Insights into the functional roles of α₁-adrenoceptor subtypes in mouse carotid arteries using knockout mice

Abstract

1. α₁-Adrenoceptor (AR) subtypes in mouse carotid arteries were characterised using a combination of agonist/antagonist pharmacology and knockout (KO) mice.

2. Phenylephrine (PE) was most potent in the α_1B-KO (pEC₅₀=6.9±0.2) followed by control (pEC₅₀=6.3±0.06) and α_1D-KO (pEC₅₀=5.5±0.07). Both N-[5-(4,5-dihydro-1H-imidazol-2yl)-2-hydroxy-5,6,7,8-tetrahydronaphthalen-1-yl] methanesulphonamide hydrobromide (A-61603) and 5-hydroxytryptamine (5-HT) were more potent in the α_1D-KO (pEC₅₀=7.4±0.27 and 7.4±0.05, respectively) than the control (pEC₅₀=6.9±0.09 and 6.9±0.08, respectively) and equipotent with the control in the α_1B-KO (pEC₅₀=6.7±0.07 and 6.8±0.04). Maximum responses to PE and A-61603 were reduced in the α_1D-KO compared to control; there was no difference in maximum responses to 5-HT.

3. In control arteries, prazosin and 5-methylurapidil acted competitively with pA₂ of 9.6 and 7.5, respectively. BMY7378 produced antagonism only at the highest concentration used (100 nM; pK_B 8.3).

4. Prazosin, 5-methylurapidil and BMY7378 acted competitively in α_1B-KO carotid arteries with pA₂ of 10.3, 7.6 and 9.6, respectively.

5. In the α_1D-KO, against PE, 5-methylurapidil produced a pA₂ of 8.1. pK_B values were calculated for prazosin (10.6) and BMY7378 (7.0). Against A-61603, 5-methylurapidil had a pA₂ of 8.5, prazosin 8.6, while BMY7378 had no effect.

6. In conclusion, the α_1B-KO mediates contraction solely through α_1D-ARs and the α_1D-KO through α_1A-ARs. Extrapolating back to the control from the knockout data suggests that all three subtypes could be involved in the responses, but we propose that the α_1D-AR causes the contractile response and that the role of the α_1B-AR is mainly regulatory.

Introduction

Three native α₁-adrenoceptor (AR) subtypes, defined by ligand binding and functional pharmacology, α_1A, α_1B and α_1D, correspond to three cloned subtypes, α_1a, α_1b and α_1d (Bylund et al., 1994). It is not known whether the three subtypes have different biological roles. Several tissues, including arteries, express more than one subtype. The mRNA and protein for all three α₁-AR subtypes are expressed in the major blood vessels of the rat (Piascik et al., 1995; Scofield et al., 1995; Piascik et al., 1997; Hrometz et al., 1999). However, separating the responses mediated by these subtypes has proved difficult, due to the limitations of selectivity of antagonists between the three receptors and the proposition that they might all be involved in the same type of response, namely contraction of vascular smooth muscle.

A handful of α_1A-AR-selective antagonists are available, such as 5-methylurapidil, WB4101 and RS100329 (Gross et al., 1988; Schwinn et al., 1995; Williams et al., 1999), while BMY7378 is the only widely accepted α_1D-AR-selective antagonist (Saussy et al., 1994; Goetz et al., 1995; Kenny et al., 1995). A major pharmacological complication when attempting to subtype α₁-ARs is the lack of a selective competitive antagonist for the α_1B-AR. This appears to be a situation in which receptor knockouts might simplify the pharmacological analysis.

Most studies of vascular α₁-ARs, either as an undivided class or as subtypes, have been carried out in rats, rabbits and dogs and until recently little data has been available for the mouse. However, α₁-AR knockout (KO) mice are now available for the α_1B- and α_1D-ARs (Cavalli et al., 1997; Tanoue et al., 2002). These provide novel environments to study and subtype the remaining two possible α₁-ARs.

We have chosen to study the carotid artery since this vessel has greater potential as an experimental model, being accessible to surgical manipulation in vivo and amenable to perfusion studies in vitro. There is also controversy over whether contraction is mediated by the α_1B- or α_1D-ARs according to species (dog: α_1B (Muramatsu et al., 1991; Kohno et al., 1994); rabbit: α_1B (Muramatsu et al., 1995); rat: α_1D (Villalobos-Molina & Ibarra, 1996; de Oliveira et al., 1998)). Theoretically, this presents a relatively straightforward scenario for observing the consequences of knocking out each of these subtypes. Previous work has shown that (1) knockout of the α_1B-AR produces little change in the size or sensitivity of responses to phenylephrine (PE) in the aorta and carotid arteries; the antagonist pharmacology is more consistent with α_1D-AR pharmacology, suggesting a major role for the α_1D-AR and a minor one for the α_1B-AR (Daly et al., 2002), and (2) knockout of the α_1D-AR produced a significant reduction in sensitivity and maximum response to PE in the aorta, consistent with the loss of the major contractile α₁-AR (Tanoue et al., 2002).

The objective of the present study was to apply a consistent antagonist analysis using the two knockouts to allow us to explore the functional relationship between α₁-AR subtypes; for example, what are the consequences of deleting each receptor? Does this show how they interact? Do other subtypes upregulate to compensate? We used the ‘definitive' antagonists (prazosin, 5-methylurapidil and BMY7378) and the α₁-AR-selective agonist PE (eliminates possible complications from α₂- and β-ARs). We also used the α_1A-AR-selective agonist N-[5-(4,5-dihydro-1H-imidazol-2yl)-2-hydroxy-5,6,7,8-tetrahydronaphthalen-1-yl] methanesulphonamide hydrobromide more commonly known as A-61603 (Knepper et al., 1995) to reinforce the antagonist analysis.

Latterly, we applied the knowledge obtained from mouse carotid arteries to data collected from our laboratory a number of years ago on the rat carotid artery, which at the time were difficult to interpret.

Methods

Animals used and set-up procedure

All transgenic mice (C57 Black genetic background; for a full description of genetic background see Cavalli et al. (1997) and Tanoue et al. (2002)) were bred at the University of Glasgow. Mice were killed by lethal overdose of carbon dioxide. The common carotid arteries were removed, placed in cold oxygenated Krebs and dissected free of connective tissue with the aid of a dissecting microscope.

Experiments were carried out in a four-chamber wire myograph (J.P. Trading, Aarhus, Denmark). Arteries were cut into approximately 2 mm lengths and mounted on two 40 μm wires. One wire was attached to a fixed head, while the other was attached to a head connected to a force transducer. The force transducer was in turn connected to a Linseis pen recorder to allow recordings of the force achieved.

Vessels were allowed to equilibrate in Krebs (37°C and gassed with 95% O₂, 5% CO₂) for 15 min after which time the vessels were set under their optimal resting length tensions: previously calculated to be 250 mg for the control, α_1B- and α_1D-KO mouse carotid arteries (Deighan, 2001). The vessels were left to equilibrate at this tension for 30–45 min with washes every 15 min. Prior to the start of each experiment, vessels were challenged with a sensitising concentration of 0.3 μM PE (control and α_1B-KO mouse), 10 μM PE (α_1D-KO), 10 μM A-61603 (α_1D-KO) or 1 μM 5-hydroxytryptamine (5-HT) (all mice strains). The contraction was allowed to plateau and then washed with Krebs. This was repeated three times to minimise changes in sensitivity to further challenges with agonists. Cumulative concentration–response curves (CRC) were carried out to either PE (1 nM–1 mM), A-61603 (1 nM–300 μM) or 5-HT (1 nM–30 μM). Subsequent CRCs to PE or A-61603 were carried out in the presence of antagonists (prazosin, 5-methylurapidil and BMY7378), which were equilibrated with the tissue for 30 min prior to beginning the CRC. Time controls were carried out in parallel with antagonist curves.

A similar analysis was carried out on carotid artery rings from male Wistar rats (320–400 g) suspended between two wire hooks and recorded isometrically. The protocol was identical to that in mice except that the vessels were equilibrated under 2.5 g of tension and noradrenaline (NA; 1 nM–10 μM) was used as the agonist.

Data analysis

Responses to agonists are expressed as tension in grams or as a percentage of the maximum response of the first CRC. The pEC₅₀ was calculated as the negative logarithm of the concentration of agonist that produces half the maximal response. pEC₅₀ values for PE, A-61603 and 5-HT in control, α_1B- and α_1D-KO mice were analysed using a one-way analysis of variance (ANOVA) followed by a Bonferroni post test.

The pEC₅₀ values, Hill slopes and maximum responses calculated from the antagonist data in mouse carotid arteries were analysed using a two-way ANOVA followed by a Bonferroni post test. For both one- and two-way ANOVA, a P-value of less than 0.05 was considered significant. The agonist concentration ratios (CRs) were determined from the ratio of the EC₅₀ of the agonist in the presence and absence of the antagonist and used for Schild analysis where the log[antagonist] is plotted against log(CR−1) (Arunlakshana & Schild, 1959). Linear regression produces an x-intercept that is equal to the pA₂ of the antagonist. If the slope of the Schild plot is equal to 1, then pA₂=pK_B and is indicative of competitive binding. Where a pA₂ value could not be calculated (e.g. where there is only a small shift with antagonist), a pK_B value was calculated instead using the equation

where pK_B is the negative logarithm of the dissociation constant K_B and [B] is the concentration of antagonist.

Solutions and drugs

The Krebs–Henseleit solution was of the following composition (mM): NaCl (119), KCl (4.7), MgCl₂ (1.2), CaCl₂ (2.5), NaHCO₃ (25), NaHPO₄ (1.2), glucose (11.5) and Na₂EDTA (0.023).

The following compounds were used: A-61603 hydrobromide (Tocris, U.K.), BMY 7378 dihydrochloride (8-[2-[4-(2-methoxyphenyl)-1-piperazinyl]ethyl]-8-azaspiro[4,5]decane-7,9-dione; Research Biochemicals International, U.K.), 5-HT (Sigma, U.K.), 5-methylurapidil (Research Biochemicals International, U.K.), noradrenaline hydrochloride (Sigma, U.K.), phenylephrine hydrochloride (Sigma, U.K.) and prazosin hydrochloride (Sigma, U.K.).

All drugs were dissolved in deionised water and then diluted (1 : 10) to give the concentrations used for the CRCs.

Results

The genetic controls for the α_1B- and α_1D-KO were found to be pharmacologically similar in their sensitivities and maximum responses to PE and 5-HT (Figure 1). Therefore, only one set of control experiments was required.

Figure 1.

Mean concentration response data to (a) PE and (b) 5-HT in α_1B-KO and α_1D-KO control mouse carotid arteries expressed as tension in grams. Both agonists produced similar responses in the two strains of control mice. Mean curves were generated using nonlinear regression upon which the mean±s.e.m. data have been superimposed (n>9).

Control, α_1B-KO and α_1D-KO mouse carotid arteries

All agonist data (pEC₅₀ values, maximum responses, Hill slopes and statistical comparisons) are presented in Table 1.

Table 1.

Comparison of pEC₅₀ values, maximum responses and Hill slopes for agonists producing contractions in (a) control, (b) α_1B-KO and (c) α_1D-KO mouse carotid arteries

Agonist	pEC50	Max. response (g)	Hill slope (95% CI)
(a) Control
PE	6.3±0.06	0.37±0.01	0.5 (0.4–0.6)#
(R)-A-61603	6.9±0.09	0.27±0.01	0.65 (0.5–0.8)#
5-HT	6.9±0.08	0.27±0.04	1.15 (0.6–1.7)
(b) α1B-KO
PE	6.9±0.22*	0.33±0.01	0.4 (0.2–0.6)#
(R)-A-61603	6.7±0.07	0.28±0.01	0.65 (0.5–0.8)#
5-HT	6.8±0.04	0.27±0.05	1.3 (0.9–1.7)
(c) α1D-KO
PE	5.5±0.07*	0.21±0.02*	0.85 (0.6–1.1)
(R)-A-61603	7.4±0.27*	0.11±0.01*	0.4 (0.1–0.7)#
5-HT	7.4±0.05*	0.26±0.06	1.0 (0.7–1.3)

pEC₅₀ values and maximum responses are expressed as mean±s.e.m. and the Hill slopes are given along with their 95% confidence intervals (95% CI).

^*P<0.05 compared to control;

^#Hill slope significantly different from unity.

PE produced concentration-dependent contractions in carotid arteries from all three strains of mice. Sensitivity to PE was found to vary between mouse strains (Figure 2a). The α_1B-KO was the most sensitive to PE, followed by the control carotid artery and finally the least sensitive was the α_1D-KO. All pEC₅₀ values were significantly different between the three mouse strains. Control and α_1B-KO carotid arteries produced similar maximum responses, whereas the maximum response from the α_1D-KO carotid artery was significantly smaller. CRCs to PE demonstrated shallow Hill slopes significantly different from unity in both control and α_1B-KO carotid arteries. This was not the case in α_1D-KO carotid arteries; the PE CRC had a Hill slope that was not significantly different from unity.

Figure 2.

Mean concentration response data to (a) PE, (b) A-61603 and (c) 5-HT in carotid arteries from control, α_1B-KO and α_1D-KO mice expressed as a percentage of their own maximum response. Mean curves were generated using nonlinear regression upon which the mean±s.e.m. data have been superimposed (n>9).

A-61603 produced concentration-dependent contractions in all three mouse strains (Figure 2b). The α_1D-KO was more sensitive to A-61603 than the control or the α_1B-KO carotid arteries. However, the efficacy of A-61603 in the α_1D-KO was reduced compared to the other two mouse strains. The maximum responses produced by the control and α_1B-KO carotid arteries were similar, while the α_1D-KO response was smaller. All three strains of mice produced shallow Hill slopes significantly different from unity.

5-HT produced concentration-dependent contractions in all three mouse strains (Figure 2c). Desensitisation occurred at the higher concentrations of 5-HT; therefore, CRCs were stopped as soon as the maximum response began to decline. All three mouse strains produced similar responses to 5-HT, with no differences observed in maximum responses or Hill slopes. However, α_1D-KO carotid arteries were found to be more sensitive to 5-HT than either control or α_1B-KO arteries.

In control, α_1B-KO and α_1D-KO carotid arteries, the subtype-selective antagonists produced a rightward displacement of the PE curve without a depression in the maximum response (Figure 3). Prazosin was found to cause a decrease in the maximum response at 1 and 10 nM in α_1D-KO arteries and at 100 nM in control and α_1B-KO arteries (Figure 3). The pA₂ values for prazosin and 5-methylurapidil in control tissue were calculated to be 9.6 and 7.5, respectively, with slopes that were not significantly different from unity, indicating competitive antagonism (Table 2). Only the highest concentration of BMY7378 (100 nM) produced a significant shift in the CRC to PE. Therefore, a pA₂ value could not be calculated. Instead, a pK_B value was calculated at 100 nM BMY7378 and was found to be 8.3. In the α_1B-KO carotid artery, all antagonists acted competitively. The pA₂ values for prazosin, 5-methylurapidil and BMY7378 were found to be 10.3, 7.6 and 9.6, respectively (Table 2). In the α_1D-KO, it was possible to calculate a pA₂ value only for 5-methylurapidil, which was 8.1. Prazosin antagonised the contractions to PE with such potency that we had to dilute the concentrations used in control and α_1B-KO by a factor of 10 to allow us to obtain four consecutive CRCs. Despite this, it was still possible to calculate EC₅₀ values only for the smallest concentration used (0.1 nM). This was the concentration used to calculate a pK_B for prazosin, which was 10.6. In contrast BMY7378 could only weakly antagonise the contraction to PE in α_1D-KO carotid arteries. A pK_B value could be calculated only at 100 nM, which was 7.0 (Table 2).

Figure 3.

Mean concentration response data to PE in the presence of increasing concentrations of antagonists in (a) control, (b) α_1B-KO and (c) α_1D-KO mouse carotid arteries. Mean curves were generated using nonlinear regression upon which the mean±s.e.m. data have been superimposed (n>6).

Table 2.

pA₂ or pK_B values and slope parameters of antagonists in control, α_1B-KO or α_1D-KO mouse carotid arteries

	Control		α1B-KO		α1D-KO
Antagonist	pA2/pKB	Slope	pA2/pKB	Slope	pA2/pKB	Slope
Prazosin	9.6	0.93 (0.77–1.08)	10.3	0.92 (0.68–1.2)	10.6	NA
5-MeU	7.5	1.1 (0.73–1.5)	7.6	1.1 (0.77–1.5)	8.1	0.82 (0.4–1.3)
BMY7378	8.3	NA	9.6	0.9	7.0	NA

CRCs were constructed to PE. Values in parentheses are the 95% confidence limits for the slope value. 5-MeU, 5-methylurapidil; NA, not applicable.

In addition to the antagonist data obtained to PE, we repeated these experiments with A-61603 as the agonist in the α_1D-KO (Figure 4). As with PE, prazosin produced a rightward displacement of the curve. However, unlike PE, there was no depression in the maximum response. A pA₂ was calculated and found to be 8.6 with a slope not significantly different from unity, indicating competitive antagonism. 5-Methylurapidil potently inhibited contractions to A-61603 in the α_1D-KO with a higher pA₂ than had been previously calculated for any of the strains of mice used, including the pA₂ obtained against PE contractions in the α_1D-KO. As with prazosin, 5-methylurapidil acted competitively. Relative to the time control, no significant shift occurred with BMY7378. Therefore, no pA₂ or pK_B could be calculated. pA₂ values and slope parameters are presented in Table 3.

Figure 4.

Mean concentration response data to A-61603 in the presence of increasing concentrations of antagonists in α_1D-KO mouse carotid arteries. Mean curves were generated using nonlinear regression upon which the mean±s.e.m. data have been superimposed (n>6).

Table 3.

pA₂ values and slope parameters of antagonists in α_1D-KO mouse carotid arteries

Antagonist	pA2	Slope
Prazosin	8.6	1.2 (0.7–1.8)
5-MeU	8.5	0.9 (0.2–1.8)
BMY7378	ND	ND

CRCs were constructed to A-61603. Values in parentheses are the 95% confidence limits for the slope value. ND, not determined.

Time controls for control, α_1B-KO and α_1D-KO carotid arteries showed no significant change in sensitivity or maximum response to PE or A-61603 (data not shown).

Rat carotid arteries

NA produced concentration-dependent contractions in rat carotid arteries with a pEC₅₀ of 7.9±0.06 and a maximum response of 1.1±0.06 g (n=6) (graphs not shown).

All antagonists used shifted the CRC to NA to the right in a concentration-dependent manner. There was no decrease in the maximum response for any of the antagonists. Schild regression produced pA₂ values of 10.0 for prazosin with a Schild slope indicative of competitive antagonism. 5-Methylurapidil and BMY7378 had pA₂ values of 9.1 and 9.2, respectively, accompanied by shallow Schild slopes, significantly different from unity (Table 4). Time controls showed no significant change in sensitivity or maximum response to NA (data not shown).

Table 4.

pA₂ values and slope parameters of antagonists in rat carotid artery

Antagonist	pA2	Slope
Prazosin	10.7	1.1 (0.4–1.4)
5-MeU	9.0	0.6 (0.3–0.8)
BMY7378	9.8	0.5 (0.3–0.7)

Values in parentheses are the 95% confidence limits for the slope value.

Discussion

This study demonstrates that when the α_1B-KO and α_1D-KO strains of mice are used in conjunction with antagonists, a different pharmacological situation emerges relative to control mice and to each other. However, the pharmacological differences between strains cannot simply be explained in terms of the effects of removing one of the subtypes. Interpreting the pharmacology of the control remains complex and suggests interactions between the subtypes beyond their effects on smooth muscle contraction.

In the α_1B-KO, the α_1D-ARs were apparently isolated, producing robust vasoconstrictor responses that were amenable to classical pharmacological analysis. In contrast, the α_1D-KO responses were less sensitive to PE (though not the α_1A-AR-selective A61603), had a smaller maximum response and responded to selective antagonists with the characteristics of an α_1A-AR. This raises a few related questions: (1) why is there no evidence for an α_1A-AR-mediated response in the control or α_1B-KO? (2) is the α_1A-AR-mediated response present in these arteries but has not been identified with the antagonists used or (3) has it been upregulated?

Receptor subtypes as revealed by agonists

The pattern of the relative potency of agonists for the three mouse strains was inverted for the two agonists tested. For PE, the order of potency was α_1B-KO>control>α_1D-KO and, for A-61603, it was α_1D-KO>control>α_1B-KO (Figure 2 and Table 1). Knepper et al. (1995) have shown that A-61603 is much more potent than PE at α_1A-AR and less potent than PE at α_1D- and α_1B-ARs. In these terms, the responses in the α_1D-KO are strongly correlated with α_1A-AR, while the control and the α_1B-KO correlate better with either α_1B- or α_1D-ARs.

The α_1D-KO does not produce such large contractions as the control or α_1B-KO in response to the α₁-AR agonists PE or A-61603. The maximum response achieved is approximately half of the response produced by the control and the α_1B-KO, yet the responses to 5-HT produced by the three strains of mice are not significantly different. This suggests that the α₁-AR(s) mediating contraction in control and α_1B-KO carotid arteries are either more efficiently coupled to contraction than the α₁-AR mediating contraction in the α_1D-KO or that there are fewer receptors present in the α_1D-KO to mediate a response.

The maximum to 5-HT was not significantly different between the three mouse strains, indicating that the decreased response to PE and A-61603 in the α_1D-KO is not due to a general decline in agonist-mediated responses (indeed sensitivity to 5-HT was enhanced) but seems likely to be a consequence of deleting the α_1D-AR. This deserves closer analysis to determine whether it represents engagement of a subpopulation of smooth muscle cells or a submaximal excitation of each cell; however, the data presented here seem to show that the remaining α_1A- and/or α_1B-ARs are not as efficient as the α_1D-AR when it comes to mediating contraction in the carotid artery, perhaps consistent with their different physiological roles as discussed below.

The increase in sensitivity to 5-HT in the α_1D-KO suggests heterologous upregulation in response to the loss of sensitivity to catecholamines via α₁-ARs. A similar observation has been made in the aorta of the α_1D-KO mouse by Tanoue et al. (2002). Both 5-HT_1A receptors and α₁-ARs are coupled to G_q/11 (Alexander et al., 2004). Therefore, it seems likely that they will share common pathways that are subject to feedback modulation and may be capable of compensating for one another in a KO mouse.

Receptor subtypes as revealed by agonist–antagonist interactions

In the α_1B-KO, the estimated affinity for BMY7378 increased compared with controls. This would be expected if the primary response in the control is α_1D-AR mediated and a secondary (α_1A- or α_1B-AR) component is present. There is no positive evidence for the presence of α_1A-ARs in the control; 5-methylurapidil has lower affinity than in vessels believed to utilise α_1A-AR (Jarajapu et al., 2001a, 2001b; Daly et al., 2002) and the control shows a relatively low sensitivity to A-61603 (Knepper et al., 1995). The analysis of control and α_1B-KO data together suggest that control carotid arteries mediate contraction through α_1D-ARs (primary response) and α_1B-ARs (secondary response) while α_1B-KO carotid arteries mediate contraction solely through α_1D-ARs.

In the α_1D-KO, where only α_1A- or α_1B-AR can be present, the affinity of 5-methylurapidil increases compared with the control, pointing to the presence of α_1A-ARs. To test this hypothesis, 5-methylurapidil was tested against the α_1A-AR agonist A-61603 in the α_1D-KO. 5-Methylurapidil showed still higher affinity, suggesting that α_1A-ARs were indeed contributing to contraction.

There is no positive evidence from control or α_1B-KO data to suggest an α_1A-AR component to their contractions. Therefore, if the α_1A-AR is involved in the functional response of the α_1D-KO, it may be as a result of upregulation of the α_1A-AR from a subfunctional level as a consequence of losing the preferred, dominant receptor, the α_1D-AR.

Comparison with published studies

Evidence has been presented for and against the presence of functional α_1A-ARs in large arteries. Several analyses have favoured α_1D- and/or α_1B-ARs over α_1A-ARs (Muramatsu et al., 1991; 1995; Aboud et al., 1993; Kohno et al., 1994; Testa et al., 1995a, 1995b; Villalobos-Molina & Ibarra, 1996; de Oliveira et al., 1998; Martinez et al., 1999). Furthermore, Rokosh & Simpson (2002) created an α_1A-KO mouse and showed histochemically that Lac-Z, whose gene substituted for the α_1A-ARs gene, was not detected in the major conducting arteries, including the carotid artery. However, there is some evidence in favour of an α_1A-AR response in rat conducting arteries. Gisbert et al. (2003) showed in rat aorta that there was an α_1A-AR-mediated component to the production of inositol phosphates by NA. In addition, our own data from the rat carotid artery show that it is difficult to define the subtype involved in the contractile response to NA. The subtype-selective antagonists BMY7378 and 5-methylurapidil both produced high pA₂ values, although low slopes indicate a complex response consistent with multiple subtypes (Table 4). In the light of the mouse data, we now propose that the rat carotid artery expresses a mixture of α_1A- and α_1D-ARs. The high sensitivity of antagonists to both α_1A- and α_1D-ARs implies not only the presence of both of these subtypes but that they may act synergistically, allowing all antagonists to be effective.

Overall, the present data suggest that large conducting arteries have the potential to express and employ all three α₁-AR subtypes. There is positive evidence for the α_1D-AR in the conducting arteries of mice and rats. There is more controversial evidence that it can be accompanied by an α_1B-AR in normal mice and by either or both of the α_1B-AR and the α_1A-AR in rats. In the α_1B-KO mouse, the α_1D-AR component becomes dominant as expected from simple removal of the α_1B-AR. However, in the α_1D-KO, the remaining response shows characteristics of the α_1A-AR. This suggests that the vessel can withstand the loss of its minor receptor without compensation, but that when its major receptor is lost it compensates by upregulating the α_1A-AR. This is not the first time we have observed this phenomenon. The α_1A-AR is upregulated in the liver of the α_1B-KO mouse (Deighan et al., 2004), which in control mice expresses only the α_1B-AR (Garcia-Sainz et al., 1994; Cavalli et al., 1997; Deighan et al., 2004). Compensatory upregulation of another α₁-AR subtype may be a general response to loss of the major subtype in any particular tissue, be it α_1B- or α_1D-ARs.

Physiological relevance

In general, there seems to be a reciprocal presence of the α_1A-AR and the α_1D-AR in blood vessels. Large noninnervated conductance arteries are associated with expressing the α_1D-AR (Kenny et al., 1995; de Oliveira et al., 1998; Daly et al., 2002; Tanoue et al., 2002), while small innervated resistance vessels are associated with the α_1A-AR (Stassen et al., 1998; Jarajapu et al., 2001b; Daly et al., 2002). The innervated vessels, on the whole, are less sensitive to agonists. This suggests a physiological basis for the relative balance of subtypes based on the balance of humoral and neurogenic control; that is, developmentally, when vessels become innervated they lose α_1D-ARs and gain α_1A-ARs, making them less sensitive to circulating catecholamines but gaining a graded sensitivity to locally released NA, at a higher concentration range. Conducting arteries do not become innervated, so remain sensitive to catecholamines through their α_1D-AR. This would explain why α₁-AR agonists, such as NA or PE, are more potent at α_1D-ARs than at the other subtypes (Minneman et al., 1994; Knepper et al., 1995; Yang et al., 1997). The present work shows that if they are deprived of this natural selection process by deletion of the preferred receptor's gene, they upregulate the alternative catecholamine contractile-signalling receptor that is best able to contract vascular smooth muscle, the α_1A-AR. However, the lower efficiency of the α_1A-AR can only partly compensate in functional terms, as witnessed by the weak submaximal contractions elicited to both PE and A-61603 in the α_1D-KO carotid artery.

Our data seem to suggest the presence of α_1B-ARs in the control; yet, when it is conclusively absent in the α_1B-KO, this has little effect on the artery's contractile ability and in the α_1D-KO there is no evidence for its presence. If it is indeed present in the control, then what functional role does the α_1B-AR have in these blood vessels? Regulatory interactions between the subtypes involving heterodimerisation of α_1B-ARs and the other two subtypes have been proposed from studies of fluorescent-labelled recombinant receptors. The α_1B-AR, which of the three subtypes is the most susceptible to agonist-mediated endocytosis (Chalthorn et al., 2002), can form heterodimers with the other subtypes that can affect their cellular location and expression levels (Uberti et al., 2003; Stanasila et al., 2003). The formation of the α_1A/α_1B heterodimer allowed the α_1A- and α_1B-ARs to cointernalise and consequently increased the extent of agonist-mediated internalisation of the α_1A-AR (Stanasila et al., 2003). In the case of the α_1B/α_1D heterodimer, the presence of the α_1B-AR was found to relocate the α_1D-AR from intracellular sites to the plasma membrane (Uberti et al., 2003; Hague et al., 2004). The present data are the first to show that the functions of α₁-ARs are influenced by the presence of the other subtypes.

The heterodimersation studies may show potential interactions that play a part in long-term receptor regulation involving other factors. The absence of the α_1B-AR seems to cause some sensitisation of the mainly α_1D-AR-mediated contraction of mouse carotid, which might indicate the loss of a regulation of receptor expression or of some other essentially negative effects of the α_1B-AR. The apparent absence of an α_1B-AR-mediated contraction in the α_1D-KO, where the α_1A-AR has taken over function, may indicate an adaptation to counteract the negative influence of the α_1B-AR since α₁-AR function is already compromised. Overall, it seems probable that the α_1B-AR will emerge as a regulatory receptor capable of fine-tuning the properties of the α_1A- and α_1D-ARs.

To summarise, normal mouse carotid arteries have antagonist absolute affinity values that are not consistent with a single α₁-AR subtype but correspond to those expected from a mixed population of at least two and possibly all three subtypes. The α_1B-KO mouse presents a straightforward picture of contractile α_1D-ARs, while the α_1D-KO mouse utilises α_1A-ARs. The emergence of α_1A-ARs when the major subtype, α_1D-AR, is knocked out suggests compensatory upregulation of the α_1A-AR. The α_1B-AR may have a regulatory role to play in control carotid artery by influencing the expression and cellular location of the other subtypes.

Abbreviations

AR

adrenoceptor

CRC

concentration–response curve

5-HT

5-hydroxytryptamine

KO

knockout

NA

noradrenaline

PE

phenylephrine

A-61603

N-[5-(4,5-dihydro-1H-imidazol-2yl)-2-hydroxy-5,6,7,8-tetrahydronaphthalen-1-yl] methanesulphonamide hydrobromide