Pharmakologisches Institut, Universität Freiburg, Hermann-Herder-Strasse 5, D-79104 Freiburg i. Br., Germany. E-mail: email@example.com
Release-modulating opioid and cannabinoid (CB) receptors, β-adrenoceptors and bradykinin receptors at noradrenergic axons were studied in mouse tissues (occipito-parietal cortex, heart atria, vas deferens and spleen) preincubated with 3H-noradrenaline.
Experiments using the OP1 receptor-selective agonists DPDPE and DSLET, the OP2-selective agonists U50488H and U69593, the OP3-selective agonist DAMGO, the ORL1 receptor-selective agonist nociceptin, and a number of selective antagonists showed that the noradrenergic axons innervating the occipito-parietal cortex possess release-inhibiting OP3 and ORL1 receptors, those innervating atria OP1, ORL1 and possibly OP3 receptors, and those innervating the vas deferens all four opioid receptor types.
Experiments using the non-selective CB agonists WIN 55,212-2 and CP 55,940 and the CB1-selective antagonist SR 141716A indicated that the noradrenergic axons of the vas deferens possess release-inhibiting CB1 receptors. Presynaptic CB receptors were not found in the occipito-parietal cortex, in atria or in the spleen.
Experiments using the non-selective β-adrenoceptor agonist isoprenaline and the β2-selective agonist salbutamol, as well as subtype-selective antagonists, demonstrated the occurrence of release-enhancing β2-adrenoceptors at the sympathetic axons of atria and the spleen, but demonstrated their absence in the occipito-parietal cortex and the vas deferens.
Experiments with bradykinin and the B2-selective antagonist Hoe 140 showed the operation of release-enhancing B2 receptors at the sympathetic axons of atria, the vas deferens and the spleen, but showed their absence in the occipito-parietal cortex.
The experiments document a number of new presynaptic receptor locations. They confirm and extend the existence of marked tissue and species differences in presynaptic receptors at noradrenergic neurons.
Neurotransmitter release is subject to modulation by presynaptic receptors. Central and peripheral noradrenergic neurons, for example, possess release-inhibiting α2-autoreceptors and many other presynaptic receptors which either induce, enhance or inhibit transmitter release (for review see Starke, 1977; Langer, 1981; Fuder & Muscholl, 1995). Presynaptic receptors have been extensively studied in many human and animal tissues. Molecular genetic techniques have now moved the mouse species into the centre of interest: transgenic animals are almost exclusively mice. However, in contrast to other species, little information is available about presynaptic receptors in mice.
In a first attempt to enlarge our knowledge about receptors on noradrenergic neurons in the mouse, we recently investigated presynaptic receptors for angiotensin in several tissues (Cox et al., 1999). The aim of the present study was to search for, and if detected subclassify, two release-inhibiting receptors, namely opioid and cannabinoid receptors, and two other release-enhancing receptors, namely β-adrenoceptors and bradykinin receptors, on the noradrenergic axons of the mouse occipito-parietal cortex, atria, vas deferens and spleen.
Tissues and superfusion
Male NMRI mice weighing 35–45 g were killed by cervical dislocation. Either six to seven slices of the occipito-parietal cortex (Limberger et al., 1995), six to eight pieces of the atria (Wahl et al., 1996), eight to 12 pieces of the vas deferens (Trendelenburg et al., 1999), or 12 to 16 pieces of the spleen (Cox et al., 1999) were obtained from one animal. The tissue pieces were preincubated in 2 ml medium containing 0.2 μM3H-noradrenaline for 30 min at 37°C. One tissue piece was then placed in each of 12 superfusion chambers between platinum electrodes, where it was superfused with 3H-noradrenaline-free medium at a rate of 1.2 ml min−1. 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 the experiments, tissues were dissolved and tritium was determined in superfusate samples and tissues.
The superfusion medium contained (mM): NaCl 118, KCl 4.8, CaCl2 1.3 (brain slices) or 2.5 (peripheral tissues), 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, for peripheral tissues, only 0.2 mM CaCl2.
There were seven periods of electrical stimulation. Each stimulation consisted of rectangular pulses of 1 ms width and 35 V cm−1 (brain slices) or 47 V cm−1 (peripheral tissues) voltage drop between the electrodes of each chamber, yielding a current strength of 60 and 80 mA, respectively. The first stimulation period was delivered at t=30 min and was not used for determination of tritium overflow. The subsequent stimulation periods (S1 to S6) were applied at t=54, 72, 90, 108, 126 and 144 min and differed, depending on the tissue and type of experiment, as indicated in the Results section.
Agonist concentration-response curves were obtained by introducing the agonist at increasing concentrations after S1, 12 min before S2, S3, S4, S5 and S6. Antagonists were present throughout superfusion at a fixed concentration.
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 time of stimulation (see Trendelenburg et al., 1997). For further evaluation, overflow ratios (Sn/S1) were calculated. Overflow ratios obtained in the presence of agonist were also calculated as a percentage of the corresponding ratio in controls in which no agonist was added after S1. Effects of agonists on basal tritium outflow were evaluated similarly (Trendelenburg et al., 1997).
Concentration-response data for agonists given alone were evaluated by sigmoid curve fitting (eq. 25 of Waud, 1976). This yielded the Emax (maximal effect) of the agonist and its EC50 (concentration causing a half-maximal effect) in the absence of antagonist. The EC50 values of agonists in the presence of antagonists were interpolated from the nearest points of the respective concentration-response curves, assuming that the Emax of the agonist had not changed. When only one or two concentrations of an antagonist were tested against an agonist, the negative logarithm of the apparent Kd value of the antagonist, i.e. the apparent pKd, was calculated from the EC50 increase. When three concentrations of an antagonist were tested against an agonist (experiments on β-adrenoceptors in atria), results were evaluated as described by Arunlakshana & Schild (1959).
Unless stated otherwise, results are expressed as arithmetic means±s.e. mean or, in the case of EC50 or Emax values, the standard errors as defined by Waud (1976). Groups were tested for significant differences by the Mann–Whitney test with Bonferroni correction. P<0.05 was taken as limit of statistical difference. n represents the number of tissue pieces.
Drugs were (−)-[ring-2,5,6- 3H]-noradrenaline, specific activity 46.8–62.3 Ci mmol−1 (DuPont, Dreieich, Germany); bradykinin, bremazocine HCl, [D-Ala2, N-Me-Phe4, Gly5-ol]-enkephalin (DAMGO), desipramine HCl, [D-Pen2,5]-enkephalin (DPDPE), [D-Ser2]-Leu-enkephalin-Thr6 (DSLET), (±)-1-[2,3- (dihydro-7-methyl-1H-inden-4-yl) oxy]-3-[(1-methylethyl)amino]-2-butanol HCl (ICI 118,551), (−)-isoprenaline (+)-bitartrate, naltriben methanesulphonate, naltrindole HCl, salbutamol hemisulphate, trans-(±)-3,4-dichloro-N-methyl-N-(2-[1-pyrrolidinyl]cyclohexyl)-benzeneacetamide methanesulphonate (U50488H), (5α,7α,8β) - (+) -N-methyl-N-[7-[1-pyrrolidinyl] −1- oxaspiro [4,5] dec-8-yl) - benzeneacetamide (U69593; Sigma, Deisenhofen, Germany); 7-benzylidenenaltrexone maleate (BNTX), (−)-cis-3-[2-hydroxy-4-(1,1-dimethylheptyl)phenyl]-trans-4- (3-hydroxypropyl) cyclohexanol (CP 55,940), nociceptin, [Phe1Ψ(CH2-NH)Gly2]nociceptin(1-13)NH2 (PheΨ), (R)-(+)-[2,3-dihydro-5-methyl-3-(4-morpholinylmethyl) pyrrolo [1,2,3-de]-1,4-benzoxazin-6-yl]-1-naphthalenylmethanone (WIN 55, 212-2; Tocris, U.K.). The following drugs were kindly provided by the producer: D-Arg[Hyp3, Thi5, D-Tic7,Oic8]bradykinin (Hoe 140; Hoechst, Frankfurt am Main, Germany); (±)-2-hydroxy-5-[2-[[2-hydroxy-3-[4-[1-methyl −4- (trifluoromethyl) - 1H-imidazol-2-yl] phenoxy]propyl] amino] ethoxy] -benzamide methanesulphonate (CGP 20712A), phentolamine methanesulphonate (Ciba-Geigy, Basel, Switzerland); naloxone HCl (Gödecke, Freiburg, Germany) and N-piperidino-5- (4-chlorophenyl)-1- (2,4-dichlorophenyl) −4-methyl-3-pyrazole-carboxamide (SR 141716A; Sanofi, Montpellier, France). Drugs were dissolved in distilled water except naltriben (10 mM HCl) and CP 55,940, SR 141716A and WIN 55,212-2 (dimethylsulphoxide).
Common features of the experiments will be summarized here. The operation of presynaptic opioid receptors, cannabinoid receptors and β-adrenoceptors is blunted whenever there is α2-autoinhibition. Conversely, bradykinin requires ongoing α2-autoinhibition for a major release-enhancing effect (see Schlicker & Göthert, 1998; Cox et al., 2000). For this reason, the former three receptors were studied with stimulation conditions leading to little autoinhibition, whereas bradykinin was studied under autoinhibition-rich conditions. Marked autoinhibition was created with 36 pulses (brain slices) or 120 pulses at 3 Hz (peripheral tissues) per stimulation period, conditions under which the α-adrenoceptor antagonist phentolamine (1 μM), when present throughout superfusion, greatly increased the overflow evoked by S1 (Table 1). Minor autoinhibition was obtained either with single pulses (brain slices) or with brief trains of 20 pulses at 50 Hz (peripheral tissues) per stimulation period (see Singer, 1988; Limberger et al., 1995), conditions under which phentolamine (1 μM) increased S1 only by maximally 70% (Table 1). Some tissue preparations were also superfused throughout the experiment with medium containing phentolamine (1 μM) which, of course, assured freedom from α2-autoinhibition.
Table 1. Overflow of tritium (S1) elicted by different pulse patterns in the absence and presence of phentolamine (1 μM)
Apart from phentolamine (Table 1), most of the other antagonists, when added throughout superfusion, did not per se change the evoked overflow of tritium (S1). The greatest change was a 56% increase by 0.1 μM naltriben. In control experiments in which no drug was added after S1, the overflow remained similar from S1 to S6 (n=5–65).
The basal outflow of tritium (for typical values see Cox et al., 1999) was not changed by the antagonists or agonists, with very few exceptions. The greatest change was a 50% increase by 1 μM of the cannabinoid (CB) agonist WIN 55,212-2.
Presynaptic opioid receptors
Presynaptic opioid receptors were investigated in the occipito-parietal cortex, in atria and in the vas deferens. Noradrenaline release was elicited either by single pulses (occipito-parietal cortex) or by 20 pulses at 50 Hz (peripheral tissues), conditions with little α2-autoinhibition (Table 1). The following opioid receptor ligands were used: the OP1 (δ) -selective agonists DPDPE and DSLET, the OP2 (κ) -selective agonists U50488H and U69593, the OP3 (μ) -selective agonist DAMGO, the ORL1-selective agonist nociceptin, the OP1-selective antagonists naltrindole, naltriben (δ1) and BNTX (δ2), the OP2-selective antagonist bremazocine, the slightly OP3-selective antagonist naloxone and the selective ORL1 antagonist PheΨ.
In the occipito-parietal cortex, DAMGO caused concentration-dependent inhibition of the evoked overflow of tritium, whereas the OP1-selective agonists DPDPE and DSLET and the OP2-selective agonist U50488H were inactive (Figure 1A). The EC50 and Emax values are shown in Table 2. The antagonists at classical opioid receptors naltrindole, bremazocine and naloxone all shifted the concentration-response curve of DAMGO to the right (Figure 1B). Apparent pKd values are in Table 3. Bremazocine was tested at two concentrations, 0.001 and 0.01 μM; the identical apparent pKd values are compatible with a competitive mode of interaction (Figure 1B, Table 3). Naltrindole, bremazocine and naloxone were approximately equipotent against DAMGO (Table 3).
Table 2. EC50 and Emax values of agonists at presynaptic opioid receptors
Table 3. Apparent pKd values of antagonists at presynaptic opioid receptors
The ORL1-selective agonist nociceptin was studied in the presence of 1 μM naloxone, a concentration blocking all classical opioid receptors but devoid of a noticeable effect at ORL1 receptors (Schlicker et al., 1998). Like DAMGO, nociceptin caused concentration-dependent and marked inhibition in brain cortex slices (Figure 2A, Table 2). Still in the presence of naloxone, the ORL1-selective antagonist PheΨ shifted the concentration-response curve of nociceptin to the right (Figure 2A, Table 3).
A different pattern of effects was found in atria. U50488H again was inactive, but DAMGO was a weaker agonist than in the brain cortex, and DPDPE and DSLET potently reduced the evoked overflow of tritium (Figure 3A, Table 2). Naloxone (0.1 μM) abolished any effect of DAMGO (not shown); although the interaction did not permit an exact calculation, the pKd value of naloxone against DAMGO clearly was >8 (Table 3). The concentration-inhibition curves of DPDPE and DSLET were shifted to the right by all antagonists at classical opioid receptors, naltriben, naltrindole, BNTX, bremazocine and naloxone (curves not shown; apparent pKd values in Table 3). The two concentrations of bremazocine, 0.01 and 0.1 μM, yielded similar apparent pKd values, in accord with a competitive antagonism. The potencies of each antagonist against DPDPE and DSLET were close to each other, and in contrast to the brain cortex, potencies now differed greatly between antagonists, with an order naltriben, naltrindole>BNTX, bremazocine, naloxone (Table 3).
Nociceptin, in the presence of naloxone, had a lower maximal effect than in the brain cortex (Figure 2B). PheΨ antagonized its effect (Figure 2B), but due to the small agonist Emax the apparent pKd value is a little in doubt (Table 3).
Finally, the pattern differed again in the vas deferens. All agonists, including the selective OP2 agonists U50488H and U69593 and (in the presence of naloxone) the ORL1 agonist nociceptin, caused considerable inhibition (Figures 2C and 3B, Table 2). Wherever tested, antagonists shifted agonist concentration-inhibition curves to the right (for example Figure 2C). Where two antagonist concentrations were examined (naltrindole against U50488H and DAMGO; bremazocine against DSLET and DAMGO), the shifts were compatible with a competitive interaction (Table 3). As in atria, the potencies of each antagonist against DPDPE and DSLET were close to one another. Apart from this, however, the potency orders of the antagonists depended greatly on the agonist; the antagonist potency order was naltriben, naltrindole>BNTX, bremazocine, naloxone against DPDPE and DSLET, as in atria; the order was bremazocine>naltrindole, naloxone against U50488H; and all antagonists were approximately equipotent against DAMGO, as in the occipito-parietal cortex (Table 3).
Presynaptic cannabinoid receptors
Presynaptic CB receptors were investigated in the occipito-parietal cortex, atria, vas deferens and spleen. Noradrenaline release was elicited by either single pulses (brain cortex), or 20 pulses at 50 Hz (atria and vas deferens), or 120 pulses at 3 Hz in the presence of phentolamine (1 μM; spleen), again conditions with no or little α2-autoinhibition (Table 1). In the spleen, the brief high-frequency pulse trains used in atria and the vas deferens to avoid autoinhibition produced too small tritium overflow peaks – hence the 120 pulses at 3 Hz in the presence of phentolamine. The CB receptor ligands tested were the agonists WIN 55,212-2 and CP 55,940 and the CB1-selective antagonist SR 141716A.
In occipito-parietal cortex, atria and spleen WIN 55,212-2 (0.1 nM–1 μM) did not change the evoked overflow of tritium (data not shown; n=6–10).
In the vas deferens, WIN 55,212-2 (Figure 4) and CP 55,940 (not shown) produced concentration-dependent inhibition. The EC50 of WIN 55,212-2 was 2.7±3.0 nM and its Emax 79±11% inhibition; the EC50 of CP 55,940 was 0.27±0.13 nM and its Emax 64±5% inhibition (n=10–12). SR 141716A was examined against WIN 55,212-2 at concentrations of 0.001 and 0.01 μM and against CP 55,940 at concentrations of 0.01 and 0.1 μM. It shifted the concentration-inhibition curves of both agonists to the right (effect against WIN 55,212-2 in Figure 4). The four apparent pKd values obtained were in the range of 9.7 and 10.5, i.e. similar, in accord with a competitive antagonism (n=6–13).
β-Adrenoceptors were investigated in the occipito-parietal cortex, atria, vas deferens and spleen. As in the CB receptor study, noradrenaline release was elicited by single pulses (occipito-parietal cortex), 20 pulses at 50 Hz (atria, vas deferens) or 120 pulses at 3 Hz (spleen). In contrast to the CB receptor part, all experiments were carried out in the presence of phentolamine (1 μM); this was done in order to prevent any activation by β-adrenoceptor agonists of presynaptic α2-adrenoceptors, a known effect at least of isoprenaline (Endo et al., 1977). The presence of phentolamine of course assured autoinhibition-free release. For comparison, 120 pulses at 3 Hz were also applied to a series of atria (in the presence of phentolamine). The drugs tested were isoprenaline, the β2-selective agonist salbutamol, the selective β1 antagonist CGP 20712A and the β2-selective antagonist ICI 118,551.
In the occipito-parietal cortex and vas deferens neither isoprenaline (0.1 nM–1 μM; n=5–7) nor salbutamol (0.1 nM–10 μM; n=2–3) had any effect on the evoked overflow of tritium (data not shown).
In atria both isoprenaline (not shown) and salbutamol (Figure 5A) increased the overflow of tritium elicited by 20 pulses at 50 Hz. EC50 and Emax values are given in Table 4. The β1-selective antagonist CGP 20712A (0.1 μM) did not alter the concentration-response curves of either agonist whereas ICI 118,551 (0.01 μM) produced a shift to the right (effects against salbutamol in Figure 5A). The apparent pKd values against isoprenaline and salbutamol agreed well (Table 5). Similar concentration-response curves for isoprenaline and salbutamol were obtained when atria were stimulated by 120 pulses at 3 Hz; both EC50 and Emax values were somewhat lower than those found with 20 pulses at 50 Hz (Table 4). In experiments with 120 pulses at 3 Hz, ICI 118,551 was tested against salbutamol at three concentrations (0.001, 0.01 and 0.1 μM), and results were subjected to a Schild analysis: the slope of the Schild plot was 0.95 (95% confidence limits, 0.88–1.02), indicating a competitive antagonism, and the pA2 value was 10.5 (10.3–10.7), identical with the pKd value against salbutamol found with 20 pulses at 50 Hz (Table 5).
Table 4. EC50 and Emax values of agonists at presynaptic β-adrenoceptors
Table 5. Apparent pKd values of antagonists at presynaptic β-adrenoceptors
Results in the spleen were similar. Both isoprenaline (not shown) and salbutamol (Figure 5B) increased the overflow of tritium elicited by 120 pulses at 3 Hz. The EC50 and Emax values were lower than in atria stimulated by 20 pulses at 50 Hz or by 120 pulses at 3 Hz (Table 4). As in atria, CGP 20712A (0.1 μM) did not change the agonist concentration-response curves whereas ICI 118,551 (0.01 μM) caused a shift to the right (effect against salbutamol in Figure 5B); pKd values were virtually identical with those found in atria (Table 5).
Presynaptic bradykinin receptors
These were once again studied in the occipito-parietal cortex, in atria, the vas deferens and the spleen. Noradrenaline release was triggered by either 36 (brain slices) or 120 pulses at 3 Hz (peripheral tissues) in the absence of phentolamine, i.e. under autoinhibition-rich conditions (Table 1). The drugs used were bradykinin and the B2 receptor antagonist Hoe 140.
Bradykinin (0.01–100 nM) did not significantly alter the stimulation-evoked overflow of tritium in the occipito-parietal cortex (data not shown; n=4–8). In atria, vas deferens and spleen bradykinin caused increases (Figure 6A–C). It was distinctly more potent and had distinctly greater maximal effects in the atria and the vas deferens than in the spleen: the EC50 values in atria, vas deferens and spleen were 0.05±0.01, 0.05±0.01 and 0.3±0.2 nM, respectively; Emax values were 101±2, 87±3 and 55±8% increases, respectively (n=5–33). The concentration-response curves of bradykinin were shifted to the right by Hoe 140, but to a greater extent in atria and the vas deferens than in the spleen (Figure 6A–C). The apparent pKd values were 11.4, 12.2 and 10.3 in atria, vas deferens and spleen, respectively.
We made two attempts to ensure that our experimental conditions were adequate for the detection of presynaptic receptors and that the pKd values were valid estimates of antagonist affinity. First, α2-autoinhibition was minimized whenever opioid receptors, CB receptors and β-adrenoceptors were examined, which loose much of their release-modulating power when autoinhibition operates (see Schlicker & Göthert 1998); conversely, strong α2-autoinhibition was created when bradykinin receptors were examined, which require autoinhibition for a major effect (Cox et al., 2000). Disregard of these conditions may lead to false negative results (Ramme et al., 1986; Cox et al., 2000). Second, although most pKd values were obtained with a single antagonist concentration and hence were apparent pKd values, we used two or three antagonist concentrations in several cases: almost identical apparent pKd values (two antagonist concentrations), or a slope close to unity of the Schild plot (three antagonist concentrations), always confirmed a competitive kinetic and, hence, the validity of the pKd as an affinity measure.
Presynaptic opioid receptors
Our results show that the noradrenergic axons of the occipito-parietal cortex of the mouse possess OP3 (μ) and ORL1 receptors, the sympathetic axons of atria OP1 (δ) and ORL1 receptors (OP3 receptors are questionable), and those of the vas deferens OP1, OP2 (κ), OP3 and ORL1 receptors. The evidence comes from findings with both agonists and antagonists.
As to agonists, only the OP3-selective DAMGO and the ORL1 agonist nociceptin inhibited the release of noradrenaline in the occipito-parietal cortex (Figures 1A and 2A), whereas in atria the OP1 agonists DPDPE and DSLET (Figures 2B and 3A), and in the vas deferens DPDPE, DSLET, and the OP2 agonists U50488H and U69593 (Figures 2C and 3B) also caused presynaptic inhibition. In atria, the EC50 of DAMGO was five to 15 times higher than in the two other tissues (Table 2); since DAMGO also binds to OP1 receptors at high concentrations (Emmerson et al., 1994), an action at the atrial presynaptic OP1 receptors cannot be excluded.
As to antagonists, the pKd values (Table 3) show that the tissues contained all three classical opioid receptors, because the orders of potency of the antagonists against, first, DPDPE and DSLET, second U50488H, and third DAMGO, differed distinctly. A second revealing observation refers to the vas deferens. In that tissue, the OP1-selective antagonist naltrindole acted with the highest potency against DPDPE and DSLET and with much lower potency against U50488H and DAMGO, confirming that the receptor for the former two peptides was OP1 (the high potency of naltrindole against DPDPE and DSLET recurred in atria); the OP2-selective antagonist bremazocine acted with the highest potency against U50488H, confirming that the receptor for U50488H was OP2; and the OP3-selective antagonist naloxone acted with the highest potency against DAMGO, confirming that the receptor for DAMGO was OP3. Final support for the receptor diagnoses comes from a comparison with published affinity data: the present pKd values agree excellently with those obtained, for example, by binding studies at monkey brain OP1, OP2 and OP3 receptors (Emmerson et al., 1994). Similarly, the pKd values of PheΨ against nociceptin in the occipito-parietal cortex (7.8) and vas deferens (7.4) agree reasonably with published values for ORL1 receptors (6.8–7.2; Guerrini et al., 1998; Schlicker et al., 1998). Two uncertain antagonist results in atria remain to be mentioned. Due to the only weak inhibition by DAMGO, a pKd value for naloxone against DAMGO could not be determined; the estimate of >8 (Table 3) is compatible with an OP3 receptor but, like the effect of DAMGO itself (see above), does not rule out an action at the atrial presynaptic OP1-receptors. The pKd value of PheΨ against nociceptin (8.8) was higher in atria than in the two other tissues (Table 3), possibly an overestimation due to the only slight inhibition by nociceptin (Figure 2B).
It has been suggested that each of the three classical opioid receptors comprises several subtypes (see Dhawan et al., 1996). Our results permit speculation on the potential OP1 subtypes (δ1 and δ2) and OP2 subtypes (κ1, κ2 and κ3) involved in presynaptic inhibition in atria and the vas deferens. In either tissue, the δ2-selective antagonist naltriben was much more potent than the δ1-selective antagonist BNTX against DPDPE and DSLET (see Dhawan et al., 1996), indicating that the OP1 receptor was δ2 (see also Wild et al., 1993). In the vas deferens, not only U50488H but also U69593 inhibited the release of noradrenaline, indicating that the OP2 receptor was κ1, the only subtype sensitive to U69593 (see Dhawan et al., 1996). The higher antagonist potency of both naltriben and naltrindole in atria than in the vas deferens (Table 3) remains unexplained.
Prior to our study there has been little work on presynaptic opioid receptors at noradrenergic neurons of mice. All four receptors, it is true, have been detected in the vas deferens, although mostly by measurement of contractions and rarely by determination of transmitter overflow (see Introduction and the review by Illes, 1989). However, presynaptic opioid receptors have never been shown in the heart, and only ORL1 receptors (Schlicker et al., 1998) but none of the classical receptors have been demonstrated at cerebral noradrenergic axons.
Presynaptic cannabinoid receptors
Presynaptic cannabinoid receptors were found only in the vas deferens, where the agonists WIN 55,212-2 and CP 55,940 decreased the release of noradrenaline (Figure 4). The effects were attenuated by the CB1-selective antagonist SR 141716A, with apparent pKd values (9.7–10.5) slightly higher than those reported in the literature both for the mouse vas deferens and for other tissues and species (7.7–9.2; Rinaldi-Carmona et al., 1994; Pertwee et al., 1995; 1996; Schlicker et al., 1997). The noradrenergic axons innervating the occipito-parietal cortex, the atria and the spleen lacked CB receptors.
Presynaptic inhibition through CB1 receptors, although a known mechanism in the mouse vas deferens, has not been shown in that tissue previously by measurement of noradrenaline overflow. Data concerning the occurrence of presynaptic CB receptors at noradrenergic axons in other mouse tissues have not been reported except for a study, with negative outcome, in the hippocampus (Schlicker et al., 1997).
Presynaptic β-adrenoceptors were found in atria and the spleen but not in the occipito-parietal cortex or the vas deferens: isoprenaline and salbutamol increased the release of noradrenaline in the former but did not change it in the latter tissues (Figure 5, Table 4). Analogous experiments showed that the noradrenergic axons innervating the mouse hippocampus also lacked presynaptic β-adrenoceptors (Trendelenburg et al., unpublished observation).
The receptors were β2 as shown, first, by the effect of salbutamol, and second by results obtained with antagonists: only the β2-selective ICI 118,551 but not the β1-selective CGP 20712A antagonized the effects of the two agonists. In both tissues, and determined with two stimulation patterns (20 pulses at 50 Hz and 120 pulses at 3 Hz) and up to three antagonist concentrations, our pKd values of ICI 118,551 (10.3–10.7) are consistently somewhat higher than literature values for other species (7.8–9.3; Bilski et al., 1983; Molderings et al., 1988).
The presynaptic β-adrenoceptors in mouse atria are well known (e.g. Johnston & Majewski, 1986) but their subtype has not been determined previously. Attempts to find presynaptic β-adrenoceptors in other mouse tissues have not been reported.
Presynaptic bradykinin receptors
Presynaptic bradykinin receptors were found in the three peripheral tissues but not in the occipito-parietal cortex: bradykinin enhanced the release of noradrenaline only in atria, vas deferens and the spleen (Figure 6). Bradykinin also failed to change the release of noradrenaline in segments of the mouse hippocampus and hypothalamus (Cox et al., unpublished observation). Antagonism by the B2-selective antagonist Hoe 140 indicated that the presynaptic bradykinin receptors were B2. The pKd values of Hoe 140 (10.3–12.2) are in the range of or slightly higher than literature values (8.4–10.9; Hock et al., 1991; Falcone et al., 1993). We do not know the reason for the large pKd differences between atria (11.4), the vas deferens (12.2) and the spleen (10.3).
Tissue and species differences have been noticed ever since receptors modulating the release of noradrenaline became known. The present study adds or confirms some clear-cut examples, of which the examples for species differences will be discussed.
In rabbits, the major presynaptic opioid receptor at noradrenergic neurons is OP2; presynaptic OP1 receptors also occur, but presynaptic OP3 receptors have not been found. In rats, in contrast, OP3 receptors prevail at noradrenergic axons, with a minority of OP1 but no OP2 receptors (Illes, 1989). The mouse seems to resemble the rat in that presynaptic OP3 receptors predominate (even though their presence in atria is questionable). Contrary to both rabbit and rat, however, all three classical opioid receptors appear to be used as presynaptic receptors at mouse noradrenergic neurons: OP1 in atria and the vas deferens, OP2 in the vas deferens, and OP3 in the brain cortex, the vas deferens and possibly atria. The recently described ORL1 receptors differs from the classical opioid receptors in that it inhibits noradrenaline release in each species and tissue examined (see Henderson & McKnight, 1997; Schlicker et al., 1998).
A known species difference in presynatic CB1 receptors is their occurrence at the noradrenergic axons of human and guinea-pig but not rat and mouse hippocampus (Schlicker et al., 1997). Our experiments show an additional difference: CB1 receptors modulate the release of noradrenaline in rat atria (Ishac et al., 1996) but not in mouse atria (this study).
Presynaptic, release-enhancing β2-adrenoceptors have been detected in many but not all peripheral tissues; with few exceptions, they were found to be absent from brain tissues (see Starke, 1977; Langer, 1981; Majewski, 1983). A noticeable species difference exists between the mouse vas deferens, where we did not find the receptors, and the guinea-pig vas deferens, where they operate (see Driessen et al., 1996). It should be noted that the failure to detect the receptors can hardly be due to unsuitable experimental conditions: as mentioned above, the conditions of our experiments were optimal for presynaptic receptor operation.
A particularly remarkable species difference concerns the effect of bradykinin on cardiac sympathetic nerves. Bradykinin, acting on B2 receptors, enhances the release of noradrenaline in human, rat, mouse and guinea-pig heart preparations (Chulak et al., 1998; Vaz-da-Silva et al., 1996; Rump et al., 1997; Seyedi et al., 1997; Cox et al., 2000; and this study). In contrast, bradykinin reduced the release of noradrenaline in rabbit hearts, perhaps because in rabbit hearts stimulation of prostaglandin synthesis by bradykinin and subsequent inhibition of noradrenaline release by prostaglandins outweighs the direct presynaptic facilitatory effect (Starke et al., 1977; Chulak et al., 1998).
Knowledge on presynaptic receptors in mice is a prerequisite to understand changes in neuronal functions in transgenic mice. For example, now that mice lacking α2A-adrenoceptors (Altman et al., 1999) or both α2A- and α2C-adrenoceptors (Hein et al., 1999) are available, it will be interesting to see whether the loss of presynaptic α2-autoreceptors leads to compensatory changes in other presynaptic receptors.