INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Lamina I, the most superficial layer of the spinal or trigeminal dorsal horn, receives modality-selective A- and C-fiber afferent input from somatic and visceral tissues (Craig 2000). Anatomic findings indicate that lamina I neurons are the major source of spinobulbar projections (Krout and Craig 1996) and project to specific brain stem sites involved in homeostatic regulation (Craig 1995). Thus lamina I may provide an important link for the somato-autonomic reflex responses evoked by somatic and visceral small-diameter afferent input (Sato and Schmidt 1973).

The brain stem plays a pivotal role in homeostatic responses, including cardiovascular and respiratory responses to muscle work (Iwamoto et al. 1985). Static contraction of skeletal muscle activates small-diameter afferents that evoke a reflex increase in sympathetic nerve activity and cardiovascular function (Kaufman and Forster 1996; Mitchell and Schmidt 1983; Wilson and Hand 1997). Activation of spinobulbar dorsal horn neurons is required for this somato-autonomic reflex, the "exercise pressor reflex" (Wilson 2001). Small-diameter muscle afferents terminate in the dorsal horn of the spinal cord, particularly within lamina I (Craig and Mense 1983). Spinobulbar lamina I neurons have never been recorded. Spinothalamic lamina I neurons excited by group III muscle afferents and noxious muscle stimulation were identified previously (Craig and Kniffki 1985), but activation of lamina I neurons by muscle contraction has never been demonstrated; this requires maintaining microelectrode recordings from these small spinal neurons during hindlimb muscle contractions strong enough to elicit blood pressure changes in an unparalyzed animal under anesthesia. The purpose of this study was to determine whether excitation of identified spinobulbar lamina I neurons by static contraction of skeletal muscle can be demonstrated.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Five adult cats (3.5-4.2 kg) were premedicated with ketamine (25 mg/kg im) and anesthetized with -chloralose (80 mg/kg) and urethan (100 mg/kg) administered via a cannula in the cephalic vein. Cannulae were inserted into the carotid artery (to record blood pressure) and into the trachea (for artificial respiration). The animals were paralyzed during surgery with pancuronium (400 µg/h) and ventilated; end-tidal CO₂ was maintained near 4.0%. Body temperature was monitored using a rectal thermistor and maintained at 37.5°C.

A laminectomy was performed to expose the lumbosacral spinal cord, and a pool filled with warm (38°C) Tyrode's solution was constructed. The medulla was exposed by enlarging the foramen magnum. A bipolar electrode (Rhodes NE- or NEX-100) was inserted into the left caudal ventrolateral medulla (CVLM) to antidromically activate spinobulbar neurons. The left calcaneal bone was cut, and the Achilles tendon was connected to a transducer (Grass model FT10) to measure the tension developed during contraction of the triceps surae muscle. The patellar tendon was secured to a post to ensure an isometric contraction. The popliteal fossa was exposed, and the tibial nerve was placed on platinum wire electrodes (LeDoux and Wilson 2001). A pool made around the exposed nerves and muscles was filled with warm mineral oil. The preparation is shown diagrammatically in Fig. 1. Once the preparation was complete, the paralytic was allowed to wear off.

View larger version (19K): [in this window] [in a new window]   Fig. 1. Diagram of the preparation. The activity of single lamina I neurons was recorded extracellularly with microelectrodes (R). Electrical stimuli (S) were applied to the caudal ventrolateral medulla (CVLM) to test for neuronal projections to the brain stem and to the axons of motoneurons in the tibial nerve to evoke a static contraction of the triceps surae muscle. A tension transducer (T) attached to the free end of the muscle was used to measure the force developed during a contraction.

Glass-insulated tungsten microelectrodes were used to record extracellularly from single lamina I neurons, identified by ongoing unit activity or by antidromic search stimuli (see Craig et al. 2001). Bipolar or monopolar electrical stimuli were delivered from the electrode in the CVLM to determine whether a neuron projected to the brain stem by antidromic activation. Two stimulus pulses of 400 µA and 1-ms duration were delivered at 100 Hz; stronger stimuli that evoked movement were avoided unless pancuronium could be administered. Every unit that followed two pulses at 100 Hz also followed a train of six antidromic stimuli at 250 Hz when tested (Fig. 2A); collision was also observed between naturally and antidromically evoked impulses (Fig. 2B). Each unit was characterized using mechanical and thermal stimulation applied over the entire hindlimb (Craig et al. 2001), and each cell was tested for excitation by muscle contraction. Muscle contraction was evoked by stimulation of the tibial nerve with a 30-Hz train of 25-µs pulses at 1.3 times motor threshold for 10-30 s. These stimulus parameters excite motor axons but do not excite A or C afferent fibers (Degtyarenko and Kaufman 2000; Kaufman and Forster 1996). Only moderate contractions that evoked moderate blood pressure changes were evoked because of the need to maintain stable single-unit microelectrode recordings.

View larger version (23K): [in this window] [in a new window]   Fig. 2. Example of the properties of a lamina I spinobulbar neuron and its response to static contraction. A: a pair of traces showing reproducible 1-for-1 following of a train of 6 antidromic stimuli (180 µA, 1 ms; ) delivered at 250-Hz from an electrode in the ipsilateral CVLM. B: collision of the 1st impulse of a train of 3 antidromic action potentials (200 Hz, ) with a preceding spontaneously occurring impulse (*). , the time at which the 1st antidromic response would have occurred. C: reconstruction of the recording site in the L6 segment of the spinal cord showing an electrolytic lesion localized to lamina I. D: diagram of the stimulating site in the medulla: Cu, cuneate nucleus; EC, external cuneate nucleus; G, gracile nucleus; LR, lateral reticular formation; IO, inferior olive; V, trigeminal nucleus, X, vagal nucleus, XII, hypoglossal nucleus. E: response of the same neuron to static contraction of the triceps surae muscle. The traces are, from the top downward: arterial blood pressure; tension in the triceps surae, the neural record and a histogram of the unit's activity (1-s bins). The thickening of the neural recording is due to stimulus artifacts during muscle stimulation (1.3× motor threshold, 25 95 s, 30 Hz for 30 s). This neuron received input only from group III muscle afferents; its central conduction velocity was 2.2 m/s; and, it had no cutaneous receptive field.

Electrolytic lesions (+20-50 µA, 10 s) were made at spinal recording sites and at stimulation sites in the medulla. The tissue blocks were fixed in formalin, and the lesions were recovered in 50-µm frozen sections that were stained with thionin (Fig. 2, C and D).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Recordings were obtained from 17 lamina I neurons identified as spinobulbar neurons by antidromic activation from the CVLM and 5 lamina I cells with unidentified projections. The average central conduction velocity of the spinobulbar lamina I neurons was 11.3 ± 6.2 (SD) m/s (n = 17; range, 2.2-20.8), which is significantly faster than spinothalamic lamina I neurons (mean, 4.7 ± 2.2 m/s, n = 186, P < 10 ⁷, unpaired t-test) (Andrew and Craig 2001). Unitary recordings from 13 spinobulbar lamina I neurons and 4 other lamina I neurons were successfully maintained during muscle contraction.

Figure 2 shows the response to muscle contraction of an antidromically identified and histologically confirmed lamina I spinobulbar neuron. This unit began to respond ~15 s after the onset of static contraction of the triceps surae muscle (Fig. 2E); its activity was maintained >1 min following the contraction. It did not respond to cutaneous mechanical or thermal innocuous or noxious stimulation or to blood pressure changes these stimuli evoked or to tapping or moving the hindlimb, but it did respond to strong squeezing and stretching of the triceps surae muscle. It received time-locked (monosynaptic) input from group III (A) tibial afferent fibers (conduction velocity, 18.5 m/s) but not from group IV (C) fibers (determined during paralysis with paired 1-mA, 1-ms pulses at 120 Hz).

The responses of lamina I neurons to static muscle contraction varied; 9 of the 17 neurons tested were excited, 2 were inhibited and 6 were unaffected. Of the 13 spinobulbar lamina I neurons tested, 5 were excited by muscle contraction. Of these, 3 had no detectable cutaneous receptive field, whereas 1 was a polymodal nociceptive neuron (HPC) (Craig et al. 2001) that responded to noxious heat and pinch and noxious cold on the hindpaw, and one was a wide-dynamic-range (WDR) neuron that responded to low- and high-threshold stimulation of the foot. Of the four lamina I neurons with unidentified projections that were excited by contraction, one received input only from muscle afferents and three had HPC cutaneous receptive properties. Most (4/6) of the neurons that were unaffected by contraction were nociceptive-specific (NS) cells, all of which projected to the CVLM. (Two had receptive fields on the leg near the muscle exposure, indicating that muscle-responsive cells without cutaneous fields did not result from surgical damage to hindlimb cutaneous afferents.) The mean central conduction velocity of the spinobulbar neurons excited by contraction (9.7 m/s; range, 2.2-19.3, SD 7.4) was not different from that of the other spinobulbar neurons (mean 10.9; range, 4.1-16.9, SD 6.0; P > 0.7, unpaired t-test).

Various responses of spinobulbar and other lamina I neurons to muscle contraction are shown in Figs. 2 and 3. These neurons responded during the contraction (Fig. 3A), after the contraction (Fig. 3B) or both (Figs. 2E and 3C), and their responses paralleled the tension developed in the muscle (Fig. 3, A and C), the centrally evoked cardiovascular reflex (Fig. 3A), or both (Fig. 3A). Interestingly, rhythmic muscle contraction (5 Hz), which is a potent stimulus for group IV (C) afferents (Adreani et al. 1997), was the most effective stimulus for activation of the cell shown in Fig. 3C.

View larger version (20K): [in this window] [in a new window]   Fig. 3. Examples of the variety of responses of lamina I neurons that were excited by muscle contraction. The traces show from the top downward: the tension developed in the triceps surae muscle, the discharge of the single neuron histogrammed in 1 s bins and either mean blood pressure (A) or the neural recording (B and C). A: response from an HPC lamina I neuron with unidentified projections that was excited by contraction that paralleled the tension developed in the muscle and the centrally-evoked cardiovascular reflex. B: response of another HPC lamina I neuron with unidentified projections that was only excited after the contraction. C: response of a spinobulbar lamina I neuron (central CV 2.2 m/s) that was excited during and after the contraction; this neuron had no identifiable cutaneous receptive field but responded to stretching the triceps muscle and also to rhythmic contractions at 5 Hz.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

These are the first records of lamina I responses to static contraction of skeletal muscle as well as the first recordings from identified spinobulbar lamina I neurons. These technical advances notwithstanding, the most striking and novel result is the documentation of spinobulbar lamina I neurons that responded to muscle contraction, particularly those which responded selectively and had no cutaneous receptive field. Such neurons provide an ascending pathway for the cardiovascular and respiratory responses that are reflexly elicited by muscular work. This evidence directly supports the proposal that modality-selective lamina I neurons are an afferent link for homeostatic responses to changes in the physiologic condition of the body, consistent with a fundamental role in homeostasis and interoception, of which pain, temperature, itch and muscular sensations are particular aspects (Andrew and Craig 2001; Craig 1996, 2000; Craig et al. 2001; Sherrington 1900).

Anatomic evidence indicates that lamina I receives dense input from small-diameter muscle afferents and also that lamina I provides the predominant spinal input to specific homeostatic brain stem sites bilaterally (Craig 1995, 2000; Craig and Mense 1983; Krout and Craig 1996). We identified these spinobulbar lamina I projection neurons by using antidromic activation from the ipsilateral CVLM because lamina I axons that ascend to the mesencephalon or thalamus are almost all contralateral (Craig 2000). The lamina I neurons we recorded that had unidentified projections (i.e., not antidromically activated from the ipsilateral CVLM nor in some cases from the contralateral thalamus) could also have included contralateral spinobulbar cells. Lamina I spinobulbar axons terminate in the caudal and rostral ventrolateral medulla, the nucleus of the solitary tract, the retrotrapezoidal region, and other homeostatic sites in the brain stem (Craig 1995; Westlund and Craig 1996). The neurons we antidromically activated from the ipsilateral CVLM could have projected to any of these sites (and are therefore termed spinobulbar rather than spinomedullary), and the anatomic data indicate that they probably did not project any further rostrally (Craig 1995). Therefore a role for these neurons in somato-autonomic reflexes (Li et al. 2001; Stornetta et al. 1989) and homeostatic integration is highly likely.

A recent study reported contraction-evoked activation of neurons in the deep dorsal horn (Degtyarenko and Kaufman 2000) that was reduced by stimulation of the mesencephalic locomotor region, similar to the effects of motor commands on afferent-induced cardiovascular responses. In contrast to modality-selective lamina I neurons, deep dorsal horn neurons are modality-ambiguous and integrate nearly all somatic afferent inflow (Carstens 1997). Further, the axonal projections of such cells were not identified, and thus their role in brain stem somato-autonomic integration is unclear.

The present findings show that spinobulbar lamina I neurons have appropriate characteristics to engage homeostatic responses to muscle exercise. Exercise demands rapid cardiovascular and respiratory adjustments. Exercise-evoked cardiorespiratory responses are mediated in part by central motor commands and in part by the "exercise pressor reflex," which originates in small-diameter group III and IV (A and C) afferents from active muscles (Iwamoto et al. 1985; Mitchell and Schmidt 1983; Wilson and Hand 1997). These receptors are either directly excited by contraction (mechanosensitive) or indirectly excited by metabolites released from the contracting muscle (metabo-receptive; Kaufman and Forster 1996; Kaufman et al. 1983; Kniffki et al. 1981; McCloskey and Mitchell 1972). The central neural correlates of both mechanisms were observed in the present study. Some lamina I neurons were excited immediately following the onset of contraction and their activity paralleled muscle tension, similar to mechanosensitive small-diameter muscle afferents. Other neurons were excited late during a contraction or after the contraction, similar to metabo-receptive muscle afferents.

In addition, most (3/5) of the spinobulbar lamina I neurons in this initial sample that responded to muscle contraction had no cutaneous input, and conversely, the spinobulbar cells that responded selectively to cutaneous nociceptors (NS cells) did not respond to contraction. This observation supports the possibility that the muscle afferents that initiate reflex cardiorespiratory changes produced by muscular work are "ergoreceptors" that are distinct from nociceptors (see Kniffki et al. 1981; Mitchell and Schmidt 1983), an idea that has received recent experimental support (LeDoux and Wilson 2001). This observation is also consistent with the general modality-selectivity of lamina I neurons (Craig 2000). On the other hand, muscle-responsive lamina I cells that did have cutaneous input were nearly all HPC cells, consistent with the possibility that the activity in such polymodal nociceptive cells generally represents increased regional metabolic needs (Craig 1996). Further investigations can address these issues more fully.

Our observations demonstrate that spinobulbar lamina I neurons associated with homeostasis can be investigated. Detailed analyses of this unexplored pathway are needed. Only lamina I neurons that project to the mesencephalon (Hylden et al. 1986) or thalamus (Craig et al. 2001) have been characterized before. Lamina I receives dense afferent input not only from skin and muscle but also from the viscera and other tissues of the body, and it receives descending controls from homeostatic brain stem sites and from the hypothalamus (Craig 2000). Thus lamina I spinobulbar neurons could be involved in many somato-autonomic and homeostatic reflexes. Significantly, recent retrograde labeling results indicate that contralateral spinothalamic and ipsilateral spinobulbar lamina I neurons are completely separate populations (Andrew and Craig 2000). Our present physiologic observations support this finding, because spinobulbar lamina I neurons have characteristics not observed in spinothalamic lamina I neurons (i.e., contraction-responsive, wide-dynamic-range, faster conduction velocities), and because spinobulbar lamina I neurons are not antidromically activated from the contralateral thalamus (0 of 10 tested) (Andrew and Craig, unpublished observations).

To conclude, we demonstrated that spinobulbar lamina I neurons can be identified that are activated by muscle contraction; such neurons provide an ascending pathway for the automatic cardiorespiratory adjustments to muscular work, that is, the exercise pressor reflex. This study provides the foundation for investigations of spinobulbar mechanisms that mediate homeostatic adjustments elicited by small-diameter afferent activity from all tissues of the body. It remains to be established how other types of somatic and visceral afferent activity are represented in this unexplored pathway.