THE JOURNAL OF COMPARATIVE NEUROLOGY 311:168-178 (1991)
Appearance of Somatostatin and Vasoactive Intestinal Peptide Along the Developing Chicken Gut MILES L. EPSTEIN AND KRISTIAN T. POULSEN Department of Anatomy and Neuroscience Training Program, University of Wisconsin Medical School, Madison, Wisconsin 53706
ABSTRACT The appearance of somatostatin (SOMI-immunoreactive (IR) and vasoactive intestinal peptide (VIP)-IR neurons in different regions of the embryonic chicken gut was studied by immunostaining wholemounts. The patterns of expression of these peptides in myenteric neurons showed a number of similarities. Both peptides first appeared in the region of the proventriculus-gizzard: SOM at embryonic day (E)4, VIP at E5.5. At later times both peptides were found in positions both rostral and caudal to the gizzard. Both peptides appeared independently in cells at a second site, the cecum of the hindgut: SOM was observed at E6.5 and VIP at E7.5. VIP-IR and SOM-IR cells appear throughout the cecum, then in the rectum, and finally in the ileum. Differences in the patterns of expression were also found. SOM- and VIP-IR neurons appeared at different times along the length of the gut. VIP-IR cells populated the entire gut by E11.5, whereas SOM-IR cells were not present throughout the gut until E13.5. SOM-IR cells appeared in the terminal part of the ganglion of Remak at E4.0. At E6 these SOM-IR cells sent fibers into the wall of the hindgut and later into the midgut. No VIP-IR cells were found in the ganglion of Remak. These findings suggest that neural crest-derived cells first express SOM- and VIP-IR in particular regions of the gut, namely, the proventriculus-gizzard and the cecum. Certain conditions must exist at these sites which favor the expression of these neuropeptides by neural crest-derived cells. The observation of SOM- and VIP-IR cells in the cecum at a stage of development before cells are seen in the ileum supports the concept that sacral neural crest cells contribute precursors for enteric neurons of the avian hindgut. Key words: enteric nervous system, neural crest, neuropeptides, immunocytochemistry, differentiation, ganglion cell, ganglion of Remak
The intrinsic neurons of the gut produce a variety of neurotransmitters and neuromodulators, including acetylcholine, gamma-aminobutyric acid, and a number of neuropeptides (Costa et al., '87). Some of these molecules have been localized to distinct sets of neurons in the guinea pig ileum. For instance, vasoactive intestinal peptide (VIP) is thought to be in myenteric inhibitory motor neurons that project to the muscle and in submucosal secretomotor neurons that project to epithelial cells in the mucosa (Furness and Costa, '87). The VIP-containing submucosal neurons also contain dynorphin and galanin (Bornstein and Furness, '88). A diverse array of neurotransmitters is found within any enteric ganglion and furthermore, a given transmitter may co-exist in the same neuron with a number of other transmitters. The question of interest is how precursor cells generate this variety of transmitter phenotypes within the enteric ganglia. The precursors of enteric neurons are derived from neural crest cells (Yntema and Hammond, '54; Le Douarin
o 1991 WILEY-LISS, INC.
and Teillet, '73). Although cells could be preprogrammed, a large body of evidence indicates that the transmitter phenotype of neurons derived from crest cells is dependent on the environment where they ultimately reside (Le Douarin, '82). The influence of the tissue environment could be exerted by cell-cell interactions, the extracellular matrix, or factors elaborated by the surrounding tissues or the crest cells themselves. The avian gut receives neural crest cells from two sources. Crest cells from the rhombencephalon (vagal crest) appear to enter the pharynx at embryonic day (E)2.5 and to reach the gizzard at E3.5-4.0 (Tucker et al., '86). These vagal crest cells form a complex network within the wall of the gut. Aggregates of cells develop within the cellular network, probably as a result of proliferation, and give rise Accepted May 13,1991. Address reprint requests to Miles Epstein, Dept. of Anatomy, UWMadison, 1300 University Ave., Madison, WI 53706.
NEUROPEPTIDE DEVELOPMENT ALONG GUT
169
mounts were examined and photographed on Technical Pan film (Kodak, Rochester, NY) with a Leitz Orthoplan microscope. Some immunostained tissues were embedded in plastic (Epon, Polyscience, Warrington, PA) and sectioned at 2 pm. Sections were counterstained with toluidine blue. Antiserum to VIP was produced as follows: porcine VIP (3 mg) was conjugated to bovine thyroglobulin (15 mg) with l-ethyl-3-(3-dimethylaminopropyl)-carbodimide (EDAC, 3 mg). After lyophilization, the conjugate was reconstituted in distilled H,O and injected intradermally into rabbits. Each rabbit received an initial injection of 300 pg of conjugate in Freund's complete adjuvant, and booster injections of 300 pg given at monthly intervals in incomplete adjuvant. The cross-reactivity of the antisera was tested by the method of Larsson ('81). The antisera used showed no cross-reactivity to any of the following: secretin, peptide histidine isoleucine (PHI), gastric inhibitory peptide, rat pancreatic polypeptide, motilin, somatostatin, glucagon, insulin, histamine, ACTH, gastrin 34, FMRF-amide, and serotonin (Mary Frick, personal communication). For control studies, this antiserum was preabsorbed with either 2 pg porcine VIP or 1 pg porcine PHI per ml of diluted antiserum and applied to wholemounts at different stages of development. No staining was observed after absorption with VIP; staining was intact after absorption with PHI. A monoclonal antibody against somatostatin was produced by Dr. A. Buchan by injecting mice with a conjugate of keyhole limpet hemocyanin and synthetic cyclic somaMETHODS tostatin 15-28 (antibody s-8, Buchan et al., '85). This Fertilized Leghorn chicken (Gallus gallus L ) eggs were antibody recognizes all three forms of somatostatin (SOM incubated at 38" +- 0.2"C for various times in a forced air 28, 14 cyclic, 14 linear), but does not recognize gastrin 17, incubator. Embryos incubated for 3-17.5 days (E3-17.5) motilin, or gastric inhibitory peptide. For control studies were staged (Hamburger and Hamilton, '51) and whole this antibody was preabsorbed with 1 pg of SOM 14 linear embryos or isolated gastrointestinal tracts were fixed in 4% per ml of diluted antibody and incubated with gut preparaparaformaldehyde containing 15% saturated picric acid in tions from E8 and E l 5 embryos; no staining was observed. 0.1 M phosphate buffer, pH 7.4, for 6-24 hours at 4°C. Pieces of gut were also removed and fixed from newly hatched chicks. After fixation, tissue was washed repeatRESULTS edly in PBS and stored in PBS containing 0.1% sodium Because the antisera used in this study may recognize azide. The entire gastrointestinal tracts from young embryos unknown peptides or proteins with amino acid sequences (E3-10.5) were incubated in 3% H,O,, washed, and incu- similar to the antigens used to generate the antisera, it is bated either in antiserum to vasoactive intestinal peptide appropriate to describe the immunoreactive material as (VIP, rabbit 2261, 1/1000 dilution) or a monoclonal anti- SOM- and VIP-like. For brevity we use the terms SOM- and body to somatostatin (SOM, 10 pg/ml), both diluted in PBS VIP-immunoreactive(1R).We are also unable to distinguish containing 0.3% Triton X-100, 5% goat serum, and 0.1% whether the immunoreactive material is the peptide precursodium azide for periods ranging from 2-4 days. For sor or the peptide itself. The progressive appearance of SOM-IR cells in myenteric embryos E11.5 to 15.5, the combined proventriculus and gizzard were separated from the combined small and large ganglia along the length of the gut over the course of chick intestine (duodenum to rectum); all were incubated in PBS embryonic development is depicted in Figure 1. SOM-IR with 1%Triton X-100, goat serum, and sodium azide for 7 appears independently in neurons in the foregut and hinddays, and then in antiserum diluted in PBS with 0.3% gut as well as in a ganglionated nerve, the ganglion of Triton X-100 for 7 days. The mucosa was removed from Remak, which is associated with the hindgut. E17.5 gut prior to its incubation in 1%Triton X-100. The In the foregut, SOM-IR is first observed in cells located at entire gastrointestinal tract in at least three animals at two sites: mesenchyme of the gizzard primordium at E4 each age between E3-15.5. and pieces of gut from at least (stage 24), and in the developing pancreas (Fig. 2a). The three E17.5 and newly hatched preparations were observed. cells at these sites are solitary or exist as clusters of 2 or 3. After incubation in primary antisera, the wholemounts At this time the cell boundaries are not apparent, and the were washed in PBS, incubated in biotinylated goat anti- immunoreactivity appears concentrated at one pole of the rabbit IgG or goat anti-mouse IgG and avidin-peroxidase cell (Fig. 2b). At higher magnification the cytoplasm of the (Vector Labs, Burlingame, CA). Reaction product was cell can be visualized (Fig. 2c), and the presence of SOM-IR developed by exposure to the chromagen 3,3'-diaminobenzi- in discrete granule-like structures is apparent. At E5.5 the number of cells in the gizzard has increased, dine and H,O,. After immunostaining, wholemounts were mounted on slides and coverslipped in glycerol. Whole- and at E5.5-6.5 a few of the cells show short immunoreac-
to a pattern of ganglia. During development the network is more complex and contains more cells at its cranial extent than its caudal pole (Epstein et al., '91). A second source of cells, the sacral crest, which populates the hindgut, was identified from quail-chick transplantation studies (Le Douarin and Teillet, '73). Recent studies have indicated that a portion of these crest cells enter the wall of the cloaca and ascend toward the midgut (Pomeranz and Gershon, '90) where they form neurons in the postumbilical gut (Pomeranz et al., '91). Although a number of studies of the development of neurotransmitters and neuropeptides in avian enteric neurons have been carried out, our understanding of the spatial and temporal development of transmitters along the length of the gut is incomplete. Saffrey et al. ('82) examined neuropeptides in preparations already midway through development. Epstein et al. ('83) confined their observations to the foregut. Despite these studies it is not clear whether neuropeptide expression begins at a particular site or occurs concurrently along the length of the gut. Our objective is to characterize the time course and location of appearance of the neuropeptides somatostatin and vasoactive intestinal peptide along the entire length of the developing gut to gain insight into the regulation of enteric neuron differentiation. By immunostaining whole mounts ofembryonic gut, we are able to follow the course of neuropeptide expression.
M.L. EPSTEIN AND K.T. POULSEN
170
SOM 4.0
6.5
8.5
10.5
11.5
13.5
8.5
10.5
11.5
VIP 5.5
6.5
7.5
Fig. 1. A schematic representation of the progressive expression of somatostatin-immunoreactive (SOM-IR) and vasoactive intestinal peptide immunoreactive (VIP-IR) cells in myenteric ganglia along the gut over the course of development. CR, crop; E, esophagus; P, proventriculus; G, gizzard; D, duodenum; B, bile duct; Y, yolk stalk C, cecum; R, rectum; RG, ganglion of Remak; circles represent cell bodies. Number indicates embryonic age.
NEUROPEPTIDE DEVELOPMENT ALONG GUT
171
Fig. 2. A. Montage showing the gut of an E4 embryo. Clusters of SOM-IR cells (arrows) are found in the mesenchyme around the gizzard primordium ( G )and in the developing pancreas (P). SOM-IR cells are also found in the ganglion of Remak (arrowhead). B. Micrograph showing the lower portion of the gizzard primordium at high magnification. Cell boundaries are not distinct but SOM-IR (arrows) is found as
dense reaction product adjacent to the nucleus. C. Photomicrograph showing SOM-IR in a plastic section stained with toluidine blue from an E6 gizzard. The reaction product appears as dense granule-like structures in the cytoplasm (arrow). The nuclei containing darkly stained nucleoli are clearly shown with the toludine blue. Bars = 200 pm for A, 10 pm for B, C.
tive processes. At this time SOM-IR is found around the circumferenceof the gizzard and has expanded into the very proximal portion of the duodenum and the isthmus region between the gizzard and proventriculus (Fig. 1).At E7.5 and E8.5 SOM-IR cells occupy a greater area in the gizzard (Fig. 3A,B), and cells are now prominent in regions both proximal and distal to the gizzard; cells are found in the proventriculus (Fig. 3C,D) and distal esophagus (Fig. 3E), and along the descending duodenum (Fig. 3F). However, at E9.5 the number and staining intensity of SOM-IR cells decreased in the proventriculus. By E10.5, this decrease is more apparent, such that only a few SOM-1R cells are found in the esophagus or proventriculus; cells remain in the gizzard and duodenum (Fig. 11, although their staining intensity appears reduced. At E13.5 the number of SOM-IR cells has increased in the proventriculus and esophagus compared to earlier stages. At E10.5 SOM-IR cells extend distally only to the duodenum although a fiber-like network extends a variable
length down the small intestine. This network ramifies in the regions of the myenteric and submucosal ganglia and circular muscle. By E11.5 a small number of cell bodies are found along the small intestine beyond the yolk stalk, and at E12.5 they extend to the level of the distal tip of the cecum. In the remainder of the small intestine, small discrete bits of immunoreactivity are found. These bits are probably SOM-IR within cells but lack sufficient size to fill the cytoplasm and give the classic appearance of immunoreactive cells. At E13.5 a small number of cells are found along the entire small intestine in 2/3 of the preparations, and at E15.5 numerous cells are found along the entire small intestine. No change in distribution occurred at the time of hatching. No SOM immunostaining was found in the vagus of E6 or newly hatched preparations. In the hindgut, SOM-IR cells first appear in the ganglion of Remak at E4 (Fig. 4A). At E5.5 a small number of cells are found along the entire length of the ganglion from the rectum to the duodenum (Figs. 1, 4B,C). At E6.5, SOM-IR
Fig. 3. Photomicrographs of wholemounts of regions of a n E8.5 gut. A. Cords of SOM-IR cells (arrow) are found in the gizzard. B. SOM-IR cells (arrow) in the gizzard are shown at higher magnification. C. Photomicrograph shows SOM-IR cells (arrow) in myenteric ganglia throughout the proventriculus. D. Portion of C at high magnification
shows a cluster of SOM-IR cells (arrow) in a myenteric ganglion. E. SOM-IR cells (arrow) are found throughout the esophagus. F. The duodenum contains SOM-IR cells (arrow). Bars = 200 km for A, C, E, F, 10 pm for B, D.
NEUROPEPTIDE DEVELOPMENT ALONG GUT
173
Fig. 4. A. High magnification micrograph of a portion of Figure 2A shows SOM-IR cells (arrow) in the ganglion of Remak. B. Wholemount of an E5.5 hindgut shows SOM-IR as faint black dots along the length of the ganglion of Remak. Arrow shows a typical neuron. C. Portion of B at higher magnification shows a SOM-IR neuron in the ganglion of Remak. D. Wholemount of an E6.5 hindgut shows SOM-IR processes
(arrow) running from the ganglion of Remak (RG) to the wall of the proximal rectum. SOM-IR cells (arrowhead) are also found in the cecum. E. Portion of the cecum in D at higher magnification shows SOM-IR cells (arrow). Bars = 10 pm for A, C, E, 200 pm for B, 100 pm for D.
fibers extend from the ganglion into the wall of the hindgut (Fig. 4D); they are found in greatest numbers at the cecal-rectaljunction and in smaller numbers at the terminal rectum. At this time, SOM-IR cells are found for the first time in the proximal cecum of the hindgut (Fig. 4D,E). At E7.5 the number of SOM-IR cells in the cecum and the number of fibers from the ganglion of Remak projecting into the rectum have increased. At E8.5 SOM-IR fibers leave the ganglion along its entire length and enter the wall of the intestine from the rectum to the duodenum (Fig. 5A) but do not project to the ceca. At this time large branches of the ganglion overlay the serosa (Fig. 5B) and could serve as sources for neurons in the rectum. SOM-IR cells first
appear in the rectum at E11.5, and at E17.5 are found in a large number of ganglia in the rectum. In contrast, few cell bodies but intense fiber-staining is found in the ceca. At the time of hatching, SOM-IR cells are found in large numbers in the ganglion of Rernak and in ganglia within the gizzard, rectum, ileum, and jejunum. The progressive appearance of VIP-IR cells in myenteric ganglia along the developing gut is schematically summarized in Figure 1. VIP-IR cells are first localized to the junction of the proventriculus and gizzard at E5.5 (Fig. 6A,B). The boundaries of the cell bodies are clearly distinguished; the immunostain is evenly distributed throughout the cytoplasm in contrast to the particulate staining seen in
174
M.L. EPSTEIN AND K.T. POULSEN
Fig. 5. A. Wholemount of an E8.5 rectum. The ganglion of Remak (RG) contains immunostained neurons which project SOM-IR processes (arrow) into the wall of the rectum. B. Portion of A at higher magnification. A segment of the ganglion of Remak containing SOM-IR
cells (arrow) extends onto the gut wall. The fate of these extensions is unknown, but it is possible that cell bodies may move into the gut wall from these sites. Bars = 200 km for A, 10 km for B.
SOM-IR cells in the gizzard. Many of the cells contain immunoreactive processes that form a nascent network in the proventriculus. VIP-IR cells subsequently appear both proximal and distal to the proventriculus-gizzard junction; by E6.5 cells are found both in the middle of the proventriculus and in the most proximal portions of the gizzard and duodenum (Fig. 1).At E7.5, VIP-IR cells are seen in the distal esophagus and the most proximal jejunum. Also at this time, isolated cells appear in the hindgut, in the middle of the ceca (Figs. 1,6,C,D). At E8.5, VIP-IR cells and processes are found proximal to the level of the crop and distal to a site beyond the yolk stalk. The length of gut with VIP expression has expanded rapidly compared to SOM. Extensive differences in the maturity of the network of VIP-IR cells and fibers are found along the length of the gut as shown in Figure 7. A large number of cells are found in the gizzard (Fig. 7A,B); the VIP-IR is particulate, as was the SOM-IR. In the proventriculus (Fig. 7C,D), an extensive network of cells and fibers is found, whereas only cells are seen in the proximal esophagus (Fig. 7E). The VIP-IR in the proventriculus and esophagus is evenly distributed in the cell bodies. The descending duodenum (Fig. 7F) contains an extensive array of connectives but few cell bodies. At the level of the yolk stalk (Fig. 8A,B), the network is not prominent, and the number of fibers in the connectives appears to be less than
in the duodenum. In the hindgut, a small number of cells and an extensive fiber array fill the ceca. A small number of cells are also found in the terminal rectum; the middle portion of the rectum shows no immunostaining (Fig. 8C,D). By E9.5, isolated VIP-IR cells are seen in the ileum distal to the yolk stalk. The middle of the rectum still shows an absence of immunostain. At E10.5, the entire gut is stained except for a small region of distal ileum stretching from the tip of the ceca to the ileocecal junction. In most cases VIP-IR fibers and cells do not extend along the gut from the rectum into the ileum. In a few cases VIP-IR fibers and cells are found in the most distal length but are absent in more proximal segments of ileum. This unstained length probably represents the site where the vagal and sacral crest-derived cells intersect. By E11.5 cells and fibers are found along the entire gut length, and the distribution is unchanged through hatching. At no time were VIP-IR cells found in the ganglion of Remak.
DISCUSSION We have examined the time course and pattern of expression of two neuropeptides along the gut in order to gain insight into the mechanisms responsible for the phenotypic diversity found in enteric ganglia. Our observations rely on the detection of immunoreactivity, and our results may
NEUROPEPTIDE DEVELOPMENT ALONG GUT
175
Fig. 6. A. Network of VIP-IR cells and processes is seen in a whole mount of an E6.5 proventriculus. B. Portion of A at high magnification. Micrograph shows VIP-IR in cytoplasm and processes of myenteric neurons (arrow). C. VIP-IR cells (arrow) are seen as small dots over the
length of a n E7.5 cecum. D. Portion of C at high magnification. Micrograph shows VIP-IR cells (arrow) in the cecum. Bars = 200 p m for A, C, 10 pm for B, D.
reflect the sensitivity of our antisera rather than differences in expression. Because of the variability in the appearance of the SOM-IR in the midgut, we are uncertain of the exact time of appearance of SOM-IR cells. Verification of the reported differences will require either specific radioimmunoassays in conjunction with high pressure liquid chromatography and/or hybridization with specific nucleotide probes. The cells expressing these peptides are derived from the vagal neural crest which enter the pharynx at E2.5 (Tucker et al., '86) and approach the ileocecal junction by E5.5 (Epstein et al., '91); these precursors give rise to aggregates of cells, which are the forerunners of the enteric ganglia. Within these aggregates are cells which will differentiate into neurons with different transmitter phenotypes and into glial cells. The signals that regulate the expression of these phenotypes are unknown, although humoral factors
have been isolated that support peptide expression in cultures of sympathetic neurons (Nawa and Patterson, '90; Nawa and Sah, '90). We see similarities in the development of SOM and VIP. Both peptides first appear in the region of the proventriculus-gizzard. SOM is present in the region of the gizzard primordium at E4.0 andVIP at the bottom of the proventriculus at E5.5. It is interesting that the gizzard is the site where these peptides are first detected. Our expectation was that expression would parallel the time course of the movement of the crest-derived cells down the gut, i.e., expression would be seen first in the esophagus, followed by the gizzard, duodenum, and midgut. However, that is not the case, and the fact that expression occurs first in the gizzard suggests this organ can promote differentiation. One possibility is that the large amount of mesenchyme associated with the gizzard primordium produces factors
176
M.L. EPSTEIN AND K.T. POULSEN
Figure I
NEUROPEPTIDE DEVELOPMENT ALONG GUT
177
Fig. 8. A. Micrograph shows a few VIP-IR cells and processes in the ES.5 gut at the level of the yolk stalk. B. Portion of A at high magnification shows a VIP-IR cell (arrow). C. Low-power micrograph shows VIP-IR (arrow) in the terminal portion of an E8.5 rectum. Compared to the terminal region, the amount of immunostain is
decreased in the middle of the rectum, shown in the left portion of the micrograph. D. Portion of C at high magnification shows a VIP-IR neuron (arrow) in the terminal rectum. Bars = 200 km for A, C, 10 ym forB,D.
that activate neuropeptide expression in the neuronal precursors. Another possibility is that the mesenchyme supports a large number of neuronal precursors and cellcell interactions between precursor cells trigger neuropeptide expression. Both VIP- and SOM-IR appear in cells in the hindgut prior to their appearance in the midgut.
Isolated SOM- and VIP-IR cells are found in the proximal ceca at E6.5 and E7.5, respectively. These cells presumably arise from precursors derived from sacral neural crest, which arise posterior to somite 28. In recent studies Pomeranz and Gershon ('90) have described two populations of sacral crest cells which enter the gut. One group remained in the dorsal mesentery to form the ganglion of Remak, whereas the other invaded the hindgut and ascended toward the midgut. Our data suggest that the latter group may provide the precursors for the SOM and VIP-IR cells we describe in the ceca. Although vagal precursors are present in the terminal ileum at E5.5 (Epstein et al., '911, expression of these peptides is delayed at least 5 days in this gut region. It is curious that expression occurs so early in the cecum and so late in the ileum, a contiguous area. It is of interest that in both the hindgut and foregut, neuropeptide
Fig. 7. Micrographs showing regions of E8.5 gut. A. VIP-IR cells are found in large numbers in the gizzard. B. Portion of A at high magnification. Cells (arrow) are found in myenteric ganglia. C. Ganglia in the proventriculus contain numerous VIP-IR cells. D. Portion of C at high magnification. VIP-IR neurons (arrow) are shown in myenteric ganglia. E. VIP-IR cells are found in the esophagus (arrow). Inset: VIP-IR cells are shown at high magnification. F. The descending duodenum contains prominent VIP-IR connectives (arrow) but few cell bodies. Bars = 200 ym for A, C, D, F, 10 pm for B, D, E inset.
178
M.L. EPSTEIN AND K.T. POULSEN
expression should first occur at a location which is distant findings suggest that these sites provide environmental from the site of precursor immigration into the gut. Our cues which lead to expression of these neuropeptides. observations of the ceca as well as the gizzard suggests a Additional work will be necessary to uncover these cues. regional heterogeneity in the ability of the gut to initiate neuropeptide expression. In the foregut the expression of these peptides proceeds ACKNOWLEDGMENTS in both oral and anal directions from the site of initial We are grateful to Dr. Alison Buchan for providing the appearance despite the fact that the neural crest precursors antibody to somatostatin and to S. Presley, J. Dahl, K. advance and form aggregates in an oral-to-and direction. Expression moving in an oral direction is seen in the Biefeldt, E. Schultz for reading the manuscript. Supported esophagus where the precursors produce both neuropep- by NSF grant BNS-8820658. tides at E8.5, but neither at E6.5. It is not clear if expression at E8.5 in the esophagus is related to the LITERATURE CITED maturation of the neuronal precursors or mesenchyme or both. Bornstein, J.C., and J.B. Furness (1988) Correlated electrophysiologicaland Differences in the expression of SOM and VIP were also histochemical studies of submucous neurons and their contribution to understanding enteric neural circuits. J. Auto. Nervous System 25:l-13. apparent. VIP expression progresses more rapidly through the midgut and hindgut than SOM. In the midgut at E8.5, Buchan, A.M.J.,L.K.J. Sikora, J.G. Levy, C.H.S. McIntosh, I. Dyck, and J.C. Brown (1985) An immunocytochemical investigation with monoclonal VIP is found distal to the yolk stalk, whereas SOM is only in antibodies to somatostatin. Histochemistry 83: 175-180. the descending duodenum. Although SOM-IR cells were Cavanaugh, M.E., and J.G. Parnavelas (1988) Development of somatostatin visible in the ceca of the hindgut at E6.5, cells were not immunoreactive neurons in the rat occipitalcortex: A combined immunocytochemical-autoradiographicstudy. J. Comp. Neurol. 268;l-12. observed in the rectum until E11.5. Hence, there is a large delay between the initial appearance of SOM-IR cells in ceca Costa, M., J.B. Furness, and I.J. Llewellyn-Smith (1987) Histochemistry of the enteric nervous system. In L.R. Johnson (ed): Physiology of the and their appearance in the rectum. In contrast, the VIP-IR Gastrointestinal Tract, 2nd ed. New York: Raven Press. cells populated the entire ceca at E8.5 and the entire rectum Epstein, M.L., D. Sherman, and M.D. Gershon (1980) Development of by E10.5. serotonergic neurons in the chick duodenum. Dev. Biol. 77:2240. It is not clear why VIP is expressed earlier than SOM Epstein, M.L., J. Hudis, and J.L. Dahl(1983)The development of peptidergic along the small intestine and in the hindgut. One possibility neurons in the foregut of the chick. J. Neurosci. 33431-2447. is that SOM expression in myenteric neurons is retarded by Epstein, M.L., K.T. Poulsen, and R. Thiboldeaux (1991) Formation of ganglia in the gut of the chick embryo. J. Comp. Neurol307:l-ll. the presence of SOM-IR extrinsic fibers. The small intestine contains SOM-IR extrinsic fibers, which arise from neurons Furness, J.B., and M. Costa (1987) The Enteric Nervous System. Edinburgh: Churchill Livingstone. in the ganglion of Remak. Extrinsic fibers are not found in the ceca where the SOM first appears in the hindgut. Garcia-Arraras, J.E., M. Chanconie, and J. Fontaine-Perus (1984) In vivo and in vitro development of somatostatin-like immunoreactivity in the VIP-IR progresses rapidly down the small intestine, which peripheral nervous system of quail embryos. J. Neurosci. 4r1549-1558. contains no major extrinsic VIP-IR fibers at E8-13. VIP-IR Hamburger, V., and H.L. Hamilton (1951) A series of normal stages in the cells are not found in the ganglion of Remak. The expresdevelopment of the chick embryo. J. Embryol. Exp. Morphol. 88:49-92. sion of SOM and VIP progresses at similar rates up the Larsson, L.-I. (1981) A novel immunocytochemical model system for specificity and sensitivity screening of antisera against multiple antigens. J. foregut between E6.5 and E8.5. The extrinsic fibers to the Histochem. Cytocbem. 29:408410. foregut are the vagus and sympathetic nerves; we did not observe VIP or SOM-IR in the vagus. Sympathetic fibers, Le Douarin, N.M. (1982) The Neural Crest. Cambridge: Cambridge University Press. which are known to contain SOM (Maxwell et al., '84), do Le Douarin, N.M., and M.-A. Teillet (1973) The migration of neural crest not reach the gut until later in development (Epstein et al., cells to the wall of the digestive tract in avian embryo. J. Embryol. Exp. '80). Further experiments will be necessary to determine if Morphol. 30:3148. the extrinsic fibers or other local environmental cues Maxwell, G.D., P.D. Sietz, and P.H. Chenard (1984) Development of influence expression. somatostatin-like immunoreactivity in embryonic sympathetic ganglia. J. Neurosci. 4:576-584. Others have found SOM early in the development of the peripheral nervous system. SOM-IR appears in the develop- Nawa, Hiroyuki, and P.H. Patterson (1990) Separation and partial characterization of neuropeptide-inducing factors in heart cell conditioned meing sympathetic chains of quail embryos at E4 (Garciadium. Neuron 4.269-27'7. Arraras et al., '84; Maxwell et al., '84). However, on the Nawa, Hiroyuki, and D.W.Y. Sah (1990) Different biological activities in basis of immunostaining and radioimmunoassay, the conditioned media control the expression of a variety of neuropeptides in cultured sympathetic neurons. Neuron 4.279-287. SOM-IR appears to decrease and reach a minimum value at about E10. We also observed a loss of immunostaining in Pomeranz, H.D., and M.D. Gershon (1990) Colonization of the avian hindgut by cells derived from the sacral neural crest. Dev. Biol. 137:378-394. the proventriculus and esophagus at about E10. A decrease in the number of SOM-IR neurons has also been observed Pomeranz, H.D., T.P. Rothman, and M.D. Gershon (1991) Colonization of the post-umbilical bowel by cells derived from the sacral neural crest: during the development of the rat occipital cortex; this Direct tracing of cell migration using an intercalating probe and a decrease resulted from cell death (Cavanaugh and Parnareplication-deficient retrovirus. Development 11I:647-655. velas, '88). However, unlike the sympathetic chains and rat Saffrey, M.J., J.M. Pol&, and G. Burnstock (1982) Distribution of vasoactive intestinal polypeptide-, substance P-, enkephalin-, and neurotensin-like cortex, SOM-IR in the gut is restored. We do not know if the immunoreactive nerves in the chicken gut during development. Neurodecrease in the gut is the result of cell death or a change in sci. 7:279-293. expression. G.C., G. Ciment, and J.-P. Thiery (1986) Pathways of avian neural In conclusion, our observations indicate that the expres- Tucker, crest cell migration in the developing gut. Dev. Biol. 116r439-450. sion of the neuropeptides SOM and VIP is initiated at two Yntema, C.L., and W.S. Hammond (1954) The origin of intrinsic ganglia of discrete and discontinuous locations, the proventriculustrunk viscera from vagal neural crest in the chick embryo. J. Comp. Neurol. 101:515-541. gizzard of the foregut and the ceca of the hindgut. These
Appearance of Somatostatin and Vasoactive Intestinal Peptide Along the Developing Chicken Gut MILES L. EPSTEIN AND KRISTIAN T. POULSEN Department of Anatomy and Neuroscience Training Program, University of Wisconsin Medical School, Madison, Wisconsin 53706
ABSTRACT The appearance of somatostatin (SOMI-immunoreactive (IR) and vasoactive intestinal peptide (VIP)-IR neurons in different regions of the embryonic chicken gut was studied by immunostaining wholemounts. The patterns of expression of these peptides in myenteric neurons showed a number of similarities. Both peptides first appeared in the region of the proventriculus-gizzard: SOM at embryonic day (E)4, VIP at E5.5. At later times both peptides were found in positions both rostral and caudal to the gizzard. Both peptides appeared independently in cells at a second site, the cecum of the hindgut: SOM was observed at E6.5 and VIP at E7.5. VIP-IR and SOM-IR cells appear throughout the cecum, then in the rectum, and finally in the ileum. Differences in the patterns of expression were also found. SOM- and VIP-IR neurons appeared at different times along the length of the gut. VIP-IR cells populated the entire gut by E11.5, whereas SOM-IR cells were not present throughout the gut until E13.5. SOM-IR cells appeared in the terminal part of the ganglion of Remak at E4.0. At E6 these SOM-IR cells sent fibers into the wall of the hindgut and later into the midgut. No VIP-IR cells were found in the ganglion of Remak. These findings suggest that neural crest-derived cells first express SOM- and VIP-IR in particular regions of the gut, namely, the proventriculus-gizzard and the cecum. Certain conditions must exist at these sites which favor the expression of these neuropeptides by neural crest-derived cells. The observation of SOM- and VIP-IR cells in the cecum at a stage of development before cells are seen in the ileum supports the concept that sacral neural crest cells contribute precursors for enteric neurons of the avian hindgut. Key words: enteric nervous system, neural crest, neuropeptides, immunocytochemistry, differentiation, ganglion cell, ganglion of Remak
The intrinsic neurons of the gut produce a variety of neurotransmitters and neuromodulators, including acetylcholine, gamma-aminobutyric acid, and a number of neuropeptides (Costa et al., '87). Some of these molecules have been localized to distinct sets of neurons in the guinea pig ileum. For instance, vasoactive intestinal peptide (VIP) is thought to be in myenteric inhibitory motor neurons that project to the muscle and in submucosal secretomotor neurons that project to epithelial cells in the mucosa (Furness and Costa, '87). The VIP-containing submucosal neurons also contain dynorphin and galanin (Bornstein and Furness, '88). A diverse array of neurotransmitters is found within any enteric ganglion and furthermore, a given transmitter may co-exist in the same neuron with a number of other transmitters. The question of interest is how precursor cells generate this variety of transmitter phenotypes within the enteric ganglia. The precursors of enteric neurons are derived from neural crest cells (Yntema and Hammond, '54; Le Douarin
o 1991 WILEY-LISS, INC.
and Teillet, '73). Although cells could be preprogrammed, a large body of evidence indicates that the transmitter phenotype of neurons derived from crest cells is dependent on the environment where they ultimately reside (Le Douarin, '82). The influence of the tissue environment could be exerted by cell-cell interactions, the extracellular matrix, or factors elaborated by the surrounding tissues or the crest cells themselves. The avian gut receives neural crest cells from two sources. Crest cells from the rhombencephalon (vagal crest) appear to enter the pharynx at embryonic day (E)2.5 and to reach the gizzard at E3.5-4.0 (Tucker et al., '86). These vagal crest cells form a complex network within the wall of the gut. Aggregates of cells develop within the cellular network, probably as a result of proliferation, and give rise Accepted May 13,1991. Address reprint requests to Miles Epstein, Dept. of Anatomy, UWMadison, 1300 University Ave., Madison, WI 53706.
NEUROPEPTIDE DEVELOPMENT ALONG GUT
169
mounts were examined and photographed on Technical Pan film (Kodak, Rochester, NY) with a Leitz Orthoplan microscope. Some immunostained tissues were embedded in plastic (Epon, Polyscience, Warrington, PA) and sectioned at 2 pm. Sections were counterstained with toluidine blue. Antiserum to VIP was produced as follows: porcine VIP (3 mg) was conjugated to bovine thyroglobulin (15 mg) with l-ethyl-3-(3-dimethylaminopropyl)-carbodimide (EDAC, 3 mg). After lyophilization, the conjugate was reconstituted in distilled H,O and injected intradermally into rabbits. Each rabbit received an initial injection of 300 pg of conjugate in Freund's complete adjuvant, and booster injections of 300 pg given at monthly intervals in incomplete adjuvant. The cross-reactivity of the antisera was tested by the method of Larsson ('81). The antisera used showed no cross-reactivity to any of the following: secretin, peptide histidine isoleucine (PHI), gastric inhibitory peptide, rat pancreatic polypeptide, motilin, somatostatin, glucagon, insulin, histamine, ACTH, gastrin 34, FMRF-amide, and serotonin (Mary Frick, personal communication). For control studies, this antiserum was preabsorbed with either 2 pg porcine VIP or 1 pg porcine PHI per ml of diluted antiserum and applied to wholemounts at different stages of development. No staining was observed after absorption with VIP; staining was intact after absorption with PHI. A monoclonal antibody against somatostatin was produced by Dr. A. Buchan by injecting mice with a conjugate of keyhole limpet hemocyanin and synthetic cyclic somaMETHODS tostatin 15-28 (antibody s-8, Buchan et al., '85). This Fertilized Leghorn chicken (Gallus gallus L ) eggs were antibody recognizes all three forms of somatostatin (SOM incubated at 38" +- 0.2"C for various times in a forced air 28, 14 cyclic, 14 linear), but does not recognize gastrin 17, incubator. Embryos incubated for 3-17.5 days (E3-17.5) motilin, or gastric inhibitory peptide. For control studies were staged (Hamburger and Hamilton, '51) and whole this antibody was preabsorbed with 1 pg of SOM 14 linear embryos or isolated gastrointestinal tracts were fixed in 4% per ml of diluted antibody and incubated with gut preparaparaformaldehyde containing 15% saturated picric acid in tions from E8 and E l 5 embryos; no staining was observed. 0.1 M phosphate buffer, pH 7.4, for 6-24 hours at 4°C. Pieces of gut were also removed and fixed from newly hatched chicks. After fixation, tissue was washed repeatRESULTS edly in PBS and stored in PBS containing 0.1% sodium Because the antisera used in this study may recognize azide. The entire gastrointestinal tracts from young embryos unknown peptides or proteins with amino acid sequences (E3-10.5) were incubated in 3% H,O,, washed, and incu- similar to the antigens used to generate the antisera, it is bated either in antiserum to vasoactive intestinal peptide appropriate to describe the immunoreactive material as (VIP, rabbit 2261, 1/1000 dilution) or a monoclonal anti- SOM- and VIP-like. For brevity we use the terms SOM- and body to somatostatin (SOM, 10 pg/ml), both diluted in PBS VIP-immunoreactive(1R).We are also unable to distinguish containing 0.3% Triton X-100, 5% goat serum, and 0.1% whether the immunoreactive material is the peptide precursodium azide for periods ranging from 2-4 days. For sor or the peptide itself. The progressive appearance of SOM-IR cells in myenteric embryos E11.5 to 15.5, the combined proventriculus and gizzard were separated from the combined small and large ganglia along the length of the gut over the course of chick intestine (duodenum to rectum); all were incubated in PBS embryonic development is depicted in Figure 1. SOM-IR with 1%Triton X-100, goat serum, and sodium azide for 7 appears independently in neurons in the foregut and hinddays, and then in antiserum diluted in PBS with 0.3% gut as well as in a ganglionated nerve, the ganglion of Triton X-100 for 7 days. The mucosa was removed from Remak, which is associated with the hindgut. E17.5 gut prior to its incubation in 1%Triton X-100. The In the foregut, SOM-IR is first observed in cells located at entire gastrointestinal tract in at least three animals at two sites: mesenchyme of the gizzard primordium at E4 each age between E3-15.5. and pieces of gut from at least (stage 24), and in the developing pancreas (Fig. 2a). The three E17.5 and newly hatched preparations were observed. cells at these sites are solitary or exist as clusters of 2 or 3. After incubation in primary antisera, the wholemounts At this time the cell boundaries are not apparent, and the were washed in PBS, incubated in biotinylated goat anti- immunoreactivity appears concentrated at one pole of the rabbit IgG or goat anti-mouse IgG and avidin-peroxidase cell (Fig. 2b). At higher magnification the cytoplasm of the (Vector Labs, Burlingame, CA). Reaction product was cell can be visualized (Fig. 2c), and the presence of SOM-IR developed by exposure to the chromagen 3,3'-diaminobenzi- in discrete granule-like structures is apparent. At E5.5 the number of cells in the gizzard has increased, dine and H,O,. After immunostaining, wholemounts were mounted on slides and coverslipped in glycerol. Whole- and at E5.5-6.5 a few of the cells show short immunoreac-
to a pattern of ganglia. During development the network is more complex and contains more cells at its cranial extent than its caudal pole (Epstein et al., '91). A second source of cells, the sacral crest, which populates the hindgut, was identified from quail-chick transplantation studies (Le Douarin and Teillet, '73). Recent studies have indicated that a portion of these crest cells enter the wall of the cloaca and ascend toward the midgut (Pomeranz and Gershon, '90) where they form neurons in the postumbilical gut (Pomeranz et al., '91). Although a number of studies of the development of neurotransmitters and neuropeptides in avian enteric neurons have been carried out, our understanding of the spatial and temporal development of transmitters along the length of the gut is incomplete. Saffrey et al. ('82) examined neuropeptides in preparations already midway through development. Epstein et al. ('83) confined their observations to the foregut. Despite these studies it is not clear whether neuropeptide expression begins at a particular site or occurs concurrently along the length of the gut. Our objective is to characterize the time course and location of appearance of the neuropeptides somatostatin and vasoactive intestinal peptide along the entire length of the developing gut to gain insight into the regulation of enteric neuron differentiation. By immunostaining whole mounts ofembryonic gut, we are able to follow the course of neuropeptide expression.
M.L. EPSTEIN AND K.T. POULSEN
170
SOM 4.0
6.5
8.5
10.5
11.5
13.5
8.5
10.5
11.5
VIP 5.5
6.5
7.5
Fig. 1. A schematic representation of the progressive expression of somatostatin-immunoreactive (SOM-IR) and vasoactive intestinal peptide immunoreactive (VIP-IR) cells in myenteric ganglia along the gut over the course of development. CR, crop; E, esophagus; P, proventriculus; G, gizzard; D, duodenum; B, bile duct; Y, yolk stalk C, cecum; R, rectum; RG, ganglion of Remak; circles represent cell bodies. Number indicates embryonic age.
NEUROPEPTIDE DEVELOPMENT ALONG GUT
171
Fig. 2. A. Montage showing the gut of an E4 embryo. Clusters of SOM-IR cells (arrows) are found in the mesenchyme around the gizzard primordium ( G )and in the developing pancreas (P). SOM-IR cells are also found in the ganglion of Remak (arrowhead). B. Micrograph showing the lower portion of the gizzard primordium at high magnification. Cell boundaries are not distinct but SOM-IR (arrows) is found as
dense reaction product adjacent to the nucleus. C. Photomicrograph showing SOM-IR in a plastic section stained with toluidine blue from an E6 gizzard. The reaction product appears as dense granule-like structures in the cytoplasm (arrow). The nuclei containing darkly stained nucleoli are clearly shown with the toludine blue. Bars = 200 pm for A, 10 pm for B, C.
tive processes. At this time SOM-IR is found around the circumferenceof the gizzard and has expanded into the very proximal portion of the duodenum and the isthmus region between the gizzard and proventriculus (Fig. 1).At E7.5 and E8.5 SOM-IR cells occupy a greater area in the gizzard (Fig. 3A,B), and cells are now prominent in regions both proximal and distal to the gizzard; cells are found in the proventriculus (Fig. 3C,D) and distal esophagus (Fig. 3E), and along the descending duodenum (Fig. 3F). However, at E9.5 the number and staining intensity of SOM-IR cells decreased in the proventriculus. By E10.5, this decrease is more apparent, such that only a few SOM-1R cells are found in the esophagus or proventriculus; cells remain in the gizzard and duodenum (Fig. 11, although their staining intensity appears reduced. At E13.5 the number of SOM-IR cells has increased in the proventriculus and esophagus compared to earlier stages. At E10.5 SOM-IR cells extend distally only to the duodenum although a fiber-like network extends a variable
length down the small intestine. This network ramifies in the regions of the myenteric and submucosal ganglia and circular muscle. By E11.5 a small number of cell bodies are found along the small intestine beyond the yolk stalk, and at E12.5 they extend to the level of the distal tip of the cecum. In the remainder of the small intestine, small discrete bits of immunoreactivity are found. These bits are probably SOM-IR within cells but lack sufficient size to fill the cytoplasm and give the classic appearance of immunoreactive cells. At E13.5 a small number of cells are found along the entire small intestine in 2/3 of the preparations, and at E15.5 numerous cells are found along the entire small intestine. No change in distribution occurred at the time of hatching. No SOM immunostaining was found in the vagus of E6 or newly hatched preparations. In the hindgut, SOM-IR cells first appear in the ganglion of Remak at E4 (Fig. 4A). At E5.5 a small number of cells are found along the entire length of the ganglion from the rectum to the duodenum (Figs. 1, 4B,C). At E6.5, SOM-IR
Fig. 3. Photomicrographs of wholemounts of regions of a n E8.5 gut. A. Cords of SOM-IR cells (arrow) are found in the gizzard. B. SOM-IR cells (arrow) in the gizzard are shown at higher magnification. C. Photomicrograph shows SOM-IR cells (arrow) in myenteric ganglia throughout the proventriculus. D. Portion of C at high magnification
shows a cluster of SOM-IR cells (arrow) in a myenteric ganglion. E. SOM-IR cells (arrow) are found throughout the esophagus. F. The duodenum contains SOM-IR cells (arrow). Bars = 200 km for A, C, E, F, 10 pm for B, D.
NEUROPEPTIDE DEVELOPMENT ALONG GUT
173
Fig. 4. A. High magnification micrograph of a portion of Figure 2A shows SOM-IR cells (arrow) in the ganglion of Remak. B. Wholemount of an E5.5 hindgut shows SOM-IR as faint black dots along the length of the ganglion of Remak. Arrow shows a typical neuron. C. Portion of B at higher magnification shows a SOM-IR neuron in the ganglion of Remak. D. Wholemount of an E6.5 hindgut shows SOM-IR processes
(arrow) running from the ganglion of Remak (RG) to the wall of the proximal rectum. SOM-IR cells (arrowhead) are also found in the cecum. E. Portion of the cecum in D at higher magnification shows SOM-IR cells (arrow). Bars = 10 pm for A, C, E, 200 pm for B, 100 pm for D.
fibers extend from the ganglion into the wall of the hindgut (Fig. 4D); they are found in greatest numbers at the cecal-rectaljunction and in smaller numbers at the terminal rectum. At this time, SOM-IR cells are found for the first time in the proximal cecum of the hindgut (Fig. 4D,E). At E7.5 the number of SOM-IR cells in the cecum and the number of fibers from the ganglion of Remak projecting into the rectum have increased. At E8.5 SOM-IR fibers leave the ganglion along its entire length and enter the wall of the intestine from the rectum to the duodenum (Fig. 5A) but do not project to the ceca. At this time large branches of the ganglion overlay the serosa (Fig. 5B) and could serve as sources for neurons in the rectum. SOM-IR cells first
appear in the rectum at E11.5, and at E17.5 are found in a large number of ganglia in the rectum. In contrast, few cell bodies but intense fiber-staining is found in the ceca. At the time of hatching, SOM-IR cells are found in large numbers in the ganglion of Rernak and in ganglia within the gizzard, rectum, ileum, and jejunum. The progressive appearance of VIP-IR cells in myenteric ganglia along the developing gut is schematically summarized in Figure 1. VIP-IR cells are first localized to the junction of the proventriculus and gizzard at E5.5 (Fig. 6A,B). The boundaries of the cell bodies are clearly distinguished; the immunostain is evenly distributed throughout the cytoplasm in contrast to the particulate staining seen in
174
M.L. EPSTEIN AND K.T. POULSEN
Fig. 5. A. Wholemount of an E8.5 rectum. The ganglion of Remak (RG) contains immunostained neurons which project SOM-IR processes (arrow) into the wall of the rectum. B. Portion of A at higher magnification. A segment of the ganglion of Remak containing SOM-IR
cells (arrow) extends onto the gut wall. The fate of these extensions is unknown, but it is possible that cell bodies may move into the gut wall from these sites. Bars = 200 km for A, 10 km for B.
SOM-IR cells in the gizzard. Many of the cells contain immunoreactive processes that form a nascent network in the proventriculus. VIP-IR cells subsequently appear both proximal and distal to the proventriculus-gizzard junction; by E6.5 cells are found both in the middle of the proventriculus and in the most proximal portions of the gizzard and duodenum (Fig. 1).At E7.5, VIP-IR cells are seen in the distal esophagus and the most proximal jejunum. Also at this time, isolated cells appear in the hindgut, in the middle of the ceca (Figs. 1,6,C,D). At E8.5, VIP-IR cells and processes are found proximal to the level of the crop and distal to a site beyond the yolk stalk. The length of gut with VIP expression has expanded rapidly compared to SOM. Extensive differences in the maturity of the network of VIP-IR cells and fibers are found along the length of the gut as shown in Figure 7. A large number of cells are found in the gizzard (Fig. 7A,B); the VIP-IR is particulate, as was the SOM-IR. In the proventriculus (Fig. 7C,D), an extensive network of cells and fibers is found, whereas only cells are seen in the proximal esophagus (Fig. 7E). The VIP-IR in the proventriculus and esophagus is evenly distributed in the cell bodies. The descending duodenum (Fig. 7F) contains an extensive array of connectives but few cell bodies. At the level of the yolk stalk (Fig. 8A,B), the network is not prominent, and the number of fibers in the connectives appears to be less than
in the duodenum. In the hindgut, a small number of cells and an extensive fiber array fill the ceca. A small number of cells are also found in the terminal rectum; the middle portion of the rectum shows no immunostaining (Fig. 8C,D). By E9.5, isolated VIP-IR cells are seen in the ileum distal to the yolk stalk. The middle of the rectum still shows an absence of immunostain. At E10.5, the entire gut is stained except for a small region of distal ileum stretching from the tip of the ceca to the ileocecal junction. In most cases VIP-IR fibers and cells do not extend along the gut from the rectum into the ileum. In a few cases VIP-IR fibers and cells are found in the most distal length but are absent in more proximal segments of ileum. This unstained length probably represents the site where the vagal and sacral crest-derived cells intersect. By E11.5 cells and fibers are found along the entire gut length, and the distribution is unchanged through hatching. At no time were VIP-IR cells found in the ganglion of Remak.
DISCUSSION We have examined the time course and pattern of expression of two neuropeptides along the gut in order to gain insight into the mechanisms responsible for the phenotypic diversity found in enteric ganglia. Our observations rely on the detection of immunoreactivity, and our results may
NEUROPEPTIDE DEVELOPMENT ALONG GUT
175
Fig. 6. A. Network of VIP-IR cells and processes is seen in a whole mount of an E6.5 proventriculus. B. Portion of A at high magnification. Micrograph shows VIP-IR in cytoplasm and processes of myenteric neurons (arrow). C. VIP-IR cells (arrow) are seen as small dots over the
length of a n E7.5 cecum. D. Portion of C at high magnification. Micrograph shows VIP-IR cells (arrow) in the cecum. Bars = 200 p m for A, C, 10 pm for B, D.
reflect the sensitivity of our antisera rather than differences in expression. Because of the variability in the appearance of the SOM-IR in the midgut, we are uncertain of the exact time of appearance of SOM-IR cells. Verification of the reported differences will require either specific radioimmunoassays in conjunction with high pressure liquid chromatography and/or hybridization with specific nucleotide probes. The cells expressing these peptides are derived from the vagal neural crest which enter the pharynx at E2.5 (Tucker et al., '86) and approach the ileocecal junction by E5.5 (Epstein et al., '91); these precursors give rise to aggregates of cells, which are the forerunners of the enteric ganglia. Within these aggregates are cells which will differentiate into neurons with different transmitter phenotypes and into glial cells. The signals that regulate the expression of these phenotypes are unknown, although humoral factors
have been isolated that support peptide expression in cultures of sympathetic neurons (Nawa and Patterson, '90; Nawa and Sah, '90). We see similarities in the development of SOM and VIP. Both peptides first appear in the region of the proventriculus-gizzard. SOM is present in the region of the gizzard primordium at E4.0 andVIP at the bottom of the proventriculus at E5.5. It is interesting that the gizzard is the site where these peptides are first detected. Our expectation was that expression would parallel the time course of the movement of the crest-derived cells down the gut, i.e., expression would be seen first in the esophagus, followed by the gizzard, duodenum, and midgut. However, that is not the case, and the fact that expression occurs first in the gizzard suggests this organ can promote differentiation. One possibility is that the large amount of mesenchyme associated with the gizzard primordium produces factors
176
M.L. EPSTEIN AND K.T. POULSEN
Figure I
NEUROPEPTIDE DEVELOPMENT ALONG GUT
177
Fig. 8. A. Micrograph shows a few VIP-IR cells and processes in the ES.5 gut at the level of the yolk stalk. B. Portion of A at high magnification shows a VIP-IR cell (arrow). C. Low-power micrograph shows VIP-IR (arrow) in the terminal portion of an E8.5 rectum. Compared to the terminal region, the amount of immunostain is
decreased in the middle of the rectum, shown in the left portion of the micrograph. D. Portion of C at high magnification shows a VIP-IR neuron (arrow) in the terminal rectum. Bars = 200 km for A, C, 10 ym forB,D.
that activate neuropeptide expression in the neuronal precursors. Another possibility is that the mesenchyme supports a large number of neuronal precursors and cellcell interactions between precursor cells trigger neuropeptide expression. Both VIP- and SOM-IR appear in cells in the hindgut prior to their appearance in the midgut.
Isolated SOM- and VIP-IR cells are found in the proximal ceca at E6.5 and E7.5, respectively. These cells presumably arise from precursors derived from sacral neural crest, which arise posterior to somite 28. In recent studies Pomeranz and Gershon ('90) have described two populations of sacral crest cells which enter the gut. One group remained in the dorsal mesentery to form the ganglion of Remak, whereas the other invaded the hindgut and ascended toward the midgut. Our data suggest that the latter group may provide the precursors for the SOM and VIP-IR cells we describe in the ceca. Although vagal precursors are present in the terminal ileum at E5.5 (Epstein et al., '911, expression of these peptides is delayed at least 5 days in this gut region. It is curious that expression occurs so early in the cecum and so late in the ileum, a contiguous area. It is of interest that in both the hindgut and foregut, neuropeptide
Fig. 7. Micrographs showing regions of E8.5 gut. A. VIP-IR cells are found in large numbers in the gizzard. B. Portion of A at high magnification. Cells (arrow) are found in myenteric ganglia. C. Ganglia in the proventriculus contain numerous VIP-IR cells. D. Portion of C at high magnification. VIP-IR neurons (arrow) are shown in myenteric ganglia. E. VIP-IR cells are found in the esophagus (arrow). Inset: VIP-IR cells are shown at high magnification. F. The descending duodenum contains prominent VIP-IR connectives (arrow) but few cell bodies. Bars = 200 ym for A, C, D, F, 10 pm for B, D, E inset.
178
M.L. EPSTEIN AND K.T. POULSEN
expression should first occur at a location which is distant findings suggest that these sites provide environmental from the site of precursor immigration into the gut. Our cues which lead to expression of these neuropeptides. observations of the ceca as well as the gizzard suggests a Additional work will be necessary to uncover these cues. regional heterogeneity in the ability of the gut to initiate neuropeptide expression. In the foregut the expression of these peptides proceeds ACKNOWLEDGMENTS in both oral and anal directions from the site of initial We are grateful to Dr. Alison Buchan for providing the appearance despite the fact that the neural crest precursors antibody to somatostatin and to S. Presley, J. Dahl, K. advance and form aggregates in an oral-to-and direction. Expression moving in an oral direction is seen in the Biefeldt, E. Schultz for reading the manuscript. Supported esophagus where the precursors produce both neuropep- by NSF grant BNS-8820658. tides at E8.5, but neither at E6.5. It is not clear if expression at E8.5 in the esophagus is related to the LITERATURE CITED maturation of the neuronal precursors or mesenchyme or both. Bornstein, J.C., and J.B. Furness (1988) Correlated electrophysiologicaland Differences in the expression of SOM and VIP were also histochemical studies of submucous neurons and their contribution to understanding enteric neural circuits. J. Auto. Nervous System 25:l-13. apparent. VIP expression progresses more rapidly through the midgut and hindgut than SOM. In the midgut at E8.5, Buchan, A.M.J.,L.K.J. Sikora, J.G. Levy, C.H.S. McIntosh, I. Dyck, and J.C. Brown (1985) An immunocytochemical investigation with monoclonal VIP is found distal to the yolk stalk, whereas SOM is only in antibodies to somatostatin. Histochemistry 83: 175-180. the descending duodenum. Although SOM-IR cells were Cavanaugh, M.E., and J.G. Parnavelas (1988) Development of somatostatin visible in the ceca of the hindgut at E6.5, cells were not immunoreactive neurons in the rat occipitalcortex: A combined immunocytochemical-autoradiographicstudy. J. Comp. Neurol. 268;l-12. observed in the rectum until E11.5. Hence, there is a large delay between the initial appearance of SOM-IR cells in ceca Costa, M., J.B. Furness, and I.J. Llewellyn-Smith (1987) Histochemistry of the enteric nervous system. In L.R. Johnson (ed): Physiology of the and their appearance in the rectum. In contrast, the VIP-IR Gastrointestinal Tract, 2nd ed. New York: Raven Press. cells populated the entire ceca at E8.5 and the entire rectum Epstein, M.L., D. Sherman, and M.D. Gershon (1980) Development of by E10.5. serotonergic neurons in the chick duodenum. Dev. Biol. 77:2240. It is not clear why VIP is expressed earlier than SOM Epstein, M.L., J. Hudis, and J.L. Dahl(1983)The development of peptidergic along the small intestine and in the hindgut. One possibility neurons in the foregut of the chick. J. Neurosci. 33431-2447. is that SOM expression in myenteric neurons is retarded by Epstein, M.L., K.T. Poulsen, and R. Thiboldeaux (1991) Formation of ganglia in the gut of the chick embryo. J. Comp. Neurol307:l-ll. the presence of SOM-IR extrinsic fibers. The small intestine contains SOM-IR extrinsic fibers, which arise from neurons Furness, J.B., and M. Costa (1987) The Enteric Nervous System. Edinburgh: Churchill Livingstone. in the ganglion of Remak. Extrinsic fibers are not found in the ceca where the SOM first appears in the hindgut. Garcia-Arraras, J.E., M. Chanconie, and J. Fontaine-Perus (1984) In vivo and in vitro development of somatostatin-like immunoreactivity in the VIP-IR progresses rapidly down the small intestine, which peripheral nervous system of quail embryos. J. Neurosci. 4r1549-1558. contains no major extrinsic VIP-IR fibers at E8-13. VIP-IR Hamburger, V., and H.L. Hamilton (1951) A series of normal stages in the cells are not found in the ganglion of Remak. The expresdevelopment of the chick embryo. J. Embryol. Exp. Morphol. 88:49-92. sion of SOM and VIP progresses at similar rates up the Larsson, L.-I. (1981) A novel immunocytochemical model system for specificity and sensitivity screening of antisera against multiple antigens. J. foregut between E6.5 and E8.5. The extrinsic fibers to the Histochem. Cytocbem. 29:408410. foregut are the vagus and sympathetic nerves; we did not observe VIP or SOM-IR in the vagus. Sympathetic fibers, Le Douarin, N.M. (1982) The Neural Crest. Cambridge: Cambridge University Press. which are known to contain SOM (Maxwell et al., '84), do Le Douarin, N.M., and M.-A. Teillet (1973) The migration of neural crest not reach the gut until later in development (Epstein et al., cells to the wall of the digestive tract in avian embryo. J. Embryol. Exp. '80). Further experiments will be necessary to determine if Morphol. 30:3148. the extrinsic fibers or other local environmental cues Maxwell, G.D., P.D. Sietz, and P.H. Chenard (1984) Development of influence expression. somatostatin-like immunoreactivity in embryonic sympathetic ganglia. J. Neurosci. 4:576-584. Others have found SOM early in the development of the peripheral nervous system. SOM-IR appears in the develop- Nawa, Hiroyuki, and P.H. Patterson (1990) Separation and partial characterization of neuropeptide-inducing factors in heart cell conditioned meing sympathetic chains of quail embryos at E4 (Garciadium. Neuron 4.269-27'7. Arraras et al., '84; Maxwell et al., '84). However, on the Nawa, Hiroyuki, and D.W.Y. Sah (1990) Different biological activities in basis of immunostaining and radioimmunoassay, the conditioned media control the expression of a variety of neuropeptides in cultured sympathetic neurons. Neuron 4.279-287. SOM-IR appears to decrease and reach a minimum value at about E10. We also observed a loss of immunostaining in Pomeranz, H.D., and M.D. Gershon (1990) Colonization of the avian hindgut by cells derived from the sacral neural crest. Dev. Biol. 137:378-394. the proventriculus and esophagus at about E10. A decrease in the number of SOM-IR neurons has also been observed Pomeranz, H.D., T.P. Rothman, and M.D. Gershon (1991) Colonization of the post-umbilical bowel by cells derived from the sacral neural crest: during the development of the rat occipital cortex; this Direct tracing of cell migration using an intercalating probe and a decrease resulted from cell death (Cavanaugh and Parnareplication-deficient retrovirus. Development 11I:647-655. velas, '88). However, unlike the sympathetic chains and rat Saffrey, M.J., J.M. Pol&, and G. Burnstock (1982) Distribution of vasoactive intestinal polypeptide-, substance P-, enkephalin-, and neurotensin-like cortex, SOM-IR in the gut is restored. We do not know if the immunoreactive nerves in the chicken gut during development. Neurodecrease in the gut is the result of cell death or a change in sci. 7:279-293. expression. G.C., G. Ciment, and J.-P. Thiery (1986) Pathways of avian neural In conclusion, our observations indicate that the expres- Tucker, crest cell migration in the developing gut. Dev. Biol. 116r439-450. sion of the neuropeptides SOM and VIP is initiated at two Yntema, C.L., and W.S. Hammond (1954) The origin of intrinsic ganglia of discrete and discontinuous locations, the proventriculustrunk viscera from vagal neural crest in the chick embryo. J. Comp. Neurol. 101:515-541. gizzard of the foregut and the ceca of the hindgut. These
