Anatomical evidence for interactions between somatostatin neurites in lamina II of the rat spinal cord

Peptides, Vol. 13, pp. 329-337, 1992

0196-9781/92 $5.00 + .00 Copyright © 1992 Pergamon Press Ltd.

Printed in the USA.

Anatomical Evidence for Interactions Between Somatostatin Neurites in Lamina II of the Rat Spinal Cord LAWRENCE

J. H A N N A N

AND

RAYMOND

H. H O ~

Department o f Cell Biology, Neurobiology and Anatomy, and Neuroscience Program, The Ohio State University, Columbus, O H 43210 R e c e i v e d 19 S e p t e m b e r 1991 HANNAN, L. J. AND R. H. HO. Anatomical evidencefor interactions between somatostatin neurites in lamina II of the rat spinal cord. PEPTIDES 13(2) 329-337, 1992.--Light microscopic analysis of adult and 10-14-day-old rat spinal cords suggested that somatostatin-immunoreactive (SOM-I) fibers apposed SOM-I cell bodies in lamina (L) II. Electron microscopic analysis of these relationships at both ages showed the presence of direct appositions between SOM-I fibers and SOM-I cells. However, synapse formation between SOM-I fibers and cells was observed only in the young rat. Similarly, synapses between SOM-I fibers and SOM negative cell bodies were only found in the young animal. Adjacent SOM-I perikarya directly contacted each other, but, again, membrane specializations were evident only in the young rat. Within L I of the adult dorsal horn, a SOM-I fiber directly apposed an unlabeled cell body. Despite analysis of serial sections through the apposition, no synaptic contacts were observed. Spinal cord Electron microscopy Nonsynaptic interactions

Somatostatin

Immunohistochemistry

Lamina I

Lamina II

with Zamboni's fixative (24) for 30 minutes. Each spinal cord was removed and immersed in Zamboni's fixative (4°C) for 2 6 hours. The spinal cord was stored in Sorenson's phosphate buffer containing 5% sucrose (4°C) overnight. Representative spinal cord segments (cervical enlargement, midthoracic cord, and midlumbosacral enlargement) were blocked and transverse sections were cut at 10 # m using a cryostat. Sections were mounted on chrome alum, gelatin-coated slides and processed by the indirect antibody peroxidase-antiperoxidase (PAP) technique of Sternberger (25) as described previously (8). Briefly, the tissue sections were rehydrated with a phosphate-buffered saline (PBS, pH 7.2) and incubated consecutively with the following antisera: 1) either anti-somatostatin- 14 serum, 11C (10), diluted 1:1000, or control serum, for 48 hours at 4°C; 2) sheep anti-rabbit IgG antiserum (diluted 1:300) for 1 hour at room temperature; and 3) a rabbit peroxidase-antiperoxidase (rPAP) complex (diluted 1:500) for 1 hour at room temperature. Sections were washed in borate-buffered saline (BBS, p H 8.2) after each incubation. After incubation with the rPAP complex, sections were immersed in a PBS solution containing 3,3' diaminobenzidine tetrahydrochloride (DAB) (0.3 g/600 ml) and 0.006%

C L I N I C A L studies have demonstrated somatostatin's analgesic effects in the h u m a n [see citations in (9)] and immunohistochemistry has revealed somatostatin-immunoreactive (SOM-I) neurons in lamina (L) II of the h u m a n spinal cord (21). The distribution of SOM-I perikarya and fibers is well known in the rat [e.g., (5,9,1 1,22)], and physiological studies have implicated SOM in nociception (12,16,27,28). However, little information is available on the connections between SOM neurites in L II and other systems, especially those which exhibit SOM i m m u noreactivity. For that reason, we sought to determine if SOMSOM interactions exist in L II, an area containing moderate to large numbers of SOM-I neurites (5,9,11,22). Since some peptidergic cell bodies are better visualized in tissues from young animals (6,9,15), we have examined 10-14-day-old rats as well as adults. METHOD

Light Microscopy Three 14-day-old rats were deeply anesthetized with sodium pentobarbital (IP) and sacrificed by an intracardiac perfusion

Preliminary observations from this study were presented at the 65th International Association of Dental Research, 1987 (7). This study was presented in partial fulfillment of the requirements for the degree Master of Science by L. J. Hannan to the Graduate School of The Ohio State University. Requests for reprints should be addressed to Dr. Raymond H. Ho, Department of Cell Biology, Neurobiology and Anatomy, The Ohio State University, 333 West Tenth Avenue, Columbus, OH 43210.

329

330

HANNAN AND HO

FIG. 1. (A) Differential interference photomicrograph of a transverse section from the cervical enlargement of a 14-day-old rat showing SOM-I neurons in lamina (L) II (arrows). Higher magnifications of neurons labeled b and c are provided in (B) and (C), respectively. The open arrow points laterally and the bar represents 20 urn. (B) A SOM-I neuron is apposed by SOM-I fibers (arrowheads). (C) A SOM-I fiber (arrowhead) can be traced to the surface of a SOM-I neuron where it branches. (D) Differential interference photomicrograph of L II SOM-I neurons from the adult with their associated SOM-I fibers (arrowheads). Bars in (B), (C), and (D) represent l0 um.

hydrogen peroxide for 3-10 minutes at room temperature. The sections were washed in BBS, dehydrated, and coverslipped with Permount. Prior to dehydration, selected sections were counterstained with cresyl violet. Adult Sprague-Dawley rats were pretreated with colchicine (200-360 vg, intrathecal, 24-30 hours) according to Yaksh and Rudy (29). Animals were deeply anesthetized with sodium pentobarbital (IP) and sacrificed by an intracardiac perfusion with PBS followed by either 4% paraformaldehyde (4°C) (N = 2) or Z a m b o n i ' s (24) solution (N = 2). Following perfusion fixation, spinal cords were removed and immersed in the same type of fixative for 2-6 h. Fifty- and 70-#m transverse or longitudinal sections from the cervical enlargement, the midthoracic cord, and the lumbosacral enlargement were cut using a freezing mi-

crotome and processed for immunohistochemistry according to Rosenthal and Ho (20). The anti-somatostatin-14 serum (11 C) was used at a dilution of 1:5000; the sheep anti-rabbit IgG antiserum (Antibodies Incorporated) was used at 1:300 and 1:600; and the rabbit peroxidase-antiperoxidase complex (Cappel Laboratories) was used at l: 1000. Control sera consisted of the primary antibody pretreated overnight (4°C) with 20 #g of SOM14 per ml of the diluted serum. Control sera did not produce immunostaining. The diluent for the antisera has been published (9), and 3,3' diaminobenzidine tetrahydrochloride was used as the chromogen. Tissue sections were analyzed and photographed using a bright field Zeiss Standard Microscope 14 and a Leitz Orthoplan 2 photomicroscope equipped with differential interference optics.

SOM-SOM INTERACTIONS IN SPINAL CORD

331

FIG. 2. An electron micrograph ofa SOM-I profile (open arrow) apposing a SOM-i L II cell body (N) in the adult rat. A second profile (*) exhibits a junctional complex (solid arrow) on a dendrite, but not on the soma. Bar represents 0.5 urn.

Electron Microscopy Two adult female (250-300 g) and two young (10 days old) Sprague-Dawley rats were used. All animals were deeply anesthetized with sodium pentobarbital and perfused transcardially with PBS followed by 4% paraformaldehyde-0.2% glutaraldehyde in 0.1 M phosphate buffer at 4°C for 30 minutes. The spinal cord was left in situ for 1 hour (h) (4°C) and then removed and placed in 4% paraformaldehyde fixative at 4°C for 1 h. Tissue blocks from the 5th lumbar spinal segment of the adult and the lumbosacral enlargement of the young rat were then rinsed in PBS and 60-#m transverse sections were cut on a vibratome. Sections were collected in PBS and exposed to a PBS-0.3% Triton X-100 solution for 30 min (4°C) and processed using the indirect antibody peroxidase-antiperoxidase method of Sternberger (25) as we have reported (20). Briefly, tissue sections were incubated consecutively in the somatostatin-14 antibody (11 C) (diluted 1: 1000) or the appropriate control serum (24 h, 4°C with agitation), the sheep anti-rabbit IgG serum (diluted 1:300, 1 h, room temperature with agitation), and finally the rabbit peroxidase-antiperoxidase complex (diluted 1:1000, 1 h, room temperature with agitation). Control sera were prepared as described above. Control sections were processed in parallel with the experimental sections and did not exhibit immunostaining. The sections were thoroughly washed with PBS after each incubation. All tissue sections were then exposed to PBS containing 0.05% 3,3' diaminobenzidine tetrahydrochloride and 0.006% hydrogen per-

oxide for 5-15 min at room temperature. The tissue sections were washed in PBS, postfixed for 20 minutes with 1.0% OsO4 (room temperature), rinsed with phosphate buffer (pH 7.2), dehydrated in a graded series of acetones, and flat embedded in Spurr's resin between two sheets of aclar. The 60-#m thick plastic sections were examined first with the light microscope and selection for electron microscopic examination was based on the staining intensity of cell bodies and the presence of SOM-I fibers that can be traced to SOM-I somata. Chosen sections were glued to beam capsules and 90-nm serial sections were cut on a Sorvall MT2-B ultramicrotome. The sections were picked up on Formvar-coated slot grids, and stained with lead citrate and uranyl acetate prior to examination with a Philips 300 electron microscope. The following criteria were used to delineate laminae I, II, and III. Lamina (L) I was most superficial in the dorsal horn and had fibers coursing predominately in a tangential direction. The outer portion of L II (IIo) was deep to L I and exhibited SOM-I fibers with only occasional SOM-I cell bodies. The inner portion of L II (IIi) was characterized by the presence of many cell bodies, some of which stained for SOM immunoreactivity. There were few myelinated fibers in L II. The presence of more myelinated fibers and the absence of L II SOM-I cell bodies helped identify L III. Labeled cell bodies that were previously selected with the light microscope were identified and serial sections through them were analyzed using the electron microscope. Other SOM-I cell bodies in these sections were also examined, giving a total of 7 adult and 10

332

HANNAN

i,~i¸~

~ii~

~'~ i~ iii? !!~ii~ill'

AND

~ ~i

FIG. 3. Electron micrograph o f a SOM-I soma (N) in L lli from a 10-day-old rat. A synaptic junction is present between a SOM-I profile with vesicles (block arrow) and the SOM-I soma. Note the dense cytoplasmic SOM-I labeling (arrows). Bar represents 0.5 #m.

HO

SOM-SOM

INTERACTIONS

IN SPINAL CORD

FIG. 4. Electron micrograph of an unlabeled s o m a in L lIi from a 10-day-old rat (N). This cell is postsynaptic to a SOM-I axon terminal (solid block arrow). Other SOM-I varicosities close to the same cell body are indicated (small arrows) and a SOM-I profile (open arrow) exhibits a synapse on a dendrite. Bar represents 0.5 urn.

333

334

HANNAN

FIG. 5. (A) Electron micrograph of two SOM-I somata (N) in L lli from a 10-day-old rat. These cells appose each other and a membrane specialization (white arrows) is apparent. Note the cytoplasmic labeling of cells (black arrows). Bar represents 2/am. (B) Higher magnification of the membrane specialization (block arrows) from an adjacent section. Bar represents 1 /am.

AND HO

SOM-SOM INTERACTIONS IN SPINAL CORD young cell bodies. Since the penetration of immunogens into the tissue sections was estimated to be 10 #m for the adult and the young rat, only varying portions of these cell bodies were labeled. Although our major objective was to determine whether the SOM-I fibers form synapses with SOM-I cell bodies, SOM-I profiles throughout L I - I I were also examined and electron micrographs of representative synapses were taken. RESULTS

Material Processedfor Light Microscopy Within the young spinal cords, SOM-I fibers cut in selected planes of section were present in L I - I I (Fig. 1A) of all representative spinal levels examined. SOM-I cell bodies were present throughout L IIi (Fig. 1A). They were easily detectable even at low magnifications and appeared as brown spherical profiles containing a negatively stained, rounded region interpreted to be the nucleus (Fig. 1A and B). Closely associated with most if not all SOM-I cell bodies were dot-like and fiberlike structures that immunostained more intensely than the associated cell body (Fig. 1A). Higher magnification indicated that many of the SOM-I fibers appeared to be on the external surfaces of the SOM-I cell bodies (Fig. IB). In addition, some SOM-I fibers branched as they approached a SOM-I cell body (Fig. 1C). As just described for the young spinal cords, SOM-I fibers were densest in L Iio of the adult spinal cord. Although the adult SOM-I cell bodies were harder to visualize, many have SOM-I fibers in close proximity to them (Fig. 1D) in a manner similar to that described for the neonate (Fig. 1B).

Materials Processedfor Sequential Light and Electron Microscopy The light microscopic morphology and distribution of SOM-I fibers and cell bodies in plastic sections from the adult and the young rat were comparable to those described above.

Electron Microscopy SOM-I cell bodies in the adult and young rats were identified by the immunohistochemical reaction product in the cytoplasm. In L II of the adult and young rats, many SOM-I profiles were separated from the surface of SOM-I cells by one or more unlabeled profiles. Serial section analysis indicated that some SOMI profiles were in direct contact with SOM-I cell bodies (Fig. 2). However, these fibers exhibited synaptic relationships with dendrites (Fig. 2). Synapses between these SOM-I profiles and SOMI cell bodies were only observed in the young rats (Fig. 3). Similarly, although SOM-I fibers apposed unlabeled L II perikarya, synapses between them were only apparent in the young rat (Fig. 4). Some SOM-I cell bodies in the adult and the young rats (Fig. 5) were in apposition along a portion of their membranes. However, membrane specializations were apparent only in the young rats (Fig. 5). As we have reported previously (20), SOM-I profiles in L II of the adult synapsed with unlabeled profiles. In addition, our observations for the young animals indicate that similar relationships are already present in the 10-day-old animal.

335 Within L I of the adult dorsal horn, a SOM-I profile was in apposition to an unlabeled cell body (Fig. 6C). Serial section analysis through this apposition did not reveal a definitive synaptic density. Since the adult spinal cord contains very few SOMI cells in L I, our samples for electron microscopy did not contain them. However, a SOM-I cell body was examined in L I of the young rat (Fig. 6A). The immunohistochemical reaction product within this cell either lined the external surface of vesicles, Golgi apparatus, and endoplasmic reticulum (Fig. 6B), or they appeared free in the cytoplasm unassociated with any organelles. DISCUSSION At the light microscopic level, the more intensely labeled SOM-I fibers intimately surrounding the lightly labeled SOM-I cell bodies draw attention to the latter in L II. Electron microscopic examination of such relationships revealed SOM-I fibers directly contacting SOM-I cell bodies in L II. Although these cell bodies have been reported by many laboratories [e.g., (5,9,11,22)], their associations with SOM-I fibers have not been illustrated. Unlike the synaptic relationships exhibited by SOM-I processes in L II [e.g., (18-20)], the predominant nonsynaptic association between SOM-I cell bodies and SOM-I fibers may indicate a lack of functional interaction between them. Alternatively, these relationships may represent anatomical substrates for nonsynaptic function. We propose that SOM may be released from SOM-I fibers into the extracellular space and be diffused to influence their targets, including L II SOM-I cell bodies. If this is true, one may expect SOM receptors on SOM-I cell bodies. Consistent with this proposal, SOM receptors have been demonstrated in the substantia gelatinosa of the rat (17). Unfortunately, the cellular location of these receptors was not determined. Thus, we may not assume that they are located on the cell bodies of SOM-I cells. The mechanism of nonsynaptic intercellular communication just proposed has been suggested for other systems. Within the spinal cord, substance P (SP)-immunoreactive fibers which exhibit granular vesicle-containing varicosities that lack specialized synaptic contacts (2) have been implicated as sites of SP release to influence nonsynaptic targets (2). Serotonin-I (3,14) and enkephalin-I (4) fibers in the spinal cord as well as various peptidergic fibers that surround trigeminal ganglion cell bodies (13) have been implicated in nonsynaptic functions. Based on these examples and others [see (4,26)], it is reasonable to propose that the SOM-I cell bodies in L II of the rat dorsal horn participate in nonsynaptic SOM-SOM interactions. Although nonsynaptic relationships are prevalent between SOM-I cells bodies and SOM-I fibers, one synapse between a SOM-I fiber and a SOM-I cell was observed in the young spinal cord. Since this is so rare, it is unlikely that synaptic interactions constitute a major mechanism for communications between the structures discussed. Despite analyzing serial sections, comparable synapses were not observed in the adult. The negative results from the adult must be interpreted with caution, however, as they may reflect a sampling limitation. Thus, it remains to be determined if these relationships are transient. The functional significance of the homologous nonsynaptic and synaptic relationships reported here are unknown, and, to our knowledge, only a few examples have been reported in other adult systems. A light microscopic investigation (23) suggested that methionine-enkephalin-Arg-Gly-Leu-I fibers appose methionine-enkephalin-Arg-Giy-Leu-I liquor-contacting neurons in

336

H A N N A N AND HO

FIG. 6. (A) Low magnification electron micrograph ofa SOM-I soma (N) in L I of the 10-day-old rat. Bar represents 1 um. (B) Higher magnification of the cell body in A showing immunostaining associated with vesicles (block arrows) and membranes (open arrows). Bar represents 0.25/~m. (C) An electron micrograph ofa SOM-I dendrite (open arrows) apposing an unlabeled L I soma (N) of the adult. A clearly defined postsynaptic density was lacking in this and adjacent serial sections. Bar represents 1 ~m.

S O M - S O M I N T E R A C T I O N S IN S P I N A L C O R D

337

L X of the rat spinal cord. Electron microscopic e x a m i n a t i o n o f these relationships is needed to d e t e r m i n e if there are direct m e m b r a n e contacts a n d if synapses exist. In a h y p o t h a l a m i c n e u r o e n d o c r i n e system (1), S O M - S O M interactions were observed in the periventricular nucleus. Nonsynaptic direct contacts between SOM-I dendrites a n d somata, between adjacent SOMI perikarya, a n d between SOM-I dendrites a n d between SOMI n o n t e r m i n a l axonal profiles were suggested. It was proposed that the homologous SOM-I cells m a y synchronize their electrical activity by ephaptic interactions via direct m e m b r a n e appositions without structural specializations. T h e authors (1) have also reported synaptic contacts between SOM-I neurites in the periventricular nucleus. T h e potential role o f such h o m o l o g o u s interaction in the n e u r o e n d o c r i n e system was discussed (1). T h e cell bodies o f origin of the SOM-I fibers apposing the SOM-I cells in L II reported here are u n k n o w n . Some o f these

appositions m a y be a n a t o m i c a l substrates for S O M - i n d u c e d analgesia. It is also possible that they m a y have other functions. This study has provided a n a n a t o m i c a l basis to investigate the m e c h a n i s m s of interactions between presumably n o n n e u r o e n docrine neurites that contain h o m o l o g o u s peptidergic-like substances. ACKNOWLEDGEMENTS This work was supported by NIH grants NS 23159 (R.H.H.) and DE 07155 (L.J.H.). We thank Dr. James S. King for his guidance in all phases of this investigation. We thank Dr. George F. Martin for his helpful suggestions in the preparation of this manuscript. The author is grateful for the technical assistance of Ms. Caryl S. Mather, Kathleen Wolken, and the photographic assistance of Mr. Karl Rubin. We thank Mr. Steve R. Zimmerman for his generous help with the computer system.

REFERENCES 1. Alonso, G.; Tapia-Arancibia, L.; Assenmacher, I. Electron microscopic immunocytochemical study of somatostatin neurons in the periventricular nucleus of the rat hypothalamus with special reference to their relationships with homologous neuronal processes. Neuroscience 16:297-306; 1985. 2. Barber, R. P.; Vaughn, J. E.; Slemmon, J. R.; Salvaterra, P. M.; Roberts, E.; Leeman, S. E. The origin, distribution and synaptic relationships of substance P axons in rat spinal cord. J. Comp. Neurol. 184:331-351; 1979. 3. Christenson, J.; Cullheim, S.; Grillner, S.; Hrkfelt, T. 5-Hydroxytryptamine immunoreactive varicosities in the lamprey spinal cord have no synaptic specializations--an ultrastructural study. Brain Res. 512:201-209; 1990. 4. Cuello, A. C. Peptides as neuromodulators in primary sensory neurons. Neuropharmacology 26:971-979; 1987. 5. Dalsgaard, C.-J.; Hrkfelt, T.; Johansson, O.; Elde, R. Somatostatin immunoreactive cell bodies in the dorsal horn and the parasympathetic intermediolateral nucleus of the rat spinal cord. Neurosci. Lett. 27:335-339; 1981. 6. DiTirro, F. J.: Martin, G. F.; Ho, R. H. A developmental study of substance-P, somatostatin, enkephalin, and serotonin immunoreactive elements in the spinal cord of the North American opossum. J. Comp. Neurol. 213:241-261; 1983. 7. Hannan, L. J.; Ho, R. H. Potential anatomical substrate for somatostatin (SOM*) induced analgesia. J. Dent. Res. 66:157; 1987. 8. Ho, R. H. WidespreaddistributionofsubstancePandsomatostatinimmunoreactive elements in the spinal cord of the neonatal rat. Cell Tissue Res. 232:471-486; 1983. 9. Ho, R. H. Somatostatin immunoreactive structures in the developing rat spinal cord. Brain Res. Bull. 21:105-116; 1988. 10. Ho, R. H.; DePalatis, L. R. Substance P immunoreactivity in the median eminence of the North American opossum and domestic fowl. Brain Res. 189:565-569; 1980. 11. Johansson, O.; Hrkfelt, T.; Elde, R. P. Immunohistochemical distribution of somatostatin-like immunoreactivity in the central nervous system of the adult rat. Neuroscience 13:265-339; 1984. 12. Kuraishi, Y.; Hirota, N.; Sato, Y.; Hino, T.; Satoh, M.; Takagi, H. Evidence that substance P and somatostatin transmit separate information related to pain in the spinal dorsal horn. Brain Res. 325: 294-298; 1985. 13. Kuwayama, Y.; Stone, R. A. Neuropeptide immunoreactivity of pericellular baskets in the guinea pig trigeminal ganglion. Neurosci. Lett. 64:169-172; 1986. 14. Maxwell, D. J.; Leranth, C.; Verhofslad, A. A. J. Fine structure of serotonin-containing axons in the marginal zone of the rat spinal cord. Brain Res. 266:253-259; 1983.

15. Pickel, V. M.; Sumal, K. K.; Miller, R. J. Early prenatal development of substance P and enkephalin-containing neurons in the rat. J. Comp. Neurol. 210:411-422; 1982. 16. Randic, M.; Miletic, V. Depressant actions of methionine-enkephalin and somatostatin in cat dorsal horn neurons activated by noxious stimuli. Brain Res. 152:196-202; 1978. 17. Reubi, J. C.; Maurer, R. Autoradiographic mapping ofsomatostatin receptors in the rat central nervous system and pituitary. Neuroscience 15:1183-1193; 1985. 18. Ribeiro-da-Silva, A.; Cuello, A. C. Substance-P and somatostatinlike immunoreactivities in synaptic glomeruli of the rat substantia gelatinosa, as revealed by bi-specific monoclonal antibodies. In: Henry, J. L., et al., eds. Substance P and neurokinins. New York: Springer Verlag; 1987:313-317. 19. Ribeiro-da-Silva, A.; Cuello, A. C. Ultrastructural evidence for the occurrence of two distinct somatostatin-containing systems in the substantia gelatinosa of the rat spinal cord. J. Chem. Neuroanat. 3: 141-153; 1990. 20. Rosenthal, B. M.; Ho, R. H. An electron microscopic study of somatostatin immunoreactive structures in lamina II of the rat spinal cord. Brain Res. Bull. 22:439-451; 1989. 21. Schoenen, J.; Lostra, F.; Vierendeels, G.; Reznik, M.; Vanderhaeghen, J. J. Substance P, enkephalins, somatostatin, cholecystokinin, oxytocin, and vasopressin in human spinal cord. Neurology 35:881-890; 1985. 22. Schroder, H. D. Somatostatin in the caudal spinal cord: An immunohistochemical study of the spinal centers involved in the innervation of pelvic organs. J. Comp. Neurol. 223:400-414; 1984. 23. Shimosegawa, T.; Koizumi, M.; Toyota, T.; Goto, Y.; Yanaihara, C.; Yanaihara, N. An immunohistochemical study of methionineenkephalin-Arg6-Gly7-Leu8-1ike immunoreactivity-containing liquor-contacting neurons (LCNs) in the rat spinal cord. Brain Res. 379:1-9; 1986. 24. Stephanini, M.; DeMartino, C.; Zamboni, L. Fixation of ejaculated spermatozoa for electron microscopy. Nature 216:173-174; 1967. 25. Sternberger, L. A. Immunocytochemistry. Second edition. New York: John Wiley and Sons; 1979. 26. Vizi, E. S.; Labos, E. Non-synaptic interactions at presynaptic level. Prog. Neurobiol. 37:145-163; 1991. 27. Wiesenfeld-Hallin, Z. Intrathecal somatostatin modulates spinal sensory and reflex mechanisms: Behavioural and electrophysiological studies in the rat. Neurosci. Lett. 62:69-74; 1985. 28. Wiesenfeld-Hallin, Z. Substance P and somatostatin modulate spinal cord excitability via physiologically different sensory pathways. Brain Res. 372:172-175; 1986. 29. Yaksh, T. L.; Rudy, T. A. Chronic catheterization of the spinal subarachnoid space. Physiol. Behav. 17:1031-1036; 1976.