resulted in a pronounced increase in cutaneous blood ... only under conditions of noxious stimulation, there is ... ducer LM-22, LIST, Darmstadt, Germany).
Exp Brain Res (1997) 113:402–410
© Springer-Verlag 1997
R E S E A R C H A RT I C L E
&roles:H.-J. Häbler · G. Wasner · W. Jänig
Interaction of sympathetic vasoconstriction and antidromic vasodilatation in the control of skin blood flow
&misc:Received: 24 April 1996 / Accepted: 3 September 1996
&p.1:Abstract We studied the interaction between the vasoconstriction evoked by postganglionic sympathetic neurones (sympathetic vasoconstriction) and the vasodilatation mediated by small-diameter afferent neurones (antidromic vasodilatation) in hairless skin of anaesthetized rats kept under controlled conditions. In all animals both the lumbar sympathetic trunk (LST) and the ipsilateral dorsal root (DR) L5 were surgically exposed, sectioned and electrically stimulated using different protocols. This experimental approach results in the exclusive and selective activation of sympathetic efferents and primary afferents respectively. Blood flow responses were measured using laser Doppler flowmetry. Sectioning the LST resulted in a pronounced increase in cutaneous blood flow by 112±15% (mean±SEM, n=25) indicating that ongoing sympathetic vasoconstrictor activity had been abolished. When a brief antidromic vasodilatation was produced by DR stimulation with 10–15 pulses at 1 Hz with C-fibre intensity during a sustained sympathetic vasoconstriction, peak blood flow reached preconstriction levels at LST stimulation frequencies of ≤3 Hz. By contrast, antidromic vasodilatation was reduced at sympathetic stimulation frequencies of ≥5 Hz and absent when stimulating the LST with 20 Hz. A similar response characteristic was obtained when LST and DR stimulation were started simultaneously. Continuous DR stimulation with 0.1 Hz evoked a substantial increase in cutaneous blood flow by 38±10% (mean±SEM, n=8) to a new baseline level. When sympathetic vasoconstriction was elicited on this background DR stimulation, the responses were smaller at all sympathetic frequencies. However, the maximum decrease in blood flow was significantly smaller than the controls at LST stimulation with ≤3 Hz but not at higher frequencies. We conclude that sympathetic vasoconstriction and antidromic vasodilatation are competitive influences in the control of cutaneous blood flow. At low levels of cutaneous sympathetH.-J. Häbler (✉) · G. Wasner · W. Jänig Physiologisches Institut, Christian-Albrechts-Universität, Olshausenstrasse 40, D-24098 Kiel, Germany; Tel.: +49 (431) 880 2037, Fax: +49 (431) 880 4580&/fn-block:
ic vasoconstrictor activity, which probably prevail under resting conditions in the absence of cold stress, antidromic vasodilatation overrides sympathetic vasoconstriction. At high levels of cutaneous sympathetic activity, which may be reached in normal life under the conditions of severe cold, sympathetic vasoconstriction can suppress antidromic vasodilatation almost totally. &kwd:Key words Vasoconstriction · Antidromic vasodilatation · Sympathetic nervous system · Axon reflex&bdy:
Introduction In mammals peripheral endings of small-diameter nociceptive afferents have not only sensory but also “efferent” functions (Holzer 1988, 1992). Excitation of these afferents elicits vasodilatation of precapillary resistance vessels and plasma extravasation from postcapillary venules. The afferent-mediated vasodilatation is the basis of the “flare” response in skin following noxious stimulation (Lewis 1927; Lisney and Bharali 1989; Chahl 1991). It can also be evoked experimentally by antidromic excitation of the afferents (Stricker 1876; Bayliss 1901; Szolcsányi 1988). Blood vessels in skin and in other tissues are supplied by capsaicin-sensitive sensory nerve fibres which contain the neuropeptides substance P and calcitonin gene-related peptide (CGRP) (Gibbins et al. 1985; Franco-Cereceda et al. 1987; Kruger et al. 1989; Kawasaki et al. 1990a; Maynard et al. 1990), and it may principally be the release of CGRP from the afferent endings which is responsible for the evoked vasodilatation (Brain et al. 1985; Kawasaki et al. 1988; Fujimori et al. 1989; Hughes and Brain 1991; Delay-Goyet et al. 1992; Kerezoudis et al. 1994). While it was originally thought that neuropeptides are released from afferents only under conditions of noxious stimulation, there is now evidence suggesting a basal release of CGRP, at least in some tissues (Gardiner et al. 1990; Han et al. 1990; Zochodne and Ho 1991). Furthermore, unmyelinated afferents with low mechanical thresholds have
403
been described in humans (Vallbo et al. 1993) which might be able to release neuropeptides, although it is still unclear whether all or only a subpopulation of thin afferents can evoke vasodilatation (Lynn and Cotsell 1992; Lynn et al. 1996). Resistance vessels that dilate during activation of unmyelinated afferents are also under control of postganglionic vasoconstrictor fibres. It is unclear in which way the two neural influences interact on the vascular bed. This question has been addressed in two studies on humans using laser Doppler flowmetry and thermography, respectively, to measure changes in blood flow through skin. Hornyak et al. (1990) found that manoeuvres which activated the sympathetic outflow to skin could diminish electrically evoked antidromic vasodilatation. Ochoa et al. (1993) concluded that sympathetic vasoconstriction overrides antidromic vasodilatation during simultaneous stimulation of both afferent and sympathetic fibres. After terminating stimulation they observed a long-lasting vasodilatation. Vasoconstriction induced by transmural nerve stimulation in the isolated rat mesenteric artery and rabbit central ear artery is enhanced by depleting neuropeptides from unmyelinated afferents with capsaicin (Kawasaki et al. 1990b; Moritoki et al. 1990; Li and Duckles 1991), while exogenously applied CGRP diminished vasoconstriction induced by electrical field stimulation or by exogenously applied noradrenaline (Maynard et al. 1990). These findings support the view that sympathetic vasoconstriction and antidromic vasodilatation are functionally competitive. Here we describe the interaction between vasoconstriction evoked by activation of sympathetic fibres and vasodilatation evoked by activation of unmyelinated afferent fibres in hairless skin of the rat hindpaw. Both nerve supplies were activated selectively and independently of each other by stimulation of the appropriate dorsal root (DR) and the lumbar sympathetic trunk (LST). The LST probably does not contain axons of afferents which project to the hindlimb (Baron et al. 1988). The changes in blood flow through skin were measured with a laser Doppler flowmeter. Preliminary results of the study have been published (Häbler and Wasner 1995).
nulated low in the neck. Animals were immobilized (pancuronium, Organon, initially 1 mg/kg i.v. and further doses of 0.4 mg/kg i.v. when necessary) and artificially ventilated with O 2-enriched room air. Tracheal pressure was monitored continuously. Blood gases and blood acid-base status were assessed (ABL 30, Radiometer, Copenhagen) every 2 h. CO2 tension was maintained between 30 and 40 mmHg, and pH was kept in the range 7.38–7.46 (Brun-Pascaud et al. 1982). Rectal temperature was maintained at 37°C by means of a servo-controlled heating blanket. At the end of the experiments the animals were killed under deep anaesthesia by an intravenous injection of a saturated solution of potassium chloride. All experiments had been approved by the local animal care committee of the state administration and were conducted in accordance with German Federal Law. The NIH principles of laboratory animal care were followed. Surgery The left LST was exposed between paravertebral ganglia L2 and L4 (Baron et al. 1988) using a retroperitoneal approach, carefully freed from connective tissue, cut caudally to ganglion L2 or L3 and placed on a pair of platinum hook electrodes for electrical stimulation. A laminectomy was performed extending from the lumbosacral transition to vertebra L2. The dura mater was opened, the left DR L5 was identified and cut as proximal as possible and also placed on a pair of stimulation electrodes. The arrangement of nerves and stimulation electrodes is schematically shown in Fig. 1. Cutting the DR L5 elicited a short-lived antidromic vasodilatation, confirming that the correct root had been chosen. A pool was formed from skin flaps and the exposed tissue was covered with warm mineral oil. Experimental protocols Blood flow was measured within the L5 innervation territory on the plantar skin of the left hindpaw using a laser Doppler flowmeter device (MBF3D, Moor Instruments, Axminster, Devon, UK). The time constant was set to 3 s. The output of this device does not provide absolute values of blood flow, but arbitrary units
Materials and methods Anaesthesia and animal maintenance The experiments were conducted on Wistar rats of either sex weighing 215–450 g, anaesthetized with pentobarbital sodium (Nembutal, initially 60 mg/kg i.p.). Further doses of anaesthetic (pentobarbital sodium in Tyrode’s solution 1:4 v/v, 10 mg/kg) were administered intravenously every hour through a catheter placed in the jugular vein. A sufficient depth of anaesthesia was judged from the absence of gross fluctuations in blood pressure and heart rate. In addition, when rats were in the paralysed state we let muscular paralysis wear off at intervals and tested for the absence of flexor reflexes. A second catheter was inserted into the ventral tail artery for continuous blood pressure recording (transducer LM-22, LIST, Darmstadt, Germany). The trachea was can-
Fig. 1 Schematic drawing of preparation and nerves which were stimulated electrically. After a lumbosacral laminectomy the cauda equina and the left lumbar sympathetic trunk (LST) were exposed. The peripheral ends of the centrally cut left dorsal root (DR) L5 and the ipsilateral LST between paravertebral ganglia L2 and L4 were stimulated electrically with trains of pulses at C-fibre strength. The axons of an afferent neurone and of a sympathetic preganglionic neurone connected to a postganglionic neurone are indicated by a continuous line and a broken line, respectively&ig.c:/f
404 which are called “flux”. In the following, “flow” and “flux” will be used synonymously. In protocol A (nine animals), a continuous vasoconstriction was performed for 6 min. The LST was stimulated with supramaximal pulses (10–15 V, pulse width 0.5 ms) at variable frequencies (0.1–20 Hz). One minute after starting stimulation (when steadystate vasoconstriction was reached) the DR was stimulated with a standard train of 10 or 15 pulses supramaximal for activating C-fibres (15–25 V, pulse width 0.5 ms) at a frequency of 1 Hz. In each experiment the number of pulses, once chosen, was then kept constant. This moderate number of stimuli applied to the DR was chosen because we wished to avoid the exhaustion of antidromic vasodilatation which is known to occur when frequent stimulations with long trains of pulses are used (Szolcsányi et al. 1992). Care was taken that vasoconstriction was initiated from similar values of baseline flux. To this end intervals of 10–15 min were allowed between stimulations. In a second series of experiments (protocol B, eight animals) stimulation of the DR L5 (standard stimulation with 10 or 15 pulses at 1 Hz, as in protocol A) and stimulation of the LST were started simultaneously. LST stimulation had a standard 50 s duration which in earlier experiments had proved appropriate to lead to a steady-state vasoconstriction in most cases. LST stimulation frequency was varied as in protocol A. Both a control vasoconstriction and a control vasodilatation were performed prior to or following each simultaneous stimulation, using identical stimulation parameters. In protocol C (eight animals) a continuous background vasodilatation was elicited by repetitively stimulating the DR L5 at 0.1 Hz. When a steady-state flow was reached the LST was stimulated with the different frequencies as above. A series of control vasoconstrictions was also performed. Quantitative analysis of responses and statistical analysis Responses were calculated in the following way. In protocol A, the flux within the 30 s preceding each LST stimulation was averaged and set at 100% (baseline flow). After vasoconstriction had reached a steady state during LST stimulation, blood flux over the 30-s period before DR L5 was stimulated was also averaged (control vasoconstriction). Starting with DR stimulation flux was averaged over a period of 2 min (test vasodilatation). Control vasodilatation was determined in the same way. Averaged fluxes were expressed as percentages of baseline flow (see Fig. 4). To compare the absolute magnitude of antidromic vasodilatation during ongoing LST stimulation with control vasodilatation, control vasoconstriction flux and baseline flux, respectively, were subtracted from fluxes during test vasodilatation and control vasodilatation, respectively. The resulting fluxes were again expressed as a percentage of baseline flux (see Fig. 3). Since averaging of blood flow over time is equivalent to integration, responses obtained in this way will be called integrated responses. Additionally, the same calculations were done with the maximum flux values occurring during antidromic vasodilatation. In protocol B, average fluxes were determined separately during the first and second minutes after starting stimulation and expressed as percentages of baseline flow. Control vasoconstriction and control vasodilatation were quantified in the same way (see Figs. 6, 7). Again, similar calculations were done with peak flow responses. In protocol C, average fluxes were calculated during the 60 s after starting LST stimulation and expressed as percentages of baseline flow (see Fig. 9). The latter was higher with than without continuous DR stimulation. A similar quantification was made using the flux value at the end of the 50 s LST stimulation (peak vasoconstriction). Results are expressed as means±SEM. All statistical analyses were carried out using paired t-test.
Results Effects of sectioning DR and LST on blood flow Cutting the DR L5 always resulted in a transient increase in blood flow. This was used to ascertain that the visual identification of the root was correct. Because this was done before setting up the pool with paraffin oil, we did not follow the flow responses for very long. Therefore we do not know whether sectioning of the DR induced long-term changes in cutaneous blood flow. By contrast, sectioning the LST always resulted in a profound increase in cutaneous flow by 112±15% (mean±SEM, n=25) of the initial flow, indicating that ongoing sympathetic vasoconstriction was abolished. Properties of antidromic vasodilatation and sympathetic vasoconstriction The vasodilatation evoked by stimulating the DR L5 with a short train of 10–15 supramaximal pulses at 1 Hz had the typical appearance previously described by other authors (e.g. Magerl et al. 1987; Szolcsányi 1988). Thus, with a latency of 5–10 s blood flow increased steeply, reached its maximum about 30 s after termination of stimulation, and then fell off exponentially (Fig. 2, lefthand panels). The response was long-lasting, but its duration varied amongst animals from 2 min to more than 10 min, while it was fairly constant in an individual animal. We sometimes found it difficult to determine its duration exactly because of small spontaneous shifts in arterial blood pressure. By contrast, neurogenic vasoconstriction commenced promptly. A constant finding was that low stimulation rates of
