- Email: [email protected]
Distribution and characterisation of neuromedin U-like immunoreactivity in rat brain and intestine and in guinea pig intestine
Regulatory Peptides, 20 (1988) 281-292 Elsevier
281
RPT 00674
Distribution and characterisation of neuromedin Ulike immunoreactivity in rat brain and intestine and in guinea pig intestine S.J. Augood, J.R. Keast and P.C. Emson MRC Group, Department of Neuroendocrinology, AFRC, Institute of Animal Physiology and Genetics Research, Babraham ( U.K.) (Received 22 July 1987; revised version received 14 October 1987; accepted 26 october 1987)
Summary Neuromedin U-8 (NMU-8) is a peptide isolated from porcine spinal cord which contracts blood vessels and the uterus. Antisera were raised against NMU-8 and used in a radioimmunoassay (RIA) together with HPLC to characterize NMU-like immunoreactivity (NMU-LI) in tissues extracts of rat brain and gut and guinea pig gut. Samples of duodenum, ileum and distal colon were taken from both species, and processed for detection of NMU-LI by fluorescence immunohistochemistry. In RIA the antiserum had no cross-reactivity with neuropeptide Y, vasoactive intestinal peptide or the C-terminal hexapeptide of pancreatic polypeptide. Preincubation of antiserum with any of these pepttdes had no effect on the NMU-LI staining. In rats the highest content of NMU-L! was found in the ileum and the lowest in the cerebral cortex and striatum. HPLC studies showed that at least two molecular forms of NMU-LI were present in both species. In rat small intestine, subpopulations of submucous and myenteric neurones were stained; nerve fibres and terminals within these ganglia and in the mucosa were also seen. NMU-LI was sparse in the muscle. In guinea pig ileum small populations of nerve terminals were seen in both myenteric and submucous ganglionated plexuses. No endocrine cells were stained in either species. Immunohistochemistry; Neuromedin; Enteric nervous system; Radioimmunoassay; High-performance liquid chromatography
Correspondence." P.C. Emson, MRC Group, Department of Neuroendocrinology, AFRC, Institute of Animal Physiology and Genetics Research, Babraham, Cambridge, CB2 4AT, U.K.
282 Introduction
Recently Minamino et al. have purified and sequenced a series of neuropeptides (termed neuromedins) which they characterized by their ability to contract smooth muscle preparations [1]. This series includes a number of novel neuropeptides ineluding neuromedin B and C (bombesin-like) [2,3], neuromedin K and L (kassininlike) [4,5] and neuromedin N (neurotensin-like) [6]. The most recent additions to this series are two peptides with a carboxy-terminal asparagine amide which they termed neuromedin U, from their ability to contract the uterus [7]. The latter peptides isolated and sequenced from porcine spinal cord were an octapeptide (neuromedin U8; NMU-8) and a longer peptide (NMU-25) containing the neuromedin U-8 sequence at its carboxy-terminal end sequence preceded by paired basic residues, suggesting that NMU-8 may arise from NMU-25 by tryptic cleavage [7]. Apart from their ability to stimulate uterine smooth muscle the peptides also have vasoconstrictor (hypertensive) properties [1,7]. To investigate the distribution, content and localization of NMU-like peptides in the rat we have established sensitive radioimmunoassays for NMU-8 and NMU-25. We have shown that NMU-like immunoreactivity (NMU-LI) is present in brain and intestine extracts, but at a higher concentration in the latter. Immunohistochemical studies showed that the NMU-LI is localized in neurones and their processes in the enteric nervous system. An initial report of the extraction and assay of NMU-LI in several species has appeared [8], although no histochemical data were included.
Materials and Methods
Peptides All synthetic peptides, including synthetic porcine NMU-8 and NMU-25, were purchased from Peninsula Laboratories, U.K. All other reagents were of analytical grade. Antisera NMU-directed antisera were raised in New Zealand white rabbits. For immunization NMU-8 (224 nmol) was conjugated to bovine thyroglobulin (1 mg, Sigma T1001) with 1-ethyl-3-(3-dimethyl-amino propyl)-carbodiimide (30 mg, Sigma E-7750) and emulsified in complete Freund's adjuvant. Rabbits were boosted 8 weeks later using NMU-8 (224 nmol) conjugated in the same manner but emulsified with incomplete Freund's adjuvant. Useable antisera were obtained 4 weeks after the first boost, and aliquots were stored at -70°C. Iodination Radioiodination of NMU-L8 was carried out using the iodogen method as described by Salacinski et al. [9]. This method uses 1,3,4,6-tetrachloro-3ct6~t-diphenylglycouril (lodogen) as the solid-phase catalyst, and Na 12~I as the free iodine donor. Purification of the iodinated peptide was achieved using a prewashed 1.0 × 0.5 cm
283 Sep-pak cartridge (Waters Associates, p-Bondapak C18 reverse-phase) with methanol/water/1% TFA as the eluting buffer. The cartridge was pre-equilibrated with 1 mg Polypep to reduce non-specific peptide absorption. The purified tracer was stored at -200C in 60% methanol.
Assay procedure Radioimmunoassays were routinely performed in a total volume of 0.6 ml of 50 mM phosphate buffer, pH 7.4 containing 0.1% BSA (RIA grade, Sigma). Stock peptide solutions were stored at -20°C in phosphate buffer, pH 7.4, with samples for standard curves dissolved in assay buffer and stored at -200C for no longer than 8 weeks. Duplicate standard curves were set up by serial dilution with 3.6-1800 fmol NMU-8/tube. Standards and assay samples were made up to a total volume of 0.4 ml in assay buffer to which 0.1 ml of 12sI NMU-8 (3-8000 dpm) and 0.1 ml of antiserum (Rb 59) were added to give a final antiserum dilution of 1:60,000. The tubes were vortexed and incubated at 4°C (24-72 h) followed by separation of bound and free NMU-8 tracer by the addition of 0.25 ml/tube of ice-cold charcoal suspension (25% gelatine, 3.2% charcoal [Norit GSX], 0.32% dextran in 50 mM phosphate buffer, pH 7.4). All tubes were centrifuged immediately at 1000 g, 4°C for 20 min in order to obtain a charcoal pellet. The supernatant (bound, B) was decanted and both the supernatant and charcoal pellet (free, F) were counted in a gamma spectrometer for radiation content. The ratio B/(B + F) was calculated for each tube and displacement curves obtained by plotting B/(B + F) as percentage bound against log NMU-8 (fmol). The percentage bound figures were corrected for apparent tracer binding in the absence of antiserum. The concentration of peptide standards was checked by amino acid analysis. Tissue extraction Albino male Sprague-Dawley rats (180--250 g) were killed by decapitation. Spinal cord, hypothalamus and gastrointestinal tract were immediately dissected out and frozen. Weighing and extractions were carried out within a 24-h period. Two different extraction media were tested for their ability to extract NMU-LI: (i) 0.5 M acetic acid: frozen tissue pieces were added to boiling acid, boiled for a further 10 min, homogenised (Ultra-turrax, setting 4, 5 rain) and boiled for a further 5 rain; and (ii) 0.1 N hydrochloric acid: frozen tissue pieces were homogenised in 10 vols. cold acid. Both groups of tissue extracts were then centrifuged in the cold (4°C) at 20,000 g for 30 minutes, supernatants decanted, freeze-dried overnight, reconstituted in assay buffer and finally recentrifuged (1800 g, 10 min) before being used for radioimmunoassay. Samples extracted using acetic acid did not dilute in parallel with the standards in the assay. In contrast the hydrochloric acid extracts diluted in parallel with the standards, and this method of extraction was used for all subsequent sample preparations. HPLC sample extractions Albino male Sprague-Dawley rats (180-250 g) and Duncan-Hartley guinea pigs were killed by decapitation. Cortex, striatum, hypothalamus, spinal cord and gas-
284 trointestinal tract were dissected out, frozen, then weighed and extracted in 0.1 N hydrochloric acid as described above. The supernatant obtained after the first centrifugation was purified using disposable reverse phase ~-Bondapak C18 cartridges (Sep-pak, Waters Associates) using 2 ml aliquots of 80% acetonitrile/water to etute the peptide. The eluants collected were evaporated to dryness overnight, reconstituted in 1 ml water/0.1% TFA then applied to a/t-Bondapak C18 column (30 × 0.39 cm). Elution was achieved using a flow rate of 2 ml/min and a linear gradient of 16-44% acetonitrile/water/0.1% TFA over 25 rain. 2 ml fractions were collected, freeze-dried and reconstituted in assay buffer for determination of their NMU-LI content by radioimmunoassay. Immunohistochernistry Albino Sprague-Dawley rats and Duncan-Hartley guinea pigs were killed by decapitation. Segments of intestine were excised, including duodenum, ileum and distal colon. The tissues were then washed briefly in 0.1 M phosphate-buffered saline (PBS), pH 7.4, to remove any gut contents. Some tissue segments were prepared for studying wholemounts of myenteric and submucous plexuses by cutting open along the mesenteric border and pinning out, slightly stretched, on dental wax, prior to fixation [10]. All tissues were fixed overnight by immersion in a picric acid-formaldehyde solution at 4°C [11] and the following day washed free of picric acid with 80% alcohol, dehydratedd, cleared in xylene and rehydrated. Tissues to be used for wholemounts were stored in PBS containing 0.01% sodium azide, whereas the solution used for storing tissues to be sectioned also contained 30% sucrose. All fixed tissues were stored at 4°C and examined immunohistochemically within 3 weeks of fixation. Cryostat sections (14/~m) were placed on chrome alum-gelatine-coated slides and preincubated in non-immune sheep serum (1:10 in PBS) for 30 min. Sections were then washed, incubated with the primary antisera (Rb 59, Rb 62; 1:200 (wholemounts) or 1:400 (sections) raised in rabbits against NMU-8) and left overnight at room temperature in humidified chambers. The following day all sections were washed in PBS and incubated for 1 h at room temperature with sheep anti-rabbit immunoglobulin conjugated to fluorescein isothiocyanate (FITC, 1:80, Wellcome). Finally sections were re-washed in PBS and mounted in phosphate-buffered glycerol (pH 8.6) for fluorescence microscopy. Using whole-mounts of submucous and myenteric plexus of both species, absorption tests were carried out for NMU-8, VIP, NPY and PP-6 (10-7-10 -s M).
Results RIA and HPLC Both rabbits immunized with NMU-8 developed useful antisera; in the studies discussed here N M U - R b 59 (data 3.11.86) was used. This assay, using tracer made by the iodogen method, had the properties characterized in Table I. The antiserum showed essentially no cross-reactivity with any neuropeptide tested, with the exception of the slightly related pancreatic polypeptide carboxy-terminal sequence, which
285 TABLE I Neuromedin U-8 radioimmunoassaycharacteristics Final antiserum dilution (Rb59) ICs0 (fmol)
1:60,000 117.0 + 19.0 (n = 6)
Minimum detectable amount/tube (fmol) % Binding range Incubation temperature Incubation time Total tube volume Added 12Si_NMU.8/tube
18 5-65% 4°C ~48 h 0.6 ml 3-8000 dpm
had <0.001% cross-reactivity with the antiserum. The antiserum completely crossreacted with NMU-25 (Fig. 1A). Tissue samples diluted serially and gave parallel displacement with the NMU-8 tracer. Initial regional distribution studies (Table II) carried out on the rat brain and intestines showed that, in contrast to many neuropeptides, NMU-8 LI was most concentrated in the intestines and relatively smaller amounts were found in the brain. Extraction o f these tissues using 0.1 N HCI to reduce the possibilities of tryptic cleavage of the dibasic group in NMU-25, followed by H P L C on a/~-Bondapak C18 column, showed that NMU-like material was present in at least two molecular forms and in some tissues up to 4-5 peaks of N M U - L I were detected. Calibration of the H P L C with the synthetic porcine peptides (NMU-8 and NMU-25) revealed that none
70
60
50
%B
\
\
.NMo8
40
B+F 30 20
10
100
101
102
103
fmol
Fig. 1. Cross-reactivity of NMU-8 and NMU-25 in the N M U radioimmunoassay.
286 TABLE I1 Regional distribution of NMU-L1 in rat and guinea pig tissues n
NMU-LI content (pmol/g)
Rat Cerebral Cortex Striatum Hypothalamus Cervical Cord Lumbar Cord Thoracic Cord Sacral Cord Stomach Ileum Duodenum Distal Colon
4 4 7 3 3 4 4 3 5 5 4
0.51 0.51 3.05 1.76 1.33 1.18 0.81 0.14 7.96 5.51 1.35
Guinea Pig Ileum
4
1.3 4- 0.20
44444444444-
0.12 0.15
0.41 0.08 0.01 0.19 0.22 0.03 0.95 1.66 0.16
Values are means -4- S.E.M.
of the peaks detected in rat (Fig. 2A-C) or guinea pig (Fig. 2D) samples eluted at exactly the positions expected for the porcine synthetic standards. Extracts of both rat and guinea pig intestine (Fig. 2A, B and D) were more heterogenous than rat CNS samples (Fig. 2C).
lmmunohistochernistry Initially two NMU antisera were used (Rb 59 and Rb 62) and, although both r50 U8
U25
600
400 3o
200 ~ 20
0
r
~
5
Fig. 2A.
=
=
,
10
15
FRACTION NUMBER
J
20
"
-
25
30
287
800"
U8
U25
1
1
-50
600 40
400
~,
|
, V
30
0 I-w 0
200 20
,
5
10
15
20
~
-
"
7
25
.
.
.
.
30
FRACTION NUMBER
600
U8
U25
1
1 40
v
400 30 z
0 Iw 0 ,<
200 20
,
5
,
10
,
15
,
,
20
25
30
FRACTION NUMBER
800,
U8
U25
l
1
.50
600 tr
E '-i I ::::)
400
o 30
Z
,<
200
20
10
15
20
25
30
FRACTION NUMBER
Fig. 2. Elution pattern of N M U - L I from extracts o f (A) rat duodenum, (B) rat ileum, (C) rat thoracic cord and (D) guinea pig ileum. Arrows indicate elution positions o f synthetic porcine NMU-8 (U8) and NMU-25 (U25). Fraction volume was 1 ml.
288 showed the same distribution of immunoreactivity, Rb 59 gave a much brighter reaction and was therefore used in all subsequent studies. For optimal staining it was
289
necessary to cut and process tissue within 2-3 weeks of fixation. In both rat and guinea pig N M U - L I was localized to neurones in the intestine (Fig. 3). In all tissues, the N M U - L I was much brighter in rats than in guinea pigs. No N M U - L I was present in epithelial endocrine cells. In both rat and guinea pig ileum, N M U - L I was partially absorbed by 10 - 7 M NMU-8 and abolished by preincubation with 10 - 6 M NMU-8. Preincubation with VIP, NPY and PP-6 (up to 10 -5 M) had no effect on NMU-LI. Rat intestine
In the ileum positive nerve cell bodies were found in both myenteric and submucous ganglia. A small population ( < 20%) of myenteric neurones were stained by this antiserum; a moderate supply of nerve terminals were present in all myenteric ganglia, but were not distributed evenly in them (Fig. 3A). Each ganglion usually contained one or two 'baskets' of dense nerve terminals surrounding unstained cell bodies, as well as smaller numbers of fine nerve terminals surrounding most other neurones. A small proportion of myenteric neurones had no N M U - L I in nerve terminals near them; this group included both stained and unstained neurones. Nerve fibres were found only rarely in the internodal strands or external muscle. In the submucous plexus, a much larger proportion (ca. 50-70%) of the total number of neurones had moderately bright N M U - L I (Fig. 3B, C). Quite a few nerve terminals with N M U - L I were found around most submucous neurones. As seen in the myenteric plexus, these were not evenly distributed, with some unstained neurones being surrounded by quite dense baskets of terminals and finer networks surrounding most of the remaining neurones; a small number of neurones received no nerve terminals. A substantial population of nerve fibres were found at all levels of the mucosa, surrounding the crypts and entering the villi (Fig. 3E, F). At both sites, most of the nerve fibres were close to the epithelium. Few were in the lamina propria. In the duodenum the distribution of N M U - L I was very similar to that in the ileum except that a slightly lower proportion of neurones in both ganglionated plexuses was stained; In the colon the innervation of the myenteric plexus and muscle was similar to the small intestine (Fig. 3D). However, that of the submucous plexus and mucosa was much sparser. Very few neurones were stained in the submucous ganglia and
Fig. 3. Localization of NMULI in the guinea pig and rat intestine. A: rat ileum, myenteric plexus. Two baskets of nerve terminals, surrounding unstained cells, are also shown (arrowheads). B,C: rat ileum, submucous plexus. NMU-LI in some nerve cell bodies (examples shown by arrowheads) and in nerve fibres and terminals. D: rat distal colon, myenteric plexus. One neurone containing NMU-LI is evident (arrowhead). Many nerve fibres and terminals are also present. E: rat ileum, mucosa. Nerve fibres surround many of the glands, shown here in transverse section. F: rat ileum, mucosa. Oblique section through a villus and some intestinal glands, showing many nerve fibres close to the epithelium. G: guinea pig ileum, myenteric plexus. Part of a myenteric ganglion, showing many fine nerve terminals surrounding unstained neurones. Micrographs A-D,G are from wholemounts; E,F from 14-#m sections. Bar = 10 #m.
290 nerve terminals were rare. In the mucosa only occasional single fibres were seen, mainly in the basal level of mucosa.
Guinea pig intestine In all layers and tissues the N M U - L I was much paler than in rats, and, in some cases, absent altogether. No nerve fibres were found in any layer of the duodenum or colon. In the ileum quite a dense supply of relatively faint nerve terminals was found in the myenteric plexus (Fig. 3G) and a less dense one in the submucous plexus. In both cases, the nerve terminals with N M U - L I surrounded most neurones in the ganglia. No N M U - L I was seen in nerve cell bodies. There were only very rare nerve fibres in the external muscle and none were found in the mucosa.
Discussion The regional distribution of N M U - L I in the rat indicates clearly that this peptide is, as reported by Domin et al. [8], particularly concentrated in the intestines. The brain areas assayed here contained relatively small amounts of NMU-LI, compared to the gut samples (although the hypothalamus was an exception; Table II). Our results show that the N M U - L I is concentrated in neurones and their processes within the intestine. H P L C separations of brain and intestinal extracts showed that the N M U - L I was particularly heterogeneous, and that none of the peaks of immunoreactivity eluted at the position expected for porcine NMU-8 or NMU-25. Similar results were obtained by Domin et al. [8] who also noticed that guinea pig gut extracts were the most heterogeneous, and that in the rat one major peak of N M U - L I eluted after NMU-25 on reverse phase. The different elution positions probably reflect differences between the amino acid sequences of rat guinea pig N M U peptides and the porcine ones; clearly sequencing of the relevant rat and guinea pig peptides is required. Another possibility is that some of the peaks of N M U - L I represent precursors or metabolites of rat and guinea pig NMU-25 or NMU-8, although our choice of a strong mineral acid (0.1 N HCI pH 1-2) extraction was designed to try to reduce this possibility. Our previous experience with boiling water-acetic acid extractions of small gut peptides has shown that under certain circumstances acid proteases may still be active even at pH 3-4 and will generate large amounts of spurious immunoreactivity [12]. In this respect it should be noted that the values for N M U - L I reported here are significantly lower than those observed by Domin et al. [8]. The reason for this discrepancy is not clear and may perhaps relate to the choice of the extraction method or the specificity of the various antisera used by our two laboratories. This can only be determined by further work. The immunohistochemical studies describe the cellular distribution of N M U - L I in the rat and guinea pig intestines, where it is found only in neural tissue. In contrast to some other peptides, for example CCK [13], it does not have a dual distribution in endocrine and neural tissue in these two species. The marked difference between
291 rats and guinea pigs in the brightness of staining also suggests, as do the chromatographic data, that there may be sequence differences between the NMU peptides. It appears that neither of the two antisera used here recognised the guinea pig NMU very well in fixed tissue and we have therefore not commented in detail on the difference in distribution of NMU-LI between rats and guinea pigs. The major projection of the myenteric ganglia is to the external muscle, to stimulate or inhibit contraction of smooth muscle, whereas the major projection of the submucous ganglia is to the mucosa, to modify absorptive/secretory processes. In rat small intestine NMU-LI is found in neurones of both ganglionated plexuses. However, although substantial numbers of immunoreactive fibres are found in the mucosa, very few are found in the internodal strands of the myenteric plexus or in the muscle. It is not known whether the NMU-LI neurones in the myenteric plexus do not project to these sites or whether the levels of NMU-LI are below the threshold for detection with this technique. Within the ganglionated plexuses there is an uneven distribution of nerve terminals, suggesting specific sites of action in each ganglion. The most extensive innervation is of a small population of unstained submucous and myenteric neurones, usually only one or two per ganglion. It would be interesting to determine (e.g. by double staining methods) which substances these neurones contain. Studies over the last few years, in particular those using double staining immunohistochemical methods, have shown that the majority of enteric neurones in the guinea pig and many in the rat contain more than one peptide [14,15]. In guinea pigs it has been shown that different subpopulations of enteric neurones can be defined by the combination of peptides contained in them; these combinations can be used to identify neurones with different projections [15]. It is not yet known which subpopulations of enteric neurones contain NMU, or the projection pattern of these neurones. There is still very little information available on the physiological actions of NMU. In contrast to tachykinins, neither porcine NMU-8 nor NMU-25 have an effect on contraction of guinea pig ileum smooth muscle [7]. However, the present studies suggest that the guinea pig NMU differs from porcine NMU and the experiments should therefore be repeated using the native peptide sequence, as found in guinea pigs. In addition, there appear to be very few nerve fibres containing NMU-LI in the external muscle. The action of NMU on myenteric and submucous neurones, where extensive networks of nerve terminals are found, and the action on the epithelium, close to which many fibres are found, have not yet been reported. The biochemical and immunohistochemical data provided here comprise a basis for many further studies on the possible role of NMU in the enteric nervous system and the regulation of gastrointestinal function.
Acknowledgements J.R. Keast gratefully acknowledges the support of the Nuffield Foundation and the Gowrie Foundation in providing a travel grant. Amino acid analysis of standard
292
samples was carried out by Dr J. McMurray, MRC Laboratory of Molecular Biology. This manuscript could not have been produced without the hard work of Mrs B. Waters.
References 1 Minamino, N., Sudoh, T., Kangawa, K. and Matsu, O., Neuromedins: novel neuropeptides identified in porcine spinal cord. In C.M. Deber, V.J. Hruby and K.D, Kopple (Eds.), Peptides, Structure and Function. Proceedings of the 9th American Peptide Symposium, Pierce Chem. Co., Rockford, IL, 1985, pp. 643-646. 2 Minamino, N., Kangawa, K. and Matsuo, H., Neuromedin B: a novel bombesin-like peptide identified in porcine spinal cord, Biochem. Biophys. Res. Commun., 114 (1983) 541-548. 3 Minamino, N., Kangawa, K. and Matsuo, H., Neuromedin C: a bombesin-like peptide identified in procine spinal cord, Biochem. Biophys. Res. Commun., 119 (1984) 14-20. 4 Kangawa, K., Minamino, N., Fukuda, A. and Matsuo, H., Neuromedin K: a novel mammalian tachykinin identified in porcine spinal cord, Biochem. Biophys. Res. Commun., 114 (1983) 533-540. 5 Minamino, N., Kangawa, K., Fukuda, A. and Matsuo, H., Neuromedin L: a novel mammalian tachykinin identified in porcine spinal cord, Neuropeptides 4 (1984) 157-166. 6 Minamino, N., Kangawa, K. and Matsuo, H., Neuromedin N: a novel neurotensin-like peptide identified in porcine spinal cord, Biochem. Biophys. Res. Commun., 122 (1985) 542-549. 7 Minamino, N., Kangawa, K. and Matsuo, H., Neuromedin U-8 and U-25: novel uterus stimulating and hypertensive peptides identified in porcine spinal cord, Biochem. Biophys. Res. Commun., 130 (1985) 1078-1085. 8 Domin, J., Ghatei, M.A., Chohan, P. and Bloom, S.R. Characterization of neuromedin U-like immunoreactivity in rat, porcine, guinea-pig and human tissue extracts using a specific radioimmunoassay, Biochem. Biophys. Res. Commun., 140 (1986) 1127-1134. 9 Salacinski, P., Hope, J., McLean, C., Clement-Jones, V., Sykes, J., Price, J. and Lowry, P.J., A new simple method which allows theoretical incorporation of radioiodine into proteins and peptides without damage, J. Endocrinol., 81 (1979) 1-13l. 10 Costa, M. and Furness, J.B., Immunohistochemistry on whole mount preparations. In Cuello, A.C. (Ed.), Immunohistochemistry, Wiley, Chichester, pp. 373-397. 11 Stefanini, M., DeMartino, C. and Zamboni, L., Fixation of ejaculated spermatozoa for electron-microscopy, Nature (Lond.), 216 (1967) 173-174. 12 Goedert, M., Neurotensin in neuronal and endocrine tissues, PhD Thesis, University of Cambridge, 1984. 13 Dockray, G.J. and Gregory, R.A., Relations between neuropeptides and gut hormones, Proc. R. Soc. Lond. Ser, B, 210 (1980) 151-164. 14 Ekblad, E., Winther, C., Ekman, R., H~.kanson, R. and Sundler, F,, Projections of peptide-containing neurones in rat small intestine. Neuroscience, 20 (1987) 169-188. 15 Costa, M., Furness, J.B. and Llewellyn-Smith, l.J., Histochemistry of the enteric nervous system. In Johnson, L.R. (Ed.), Physiology of the gastrointestinal tract, 2nd edn., Raven, New York, 1987, pp. 1-40.
281
RPT 00674
Distribution and characterisation of neuromedin Ulike immunoreactivity in rat brain and intestine and in guinea pig intestine S.J. Augood, J.R. Keast and P.C. Emson MRC Group, Department of Neuroendocrinology, AFRC, Institute of Animal Physiology and Genetics Research, Babraham ( U.K.) (Received 22 July 1987; revised version received 14 October 1987; accepted 26 october 1987)
Summary Neuromedin U-8 (NMU-8) is a peptide isolated from porcine spinal cord which contracts blood vessels and the uterus. Antisera were raised against NMU-8 and used in a radioimmunoassay (RIA) together with HPLC to characterize NMU-like immunoreactivity (NMU-LI) in tissues extracts of rat brain and gut and guinea pig gut. Samples of duodenum, ileum and distal colon were taken from both species, and processed for detection of NMU-LI by fluorescence immunohistochemistry. In RIA the antiserum had no cross-reactivity with neuropeptide Y, vasoactive intestinal peptide or the C-terminal hexapeptide of pancreatic polypeptide. Preincubation of antiserum with any of these pepttdes had no effect on the NMU-LI staining. In rats the highest content of NMU-L! was found in the ileum and the lowest in the cerebral cortex and striatum. HPLC studies showed that at least two molecular forms of NMU-LI were present in both species. In rat small intestine, subpopulations of submucous and myenteric neurones were stained; nerve fibres and terminals within these ganglia and in the mucosa were also seen. NMU-LI was sparse in the muscle. In guinea pig ileum small populations of nerve terminals were seen in both myenteric and submucous ganglionated plexuses. No endocrine cells were stained in either species. Immunohistochemistry; Neuromedin; Enteric nervous system; Radioimmunoassay; High-performance liquid chromatography
Correspondence." P.C. Emson, MRC Group, Department of Neuroendocrinology, AFRC, Institute of Animal Physiology and Genetics Research, Babraham, Cambridge, CB2 4AT, U.K.
282 Introduction
Recently Minamino et al. have purified and sequenced a series of neuropeptides (termed neuromedins) which they characterized by their ability to contract smooth muscle preparations [1]. This series includes a number of novel neuropeptides ineluding neuromedin B and C (bombesin-like) [2,3], neuromedin K and L (kassininlike) [4,5] and neuromedin N (neurotensin-like) [6]. The most recent additions to this series are two peptides with a carboxy-terminal asparagine amide which they termed neuromedin U, from their ability to contract the uterus [7]. The latter peptides isolated and sequenced from porcine spinal cord were an octapeptide (neuromedin U8; NMU-8) and a longer peptide (NMU-25) containing the neuromedin U-8 sequence at its carboxy-terminal end sequence preceded by paired basic residues, suggesting that NMU-8 may arise from NMU-25 by tryptic cleavage [7]. Apart from their ability to stimulate uterine smooth muscle the peptides also have vasoconstrictor (hypertensive) properties [1,7]. To investigate the distribution, content and localization of NMU-like peptides in the rat we have established sensitive radioimmunoassays for NMU-8 and NMU-25. We have shown that NMU-like immunoreactivity (NMU-LI) is present in brain and intestine extracts, but at a higher concentration in the latter. Immunohistochemical studies showed that the NMU-LI is localized in neurones and their processes in the enteric nervous system. An initial report of the extraction and assay of NMU-LI in several species has appeared [8], although no histochemical data were included.
Materials and Methods
Peptides All synthetic peptides, including synthetic porcine NMU-8 and NMU-25, were purchased from Peninsula Laboratories, U.K. All other reagents were of analytical grade. Antisera NMU-directed antisera were raised in New Zealand white rabbits. For immunization NMU-8 (224 nmol) was conjugated to bovine thyroglobulin (1 mg, Sigma T1001) with 1-ethyl-3-(3-dimethyl-amino propyl)-carbodiimide (30 mg, Sigma E-7750) and emulsified in complete Freund's adjuvant. Rabbits were boosted 8 weeks later using NMU-8 (224 nmol) conjugated in the same manner but emulsified with incomplete Freund's adjuvant. Useable antisera were obtained 4 weeks after the first boost, and aliquots were stored at -70°C. Iodination Radioiodination of NMU-L8 was carried out using the iodogen method as described by Salacinski et al. [9]. This method uses 1,3,4,6-tetrachloro-3ct6~t-diphenylglycouril (lodogen) as the solid-phase catalyst, and Na 12~I as the free iodine donor. Purification of the iodinated peptide was achieved using a prewashed 1.0 × 0.5 cm
283 Sep-pak cartridge (Waters Associates, p-Bondapak C18 reverse-phase) with methanol/water/1% TFA as the eluting buffer. The cartridge was pre-equilibrated with 1 mg Polypep to reduce non-specific peptide absorption. The purified tracer was stored at -200C in 60% methanol.
Assay procedure Radioimmunoassays were routinely performed in a total volume of 0.6 ml of 50 mM phosphate buffer, pH 7.4 containing 0.1% BSA (RIA grade, Sigma). Stock peptide solutions were stored at -20°C in phosphate buffer, pH 7.4, with samples for standard curves dissolved in assay buffer and stored at -200C for no longer than 8 weeks. Duplicate standard curves were set up by serial dilution with 3.6-1800 fmol NMU-8/tube. Standards and assay samples were made up to a total volume of 0.4 ml in assay buffer to which 0.1 ml of 12sI NMU-8 (3-8000 dpm) and 0.1 ml of antiserum (Rb 59) were added to give a final antiserum dilution of 1:60,000. The tubes were vortexed and incubated at 4°C (24-72 h) followed by separation of bound and free NMU-8 tracer by the addition of 0.25 ml/tube of ice-cold charcoal suspension (25% gelatine, 3.2% charcoal [Norit GSX], 0.32% dextran in 50 mM phosphate buffer, pH 7.4). All tubes were centrifuged immediately at 1000 g, 4°C for 20 min in order to obtain a charcoal pellet. The supernatant (bound, B) was decanted and both the supernatant and charcoal pellet (free, F) were counted in a gamma spectrometer for radiation content. The ratio B/(B + F) was calculated for each tube and displacement curves obtained by plotting B/(B + F) as percentage bound against log NMU-8 (fmol). The percentage bound figures were corrected for apparent tracer binding in the absence of antiserum. The concentration of peptide standards was checked by amino acid analysis. Tissue extraction Albino male Sprague-Dawley rats (180--250 g) were killed by decapitation. Spinal cord, hypothalamus and gastrointestinal tract were immediately dissected out and frozen. Weighing and extractions were carried out within a 24-h period. Two different extraction media were tested for their ability to extract NMU-LI: (i) 0.5 M acetic acid: frozen tissue pieces were added to boiling acid, boiled for a further 10 min, homogenised (Ultra-turrax, setting 4, 5 rain) and boiled for a further 5 rain; and (ii) 0.1 N hydrochloric acid: frozen tissue pieces were homogenised in 10 vols. cold acid. Both groups of tissue extracts were then centrifuged in the cold (4°C) at 20,000 g for 30 minutes, supernatants decanted, freeze-dried overnight, reconstituted in assay buffer and finally recentrifuged (1800 g, 10 min) before being used for radioimmunoassay. Samples extracted using acetic acid did not dilute in parallel with the standards in the assay. In contrast the hydrochloric acid extracts diluted in parallel with the standards, and this method of extraction was used for all subsequent sample preparations. HPLC sample extractions Albino male Sprague-Dawley rats (180-250 g) and Duncan-Hartley guinea pigs were killed by decapitation. Cortex, striatum, hypothalamus, spinal cord and gas-
284 trointestinal tract were dissected out, frozen, then weighed and extracted in 0.1 N hydrochloric acid as described above. The supernatant obtained after the first centrifugation was purified using disposable reverse phase ~-Bondapak C18 cartridges (Sep-pak, Waters Associates) using 2 ml aliquots of 80% acetonitrile/water to etute the peptide. The eluants collected were evaporated to dryness overnight, reconstituted in 1 ml water/0.1% TFA then applied to a/t-Bondapak C18 column (30 × 0.39 cm). Elution was achieved using a flow rate of 2 ml/min and a linear gradient of 16-44% acetonitrile/water/0.1% TFA over 25 rain. 2 ml fractions were collected, freeze-dried and reconstituted in assay buffer for determination of their NMU-LI content by radioimmunoassay. Immunohistochernistry Albino Sprague-Dawley rats and Duncan-Hartley guinea pigs were killed by decapitation. Segments of intestine were excised, including duodenum, ileum and distal colon. The tissues were then washed briefly in 0.1 M phosphate-buffered saline (PBS), pH 7.4, to remove any gut contents. Some tissue segments were prepared for studying wholemounts of myenteric and submucous plexuses by cutting open along the mesenteric border and pinning out, slightly stretched, on dental wax, prior to fixation [10]. All tissues were fixed overnight by immersion in a picric acid-formaldehyde solution at 4°C [11] and the following day washed free of picric acid with 80% alcohol, dehydratedd, cleared in xylene and rehydrated. Tissues to be used for wholemounts were stored in PBS containing 0.01% sodium azide, whereas the solution used for storing tissues to be sectioned also contained 30% sucrose. All fixed tissues were stored at 4°C and examined immunohistochemically within 3 weeks of fixation. Cryostat sections (14/~m) were placed on chrome alum-gelatine-coated slides and preincubated in non-immune sheep serum (1:10 in PBS) for 30 min. Sections were then washed, incubated with the primary antisera (Rb 59, Rb 62; 1:200 (wholemounts) or 1:400 (sections) raised in rabbits against NMU-8) and left overnight at room temperature in humidified chambers. The following day all sections were washed in PBS and incubated for 1 h at room temperature with sheep anti-rabbit immunoglobulin conjugated to fluorescein isothiocyanate (FITC, 1:80, Wellcome). Finally sections were re-washed in PBS and mounted in phosphate-buffered glycerol (pH 8.6) for fluorescence microscopy. Using whole-mounts of submucous and myenteric plexus of both species, absorption tests were carried out for NMU-8, VIP, NPY and PP-6 (10-7-10 -s M).
Results RIA and HPLC Both rabbits immunized with NMU-8 developed useful antisera; in the studies discussed here N M U - R b 59 (data 3.11.86) was used. This assay, using tracer made by the iodogen method, had the properties characterized in Table I. The antiserum showed essentially no cross-reactivity with any neuropeptide tested, with the exception of the slightly related pancreatic polypeptide carboxy-terminal sequence, which
285 TABLE I Neuromedin U-8 radioimmunoassaycharacteristics Final antiserum dilution (Rb59) ICs0 (fmol)
1:60,000 117.0 + 19.0 (n = 6)
Minimum detectable amount/tube (fmol) % Binding range Incubation temperature Incubation time Total tube volume Added 12Si_NMU.8/tube
18 5-65% 4°C ~48 h 0.6 ml 3-8000 dpm
had <0.001% cross-reactivity with the antiserum. The antiserum completely crossreacted with NMU-25 (Fig. 1A). Tissue samples diluted serially and gave parallel displacement with the NMU-8 tracer. Initial regional distribution studies (Table II) carried out on the rat brain and intestines showed that, in contrast to many neuropeptides, NMU-8 LI was most concentrated in the intestines and relatively smaller amounts were found in the brain. Extraction o f these tissues using 0.1 N HCI to reduce the possibilities of tryptic cleavage of the dibasic group in NMU-25, followed by H P L C on a/~-Bondapak C18 column, showed that NMU-like material was present in at least two molecular forms and in some tissues up to 4-5 peaks of N M U - L I were detected. Calibration of the H P L C with the synthetic porcine peptides (NMU-8 and NMU-25) revealed that none
70
60
50
%B
\
\
.NMo8
40
B+F 30 20
10
100
101
102
103
fmol
Fig. 1. Cross-reactivity of NMU-8 and NMU-25 in the N M U radioimmunoassay.
286 TABLE I1 Regional distribution of NMU-L1 in rat and guinea pig tissues n
NMU-LI content (pmol/g)
Rat Cerebral Cortex Striatum Hypothalamus Cervical Cord Lumbar Cord Thoracic Cord Sacral Cord Stomach Ileum Duodenum Distal Colon
4 4 7 3 3 4 4 3 5 5 4
0.51 0.51 3.05 1.76 1.33 1.18 0.81 0.14 7.96 5.51 1.35
Guinea Pig Ileum
4
1.3 4- 0.20
44444444444-
0.12 0.15
0.41 0.08 0.01 0.19 0.22 0.03 0.95 1.66 0.16
Values are means -4- S.E.M.
of the peaks detected in rat (Fig. 2A-C) or guinea pig (Fig. 2D) samples eluted at exactly the positions expected for the porcine synthetic standards. Extracts of both rat and guinea pig intestine (Fig. 2A, B and D) were more heterogenous than rat CNS samples (Fig. 2C).
lmmunohistochernistry Initially two NMU antisera were used (Rb 59 and Rb 62) and, although both r50 U8
U25
600
400 3o
200 ~ 20
0
r
~
5
Fig. 2A.
=
=
,
10
15
FRACTION NUMBER
J
20
"
-
25
30
287
800"
U8
U25
1
1
-50
600 40
400
~,
|
, V
30
0 I-w 0
200 20
,
5
10
15
20
~
-
"
7
25
.
.
.
.
30
FRACTION NUMBER
600
U8
U25
1
1 40
v
400 30 z
0 Iw 0 ,<
200 20
,
5
,
10
,
15
,
,
20
25
30
FRACTION NUMBER
800,
U8
U25
l
1
.50
600 tr
E '-i I ::::)
400
o 30
Z
,<
200
20
10
15
20
25
30
FRACTION NUMBER
Fig. 2. Elution pattern of N M U - L I from extracts o f (A) rat duodenum, (B) rat ileum, (C) rat thoracic cord and (D) guinea pig ileum. Arrows indicate elution positions o f synthetic porcine NMU-8 (U8) and NMU-25 (U25). Fraction volume was 1 ml.
288 showed the same distribution of immunoreactivity, Rb 59 gave a much brighter reaction and was therefore used in all subsequent studies. For optimal staining it was
289
necessary to cut and process tissue within 2-3 weeks of fixation. In both rat and guinea pig N M U - L I was localized to neurones in the intestine (Fig. 3). In all tissues, the N M U - L I was much brighter in rats than in guinea pigs. No N M U - L I was present in epithelial endocrine cells. In both rat and guinea pig ileum, N M U - L I was partially absorbed by 10 - 7 M NMU-8 and abolished by preincubation with 10 - 6 M NMU-8. Preincubation with VIP, NPY and PP-6 (up to 10 -5 M) had no effect on NMU-LI. Rat intestine
In the ileum positive nerve cell bodies were found in both myenteric and submucous ganglia. A small population ( < 20%) of myenteric neurones were stained by this antiserum; a moderate supply of nerve terminals were present in all myenteric ganglia, but were not distributed evenly in them (Fig. 3A). Each ganglion usually contained one or two 'baskets' of dense nerve terminals surrounding unstained cell bodies, as well as smaller numbers of fine nerve terminals surrounding most other neurones. A small proportion of myenteric neurones had no N M U - L I in nerve terminals near them; this group included both stained and unstained neurones. Nerve fibres were found only rarely in the internodal strands or external muscle. In the submucous plexus, a much larger proportion (ca. 50-70%) of the total number of neurones had moderately bright N M U - L I (Fig. 3B, C). Quite a few nerve terminals with N M U - L I were found around most submucous neurones. As seen in the myenteric plexus, these were not evenly distributed, with some unstained neurones being surrounded by quite dense baskets of terminals and finer networks surrounding most of the remaining neurones; a small number of neurones received no nerve terminals. A substantial population of nerve fibres were found at all levels of the mucosa, surrounding the crypts and entering the villi (Fig. 3E, F). At both sites, most of the nerve fibres were close to the epithelium. Few were in the lamina propria. In the duodenum the distribution of N M U - L I was very similar to that in the ileum except that a slightly lower proportion of neurones in both ganglionated plexuses was stained; In the colon the innervation of the myenteric plexus and muscle was similar to the small intestine (Fig. 3D). However, that of the submucous plexus and mucosa was much sparser. Very few neurones were stained in the submucous ganglia and
Fig. 3. Localization of NMULI in the guinea pig and rat intestine. A: rat ileum, myenteric plexus. Two baskets of nerve terminals, surrounding unstained cells, are also shown (arrowheads). B,C: rat ileum, submucous plexus. NMU-LI in some nerve cell bodies (examples shown by arrowheads) and in nerve fibres and terminals. D: rat distal colon, myenteric plexus. One neurone containing NMU-LI is evident (arrowhead). Many nerve fibres and terminals are also present. E: rat ileum, mucosa. Nerve fibres surround many of the glands, shown here in transverse section. F: rat ileum, mucosa. Oblique section through a villus and some intestinal glands, showing many nerve fibres close to the epithelium. G: guinea pig ileum, myenteric plexus. Part of a myenteric ganglion, showing many fine nerve terminals surrounding unstained neurones. Micrographs A-D,G are from wholemounts; E,F from 14-#m sections. Bar = 10 #m.
290 nerve terminals were rare. In the mucosa only occasional single fibres were seen, mainly in the basal level of mucosa.
Guinea pig intestine In all layers and tissues the N M U - L I was much paler than in rats, and, in some cases, absent altogether. No nerve fibres were found in any layer of the duodenum or colon. In the ileum quite a dense supply of relatively faint nerve terminals was found in the myenteric plexus (Fig. 3G) and a less dense one in the submucous plexus. In both cases, the nerve terminals with N M U - L I surrounded most neurones in the ganglia. No N M U - L I was seen in nerve cell bodies. There were only very rare nerve fibres in the external muscle and none were found in the mucosa.
Discussion The regional distribution of N M U - L I in the rat indicates clearly that this peptide is, as reported by Domin et al. [8], particularly concentrated in the intestines. The brain areas assayed here contained relatively small amounts of NMU-LI, compared to the gut samples (although the hypothalamus was an exception; Table II). Our results show that the N M U - L I is concentrated in neurones and their processes within the intestine. H P L C separations of brain and intestinal extracts showed that the N M U - L I was particularly heterogeneous, and that none of the peaks of immunoreactivity eluted at the position expected for porcine NMU-8 or NMU-25. Similar results were obtained by Domin et al. [8] who also noticed that guinea pig gut extracts were the most heterogeneous, and that in the rat one major peak of N M U - L I eluted after NMU-25 on reverse phase. The different elution positions probably reflect differences between the amino acid sequences of rat guinea pig N M U peptides and the porcine ones; clearly sequencing of the relevant rat and guinea pig peptides is required. Another possibility is that some of the peaks of N M U - L I represent precursors or metabolites of rat and guinea pig NMU-25 or NMU-8, although our choice of a strong mineral acid (0.1 N HCI pH 1-2) extraction was designed to try to reduce this possibility. Our previous experience with boiling water-acetic acid extractions of small gut peptides has shown that under certain circumstances acid proteases may still be active even at pH 3-4 and will generate large amounts of spurious immunoreactivity [12]. In this respect it should be noted that the values for N M U - L I reported here are significantly lower than those observed by Domin et al. [8]. The reason for this discrepancy is not clear and may perhaps relate to the choice of the extraction method or the specificity of the various antisera used by our two laboratories. This can only be determined by further work. The immunohistochemical studies describe the cellular distribution of N M U - L I in the rat and guinea pig intestines, where it is found only in neural tissue. In contrast to some other peptides, for example CCK [13], it does not have a dual distribution in endocrine and neural tissue in these two species. The marked difference between
291 rats and guinea pigs in the brightness of staining also suggests, as do the chromatographic data, that there may be sequence differences between the NMU peptides. It appears that neither of the two antisera used here recognised the guinea pig NMU very well in fixed tissue and we have therefore not commented in detail on the difference in distribution of NMU-LI between rats and guinea pigs. The major projection of the myenteric ganglia is to the external muscle, to stimulate or inhibit contraction of smooth muscle, whereas the major projection of the submucous ganglia is to the mucosa, to modify absorptive/secretory processes. In rat small intestine NMU-LI is found in neurones of both ganglionated plexuses. However, although substantial numbers of immunoreactive fibres are found in the mucosa, very few are found in the internodal strands of the myenteric plexus or in the muscle. It is not known whether the NMU-LI neurones in the myenteric plexus do not project to these sites or whether the levels of NMU-LI are below the threshold for detection with this technique. Within the ganglionated plexuses there is an uneven distribution of nerve terminals, suggesting specific sites of action in each ganglion. The most extensive innervation is of a small population of unstained submucous and myenteric neurones, usually only one or two per ganglion. It would be interesting to determine (e.g. by double staining methods) which substances these neurones contain. Studies over the last few years, in particular those using double staining immunohistochemical methods, have shown that the majority of enteric neurones in the guinea pig and many in the rat contain more than one peptide [14,15]. In guinea pigs it has been shown that different subpopulations of enteric neurones can be defined by the combination of peptides contained in them; these combinations can be used to identify neurones with different projections [15]. It is not yet known which subpopulations of enteric neurones contain NMU, or the projection pattern of these neurones. There is still very little information available on the physiological actions of NMU. In contrast to tachykinins, neither porcine NMU-8 nor NMU-25 have an effect on contraction of guinea pig ileum smooth muscle [7]. However, the present studies suggest that the guinea pig NMU differs from porcine NMU and the experiments should therefore be repeated using the native peptide sequence, as found in guinea pigs. In addition, there appear to be very few nerve fibres containing NMU-LI in the external muscle. The action of NMU on myenteric and submucous neurones, where extensive networks of nerve terminals are found, and the action on the epithelium, close to which many fibres are found, have not yet been reported. The biochemical and immunohistochemical data provided here comprise a basis for many further studies on the possible role of NMU in the enteric nervous system and the regulation of gastrointestinal function.
Acknowledgements J.R. Keast gratefully acknowledges the support of the Nuffield Foundation and the Gowrie Foundation in providing a travel grant. Amino acid analysis of standard
292
samples was carried out by Dr J. McMurray, MRC Laboratory of Molecular Biology. This manuscript could not have been produced without the hard work of Mrs B. Waters.
References 1 Minamino, N., Sudoh, T., Kangawa, K. and Matsu, O., Neuromedins: novel neuropeptides identified in porcine spinal cord. In C.M. Deber, V.J. Hruby and K.D, Kopple (Eds.), Peptides, Structure and Function. Proceedings of the 9th American Peptide Symposium, Pierce Chem. Co., Rockford, IL, 1985, pp. 643-646. 2 Minamino, N., Kangawa, K. and Matsuo, H., Neuromedin B: a novel bombesin-like peptide identified in porcine spinal cord, Biochem. Biophys. Res. Commun., 114 (1983) 541-548. 3 Minamino, N., Kangawa, K. and Matsuo, H., Neuromedin C: a bombesin-like peptide identified in procine spinal cord, Biochem. Biophys. Res. Commun., 119 (1984) 14-20. 4 Kangawa, K., Minamino, N., Fukuda, A. and Matsuo, H., Neuromedin K: a novel mammalian tachykinin identified in porcine spinal cord, Biochem. Biophys. Res. Commun., 114 (1983) 533-540. 5 Minamino, N., Kangawa, K., Fukuda, A. and Matsuo, H., Neuromedin L: a novel mammalian tachykinin identified in porcine spinal cord, Neuropeptides 4 (1984) 157-166. 6 Minamino, N., Kangawa, K. and Matsuo, H., Neuromedin N: a novel neurotensin-like peptide identified in porcine spinal cord, Biochem. Biophys. Res. Commun., 122 (1985) 542-549. 7 Minamino, N., Kangawa, K. and Matsuo, H., Neuromedin U-8 and U-25: novel uterus stimulating and hypertensive peptides identified in porcine spinal cord, Biochem. Biophys. Res. Commun., 130 (1985) 1078-1085. 8 Domin, J., Ghatei, M.A., Chohan, P. and Bloom, S.R. Characterization of neuromedin U-like immunoreactivity in rat, porcine, guinea-pig and human tissue extracts using a specific radioimmunoassay, Biochem. Biophys. Res. Commun., 140 (1986) 1127-1134. 9 Salacinski, P., Hope, J., McLean, C., Clement-Jones, V., Sykes, J., Price, J. and Lowry, P.J., A new simple method which allows theoretical incorporation of radioiodine into proteins and peptides without damage, J. Endocrinol., 81 (1979) 1-13l. 10 Costa, M. and Furness, J.B., Immunohistochemistry on whole mount preparations. In Cuello, A.C. (Ed.), Immunohistochemistry, Wiley, Chichester, pp. 373-397. 11 Stefanini, M., DeMartino, C. and Zamboni, L., Fixation of ejaculated spermatozoa for electron-microscopy, Nature (Lond.), 216 (1967) 173-174. 12 Goedert, M., Neurotensin in neuronal and endocrine tissues, PhD Thesis, University of Cambridge, 1984. 13 Dockray, G.J. and Gregory, R.A., Relations between neuropeptides and gut hormones, Proc. R. Soc. Lond. Ser, B, 210 (1980) 151-164. 14 Ekblad, E., Winther, C., Ekman, R., H~.kanson, R. and Sundler, F,, Projections of peptide-containing neurones in rat small intestine. Neuroscience, 20 (1987) 169-188. 15 Costa, M., Furness, J.B. and Llewellyn-Smith, l.J., Histochemistry of the enteric nervous system. In Johnson, L.R. (Ed.), Physiology of the gastrointestinal tract, 2nd edn., Raven, New York, 1987, pp. 1-40.