Pharmac. Ther.Vol. 54, pp. 231-248, 1992 Printed in Great Britain. All rights reserved

0163-7258/92$15.00 © 1992 Pergamon Press Ltd

Associate Editor: M. J, LEwis

THE ROLE OF IMIDAZOLINE RECEPTORS IN BLOOD PRESSURE REGULATION CARLENE A. HAMILTON Department of Medicine and Therapeutics, Western Infirmary, Glasgow, G I I 6NT, U.K. Abstract--Using the ligands [3H] clonidine and [3H] idazoxan, nonadrenergic imidazoline preferring binding sites have been identified in a range of tissues from several species including man. These sites may represent a new family of receptors. An endogenous ligand and potential clonidine displacing substance has been identified. There is strong evidence for an involvement of the nonadrenergic imidazoline [3H] clonidine labelled sites in the nucleus reticularis lateralis in blood pressure regulation, and some evidence for a role in sodium regulation in the kidney for the [3HI idazoxan labelled sites. Some drugs which were previously thought to act via ~t2-adrenoceptors, may mediate their effects in part via these imidazoline sites.

CONTENTS 1. Background 2. Identification of Imidazoline Preferring Sites 2.1. Sites labelled by [3H] clonidine and analogs 2.2. Sites labelled by [3H] idazoxan 3. Characterization of Imidazoline Preferring Sites 4. Physiological Responses Linked to the Imidazoline Sites 4.1. Blood pressure and heart rate regulation 4.2. Regulation of Na+/H ÷ exchange 4.3. Modulation of noradrenaline release from postganglionic sympathetic nerve endings 4.4. Hormonal regulation 4.4.1. Prolactin secretion 4.4.2. Insulin release 4.5. Effects on gastric ulceration 4.6. Effects related to cerebral ischemia 5. Endogenous Ligands for the Imidazoline Sites--Clonidine Displacing Substance 6. Imidazoline Preferring Drugs and Blood Pressure Regulation 6.1. Central effects 6.2. Renal effects 7. Summary and Conclusions References

231 232 232 234 236 237 237 239 240 240 240 240 240 241 242 243 243 244 244 245

1. B A C K G R O U N D In the 1970s and 1980s n u m e r o u s differences in biological responses between ch-adrenoceptor agonists o f the phenethylamine and imidazoline classes were reported (Ruffolo et al., 1976, 1977, 1983; Bousquet et al., 1984). It was suggested that agonists with imidazoline-like structures acted at a different site on the ct-adrenoceptor to their phenethylamine counterparts (Ruffolo et al., 1977). Note on nomenclature--These nonadrenergic imidazoline preferring sites have been referred to by various names by the different groups working in the field. These include imidazole, imidazoline and imidazoline-guanidinium receptive sites, imidazole, imidazoline and idazoxan or I receptors. For consistency, one term--the imidazoline preferring site has been used throughout in this review. 231

232

C.A. HAMILTON

CI NH

O-CH2 H ClRAZOLINE

~N--N~

CLONIDINE

r N (i..~N

03 ""'-

H UK 14304

IDAZOXAN

cI V

O

NH II " CH "- N ~ N H --C - - NH2

N

A

RILMENIDINE

GUANABENZ

CH2--CH-- CH2"--N."'~1

N

BHT 920

FIG. 1. Some compounds with imidazoline or closely related structures. In 1986 using [3H] p-aminoclonidine (a partial ~2-adrenoceptor agonist) Meeley and colleagues identified nonadrenergic 'imidazoline preferring' sites in bovine brain. Numerous other reports of ct2-adrenergic drugs binding at nonadrenergic imidazoline preferring sites in a range of tissues and species followed and the concept evolved that rather than acting at a separate site on the ~t2-adrenoceptor, these compounds could act at a distinct class of receptor--the imidazoline receptor to produce physiological effects. A number of compounds reported to have affinity for these sites are shown in Fig. 1.

2. IDENTIFICATION OF IMIDAZOLINE PREFERRING SITES 2.1. SITESLABELLEDBY [3H] CLONIDINEAND ANALOGS Nonadrenergic binding sites were first identified using the ligand [3H]p-aminoclonidine in bovine ventrolateral medulla by Meeley and coworkers (1986). They showed that noradrenaline only

Imidazoline receptors and blood pressure

233

1:I i

40

20

-20 -10

-9

-8

-7

-6

-5

-4

Log[] FIG. 2. Inhibition of [3H] p oaminoclonidine binding to distinct subpopulations of sites in ventrolateralla medulla membranes. 0 , Noradrenaline (n = 7); C), clonidine (n = 3). Reprinted from Ernsberger et al. (1987), with permission of the authors and the copyright holder, Elsevier Science Publishers BV, Amsterdam. displaced 70-75% o f the specifically bound ligand although clonidine and phentolamine displaced all specifically bound [3H] p-aminoclonidine (Fig. 2). Further characterization of these nonadrenergic sites demonstrated a relatively high affinity for imidazoles and imidazolines, but a lack of affinity for phenylethylamines such as noradrenaline and alkaloids such as yohimbine (Ernsberger et al., 1987). Nonadrenergic imidazoline preferring sites have also been identified in human brain stem (Bricca et al., 1988), rat brain (Kamisaki et al,, 1990) and kidney (Ernsberger et al., 1988a)

TABLE1. Drug Affinities at Imidazoline Sites Labelled by [SH] Clonidine and [SH] p-Aminoclonidine

Species Cow Rat Rabbit Cow Rabbit Man Rabbit

Tissue Ventrolateral medulla Striatum Kidney Forebrain Ventrolateral medulla Kidney Forebrain Nucleus reticularis lateralis Forebrain Kidney

Ligand

Displacing

PAC

Cimetidine

480 + 88t

PAC CI CI PAC

Cimetidine Cimetidine Cimetidine Guanabenz

6100"~ > 100,000~ > 100,000~ > 1,000,000" II

CI CI C1

Guanabenz Guanabenz Cirazoline

1.5 __+0.2§ 1.3 _+ 1.3 20¶

CI C1

Cirazoline Cirazoline

0.6 _+0.1§ 0.8 + 0.1§

*Results expressed as Ki rather than ICs0. ~'Ernsberger et al. (1987). ~Kamisaki et al. (1990). §Hamilton et al. (1991). [[Ernsberger et al. 0990). ¶Bricca et al. 0989). PAC, p-aminoclonidine; C1, clonidine.

Ics0 (nM)

234

C.A. HAMILTON

loo!

,.,=

0

10

9

8

7

6

5

4

- LOG [C_,OMPETITOFI(M)I FIG. 3. Inhibition of [3H] RX 781094 (idazoxan) binding to basolateral membranes from rabbit kidney. O, Tramazoline; O, tolazoline; O, ( - ) adrenaline; ~, rauwolscine; A, serotonin; A, phenylephrine; n = 3. Reprinted from Coupry et al. (1987), with permission of the authors and the copyright holder, Academic Press, Orlando. and rabbit brain and kidney (Hamilton et al., 1991) using either [3H] p-aminoclonidine or [31-I] clonidine. However, differences between the sites identified by the various groups emerge on comparison of the data (Table 1). In rat kidney nonadrenergic [3H] p-aminoclonidine binding represented 29% of the total specific sites and displayed a high affinity for a number of imidazolines and imidazoles including cimetidine (Ernsberger et al., 1988b). In contrast, in rabbit kidney using [3H] clonidine as the ligand, Hamilton et al. (1991) reported that the majority of sites labelled were nonadrenergic and that the affinity of cimetidine for the site was low. Similar discrepancies emerge when comparing nonadrenergic binding sites in the brain using [3H] clonidine and p-aminoclonidine. Bricca and colleagues (1989) were unable to identify any nonadrenergic binding in rat medulla oblongata using [3H] clonidine as the ligand, whereas Kamisaki et al. (1990) found that approximately one third of the sites in rat medulla oblongata labelled by [3HI p-aminoclonidine were nonadrenergic imidazoline preferring sites. Moreover, the group reported a Ki for cimetidine at this site of 6.11 p M, whereas using the same ligand an IC50 of 480 nu was observed in bovine ventrolateral medulla (Ernsberger et al., 1987). Whether these differences can be related to the ligand ([3H] clonidine vs [3H] p-aminoclonidine), the species or technical differences in assays is not clear. However, the possibility of multiple binding sites cannot be ruled out.

2.2. SITES LABELLEDBY [3H] IDAZOXAN Idazoxan is an ~2-imidazoline type adrenoceptor antagonist. It has been reported to have a higher potency and specificity for 0tE-adrenoceptors than yohimbine and its isomer rauwolscine (Doxey et al., 1984) and has been used to label ~2-adrenoceptors (Howlett et al., 1982). However, in 1987 Coupry and colleagues demonstrated that in addition to labelling ~tE-adrenergic sites, [3H] idazoxan binds to nonadrenergic imidazoline preferring sites in rabbit basolateral membranes (Fig. 3). Adrenaline and rauwolscine displaced less than 25% of specifically bound [3H] idazoxan whereas the imidazolines tramazoline and tolazoline displaced all specifically bound idazoxan. Moreover, whereas [3I-/] idazoxan gave a binding site density of 566 _+ 118 fmol/mg protein Bm,x

Imidazoline receptors and blood pressure

235

TABLE 2. Drug Affinities at Imidazoline Sites Labelled by [3HI Idazoxan

Species

Tissue

Man Rabbit Pig Rat Man Rabbit Rabbit Rabbit Man Rabbit Pig Rat Man Rabbit Rabbit Rabbit

Kidney Kidney Kidney Kidney Liver Liver Urethra Forebrain Kidney Kidney Kidney Kidney Liver Liver Urethra Forebrain

Drug Guanabenz Guanabenz Guanabenz Guanabenz Guanabenz Guanabenz Guanabenz Guanabenz Clonidine Clonidine Clonidine Clonidine Clonidine Clonidine Clonidine Clonidine

Ki (nu) 9.6 + 1.55" 1.8 ___0.5"t" 18 + 5"~: 87§ 172.5 + 50.7L[ 4.7 + 0.711 4.4 ___0.9¶ 1.3 + 0.1 t 2575 + 710t 303 + 311" 10,000 + 3,000":~ > 10,000~ 6763 +_64211 7406 + 12991[ 1265 4- 170¶ 543 _ 80*l

*Results expressed as IC50 rather than Ki. tLachaud-Pettiti et al. (1991). :~Vigne et al. (1989). §Michel et al. (1989). lITesson et al. (1991). ¶Yablonsky and Dausse (1989). ~Hamilton et al. (1991).

for [3H] rauwolscine binding under the same conditions was only 155 + 28 fmol/mg protein consistent with idazoxan labelling additional sites. Reports of [3H] idazoxan binding to nonadrenergic sites in other tissues followed rapidly. In the rabbit nonadrenergic binding to brain (Hamilton et al., 1988), smooth muscle urethra (Yablonsky et al., 1988) and adipocyte membranes (Langin and Lafontan, 1989) was demonstrated. In addition, nonadrenergic binding of [3H] idazoxan to rat and human renal cortical membranes (Michel et al., 1989), human platelets and myometrium (Michel et al., 1989) and guinea pig cerebral cortex (Wikberg, 1989), pig kidney (Vigne et al., 1989), rat brain (Brown et al., 1990) and human and rabbit liver (Tesson et al., 1991) has been shown. Differences appear to exist between the nonadrenergic sites labelled by idazoxan in different species (Table 2). For example, clonidine and phentolamine did not displace [3H] idazoxan from human kidney (Michel et al., 1989) but did from rabbit kidney (Hamilton et al., 1988) and guinea pig brain (Wikberg, 1989). Differences in binding affinity in human and rabbit liver have also been reported for guanabenz and UK 14304 (Tesson et al., 1991), and as for nonadrenergic [3H] clonidine binding the possibility of species heterogeneity has been suggested for nonadrenergic [3HI idazoxan binding (Michel and Insel, 1989). There is also evidence for within species tissue heterogeneity. Wikberg and colleagues (1991) have suggested that subtypes of [3H] idazoxan imidazoline sites can be delineated using stereoisomers of medetomidine in guinea pig tissues. Whereas the affinities of the ( - ) and ( + ) isomers for the [3H] idazoxan binding site were similar in the ileum the affinity of the ( - ) isomer was approximately ten times lower than that of the ( + ) isomer in the cortex. In addition to heterogeneity within [3H] clonidine and [3HI idazoxan binding sites differences in the affinity of a number of compounds between these binding sites has been reported. Whether the nonadrenergic sites labelled by [3H] clonidine and [3H] idazoxan are one and the same or have separate identities and function has been the subject of some controversy (Bousquet, 1989; Michel and Insel, 1989). However, recently Hamilton and colleagues (1991) demonstrated differences in nonadrenergic [3H] clonidine and [3H] idazoxan binding in rabbit forebrain and kidney membranes under identical experimental conditions. On the basis of these studies these authors suggested that a whole family of imidazoline receptors may exist (Hamilton, 1992).

236

C.A. HAMILTON 3. CHARACTERIZATION OF IMIDAZOLINE PREFERRING SITES

In most tissues studied ~2 and imidazoline preferring sites can be separated from each other in a number of ways. The nonadrenergic imidazoline site labelled by [3H] idazoxan has been partially purified and characterised and shown to be distinct from the 0t2-adrenoceptor. An anatomical location of [3H] idazoxan sites distinct from [3H] rauwolscine sites has been reported in rat brain (Boyajian et al., 1987). Using autoradiographic techniques the authors showed that the densest areas for [3H] idazoxan binding appeared over anterior olfactory nuclei, fundus striatum, septum, thalamus, hypothalamus, amygdala, entorhinal cortex, central gray, inferior colliculus, dorsal parabrachial nucleus, locus ceruleus and nucleus of the solitary tract. The most dense [3H] rauwolscine labelling appeared over nucleus caudate-putamen, nucleus accumbens, olfactory tubercle, islands of Colleja, hippocampus, parasubiculum, basolateral amygdaloid nucleus and substantia nigra. The maximum binding values for [3HI idazoxan were consistently higher than those for [3H] rauwolscine and the pharmacological data (Boyajian and Leslie, 1987) was consistent with [3H] rauwolscine labelling adrenergic sites while [3H] idazoxan labelled both adrenergic and imidazoline preferring sites. Further evidence that the 0t2 and imidazoline binding sites labelled by [3H] idazoxan are distinct comes from studies in which COS 7 cells transfected with genes for the ~2C4 and ~2C~0 subtypes were shown to contain no nonadrenergic [3H] idazoxan sites (Michel et aL, 1990). The imidazoline sites labelled by [3HI idazoxan in guinea pig cerebral cortex (Wikberg and Uhlen, 1990) and rabbit kidney (Parini et al., 1989) have been solubilized and could be physically separated from ~2-adrenoceptors, indicating that the ct2 and imidazoline sites labelled by [3HI idazoxan are discrete entities and not different sites as the same protein. The imidazoline site was trypsin sensitive showing it to be a protein not a lipid component of the membrane, and in contrast to ~2-adrenergic sites was not bound to heparin-agarose or lectin resin indicating that it was devoid of complex N linked oligosaccarides (Parini et al., 1989). Unlike ~2-adrenoceptor sites the imidazoline sites labelled by idazoxan may not be sensitive to changes in Na ÷ concentration or GTP and its analogs (Michel et al., 1989). However, GTP and its analogs only modulate agonist binding to the receptor and, as Michel pointed out, it is possible that although UK 14304 (5 bromo-6-(imidazolin-2-ylamino)-quinoxaline) is an agonist at ~2-adrenoceptors it was not acting as an agonist at the imidazoline sites labelled by idazoxan in rat kidney. Additional evidence for a lack of sensitivity to GTP has been obtained recently in that GTP analogs also failed to modulate binding of a number of other compounds including guanabenz, clonidine and cirazoline at the imidazoline site (Zonnenschein et al., 1990). Although the [3HI idazoxan imidazoline preferring site is insensitive to Na ÷ and divalent cations it has been reported to be sensitive to K ÷ (Zonnenschein et al., 1990; Lachaud-Pettiti et al., 1991). In basolateral membranes from human kidney, K + inhibited binding by up to 70% with an ECs0 of 70 mM and acted as a competitive inhibitor (Lachaud-Pettiti et al., 1991). In rat liver 4 aminopyridine, a K ÷ channel blocker in neuronal cells inhibited specific idazoxan binding with an IC50 of 0.34 mM, a concentration which is effective in blocking K + channels. Cs ÷ and NH~- also interfered with [3H] idazoxan binding and it has been suggested that these imidazoline sites might be coupled to K ÷ gating (Zonnenschein et al., 1990). Recently an intracellular location for the imidazoline sites labelled by [3H] idazoxan has been proposed. In human platelets Michel and coworkers (1990) showed that, although imidazoline preferring sites could be detected in crude membrane preparations, they were not found in purified plasma membranes suggesting that these sites were located in an intracellular compartment. An intracellular location for these sites in human renal proximal tubules has also been proposed (Lachaud-Pettiti et al., 1991); while the major localization of 0q-adrenoceptors was in the basolateral membranes, imidazoline sites were present in both basolateral and intracellular membranes. Moreover, in human and rabbit liver the imidazoline preferring site labelled by [3H] idazoxan have been localized to the outer mitochondrial membrane (Tesson et al., 1991). Interestingly, most of the drugs reported to have a high affinity for the nonadrenergic sites labelled by idazoxan are relatively lipophylic and thus should have access to intracellular binding sites.

Imidazoline receptors and blood pressure Clonidine 0.1

1

¢t-MNE 0.1

1

237 ST 587

Cirazollne 10

0.01 0.1

1

3

lo~g

-10

._. -20

-30

7..

FIG. 4. Effects on the mean blood pressure (MAP) of various ct-adrenergic drugs injected bilaterally into the NRL of anesthetised normotensive cats. r = 4; *p < 0.05; **p < 0.01; ***p < 0.001. Reprinted from Bosquet et al. (1984), with permission of the authors and the copyright holder, American Society for Pharmacology and Experimental Therapeutics, Bethesda. 4. PHYSIOLOGICAL RESPONSES LINKED TO THE IMIDAZOLINE SITES If these nonadrenergic imidazoline binding sites are true receptors they should be linked to some physiological response. A number of roles for both the sites labelled by [3H] clonidine and those labelled by [3HI idazoxan have been proposed with varying amounts of supporting evidence as discussed below.

4.1. BLOOD PRESSURE AND HEART RATE REGULATION

There is good evidence that the sites in the brain labelled by [3H] clonidine and [3H] p-aminoclonidine are involved in the regulation of blood pressure. As long ago as 1984 Bousquet and colleagues reported that injection of imidazolines (clonidine, cirazoline and ST 87) caused hypotension when injected into the nucleus reticularis lateralis of anaesthetised cats while the • 2-adrenoceptor selective catecholamine ~ methyl-noradrenaline had no effect (Fig. 4). Subsequently other publications indicated that effects of ~2-adrenoceptor agonists on blood pressure and heart rate might be related to their structure. In 1986 Sasaki and colleagues reported that although injection of both clonidine and guanfacine into the lateral ventricle of anesthetized rats resulted in a decrease in blood pressure and heart rate, only responses to clonidine were blocked by yohimbine. Local injections into the nucleus tractus solitarius (NTS) and ventromedial hypothalamus (VMH) suggested that clonidine was acting on the NTS while guanafacine acted more rostrally on the VMH. King and Pang (1988) suggested that clonidine (imidazolidine) and BHT 920 (thiazoloazepin) had different cardiovascular effects on intracerebroventricular injection, clonidine significantly decreased blood pressure and had a small effect on heart rate, while BHT 920 increased blood pressure and decreased heart rate. Neither the pressor effect of BHT 920 nor the depressor effect of clonidine was abolished by rauwolscine. Although these results suggested that not all the depressor and bradycardic effects of centrally acting antihypertensive drugs are mediated by ~t2-adrenoceptors, no overall pattern was apparent. However, more recently it has been shown that in anesthetized rats the hypotension and bradycardia induced by drugs injected into the rostral ventrolateral medulla correlated with affinity at imidazoline sites labelled by [3H] p-aminoclonidine but not with ct2-adrenoceptor affinity (Ernsberger et al., 1990) (Fig. 5). Furthermore, the imidazole idazoxan selectively reversed the fall in arterial pressure elicited by clonidine

238

C.A. HAMILTON

10Pl~nym ~

PIw~rm ~

o -

o

..............................

N~=~zo~

•

"1,--'-

N~mro=n= HTM

-10

a-MIe~E

-10

•

Kolne~ i NE O

-20

NE

-

-20

Oxymmzoene

g

-30

C~dme

-30

I

p

o

CNnet~e

PAC

/Tc-.

.or - 0.837

r - -0.051

-50 "~-1o

-9

-8

-7

Affinity at Imidazole

-s

-s

binding sites

> -4

,

-10

(pKi)

I

I

I

-9

-8

-7

Affinity

1 i -6

, 4.,,,...J g >4

-5

at 0.2 -receptors (pKi)

FIG. 5. Relationships between vasodepressor potency and binding affinity at imidazole binding sites or c