Bioorgnnic & Medicinal Chemistry Letters. Vol.2, No.8, pp. 839-844. 1992

Printed in Great Britain 0960-894X,92 $5.00 + .OO

0 1992 Peqamon Press Ltd

RIGID ANALOGS OF ACETYLCHOLINE CAN BE Ml-SELECTIVE AGONISTS: IMPLICATIONS FOR A RATIONAL TREATMENT STRATEGY IN

ALZHEIMER’S DISEASE

Abraham Fisher’, David Gurwitz, Dov Barak, Rachel Haring, Ishai Karton, Rachel Brandeis, Zipora Pittel, Daniele Marciano, Haim Meshulam, Zvi Vogel’and Eliahu Heldman.

Israel Institute for Biological Research, P.O.Box 19, Ness-Ziona 70450, and The Weizmarm Institute, Rehovot’, ISRAEL.

(Received 1 May 1992, accepted 25 June 1992)

Abstract

Rigid analogs of acetylcholine offer an opportunity for selective actions at muscarinic receptor subtypes, since restricted conformational mobility alters the capacity of ligands to adapt to subtle differences in receptor structure. AFlOZB, a highly rigid analog of acetylcholine, is a centrally active Ml agonist and is evaluated in light of some current- ly available therapeutic strategies in Alzheimer’s disease.

Introduction

Molecular biology has revealed the existence of five distinct human muscarinic receptor (mAChR) subt s

F . (ml-m5)‘. It is likely that the ml, m2, m3, m4 AChRs fit the phar-

macological de motion of the Ml, M2, M3 and M4 AChRs, respectively’. According to the cholinergic hypothesis in Alzheimer’s disease (AD)*, the demented state might be alleviated by treatment with Ml or ml agonists?, M2 or m2 antagonists3’,b, m3 agonistssb, or m3 antag~nists~~***.

An implicit assumption, in the search for Ml or ml specific muscarinic agonists, has been that there are differences in the binding sites for the mAChR subt pes1v3.

K In this

regard, it has been suggested that for rigid analogs of acetylcholine (AC ), the confor- mational and stereochemical restraints of a articular mAChR subtype can uniquely be satisfied while others are partially excluded or the other receptor subtypes4. In addition, P rigid ligands can display different receptor association and dissociation characteristics, when compared with analogous flexible structure$. This ma has nothing to do with the specific pharmacophoric pattern o iy

result in selectivity which the rigid ligand. Moreover,

selectivity may not result from the agonist/receptor interaction alone, but also from the properties of the agonist/receptor/G-protein(s) complexes. Those questions will be ad- dressed through examination of the structural and pharmacological profile of (&‘i.r-Z methyl-sgiro(l,3-oxathiolane_5,3’)quinuclidine, AF102B, which is a selective Ml ago- nist?** 9. In this paper some current features of AFI02B, which may be relevant to a rational treatment strategy in AD are presented. Comparison is made, whenever possible, with some new and old muscarinic agonists (Fig 1).

Materials and Methods

Chinese hamster ovary (CHO) cells stably transfected with rat ml AChR or m3 AChR and human ml-m5 AChRs were used in this study. In these cell cultures tested agonists were assayed on the following parat?eters and according to the mEthods in cited refs: phosphoinositide (PI) hydrolysis ; cyclic AMP accumulation mobikztion I’; intmcelZular &’ accumulation Id.

; amchidonic acid

* To whom corms ndence should b &

e addressed ** The terms ml, m2 3 4 5 andMl,M2or hgands (agonists or antagonisk) are used for ligands &ek&?tzi ?te cloned ml-m5 and Ml, M2 or M3 pharmacologically characterized mAChRs, respective- ly*

839

840 A. FISHER~~ al.

CH3 0,,.+N.‘CH3

x , “CH3

CH3

Acatylchaline

N4;2 et d

N

L-670-207

ml + ,‘% “: “‘CH 3 CH3

X-0 Dioxolane X-s Oxmhiilafle

CH3rx

-2s NL

x-0 ds-AF30 X-S AFlOP

CH I 3

A - ,2”3 w

“u

, 0 s N CH,-N-4 CH,_N’z - - “’

x Y z R

: Me SND210066

C:CH :; Et YM796 C-C GO= N Et RSBB

AF150 AF151 SDZ ENS 163

Rest&s and Discussion

The findings obtained to date indicate that AFlO2B is a rather selective Ml agonist, exhibiting unique models for AD3L,4*S.

roper-ties in restoring memory impairments in a variety of animal Notably, AFlO2B can act as a full agonist, a partial agonist, or an

antagonist depending on the trssue, Examination of

on the mAChR subtype and functional assay studied Table 1 and the cited refere&*$@ indicates that AF’IO2B can best be

considered as a selective Ml or m 1 agonist when defined through binding and function- a13’pq6*89 assays. Yet, it is possible that due to different receptor reserves, AFIO2B might show full agonistic activity on certain Ml AChRs (e.g. those modulatin es) and partial agonistic or even antagonistic

memory process- activi gon

by peripheral or central (Ml, M3?) AC~RS”+~*“* other type of Eln ctions mediated

Albeit plausible, this explanation cannot clarify completely why AF102B does not induce mainly central effects such as tremors, hypothermia and purposeless chewing (M2 or M3 agonistic effects)k*7. In fact, the M2 are more efficiently coupled to effector systems than the Ml AChRs”. Thus, a more reasonable explanation would be that AFlO2B is a unique functionally selective Ml and ml agonist. In this context, in CHO cells transfected with ml AChR, AF102B in- creased intracellular Ca” almost to the same extent as carbachol, while activating only partially PI hydrolysis and arachidonic acid release, and completely failing to increase c.4MP accumulation (Table 1). By comparison, AFIO;?B, showed only marginal agonis- tic effects on PI hydrolysis and arachidonic acid release and no effect on cAlUP accumu- lation in CHO cells transfected with the m3 AChR subtype. Results obtained with the m2, m4 and m5 AChRs also emphasize the functiona selectivity of AFlO2B towards ml AChRs (Table 1).

Rigid analogs of acetylcholine 841

Table 1. The effects of McN-A-343*, AFlOZB, AFl50 and AFlSl on PI bydmiysk, omchidordc acid rekose, CAMP accumulation and bzreese In intracellular Ca*+ levels in CHO celk stably tramfected with human ml-m5 AChRs. ResoIts in 91 are expmsed mktive to cubachol-induced effects.

The structure of AF102B embodies the muscarinic pharmacophore in a framework *of high rigidity being closely related to the rigid analog cis-AF30, to the semirigid muscann- ic a onist acceclidine and to ACh. However, unlike AF102B the active diastereomer of AF 3” 0 is somewhat less Ml selective &et”. We investigated in this study whether variation in selectivity can result from differences in the pharmacophoric pattern of these and some other published muscarinic agonists. This was done using a slightly modified pharmaco- phoric model for muscarinic activity proposed by Schulman and Sabio”. In this model the cationic head, the ether oxygen and the termmal methyl group were assumed to be essential for interaction with mAChRs. The question of subtype selectivity in the proposed pharmacophore was not addressed in construction of the model. Application of the model to a variety of muscarinic agonists, regardless of their selectivity, shows that satisfactory parameters for all cases could be obtained (see Table 2). Thus it appears that the musca- rinic pharmacophore (as defined here) is the common denominator of the structural re- quirements for agonistic activity. However, in the case of pilocarpine and some of its ana- logs (e.g. SDZ ENS 163’“), the validity of such a generalized pharmacophore has been questioned”.

842 A. FISHER et al.

In Table 2, a large spatial variability for the other heteroatom (or its equivalent) is observed for the evaluated compounds. The common pharmacophoric pattern, shared by these agonists, suggests that additional interactions might be involved in dictatmg selectiv- ity. Yet no obvious correlation between the model parameters and selectivity can be found. Albeit, there seems to be a direct correlation between potency and P-O distance, neither the P-O distance nor the P-0-C-Het angle correspond to the Ml/M2 relation. Therefore, selectivity in an agonist appears to depend on a more complex summation and balance of parameters including, inter alia: conformational rigidity Het,

., p”tenFy, the natureof its spatial location., the ml-m5 mAChRs subtypes and their drf erenhal couplmg wrth

a repertoire of G-proteins in various tissues. Consequently, most of the known putative Ml selective agonists show partial activities, at least in some of the standard assays. On the other hand, full muscarinic agonists appear to be M2 selective (for highly potent M2 agonists see Ref 14), perhaps due to a more efficient coupling of M2 AChRs to effector systemsU.

Table 2. Moleculnr modeling of some muscminic agonists

h’ote: Point P corresponds to the anionic site as defined in Schulman’s model”. Distances behveen the oharmacoDhmic elements were used instead the original P, Q parameters. All compounds zwerc sup&imposed with respect to points P,O,e resultii~rms < 0.3A. Dihedral angle P-0-C-Het is used to demonstrate the varying position of hetematom or its equivalent @et) for the diierent compouods. The modelling study, which included +c- tural optimizations by means of molecular mechanics, structural comparisons and nsual iqechons, %vas performed on IRIS 4D graphic workstation. * not defined uniformly by thczamemethodandinthesamelab.

Rigid analogs of acetylcholine 843

According to the presently accepted models of receptor activation and signat transduc- tion, the efficiency of coupling between the agonist/rece repertoire of G proteins will determine the magnitude o F

tor complex and the cellular the signal transmitted. Such

coupling depends on the structure of the complex (other cellular environment being con- stant), and for complexes of a receptor with different agonists it will depend upon the agonist structure and its molecular properties. It follows that while different mAChR subtypes may have similar binding parameters for a given agonist, the resulting receptor/agonist complexes can display differing reactivities towards various G-proteins. That is, the structuraI heterogeneity of mAChRs can be manifested during the conforma- tional changes taking place upon binding of an agonist. In the case of a flexible agonist, mutual conformational changes can occur in both receptor and l&and, yet such conforma- tional freedom is prohibited in a rigid structure. Therefore, this is a that can be most influenced by rigidity in a given agonist, inasmuc E””

ently the event as rigid agonists

can impose only certain conformational changes in the receptor structure. Thus only dis- tinct G-proteins are in fact activated by such agonists, followed by activation of the G- protein GTPase, (e.g. can form agonist/receptor/G-protein complexes), while others are apparently “silerrt”. Notably, for the ml AChR ysd m case of compounds like AF102B and two new ngrd agonrsts AF150 and AF151 ipg denced by a selective increase in intracellular Ca

1, Table l), this feature is evi- actrvahon of PI hydrolyses and

amch&mc acid release, yet lack of significant effect b CAMP accumulation (Table 1)‘. It is plausible, however, that the CAMP signaling pathway via ml AChRs may exhibit less “spare receptors” compared with the other signals; this may explain the higher carbachol concentratrons required for stimulating CAMP accumulation compared with other sigMllSQb.

Finally, rigidity in an agonist can lead to discov level of the receptors, as well as along distinct * 3

of selective ml agonists both at the transduction pathways. In this con-

text, the highly rigid ml agonist AFlMB, unlike carbachol, had no effect on either cAMP levels in intact mlAChR-transfected cells, or adenylyl cyclase activity in isolated mem- branes prepared from the same cells’. AFlO2B improved memory and learning deficits in a varietv of animal models. which mimic cholinereic deficits renorted in AD. without p$$ts~&~advlv central c$ peripheral side-effects-at the effecti\;e doses *and showing-a

“v4*‘. Tlus compound is a prormsmg candrdate drug and 1s presently m Phase II clinical trials in AD natients. Notablv. mRNA for both mlAChR’ ‘and Galgb were markedly elevated in po&-mortem brain &sues of AD Gs anddecrease

patient~~and an elevation of of Gi moteins were moorted in aged human brains . Taken together.

these observations ma in these situations. Al ert speculative, it is thus possible that the desired Ml or ml-selec- %.

Imply increased &isitivity sf mlAChR-mediated adenylyl &cla&

tive agonists for the treatment of AD, should not stimulate adenylyl cyclase via mlAChR. Cholinergic replacement is currently being evaluated for the treatment of ADs. In this context both selectivity, at the level of various receptor subtypes, as well as at the level of signal transduction for the same mAChR subtype Should be considered in rational drug design.

Acknowledgement: Supported in part by Snow Brand Milk Products, Japan.

1.

2. 3.

(a) E. C. Hulme, N.J.M.Birdsall and N.J.Buckley; ti Rev. Phar& Toxicol,, 1990, Nn, 633. (b) R. Stein, R. Pinkas-Kramarski and M. Sokolovsky; EMBO L, 1988,2, 3031. (c) C. C. Felder, R.Y. Kanterman, A.L. Ma and J. Axelrod; J. Biol. Chea, 1989,2&, 28356. (a) C. C. Felder, M. 0. Poulter and J. Wess; Proc. Nat w SEI;, 1992, @, 509. R.T. Bartus; L $n, ti a, 1989,2& 680. (a) A. Fisher and E. Heldman; In: &&Clinical anp -tical m gf

844 A. FISHER etal.

4.

5.

6.

7.

8.

9.

10.

:4 13.

14.

15.

16.

17.

18.

19.

imer’g gnd Parkinson’s Diseases, T. Nagatsu, A. Fisher A and M. Yoshida M, eds., New York and London, Plenum Press, 1991, m, 309. (b) L. T. Potter; In: &heimer’s Disease m Treatment m. Z. S. Khachaturian and J. P. Blass, eds. Marcel Dekker, New York, 1992, 57. (c) S. Sugita, N. LJJ~mura, Z. -G. Jiang and R.A. North. Proc. N& &g& Sci. US, 1991, 88,

(a) A. Fisher, R. Brandeis, I. Karton, Z. Pittel, D. Gurwitz, R. Hating, M. Sapir, A. Levy and E. Heldman; macol. ErptL Therau,, 1991,252, 392. (b) A. Fisher, B. Brandeis, Z. Pittel, I. Karton, M. Sapir, S. Dachir, A. Levy and E. Heldman E; Neurosci. m, 1989, .&Q, 325. (c) R. Btandeis, S. Dachir. M. Sapir, A. Levy and A. Fisher; Pharmacol. Biochem. B&g!& 1990,%,89. (d) M. McKinney, D. Anderson and L. Vella-Rountree; Eacrl, N., 1988, 3.L 39. (a) Y. Lass, S. Axelrod, B. Gavish, S. Cohen and A. Fisher; Fxnerienti& 1979, ;15, 650. (b) A. Miklavc, D. Koqan, F. Avbelj and D. HadzifL Chem. && Faradav w 2,!988, M, 763. S8~occ,9F. Mizobe, A. Fisher, G. Kawanishi and H. Kobayashi; I&& &&,

(a) k Nakakua, Y. Iga, F. Mlzobe and G. Kawanishi; &T& L Pharmacol., 1988, 48,502. (b) N. Nakahara, Y. Iga, Y. Saito, F. Mizobe and G. Kawanishi G; Zpn. L Pharmacok, 1989, a, 539. (a) S. Ono, Y. Saito, N. Ohgane, G. Kawanishi and F. Mizqbei E1K,Lgbana_ $,o~!f9k~l~~.~~:“) N. Ogane, Y. Takada, G. Kawanishi and F. Mlzobe;

(a) h. Gurwiiz, RI Haring, C. M. Fraser, E Heldman and A. Fisher && Neurosci. Abstr., 1991, 12, 388. (b) D. Gun&, R.Haring, C. M. Frank, E. Heldman and A. Fisher, submitted. J. Saunders, G.A. Showell, R. Baker, S.B. Freedman, D. Hill, A. M&night, N.

Newberry, J.D. Salamone, J. Hirshfeld and J.P. Springer. J. Med. Chem, 1987, 3,969. J. M. Schulman, M. L. Sabio and R.L. Disch; L &&& Cliem,, 1983,z, 817.

N. J. M Birdsall and E. C. Hulme; In: T&MuscwReceptors, J. H. Brown, ed. The Humana Press, Clifton NJ,1989, 31. P. Angeli, I. Brasili, F. Cantalamessa, G. Marucci and J. Wess; Naunvn,

Schmied,Brr;h,Pharmacol., 1990, J&& 625. S. Freedman, E. A. Harley, S. Patel, N. R. Newberry, M. J. Gilbert, A. T.

McKnight, J. K. Tang, J. J. Maguire, N. T. Mudunkotuwa, R. Baker, L. J. Street, A. M. MacLeod, J. Saunders and L. L. Iversen; &L Pharmacol, ,19% m,575. A. Enz, J.A. Gray and R. Spiegel; In: Cholinera ic && fiar Alzheimer Therapy, R. Becker and E. Giacobini, eds., Birkhauser,l991,370.

F. Wanibuchi, T. Konishi, M. Harada, M. Terai, K. Hidaka, T. Tamura, S. Tsukamoto and S. Usuda, m k pharmacol., 1990, m, 479. (a) G. Shapiro, P. Floersheim, J. Boelsterli, R. Amstutz, G. Bolliger, H. Gammenthaler, G. Gmelin, P. Supavilai and M. Walinshaw; L&&d. Chem., 1992,% 15. (b)J. M. Schulman , R. C. Peck and R.L.Disch; L Med. Chem,, 1991, g, 1455. A. Fisher, Y. Segall, H. Meshulam, E. Shirin, D. Gurwitz, R. Haring, R. Bran- deis, M. Segal, H. Marloam, Z. Pitt& C. M. Fraser, E. Heldman and I. Karton; &% Neurosci. Al?&, 1991,12,388. (a) P. I. Harrison , A. J. L. Barton, B. Najlerahim, B. McDonald and R. C. A.

Pearson; Mol. Brain Res,, 1991, P, 15. 0) P. J. Harrison , A. J. L. Barton, B. McDonald and R. C. A. Pearson; Mol. w& 1991, fi, 71. (c) L. T. Young, J. J. Warsh, P. P. Li, K. P. Siu, L. Becker, I. Gilbert, 0. Hornykiewicz and S. J. Kish; Dev. Bti &, 1991, a, 243.