A novel muscarinic receptor ligand which penetrates the blood brain barrier and displays in vivo selectivity for the m2 subtype

Life Sciences, Vol. 53, pp. 1743-1751 Printed in the USA

Pergamon Press

A NOVEL MUSCARINIC RECEPTOR LIGAND WHICH PENETRATES THE BLOOD BRAIN BARRIER AND DISPLAYS IN VIVO SELECTIVITY FOR THE m2 SUBTYPE Miriam S. Gitler 1 , Victor I. Cohen 1, Rosanna De La Cruz 1 , Sheila F. Boulay 1, Biyun Jin 1, Barry R. Zeeberg 1 , and Richard C. Reba 1,2 1 Department of Radiology, Section of Radiopharmaceutical Chemistry, George Washington University Medical Center, 2300 Eye St., N.W., Washington D.C. and 2Department of Radiology, Nuclear Medicine Section, University of Chicago Hospital, 5841 S. Maryland Avenue, Chicago IL (Received in final form September 27, 1993)

Summary Alzheimer's disease (AD) involves selective loss of muscarinic m2, but not ml, subtype neuroreceptors in the posterior parietal cortex of the human brain. Emission tomographic study of the loss of m2 receptors in AD is limited by the fact that there is currently no available m2-selective radioligand which can penetrate the blood-brain barrier. In our efforts to prepare such a radioligand, we have used competition studies against currently existing muscarinic receptor radioligands to infer the in vitro and in vivo properties of a novel muscarinic receptor ligand, 5[[4-[4-(diisobutylamino)butyl]-I -phe nyl]acetyl]- 10,11 -dihydro-5 H-dibenzo [b,e][1,4]diazepin-11-one (DIBD). In vitro competition studies against [3H](R)-3-quinuclidinylbenzilate ([3H]QNB) and [3H]N-methylscopolamine ([3H]NMS), using membranes derived from transfected cells expressing only ml, m2, m3, or m4 receptor subtypes, indicate that DIBD is selective for m2/m4 over ml/m3. In vivo competition studies against (R,R)-[1251]IQNB indicate that DIBD crosses the blood brain barrier (BBB). The relationship of the regional percentage decrease in (R,R)-[1251]IQNB versus the percentage of each of the receptor subtypes indicates that DIBD competes more effectively in those brain regions which are known to be enriched in the m2, relative to the ml, m3, and m4, receptor subtype; however, analysis of the data using a mathematical model shows that caution is required when interpreting the in vivo results. We conclude that a suitably radiolabeled derivative of DIBD may be of potential use in emission tomographic study of changes in m2 receptors in the central nervous system. It appears that selective loss of m2 subtype neuroreceptors in the posterior parietal cortex of the human brain occurs in patients with Alzheimer's disease (AD) (1-5). Several studies (6-9) have recently attempted to determine whether single photon emission computed tomogral3hic (SPECT) imaging of the distribution of (R)-3-quinuclidinyl (R)-4-iodobenzilate ((R,R)-[1231]IQNB) could be useful for detecting pathological changes in muscarinic neuroreceptor concentration in patients with dementia. These studies are subject to the limitation that the m2 subtype constitutes only about 19% of the total muscarinic acetylcholine receptor Corresponding Author: B. Zeeberg, Radiopharmaceutical Chemistry, the George Washington University Medical Center, Room 662 Ross Hall, 2300 Eye St., N.W., Washington, D.C. 20037 0024-3205/93 $6.00 + .00 Copyright © 1993 Pergamon Press Ltd All rights reserved.

1744

h~ Vivo Muscarinic m2 Selective Ligand

Vol. 53, No. 23, 1993

(mAChR) in the posterior parietal cortex (10). Since (R,R)-[1231]IQNB is nonsubtypeselective (11), even a complete loss of the small fraction of m2 receptors would correspond to a small relative change in the observed (R,R)-[1231]IQNB accumulation. Thus, it is essential to develop a radioligand which can penetrate the blood brain barrier (BBB) and which has high in vivo selectivity for the m2 subtype. The m2-selective compounds AF-DX 116 (12-15) and DIBA (15) do not significantly penetrate the BBB (unpublished observations). The m2-selective compounds AF-DX 384 (16) and AQ-RA 741 (15,17) are also expected not to be able to penetrate the BBB. We describe here the in vitro and in vivo characterization of a novel mAChR ligand, 5-[[4-[4-(diisobutylamino)butyl]-1-phenyl]acetyl]-10,1 1-dihydro-5H-dibenzo [b,e][1,4]diazepin-11-one (DIBD) (Fig. 1), which appears to penetrate the BBB and which has high in vivo selectivity for the m2 subtype.

C=O ~

,C=O CH2NEt2~ C H 2 N P r 2

AF-DX 116

AF-DX 384

C=O,

C=O

C=O

(~iH2)4

(C~H2)4

(~H2)4

NEt2

NEt2

AQ-RA 741

DIBA

Ni-Bu2 DIBD

FIG. 1 Structural formulas of AF-DX 116, AF-DX 384, AQ-RA 741, DIBA, and DIBD. Methocf$ Radiopharmaceuticals

and

chemicals

The (R,R)-diastereomer of IQNB, (R,R)-[1251]IQNB, was radioiodinated (379 Ci/mmol) and characterized according to the method of Rzeszotarski et al. (18). (R,R)-[1251]IQNB was obtained by the reaction of an (R,R)-QNB triazene precursor with Na1251. The product was purified by high performance liquid chromatography (HPLC) using a Z-module C-18 Bondapak reversed-phase cartridge, eluted with MeOH:water (60:40) containing 10 mM formic acid and 1% octane sulfonic acid as an ion pair agent to improve resolution. Final radiochemical purity was _> 95% as determined by thin layer chromatography (TLC). Specific activity of (R,R)[1251]IQNB was determined by comparing,., the radioactivity specifically bound to a caudate nucleus homogenate to that when using [°H]QNB of known specific activity (19). DIBD was prepared as described elsewhere (20). In

vivo

studies

General Experimental Procedures Male Sprague-Dawley rats weighing 200-250 g were used in the experiments. Animals were anesthetized with ketamine:xylazine (100:10 mg/kg i.p.) and the right jugular vein was exposed for intravenous injection of all compounds (see Experimental Design section). Animals were maintained under anesthesia until time of sacrifice. At the end of each study, animals were sacrificed by decapitation and the brains were rapidly removed, blotted free of excess blood and placed on ice. Tissue samples (20-70 mg) of specific

Vol. 53, No. 23, 1993

In Vivo Muscarinic m2 Selective Ligand

1745

brain regions were counted for 2 min in an autogamma counter (GammaTrac 1193, Tm Analytic; Elk Grove Village, II) with a counting efficiency of 78% for 1251. The brain regions of interest included the cerebral cortex, corpus striatum, thalamus, hippocampus, ports, medulla, and cerebellum. In order to determine if there were any differences in the in vivo accumulation of (R,R)-[1251]IQNB between the right versus left cerebral hemispheres, the left and right cerebral cortex, corpus striatum, and thalamus were dissected and studied as separate entities.

Experimental Design Animals were injected with a single bolus of DIBD (1000 or 3000 nmol in a final volume of 0.10 ml distilled water containing 10% N,N-Dimethylformamide (DMF) and 10% Emulphor EL-620 (Rhone-Poulenc, Cranbury, NJ) or the solvent alone. Immediately after the injection of DIBD, animals received a bolus injection of (R,R)-[ 1251]IQNB (6 uCi in a final volume of 0.10 ml normal saline containing up to 55% ethanol). After a 1 h interval, animals were sacrificed. The brains were rapidly removed, and tissue samples were dissected and treated as described above. Data Analysis In all studies, at least 6 animals were used per reported result. The percent of (R,R)-[1251]IQNB injected dose per gram (% dose/gram) of tissue (wet weight) was calculated. The mean of each group was used to determine the percent displacement of (R,R)[1251]IQNB from specific brain regions by DIBD. For each brain region studied, the percent displacement in that region was plotted (Figs. 2, 3) against the percent of each subtype in that region (10, 21-23), and linear fits were performed using the built-in curve-fitting function of a commercial graphics program (Cricket Graph). Iterative

Curve-Fitting

The ratio Rj = {specific + nonspecific binding of [1251]lQNB}/injected dose of [1251]IQNB (nM/nmol)

(Eq. 1)

for binding to multiple receptor subtypes within the jth brain region, in the presence or absence of cold competing DIBD, was computed as m

Rj = q + ~,, (5i Bi,j (1 + c/IC50i) -1 i=1

(Eq. 2)

where T1 (nM/nmol) is the nonspecific binding of (R,R)-[1251]IQNB (including unbound radioactive ligand), m is the number of receptor subtypes, Bi,j (nM) is the concentration of the ith receptor subtype within the jth brain re~ion, (5 i (/nmol) is a parameter which characterizes the amount of specific binding of (R,R)-[ 1 51]IQNB to the ith receptor subtvp_e, IC5o (nmol) i.s the C5o for the nhibton effected by DIBD upon the bnding of (R,R)-[1251]QNB to the th receptor subtype, and c (nmol) is the amount of DIBD injected. The linear relationship between R i a n d (~i results from the fact that there is insignificant receptor occupancy by (R,R)[1251]IQNB at the dose injected here. Since there is no assumption of equilibrium, the constant parameters (~i and IC5oi should be interpreted as applying only to the specific conditions of the particular experiment in which they were determined (that is, only for the particular schedule for injection of (R,R)-[1251]IQNB, injection of DIBD, and sacrificing the animals). An assumption is made that there are no "cross terms" such as (~i,k Bi,j Bk,j; this assumption is subsequently justified by the fit of the data to the model (Eq. 2). The sum of squares (SSQ) relative error was computed as n

SSQ =

[(Rmeas,j - Rj)/Rmeas,j] 2 j=l

(Eq. 3)

1746

hz Vivo Muscarinic m2 Selective Ligand

Vol. 53, No. 23, 1993

where n is the number of data points, and Rmeas,j (nM/nmol) is the measured quantity analogous to that described by Eq. 1. SSQ was minimized by nonlinear iterative least squares (24). The floating parameters were q, the (~i's, and the IC5oi's. The fixed parameters were the values for the Bi,j within each brain region (Table 1). In preliminary studies, it was found that the ml and m3 subtypes could be combined into a single pool (data not shown); the values of Bi,j used for this pool are the sum of the values of Bi d for the ml and m3 subtypes. Combining the ml, m3, and m4 subtypes into a single pool yielded similar parameter estimates, but with no improvement in the high standard errors for the estimate of IC5Ol 3 4 (data not shown). In preliminary studies, it was found that (3"1/3, and (~4 attained unbounde'dl'y high values. Thus, we artificially imposed an upper limit during the iteration procedure. We systematically varied this upper limit from one set of iterations to the next. Results and Discussion Although the in vitro selectivity of DIBD for m2 versus ml is comparable to the selectivity exhibited by AF-DX 116, AQ-RA 741, and DIBA, it appears to be nonselective for m2 versus m3 or m4 (Table 2). This nonselectivity is a potential disadvantage, since there is a substantial proportion of m3 and m4 (23) in the brain regions of interest in AD (1-5). Thus, we are involved in an ongoing effort to prepare derivatives of DIBD which might exhibit the desired m2 versus m4 selectivity. However, since DIBD appeared to be among the most promising of the compounds prepared to date (20), it is an appropriate candidate for initial in vivo studies. The purpose of the in vivo studies is to determine whether a proposed parent ligand crosses the BBB, and whether it exhibits an in vivo m2 selectivity. Since the potential ligand is nonradioactive, its properties must be studied by competition against a radioactive ligand. There is one essential property required of the radioactive ligand: it must have a high pharmacokinetic TABLE 1 Fixed Values for Bi,j's (nM) used in the Iteration Procedure 1 csl Bin1 Bm2 Bin3 Bin4

csr

72.79 30.12 16.32 112.95

hipp

72.79 76.14 30.12 27.54 16.32 16.20 112.95 30.78

ccl

ccr

thl

thr

med

pons

cb

70.38 39.33 20.70 49.68

70.38 39.33 20.70 49.68

13.76 36.12 5.16 17.20

13.76 36.12 5.16 17.20

2.60 36.40 2.86 2.60

2.60 36.40 2.86 2.60

0.42 15.75 1.05 0.53

1Computed from the data reported in refs. 10, 21-23. TABLE 2 KI Values and Relative Binding Indices (RBI) of Some mAChR Ligands LIGAND

ml

KI (nM) 1 m2 m3

m4

RBI (IC5o/IC5o QNB) 2 ml m2 m3 m4

QNB AF-DX 116 AQ-RA 741 DIBA DIBD

ND 3 740 34.0 4.0 ND

ND 73.4 3.7 0.3 ND

ND 545 15.0 2.0 ND

1.00 ND 310 41.4 22.7

ND 786 86.0 11.0 ND

1.00 ND 43.3 4.33 4.34

1.00 ND 820 80.2 12.9

1.00 ND ND ND 2.40

1Determined in competiton studies against [3H]N-methylscopolamine ([3H]NMS) (15). 2Determined in competiton studies against [3H](R)-3-quinuclidinylbenzilate ([3H]QNB) (20). 3Not determined.

Vol. 53, No. 23, 1993

In Vivo Muscarinic m2 Selective Ligand

1747

sensitivity with respect to each of the subtypes present. That is to say, if a given dose of the potential ligand results in 50% occupancy of the total ml+m2+m3+m4 receptors, then the total observed radioactivity must be decreased by approximately 50%, regardless of how the 50% occupancy happens to be divided up among the four subtypes. A detailed theoretical treatment of this issue has been published (25). It has been shown that (R,R)-[1251]IQNB is relatively non subtype-selective (11) and has the appropriate pharmacokinetic sensitivity (26). In vivo rat studies in the presence or absence of a preinjection of 708 nmol of (R)-QNB show that there is an insignificant percentage of nonspecific binding of (R,R)-[1251]IQNB in all brain regions, with the exception of the cerebellum (26). In vivo studies in the rat show that there is significant blockade of (R,R)-[1251]IQNB by 1000 (Fig. 2) or 3000 (Fig. 3) nmol of preinjected DIBD, and that the percentage blockade is correlated with the percentage of m2 subtype, but not with the percentage ml, m3, or m4 subtype. Because of the high percentage of nonspecific binding in the cerebellum, the cerebellum data are omitted; however, correction for the nonspecific binding in the cerebellum, which contains 75% m2 subtype (10), indicates close to a 100% decrease of the specifically bound (R,R)-[1251]IQNB (data not shown). These results indicate that DIBD crosses the BBB, and that DIBD exhibits an effective in vivo m2 selectivity. Furthermore, the low % displacement in several brain regions (such as hippocampus (Figs. 2, 3)) indicates that the injected DIBD does not induce a significant effect upon the delivery of (R,R)-[1251]IQNB. There is precedent for an in vivo selectivity which differs from the the in vitro selectivity: Frost et al (27) have shown that " . . . although diprenorphine appears to bind to the opiate receptor subtypes with approximately equal affinity in vitro, it may bind to only one subtype in vivo . . . " The results of iterative curve-fitting indicate that the measured data and the estimated parameter values are consistent with the model (Eq. 2): the estimated parameter values, when used in the model (Eq. 2), generated simulated values for Rj which are essentially identical to the corresponding Rmeas,j (Fig 4~ The estimated parameter values (Table 3, "no constraint") suggest that, whereas ~.,,R) L1251] IQNB exhibits a modest in vivo m2-selectivity, DIBD exhibits a large m2-selectivity. Although artificially imposing an upper limit to the estimated parameter values during the iteration procedure has a modest effect upon the SSQ error, the apparent m2-selectivity of DIBD is decreased and eventually essentially disappears as the upper limit is decreased. It would thus seem that we have no criterion for choosing between the candidate solution sets. However, the simulated data sets, generated from the various candidate solution sets, can be used to compute the percentage decrease in (R,R)-[ 1251]IQNB versus the percentage of m2 subtype (Fig. 5). Because of the poor error propagation properties of the subtraction process used in computing the % decrease in (R,R)-[1251]IQNB for the measured data, none of the candidate solution sets produces an exact fit to the measured data. However, it is clear that candidate solution sets with the lower values for the upper limit should be rejected, since they fail to exhibit the positive dependence upon % m2 which is apparent in the measured data. Furthermore, based upon either the SSQ errors for the iteration (top line of Table 3) or the SSQ errors for the fit to the % decrease in (R,R)-[1251]IQNB (ninth line of Table 3), it would be difficult to justify accepting candidate solution sets with upper limit < 1.o x lO5. Computation of the selectivity in terms of the ratios of the equilibrium dissociation constants (Table 3) shows that for our most conservative estimate (that is, for the candidate solution sets with upper limit = 1.o x lO5), DIBD has a substantial m2 versus ml/m3 selectivity of 20.53, and an insignificant m2 versus m4 selectivity. The ratio of binding potentials for m2 versus ml/m3 in the cortex, computed as (B2/IC5o2)/(B1/3/ICso1/3) = 8.9, shows that DIBD ispotentially useful in distinguishing between m2 versus ml/m3 in the cortex. Since (R,R)[12"51]IQNB appears to be p_otentially useful in distinguishing between m2 versus m4, utilization of both DIBD and (R,R)-[ 1251]IQNB may permit us to determine changes in m2, even in the presence of excess ml, m3, and m4 subtypes. A careful mathematical analysis is, therefore, required, even though inspection of Figs. 2 and 3 indicates an m2-selectivity for DIBD. The reason that intuition may be unreliable is that (R,R)-[1251]IQNB itself exhibits a degree of m2-selectivity. Thus, analysis of the results requires considering the subtype selectivity of the radioactive, as well as the cold, ligand.

1748

ht Vivo Muscarinic m2 Selective Ligand

20

~,

Vol. 53, No. 23, 1993

40

f

==30

.= O

_o

"7

10

• []

g

m2 m3 - THL • CSR

o~ o

O O 0

,'~ •

0

20

60

ME

THR THL CSL CCR CCL CSR

lo

CSL CCL HI

,

40 % subtype

PO

HI

80

2O

40 %

60

80

subtype

FIG. 2

FIG. 3

Percentage decrease of (R,R)-[1251]IQNB as a function of the percentage ml, m2, m3, or m4 subtype, for an injection of 1000 nmol of DIBD. Regions to which each % decrease corresponds are: PO (pons), ME (medulla), THR (right thalamus), THL (left thalamus), CSR (right corpus striatum), CSL (left corpus striatum), CCR (right cerebral cortex), CCL ( l e f t cerebral cortex), and HI (hippocampus).

Percentage decrease of (R,R)-[ 1251]IQNB as a function of the percentage ml, m2, m3, or m4 subtype, for an injection of 3000 nmol of DIBD. The brain region to which each % decrease of (R,R)-[125t]IQNB corresponds is indicated. Abbreviations are given in the legend to Fig. 2.

4-

y - - 3.2699e-2 * 1.001

R^2 =~.96 ! 30

3 z O

o

E

c •

c

202

O

n-

Oo lo 10

20 R

meas,j

(nM/nmol)

FIG. 4 Simulated (corresponding to the column "no constraint" in Table 3) versus measured values for (R,R)-[1251]IQNB combining the data for 0, 1000, and 3000 nmol injections of DIBD. The data for the cerebellum are included.

40

60

80

% m2

FIG. 5 Percentage decrease of (R,R)-[ 1251]IQNB as a function of the percentage m2 subtype, for an injection of 3000 nmol of DIBD. The measured data are shown as open circles; the solid lines represent the fits for simulations using (from top to bottom) the parameter values in columns in Table 3 entitled "5 x 103," "1 x 104," "2.5 x 104," "1 x 105," "5 x 107," and "no constraint." The data for the cerebellum are incruded.

(Inmol)

x

imposed

refer to the artificially

lo3

1.47

20.53

5.08

2.99

1338

(2.29 x 104)

x

upper limit for the parameter

1.87

> 7.50 x 104

in (R,R)-[l 251]IQNB.

1.01 x 104

> 2.37 x lo4

headings

x

x 104)

4.87

lo3

105

(1.4 x 106)

1.0 x

(332

Ic50,,3/Ic502

1Column

lo- 3

7.18

103

5.23

the fit to the % decrease

x

(3.6 x 10.3)

6.6

x

2.91

Ic504/Ic502

lo- 2

9.27

x

x

(42 x 10-3)

3.4

(1.74 x 103)

4.95

lo- 2

(3.0 x 10-3)

x

0.274

(1.69 x 103)

103

107

5.77

x

x

(3.2 x 10")

5.0

2.58

4.18

lo- 3

(3.4 x 10.3)

x

02104

ND

>

lo8

103

(1.59 x 103)

x

1.32 x

5.57

ND3

>

108

10-3

(2.0 x 10.3)

cJ2/01/3

(Fig.5)

SSQ4

lo- 2

1022

(nmol)

ho4

x

6.4

3.4

5.7

10-2 (4.0 x 10-3)

x

(3.5 x 10.3)

3.3

(2.9 x 10-3)

640

(nmol)

(nmol)

IC502

ICsoj/3

CT4 (hmol)

02

(1.8 x 10.3)

10-2 1.1

Vnmol)

x

1.2

1.3

0113

10-2

(0.077)2

x

(0.087)

(0.082)

0.274

0.355

0.320 0.273

0.278

(nM/nmol)

SSQ

q

1.0 x 105

5.0 x 107

no constraint1

x

10-2

x

10-2

x

10-3 x

lo4

x

lo3

x

lo3

1.32

5.17

4.98

3.04

1871

2Standard

(2.06 x 104)

6.37

(1.89 x 103)

4.84

(1.14 x 105)

2.50

(4.0 x 10.3)

6.8

(4 5 x 10.3)

3.4

(3.4 x 10.3)

1.1

(0.094)

0.274

0.422

2.5 x 104

x

lo- 2

x

lo- 2

x

lo- 3 x

104

x

lo3 x

lo3

error. 3Not

1.42

2.06

5.02

3.03

2757

determined.

(2.71 x 104)

6.89

(1.74 x 103)

4.85

(2.68 x 104)

1.0

(4.5 x 10.3)

6.7

(5.1 x 10.3)

3.4

(4.1 x 10.3)

1.1

(0.106)

0.274

0.533

1.0 x 104

by the Iteration Procedure

estimate.

SSQ Errors and Parameter Estimates Determined

TABLE 3

x

10-2 x

10‘2 x

lo- 3

103

x

103

4SSQ

1.00

1.00

5.30

2.48

4477

error for

(1.93 x 104)

5.0

(2.43 x 103)

x

r

s

E.

kc F

B N

0

9. 1.

$ 5.0

103

E

x

2 2

(8.16 x 103)

5.0

(4.9 x 10.3)

6.2

(5.4 x 10.3)

3.3

(4.4 x 10.3)

1.3

(0.113)

0.282

0.603

5.0 x 103

1750

b~ Vivo Muscarinic m2 Selective Ligand

Vol. 53, No. 23, 1993

We conclude that the relationship of the regional percentage decrease in (R,R)[1251]IQNB versus the percentage of each of the receptor subtypes indicates that DIBD competes more effectively in those brain regions which are known to be enriched in the m2, relative to the ml, m3, and m4, receptor subtype; however, analysis of the data using a mathematical model shows that caution is required when interpreting the in vivo results. A definitive estimate of the in vivo m2-selectivity awaits the development of [3H]DIBD. A suitably radiolabeled derivative of DIBD may be of potential use in emission tomographic study of the change of m2 receptors in AD. Acknowledaements The authors gratefully acknowledge Dr. J. Baumgold for providing the transfected CHO membranes, and Drs. Wolfe, Eckelman, Gibson, and Baumgold for their comments. This work was supported by a grant from the National Institutes of Health (NS22215) and, in part, a grant from the Department of Energy (DE FG05 88ER60649). References

3 4 5 6 7 8 9 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

20.

D.C. MASH, D.D. FLYNN, and L.T. POTTER, Science 31 1115-1117 (1985) J.X. WANG, W.R. ROESKE, L. MEI, E. MALATYNSKA, W. WANG, E.K. PERRY, R.H. PERRY, and H.I. YAMAMURA, Excerpta Medica, M.J. Rand and C. Raper (eds), 83-86, Elsevier Science Publications, Amsterdam (1987) D.M. ARAUJO, P.A. LAPCHAK, Y. ROBITAILLE, S. GAUTHIER, and R. QUIRION, J. Neurochem. 50 1914-1923 (1988) R. QUIRION, I. AUBERT, P.A. LAPCHAK, R.P. SCHAUM, S. TEOLIS, S. GAUTHIER, and D.M. ARAUJO, Trends Pharmacol. Sci. SUDDI. IV 80-84 (1989) I. AUBERT, D.M. ARAUJO, D. CFCYRE, Y. ROBITAILLE, S. GAUTHIER, and R. QUIRION, J. Neurochem. 58 1529-541 (1992) B.L. HOLMAN, R.E. GIBSON, T.C. HILL, W.C. ECKELMAN, M. ALBERT, and R.C. REBA, J. Am. Med, Assoc. 254 3063-3066 (1985) D.R. WEINBERGER, U. MANN, R.E. GIBSON, R. COPPOLA, D.W. JONES, A.R. BRAUN, K.F. BERMAN, T. SUNDERLAND, R.C. REBA, and T.N. CHASE, Adv. Neurol. 51 147-150 (1990) D.R. WEINBERGER, R.E. GIBSON, R. COPPOLA, D.W. JONES, S. MOLCHAN, T. SUNDERLAND, K.F. BERMAN , and R.C. REBA, Archives Neurol. 48 169-176 (1991) H.-J. KIM, B.R. ZEEBERG, R.E. GIBSON, P. HOSAIN, R. WESLEY, and R.C. REBA, J. Nucl. Med. 31 729 (1990) M. LI, R.P. YASUDA, S.J. WALL, A. WELLSTEIN, and B. WOLFE, Mol. Pharmacol. 40 28-35 (1991) B.R. ZEEBERG, M.S. GITLER, J. BAUMGOLD, R.A. de la CRUZ, and R.C. REBA, Biochem. Biophys. Res. Commun. 17~) 768-775 (1991) W. REGENOLD, D.M. ARAUJO, and R. QUIRION, Synapse 4 115-125 (1989) J.X. WANG, W.R. ROESKE, K.N. HAWKINS, D.R. GEHLERT, and H.I. YAMAMURA, Brain Res. 477 322-326 (1989) E. GIRALDO, R. HAMMER, and H. LADINSKY, Life Sci. 40 833-840 (1987) M.S. GITLER, R.C. REBA, V.I. COHEN, W.J. RZESZOTARSKI, and J. BAUMGOLD, Brain Res. 582 253-260 (1992) S. GAUTHIER, D. CECYRE, I. AUBERT, and R. QUIRION, Soc. Neurosci. Abstr. 16 1060 (1990) W.G. EBERLEIN, W. ENGEL, G. MIHM, K. RUDOLPH, B. WETZEL, M. ENTZEROTH, N. MAYER, and H.N. DOODS, Trends Pharmacol. Sci., Suppl IV 50-54 (1989) W.J. RZESZOTARSKI, W.C. ECKELMAN, B.E. FRANCIS, D.A. SIMMS, R.E. GIBSON, E.M. JAGODA, M.P. GRISSOM, R.R. ENG, J.J. CONKLIN, and R.C. REBA, J. Med. Chem. 27 156160 (1984) V.I. COHEN, W.J. RZESZOTARSKI, R.E. GIBSON, L.H. FAN, and R.C. REBA, J. Pharm. Sci. 78 833-836 (1 989) V.I. COHEN, B. JIN, M.S.GITLER, R. DE LA CRUZ, S.F. BOULAY, B.R. ZEEBERG, and R.C. REBA, submitted to J. Med. Chem.

Vol. 53, No. 23, 1993

21. 22. 23. 24. 25. 26. 27.

In I."ivo Muscarinic m2 Selective Ligand

1751

S.J. WALL, R.P. YASUDA, F. HORY, S. FLAGG, B.M. MARTIN, E.I. GINNS, and B.B. WOLFE Mol. Pharmacol. 39 643-649 (1991) S.J. WALL, R.P. YASUDA, M. LI, and B.B. WOLFE, Mol. Pharmacol. 40 783-789 (1991) R.P. YASUDA, W. CIESLA, L.R. FLORES, S.J. WALL, M. LI, S.A. SATKUS, J.S. WEISSTEIN B.V. SPAGNOLA, and B.B. WOLFE, Mol. Pharmacol. 43 149-157 (1993) D.W. MARQUARDT, J. Soc. Ind. Appl. Math. 11 431-441 (1963) H.J. KIM, B.R. ZEEBERG, and R.C. REBA, IEEE Trans. Med. Imag. 9 247-261 (1990) M.S. GITLER, B.R. ZEEBERG, and R.C. REBA, J. Nucl. Med. 33 883 (1992) J.J. FROST, R.F. DANNALS, T. DUELFER, H.D.BURNS, H.T.RAVERT, B. LANGSTROM, V BALASUBRAMANIAN, and H. WAGNER JR., Ann. Neurol. 15 (suppl.) $85-$92 (1984)