Parallel nociceptive reflex pathways with negative and positive feedback functions to foot extensors in the cat

Authors


  • Author's present address N. Wada: Department of Veterinary Physiology, Yamaguchi University, Yamaguchi 753, Japan.

Abstract

  • 1Nociceptive reflex pathways to foot extensors were investigated with particular attention given to those not following a flexor reflex (FRA) or withdrawal pattern.
  • 2In anaemically decapitated, high spinal paralysed cats nociceptive afferents of the foot pad were activated by noxious radiant heat (48–60 °C), while for comparison non-nociceptive afferents were activated by weak mechanical stimulation of the skin or graded electrical nerve stimulation. The reflex action of the afferents on hindlimb motoneurones, innervating plantaris and intrinsic foot extensors (tibial nerve), was investigated by intracellular recording, by monosynaptic reflex testing and by recording of neurograms during fictive locomotion. A possible descending control of the nociceptive and non-nociceptive pathways was tested by application of opioidergic and monoaminergic compounds.
  • 3Beside the typical FRA pattern evoked in the majority of hindlimb motoneurone pools by nociceptive afferents from different skin areas of the foot, the results revealed parallel excitatory and inhibitory nociceptive reflex pathways from the central pad and partly from the toe pads to foot extensors. The excitatory pathways, which did not follow the FRA pattern, were predominantly to plantaris and intrinsic foot extensors. They were distinctly less depressed by opioids and monoaminergic compounds than FRA pathways.
  • 4While the nociceptive FRA pathways have a general nocifensive withdrawal function, the nociceptive excitatory non-FRA pathway to the foot extensors causes a movement of the affected area towards the stimulus or at least a resistance against the stimulus, i.e. it mediates a positive feedback.

The classical reflex evoked by nociceptive afferents can be considered to be the Sherringtonian flexion reflex with its function ‘to withdraw the limb from contact with injurious agents’ (Creed et al. 1932), i.e. with a typical negative feedback function. More recent investigations by Schouenborg and co-workers in the rat questioned the general flexion reflex response in favour of multiple modular withdrawal reflexes to nociceptive stimuli. However, the authors adhered to the concept of a principally negative feedback role of these reflexes even though both excitatory and inhibitory effects could be evoked from overlapping skin areas (Schouenborg et al. 1990, 1994; Weng & Schouenborg, 1996). As in most biological systems, positive feedback in motor control had generally been assumed to have a destabilising function rather than a stabilising one. However, previously it had already been suggested that flexor reflex afferents (FRAs, in the sense of Lundberg, 1979), which are activated during normal active movements, may contribute positive feedback to the ongoing movement (Lundberg, 1979; Jankowska & Lundberg, 1981; Lundberg et al. 1987b; Schomburg, 1990). This possibility became more evident when positive force feedback from Golgi tendon organs to extensors could be demonstrated under particular conditions, as during standing and during the stance phase during locomotion (Pearson & Collins, 1993; Gossard et al. 1994; Pratt, 1995). Under corresponding motor conditions this positive force feedback is probably required for stabilisation of the leg and the foot (Prochazka et al. 1997).

From the very beginning of the development of the FRA concept, Lundberg suggested that the different types of afferents belonging to the FRA group may also use ‘private’ pathways, not forming part of the FRA system in order to subserve more specific functions (Holmqvist & Lundberg, 1961; Lundberg, 1979; Lundberg et al. 1987b). For a long time it was uncertain whether there is also ‘a special spinal reflex pathway from nociceptors’ (Lundberg, 1982). We have now been able to demonstrate such a ‘private’ excitatory nociceptive non-FRA pathway from the foot pad to extensors of the foot by using various experimental approaches. This pathway acts in a positive feedback way and exists in parallel with an inhibitory nociceptive FRA pathway; it seems to be the nociceptive equivalent of the mechanoreceptive excitatory non-FRA pathway from the pad to foot extensors described by Engberg (1964). The nociceptive pathway probably subserves specific motor functions which supplement the function of the mechanoreceptive pathway. A preliminary communication of some of these findings has been presented (Schmidt et al. 1987).

METHODS

General procedure

The results were collected from different series of experiments in adult cats: 15 experiments with intracellular recording; five experiments with monosynaptic reflex testing and systematic analysis of local specificity of stimulation; some 40 experiments with monosynaptic reflex testing on the influence of opioids and l-DOPA (l-3′4-dihydroxyphenylalanine) on the transmission in the segmental reflex pathways; and four experiments with recording of neurograms during fictive locomotion. The experiments were carried out with official permission in accordance with national guidelines.

The basic procedure was identical in all experiments. Under general ether-halothane-nitrous oxide anaesthesia (O2:N2O, 1:2, halothane initially 2.5 %, then increasingly replaced by ether, to obtain full anaesthesia, as assessed by a complete loss of muscle tone and by a lack of blood pressure, heart rate or motor responses to any stimuli) the cats were anaemically decapitated by a permanent bilateral ligature of both common carotid arteries and their main branches, and of the ascending vertebral arteries. An irreversible interruption of the spontaneous respiration resulted from this procedure together with persistent large, non-reacting pupils. These effects were taken as a sign of the completion of anaemic decapitation.

To prove the reliability of the procedure, in another, former series of experiments 100 ml of an Evans Blue solution (0.5 g (100 ml)−1, 2000 i.u. heparin added) was infused over a period of 3 min. Then the cat was killed and dissected 5 min after the infusion. In contrast to spinal structures and the muscles of the body and limbs, no dye was found within the superficial (pial) or deep brain vessels (for further details see Kniffki et al. 1981).

After anaemic decapitation, the cat was artificially ventilated, spinalised at C1 and paralysed with pancuronium bromide (Pancuronium ‘Organon’ about 0.15 mg kg−1 every hour i.v. as required). An end-expiratory CO2 concentration of 3.5–4.5 % was regulated via the respiratory volume. The arterial blood pressure was maintained above 80 mmHg, if necessary by infusion of a dextran solution. Rectal temperature was maintained close to 37–38 °C. At the end of the experiments the cats were killed by injection of 5 ml of 3 m KCl solution, which induced immediate cardiac arrest.

Preparation and stimulation

One hindlimb was completely denervated except for the plantar division of the tibial nerve (Tib) which innervates the foot pad. For comparison, in six experiments the sural nerve (Sur, innervating the heel and the lateral side of the foot) or the cutaneous branch of the superficial peroneal nerve (SPC, innervating the dorsum of the foot) was also kept intact. The proximal branches of the transected nerves, posterior biceps semitendinosus (PBSt), anterior biceps semimembranosus (ABSm), quadriceps (Q), gastrocnemius-soleus (GS, partly separated as lateral gastrocnemius-soleus, LGS, and medial gastrocnemius, MG), plantaris (Pl), flexor digitorum et hallucis longus (FDL), deep peroneal (DP), superficial peroneal nerve muscular branch (SPm), saphenus (Saph), Sur and SPC (if not intact), posterior nerve to the knee joint (joint), as well as the Tib nerve (mobilised if left in continuity) were mounted on bipolar electrodes and electrically stimulated with single or double rectangular pulses of 0.1 ms duration, with a recurrence frequency of 1 Hz (intracellular recording) or 0.5 Hz (monosynaptic reflex testing) and a stimulation strength indicated in multiples of the threshold strength of the nerves (‘T‘) as indicated by the afferent volley recorded from the dorsal root L7 close to the root entry. Nociceptive cutaneous afferents were activated by noxious radiant heat (48–55 °C and up to 60 °C, in four specifically cited experiments). The radiated area was a circle of 1 cm2 and its temperature was measured at the skin surface (cf. Schomburg & Steffens, 1986). Heat stimulation was applied to areas innervated by the intact nerve or nerves, respectively, in the case of the intact Tib nerve to the central foot pad and for comparison to the different toe pads. Low threshold mechanoreceptors were activated by light stroking of the plantar side of the foot.

Except for the experiments concerning fictive locomotion, the ventral roots L5 or L6 to S2 were cut and the ventral roots L7/S1 were mounted for stimulation in experiments with intracellular recording and for recording in experiments with monosynaptic reflex testing.

Intracellular recording of α-motoneurones (DC recording bandwidth 0–3 kHz, AC recording 0.1–10 kHz) was performed with microelectrodes filled with 2 m potassium citrate. Results were obtained from 126 motoneurones of PBSt (41), GS (38), ABSm (23), Tib (12), Pl (9) and SPm (3).

Monosynaptic reflex testing. In order to be able to test different reflex pathways to different motor nuclei in parallel for a systematic analysis of local specificity of noxious stimulation or before, during and after drug application, the technique of monosynaptic reflex testing was used. The monosynaptic reflexes were recorded from the ventral roots L7/S1. The reflex recordings were rectified and averaged over eight samples with a time resolution of 20 μs per point.

Different peripheral motor nerves (PBSt, GS, Pl, Tib) and occasionally FDL, SPm and DP were stimulated alternately. The strength of the test stimuli was routinely 5T i.e. well above group I maximum in order to avoid any influence of a changing group I excitability during the experiment. Single pulses were used. Only if single pulses were insufficient to evoke a reflex were double pulses applied, at an interval of 1.0–1.2 ms. Any polysynaptic responses which occurred at the stimulus strength used and identified by their longer delay, were excluded from the data analysis.

The conditioning stimuli were as follows: noxious radiant heat applied to the central foot pad (and also to the different toe pads for analysis of local specificity) or to the dorsum or the lateral side of the foot for activation of nociceptive cutaneous afferents; stimulation of cutaneous nerves (Saph, Sur, SPC) with low strength (generally less than 1.2T) for activation of low threshold cutaneous afferents; stimulation of the joint nerve (strength not higher than 2.2T) for activation of mainly non-nociceptive joint afferents; and stimulation of the Q and the GS nerve (5.5T) for activation of group Ib and group II muscle afferents (conditioning-test interval with electrical conditioning nerve stimulation 5–7 ms). A differentiation between group Ib and group II effects was performed by grading the strength (cf. Fu et al. 1974; Lundberg et al. 1987a). Antagonistic group Ia effects from Q to PBSt were largely excluded by the long conditioning-test interval, but a contribution of Ia fibres to the Ib effects from Q and GS via common interneurones (Jankowska, 1979; Harrison et al. 1983) cannot be excluded.

The influence of the following drugs was tested: (d-Ser2)-leuenkephalin (Thr6) (DSLET) an opioid with a δ-morphine receptor agonistic action; (d-Ala2,N-Me-Phe4,Gly5-ol)-enkephalin (DAMGO), an opioid with a μ-morphine receptor agonistic action; a benzomorphan derivate (MR 2034) with a prevailing κ-morphine receptor agonistic action, l-DOPA (40–100 mg kg−1, i.v.), and the α2-agonist clonidine (0.02–0.045 mg kg−1, i.v.). The application of the opioids was performed in two different ways. The lumbar spinal cord (segment L3 to cauda equina) was suffused with opioids (‘local application’, seven experiments with DSLET, five experiments with DAMGO) diluted in Ringer solution (concentration of 10−6-10−3m). After suffusion the exposed spinal cord remained covered with the solution. In other experiments the drugs were injected intravenously (vena cava superior, dosage: DSLET, 0.5–3.6 mg kg−1; DAMGO, 1.2–2 mg kg−1; MR2034, 1–3 mg kg−1).

In most experiments, after recovery from the first application (generally after about 3–10 h, average time 7.7 h) the drug was applied a second, sometimes even a third, time. Between local applications the opioid solution was washed out and the spinal cord was covered with Ringer's solution for at least 1 h. If not mentioned otherwise, in the figures, according to the technique used, the time course of the influence of the opioids (and naloxone) is shown for the different reflex pathways in parallel, i.e. before, during and after the same application of a drug.

Fictive locomotion. In experiments with investigations on Pl activity during fictive locomotion, neurograms were recorded from the nerves to PBSt, GS, Pl, FDL and DP (partly on both sides). Stimulation of nociceptive afferents and cutaneous nerves was performed in the same way as in the other experiments. Fictive locomotion was induced by i.v. injection of 100 mg kg−1 nialamide and 40–100 mg kg−1l-DOPA (cf. Koehler et al. 1984).

RESULTS

Excitatory and inhibitory nociceptive effects from the central foot pad

According to the prevailing flexor reflex pattern in spinal cats, flexor motoneurones (PBSt and DP) received an almost pure excitation from nociceptive afferents of the central foot pad activated by noxious radiant heat (48–55 °C). This nociceptive excitation was observed not only by intracellular recording but also by measuring the overall motoneuronal excitability by monosynaptic reflex testing (Fig. 1A). In contrast, extensor motoneurones (GS, FDL) were clearly inhibited, when their monosynaptic test reflexes were conditioned by noxious radiant heat applied to the central pad under control conditions (Fig. 1B), i.e. without any pharmacological treatment (confirming results of Schomburg & Steffens, 1986, 1998; Schmidt et al. 1991; Steffens & Schomburg, 1993). The only stable exceptions were observed in the extensor Pl and the intrinsic foot extensors innervated by Tib. In about 90 % of tests they received a pure excitation from nociceptive afferents of the pad. This excitation was the predominant result not only in experiments based on monosynaptic reflex testing (Fig. 1 and Fig. 4) but also with intracellular recording from Pl and Tib motoneurones (Fig. 2A). In Fig. 2 the responses of a Pl motoneurone to different stimuli are compared. Noxious radiant heat (53 °C) applied to the central pad caused a distinct depolarisation with an initial volley of discharges. Light stroking of the plantar side of the foot also caused depolarisation and some discharges. These latter effects are comparable to those described by Engberg (1964). In all other respects the Pl motoneurone reacted as a typical extensor motoneurone. It received IPSPs from electrically stimulated FRAs: from group II (Fig. 2CG, upper/ middle traces) and group III (Fig. 2CG, lowest traces) afferents of different muscles; and from group II and III cutaneous afferents (Fig. 2I and J). This extraordinary divergence between the effects from nociceptive afferents of the pad and those from other FRAs to Pl was confirmed by investigating the reaction of the mass activity of the motoneurone pools with monosynaptic reflex testing. Thus, the monosynaptic reflexes of PBSt and GS reacted in the same way with only excitation (PBSt) or inhibition (GS) to either noxious radiant heat applied to the central pad or electrical stimulation of low threshold cutaneous (SPC, Sur) and joint or medium threshold (group II and III) muscle afferents. In contrast, the Pl monosynaptic reflex was clearly facilitated by the nociceptive input, but inhibited by conditioning electrical stimulation of the other afferents (see Schmidt et al. 1991; Schomburg & Steffens, 1998).

Figure 1.

Nociceptive afferents of the central pad (A–C) and of the dorsum of the foot (D–F) exert non-specific FRA effects onto PBSt (A and D) and LGS (B andE), but locally specific antagonistic effects onto Pl (C and F), in a high spinal cat

Conditioning effect of noxious radiant heat (50.6 °C) applied to the central pad (innervated by Tib, A–C) and dorsum of the foot (innervated by SPC, D–F) on monosynaptic reflexes of PBSt (A and D, both facilitatory), LGS (B and E, both inhibitory) and of Pl (C, excitation from nociceptive afferents of the central foot pad; F, inhibition from those of the dorsum of the foot). Dots represent the amplitude of the monosynaptic reflexes expressed as a percentage of the averaged control values before heat stimulation. The continuous line connects the values of a sliding average derived by averaging the amplitudes of seven consecutive reflexes and for each point advancing the series averaged by one reflex. Due to this technique the line may rise or decline shortly before the heat stimulus. The two-peaked facilitation observed in C was not a general finding.

Figure 4.

Depressing influence of the opioid DAMGO (μ-receptor agonist) on nociceptive FRA pathways with negative feedback (A, PBSt; B, GS) but not on nociceptive non-FRA pathways with positive feedback (C, Pl; D, Tib)

Time course of DAMGO action investigated with monosynaptic reflex testing of the different motoneurone pools in parallel, i.e. all reflexes were tested before and after the same injection of DAMGO (2 mg kg−1i.v.). The dots represent the amplitude of the unconditioned monosynaptic reflexes, the vertical bars the amount of facilitation (upwards) or inhibition (downwards) induced by conditioning with nociceptive afferents of the central foot pad (noxious heat 50 °C). The scale of the ordinate is given in arbitrary units.

Figure 2.

Excitation of a Pl motoneurone by nociceptive afferents from the central foot pad (A) and low threshold mechanosensitive afferents of the plantar side of the foot (B), but inhibition from other flexor reflex afferents (group II and III muscle afferents, C–G; cutaneous afferents, I–J) of the hindlimb

Intracellular recording of a Pl motoneurone in a high spinal cat. A, activation of nociceptive afferents of the central pad by radiant heat, 53 °C; B, activation of low threshold mechanosensitive afferents of the plantar side of the foot by light stroking with a soft brush. C–J, electrical stimulation of the nerves at the indicated strength. In C–G with 2T mainly Ib afferents are maximally excited, with possibly some group II contribution; with 5T additional activation of group II afferents; 20T is well above group II maximum in the low to medium group III range. Stimulus strength is indicated in multiples of the threshold strength T.

In former experiments, it had been demonstrated by spatial facilitation that nociceptive afferents of the pad and other FRAs (low threshold cutaneous and joint afferents and medium threshold muscle afferents) converge onto common interneurones in reflex pathways to PBSt and GS (Steffens & Schomburg, 1993). In order to test whether the nociceptive afferents interact with the FRA pathways, the spatial interactions between these afferents were tested. In confirmation of the former results the FRA-evoked EPSPs in PBSt motoneurones and the IPSPs in GS motoneurones were spatially facilitated by noxious radiant heat applied to the pad. However, for Pl motoneurones, in spite of the excitatory influence of the noxious stimulus to the pad on the membrane potential, there was a spatial facilitatory effect of this stimulus on the IPSPs evoked by FRA stimulation. Such a spatial facilitation of FRA-evoked IPSPs in Pl and Tib motoneurones was a common finding. As shown in Fig. 3, the spatial facilitation not only concerned IPSPs evoked from cutaneous afferents (Fig. 3EF) but also those from other FRA, joint (Fig. 3AD) and group II muscle afferents (Fig. 3G), as well as from Ib afferents (Fig. 3H). Such facilitation could even occur if there was no clear effect of the control stimulation of a joint (Fig. 3AB), cutaneous (Fig. 3E) or muscle nerve (Fig. 3H). Moreover, a clear spatial facilitation was also observed even when radiant heat by itself only evoked a liminal depolarisation of the membrane potential. The impression of a parallel excitatory and inhibitory access from nociceptive afferents of the central foot pad to foot extensors was further strengthened by the finding in one experiment that the Pl monosynaptic reflex showed a characteristic distinct facilitation which, during the course of the experiment, turned to a stable inhibition without any recognizable reason. Moreover, in several cases predominant inhibition of the monosynaptic reflexes to the extensors turned to a predominant excitation after application of opioids (see below).

Figure 3.

Spatial facilitation of inhibitory FRA pathways (A–D, joint afferents; E and F, cutaneous afferents; G, group II afferents from Q) and a group Ib pathway (H, GS) by nociceptive afferents from the central foot pad (radiant heat, 53 °C) in a Pl motoneurone

Note that low threshold joint (A), cutaneous (E) or muscle (H) nerve stimulation did not evoke any effect in the controls without conditioning nociceptive stimulation. ‘Both’: superimposition of the control recording (upper trace) and the conditioned response (second trace).

The nociceptive heat threshold for evoking facilitation from the central pad to PBSt and Pl or Tib was identical (40–43 °C, cf. Schomburg & Steffens, 1986; Steffens & Schomburg, 1993) and each motoneurone pool received increasing facilitation when the temperature was gradually increased to about 52 °C. However, if the temperature was raised to 60 °C (as done at the end of four experiments), the facilitation of PBSt showed a distinct further increase while the facilitation of Pl and Tib was in three cases no longer enhanced but rather diminished (cf. Fig. 5).

Figure 5.

Influence of nociceptive afferents of the central foot pad on spinal locomotor activity. Fictive spinal locomotion induced in a spinal cat by injection of nialamide (100 mg kg−1i.v.) and l-DOPA (85 mg kg−1i.v.)

Stimulation period marked by a bar below the records. A, before l-DOPA (7 h after nialamide) only tonic activation of PBSt and Pl, and inhibition of GS by the nociceptive input. B, facilitation of rhythmic motor activity after l-DOPA, and crescendo type of activity during active phase in ipsilateral Pl (iPl), but not in the other nerves. C, facilitatory effect with raised locomotor frequency, but inhibition of Pl induced by increased temperature (60 °C compared to 52 °C in A and B). D, comparative stimulation of a cutaneous nerve (Sur 5T).

Local specificity of the nociceptive excitatory and inhibitory effects to foot extensors

While the predominant facilitation of flexor motoneurones (PBSt and DP) and the predominant inhibition of extensor motoneurones (GS and FDL) were evoked by nociceptive afferents from different skin areas of the hindlimb innervated by Sur (lateral side of the foot), SPC (dorsum of the foot) or Tib (central pad) (Fig. 1A, B and D, E for central pad and dorsum of the foot), the facilitation of Pl and Tib motoneurones was only stable with activation of nociceptors of the central pad of the foot (cf. Fig. 1C and F), but could partly also be evoked from nociceptors of the toe pads. The effect of noxious radiant heat stimulation of the different toe pads showed no significant dependence on location with respect to either the excitation of PBSt or the inhibition of GS. However, the facilitation to Pl was only predominant for noxious stimuli applied to the lateral toe pads (four and five), whereas stimulation of the medial toe pads caused inhibition (three experiments) or minor facilitation (two experiments). For Tib a facilitation from the lateral toe pads was small or absent, while a predominant inhibition from the medial pads occurred. Motoneurones of the peroneal muscles received a clear facilitation from the central pad and the lateral toe pads, with less or no facilitation from the medial toe pads.

Enkephalinergic and monoaminergic control of the excitatory and inhibitory nociceptive pathways

As described previously, the nociceptive (and non-nociceptive FRA) facilitation of PBSt and, to a lesser extent, the inhibition of GS are distinctly depressed by the δ-opioid agonist DSLET (Schmidt et al. 1991) and by l-DOPA (Schomburg & Steffens, 1998) while the excitatory effect from nociceptors of the central pad to Pl and Tib remained unaffected by these drugs, at least with a comparable dosage. The μ-opioid receptor agonistic enkephalin DAMGO, the κ-agonist MR 2034, and the α2-receptor agonist clonidine also have a similar differentiating effect. In Fig. 4 the nociceptive (central pad) facilitation of PBSt and the corresponding inhibition of GS were distinctly depressed by DAMGO (2 mg kg−1i.v.) while the excitation of Pl and Tib from the nociceptive afferents of the central pad was not affected or even slightly increased. Correspondingly, although naloxone had a clearly antagonising effect on all of the depressing effects of DSLET and DAMGO, it was without effect on the opioid-resistant nociceptive facilitatory action to Pl and Tib, even when it antagonised the slight depressing action of the opioids onto the size of the unconditioned monosynaptic reflexes of these muscles (cf. Schmidt et al. 1991).

Table 1 gives a summarising overview of the mean depressing action of the different opioid receptor agonists and monoamines on the facilitation evoked by noxious radiant heat to the central pad on Pl and PBSt monosynaptic reflexes. Even if the depression of the nociceptive excitatory non-FRA pathway to Pl is different for the different substances, it is evident that in all cases, at the doses used, this depression is distinctly smaller than the depression of the facilitatory FRA pathway to PBSt (the greatest difference being observed with l-DOPA). Similar results to those for Pl were obtained for the excitatory nociceptive pathway to Tib. With higher i.v. doses, or higher concentrations of the opioids superfused over the spinal cord, the transmission in the excitatory pathway to Pl or Tib was also depressed to a distinct degree.

Table 1. Amount of mean percentile depression of the transmission in nociceptive excitatory pathways from the central pad to Pl and PBSt by different opioidergic and monoaminergic drugs determined by monosynaptic reflex testing
 DSLET*δ-agonistDAMGO μ-agonistMR2034 κ-agonistL-DOPA**Clonidine α2-agonist
 0.5–3.6 mg kg−1 or 10−5–10−4m1.2–2 mg kg−1 or 10−5–10−4m1–3 mg kg−140–100 mg kg−10.02–0.045 mg kg−1
  1. Doses of intravenous injection are given in mg kg−1; in cases of spinal superfusion with opioid solutions the concentration is indicated in m. Values in parentheses are number of experiments. * From Schmidt et al. 1991; ** from Schomburg & Steffens, 1998.

 (14)(11)(6)(9)(8)
Pl614252.517
PBSt7482549189

Two observations with noxious radiant heat stimulation of the central pad should be particularly mentioned in this context. (1) The facilitation of Pl or Tib was enhanced after application of the opioids (DSLET, DAMGO and MR2034) or l -DOPA in about one-fifth of cases with the low dosage given in the table; cf. Fig. 4. (2) In three experiments a slight inhibition of Pl in the controls turned to a clear excitation after opioids. Similarly, in several cases the inhibition of GS (eight cases) or FDL (11 cases) monosynaptic reflexes turned to a slight facilitation after the application of the opioids (DSLET, DAMGO, MR2934).

Nociceptive influence on locomotion

In order to reveal the functional importance of the parallel inhibitory and excitatory nociceptive pathways from the pad to foot extensors, the influence of nociceptive afferents from the pad on fictive locomotor activity, particularly of Pl, was investigated in four experiments. In the control records of all of these experiments (after nialamide, but before l-DOPA) the neurograms of PBSt and Pl showed a long-lasting tonic activation, while the activity in the GS nerve was clearly depressed (Fig. 5A). After l-DOPA the already pronounced rhythmic modulation of the efferent activity was clearly facilitated and accentuated by noxious stimulation of the pad with some characteristic features shown in Fig. 5. In Fig. 5A (control before l-DOPA, 7 h after nialamide 100 mg kg−1i.v.) PBSt and Pl were tonically activated during application of noxious radiant heat to the central pad. After l-DOPA (85 mg kg−1i.v., B 48 min, C 63 min, D 56 min) in this case only a weak rhythmic modulation of the efferent activity occurred spontaneously. Noxious stimulation of the pad (Fig. 5B and C) distinctly increased the rhythmic motor activity, which developed a clear locomotor pattern with a flexor-extensor (PBSt; GS, Pl) and an ipsilateral-contralateral alternation. With a heat stimulus of 52 °C the ipsilateral Pl (iPl) activity developed a crescendo characteristic of increasing amplitude, which lasted into the flexion phase, while the activity in the other nerves, including contralateral Pl (cPl), showed a characteristic decrescendo activity in the course of their active phase. With higher heat stimulus (at 60 °C, with the duration reduced to avoid greater skin damage) the frequency of locomotor activity is more increased than with the lower temperature, but the iPl activity was distinctly reduced and the crescendo effect less pronounced. If the effectiveness of noxious heat stimulation on locomotor activation (Fig. 5B and C) is compared to that of cutaneous skin stimulation (5T, mainly within the group II range, Fig. 5D) the latter is distinctly higher. The crescendo effect of the iPl activity could also occur with Sur nerve stimulation. Compared to the latency of the action of noxious pad stimulation on locomotor activity, which generally exceeded 10 s, the latency for motoneuronal excitation or inhibition (measured with intracellular recording or monosynaptic reflex testing) was distinctly less then 10 s (mainly 1–7 s).

DISCUSSION

The results revealed parallel excitatory and inhibitory reflex pathways from the central pad and partly from toe pads to foot extensors in the spinal cat: while the excitatory pathway is predominant to Pl and Tib, the inhibitory one is predominant to GS and FDL. The existence of the inhibitory pathway to Pl and Tib only became evident by the nociceptive spatial facilitation of FRA-evoked IPSPs in these motoneurones, and also by the observation that in some experiments an inhibition due to noxious stimulation turned into a facilitation after opioids. Moreover, in one experiment an originally excitatory effect spontaneously turned into an inhibitory one in the course of the experiment. Conversley, the excitatory pathway to GS and FDL mainly became evident after application of opioids. The persistence of the excitatory nociceptive pathways to foot extensors following opioids and l-DOPA, compared to the depression of the inhibitory pathways, indicated that the former are non-FRA pathways, while the latter are FRA pathways according to the work of Schmidt et al. (1991) and Schomburg & Steffens (1998). This assumption is confirmed by the finding that nociceptive afferents of the pad contribute to the multisensorial convergence in the inhibitory nociceptive FRA pathways to Pl. Thus, the excitatory nociceptive pathway to Pl and Tib is comparable to the mechanosensitive excitatory non-FRA pathway from the foot sole to extensors described by Engberg (1964).

As stated in the introduction the existence of ‘private’ pathways for nociceptive afferents, in addition to their contribution to common FRA pathways, is well in accordance with Lundberg's FRA concept (Holmqvist & Lundberg, 1961; Lundberg, 1979; Lundberg et al. 1987b). According to this concept, the parallel existence of excitatory and inhibitory pathways from a specific type of afferent to a motoneurone pool is not a vagary of the spinal cord but rather demonstrates the great degree of functional freedom and variability in motor control, necessary for the performance of the great variety of complex movements. It can be assumed that the transmission in the two parallel pathways is controlled by descending commands and actual spinal conditions, as also is the case for transmission in the parallel excitatory and inhibitory segmental pathways from low threshold cutaneous afferents, which depends on the phase of the step cycle during fictive locomotion (Schomburg & Behrends, 1978; for further discussion and references see Schomburg, 1990).

The existence of two antagonistic nociceptive pathways to foot extensors with distinctly different sensitivity to opioids and l-DOPA could indicate that they were evoked by two different nociceptive fibre groups, Aδ- and C-fibres. However, recent experiments clearly demonstrated that both nociceptive FRA pathways and the nociceptive excitatory non-FRA pathways are evoked by Aδ-fibres and TTX-resistant C-fibres (Schomburg et al. 2000). According to these findings the long latency of the locomotor reaction to noxious heat stimulation of the pad would indicate that this effect is mainly evoked by a C-fibre input.

Functionally the inhibitory nociceptive pathway from the pad to foot extensors fits well into the withdrawal function of the nociceptive flexion reflex (Creed et al. 1932; Schouenborg et al. 1990, 1994; Weng & Schouenborg, 1996). However, by opposing this function, the excitatory nociceptive pathway would cause a movement of the affected area towards the stimulus or at least a resistance against the stimulus, i.e. it would induce a positive feedback. This positive feedback would get support from the corresponding mechanosensitive afferents of the foot sole (Engberg, 1964), particularly, since spatial facilitation between the nociceptive and mechanosensitive input from the pad can be assumed (Behrends et al. 1983). Like the positive feedback from Ib afferents (Pearson & Collins, 1993; Gossard et al. 1994; Pratt, 1995) positive feedback from nociceptive afferents of the pad may become effective during particular phases of the locomotor cycle. The facilitation of the crescendo effect of the Pl activity by nociceptive afferents of the pad in the late extension phase, and at the transition to the flexion phase during fictive locomotion, would indicate a support function of the positive feedback in a phase of maximal propulsion. During normal locomotor activity the positive feedback is probably only mediated by mechanosensitive afferents. However, during forced propulsion and even more likely during landing of the cat from greater heights, nociceptive afferents are probably additionally activated. It is during just these situations that a positive feedback is required to stabilise the foot, in order to avoid collapse, as has similarly been postulated for the positive force feedback from Ib afferents (Prochazka et al. 1997). The observed partial inhibition, or at least not increased excitation, to Pl and Tib with increasing pad temperature could indicate that the positive feedback can be overridden by the negative one if the noxious input exceeds a critical level, causing then a collapse of the animal. A similar transition from facilitation to inhibition with increasing noxious stimulation was partly observed in humans from the heel to soleus and intrinsic foot extensors (Ellrich et al. 2000).

The distinct facilitation of fictive locomotor activity by nociceptive afferents indicates that these afferents, like other FRAs (Schomburg et al. 1998), have access to the spinal locomotor generator. However, similar to the effect from group III and IV muscle afferents (Kniffki et al. 1981), their impact on locomotor activity is distinctly smaller than that of cutaneous nerve stimulation.

Acknowledgements

The careful revision of the manuscript by Professor Thomas A. Sears (King's College London) is gratefully acknowledged. The investigations were supported by the Deutsche Forschungsgemeinschaft. N.W. received a grant of the Alexander von Humboldt-Stiftung.

Ancillary