endings observed in this study should be classified as free nerve endings. The free endings in the skin represent sensors for temperature and pain (Kruger et al., 1981). Sensory nerve fibers in the peritoneum may transmit information about pain, temperature, or pressure.
Thurston-Stanfield (2002) studied the effects of temperature and volume on intraperitoneal saline-induced changes in blood pressure, nociception, and neural activity in the medulla oblongata of the brain. The ON and OFF cells of the rostral ventromedial medulla were hypothesized to modulate nociception with ON cells facilitating pain and OFF cells inhibiting pain. He reported that the injection of a large volume of cold saline-induced excitation of ON cells and inhibition of OFF cells in the medulla. Such effects may be conveyed to the medulla by the sensory nerve fibers in the peritoneum to the medulla. Thus, these sensory nerve fibers in the peritoneum may receive information about not only pain but also about pressure or temperature.
The authors thank Ms. M. Hatta and Mr. K.Gion for their technical assistence.