Tail arteries from chronically spinalized rats have potentiated responses to nerve stimulation in vitro

Abstract

Patients with severe spinal cord lesions that damage descending autonomic pathways generally have low resting arterial pressure but bladder or colon distension or unheeded injuries may elicit a life-threatening hypertensive episode. Such episodes (known as autonomic dysreflexia) are thought to result from the loss of descending baroreflex inhibition and/or plasticity within the spinal cord. However, it is not clear whether changes in the periphery contribute to the exaggerated reflex vasoconstriction. The effects of spinal transection at T7–8 on nerve- and agonist-evoked contractions of the rat tail artery were investigated in vitro. Isometric contractions of arterial segments were recorded and responses of arteries from spinalized animals (‘spinalized arteries’) and age-matched and sham-operated controls were compared. Two and eight weeks after transection, nerve stimulation at 0.1–10 Hz produced contractions of greater force and duration in spinalized arteries. At both stages, the α-adrenoceptor antagonists prazosin (10 nm) and idazoxan (0.1 μm) produced less blockade of nerve-evoked contraction in spinalized arteries. Two weeks after transection, spinalized arteries were supersensitive to the α₁-adrenoceptor agonist phenylephrine, and the α₂-adrenoceptor agonist, clonidine, but 8 weeks after transection, spinalized arteries were supersensitive only to clonidine. Contractions of spinalized arteries elicited by 60 mm K⁺ were larger and decayed more slowly at both stages. These findings demonstrate that spinal transection markedly increases nerve-evoked contractions and this can, in part, be accounted for by increased reactivity of the vascular smooth muscle to vasoconstrictor agents. This hyper-reactivity may contribute to the genesis of autonomic dysreflexia in patients.

People with severe cervical or high thoracic spinal lesions often develop an exaggerated reflex sympathetic vasoconstriction in response to bladder or colon distension or undetected injuries below the lesion. The increased vascular resistance can raise arterial pressure from low resting levels (mean ∼70 mmHg) to systolic peaks as high as 300 mmHg, provoking cerebral and coronary vascular damage (Lee et al. 1995). This condition, termed autonomic hyper- or dys-reflexia, is generally thought to result from the loss of descending inhibitory pathways and subsequent plasticity of connections within the spinal cord caudal to the lesion (Weaver et al. 2002). The possibility of modified behaviour of the peripheral sympathetic pathways has not been investigated. However, it has been repeatedly suggested that alterations in the sympathetic neurovascular function may contribute to autonomic dysreflexia (see Mathias & Frankel, 1999).

Evidence suggestive of a change in sympathetic neurovascular transmission comes from work on patients (Wallin & Stjernberg, 1984). Wallin & Stjernberg recorded ongoing and reflex-evoked nerve activity in cutaneous vasoconstrictor neurones in peroneal skin nerve fascicles while monitoring blood flow in the skin innervated by these fascicles. In patients with cervical or high thoracic lesions, normal ongoing nerve activity was not detected but bladder compression or electrical stimulation of the skin elicited a brief sympathetic discharge. Associated with this burst of nerve activity was a decrease in skin blood flow. Similar short bursts of activity in normal subjects produced a decrease in skin blood flow lasting about 10 s. In spinal patients, the decrease in blood flow lasted 30 s or more. Although these data are open to various interpretations, it appears that neurally evoked vasoconstriction is prolonged in the skin of patients with spinal lesions.

Rats with spinal transection at T5 develop autonomic dysreflexia (Krassioukov & Weaver, 1995a,b). In such animals, autonomic dysreflexia is associated with plasticity of synaptic connections within the cord below the lesion. Preganglionic neurones responded to the loss of descending pathways (mainly glutamatergic and GABAergic) by initially retracting their dendrites and then extending them over the ensuing weeks (Krassioukov & Weaver, 1996). Following transection, many of the remaining synapses initially lacked an amino acid transmitter but the proportions containing glutamate and GABA substantially recovered by 2 weeks post injury (Llewellyn-Smith & Weaver, 2001). Subsequently, the proliferation of GAP (growth associated protein)-43 immunoreactive axons (Cassam et al. 1999) in the segments caudal to the lesion suggests that novel connections are formed. These changes were accompanied by a marked increase in intraspinal nerve growth factor (NGF), causing calcitonin gene-related peptide (CGRP)-containing primary afferent terminals to sprout within the dorsal horn (Krenz et al. 1999). The attenuation of dysreflexia by intrathecal delivery of antibody to NGF (Krenz et al. 1999) led to the idea that more widespread peptidergic primary afferent inputs contribute to autonomic dysreflexia.

Another explanation for the increased reflex response might be an increased sensitivity of the vasculature to neurally released noradrenaline (NA). Consistent with this suggestion, pressor responses to intravenously infused NA are increased in patients with cervical spinal lesions (Mathias et al. 1976). In addition, the foot veins of tetraplegics have raised sensitivity to locally infused NA (Arnold et al. 1995). In contrast, studies of blood pressure responses to phenylephrine infusion in dysreflexic rats (Maiorov et al. 1997; Landrum et al. 1998) suggest that overall vascular sensitivity to circulating α-adrenoceptor agonists is unchanged. However, the junctional receptors activated by neurally released transmitters may differ from the extrajunctional receptors activated by circulating agonists, either in receptor type or in the postreceptor pathway that produces the effector response (Hirst et al. 1992).

In the present study, we investigated the effects of spinal transection at T7–8 on the mechanical responses of the rat tail artery to nerve stimulation. In rats, the tail artery is innervated by postganglionic neurones in paravertebral ganglia L6–S4 (Sittiracha et al. 1987) with the preganglionic supply arising from spinal segments T13–L2 (Rathner & McAllen, 1998; Smith & Gilbey, 1998). Transection at T7–8 removes all bulbospinal connections to the preganglionic neurones without damaging them directly. The tail artery has very few peptidergic afferent axons in its perivascular plexus (McLachlan, 1996) and lacks vasodilator responses to sensory nerve stimulation (Li & Duckles, 1993). Neurally evoked contractions of the tail artery are largely blocked by α-adrenoceptor antagonists, with both α₁- and α₂-adrenoceptors contributing to the postjunctional response (Bao et al. 1993; Brock et al. 1997; Bradley et al. 2003). A synergistic role for both ATP and neuropeptide Y (NPY) coreleased with NA, particularly during short trains of stimuli, has also been demonstrated (Bradley et al. 2003). Here, the effects of the α₁-adrenoceptor antagonist prazosin, and the α₂-adrenoceptor antagonist idazoxan, on neurally evoked contraction were assessed in arteries from spinalized animals. In addition, the effects of spinal transection on the sensitivity of the tail artery to the α₁-adrenoceptor agonist, phenylephrine, to the α₂-adrenoceptor agonist, clonidine, and to depolarization evoked by 60 mm K⁺, were evaluated.

Methods

All experimental procedures conformed to the National Health and Medical Research Council of Australia guidelines and were approved by the University of New South Wales Animal Care and Ethics Committee.

The spinal cord was transected in female inbred Wistar rats (160–210 g, ∼8 weeks of age), anaesthetized with ketamine (60 mg kg⁻¹) and xylazine (10 mg kg⁻¹), injected i.p. After infiltrating the back muscles with 0.3 ml of the long-acting local anaesthetic Marcaine (5 mg ml⁻¹ bupivacaine hydrochloride; Astra Pharmaceuticals Pty Ltd, Australia), the spinal cord was exposed by a laminectomy over T7–8 segments and the cord was cut with fine scissors. A piece of gelatine foam was placed between the cut surfaces and another over the laminectomy before applying oxytetracycline powder and closing the incision. Two millilitres of warm saline (0.9% NaCl) were injected i.p. and oxytetracycline (100 mg kg⁻¹) and benzylpenicillin (90 mg kg⁻¹) were injected s.c. The animals recovered on a heating blanket and respiration and rectal temperature were monitored frequently until they recovered consciousness. Following surgery, the animals recovered quickly, moving about their cages, eating and drinking within 4 h. The urinary bladder was emptied manually at least four times a day for the first 1–2 weeks until the bladder emptied automatically and then they were monitored at least twice a day. The animals were caged in groups of 2–4. They recovered body weight over the first 1–2 weeks and then put on weight and appeared healthy until they were killed after either 2 (2.2 ± 0.4, n = 6, mean ±s.d.) or 8 (7.9 ± 0.8, n = 7) weeks. In sham-operated controls, the dura was opened and T7–8 segments exposed but not cut. These animals were maintained for 8 (8.0 ± 0.4, n = 6) weeks. Unoperated animals age-matched to the 2-week (n = 7) and 8-week (n = 7) spinalized animals were also used.

To remove the main ventral caudal artery, the animals were exsanguinated under deep anaesthesia (pentobarbitone 100 mg kg⁻¹, i.p.) and the vessel dissected from 10–30 mm distal to the base of the tail. Once isolated, the arteries were maintained in physiological saline of the following composition (mm): Na⁺, 150.6; K⁺, 4.7; Ca²⁺, 2; Mg²⁺, 1.2; Cl⁻, 144.1; H₂PO₄⁻, 1.3; HCO₃⁻, 16.3; glucose, 7.8. This solution was gassed with 95% O₂–5% CO₂ (to pH 7.2) and warmed to 35–36°C.

Mechanical responses

Segments of artery (∼1.5 mm long) were mounted isometrically between stainless steel wires (50 μm diameter) in a 4-chamber myograph (Multi Myograph Model 610M, Danish Myo Technology, Denmark). Each chamber of the myograph contained 6 ml of physiological saline that was exchanged at intervals of 8–15 min throughout the recording period. There was some variation in the lengths (median 1.4 mm, range 1–1.7 mm) and lumen diameters (for the range of diameters see Results) of the arteries studied. To normalize the basal conditions and the isometric contractions, the measured force was converted to the effective pressure exerted on the luminal surface of the artery (Mulvany & Halpern, 1977). It was difficult to decide on the length to set up these vessels because, due to the distinctly different levels of sympathetic activation in vivo, the control arteries would tend to be more constricted than the spinalized ones whereas the distending pressure may have been slightly higher in the controls. However, in conscious rats, the mean arterial pressure returns close to control values (∼90 mmHg) 4 weeks following spinal transection (Rodenbaugh et al. 2003). We chose to set the arteries at the same distending pressure. Initially, the effective distending pressure under basal conditions was set at ∼95 mmHg and left to equilibrate for 30 min. After this, the basal effective distending pressure stabilized at ∼80 mmHg. Preliminary measurements had shown that at this pressure both control and spinalized arteries were at the peak of their length–force relation. Output from the myograph was recorded using a MacLab recording system (ADInstruments, Australia).

Experimental protocols

Two chambers of the myograph were used to record responses elicited by electrical stimulation of the perivascular nerves using a Grass SD9 Stimulator (Grass Instruments, USA). The electrical stimuli (1 ms pulse width) were applied through platinum plate electrodes mounted either side of the tissue. For each tissue, a stimulus–response curve was constructed for contractions evoked by trains of 10 stimuli at 10 Hz, with the stimulation voltage increasing in 5 V increments between 5 and 30 V. All subsequent protocols were performed at supramaximal voltage (25 V). In all cases, the peak amplitude of the contractile response recorded during the trains of stimuli was measured. For the contractions evoked by 10 stimuli at 10 Hz, the rise time was the interval between 10 and 90% of the peak contraction and the half-width was the duration at 50% of the peak amplitude.

First the tissues were stimulated with single trains of 25 stimuli at 0.1, 0.3 and 0.5 Hz, with each train separated by a 5 min interval. Subsequently, the tissues were given repeated cycles of stimulation composed of single trains of 10 stimuli at 1 Hz and 10 Hz, and 100 stimuli at 1 Hz, each separated by a 4 min interval. Each of these cycles lasted about 14 min. Measurements of responses evoked during the second and third cycles of stimulation were used for control data.

Following the third stimulation cycle, the α₁-adrenoceptor antagonist, prazosin (10 nm, Sigma Chemical Company, Australia) was added to one chamber and the α₂-adrenoceptor antagonist, idazoxan (0.1 μm, Sigma) to the other. The drug was left in contact with the tissue for two stimulation cycles and data from the second cycle was used to assess its effects. The concentrations of antagonist used in this study are approximately 10 times higher than those corresponding to the pA₂ values for prazosin at α₁-adrenoceptors and idazoxan at α₂-adrenoceptors (see Brock et al. 1997).

Following the fifth stimulation cycle, the other adrenoceptor antagonist was added so that the tissues were exposed to a combination of prazosin and idazoxan. Following the seventh stimulation cycle, phentolamine (1 μm, Ciba-Geigy, Australia) was added to both chambers and, after the ninth stimulation cycle, tetrodotoxin (0.5 μm, Sigma) was added to both chambers. In the presence of each drug combination, responses to the second stimulation cycle were used to assess their effects.

Segments of artery in the other two chambers of the myograph were used to construct concentration–response curves for the α₁- and α₂-adrenoceptor agonists l-phenylephrine (0.01–300 μm, Sigma) and clonidine (0.001–10 μm, Sigma). Non-cumulative concentration–response curves were determined by increasing the concentration of each agonist by half-log increments, with the tissue exposed to each concentration for 7 min followed by 9 min washes before the next addition of the agonist. The interval between applications of the higher concentrations of clonidine (>0.1 μm) was extended to allow time for the tension to return to the basal level. The peak response to each concentration of the agonists was determined.

Following both the nerve stimulation and the α-adrenoceptor agonist protocols, responses to high [K⁺] were recorded from all four tissues in the presence of prazosin (10 nm) and idazoxan (0.1 μm) to prevent the excitatory actions of NA released from the sympathetic nerves by depolarization. The cotransmitters ATP and NPY do not contribute to K⁺-evoked contractions of the tail artery (Chen & Rembold, 1995). The tissues were exposed to three applications of physiological saline containing 60 mm K⁺ (equimolar substitution of KCl for NaCl) for 3 min, each separated by 5–10 min washes. This concentration of K⁺ was chosen because in control tissues it produces about 60% of the maximum response to increasing [K⁺] (Chen & Rembold, 1995). Responses to the second and third applications of K⁺ were analysed.

The entire testing procedure lasted between 3 h 50 min and 4 h 40 min (median 4 h 14 min).

Statistical analysis

For each tissue the stimulus strength–response curve data was normalized to their maximum response before comparisons were made using repeated-measures analysis of variance. Unless otherwise indicated, Mann-Whitney U tests were used for pairwise comparisons and Kruskal-Wallis tests for multiple comparisons, because the groups of raw data had unequal variance. Within-group comparisons between the sequential effects of the antagonists were made with Friedman tests followed by Wilcoxon signed rank tests. P < 0.05 was taken as a significant difference.

The normalized data for the stimulus–response curves are presented as mean ±s.e.m. All other data are presented as median and interquartile range (IQR).

Results

General observations

For all the measured parameters, there were no significant differences between the control groups, i.e. arteries isolated from 8-week sham-operated animals and arteries from unoperated animals age-matched to both 2-week and 8-week spinalized animals. The effects of spinal transection were therefore assessed by comparing arteries from 2-week spinalized animals with those from unoperated age-matched controls and arteries from 8-week spinalized animals with those from 8-week age-matched sham-operated animals. For most measured parameters, there were also no differences between the 2-week and 8-week spinalized arteries but, where there were differences, these are described.

Basal effective distending pressures (about 10 × 10³ N m⁻²) varied <10% between the groups. Under these conditions, the lumen diameters (calculated from the measured internal circumference) of the 2-week spinalized arteries (753 μm, IQR 728–776 μm) were similar to those of the age-matched arteries (796 μm, IQR 746–833 μm, P = 0.39) but those of the 8-week spinalized arteries (749 μm, IQR 746–796 μm) were smaller than those of the 8-week sham-operated arteries (877 μm, IQR 858–931 μm; P < 0.05).

Stimulus–response curves

The effects of increasing stimulus voltage from 5 to 30 V were investigated for contractions evoked by 10 stimuli at 10 Hz. As the maximal force of electrically evoked contraction of the spinalized arteries was larger than that of their respective control arteries (see below), the force of their response to each stimulation voltage was normalized to that of their maximum response before comparisons were made. None of the arteries was activated at a stimulus strength of 5 V. At 10 V, the control arteries produced 35–50% of the maximum force whereas the 2-week and 8-week spinalized arteries produced about 70% of the maximum force (Fig. 1A and B; repeated measures ANOVA, groups–voltage interaction P < 0.001 for both comparisons). This difference was most pronounced 2 weeks post surgery. For all arteries, 25 V was supramaximal for evoking contraction and was used throughout.

Figure 1. Spinal transection increases the sensitivity of the tail artery to electrical stimulation.

A and B, stimulus–response curves for control and spinalized arteries at 2 weeks (A) and 8 weeks (B) after spinal transection. The force of each vessel's response to each stimulation voltage was normalized to that of their maximum response. For each stimulation voltage, significant differences (unpaired t tests) between control and spinalized arteries are indicated (*P < 0.05, **P < 0.01). Arteries from spinalized animals were more readily activated at lower stimulation voltages than control arteries.

Responses to nerve stimulation

Figure 2 shows the contractile response of an 8-week sham-operated artery and an 8-week spinalized artery to trains of 25 stimuli at 0.1, 0.3 and 0.5 Hz, 10 stimuli at 1 and 10 Hz, and 100 stimuli at 1 Hz. In comparison with their respective control arteries, both 2-week and 8-week spinalized arteries generated much larger responses to stimulation with 25 stimuli at 0.1, 0.3 and 0.5 Hz and 100 stimuli at 1 Hz (Fig. 3). At both time points, the enhancement was most marked at 0.1 Hz (>25-fold) and decreased as the frequency of stimulation was increased (10-fold at 0.3 Hz, 5-fold at 0.5 Hz and 1.5-fold at 1 Hz).

Figure 2. Spinal transection enhances contractions to trains of nerve stimuli.

A, contractions of 8-week sham-operated (upper trace) and 8-week spinalized arteries (lower trace) to stimulation of the perivascular nerves with 25 pulses at 0.1, 0.3 and 0.5 Hz. B, contractions in the same tissues evoked by 10 pulses at 1 and 10 Hz and 100 pulses at 1 Hz. In comparison with the sham-operated artery, the spinalized artery generated greater force during the trains of stimuli. In addition, the contractions of the spinalized artery were prolonged in time course (see B). The force scale bar in A applies also in B.

Figure 3. Responses to low frequency nerve stimulation are increased at both 2 and 8 weeks after spinal transection.

A and B, the peak increase in effective pressure produced by 25 pulses at 0.1, 0.3 and 0.5 Hz and 100 pulses at 1 Hz in 2-week age-matched control (open columns, n = 7) and spinalized (hatched columns, n = 6) arteries (A) and 8-week sham-operated (open columns, n = 4) and spinalized (hatched columns, n = 7) arteries (B). Statistical differences between the response of control and spinalized arteries to each frequency of stimulation are indicated (*P < 0.05, **P < 0.01). In comparison with arteries from the control animals, those from spinalized animals at both time points generated a much greater force to low frequency trains of stimuli.

Time course of contractions to short trains of stimuli

The mechanical responses of spinalized arteries to electrical stimulation were also prolonged in time course (Fig. 2B). To exemplify this change, contractions to 10 pulses at 10 Hz were compared. The contractions of the spinalized arteries produced by trains of 10 stimuli at 10 Hz were significantly increased in force and prolonged in time course compared to those of their respective control arteries (Table 1). Both force and half-width were about twice as large in the spinalized arteries. As the rise time of the contractions was increased by about a third in the spinalized arteries (Table 1), the change in time course was due primarily to an increase in the relaxation time.

Table 1.

Amplitude, rise time and half-width of contractions evoked by 10 stimuli at 10 Hz

	Maximum increase in effective pressure (103 N m−2)	Rise time (s)	Half-width (s)
Control
″Age matched (2 weeks)	8.2	1.5	4.7
″n = 7	(6.3–8.7)	(1.4–1.6)	(4.4–5.3)
″8-week sham operated	6.5	1.4	4.9
″n = 6	(6.1–6.9)	(1.4–1.5)	(4.4–5.2)
Spinalized
″2 weeks	18.5**	2.2**	17.0**
″n = 6	(16.2–19.2)	(2.1–2.2)	(16.8–18.0)
″8 weeks	20.0**	2.0**	13.4**
″n = 7	(19.4–20.2)	(1.9–2.1)	(10.2–17.0)

The rise time was the time interval between 10 and 90% of the peak contraction and the half-width was the time interval (between the rising and falling phases) at 50% of the peak contraction.

^**Significant differences between 2-week unoperated and 2-week spinalized arteries and between 8-week sham operated and 8-week spinalized arteries (P < 0.01). The numbers in parentheses indicate IQR.

Effects of blocking α-adrenoceptors on electrically evoked contractions

The effects of prazosin (10 nm) and idazoxan (0.1 μm) on contractions evoked by 100 stimuli at 1 Hz were assessed. In control arteries, prazosin and idazoxan reduced the maximum force of this contraction by about 95% and 70%, respectively (Fig. 4). The blockade produced by the combined application of prazosin and idazoxan (median 97%) was slightly greater than that of prazosin alone (median 96%, Wilcoxon signed rank test, P = 0.01). The subsequent addition of phentolamine (1 μm) had no additional effect (Wilcoxon signed rank test, P = 0.25). The non-additive effects of prazosin and idazoxan support the idea that α₂-adrenoceptor activation has a synergistic action on α₁-adrenoceptor-mediated contraction in the tail artery (Xiao & Rand, 1989; Brock et al. 1997).

Figure 4. Reduced blockade of nerve-evoked transmission in spinalized arteries by α-adrenoceptor antagonists.

A and B, percentage blockade of contractions evoked by 100 pulses at 1 Hz produced by the α-adrenoceptor antagonists prazosin (10 nm), idazoxan (0.1 μm) and phentolamine (1 μm) in 2-week age-matched control (open columns, n = 7) and spinalized (hatched columns, n = 6) arteries (A) and 8-week sham-operated (open columns, n = 6) and spinalized (hatched columns, n = 7) arteries (B). Prazosin and idazoxan were applied alone and together. The column labelled phentolamine indicates the percentage blockade after adding this agent in the presence of prazosin and idazoxan. Statistical differences between control and spinalized arteries are indicated (**P < 0.01). In comparison with the control arteries, the extent of blockade of electrically evoked contractions of the spinalized arteries produced by all α-adrenoceptor antagonists was reduced.

The effects of prazosin and idazoxan applied alone and in combination were reduced in spinalized arteries (Fig. 4) and the subsequent addition of phentolamine reduced the force of the contractions by a further 5–10% of the predrug treatment values (Wilcoxon signed rank test, P < 0.05 for both 2 and 8-week spinalized arteries). The combined application of all three antagonists reduced the force of contraction by 85–90% in the spinalized arteries, compared to 95% in controls (Fig. 4).

In all tissues, the component of contraction that was resistant to α-adrenoceptor antagonists was fully blocked by tetrodotoxin (0.5 μm), indicating that it was due to the activation of non-adrenoceptors by neurally released substances. This finding also confirms that the electrically evoked contractions were due entirely to activation of the perivascular nerves.

Responses to phenylephrine and clonidine

The 2-week spinalized arteries had significantly lower EC₅₀ values for phenylephrine than the unoperated age-matched arteries (Table 2) but the EC₅₀ values for this agent in 8-week spinalized arteries was similar to that in 8-week sham-operated arteries (Table 2, P = 0.25). In addition, the EC₅₀ values for phenylephrine in the 2-week spinalized arteries were lower than in the 8-week spinalized arteries (Table 2). However, the maximum increase in effective pressure produced by phenylephrine did not differ between the groups of spinalized arteries or between these and their respective controls.

Table 2.

Effects of spinal transection on sensitivity to phenylephrine and clonidine

	Phenylephrine		Clonidine
Group	EC50 (μm)	Maximum increase in effective pressure (103 N m−2)	EC50 (nm)	Maximum increase in effective pressure (103 N m−2)
Control
″Age matched (2 weeks)	1.69	33.9	67	22.3
″n = 7	(1.37–2.54)	(32.9–36.1)	(59–120)	(21.4–22.6)
″8-week sham operated	2.27	31.5	59	23.6
″n = 6	(2.02–2.50)	(31.0–32.3)	(51–100)	(21.0–24.5)
Spinalized
″2 weeks	0.47*a	34.7	24**	26.4*b
″n = 6	(0.36–0.79)	(33.0–36.7)	(23–25)	(24.9–28.3)
″8 weeks	1.61a	29.7	23**	19.9b
″n = 7	(1.22–2.58)	(27.1–36.7)	(19–27)	(18.9–20.8)

^*Significant differences between 2-week unoperated and 2-week spinalized arteries and between 8-week sham operated and 8-week spinalized arteries: P < 0.05

^**Significant differences between 2-week unoperated and 2-week spinalized arteries and between 8-week sham operated and 8-week spinalized arteries: P < 0.01.

^aSignificant differences between groups of spinalized arteries indicated by (P < 0.05 in both cases). The numbers in parentheses indicate IQR.

^bSignificant differences between groups of spinalized arteries indicated by (P < 0.05 in both cases). The numbers in parentheses indicate IQR.

The EC₅₀ values for clonidine were significantly lower in both groups of spinalized arteries compared to their controls (Table 2). For spinalized arteries, the maximum increase in pressure produced by clonidine was greater than control at 2 weeks but not after 8 weeks (Table 2). The maximum increase in effective pressure produced by clonidine in the 2-week spinalized arteries was also significantly greater than in the 8-week spinalized arteries.

Responses to 60 mm K⁺ solution

The magnitude of the contractile response to 60 mm K⁺ of the 2-week and 8-week spinalized arteries was significantly increased compared to that of their respective control arteries (Fig. 5). In addition, on returning to the normal bathing solution, the 50% decay of the K⁺-evoked contraction was slower in the spinalized arteries (2 weeks: age matched, 8.3 s, IQR 6.7–8.6 s; spinalized, 24.5 s, IQR 18.2–25.5 s, P < 0.01; 8 weeks: sham operated, 8.1 s, IQR 7.2–10.4 s; spinalized, 27.3 s, IQR 20.5–30.1 s, P < 0.01).

Figure 5. Enhanced responses to 60 mm K⁺ suggest a non-specific increase in reactivity in spinalized arteries.

A, contractions of 8-week sham-operated and 8-week spinalized arteries to 60 mm K⁺ in the presence of α-adrenoceptor antagonists (10 nm prazosin + 0.1 μm idazoxan). B and C, comparison of responses to 60 mm K⁺ of control and spinalized arteries after 2 and 8 weeks. Statistical differences between 2-week age-matched (n = 7) and 2-week spinalized (n = 6) arteries and between 8-week sham-operated (n = 6) and 8-week spinalized (n = 7) arteries are indicated (**P < 0.01). In comparison with the arteries from the control animals, those from spinalized animals generated a much greater force to 60 mm K⁺. In addition, on returning to the normal bathing solution, the decay of the K⁺-evoked contraction was much slower in the spinalized arteries (see A).

Correlations

The degree of association between the response of the arteries to electrical activation (100 pulses at 1 Hz) and their sensitivity to phenylephrine, clonidine and 60 mm K⁺ was assessed using Spearman's coefficient of rank correlation. While there was no correlation between the EC₅₀ for phenylephrine and the response to electrical activation (P = 0.99), a significant negative correlation was found between the EC₅₀ value for clonidine and the response to electrical activation (ρ = −0.663, P < 0.001). In addition, a significant positive correlation was found between the responses to 60 mm K⁺ and to electrical activation (ρ = 0.853, P < 0.0001).

Discussion

This study has revealed markedly enhanced responses of the rat tail artery of spinalized animals to trains of stimuli that activated the perivascular sympathetic nerves. The contractions of these vessels were larger and more prolonged than in age-matched and sham-operated controls. While there were also increases in the sensitivity to α-adrenoceptor agonists, these did not always correlate with the enhanced nerve-evoked contractions. The much greater contractions of spinalized arteries in response to 60 mm K⁺ suggest that changes in the reactivity of the vascular smooth muscle contribute to the enhanced responses. The reduced blockade by the α-adrenoceptor antagonists in spinalized arteries may indicate that NA release is also increased.

The increased reactivity of spinalized arteries to nerve stimulation was observed for all frequencies of stimulation studied (0.1–10 Hz) but was most marked at the lower frequencies of stimulation (see Fig. 2). The low frequency trains of stimuli produce contractions similar to those elicited by the asynchronous release of neurotransmitter from perivascular terminals (Brock et al. 1997). These contractions are likely to mimic the effects of ongoing nerve activity in vivo, although the frequencies of discharge during reflex bursts may be higher. However, as the level of sympathetic nerve activity after spinal injury appears to be very low (Wallin & Stjernberg, 1984; Stjernberg et al. 1986), increased reactivity like that demonstrated here would clearly exaggerate vasoconstriction in response to a burst at any frequency.

The spinalized arteries had developed a raised sensitivity to phenylephrine after 2 weeks that had disappeared by 8 weeks although both groups of spinalized arteries were equally hyper-reactive to nerve stimuli. Therefore supersensitivity to α₁-adrenoceptor activation is unlikely to be the cause of the increased response to nerve stimulation.

In contrast, supersensitivity to clonidine was present both 2 and 8 weeks after spinal transection and the response of the arteries to electrical stimulation was correlated with their EC₅₀ for clonidine. While this suggests that α₂-adrenoceptor activation is involved in the enhanced nerve-evoked responses of spinalized arteries, α₂-blockade with idazoxan was markedly less effective in reducing nerve-evoked contraction than in controls (Fig. 4). An increased sensitivity of the α₂-adrenoceptors to neurally released NA might explain the reduced amount of blockade produced by idazoxan. In fact, both prazosin and idazoxan produced less blockade of nerve-evoked contraction in spinalized arteries. As the blockade by prazosin was reduced but the sensitivity to phenylephrine was unchanged in the 8-week spinalized arteries, it is unlikely that alteration in α₁-adrenoceptors is the cause of the reduced effectiveness of prazosin. More probably, increased release of NA from the perivascular sympathetic nerves explains the reduced blockade by both prazosin and idazoxan. In spinalized tail arteries, using in situ amperometry, we have observed that idazoxan produces a greater augmentation of NA release than in control tissues (authors' unpublished observations). Therefore, the more marked reduction in the amount of blockade by idazoxan (compared with prazosin) in spinalized arteries might best be explained by an increased α₂-adrenoceptor-mediated autoinhibition of NA release (Starke, 1987) in tissues from spinalized animals.

As the size of the remaining component of neurally evoked contraction in the presence of a high concentration of phentolamine was increased from 5% to more than 10% of the control response (Fig. 4), it is also possible that cotransmitters play a greater role in neural activation in spinalized arteries. In support of this suggestion, intracellularly recorded purinergic excitatory junction potentials are increased in amplitude in 7-week spinalized rat tail arteries (Brock & McLachlan, 2003). While this increase in purinergic transmission may play an important role in potentiating the response to coreleased NA (Bradley et al. 2003), the substantial blockade produced by the α-adrenoceptor antagonists suggests that noradrenaline is primarily responsible for activation of the spinalized arteries. The contribution of neurally released ATP to the contractile responses remains to be tested.

Both 2-week and 8-week spinalized tail arteries showed larger and more prolonged responses to 60 mm K⁺ that were of similar magnitude. As the postjunctional α-adrenoceptors were blocked with prazosin and idazoxan, contractions to 60 mm K⁺ are likely to result from direct depolarization of the vascular smooth muscle. Like the EC₅₀ for clonidine, the response of the control and spinalized arteries to 60 mm K⁺ was correlated with their response to electrical stimulation. These findings suggest that hyper-reactivity of the arteries to neural activation can be attributed, at least in part, to a postjunctional change. Contraction of the tail artery to both high K⁺ and α₂-adrenoceptor agonists is dependent on the presence of extracellular Ca²⁺ (Abe et al. 1987; Chen & Rembold, 1995). Therefore, a possible explanation for the increased sensitivity to both clonidine and high K⁺ is that the contractile mechanism is selectively sensitized to Ca²⁺ entering from outside the cell. Alternatively, as both agents depolarize the muscle, depolarization-induced Ca²⁺ entry may be increased.

The major factor underlying the changed responsiveness of spinalized arteries is probably the decrease in ongoing sympathetic nerve activity following spinal transection. Silencing ongoing nerve activity by surgically removing the preganglionic input to postganglionic sympathetic neurones (i.e. decentralization) enhanced neurovascular transmission in rabbit ear artery (Tsuru & Uematsu, 1986). In this tissue, decentralization produced a marked increase in the contractile response to low frequency nerve stimulation that developed relatively slowly following transection of the preganglionic supply, reaching a plateau level at 4 weeks (Tsuru & Uematsu, 1986). However, supersensitivity to NA and K⁺ was fully developed after 1 week. As overflow of tritium from 8-week decentralized ear arteries labelled with [³H]NA was markedly increased in response to low frequency nerve stimulation (Tsuru et al. 1993), it was proposed that the enhanced response to nerve stimulation was due primarily to an increase in NA release. Silencing ongoing activity at other central and peripheral synapses with tetrodotoxin also increases synaptic efficacy, with both pre- and postsynaptic changes (Gallego & Geijo, 1987; Murthy et al. 2001).

Studies in spinalized patients have demonstrated enhanced pressor responses to various vasoconstrictor agents consistent with a generalized postjunctional hyper-reactivity (Mathias & Frankel, 1999). It has also been reported that bladder percussion in patients with cervical or high thoracic spinal lesions produces a marked increase in NA spillover in the legs (Karlsson et al. 1998). This observation is difficult to reconcile with the moderate activation of muscle and skin vasoconstrictor neurones supplying the legs produced by bladder stimulation (Wallin & Stjernberg, 1984; Stjernberg et al. 1986). However, it would be consistent with increased release of NA per impulse because the normal rate of plasma NA clearance (Karlsson et al. 1998) indicates that uptake of released NA is unlikely to be decreased.

In conclusion, spinal transection at T7–8 produces a marked increase in the response of the tail artery to perivascular nerve stimulation. Such a response may contribute to the genesis of autonomic dysreflexia. The present study has not established the cause of this change but it seems likely that both pre- and postjunctional changes are involved, following the decrease in ongoing sympathetic activity produced by spinal transection. Interestingly, dynamic exercise has been reported to decrease the reactivity of vascular smooth muscle to both sympathetic nerve activity and α-adrenoceptor agonists (Howard & DiCarlo, 1992; Halliwill et al. 1996). In rats with spinal transection at T5, a single bout of mild-to-moderate exercise reduced the pressor response to colon distension by about 50% (Collins & Dicarlo, 2002), suggesting that modulation of vascular reactivity by nerve activity is a dynamic process.