Nruroscrence Vol. 36, No. Printed m Great Britain
1. pp. 73-81,
0306-4522,90 $3.00 + 0.00 Pergamon Press plc IC\1990 IBRO
1990
CHANGING PATTERNS OF C-FOS INDUCTION SPINAL NEURONS FOLLOWING THERMAL CUTANEOUS STIMULATION IN THE RAT
IN
S. WILLIAMS,* G. I. EvANt and S. P. HUNT* *MRC Molecular Neurobiology Unit, University of Cambridge Medical School, Hills Road. Cambridge CB2 2QH, U.K. tInternational Cancer Research Fund Laboratories, Bartholomew Close, London ECIA 7BE, U.K. Abstract-Patterns of neuronal activity in the lumbar spinal cord of the anaesthetized rat were mapped by immunocytochemical localization of the c-fos gene product, Fos protein, at different timepoints following brief noxious stimulation of one hindpaw (20 s immersion in water at 52°C). After 2 h, Fos-immunoreactive neurons were seen mainly in the superficial laminae of the ipsilateral dorsal horn, with maximum somatotopic organization in lamina II. Subcutaneous injection of dilute formalin produced a similar pattern of immunostaining at 2 h, with a greater proportion of Fos-positive neurons in laminae III-VIII than with heat. With a survival time of 8 h following formalin injection, Fos immunoreactivity was virtually absent from the spinal cord. Eight hours after heat stimulation, however, the superficial pattern had given way to the appearance of a population of immunoreactive cells in the deeper laminae. The pattern of this “second wave” of heat-induced Fos-positive cells had a marked contralateral component, and was still present after 24 h, having become even more diffuse and symmetrical. The number of Fos-positive cells seen at 8 h was increased by local anaesthetic blockade of the peripheral nerve after stimulation, and reduced by continuous barbiturate anaesthesia. These findings suggest that the early stages of thermal injury trigger a complex pattern of molecular events within the spinal cord, which are initially monosynaptic and closely related to primary afferent terminal depolarization, and in the longer term the result of an induced pattern of synaptic activity set up within the spinal cord.
EXPERIMENTAL
The proto-onco~ene c-fos belongs to a large family of “immediate early onset” genes, activated in various
PROCEDURES
Male Sprague-Dawley rats (Tuck’s, 200-300 g) were deeply anaesthetized by intraperitoneal injection of 4 ml/kg Equithesin, prepared according to the following formula: 81 ml sodium pentobarbitone 60 mg/ml (“Sagatal”), 21.25 g chforat hydrate, 10.63 g ma~esium sulphate, 198 ml propylene glycol, $0 ml ethanol, distilled water to 500 ml. The left hindpaw was stimulated either by immersion for 20s in a gently-stirring water bath at 52°C. or by subcutaneous injection of 50~1 4% formalin into the distal paw pad. Anaesthesia following a single dose lasted approximately 2 h, but in some cases to achieve continuous anaesthesia, two or three injections of 2 ml/kg Equithesin were given when the cornea1 reflex returned. Local anaesthetic blockade of the sciatic nerve was produced by injection of 0.3 ml of 0.5% bupivacaine into the sciatic notch of the pelvis. At intervals of 30 min, 1, 2, 4, 8, 16 and 24 h after each stimulus, animals were terminally reanaesthetized with saturated chloral hydrate solution (1 ml, i.p.) and perfused intracardially with 200 ml normal saline followed by 400 ml freshly-depolyme~zed paraformaldehyde, 4% in 0.05 M sodium phosphate buffer (PB), pH 7.4. Spinal cords were post-fixed in the latter solution for 12 h, then washed overnight in 0.1 M PB containing 30% sucrose and 0.01% sodium azide. Forty-micron frozen sections taken in the tranversc plane from spinal cord segments Ll-L6, and in coronal and parasagittal planes from all segments caudal to T12, were incubated in rabbit antiserum to a synthetic peptide common to all c-fos and v-fos proteins,14.diluted 1: 3000 or 1: 10,000, and processed with an avidin-biotin complex-horseradish peroxidase (ABC-HRP) kit (Vector). The location of Fos-immunoreactive neurons was plotted with a camera lucida attached to a Leitz Orthomat photomicroscope viewing at x25. From approximately 100 sections from segments L3 to L4, three were chosen for maximum number of labelled neurons and viewed against a
cell types within minutes of stimulation by growth factors or neurotransmitter substances.‘r Following the suggestion that certain of these genes might play a role in the establishment of long-term functional change in the nervous system,3.” it was found that the c-J& could be induced synaptically in the spinal cord of the anaesthetized rat near the central terminals of C-fibres stimulated by noxious radiant heat or mustard oil applied to the skinI Fos immunoreactivity in this model appears to be restricted to cells with neuronal morphology under light and electron microscopy,‘4 and some can be retrogradely labelled from projection areas such as the thalamus and brainstem.lK It was assumed that these molecular events may be causally related to the changes in spinal cord physiology that follow peripheral injury or tetanic stimulation of C-fibres, which can be both complex, and persistent for many hours or days following the initial stimu1us.5.3’J6.38Here 3 we extend the original observations and report on the patterns of Fos protein immunoreactivity which evolve over many hours following brief sensory stimulation. A preliminary report of these findings has been made elsewhere.32 -Abhretkfions:
ABC, avidin-biotin complex; HRP, horseradish peroxidase; mRNA, messenger RNA; NMDA, N-methyl-o-aspartate; PB, phosphate buffer; SP, substance P. 73
s
74
wIl.I.I~MS (‘I
standard plan with the boundartes of Rexcd’s lammae adapted from Paxinos and Watson.?” For quantitative analysis all Fos-positive cells were counted separately regardless of staining intensity, on the left (ipsilateral to the stimulus) and right sides, and in superficial (Rexed’s laminae I-II) and deeper (III-X) zones of the same representative sections.
ut.
.
5H
RESULTS
Choice of’ stimuli
Surgical exposure of a nerve alone provokes massive cfos induction in the spinal cord, precluding electrical stimulation techniques for the purposes of this study. I4 Immersion of the hindpaw of the intact animal in heated water was easily and accurately administered and gave consistent results, and was therefore selected as a suitable physical stimulus for these experiments. As for a chemical stimulus, there had been considerable variation in response to cutaneous mustard oil depending on the age of the animal, probably due to differences in epidermal thickness. In contrast, plantar injection of dilute formalin is a commonly-used artificial noxious stimulus’ in which the site, dose and timing can be reliably controlled, and produces a characteristic electrophysiological response in the spinal cord which has recently been investigated.* Basal levels of Fos within the spinal cord were not revealed by these techniques, and staining was seen only following stimulation, almost exclusively in neuronal nuclei. Cells stained to variable degree regardless of position, so that densely-immunoreactive and faintly-outlined nuclei were juxtaposed. Use of higher concentrations of the polyclonal primary serum resulted in some non-specific background staining of gray, and cellular elements of white matter (Figs 6 and 7) and occasionally the cytoplasm of strongly positive cells (Fig. 7D).
1H
Thermal cutaneous stimulation
After the heat stimulus (immersion of the left hindpaw for 20 s in stirred water at 52”C), Fos immunoreactivity appeared ipsilaterally in laminae I-III and V within 30 min (n = 3) and laminae VI, VII and X by 1 h (n = 3) (Fig. 1). A smaller number of Fos-positive cells was distributed similarly on the right-hand side of the cord. After a survival time of 2 h, the Fos-positive cell count had reached a peak in superficial laminae I and II on the left, ipsilateral to the stimulus, although the number of stained contralateral cells had fallen (Fig. 4). Dense staining in lamina 11 was restricted to cord segments L2-L4, but positive neurons were seen in laminae I, III-VIII and X in transverse sections from segments Ll to L6. Longitudinal sections of spinal cord showed that an increasingly attenuated signal was present as far rostrally as Ll in lamina I, and TlO in lamina V. After 4 h, the number of immunoreactive cells in the most superficial laminae was reduced, whereas the
Fig. 1. Camera lucida drawings of rat spinal cord, fourth lumbar segment in transverse section, immunostained for Fos 0.5524 h after heat stimulation of the left hindpaw. Each circle represents three positive nuclei on a composite of three sections. Filled circles show the position of heavily stained neurons.
signal in V, VI, VII and X appeared to be increasing (n = 3). Contralaterally, a few positive cells had appeared in the deeper laminae. At 8 h (n = 6) this change had progressed with marked shift of positive cells to deep laminae, especially V. At 16 h (n = 3) there was diffuse bilateral induction of c-fos in all laminae except IX, and by 24 h the pattern was almost symmetrical, with small numbers of positive cells appearing even in the contralateral superficial laminae (Figs 1 and 4).
Changing
patterns
of C+OS expression
No Fos immunoreactivity was seen in the spinal cord following immersion of the hindpaw in water at 42 C (II = 2 at each timepoint). In the cord of a heatstimulated animal, therefore, ipsilateral Fos immunoreactivity attenuated in the superficial laminae after 2 h, but gradually increased in the deep Iaminae until at 8 h it appeared that many more deep cells were positive than at 2 h. particularly in lamina V (Figs I and 4). Secondly, Fos-positive cells were seen from 4 h onwards in the contralateral dorsal horn which had not been present at 2 h (Fig. 4). By I6 h a massive increase in the Fos-positive cell count had occurred bilaterally in laminae VI-VIII. and at 24 h there was a very diffuse bilateral pattern bearing little relation to the relatively restricted signal seen earlier (Fig. I). Stimulation
by ,formnlin
Two hours after subcutaneous formalin injection into thcdistal paw pad (tz = 3), there was Fos immunoreactivity in ipsilateral neuronal nuclei throughout the gray matter of the lumbar spinal cord segments From L2 to LS, most strongly in laminae I-VII (Fig. 2). The signal was maximal in a short segment of cord near the L3/L4 junction, but absent from lamina II elsewhere and increasingly attenuated in other laminae with increasing rostra1 or caudal distance from the L3/L4 junction. By comparison with that induced by heat stimulation of the whole foot, the formalin pattern was also more restricted mediolaterally in lamina II, although there was more signal in the deeper laminae V-VII at 2 h, especially in lateral lamina V, than with heat stimulation. Eight hours after formalin stimulation, however (n = 3). there were only a very few faintly-immunoreactive nuclei in laminae I---VI of the ipsilateral dorsal horn, and no contralateral signal (Fig. 2). Insertion of a dry needle (n = 2), or injection of 50~1 normal saline (n = 3), resulted in Fos immunoreactivity in one or two neurons per 40-l.lrn section in a few sections from the fourth lumbar cord segment.
2H
8H Fig. 2. Rat lumbar spinal cord 2 and 8 h following subcutaneous injection of formalin into the distal pad of the left hindpaw.
75
in spinal neurons
A large number of lamina II neurons were activated by heat stimulation of the entire hindpaw surface, in a pattern extending rostrocaudally for several segments and mediolaterally across most of the width of the lamina. In contrast, formalin injection produced a signal in a very restricted zone of lamina II, where the pattern of Fos-positive neurons occupied less than half of the width of the dorsal horn for approximately 200 pm in the region of the L3/L4 junction. Nonetheless, the number of stained cells in the deeper Iaminae at 2 h, centred around the patch of Fos-positive neurons in lamina II, was greater with formalin despite the fact that a much smaller area of skin had been directly stimulated. Neurogenic oedema usually developed within 1 h of stimulation by heat, more prominently in the loose dorsal tissue than on the plantar surface of the paw, and less so with the localized formalin stimulus. The oedema following heat stimulation had increased after 8 h and was still marked at 24 h, while that following formalin injection appeared to have resolved by 8 h. Animals with oedema invariably developed a marked signal in the spinal cord. Occasionally oedema did not develop but this did not preclude spinal induction of c-fffs. Thermal cutaneous stimtdation ,folfowed andlor sustained general anaesthesio
by
local
The following experiments were designed to examine the “second wave” of signal seen bilaterally in the deeper laminae, V-VII and X, 8 h after heat stimulation. After bupivacaine blockade of the sciatic nerve, the apparent motor and sensory deficit was similar to that seen after interruption of neurotransmission by sciatic nerve division or crush, and normal sensation appeared to be restricted to the area supplied by the saphenous nerve.’ Partial hindlimb paralysis, with extension of the leg but no active flexion of the leg or foot, and apparent anaesthesia of an area including the lateral three digits, occurred within 10 min and lasted for 3-4 h. During this period there was no hyperventilation. vocalization or attempted withdrawal after any degree of stimulation of the lateral three digits, but gentle seizing of the medial digit with forceps produced one or more of these signs of nociceptive activity. The second digit was sometimes included in the region of complete anaesthesia, and sometimes not. If not, a firm pinch was required to elicit a response. No Fos immunoreactivity was seen in the lumbar cord following delivery of the block alone (n = 3). When the ipsilateral sciatic nerve had been blocked I5 min before heat stimulation (n = 3), neuronal c$s induction in lamina II after 2 h was restricted to a very short length of cord in the LI!L2 region corresponding to the saphenous nerve projection.‘” In segments L?-L6 of these animals the signal in the deeper laminae was relatively attenuated.
s. WILL
76
LAMS el
Ui
The number of Fos-positive cells seen at 8 h was increased when the ipsilateral sciatic nerve was blocked with bupivacaine 1 h following heat stimulation of the hindpaw (fz = 6) (Figs 3 and 5). When these animals had recovered from general anaesthesia their gait and posture confirmed successful blockade of at least motor neurotransmission in the sciatic nerve. Conversely, animals kept anaesthetized with repeated doses of Equithesin for the full 8 h between stimuIation and perfusion (n = 3) had fewer positive neurons than those which had been allowed to recover. Those receiving both local and continuous general anaesthesia after stimulation (n = 3) developed Fos immunoreactivity in a similar pattern and number of cells to animals given only the usual initial single dose of Equithesin and no additional anaesthetic (Figs 3 and 5). DISCUSSION
Fos protein appeared in neurons of the rat spinal cord within 30min, following noxious thermal or chemical cutaneous stimulation. At 2 h the pattern of positive neurons from either mode of stimulation was topographically consistent with induction by synaptic activation by primary afferent nerve fibres, possibly in a monosynaptic fashion.29 We previously demonstrated that the appearance of Fos protein within the spinal cord was dependent upon activation of smalldiameter high-threshold primary afferents, and that
LAMINAE I/ -X
Fig. 4. Maximal numbers of Fos-positive neurons, in superficial (I-II) and deep (III-X) laminae, on the left and right sides, per section of rat lumbar spinal cord at different times following heat stimulation of the left hindpaw. Average of three representative sections each from three animals. Error bars = S.E.M.
the pattern of labelling within the first few hours following stimulation closely followed the pattern of termination of particular sensory afferents within the dorsal hom.i4 Activation of cutaneous at&rents resulted in labelling within their areas of te~ination in laminae I, II and V, while Fos-positive neurons were largely restricted to laminae I and V following activation of muscle or visceral afferents,33 which are known to terminate almost exclusively within these laminae.6 This suggested a close, perhaps monosynaptic Iink between primary afferent activation and c-fis induction within neurons. However, we report here that at longer survival times, a “second wave” of c-jbs labelling can occur in neurons distant from areas of high-threshold primary afferent termination. Eficts
LA
LA + GA Fig 3. Rat lumbar spinal cord 8 h following heat stimulation of the left hindpaw, with local anaesthetic blockade of the ipsiiateral sciatic nerve beginning 1 h post-stimulation (LA), continuous general anaesthesia (GA), and both local and continuous general anaesthesia (LA + GA).
of heat
With heat, the number of Fos-positive neurons peaked twice within the first 24 h. The first peak at 2 h, as originally reported in Hunt et ~1.,‘~was seen principally in the superficial laminae, almost exclusively ipsilateral to the stimulus, while a “second
Fig. 5. Maximal numbers of Fos-positive neurons per section of rat lumbar spinal cord 8 h following heat stimulation of the left hindpaw, with and without local anaesthetic blockade of the sciatic nerve (LA), continuous general anaesthesia (GA), or both (the same sections as those shown in Fig. 3). Average of 10 representative sections. Error bars = S.E.M.
Changing patterns of c,fos expression in spinal neurons wave” of Fos protein synthesis occurred bilaterally in the deep laminae over a much longer period of time, commenting at about 8 h, reaching a maximum number of cells at around 16 h, and still clearly present 24h post-stimulation (Figs 1 and 4). Polwynaptic
mechanisms
It would be difficult to maintain that the activation of deeper neurons was the result of a monosynaptic event. Stimulation with heat at 52’C, and probably with formalin, activates primary afferents which are almost exclusively unmyelinated C-fibres in the rat,17 80-85% of which are nociceptive.‘b These, along with the small myelinated high-threshold A-delta sensory fibrcs, terminate ipsilaterally largely within laminae I and II, and occasionally IV and V.” Yet Fos-positive cell bodies are found at 16-24 h within laminae other than these. and the delayed bilateral expression of c-f& products following unilateral stimulation suggests that polysynaptic pathways are involved in the activation of deeper neurons. For many neurons. we believe that the evolving bilateral pattern of c,fos expression within the deeper laminae of the spinal cord occurs largely independently of continuous primary afferent input (see below), and is probably polysynaptic in origin.” Thus, while the early pattern of neuronal excitation could well be the result of monosynaptic activation, following release from primary afferents of neurotransmitter candidates such as substance P (SP),’ synaptic contact with primary afferents could not explain the distribution of induced gene expression seen later. This is presumably the result of patterns of neuronal activity set up in intra- or intersegmental pathways, or possibly involving the activation of pathways descending from the brainstem. In the lumbar cord of the spinalized rat, about half the ccfls in laminae IV-VI which have a direct ipsilateral primary afferent input can be inconsistently excited by stimulation of the contralateral sciatic nerve at A-fibre strength, via a segmental pathway that tends to be inhibited by stimulation at C-fibre strength, and is hence thought to be polysynaptic.“’ Such a shortlatency mechanism, however, would not explain the occurrence of neuronal biosynthetic activity & ~UWO after 8 h or more. and a descending system, such as that postulated to account for diffuse noxious inhibitory controls,” remains a likely mediator of the delayed effects of noxious stimulation on spinal c-fijs expression. The appearance of c$is messenger RNA (mRNA) has from previous work with cell lines been regarded as a transient event, translating into protein with a substantially longer half-life.” Yet the appearance of Fos protein bilaterally over an extended period of time, in an apparently (from the distributions) novel population of cells would perhaps argue against this being the universal mechanism. Indeed, using in situ hybridization histochemistry, mRNA for c+,c is found to be heavily expressed in ipsilateral neurons of
77
iaminae I and II from 5 to 120min after cutaneous heat stimulation, before decaying over the next 120 min.35
Formalin injection resulted in a single peak of cell labelling at 2-4 h which appeared to fade uniformly thereafter, having all but disappeared at 8 h (Fig. 2). The single injection of 50~1 formalin into the distal paw pad was sufficient to induce c+s in neurons within al1 laminae of the lumbar spinal cord. In the region of cord with maximal numbers of Fos-positive neurons, transverse sections contained at least as many stained cells following formalin injection. with a greater effect in the deeper Iaminae III--X. as with immersion of the entire paw in water at 52, C. This may indicate that formalin injection activates a greater number of neurons per unit of skin area stimulated than heat, or that it activates some or all of them more intensely. Formalin injection may activate deep tissue nociceptors which cutaneous heating fails to stimulate effectively. In an electrophysiological study, Dickenson and Sullivan’ report that dorsal horn neurons activated exclusively by innocuous stimuli were not excited by formafin injection into their receptive fiefd. However, the appearance of Fos-positive neurons within laminae III-IV following formalin injection suggests that the abnormal excitation of normally non-nociceptive afferents. which are thought to be the sole input to the intermediate dorsal horn laminae,2’ may be at least partly responsible for the early pattern of labelled neurons within the dorsal horn. It may be that non-specific activation of many fibrc types by formalin, including non-nociceptive afferents, results in inhibition of the later developing pattern of c+s induction.
To test our hypothesis that the “second wave” of c+s expression in dorsal horn neurons was a delayed response to the initial stimulus and not due to continued activity in heat-sensitized aKerents, a group of animals was kept continuously anaesthetized to prevent stimulation of the heated paw by volunrary movement. The signal was indeed reduced. but these animals had received in total two to three times the dose of Equithesin required to produced anaesthesia initially. Equithesin contains sodium pentobarbitone, and was thought likely to have affected the results by inhibition of neurotransmission within the spinal cord, as low doses of barbiturate drugs seem to do in primates.” Sodium pentobarbitone abolishes the c~fis expression induced in mouse brain by Metrazole.Z” The effect of local anaesthetic applied to the sciatic nerve 1 h after heat stimulation suggested that ongoing primary afferent activity, the result of tissue damage and inflammation, was not necessary for the appearance of the “second wave” of
78
S. WILLIAMS
('I~1
Fig. 6. Photomicrogmphs of one 40qm section from LJ cord segment, 2 h after heat stimulation of the left hindpaw. (A) Low power view showing Fos immunoreactivity concentrated ipsilaterally in superficial laminae, at higher power in B. with a few positive cells in lamina V (D). No significant staining is seen contralaterally in superficial (C) or deep laminae (E). Both scale bars == 100 /urn.
Fig. 7. Sixteen hours after heat stimulation (A), Fos-positive cells are much sparser ipsilaterally in laminae I and II (B) than at 2 h. but diffuse c-fos induction has occurred contralaterally in superficial laminae (C). as well as laminae III OXon both the stimulated (D) and unstlmulated side (E). Reaction product is visible in the cytoplasm of an intensely stained cell (D: large arrowhead), but many nuclei at this stage are only faintly positive. with a granular appearance (D, small arrowheads). Both scale bars = 100 /in?.
thermally-induced Fos labelling. Bupivacaine was shown by observation and behavioural tests in conscious animals, and Fos immunohistochem~stry following heat stimulation, to produce effective blockade of the ipsilateral sciatic nerve for up to 4 h, although primary afferent input after this period could have contributed towards the observed late labelling. The increasing signal in superficial laminae suggests that c+s induction is under tonic inhibitory control by primary afTerent activity. An increase in the “spontaneous” activity within lamina I and the substantia gelatinosa has been reported following surgical deafferentation,’ which suggests that continuous activity within superficial laminae could result in the induction of C~$Xin local neurons. Since 2 h was required for the earliest signai to become maximai, it was suggested that the “second wave” might have followed some delayed afferent activity by a similar period, despite the fact that the diffuse, bilateral pattern of Fos-positive neurons emerging at 8 h was quite distinct from the arrangement of central termini of primary afferent fibres. With sciatic blockade starting at 1 h. the increased signal 8 h after heat stimulation suggests that activity in the intact sciatic nerve after thermal injury may tend to reduce the “second wave” of c-f0.r induction, rather than contribute towards it. Finally, animals which received both local and continuous general anaesthesia developed a pattern at 8 h similar to that of control animals, and it was concluded that the “second wave” of c+s induction was independent of continued activity in the sciatic nerve. This has the important implication that c-&s induction may occur as the result of transmission from neuron to neuron within the CNS, not purely as the result of synaptic contact with an active primary afferent terminal. To date, poiysynaptic neuronal activation of c;fos has been variably demonstrated
only by direct electrical stimulation of the cerebral cortex, or other means of producing seizure’-
The nature of the stimulus required for the induction of c-&s in neurons is unknown. Some evidence suggests that glutamate acting at the Ymethyl-D-aspartate (NMDA) receptor” or the peptide SP,4 both thought to be primary afferent neurotransmitters,2R may be effective stimuli, at least in primary culture. Alternatively, ~f0.r induction may be related entirely to the degree of membrane depolarization and the entry of calcium either through ligand or voltage-gated channels.” What is the significance of ~fi)s activation? Early onset genes are thought in many cases to code for transcription factors controlling the expression of batteries of downstream genes. Fos protein has been shown to form a heterodimer with the c-jun product, which then effects transcription through an identified c7jun/AP1 DNA binding domain on other genes.‘” Both c-&n A and c-f0.r genes are rapidly induced in dorsal horn neurons following noxious stimulation.” The genes affected by C-$X activation have yet to be identified. However, 8-24 h following injection of formalin or complete Freund’s adjuvant into one hindpaw, the levels of mRNA encoding SP,‘” enkephalin’” and dynorphin13,zh rise dramatically within the dorsal horn. These events may be dependent on c-$&r activation. The function of increased peptide synthesis within the cord is unclear but, as within the “second wave” of c+.s activation following brief heat stimulation, could reflect the activity 01 an endogenous pain control system brought into play following particular types of noxious stimulation. Acknowledgements--We wish to thank Mrs L. A. McNaughton for her assistance. SW. holds a Wellcome Clinical Research Training Fellowship, and received support for part of this work from the Sir Halley Stewart Trust.
REFERENCES
I. Alreja M., Mutalik P. G., Nayar U. and Manchanda S. K. (1984) The formalin test: a tonic pain model in the primate. Pain 20, 97-105.
2. Basbaum A. I. and Wall P. D. (1976) Chronic changes in the response of cells in adult cat dorsal horn following partial deaffe~nt~ation: the appearance of responding cells in a previously non-responsive region. Bruin Res. 116, 184-204. 3. Berridge M. (1986) Second messenger dualism in neuromodulation and memory. Nature 323, 294-295. 4. Bigot D., Evan G. and Hunt S, P. (1989) Rapid induction of c-$13 protein in gliai and neuronal cells studied itzz&w. Neurosci. Left., Suppl. 36, 575.
5. Cook A. J., Woolf C. J., Wall P. D. and McMahon S. B. (1987) Dynamic receptive field plasticity in rat spinal cord dorsal horn following C-primary afferent input. Nurure 325, 151-153. 6. De Groat W. C. (1986) Spinal cord projections and neuropeptides in viscera1 afferent neurons. In Kweruf Sensaiiun (eds Cervero F. and Morrison J. F. is.>: Prog. Brain Res. 67, 165-187. 7. Devor M., Schonfeld D., Seltzer Z. and Wall P. D. (1979) Two modes of cutaneous reinnervation following peripheral nerve injury. J. camp. Neuroi. 185, 211-220. 8. Dickenson A. H. and Sullivan A. F. (1987) Subcutaneous formalin-induced activity of dorsal horn neurons in the rdt: differential resnonse to an intrathecal opiate administered pre or post formalin. Pain 30, 349-360. 9. Duggan A. W.: Morton C. R., Zhao Z. 6. and Hendry I. A. (1987) Noxious heating of the skin releases immunoreactive substance P in the substantia aelatinosa of the cat: a studv with antibodv micronrobes. Bruin Res. 403, 345.-349. 10. Fitzgerald M. (1982) The con&lateral input to the dorsal kern of the spinal cord-in the decerebrate spinal rat. Brain Res. 236, D-287.
11. Goelet P., Casteltucci V. F., Schacher S. and Kandel E. R. (1986) The long and the short of long-term memory--a molecular framework. Nature 322, 41%-422.
Changing
patterns
of c:fos expression
in spinal
neurons
81
12. Greenberg M. E., Greene L. A. and Ziff E. B. (1985) Nerve growth factor and epidermal growth factor induce rapid transient changes in proto-oncogene transcription in PC12 cells. J. biol. Chem. 260, 14,101~14,110. 13. Hollt V.. Haarmann I., Millan M. J. and Herz A. (1987) Prodynorphin gene expression is enhanced in the spinal cord of chronic arthritic rats. Neurosci. Lett. 73, 90-94. 14. Hunt S. P., Pini A. and Evan G. (1987) Induction of c-/&y-like protein in spinal cord neurons following sensory stimulation. Nurture 328, 632-634. 15. Le Bars D., Dickenson A. H. and Besson J. M. (1979) Diffuse noxious inhibitory controls (DNIC)&I. Effects on dorsal horn convergent neurons in the rat. Pain 6, 2833304. 16. Lynn B. and Carpenter S. E. (1982) Primary afferent units from the hairy skin of the rat hind limb. Brain Rrs. 238, 29 43. 17. Lynn B. and Shakhanbeh J. (1988) Properties of A-delta high-threshold mechanoreceptors in the rat hairy and glabrous skin and their response to heat. Neurosci. Lelt. 85, 71-76. IX. Menetrey D., Cannon A., Levine J. D. and Basbaum A. I. (1989) Expression of c-fbs protein in interneurons and projection neurons of the rat spinal cord in response to noxious somatic, articular and visceral stimulation. J. romp. Neural. 285, 177-l 95. 19. Minami M., Kuraishi Y., Kawamura M., Yamaguchi T.. Masu Y.. Nakanishi S. and Satoh M. (1989) Enhancement of preprotachykinin A gene expression by adjuvant-induced inflammation in the rat spinal cord: possible involvement of substance P-containing spinal neurons in nociception. Neurosci. Leti. 98, I0551 IO. 20. Morgan J.. Cohen D.. Hempstead J. and Curran T. (1987) Mapping patterns of c-fos expression in the central nervous system after seizure. Science 237, 1922197. 21. Morgan J. I. and Curran T. (1986) Role of ion flux in the control of c-~‘s expression. Nature 322, 552 -555. 22. Nagy J. I., Hunt S. P., Iversen L. L. and Emson P. C. (1981) Biochemical and anatomical observations on the degeneration of peptide-containing primary afferent neurons after neonatal capsaicin. Neuroscience 6, 1923-1934. 23. Noguchi K., Morita Y., Kiyama H., Sato M., Ono K. and Tohyama M. (1989) Preproenkephalin gene expression in the rat spinal cord after noxious stimuli. Molec. Brain Res. 5, 227-234. 24. Paxinos G. and Watson C. (1982) The Rat Bruin in Srereotaxic Coordinates. Academic Press, Sydney. 25. Rauscher F. J. III. Sambucetti L. C., Curran T., Distel R. J. and Spiegelman B. M. (1988) Common DNA binding site for Fos protein complexes and transcription factor AP-I. Cell 52, 471-480. 26. Ruda M. A., Iadorola M. J., Cohen L. V. and Young W. S. III (1988) In-situ hybridisation histochemistry and immunocytochemistry reveal an increase in spinal cord dynorphin biosynthesis in a rat model of peripheral inflammation and hyperalgesia. Proc. nutn. Acad. Sci. U.S.A. 85, 622-626. 27. Sagar S. M.. Sharp F. R. and Curran T. (1988) Expression of cfos protein in brain: metabolic mapping at the cellular levjel. Science 240, 1328 1331. 28. Salt T. E. and Hill R. G. (1983) Neurotransmitter candidates of somatosensory primary afferent fibres. Neuroscience IO, 108331103. 29. Swett J. E. and Woolf C. J. (1985) The somatotopic organization of primary afferent terminals in the superficial laminae of the dorsal horn of the rat spinal cord. J. camp. Neural. 231, 66677. 30. Szekely A. M., Barbaccia M. L. and Costa E. (1987) Activation of specific glutamate receptor subtypes increases c-fos proto-oncogene expression in primary cultures of neonatal rat cerebellar granule cells. Neurophurmacology 26, 1779 1782. 31. Wall P. D. (1985) Future trends in pain research. Phil. Trans. R. Sot.. Lond. B 308, 393-401. 32. Williams S.. Evan G. and Hunt S. P. (1989) Distribution of cfos immunoreactivity in spinal cord depends on stimulus modality and survival time. Neurosci. Left., Suppl. 36, S30. 33. Williams S.. Pini A., Evan G. and Hunt S. P. (1990) Molecular events in the spinal cord following sensory stimulation. In Sensory Processing in the Superficial Dorsal Horn of the Spinal Cord (eds Cervero F., Bennett G. J. and Headley P. M.). Plenum Press. New York. 34. Willis W. D.. Trevino D. L., Coulter J. D. and Maunz R. A. (1974) Responses of primate spinothalamic tract. J. Neurophwiol. 37, 358-372. 35. Wisden W. and Hunt S. P. (1989) Induction of early onset genes in spinal cord neurons following sensory stimulation. Nwrosci. Leir. Suppl. 36, S75. 36. Woolf C. J. (1983) Evidence for a central component of post-injury pain hypersensitivity. Nrrfure 306, 686-688. 37. Woolf C. J. (1987) Central terminations of cutaneous mechanoreceptive afferents in the rat lumbar spinal cord. J. camp. Nrurol. 261, I0551 19. 38. Woolf C. (1990) Afferent induced alterations of receptive field properties. In Sensory Processing in the Supetjkial Dorsal Horn of‘rhe Spinal Cord (eds Cervero F., Bennett G. J. and Headley P. M.). Plenum Press, New York. (Accepted
II October
1989)
1. pp. 73-81,
0306-4522,90 $3.00 + 0.00 Pergamon Press plc IC\1990 IBRO
1990
CHANGING PATTERNS OF C-FOS INDUCTION SPINAL NEURONS FOLLOWING THERMAL CUTANEOUS STIMULATION IN THE RAT
IN
S. WILLIAMS,* G. I. EvANt and S. P. HUNT* *MRC Molecular Neurobiology Unit, University of Cambridge Medical School, Hills Road. Cambridge CB2 2QH, U.K. tInternational Cancer Research Fund Laboratories, Bartholomew Close, London ECIA 7BE, U.K. Abstract-Patterns of neuronal activity in the lumbar spinal cord of the anaesthetized rat were mapped by immunocytochemical localization of the c-fos gene product, Fos protein, at different timepoints following brief noxious stimulation of one hindpaw (20 s immersion in water at 52°C). After 2 h, Fos-immunoreactive neurons were seen mainly in the superficial laminae of the ipsilateral dorsal horn, with maximum somatotopic organization in lamina II. Subcutaneous injection of dilute formalin produced a similar pattern of immunostaining at 2 h, with a greater proportion of Fos-positive neurons in laminae III-VIII than with heat. With a survival time of 8 h following formalin injection, Fos immunoreactivity was virtually absent from the spinal cord. Eight hours after heat stimulation, however, the superficial pattern had given way to the appearance of a population of immunoreactive cells in the deeper laminae. The pattern of this “second wave” of heat-induced Fos-positive cells had a marked contralateral component, and was still present after 24 h, having become even more diffuse and symmetrical. The number of Fos-positive cells seen at 8 h was increased by local anaesthetic blockade of the peripheral nerve after stimulation, and reduced by continuous barbiturate anaesthesia. These findings suggest that the early stages of thermal injury trigger a complex pattern of molecular events within the spinal cord, which are initially monosynaptic and closely related to primary afferent terminal depolarization, and in the longer term the result of an induced pattern of synaptic activity set up within the spinal cord.
EXPERIMENTAL
The proto-onco~ene c-fos belongs to a large family of “immediate early onset” genes, activated in various
PROCEDURES
Male Sprague-Dawley rats (Tuck’s, 200-300 g) were deeply anaesthetized by intraperitoneal injection of 4 ml/kg Equithesin, prepared according to the following formula: 81 ml sodium pentobarbitone 60 mg/ml (“Sagatal”), 21.25 g chforat hydrate, 10.63 g ma~esium sulphate, 198 ml propylene glycol, $0 ml ethanol, distilled water to 500 ml. The left hindpaw was stimulated either by immersion for 20s in a gently-stirring water bath at 52°C. or by subcutaneous injection of 50~1 4% formalin into the distal paw pad. Anaesthesia following a single dose lasted approximately 2 h, but in some cases to achieve continuous anaesthesia, two or three injections of 2 ml/kg Equithesin were given when the cornea1 reflex returned. Local anaesthetic blockade of the sciatic nerve was produced by injection of 0.3 ml of 0.5% bupivacaine into the sciatic notch of the pelvis. At intervals of 30 min, 1, 2, 4, 8, 16 and 24 h after each stimulus, animals were terminally reanaesthetized with saturated chloral hydrate solution (1 ml, i.p.) and perfused intracardially with 200 ml normal saline followed by 400 ml freshly-depolyme~zed paraformaldehyde, 4% in 0.05 M sodium phosphate buffer (PB), pH 7.4. Spinal cords were post-fixed in the latter solution for 12 h, then washed overnight in 0.1 M PB containing 30% sucrose and 0.01% sodium azide. Forty-micron frozen sections taken in the tranversc plane from spinal cord segments Ll-L6, and in coronal and parasagittal planes from all segments caudal to T12, were incubated in rabbit antiserum to a synthetic peptide common to all c-fos and v-fos proteins,14.diluted 1: 3000 or 1: 10,000, and processed with an avidin-biotin complex-horseradish peroxidase (ABC-HRP) kit (Vector). The location of Fos-immunoreactive neurons was plotted with a camera lucida attached to a Leitz Orthomat photomicroscope viewing at x25. From approximately 100 sections from segments L3 to L4, three were chosen for maximum number of labelled neurons and viewed against a
cell types within minutes of stimulation by growth factors or neurotransmitter substances.‘r Following the suggestion that certain of these genes might play a role in the establishment of long-term functional change in the nervous system,3.” it was found that the c-J& could be induced synaptically in the spinal cord of the anaesthetized rat near the central terminals of C-fibres stimulated by noxious radiant heat or mustard oil applied to the skinI Fos immunoreactivity in this model appears to be restricted to cells with neuronal morphology under light and electron microscopy,‘4 and some can be retrogradely labelled from projection areas such as the thalamus and brainstem.lK It was assumed that these molecular events may be causally related to the changes in spinal cord physiology that follow peripheral injury or tetanic stimulation of C-fibres, which can be both complex, and persistent for many hours or days following the initial stimu1us.5.3’J6.38Here 3 we extend the original observations and report on the patterns of Fos protein immunoreactivity which evolve over many hours following brief sensory stimulation. A preliminary report of these findings has been made elsewhere.32 -Abhretkfions:
ABC, avidin-biotin complex; HRP, horseradish peroxidase; mRNA, messenger RNA; NMDA, N-methyl-o-aspartate; PB, phosphate buffer; SP, substance P. 73
s
74
wIl.I.I~MS (‘I
standard plan with the boundartes of Rexcd’s lammae adapted from Paxinos and Watson.?” For quantitative analysis all Fos-positive cells were counted separately regardless of staining intensity, on the left (ipsilateral to the stimulus) and right sides, and in superficial (Rexed’s laminae I-II) and deeper (III-X) zones of the same representative sections.
ut.
.
5H
RESULTS
Choice of’ stimuli
Surgical exposure of a nerve alone provokes massive cfos induction in the spinal cord, precluding electrical stimulation techniques for the purposes of this study. I4 Immersion of the hindpaw of the intact animal in heated water was easily and accurately administered and gave consistent results, and was therefore selected as a suitable physical stimulus for these experiments. As for a chemical stimulus, there had been considerable variation in response to cutaneous mustard oil depending on the age of the animal, probably due to differences in epidermal thickness. In contrast, plantar injection of dilute formalin is a commonly-used artificial noxious stimulus’ in which the site, dose and timing can be reliably controlled, and produces a characteristic electrophysiological response in the spinal cord which has recently been investigated.* Basal levels of Fos within the spinal cord were not revealed by these techniques, and staining was seen only following stimulation, almost exclusively in neuronal nuclei. Cells stained to variable degree regardless of position, so that densely-immunoreactive and faintly-outlined nuclei were juxtaposed. Use of higher concentrations of the polyclonal primary serum resulted in some non-specific background staining of gray, and cellular elements of white matter (Figs 6 and 7) and occasionally the cytoplasm of strongly positive cells (Fig. 7D).
1H
Thermal cutaneous stimulation
After the heat stimulus (immersion of the left hindpaw for 20 s in stirred water at 52”C), Fos immunoreactivity appeared ipsilaterally in laminae I-III and V within 30 min (n = 3) and laminae VI, VII and X by 1 h (n = 3) (Fig. 1). A smaller number of Fos-positive cells was distributed similarly on the right-hand side of the cord. After a survival time of 2 h, the Fos-positive cell count had reached a peak in superficial laminae I and II on the left, ipsilateral to the stimulus, although the number of stained contralateral cells had fallen (Fig. 4). Dense staining in lamina 11 was restricted to cord segments L2-L4, but positive neurons were seen in laminae I, III-VIII and X in transverse sections from segments Ll to L6. Longitudinal sections of spinal cord showed that an increasingly attenuated signal was present as far rostrally as Ll in lamina I, and TlO in lamina V. After 4 h, the number of immunoreactive cells in the most superficial laminae was reduced, whereas the
Fig. 1. Camera lucida drawings of rat spinal cord, fourth lumbar segment in transverse section, immunostained for Fos 0.5524 h after heat stimulation of the left hindpaw. Each circle represents three positive nuclei on a composite of three sections. Filled circles show the position of heavily stained neurons.
signal in V, VI, VII and X appeared to be increasing (n = 3). Contralaterally, a few positive cells had appeared in the deeper laminae. At 8 h (n = 6) this change had progressed with marked shift of positive cells to deep laminae, especially V. At 16 h (n = 3) there was diffuse bilateral induction of c-fos in all laminae except IX, and by 24 h the pattern was almost symmetrical, with small numbers of positive cells appearing even in the contralateral superficial laminae (Figs 1 and 4).
Changing
patterns
of C+OS expression
No Fos immunoreactivity was seen in the spinal cord following immersion of the hindpaw in water at 42 C (II = 2 at each timepoint). In the cord of a heatstimulated animal, therefore, ipsilateral Fos immunoreactivity attenuated in the superficial laminae after 2 h, but gradually increased in the deep Iaminae until at 8 h it appeared that many more deep cells were positive than at 2 h. particularly in lamina V (Figs I and 4). Secondly, Fos-positive cells were seen from 4 h onwards in the contralateral dorsal horn which had not been present at 2 h (Fig. 4). By I6 h a massive increase in the Fos-positive cell count had occurred bilaterally in laminae VI-VIII. and at 24 h there was a very diffuse bilateral pattern bearing little relation to the relatively restricted signal seen earlier (Fig. I). Stimulation
by ,formnlin
Two hours after subcutaneous formalin injection into thcdistal paw pad (tz = 3), there was Fos immunoreactivity in ipsilateral neuronal nuclei throughout the gray matter of the lumbar spinal cord segments From L2 to LS, most strongly in laminae I-VII (Fig. 2). The signal was maximal in a short segment of cord near the L3/L4 junction, but absent from lamina II elsewhere and increasingly attenuated in other laminae with increasing rostra1 or caudal distance from the L3/L4 junction. By comparison with that induced by heat stimulation of the whole foot, the formalin pattern was also more restricted mediolaterally in lamina II, although there was more signal in the deeper laminae V-VII at 2 h, especially in lateral lamina V, than with heat stimulation. Eight hours after formalin stimulation, however (n = 3). there were only a very few faintly-immunoreactive nuclei in laminae I---VI of the ipsilateral dorsal horn, and no contralateral signal (Fig. 2). Insertion of a dry needle (n = 2), or injection of 50~1 normal saline (n = 3), resulted in Fos immunoreactivity in one or two neurons per 40-l.lrn section in a few sections from the fourth lumbar cord segment.
2H
8H Fig. 2. Rat lumbar spinal cord 2 and 8 h following subcutaneous injection of formalin into the distal pad of the left hindpaw.
75
in spinal neurons
A large number of lamina II neurons were activated by heat stimulation of the entire hindpaw surface, in a pattern extending rostrocaudally for several segments and mediolaterally across most of the width of the lamina. In contrast, formalin injection produced a signal in a very restricted zone of lamina II, where the pattern of Fos-positive neurons occupied less than half of the width of the dorsal horn for approximately 200 pm in the region of the L3/L4 junction. Nonetheless, the number of stained cells in the deeper Iaminae at 2 h, centred around the patch of Fos-positive neurons in lamina II, was greater with formalin despite the fact that a much smaller area of skin had been directly stimulated. Neurogenic oedema usually developed within 1 h of stimulation by heat, more prominently in the loose dorsal tissue than on the plantar surface of the paw, and less so with the localized formalin stimulus. The oedema following heat stimulation had increased after 8 h and was still marked at 24 h, while that following formalin injection appeared to have resolved by 8 h. Animals with oedema invariably developed a marked signal in the spinal cord. Occasionally oedema did not develop but this did not preclude spinal induction of c-fffs. Thermal cutaneous stimtdation ,folfowed andlor sustained general anaesthesio
by
local
The following experiments were designed to examine the “second wave” of signal seen bilaterally in the deeper laminae, V-VII and X, 8 h after heat stimulation. After bupivacaine blockade of the sciatic nerve, the apparent motor and sensory deficit was similar to that seen after interruption of neurotransmission by sciatic nerve division or crush, and normal sensation appeared to be restricted to the area supplied by the saphenous nerve.’ Partial hindlimb paralysis, with extension of the leg but no active flexion of the leg or foot, and apparent anaesthesia of an area including the lateral three digits, occurred within 10 min and lasted for 3-4 h. During this period there was no hyperventilation. vocalization or attempted withdrawal after any degree of stimulation of the lateral three digits, but gentle seizing of the medial digit with forceps produced one or more of these signs of nociceptive activity. The second digit was sometimes included in the region of complete anaesthesia, and sometimes not. If not, a firm pinch was required to elicit a response. No Fos immunoreactivity was seen in the lumbar cord following delivery of the block alone (n = 3). When the ipsilateral sciatic nerve had been blocked I5 min before heat stimulation (n = 3), neuronal c$s induction in lamina II after 2 h was restricted to a very short length of cord in the LI!L2 region corresponding to the saphenous nerve projection.‘” In segments L?-L6 of these animals the signal in the deeper laminae was relatively attenuated.
s. WILL
76
LAMS el
Ui
The number of Fos-positive cells seen at 8 h was increased when the ipsilateral sciatic nerve was blocked with bupivacaine 1 h following heat stimulation of the hindpaw (fz = 6) (Figs 3 and 5). When these animals had recovered from general anaesthesia their gait and posture confirmed successful blockade of at least motor neurotransmission in the sciatic nerve. Conversely, animals kept anaesthetized with repeated doses of Equithesin for the full 8 h between stimuIation and perfusion (n = 3) had fewer positive neurons than those which had been allowed to recover. Those receiving both local and continuous general anaesthesia after stimulation (n = 3) developed Fos immunoreactivity in a similar pattern and number of cells to animals given only the usual initial single dose of Equithesin and no additional anaesthetic (Figs 3 and 5). DISCUSSION
Fos protein appeared in neurons of the rat spinal cord within 30min, following noxious thermal or chemical cutaneous stimulation. At 2 h the pattern of positive neurons from either mode of stimulation was topographically consistent with induction by synaptic activation by primary afferent nerve fibres, possibly in a monosynaptic fashion.29 We previously demonstrated that the appearance of Fos protein within the spinal cord was dependent upon activation of smalldiameter high-threshold primary afferents, and that
LAMINAE I/ -X
Fig. 4. Maximal numbers of Fos-positive neurons, in superficial (I-II) and deep (III-X) laminae, on the left and right sides, per section of rat lumbar spinal cord at different times following heat stimulation of the left hindpaw. Average of three representative sections each from three animals. Error bars = S.E.M.
the pattern of labelling within the first few hours following stimulation closely followed the pattern of termination of particular sensory afferents within the dorsal hom.i4 Activation of cutaneous at&rents resulted in labelling within their areas of te~ination in laminae I, II and V, while Fos-positive neurons were largely restricted to laminae I and V following activation of muscle or visceral afferents,33 which are known to terminate almost exclusively within these laminae.6 This suggested a close, perhaps monosynaptic Iink between primary afferent activation and c-fis induction within neurons. However, we report here that at longer survival times, a “second wave” of c-jbs labelling can occur in neurons distant from areas of high-threshold primary afferent termination. Eficts
LA
LA + GA Fig 3. Rat lumbar spinal cord 8 h following heat stimulation of the left hindpaw, with local anaesthetic blockade of the ipsiiateral sciatic nerve beginning 1 h post-stimulation (LA), continuous general anaesthesia (GA), and both local and continuous general anaesthesia (LA + GA).
of heat
With heat, the number of Fos-positive neurons peaked twice within the first 24 h. The first peak at 2 h, as originally reported in Hunt et ~1.,‘~was seen principally in the superficial laminae, almost exclusively ipsilateral to the stimulus, while a “second
Fig. 5. Maximal numbers of Fos-positive neurons per section of rat lumbar spinal cord 8 h following heat stimulation of the left hindpaw, with and without local anaesthetic blockade of the sciatic nerve (LA), continuous general anaesthesia (GA), or both (the same sections as those shown in Fig. 3). Average of 10 representative sections. Error bars = S.E.M.
Changing patterns of c,fos expression in spinal neurons wave” of Fos protein synthesis occurred bilaterally in the deep laminae over a much longer period of time, commenting at about 8 h, reaching a maximum number of cells at around 16 h, and still clearly present 24h post-stimulation (Figs 1 and 4). Polwynaptic
mechanisms
It would be difficult to maintain that the activation of deeper neurons was the result of a monosynaptic event. Stimulation with heat at 52’C, and probably with formalin, activates primary afferents which are almost exclusively unmyelinated C-fibres in the rat,17 80-85% of which are nociceptive.‘b These, along with the small myelinated high-threshold A-delta sensory fibrcs, terminate ipsilaterally largely within laminae I and II, and occasionally IV and V.” Yet Fos-positive cell bodies are found at 16-24 h within laminae other than these. and the delayed bilateral expression of c-f& products following unilateral stimulation suggests that polysynaptic pathways are involved in the activation of deeper neurons. For many neurons. we believe that the evolving bilateral pattern of c,fos expression within the deeper laminae of the spinal cord occurs largely independently of continuous primary afferent input (see below), and is probably polysynaptic in origin.” Thus, while the early pattern of neuronal excitation could well be the result of monosynaptic activation, following release from primary afferents of neurotransmitter candidates such as substance P (SP),’ synaptic contact with primary afferents could not explain the distribution of induced gene expression seen later. This is presumably the result of patterns of neuronal activity set up in intra- or intersegmental pathways, or possibly involving the activation of pathways descending from the brainstem. In the lumbar cord of the spinalized rat, about half the ccfls in laminae IV-VI which have a direct ipsilateral primary afferent input can be inconsistently excited by stimulation of the contralateral sciatic nerve at A-fibre strength, via a segmental pathway that tends to be inhibited by stimulation at C-fibre strength, and is hence thought to be polysynaptic.“’ Such a shortlatency mechanism, however, would not explain the occurrence of neuronal biosynthetic activity & ~UWO after 8 h or more. and a descending system, such as that postulated to account for diffuse noxious inhibitory controls,” remains a likely mediator of the delayed effects of noxious stimulation on spinal c-fijs expression. The appearance of c$is messenger RNA (mRNA) has from previous work with cell lines been regarded as a transient event, translating into protein with a substantially longer half-life.” Yet the appearance of Fos protein bilaterally over an extended period of time, in an apparently (from the distributions) novel population of cells would perhaps argue against this being the universal mechanism. Indeed, using in situ hybridization histochemistry, mRNA for c+,c is found to be heavily expressed in ipsilateral neurons of
77
iaminae I and II from 5 to 120min after cutaneous heat stimulation, before decaying over the next 120 min.35
Formalin injection resulted in a single peak of cell labelling at 2-4 h which appeared to fade uniformly thereafter, having all but disappeared at 8 h (Fig. 2). The single injection of 50~1 formalin into the distal paw pad was sufficient to induce c+s in neurons within al1 laminae of the lumbar spinal cord. In the region of cord with maximal numbers of Fos-positive neurons, transverse sections contained at least as many stained cells following formalin injection. with a greater effect in the deeper Iaminae III--X. as with immersion of the entire paw in water at 52, C. This may indicate that formalin injection activates a greater number of neurons per unit of skin area stimulated than heat, or that it activates some or all of them more intensely. Formalin injection may activate deep tissue nociceptors which cutaneous heating fails to stimulate effectively. In an electrophysiological study, Dickenson and Sullivan’ report that dorsal horn neurons activated exclusively by innocuous stimuli were not excited by formafin injection into their receptive fiefd. However, the appearance of Fos-positive neurons within laminae III-IV following formalin injection suggests that the abnormal excitation of normally non-nociceptive afferents. which are thought to be the sole input to the intermediate dorsal horn laminae,2’ may be at least partly responsible for the early pattern of labelled neurons within the dorsal horn. It may be that non-specific activation of many fibrc types by formalin, including non-nociceptive afferents, results in inhibition of the later developing pattern of c+s induction.
To test our hypothesis that the “second wave” of c+s expression in dorsal horn neurons was a delayed response to the initial stimulus and not due to continued activity in heat-sensitized aKerents, a group of animals was kept continuously anaesthetized to prevent stimulation of the heated paw by volunrary movement. The signal was indeed reduced. but these animals had received in total two to three times the dose of Equithesin required to produced anaesthesia initially. Equithesin contains sodium pentobarbitone, and was thought likely to have affected the results by inhibition of neurotransmission within the spinal cord, as low doses of barbiturate drugs seem to do in primates.” Sodium pentobarbitone abolishes the c~fis expression induced in mouse brain by Metrazole.Z” The effect of local anaesthetic applied to the sciatic nerve 1 h after heat stimulation suggested that ongoing primary afferent activity, the result of tissue damage and inflammation, was not necessary for the appearance of the “second wave” of
78
S. WILLIAMS
('I~1
Fig. 6. Photomicrogmphs of one 40qm section from LJ cord segment, 2 h after heat stimulation of the left hindpaw. (A) Low power view showing Fos immunoreactivity concentrated ipsilaterally in superficial laminae, at higher power in B. with a few positive cells in lamina V (D). No significant staining is seen contralaterally in superficial (C) or deep laminae (E). Both scale bars == 100 /urn.
Fig. 7. Sixteen hours after heat stimulation (A), Fos-positive cells are much sparser ipsilaterally in laminae I and II (B) than at 2 h. but diffuse c-fos induction has occurred contralaterally in superficial laminae (C). as well as laminae III OXon both the stimulated (D) and unstlmulated side (E). Reaction product is visible in the cytoplasm of an intensely stained cell (D: large arrowhead), but many nuclei at this stage are only faintly positive. with a granular appearance (D, small arrowheads). Both scale bars = 100 /in?.
thermally-induced Fos labelling. Bupivacaine was shown by observation and behavioural tests in conscious animals, and Fos immunohistochem~stry following heat stimulation, to produce effective blockade of the ipsilateral sciatic nerve for up to 4 h, although primary afferent input after this period could have contributed towards the observed late labelling. The increasing signal in superficial laminae suggests that c+s induction is under tonic inhibitory control by primary afTerent activity. An increase in the “spontaneous” activity within lamina I and the substantia gelatinosa has been reported following surgical deafferentation,’ which suggests that continuous activity within superficial laminae could result in the induction of C~$Xin local neurons. Since 2 h was required for the earliest signai to become maximai, it was suggested that the “second wave” might have followed some delayed afferent activity by a similar period, despite the fact that the diffuse, bilateral pattern of Fos-positive neurons emerging at 8 h was quite distinct from the arrangement of central termini of primary afferent fibres. With sciatic blockade starting at 1 h. the increased signal 8 h after heat stimulation suggests that activity in the intact sciatic nerve after thermal injury may tend to reduce the “second wave” of c-f0.r induction, rather than contribute towards it. Finally, animals which received both local and continuous general anaesthesia developed a pattern at 8 h similar to that of control animals, and it was concluded that the “second wave” of c+s induction was independent of continued activity in the sciatic nerve. This has the important implication that c-&s induction may occur as the result of transmission from neuron to neuron within the CNS, not purely as the result of synaptic contact with an active primary afferent terminal. To date, poiysynaptic neuronal activation of c;fos has been variably demonstrated
only by direct electrical stimulation of the cerebral cortex, or other means of producing seizure’-
The nature of the stimulus required for the induction of c-&s in neurons is unknown. Some evidence suggests that glutamate acting at the Ymethyl-D-aspartate (NMDA) receptor” or the peptide SP,4 both thought to be primary afferent neurotransmitters,2R may be effective stimuli, at least in primary culture. Alternatively, ~f0.r induction may be related entirely to the degree of membrane depolarization and the entry of calcium either through ligand or voltage-gated channels.” What is the significance of ~fi)s activation? Early onset genes are thought in many cases to code for transcription factors controlling the expression of batteries of downstream genes. Fos protein has been shown to form a heterodimer with the c-jun product, which then effects transcription through an identified c7jun/AP1 DNA binding domain on other genes.‘” Both c-&n A and c-f0.r genes are rapidly induced in dorsal horn neurons following noxious stimulation.” The genes affected by C-$X activation have yet to be identified. However, 8-24 h following injection of formalin or complete Freund’s adjuvant into one hindpaw, the levels of mRNA encoding SP,‘” enkephalin’” and dynorphin13,zh rise dramatically within the dorsal horn. These events may be dependent on c-$&r activation. The function of increased peptide synthesis within the cord is unclear but, as within the “second wave” of c+.s activation following brief heat stimulation, could reflect the activity 01 an endogenous pain control system brought into play following particular types of noxious stimulation. Acknowledgements--We wish to thank Mrs L. A. McNaughton for her assistance. SW. holds a Wellcome Clinical Research Training Fellowship, and received support for part of this work from the Sir Halley Stewart Trust.
REFERENCES
I. Alreja M., Mutalik P. G., Nayar U. and Manchanda S. K. (1984) The formalin test: a tonic pain model in the primate. Pain 20, 97-105.
2. Basbaum A. I. and Wall P. D. (1976) Chronic changes in the response of cells in adult cat dorsal horn following partial deaffe~nt~ation: the appearance of responding cells in a previously non-responsive region. Bruin Res. 116, 184-204. 3. Berridge M. (1986) Second messenger dualism in neuromodulation and memory. Nature 323, 294-295. 4. Bigot D., Evan G. and Hunt S, P. (1989) Rapid induction of c-$13 protein in gliai and neuronal cells studied itzz&w. Neurosci. Left., Suppl. 36, 575.
5. Cook A. J., Woolf C. J., Wall P. D. and McMahon S. B. (1987) Dynamic receptive field plasticity in rat spinal cord dorsal horn following C-primary afferent input. Nurure 325, 151-153. 6. De Groat W. C. (1986) Spinal cord projections and neuropeptides in viscera1 afferent neurons. In Kweruf Sensaiiun (eds Cervero F. and Morrison J. F. is.>: Prog. Brain Res. 67, 165-187. 7. Devor M., Schonfeld D., Seltzer Z. and Wall P. D. (1979) Two modes of cutaneous reinnervation following peripheral nerve injury. J. camp. Neuroi. 185, 211-220. 8. Dickenson A. H. and Sullivan A. F. (1987) Subcutaneous formalin-induced activity of dorsal horn neurons in the rdt: differential resnonse to an intrathecal opiate administered pre or post formalin. Pain 30, 349-360. 9. Duggan A. W.: Morton C. R., Zhao Z. 6. and Hendry I. A. (1987) Noxious heating of the skin releases immunoreactive substance P in the substantia aelatinosa of the cat: a studv with antibodv micronrobes. Bruin Res. 403, 345.-349. 10. Fitzgerald M. (1982) The con&lateral input to the dorsal kern of the spinal cord-in the decerebrate spinal rat. Brain Res. 236, D-287.
11. Goelet P., Casteltucci V. F., Schacher S. and Kandel E. R. (1986) The long and the short of long-term memory--a molecular framework. Nature 322, 41%-422.
Changing
patterns
of c:fos expression
in spinal
neurons
81
12. Greenberg M. E., Greene L. A. and Ziff E. B. (1985) Nerve growth factor and epidermal growth factor induce rapid transient changes in proto-oncogene transcription in PC12 cells. J. biol. Chem. 260, 14,101~14,110. 13. Hollt V.. Haarmann I., Millan M. J. and Herz A. (1987) Prodynorphin gene expression is enhanced in the spinal cord of chronic arthritic rats. Neurosci. Lett. 73, 90-94. 14. Hunt S. P., Pini A. and Evan G. (1987) Induction of c-/&y-like protein in spinal cord neurons following sensory stimulation. Nurture 328, 632-634. 15. Le Bars D., Dickenson A. H. and Besson J. M. (1979) Diffuse noxious inhibitory controls (DNIC)&I. Effects on dorsal horn convergent neurons in the rat. Pain 6, 2833304. 16. Lynn B. and Carpenter S. E. (1982) Primary afferent units from the hairy skin of the rat hind limb. Brain Rrs. 238, 29 43. 17. Lynn B. and Shakhanbeh J. (1988) Properties of A-delta high-threshold mechanoreceptors in the rat hairy and glabrous skin and their response to heat. Neurosci. Lelt. 85, 71-76. IX. Menetrey D., Cannon A., Levine J. D. and Basbaum A. I. (1989) Expression of c-fbs protein in interneurons and projection neurons of the rat spinal cord in response to noxious somatic, articular and visceral stimulation. J. romp. Neural. 285, 177-l 95. 19. Minami M., Kuraishi Y., Kawamura M., Yamaguchi T.. Masu Y.. Nakanishi S. and Satoh M. (1989) Enhancement of preprotachykinin A gene expression by adjuvant-induced inflammation in the rat spinal cord: possible involvement of substance P-containing spinal neurons in nociception. Neurosci. Leti. 98, I0551 IO. 20. Morgan J.. Cohen D.. Hempstead J. and Curran T. (1987) Mapping patterns of c-fos expression in the central nervous system after seizure. Science 237, 1922197. 21. Morgan J. I. and Curran T. (1986) Role of ion flux in the control of c-~‘s expression. Nature 322, 552 -555. 22. Nagy J. I., Hunt S. P., Iversen L. L. and Emson P. C. (1981) Biochemical and anatomical observations on the degeneration of peptide-containing primary afferent neurons after neonatal capsaicin. Neuroscience 6, 1923-1934. 23. Noguchi K., Morita Y., Kiyama H., Sato M., Ono K. and Tohyama M. (1989) Preproenkephalin gene expression in the rat spinal cord after noxious stimuli. Molec. Brain Res. 5, 227-234. 24. Paxinos G. and Watson C. (1982) The Rat Bruin in Srereotaxic Coordinates. Academic Press, Sydney. 25. Rauscher F. J. III. Sambucetti L. C., Curran T., Distel R. J. and Spiegelman B. M. (1988) Common DNA binding site for Fos protein complexes and transcription factor AP-I. Cell 52, 471-480. 26. Ruda M. A., Iadorola M. J., Cohen L. V. and Young W. S. III (1988) In-situ hybridisation histochemistry and immunocytochemistry reveal an increase in spinal cord dynorphin biosynthesis in a rat model of peripheral inflammation and hyperalgesia. Proc. nutn. Acad. Sci. U.S.A. 85, 622-626. 27. Sagar S. M.. Sharp F. R. and Curran T. (1988) Expression of cfos protein in brain: metabolic mapping at the cellular levjel. Science 240, 1328 1331. 28. Salt T. E. and Hill R. G. (1983) Neurotransmitter candidates of somatosensory primary afferent fibres. Neuroscience IO, 108331103. 29. Swett J. E. and Woolf C. J. (1985) The somatotopic organization of primary afferent terminals in the superficial laminae of the dorsal horn of the rat spinal cord. J. camp. Neural. 231, 66677. 30. Szekely A. M., Barbaccia M. L. and Costa E. (1987) Activation of specific glutamate receptor subtypes increases c-fos proto-oncogene expression in primary cultures of neonatal rat cerebellar granule cells. Neurophurmacology 26, 1779 1782. 31. Wall P. D. (1985) Future trends in pain research. Phil. Trans. R. Sot.. Lond. B 308, 393-401. 32. Williams S.. Evan G. and Hunt S. P. (1989) Distribution of cfos immunoreactivity in spinal cord depends on stimulus modality and survival time. Neurosci. Left., Suppl. 36, S30. 33. Williams S.. Pini A., Evan G. and Hunt S. P. (1990) Molecular events in the spinal cord following sensory stimulation. In Sensory Processing in the Superficial Dorsal Horn of the Spinal Cord (eds Cervero F., Bennett G. J. and Headley P. M.). Plenum Press. New York. 34. Willis W. D.. Trevino D. L., Coulter J. D. and Maunz R. A. (1974) Responses of primate spinothalamic tract. J. Neurophwiol. 37, 358-372. 35. Wisden W. and Hunt S. P. (1989) Induction of early onset genes in spinal cord neurons following sensory stimulation. Nwrosci. Leir. Suppl. 36, S75. 36. Woolf C. J. (1983) Evidence for a central component of post-injury pain hypersensitivity. Nrrfure 306, 686-688. 37. Woolf C. J. (1987) Central terminations of cutaneous mechanoreceptive afferents in the rat lumbar spinal cord. J. camp. Nrurol. 261, I0551 19. 38. Woolf C. (1990) Afferent induced alterations of receptive field properties. In Sensory Processing in the Supetjkial Dorsal Horn of‘rhe Spinal Cord (eds Cervero F., Bennett G. J. and Headley P. M.). Plenum Press, New York. (Accepted
II October
1989)
