Crosstalk between presynaptic angiotensin receptors, bradykinin receptors and α₂-autoreceptors in sympathetic neurons: a study in α₂-adrenoceptor-deficient mice

Abstract

1. In mouse atria, angiotensin II and bradykinin lose much or all of their noradrenaline release-enhancing effect when presynaptic α₂-autoinhibition does not operate either because of stimulation with very brief pulse trains or because of treatment with α₂ antagonists. We now studied this operational condition in α₂-adrenoceptor-deficient mice. Release of ³H-noradrenaline was elicited by electrical stimulation.

2. In tissues from wild-type (WT) mice, angiotensin II and bradykinin increased the overflow of tritium evoked by 120 pulses at 3 Hz. This enhancement did not occur or was much reduced when tissues were stimulated by 120 pulses at 3 Hz in the presence of rauwolscine and phentolamine, or when they were stimulated by 20 pulses at 50 Hz.

3. In tissues from mice lacking the α_2A-adrenoceptor (α_2AKO) or the α_2B-adrenoceptor (α_2BKO), the concentration–response curves of angiotensin II and bradykinin (120 pulses at 3 Hz) were unchanged. In tissues from mice lacking the α_2C-adrenoceptor (α_2CKO) or both the α_2A- and the α_2C-adrenoceptor (α_2ACKO), the concentration–response curves were shifted to the same extent downwards.

4. As in WT tissues, angiotensin II and bradykinin lost most or all of their effect in α_2AKO and α_2ACKO tissues when rauwolscine and phentolamine were present or trains consisted of 20 pulses at 50 Hz.

5. Rauwolscine and phentolamine increased tritium overflow evoked by 120 pulses at 3 Hz up to seven-fold in WT and α_2BKO tissues, three-fold in α_2AKO and α_2CKO tissues, and two-fold in α_2ACKO tissues.

6. Results confirm that angiotensin II and bradykinin require ongoing α₂-autoinhibition for the full extent of their release-enhancing effect. Specifically, they require ongoing α_2C-autoinhibition. The peptide effects that remain in α_2C-autoreceptor-deficient mice seem to be because of α_2B-autoinhibition. The results hence also suggest that in addition to α_2A- and α_2C- mouse postganglionic sympathetic neurons possess α_2B-autoreceptors.

Introduction

Angiotensin II and bradykinin increase exocytotic release of noradrenaline in many sympathetically innervated tissues (for a review, see Langer, 1977; Starke, 1977; Westfall, 1977; Fuder & Muscholl, 1995; Boehm & Kubista, 2002). Angiotensin II acts on presynaptic AT₁-receptors (Boehm & Kubista, 2002), which in some tissues differ from smooth muscle AT₁-receptors in their pharmacological properties (Guimarães et al., 2001; Pinheiro et al., 2002). Bradykinin acts on presynaptic B₂-receptors (Fuder & Muscholl, 1995; Boehm & Kubista, 2002). The ensuing presynaptic signal transduction probably involves G_q and activation of protein kinase C (Boehm & Kubista, 2002).

It was early observed that angiotensin II lost its release-enhancing effect in rabbit isolated hearts when the hearts had been pretreated with phenoxybenzamine, which by itself increased the release of noradrenaline because of irreversible α₂-autoreceptor blockade (Starke & Schümann, 1972). The observation lay dormant for many years. It was then extended in several ways. First, blockade of α₂-adrenoceptors also prevented or greatly attenuated the presynaptic effect of angiotensin II in guinea-pig atria (Brasch et al., 1995), mouse atria (Cox et al., 2000) and rat hearts (Kurz et al., 1997; Mota & Guimarães, 2002), and when the competitive α-adrenoceptor antagonists phentolamine, idazoxan or yohimbine were used instead of phenoxybenzamine (Brasch et al., 1995; Kurz et al., 1997; Cox et al., 2000; Mota & Guimarães, 2002). Second, an analogous interaction was observed between α-adrenoceptor antagonists and bradykinin: like angiotensin II, bradykinin lost or almost lost its release-enhancing effect in mouse atria treated with phentolamine or the selective α₂-adrenoceptor antagonist rauwolscine (Cox et al., 2000). Third, the two peptides also failed to enhance noradrenaline release when, in the absence of any α₂ antagonist, the sympathetic axons were stimulated under α₂-autoinhibition-free or -poor conditions such as by brief high-frequency pulse trains (Brasch et al., 1995; Cox et al., 2000). Finally, when the effect of angiotensin II and bradykinin in mouse atria had been minimized by phentolamine, it could be restored by administration of neuropeptide Y or [D-Ser²]-leucine enkephalin-Thr⁶ (DSLET), which by themselves reduced noradrenaline release through presynaptic neuropeptide Y Y₂- and opioid OP₁-receptors, respectively (Cox et al., 2000).

From these results, it has been concluded that (1) both angiotensin II and bradykinin require ongoing α₂-autoinhibition for their full presynaptic release-facilitating effect; (2) in the absence of α₂-autoinhibition the effect of the peptides can be restored by activation of another G_i/o-coupled presynaptic receptor; and (3) protein kinase C, the likely effector of the AT₁- and B₂-receptors, increases noradrenaline release to a large extent by inactivating a component of the G_i/o presynaptic inhibitory pathway (Brasch et al., 1995; Cox et al., 2000; see also Boehm & Kubista, 2002, Figure 5c).

Figure 5.

Effect of angiotensin II and bradykinin on the evoked overflow of tritium from α_2AKO mouse atria (a, b) and vasa deferentia (c, d) stimulated under different degrees of α₂-autoinhibition. The preparations were stimulated for six periods (S₁–S₆) by either 120 pulses at 3 Hz or 20 pulses at 50 Hz. Angiotensin II and bradykinin were added at increasing concentrations 12 min before S₂–S₆. In some experiments, phentolamine 1 μM and rauwolscine 1 μM were present throughout superfusion. Symbols represent the evoked overflow of tritium, expressed as a percentage of control (no peptide added). Values are means±s.e.m. from 5 to 19 tissue pieces. Significant differences from control, no peptide: #P<0.05. Significant differences from 120 pulses/3 Hz without α-adrenoceptor antagonists: ^*P<0.05.

There is a third approach to avoid α₂-autoinhibition, in addition to autoreceptor blockade and autoinhibition-free stimulation conditions, namely disruption of the α₂-autoreceptor genes. It is now established that the main α₂-autoreceptor is the α_2A-adrenoceptor (see Hein, 2001; Starke, 2001). However, especially in sympathetically innervated tissues the α_2C-adrenoceptor also functions as an autoreceptor; there is no firm evidence for α_2B-autoreceptors (Hein et al., 1999; see Hein, 2001; Starke, 2001).

The present experiments were devised to study the crosstalk between AT₁- and B₂-receptors on the one hand, and α₂-autoreceptors on the other by means of genetically manipulated mice: mice lacking the α_2A-adrenoceptor (α_2AKO, knockout), mice lacking the α_2B-adrenoceptor (α_2BKO), mice lacking the α_2C-adrenoceptor (α_2CKO), mice lacking both the α_2A- and the α_2C-adrenoceptor (α_2ACKO), and wild-type (WT) mice sharing the genetic background of the KO strains.

Since the crosstalk had been demonstrated previously only in the heart, we used in addition another tissue, the mouse vas deferens. Finally, we studied the interaction of bradykinin with α-adrenoceptor antagonists, previously tested in mouse atria only, in rat and rabbit atria. A mixture of phentolamine and rauwolscine was used to block α₂-autoreceptors when desired, because phentolamine is especially potent at rodent α_2A-adrenoceptors whereas rauwolscine is especially potent at α_2C-adrenoceptors, the two adrenoceptors known to operate physiologically as autoreceptors in mouse atria and vasa deferentia (Hein et al., 1999; Trendelenburg et al., 2001).

Methods

The α₂-adrenoceptor-deficient mice have been described previously (Altman et al., 1999; Hein et al., 1999). Male C57BL/6 × 129Sv (WT), α_2AKO, α_2BKO, α_2CKO and α_2ACKO mice aged 2–15 months, male Wistar rats (220–300 g) and male or female rabbits (1.8–2.7 kg) were killed by cervical dislocation. The wall of each pair of mouse, rat or rabbit atria was cut into pieces (mouse 6–8; rat 12–15; rabbit 20–30), and each mouse vas deferens was cut into six pieces. The tissue segments were preincubated in 2 ml medium containing 0.2 μM ³H-noradrenaline for 30 min at 37°C. Segments were then placed in superfusion chambers between platinum electrodes, one segment per chamber, where they were superfused with ³H-noradrenaline-free medium at a rate of 1.2 ml min⁻¹, also at 37°C. Successive 2-min samples of the superfusate were collected from t=50 min onwards (t=0 min being the start of superfusion). At the end of experiments, tissues were dissolved and tritium was determined in superfusate chambers and tissues.

The superfusion medium contained (mM): NaCl 118, KCl 4.8, CaCl₂ 2.5 unless stated otherwise, MgSO₄ 1.2, NaHCO₃ 25, KH₂PO₄ 1.2, glucose 11, ascorbic acid 0.57, disodium EDTA 0.03 and desipramine 0.001. The medium for preincubation with ³H-noradrenaline contained no desipramine and only 0.2 mM CaCl₂ (Limberger et al., 1992).

There were seven periods of electrical stimulation. Stimulation consisted of rectangular pulses of 1 ms width and 47 V cm⁻¹ voltage drop between the electrodes of each chamber, yielding a current strength of 80 mA. The first stimulation period (180 pulses at 3 Hz) was delivered at t=30 min and was not used for the determination of tritium overflow. The subsequent stimulation periods (S₁–S₆) were applied at t=54, 72, 90, 108, 126 and 144 min. They consisted of either 120 pulses at 3 Hz, 36 pulses at 3 Hz or 20 pulses at 50 Hz. Concentration–response curves for the release-enhancing effect of angiotensin II and bradykinin were determined by the addition of the peptides at increasing concentrations 12 min before S₂–S₆.

The outflow of tritium was calculated as a fraction of the tritium content of the tissue at the onset of the respective collection period (fractional rate; min⁻¹). The overflow elicited by electrical stimulation was calculated as the difference ‘total tritium outflow during and after stimulation' minus ‘basal outflow', and was then expressed as a percentage of the tritium content of the tissue at the onset of stimulation (see Trendelenburg et al., 1997). For further evaluation, overflow ratios S_n/S₁ were calculated. Overflow ratios obtained after addition of a drug were also calculated as a percentage of the corresponding ratio in controls in which no drug was added. Effects of drugs on basal tritium efflux were evaluated similarly.

Results are expressed as arithmetic means±s.e.m. Groups were tested for significant differences with the Mann–Whitney test and Bonferroni correction. An error probability P<0.05 was taken to be a significant difference, n is the number of tissue pieces.

Drugs

(−)-[Ring-2,5,6-³H]-noradrenaline, specific activity 51.8–56.0 Ci mmol⁻¹, was from DuPont, Dreieich, Germany; angiotensin II (human), bradykinin, desipramine hydrochloride, rauwolscine hydrochloride and (R)-(+)-[2,3-dihydro-5-methyl-3-(4-morpholinylmethyl)pyrrolo[1,2,3-de]-1,4-benzoxan-6-yl]-1-naphthalenylmethanone mesylate (WIN 55,212-2) were from Sigma, Deisenhofen, Germany; phentolamine hydrochloride was a gift from Ciba-Geigy, Basel, Switzerland. Drugs were dissolved in water except WIN 55,212-2 (dimethylsulfoxide).

Results

The primary results from experiments of this kind are tritium efflux-versus-time curves with the characteristic peaks caused by electrical stimulation. Such curves have been published from our laboratory for all tissues examined (mouse atria Wahl et al. (1996), mouse vas deferens Altman et al. (1999), rat atria Limberger et al. (1992), rabbit atria Limberger et al. (1995)). The fractional rates of basal tritium efflux and the size of the stimulation peaks were similar to the previous work. Values for the overflow of tritium elicited by S₁ are summarized in Table 1. In control experiments without drug administration after S₁, the peaks remained similar from S₁ to S₆ (n=8–27). None of the drugs changed the basal efflux of tritium.

Table 1.

Electrically evoked overflow of tritium from mouse, rat and rabbit tissues

Species	Mouse strain	Tissue	Overflow (% of tissue tritium) evoked by
			120 pulses/3 Hz	20 pulses/50 Hz	120 pulses/3 Hz (+phentolamine 1 μM+rauwolscine 1 μM)
Mouse	WT	Atria	0.984±0.042 (n=77)	0.367±0.013 (n=29)	4.314±0.134** (n=24)
	α2AKO		1.428±0.086* (n=63)	0.494±0.032* (n=12)	4.451±0.181** (n=18)
	α2BKO		0.965±0.071 (n=40)
	α2CKO		1.284±0.067* (n=32)
	α2ACKO		2.154±0.068* (n=30)	0.323±0.023 (n=15)	2.121±0.124 (n=24)
Mouse	WT	Vas deferens	0.313±0.008 (n=90)	0.388±0.018 (n=27)	2.314±0.174**,a (n=23)
	α2AKO		0.522±0.024* (n=61)	0.438±0.020 (n=15)	2.045±0.239** (n=25)
	α2BKO		0.275±0.011* (n=51)
	α2CKO		0.450±0.027* (n=37)
	α2ACKO		0.651±0.029* (n=48)	0.439±0.018 (n=35)	1.543±0.143** (n=19)
Rat		Atria	1.302±0.069 (n=34)	0.214±0.009 (n=22)	4.055±0.132** (n=38)
Rabbit		Atria	0.630±0.045 (n=39)	0.287±0.025 (n=24)	2.266±0.165** (n=25)

Preparations were incubated with ³H-noradrenaline and then superfused with a medium containing desipramine 1 μM and in some cases also other drugs as indicated. Values represent the overflow of tritium evoked by the first stimulation period, S₁. Means±s.e.m. from n superfusion chambers.

^aIn the presence of WIN 55,212-2 0.1 μM: 0.486±0.036 (n=23); at a lower Ca²⁺-concentration (0.65 instead of 2.5 mM): 0.286±0.024 (n=24); when S₁ consisted of 36 pulses/3 Hz instead of 120 pulses/3 Hz: 0.589±0.036 (n=24).

^*Signficant differences (P<0.05) from WT.

^**Significant differences (P<0.05) from absence of phentolamine 1 μM and rauwolscine 1 μM.

Peptide effects in WT mouse tissues

Initial experiments were carried out with tissues from WT mice. When these tissues were stimulated by trains of 120 pulses at 3 Hz, marked autoinhibition developed as shown by the large increases of S₁ overflow values caused by the combined administration of phentolamine 1 μM and rauwolscine 1 μM throughout superfusion: a more than four-fold increase in WT atria and a more than seven-fold increase in WT vas deferens (Table 1).

Under these conditions (and in the absence of α-adrenoceptor antagonists), angiotensin II and bradykinin greatly increased the evoked overflow of tritium (filled circles in Figure 1). The maximum increase caused by angiotensin II in atria was by about 80%, and the EC₅₀, that is the concentration causing a 40% increase, was 0.21 nM (calculated by interpolation from the nearest points of the concentration–response curve; Figure 1a). The maximum increase caused by bradykinin in atria also was by about 80%, and the EC₅₀ 0.16 nM (Figure 1b). The maximum increase caused by angiotensin II in the vas deferens was by about 60%, and the EC₅₀ 0.08 nM (Figure 1c). The maximum increase caused by bradykinin in the vas deferens was by about 80%, also with an EC₅₀ of 0.08 nM (Figure 1d). The values are similar to previously determined values in atria and vasa deferentia of NMRI mice (Cox et al. (2000) and Trendelenburg et al. (2000); no previous values available for angiotensin II in the vas deferens).

Figure 1.

Effect of angiotensin II and bradykinin on the evoked overflow of tritium from WT mouse atria (a, b) and vasa deferentia (c, d) stimulated under different degrees of α₂-autoinhibition. The preparations were stimulated for six periods (S₁–S₆) by either 120 pulses at 3 Hz or 20 pulses at 50 Hz. Angiotensin II and bradykinin were added at increasing concentrations 12 min before S₂–S₆. In some experiments, phentolamine 1 μM and rauwolscine 1 μM were present throughout superfusion. Symbols represent the evoked overflow of tritium, expressed as a percentage of control (no peptide added). Values are means±s.e.m. from 9 to 29 tissue pieces. Significant differences from control, no peptide: #P<0.05. Significant differences from 120 pulses/3 Hz without α-adrenoceptor antagonists: ^*P<0.05.

As in our previous work on atria of NMRI mice (Cox et al., 2000), the release-enhancing effects of angiotensin II and bradykinin in atria as well as the vas deferens of WT mice were significantly and greatly reduced or even abolished when either, in experiments with 120 pulses at 3 Hz, α₂-autoreceptors were blocked by addition of phentolamine 1 μM and rauwolscine 1 μM throughout superfusion (empty diamonds in Figure 1), or when the tissues were stimulated by trains of 20 pulses at 50 Hz (empty squares in Figure 1). Only minor α₂-autoinhibition develops during the latter brief high-frequency pulse trains (Trendelenburg et al., 1999;2000;2001; Cox et al., 2000).

In the atria of NMRI mice stimulated by trains of 120 pulses at 3 Hz and treated with phentolamine, the effects of angiotensin II and bradykinin fully reappeared when the atria were additionally treated with neuropeptide Y or DSLET (Cox et al., 2000). We carried out an analogous experiment in the WT vas deferens, activating another G_i/o-coupled receptor, the cannabinoid CB₁-receptor, by the selective agonist WIN 55,212-2 (see Trendelenburg et al., 2000). WIN 55,212-2 indeed completely re-established the facilitation by angiotensin II and bradykinin (Figure 2).

Figure 2.

Effect of angiotensin II and bradykinin on the evoked overflow of tritium from WT mouse vasa deferentia in the presence of α-adrenoceptor antagonists: influence of WIN 55,212-2. The preparations were stimulated for six periods (S₁–S₆) by 120 pulses at 3 Hz. Angiotensin II and bradykinin were added at increasing concentrations 12 min before S₂–S₆. In some experiments, phentolamine 1 μM and rauwolscine 1 μM, in other experiments these two drugs and in addition WIN 55,212-2 0.1 μM were present throughout superfusion. Symbols represent the evoked overflow of tritium, expressed as a percentage of control (no peptide added). Values are means±s.e.m. from 8 to 29 tissue pieces. Significant differences from control, no peptide: #P<0.05. Significant differences from 120 pulses/3 Hz without α-adrenoceptor antagonists and WIN 55,212-2: ^*P<0.05.

WIN 55,212-2 had a second effect: it greatly reduced the evoked overflow of tritium from the high level attained in the presence of phentolamine and rauwolscine (S₁; footnote ‘a' in Table 1). We wished to make sure that it was the presence of WIN 55,212-2 and not just this reduction in overflow that re-established the effects of the peptides. To this end, we lowered the evoked overflow in two other ways: by a decrease in the concentration of Ca²⁺ in the medium to 0.65 mM, and by a decrease in the number of pulses per train to 36. In both cases the overflow at S₁, in the presence of phentolamine and rauwolscine, was close to the overflow evoked by 120 pulses at 3 Hz in normal Ca²⁺ in the presence of phentolamine, rauwolscine and WIN 55,212-2 (S₁; footnote ‘a' in Table 1). In contrast to WIN 55,2122 (Figure 2), neither low Ca²⁺ (Figure 3) nor stimulation with 36 pulses per train (not shown; n=8) re-established the effects of angiotensin II and bradykinin: the interaction with WIN 55,212-2 was a specific one.

Figure 3.

Effect of angiotensin II and bradykinin on the evoked overflow of tritium from WT mouse vasa deferentia in the presence of α-adrenoceptor antagonists: influence of the concentration of Ca²⁺. The preparations were stimulated for six periods (S₁–S₆) by 120 pulses at 3 Hz. Angiotensin II and bradykinin were added at increasing concentrations 12 min before S₂–S₆. In some experiments, phentolamine 1 μM and rauwolscine 1 μM were present throughout superfusion. The superfusion medium contained either the normal (2.5 mM) or a lower concentration of Ca²⁺ (0.65 mM). Symbols represent the evoked overflow of tritium, expressed as a percentage of control (no peptide added). Values are means±s.e.m. from 8 to 29 tissue pieces. Significant differences from control, no peptide: #P<0.05. Significant differences from 120 pulses/3 Hz, 2.5 mM Ca²⁺, without α-adrenoceptor antagonists: ^*P<0.05.

Peptide effects in KO mouse tissues

We then proceeded to the receptor-deficient mice. Tissues were initially stimulated by trains of 120 pulses at 3 Hz in the absence of α-adrenoceptor antagonists. The S₁ overflow responses under these conditions were greater in atria and vasa deferentia from α_2AKO, α_2CKO and α_ACKO mice than in WT tissues (Table 1), in accord with the physiological function of α_2A- and α_2C-adrenoceptors as autoreceptors. Neither deletion of the α_2A-adrenoceptor nor deletion of the α_2B-adrenoceptor changed the effects of angiotensin II and bradykinin: the peptide concentration–response curves in the α_2AKO and α_2BKO tissues were practically identical with the curves in the WT tissues, with no significant difference from the latter (Figure 4). Only one of the single-gene disruptions, that of the α_2C gene, caused a change: the concentration–response curves of angiotensin II and bradykinin in α_2CKO atria and vasa deferentia were consistently shifted downwards to reach lower maxima. Elimination of the α_2A-adrenoceptor in addition to the α_2C-receptor produced no further change: the peptide concentration–response curves in α_2ACKO tissues were superimposed on the curves in the α_2CKO tissues (Figure 4).

Figure 4.

Effect of angiotensin II and bradykinin on the evoked overflow of tritium from atria (a, b) and vasa deferentia (c, d) taken from WT and various α₂-adrenoceptor-deficient mice. The preparations were stimulated for six periods (S₁–S₆) by 120 pulses at 3 Hz. Angiotensin II and bradykinin were added at increasing concentrations 12 min before S₂–S₆. Symbols represent the evoked overflow of tritium, expressed as a percentage of control (no peptide added). Values are means±s.e.m. from 5 to 29 tissue pieces. All peptide effects were significant (P<0.05) versus control, no peptide, except in some cases at 0.01 and 0.1 nM. Significant differences from WT: ^*P<0.05.

The question arose whether the facilitatory effects of the peptides that remained when one or both of the autoreceptors were deleted were independent of α₂-autoinhibition. We therefore stimulated the α_2AKO and α_2ACKO tissues also by trains of 120 pulses at 3 Hz in the presence of phentolamine and rauwolscine, or by trains of 20 pulses at 50 Hz. As shown in Table 1, the antagonist mixture increased S₁ overflow values in α_2AKO atria, α_2AKO vasa deferentia and α_2ACKO vasa deferentia although not in α_2ACKO atria, indicating the operation of autoinhibition in the former three tissues. The effects of angiotensin II and bradykinin behaved like in WT tissues: in α_2AKO atria and vasa deferentia (Figure 5) as well as in α_2ACKO atria and vasa deferentia (Figure 6) the peptide effects were attenuated or abolished when either, in experiments with 120 pulses at 3 Hz, α₂-adrenoceptors were blocked (empty diamonds in Figure 5 and Figure 6) or when, in the absence of blockers, stimulation was changed to trains of 20 pulses at 50 Hz (empty squares in Figure 5 and Figure 6). There was one exception: the α antagonists did not reduce the effect of angiotensin II in the α_2ACKO vas deferens, possibly because the effect was small in that tissue even in the absence of the blockers (Figure 6c; but angiotensin II also did not cause significant facilitation in the α-adrenoceptor-blocked α_2ACKO vas deferens as compared to the control without peptide).

Figure 6.

Effect of angiotensin II and bradykinin on the evoked overflow of tritium from α_2ACKO mouse atria (a, b) and vasa deferentia (c, d) stimulated under different degrees of α₂-autoinhibition. The preparations were stimulated for six periods (S₁–S₆) by either 120 pulses at 3 Hz or 20 pulses at 50 Hz. Angiotensin II and bradykinin were added at increasing concentrations 12 min before S₂–S₆. In some experiments, phentolamine 1 μM and rauwolscine 1 μM were present throughout superfusion. Symbols represent the evoked overflow of tritium, expressed as a percentage of control (no peptide added). Values are means±s.e.m. from 5 to 10 tissue pieces. Significant differences from control, no peptide: #P<0.05. Significant differences from 120 pulses/3 Hz without α-adrenoceptor antagonists: ^*P<0.05.

Assessment of α₂-autoinhibition in mouse tissues

The degree of α₂-autoinhibition can be estimated from the increase of S₁ overflow values caused by α₂ antagonists that are present throughout superfusion (Table 1). However, this procedure is not very reliable (see Limberger et al., 1995, p. 32). A more accurate measure of autoinhibition is the increase in overflow caused by α₂ antagonists when they are added after S₂, that is their effect on S₃–S₆. With this protocol, each tissue segment serves as its own control. When added together after S₂, phentolamine 1 μM and rauwolscine 1 μM increased the overflow of tritium evoked by 120 pulses at 3 Hz in all five mouse strains (Figure 7). The increase was greatest in WT tissues: three- to four-fold in atria and more than seven-fold in the vas deferens (Figure 7a). Deletion of the α_2B-adrenoceptor caused no significant change (Figure 7c). Deletion of both the α_2A- and the α_2C-adrenoceptor alone significantly reduced or tended to reduce the effect of the antagonists, but in either case an approximately three-fold increase remained (Figure 7b and d). While these results had been expected (see Trendelenburg et al., 2001), it was surprising to see that the antagonists continued to cause some increase, about two-fold, even in the α_2ACKO tissues, that is when both known α₂-autoreceptors had been deleted (Figure 7e). Antagonist addition after S₂ thus revealed α₂-autoinhibition also in α_2ACKO atria (filled diamonds in Figure 7e) in which the S₁ comparison had not detected it (Table 1; see above).

Figure 7.

Effect of a mixture of phentolamine 1 μM and rauwolscine 1 μM on the evoked overflow of tritium from atria and vasa deferentia taken from WT and various α₂-adrenoceptor-deficient mice. The preparations were stimulated for six periods (S₁–S₆) by 120 pulses at 3 Hz. Phentolamine 1 μM and rauwolscine 1 μM were added simultaneously 12 min before S₃. Symbols represent the evoked overflow of tritium, expressed as a percentage of control (no α antagonist added). Values are means±s.e.m. from 5 to 12 tissue pieces. Significant differences from control, no α antagonist: #P<0.05. Significant differences from WT: ^*P<0.05. The significance symbols refer to S₃–S₆.

Rat and rabbit atria

Marked autoinhibition developed in rat and rabbit atria when they were stimulated by trains of 120 pulses at 3 Hz: combined administration of phentolamine 1 μM and rauwolscine 1 μM throughout superfusion increased the overflow at S₁ three-fold in the rat and three- to four-fold in the rabbit (Table 1).

Under these conditions (and in the absence of α-adrenoceptor antagonists), angiotensin II and bradykinin greatly increased the evoked overflow of tritium from rat atria (filled circles in Figure 8). The maximum increase caused by angiotensin II was by about 60%, and the EC₅₀, that is the concentration causing a 30% increase, was 0.30 nM (calculated by interpolation from the nearest points of the concentration–response curve; Figure 8a). The maximum increase caused by bradykinin in rat atria also was by about 60%, and the EC₅₀ was 0.05 nM (Figure 8b). Still under these conditions, angiotensin II also greatly increased the evoked overflow of tritium from rabbit atria, with maximal enhancement by 130% and an EC₅₀ of 0.27 nM (filled circles in Figure 9a). Bradykinin, in contrast, failed to cause any change in rabbit atria (filled circles in Figure 9b).

Figure 8.

Effect of angiotensin II and bradykinin on the evoked overflow of tritium from rat atria stimulated under different degrees of α₂-autoinhibition. The preparations were stimulated for six periods (S₁–S₆) by either 120 pulses at 3 Hz or 20 pulses at 50 Hz. Angiotensin II and bradykinin were added at increasing concentrations 12 min before S₂–S₆. In some experiments, phentolamine 1 μM and rauwolscine 1 μM were present throughout superfusion. Symbols represent the evoked overflow of tritium, expressed as a percentage of control (no peptide added). Values are means±s.e.m. from 8 to 12 tissue pieces. Significant differences from control, no peptide: #P<0.05. Significant differences from 120 pulses/3 Hz without α-adrenoceptor antagonists: ^*P<0.05.

Figure 9.

Effect of angiotensin II and bradykinin on the evoked overflow of tritium from rabbit atria stimulated under different degrees of α₂-autoinhibition. The preparations were stimulated for six periods (S₁–S₆) by either 120 pulses at 3 Hz or 20 pulses at 50 Hz. Angiotensin II and bradykinin were added at increasing concentrations 12 min before S₂–S₆. In some experiments, phentolamine 1 μM and rauwolscine 1 μM were present throughout superfusion. Symbols represent the evoked overflow of tritium, expressed as a percentage of control (no peptide added). Values are means±s.e.m. from 8 to 11 tissue pieces. Significant differences from control, no peptide: #P<0.05. Significant differences from 120 pulses/3 Hz without α-adrenoceptor antagonists: ^*P<0.05.

As in the WT mouse tissues, the release-enhancing effects of angiotensin II and bradykinin in rat atria were greatly reduced or even abolished when either, in experiments with 120 pulses at 3 Hz, α₂-autoreceptors were blocked by addition of a mixture of phentolamine 1 μM and rauwolscine 1 μM throughout superfusion (empty diamonds in Figure 8), or when the tissues were stimulated by trains of 20 pulses at 50 Hz (empty squares in Figure 8). The effect of angiotensin II in rabbit atria equally was much depressed under these conditions (empty symbols in Figure 9a). Bradykinin remained without effect in rabbit atria (empty symbols in Figure 9b).

Discussion

Our results confirm that angiotensin II and bradykinin can enhance exocytotic release of noradrenaline from the sympathetic neurons innervating the mouse heart, the rat heart and the mouse vas deferens, and that angiotensin II can enhance noradrenaline release in the rabbit heart (see Starke (1977), Fuder & Muscholl (1995) and Boehm & Kubista (2002); for mouse vas deferens: Trendelenburg et al., (2000) and Schelb et al. (2001)). The lack of any effect of bradykinin in rabbit atria even under autoinhibition-rich conditions (Figure 9b) may be because of a dual action. Bradykinin reduces the release of noradrenaline in the rabbit perfused heart and in other rabbit tissues by promotion of the synthesis of prostaglandins, which then cause presynaptic inhibition (Starke et al., 1977; Rónai, 1991). This component may have counteracted any possible direct facilitatory effect of bradykinin in the present experiments on superfused atrial segments, thus leading to a zero net effect. The effects of angiotensin II and bradykinin in mouse atria and vasa deferentia and rat atria are mediated by AT₁- and B₂-receptors, respectively (see references above). The type of presynaptic angiotensin receptor in the rabbit heart does not seem to have been identified.

Our results also confirm that angiotensin II loses much or all of its presynaptic facilitatory effect in mouse, rat and rabbit cardiac preparations when either α₂-autoreceptors are pharmacologically blocked or the sympathetic nerves stimulated in a manner leading to no or little autoinhibition, and that the same loss occurs with the facilitatory effect of bradykinin in mouse atria (Starke & Schümann, 1972; Kurz et al., 1997; Cox et al., 2000; Mota & Guimarães, 2002). To these instances, our experiments add the demonstration of analogous conditions for the effects of angiotensin II and bradykinin in the mouse vas deferens and for the effect of bradykinin in rat atria. It should be noted that we used NMRI mice in our previous study (Cox et al., 2000) but a different strain, the WT of the genetically engineered mice, in the present work: results were the same.

We have suggested that angiotensin II and bradykinin increase the release of noradrenaline at least to a large extent by interrupting an ongoing α₂-autoinhibition. More specifically, we have suggested that one effector of AT₁- and B₂-receptors, namely protein kinase C, inactivates a protein involved in the presynaptic α₂-autoreceptor→G_i/o→presynaptic calcium channel inhibitory pathway (Cox et al., 2000). In support of this view, when α₂-autoinhibition was absent, activation of other G_i/o-coupled presynaptic receptors, namely neuropeptide Y Y₂- and opioid OP₁-receptors, enabled angiotensin II and bradykinin to produce their full scope of facilitation (Cox et al., 2000). Again we now add a further example: in the mouse vas deferens, cannabinoids reduce the release of noradrenaline by the activation of G_i/o-coupled CB₁-receptors, and exposure to the CB₁ agonist WIN 55,212-2 restored the full extent of facilitation by angiotensin II and bradykinin (Figure 2). Presynaptic CB₁-receptor do not operate in mouse atria (Trendelenburg et al., 2000), so the interaction experiment had to be limited to the vas deferens. It has been shown a number of years ago that angiotensin II caused a greater increase in noradrenaline release in rabbit atria when presynaptic muscarinic receptors were activated simultaneously (Garcia-Sevilla et al., 1985). Perhaps, this was because of an analogous mechanism, that is in this case muscarinic stimulation of the presynaptic G_i/o pathway, thus giving angiotensin II greater space for disinhibition.

The center of the present work, the experiments on α₂-adrenoceptor-deficient tissues, gave both expected and unexpected results. Since there is no firm evidence for α_2B-autoreceptors, we did not expect changes in the effects of angiotensin II and bradykinin in the α_2BKO tissues: and in fact, there was no change (Figure 4). Since the α_2A-adrenoceptor is generally considered to be the major α₂-autoreceptor, we had expected a clear attenuation of the peptide effects in the α_2AKO tissues; however, here also no change occurred (Figure 4). Of the single deletions, only deletion of the α_2C-adrenoceptor blunted the effects of the peptides (Figure 4). The most unexpected findings were those in the α_2ACKO tissues. Even when both known autoreceptors had been deleted, angiotensin II and bradykinin still facilitated the release of noradrenaline, and did so to no smaller extent than after deletion of the α_2C-adrenoceptor alone (Figure 4). All effects of the peptides remaining after gene deletion, in the α_2AKO as well as in the α_2ACKO tissues, continued to depend on an ongoing α₂-autoinhibition: when there was no autoinhibition, the effects shrank or disappeared (Figure 5 and Figure 6). Finally, this remaining autoinhibition was demonstrated directly: although reduced, it clearly operated in α_2AKO, α_2CKO and even α_2ACKO tissues (Figure 7). It should also be noted that the degree of autoinhibition in α_2AKO and α_2CKO tissues was very similar (Figure 7b and d), indicating that the selective attenuation of the peptide effects in α_2CKO but not the α_2AKO tissues was not because of a lower overall autoinhibitory tone in the former.

These results tell us something on the prerequisites for AT₁- and B₂-receptor-mediated presynaptic facilitation. They simultaneously change our concept of α₂-autoinhibition.

The first conclusion is that although the main autoreceptor is α_2A (see Hein, 2001; Starke, 2001), it is the α_2C- rather than α_2A-autoinhibition that permits or promotes the presynaptic effects of angiotensin II and bradykinin: deletion of the α_2A-adrenoceptor caused no change in the peptide effects, neither when it was the only deletion (compare α_2AKO to WT in Figure 4) nor when the α_2C-adrenoceptor was also deleted (compared α_2ACKO to α_2CKO in Figure 4). It cannot be excluded that α_2A- and α_2C-autoreceptors couple to different transduction pathways, above all to different G_i/o heterotrimers (see Hein, 2001; Starke, 2001). The present results may be an incitement to search for the difference that makes α_2C signal transduction selectively sensitive to angiotensin II and bradykinin.

The second conclusion, rather provocative for autoreceptor theory, is that there must be another autoreceptor that permits facilitation by angiotensin II and bradykinin: the autoreceptor responsible for the remaining facilitation in the α_2ACKO tissues (Figure 4). It should be remembered that this remaining facilitation was α₂-autoinhibition dependent (Figure 6) and that α₂-autoinhibition was indeed present (Figure 7e). Given the fact that the receptor cannot be α_2A or α_2C, and given the belief that only three α₂-adrenoceptor types exist (see Hein, 2001): are there α_2B-autoreceptors, and is the transduction pathway behind these, like the α_2C transduction pathway, sensitive to angiotensin II and bradykinin?

Two contra arguments immediately come to mind: deletion of the α_2B gene alone changed neither the degree of autoinhibition (Figure 7c) nor the concentration–response curves of angiotensin II and bradykinin (Figure 4). However, these arguments do not exclude the existence of α_2B-autoreceptors. There was a similar series of findings when the α_2C-autoreceptors were discovered. Initial experiments had shown that some noradrenaline release-inhibiting effect of α₂-adrenoceptor agonists persisted after α_2A-adrenoceptors had been genetically deleted (Altman et al., 1999). Deletion of α_2C-adrenoceptors alone did not change the effect of the agonists, and this seemed to argue against α_2C-autoreceptors (Hein et al., 1999; Trendelenburg et al., 1999). It was only when Hein et al. (1999) generated the double KO, α_2ACKO, that the non-α_2A-autoreceptors were identified as α_2C: deletion of the α_2C- in addition ot the α_2A-adrenoceptor abolished the remaining inhibition by α₂ agonists (Hein et al., 1999). Perhaps, certain functions that depend on two α₂-autoreceptor types remain unchanged when only one contributing type is deleted (the α_2C-adrenoceptor in the previous work on α₂ agonists, the α_2B-adrenoceptor in the present work): the remaining contributing type compensates for the loss. The α_2B-adrenoceptor gene is expressed in the rat superior cervical ganglion (Vidovic & Hill, 1997). This gives some support to the idea of α_2B-autoreceptors. In fact, in a study of the influence of α₂-autoinhibition on the stimulation frequency–noradrenaline release relation, the possibility of α_2B-autoreceptors has already cautiously been raised (Scheibner et al., 2001).

In conclusion, we haved extended the evidence for a crosstalk between presynaptic α₂-autoreceptors, angiotensin AT₁- and bradykinin B₂-receptors to further tissues. The crosstalk is a frequent phenomenon. We have also extended the evidence that stimulation of the presynaptic G_i/o inhibitory pathway through other receptors, namely neuropeptide Y Y₂-, opioid OP₁- and–this is the new finding–cannabinoid CB₁-receptors can replace G_i/o stimulation through α₂-autoreceptors to permit full presynaptic facilitation by angiotensin II and bradykinin. The results should not be interpreted as indicating that interruption of α₂-autoinhibition is the sole presynaptic mode of action of angiotensin II and bradykinin; there may be α₂-autoinhibition-independent mechanisms (see Costa & Majewki, 1988; Mota et al., 2000); indeed, small facilitatory effects of the two peptides in the presence of phentolamine plus rauwolscine or during brief high-frequency stimulation were occasionally seen throughout the present work (e.g. Figure 1 and Figure 8); however, in the preparations examined the interruption of autoinhibition prevailed. Experiments on α₂-adrenoceptor subtype-deficient mice indicate that it is the α_2C-autoreceptor and not the α_2A-autoreceptor that interacts with the AT₁- and B₂-receptors. However, α_2B-autoreceptors may also contribute to the crosstalk. The existence of α_2B- in addition to the established α_2A- and α_2C-autoreceptors is a somewhat provocative perspective opened by our study.

Abbreviations

Dslet

[D-Ser²]-leucine enkephalin-Thr⁶

KO

knockout

WIN

55,212-2, (R)-(+)-[2,3-dihydro-5-methyl-3- (4-morpholinylmethyl)pyrrolo[1,2,3-de]-1,4-benzoxan-6-yl]-1-naphthalenylmethanone mesylate

WT

wild type