Progress in Medicinal Chemistry — Vol. 41,
`Edited by ED. Ring and A_W. Oxford
`02003 Elsevier Science B.V. All rights reserved.
`
`5 Orally Bioavailable 133-Adrenergic
`Receptor Agonists as Potential
`Therapeutic Agents for Obesity
`and Type-II Diabetes
`
`BAIHUA HU and IRP L. JENNINGS
`
`Chemical Sciences, Wyeth Research, P.O. Box 42528, Philadelphia,
`PA 19101, USA
`
`INTRODUCTION (cid:9)
`
`fia-ADRENERGIC RECEPTOR: STRUCTURE AND ANTI-OBESITY ACTIVITY (cid:9)
`
`BIOLOGICAL ASSAYS (cid:9)
`
`ORALLY BIOAVA/LABLE 133-AR AGONISTS AS THERAPEUTIC AGENTS (cid:9)
`Arylethanolamines (cid:9)
`Aryloxypropanolamines (cid:9)
`Tetrahydroisoquinolines (cid:9)
`
`STATUS OF 133-AR AGONISTS IN DEVELOPMENT (cid:9)
`
`CONCLUSIONS (cid:9)
`
`REFERENCES (cid:9)
`
`167
`
`168
`
`171
`
`172
`173
`184
`187
`
`188
`
`189
`
`189
`
`INTRODUCTION
`
`Although obesity is now recognized as a common medical problem in
`industrialized societies, it remains an inadequately treated disease [I]. It is
`recognized as a major risk factor for serious health complications such as type-fl
`diabetes, high blood pressure, cardiovascular disease, altered steroid meta-
`bolism, and cancers of the breast and uterus. Obesity is estimated to cause
`
`167
`
`SAWAI EX. 1025
`Page 1 of 15
`
`

`

`168 (cid:9)
`
`ORALLY B1OAVAILABLE 03-ADRENERGIC RECEPTOR AGON1STS
`
`B. HU AND L.L. JENNINGS (cid:9)
`
`169
`
`30 million deaths per year in the United States [2]. However, health-care
`professionals generally use drugs to treat the complications of obesity rather than
`the underlying condition because of the small number of treatment options
`available for managing the disease.
`Obesity arises from an imbalance between energy intake and energy
`expenditure. The major life-style factors contributing to an increase in the
`incidence of obesity are an increasingly sedentary lifestyle and increased
`caloric intake. However, clinical studies indicate that genetic factors also
`contribute to the disease. For instance, biochemical and metabolic differences
`between lean and obese individuals have been described calling into question
`the widely held opinion that obesity is modifiable by behavioural changes
`alone [3]. The public health issues associated with obesity justify the
`development of new medications for its treatment. In parallel with the rapid
`evolution of our understanding of the molecular mechanisms that cause
`obesity, there has been a corresponding increase in efforts to discover and
`develop new anti-obesity medications. 133-Adrenergic receptor (133-AR)
`agonists are one of a number of promising categories of drugs that are under
`investigation. For recent reviews, see Refs. [4-10]. This review will focus on
`recent progress in the development of potent, selective and orally bioavailable
`P3-AR agonists for the treatment of diabetes, and more particularly, of obesity.
`
`133-ADRENERGIC RECEPTOR: STRUCTLTRE AND ANTI-OBESITY
`ACTIVITY
`
`As obesity arises from the storage of excess energy, especially in the form of
`triglycerides (TGs), weight reduction requires a period of negative energy
`balance, either by reducing food intake or by increasing energy consumption.
`However, most marketed anti-obesity drugs are appetite suppressants. An
`alternative mechanism for altering body fat composition is through increased
`energy expenditure, either by an increase in physical activity or by accelerating
`the metabolic processing of food and/or fat.
`The 133 receptor is found primarily in adipose tissue, where fat is organized,
`and is known to mediate a variety of metabolic functions, including fat
`mobilization (lip olysis) from white adipose tissue (WAT), increased fat oxidation
`(thermogenesis) in brown adipose tissue (BAT), improved sensitivity to insulin,
`and relaxation of urinary bladder detrusor tissue. (For review on structure and
`function of the 33-AR, see Ref. [11]). A number of recent studies indicate that the
`receptor is present in the human heart, skeletal muscle, gall bladder, gastro-
`intestinal (GI) tract and prostate, in addition to adipocytes [12]. The 133 receptor is
`composed of a single 408 amino acid residue peptide chain that belongs to the
`super family of G-protein-coupled receptors. As expected, it has seven
`
`hydrophobic stretches of about 22-28 residues forming seven a-helical
`transmembrane spanning domains that form the catecholamine-binding site.
`The glycosylated N-terminus is extracellular, whereas the C-terminus is
`intracellular. In contrast to the related 13 and (32 receptors, the C-terminus of
`the 133 receptor contains no serine- and du-eonine-rich regions that act as targets
`for protein kinase A phosphorylation. The absence of phosphorylation sites may
`explain the resistance of the 133 receptor to down regulate following chronic
`stimulation, a feature that distinguishes it from the Pi and 132 receptors. The
`amino acid sequence of the human 133-AR is about 50% identical to that of either
`the human 01 or 132 receptor, respectively [13]. Comparison of the 133 receptor of
`other species with that of human reveals a high degree of sequence homology —
`approximately 80-90% between human, bovine, rodent, and canine. The human,
`monkey, and bovine f33 receptors are more similar to each other than to any of the
`rodent (rat, mouse, and hamster) sequences. The human [33 receptor is distinct
`from the rodent sequences in several segments, a major one being transmembrane
`spanning domain 1 (I'm) Where a (Val-Ala-Leu) deletion was observed in the
`rodents but not in higher species.
`A naturally occurring polymorphism in the amino acid sequence of the P3-AR
`in humans (Trp64Arg) has been identified. Interestingly, this variation 'restores'
`in humans the arginine residue present at this position in animals [11]. This
`mutation has been associated with an increased propensity for weight gain in
`several populations, a feature of insulin resistance and early development of
`type-ll diabetes [14-16]. One functional study on white fat cells showed that the
`mutant receptor is as responsive to the lipolytic effects of the natural ligand
`noradrenaline as the wild-type [17]. However, it is yet to be established whether
`133-AR agonists optimized for the wild-type 03-AR are effective for the
`treatment for obesity in individuals carrying this mutation_
`The role of the 133 receptor in adipocytes is now well understood (Figure 5.1)
`[4-6]. Like the 13, and 132 receptors, the (33 receptor is fully coupled to a
`stimulatory G-protein that activates adenylate cyclase in the plasma membrane
`to generate intracellular cAMP. Measurement of an increase in cAMP levels in
`Chinese hamster ovary (CHO) cells expressing the 133 receptor is, therefore, a
`widely used screening assay for 133 agonists [18-20]. The cAMP so formed
`activates protein kinase A that in turn activates hormone sensitive lipase by
`phosphorylation. The resulting lipase-induced lipolysis converts TGs stored in
`WAT into free fatty acids (FFAs). In brown adipocytes, FFA is oxidized by the
`uncoupling protein 1 (IJCP1) into carbon dioxide and water. UCP1 mediates
`proton transport across the inner initochondrial membrane without generation of
`ATP, thus 'wasting' energy as heat. The overall effect is a loss of fat from the
`body at the expense of more oxygen consumption. Thus, measurement of oxygen
`consumption is the mostly commonly used in vivo model for [33 agonism [21].
`
`SAWAI EX. 1025
`Page 2 of 15
`
`

`

`170 (cid:9)
`
`ORALLY BIOAVAILABLEN-ADRENERGIC RECEPTOR AGON1STS
`
`B. HU AND LL. JENNINGS (cid:9)
`
`171
`
`the expression level of 133 receptors is high enough in humans (at least in young
`lean subjects) to achieve the desired (33 agonist mediated metabolic effects.
`IJCPI, which oxidizes FFA into carbon dioxide and water, is believed to be
`specifically expressed in BAT. This would imply that the thermogenic effect of
`03 agonism would be limited in the body to BAT where UCP1 is found.
`However, two homologues of UCP1 have been recently discovered that are not
`specific for BAT: UCP2, expressed in most tissues at varying levels, and UCP3,
`expressed mainly in skeletal muscle, WAT and BAT. Several studies indicate
`that these UCPs also have proton transport capacity. Given that UCP2 and UCP3
`are highly expressed in adult human tissues, this could mean that tissues other
`than BAT, such as WAT and skeletal muscle, could contribute significantly to
`energy expenditure and fat oxidation on stimulation of 133 receptors.
`Experiments have shown that chronic stimulation of the p, receptor in obese
`animals resulted in reduced adiposity, associated with an increased expression of
`UCP1. 133 Agonists also up-regulate UCP2 and UCP3 in skeletal muscle of obese
`yellow KK mice. These results suggest that the anti-obesity effects of 133 agonists
`are attributable to increased themlogenesis, not only by UCP1, but also by UCP2
`and UCP3 [24-27].
`In addition to their anti-obesity effects, 33 agonists also exert anti-diabetic
`effects, including enhancement of insulin sensitivity and improvement in
`insulin-mediated glucose uptake. Chronic treatment with (33 agonists reduces
`hyperglycemia even at doses that do not pause weight loss. The mechanisms
`underlying the anti-diabetic effect of 133 agonism are currently under
`examination and readers interested in this aspect are referred to a number of
`in-depth discussions [5-7,28].
`
`BIOLOGICAL ASSAYS
`
`With the recognition of the differentiation between the rodent and human
`adrenergic receptors, researchers have come to rely on the use of human-cloned
`133 receptor assays for the identification of 133 agonists [18-20]. E-Alts agonist
`activities are assessed in vitro by measuring the accumulation of cAMP in CHO
`cells expressing human-cloned En 132-, and f31-AR5. The results from these
`functional assays are reported in terms of potency (EC50) and efficacy (or
`intrinsic activity (IA) which is defined as a fraction of the maximal stimulation
`caused by the non-selective full agonist isoproterenol). However, compounds
`with low cAMP functional activity at the Er and 132-ARs may exhibit potent
`antagonist activity that may cause unwanted side-effects [5-10). Thus, binding
`affinities (K1) of the compounds to membranes prepared from CHO cells
`expressing human-cloned 133-, 132-, and [3t-AR5 are determined, and the K, values
`used to assess the selectivity of the agonist or antagonist.
`
`White Adipocytes
`
`Figure 5.1 Proposed mechanism underlying the anti-obesity effect of 133--AR agonists: FFAs, the
`breakdown products of 113-AR mediated lipotysis of white adipocytes, stimulate a thennogenesis
`response in brown adipocytes via the UCP1.
`
`Mice treated with a selective [33 agonist can double oxygen consumption, which
`demonstrates the remarkable capacity of this thermogenic mechanism [22].
`In contrast to pi and [32 receptors, which are primarily localized in the heart
`or on vascular, uterine, or airways smooth muscle, 133-ARs are expressed
`abundantly and predominantly on BAT. The amount of adipose tissue in
`neonates is high relative to that in adults_ However, with increasing age, the
`amount of BAT in lean humans declines, so it has been argued that the amount of
`BAT (and hence the amount of 133 receptors) in adult humans may not be enough
`to produce satisfactory thermogenesis by the activation of 133 receptors.
`However, evidence from a number of studies suggests that BAT can be restored
`in adult humans following chronic treatment with catecholamines. Other studies
`suggest ths t in addition to BAT, skeletal muscle is another tissue where the
`oxidation of FFAs occurs. Skeletal muscle represents up to 40% of total body
`weight and is endowed with significant capacity for thermogenesis. A recently
`reported clinical study demonstrated that treating young lean volunteers with a
`selective 13.3 agonist induced an increase in plasma FFA concentrations, 24 h fat
`oxidation, and stimulated glucose disposal [23]. These new findings suggest that
`
`SAWAI EX. 1025
`Page 3 of 15
`
`

`

`In (cid:9)
`
`ORALLY BLOM/ARABLE B3-ADRENERGIC RECEPTOR AGON1STS
`
`B. HU AND L.L. JENNZIGS (cid:9)
`
`173
`
`A number of in vivo assays have been developed or adapted to assay the anti-
`hyperglycemic, anti-obesity and/or anti-diabetic activity of 133-AR agonists in
`animals [9, 21]. Potent and selectiv6 human p3-AR agonists have usually been
`evaluated in vivo in db/db mice, a model of type-11 diabetes and obesity, for their
`anti-hyperglycemic properties (such as lowering plasma glucose or change in
`TG levels). Another in vivo assay measures changes in metabolic thermogenesis
`by measuring changes in oxygen consumption in transgenic mice expressing the
`human 33-AR. However, the thermogenesis assay proved to have low sensitivity
`and necessitated using high doses. A lipolysis assay that measures the trans-
`formation of TGs to glycerol and FFAs has the advantage of greater sensitivity
`over the thermogenesis model.
`
`ORALLY BIOAVAILABLE 133-AR AGON1STS AS THERAPEUTIC
`AGENTS
`
`CL-316243, BRL-37344, and CGP-12177A (compounds 1-3) are representa-
`tive of the first generation of 133 agonists that were optimized for activity and
`selectivity between 13-AR subtypes by using rodents as a model for the modu-
`lation of adipose tissue in humans [4-10]. These compounds have shown effects
`attributable to 33 receptor stimulation, such as the mobilization of fat from WAT
`deposits, increased thermogenesis, and increased fat oxidation in rodents.
`In addition to their anti-obesity effects, they exhibit potent anti-diabetic effects
`(such as an increase in insulin secretion and improvement in insulin-mediated
`glucose uptake) in the rodent model type-It diabetes. However, human clinical
`trials with these early 133 agonists were disappointing because of a lack of
`selectivity and insufficient anti-obesity effects. In the late 1980s, important
`progress was .made in the cloning and sequencing of the rat and human 133
`receptors. With the human 133-AR now available for the first time, it was soon
`apparent that these early clinical candidates were only partial agonists of this
`receptor and selectivity for the 133-AR over pr and 13 rARs in humans was
`actually a lot lower than that observed in rats. Many groups recognized that a
`cloned human receptor assay would offer major advantages over rodent models
`for the identification and optimization of future 133 agonists. Continued research
`effort led to a number of so-called second-generation compounds that are
`showing promising results in both primates and in humans. A large number of 133
`agonists have been prepared and evaluated, and theie fall basically into three
`structural classes, i.e., arylethanolamines, aryloxypropanolamines, and tetra-
`hydroisoquinolines. In the following discussion, we summarize progress in the
`discovery and optimization of orally bioavailable 133-AR agonists as agents for
`the treatment of obesity and diabetes.
`
`HO
`
`OCH,C001-1
`
`(2) BRL 37344
`
`OH
`
`N1D3d
`
`NH
`. 0
`(3) CGP 12177A
`
`ARYLETHANOLANNES
`
`•
`The phenethanolamine derivatives BRL-26830A (4) and BRL-35135 (6)
`synthesized at Beecham Research Laboratories (now GlaxoSmithKline) were
`the first 133-AR agonists to be examined in rodents. For reviews, see Refs. [29, 30].
`These esters are well absorbed and rapidly metabolized in vivo to the
`corresponding acids. In vitro the acids BRL-28410 (5) and BRL-37344 (2)
`were shown to have potent effects on rat lipolysis (133 effect) and showed
`selectivity over atrial (Pi ) and tracheal (j32) effects. BRL-37344 is the more potent
`and selective agent of the two, exhibiting 400-fold selectivity over Pi and 21-fold
`versus 132. The esters (4) and (6) were evaluated in a number of clinical trials.
`A slightly greater weight loss compared to placebo was observed. However,
`further clinical trials were halted due to poor results and the occurrence of P r and
`132-mediated side-effects.
`
`SAWAI EX. 1025
`Page 4 of 15
`
`

`

`174 (cid:9)
`
`ORALLY B1OAVAILABLE 133-ADRENERGIC RECEPTOR AGONISTS
`
`B. HU AND LL JENNINGS
`
`175
`
`CO2R
`
`(4) R = Me BRL-26830
`
`(5) R. H BRL-28410
`
`OCII2CO2Mc
`
`(6) BRL-3511S
`
`(8)
`
`CI
`
`Researchers at Glaxo (now.GlaxoSmithIGine) explored a series of aniline-
`based phenetbanolamine 133 agonists in the 1990s [31, 32]. The parent
`compound (7) (GR-9803) was found to be a potent full agonist of the human
`133-AR (ECso = 9 nM) but with low selectivity over p, and 132 receptors.
`Varying the size and acidity on the right-hand side of the phenyl substituent of
`(7) led to the acylsulphonarnide derivative (8) and biphenyl derivative (9)
`(GW-2696). Acylsulphonamide (8) has an ECso value of 1 nM and shows
`modest selectivity over the 132 and 132 receptors (500-fold over p, and 60-fold
`over 132.) [31]. Although it has a pharmacoldnetic half-life of less than 2 h, it
`does show low clearance in the dog. The biphenyl analogue is a very potent
`and selective human A, agonist (EC50 = 1 nM, 375-fold over p, and 750-fold
`over 132) [32]. This compound induces no significant stimulation of 131 and 132
`receptors. GW-2696 has a half-life of 4.4 h and 41% bioavailability in the dog.
`In the db/db mice, it reduced glucose levels by at least 50% at a dose of
`10 mg/kg for 1 or 2 weeks (route of administration unknown).
`
`COOH
`
`(7) GA-9803
`
`CL-316243 (1), optimized by the Wyeth group against rodent 133-AR models
`[33], is an extremely potent stimulant of rat BAT lipolysis ([33 effect,
`EC50 = 3 nM) with more than 100,000-fold selectivity for the guinea pig [33
`over the 131 and [32 receptors. Although in early clinical studies the compound
`increased metabolic rate and reduced body weight without eliciting side-effects,
`it had low oral bioavailability, which necessitated high doses (up to 1500 mg).
`A number of prodrugs of CL-316243 were synthesized in an effort to improve
`the oral bioavailability. A 2-3-fold increase in bioavailability was achieved via
`simple alkyl di-esters derivatives [34]. However, no clinical studies were
`conducted on these prodrug forms.
`Typical of 03 agonists optimized for thermogenic activity in the rat, CL-
`316243 was subsequently found to be a weak partial agonist of the human 133
`Eeso = 1.2 }LK 132
`receptor with much reduced potency end selectivity (p3
`= 262 p.M; p, Ecso = 111 ILM). The synthesis and activity of several new
`series of compounds with improved potency and selectivity in the human 133-AR
`have been reported [35-44]. A piperidine analogue (10), possessing a 2,4-
`thiazolidine moiety as a carboxylic acid replacement, was shown to be a potent
`and selective human 133-AR agonist (f33 BC50 = 10 nM, IA = 1.2; >110-fold
`selectivity for I33 over [31 and OD [38]. The therapeutic potential of p3 agonist
`(10) for disorders related to obesity or type-II diabetes was demonstrated in an
`in vivo procedure which compared thermogenesis in human 133-AR transgenic
`mice (Tg mice) with 133-AR knock-out mice (I(O mice). Administered 10 mg/kg
`(i.p.) to fig mice and 1C0 mice compound (10) was active (30 ± 4%
`
`SAWAI EX. 1025
`Page 5 of 15
`
`(cid:9)
`(cid:9)
`

`

`176 (cid:9)
`
`ORALLY BIOAVAILABLE133-ADRENERGIC RECEF7012 AGON1STS
`
`therrnogenesis) in Tg mice and inactive (-2 ± 4% thermogenesis) in KO mice.
`The plasma half-life [intravenous (iv.) 7 h; p.o. 5.7 was acceptable, but the
`bioavailability in rat was low (-3%) probably due to rapid glucuronidation.
`A series of amino acid analogues were synthesized and tested and of these the
`leucine derivative (11) was found to be one of the best [37]. Compound (11) is
`among the most potent and selective human 133 agonists known to date with an
`EC30 of 8 ruM at the 133 receptor, very weak .agonist activity (IA < 10%) or
`antagonist activity (IC1 > 3 p.M) at either the Pt or 02 receptor, and significant
`thermogenesis (53%) effects on human 83 Tg mice. The bioavailability of
`compound (11) was not determined. However, the bioavailability of a closely
`related sulfonamide analogue (12) with similar biological activity (3,
`Eqo = 10 nM, IA = 1.00; thennogenesis 25% in Tg mice) and selectivity
`(>1800 over Pi or 132 receptor) was determined to be 95% in the rat [41].
`
`B. HI) AND L.L. JENNINGS (cid:9)
`
`177
`
`Table 5.1 ORALLY B1OAVAILABLE 113-AR AGONISTS FROM MERCK
`
`Compound (cid:9)
`(Ref) (cid:9)
`
`(13a) [54]
`(13b) [47]
`(13c) [48]
`(134) [50]
`(13e) [51]
`(13f) [52]
`
`EC,
`(nM)
`fa. 944
`
`13 (75)
`3.6(94)
`3.1 (85)
`14 (84)
`8(90)
`23 (53)
`
`13, ICsa (cid:9)
`(nld)
`
`3,400
`2300
`20,000
`4,800
`1,290
`5,500
`
`P2 IC50 (cid:9)
`(nM) (cid:9)
`
`Bioavailablity in (cid:9)
`dogs, F (%) (cid:9)
`
`Halflife
`dogs (h)
`
`1,600
`2,300
`4,700
`1,800
`7,600
`3,000
`
`27
`38
`25
`38
`30
`62
`
`3.6
`13_5
`6
`5
`3.8
`3.6
`
`R =
`
`OCF3
`
`OH
`
`Scientists at Merck reported a series of 133 agonists in which the traditional
`carboxylic acid functionality was replaced with various sulfonamide groups
`[45-61]. Many benzenesulfonamide analogues were identified with potent
`human 133 agonist activity, excellent selectivity against Pt and 82 receptors, and
`
`SAWAI EX. 1025
`Page 6 of 15
`
`(cid:9)
`(cid:9)
`(cid:9)
`

`

`178 (cid:9)
`
`ORALLY BIOAVAILA13LE prADRENERGIC RECEPTOR AGONISTS
`
`B. LIU AND L.L. JENNINGS (cid:9)
`
`179
`
`improved pharmacokinetic profile [47-58]. Tetrazolone benzenesulfonamide
`analogue (13a), L-770644, is a potent and selective human 133 agonist (133
`EC.30 = 13 nM, IA = 0.75; 131 EC50= 1.9 p,M, IA = 0.33; R2 EC50 = 1.8 P,M,
`IA = 0.26) [54]. It has 27% oral bioavailability in both dog and rat and a half-life
`of 3.6 bin the dog. Moreover, is. administration to-monkeys evokes full agonist
`activity for hyperglycerolemia with an BD50 value of 0.21 mg/kg. The
`compound has minimal effect on heart rate at the highest dose of 10 mg/kg.
`Concurrent studies by Merck have examined a number of tetrazolone
`replacements, including thiazole (13b) [47], triazole (13c) [48], oxazole (13d)
`[50], and oxadiazoles (13e) [51] and (130 [52]. These have been found to
`possess the desired combination of potency, selectivity, and oral bioavailability
`(Table 5.1). For example, 4-(trifiuoromethyl)phenyl substituted thiazole
`analogue (13b) is a potent (EC50= 3.6 nM) and selective (>600-fold) full
`agonist with 38% orally bioavailability in dogs and an exceptionally long half-
`life of 13.5 h [47]. In a rising dose i.v. infusion study by using anaesthetized
`rhesus monkeys, the compound elicited hyperglycerolemia with an ED50 value
`of 0.26 mg/kg. Furthermore, compound (13b) is more than 100-fold selective as
`a 133 agonist when tested against a panel of receptors and ion channels. On the
`basis of this favourable activit}, and safety profile, compound (13b) was chosen"
`for Phase I clinical trials [47].
`Researchers at Bristol-Myers Squibb have also disclosed a series of
`phenethanolamines as potential anti-obesity and anti-diabetic agents that were
`optimized using a cloned human receptor assay [62-66]. Introduction of
`substituents at the a-position of the phenethylamine moiety of the original
`lead (14) produced the 3,4-dimethoxyphenyl analogue (15) (3MS-189665)
`with potent 133 agonist activity (K; = 33 riM) but with limited selectivity over
`131 and 112 receptors (69-fold and 8-fold, respectively). Further effort
`investigating the effect of substituents on the right-band side phenyl groups
`gave (16) (BMS-194449) [62,65]. This is a potent 133 agonist (K; = 160 nM)
`with good selectivity over 131 and 132 receptors. The compound was evaluated
`as a lipolytic agent in vivo in primates (African green monkey) and was
`found to increase FFA concentrations at a dose of 0.5 mg/kg i.v. but did not
`increase the heart rate (131 effect) or significantly decrease potassium levels at
`this dose (112 effect). However, the oral bioavailability of the compound was
`found to be less than 2% in the rat, primarily due to extensive glucuronida-
`lion of both the phenolic and ll-hydroxy groups at the intestinal wall and in
`the liver.
`Further modification of BMS-194449 by means of chain homologation
`produced another clinical candidate (17) (13MS-196085) [63,66]. BMS-196085
`is more potent (K; = 21 nM) and selective than BMS-194449. In vivo, BMS-
`196085 increased FFA concentrations in plasma of African green monkeys at
`a dose of 0.1 mg/kg, i.v. and it increased lipolysis at a dose of 0.02 mg/kg.
`
`In a 2-week is. study in Swiss Webster mice, BMS-196085 decreased TG
`levels in a dose-dependent manner. The glucose lowering effect of BMS-
`196085 was studied in ob/ob mice, where it normalized glucose at a dose of
`3.4 mg/kg LS,. Unfortunately, the oral bioavailability of BlvIS-196085 was
`found to be less than 5%, primarily because of extensive first pass meta-
`bolism. And, even though BMS-196085 was a weak partial agonist (IA = 0.4)
`for the 3, receptor, it increased heart rate to 26 b.p.m. at doses of 0.1 and
`0.5 mg/kg i.v.
`In an effort to improve its selectivity profile, the effect of adding or replacing
`substituents on both the left-hand and right-hand side phenyl groups of BMS-
`196085 was investigated [64]. Replacement of the para-hydroxyl group on
`the left-band side phenyl with H, F, or Cl or replacement of the Methyl
`sulfonamide group with NHCOMe, OH, NHCHO, CO2H, CO2Me, NHCONHz,
`or NHCO2Me resulted in a marked loss of IA. On the other hand, introduction of
`a meta-chloro substituent on the right-hand side phenyl provided (18) (EMS-
`210285), which is a highly selective agonist for the 133 receptor. As a pre-clinical
`backup candidate, this compound was reported to be superior to BMS-196085 in
`terms of functional selectivity in vitro and in vivo in primates. No further
`information is available.
`
`OH
`
`F10
`
`(14)12=H
`(15)R ..3,4-(Me0)2Ph BMS-189665
`
`OCHF2
`
`OCRF2
`(16) BMS-194.149
`
`SAWAI EX. 1025
`Page 7 of 15
`
`

`

`180 (cid:9)
`
`ORALLY BIOAVAILABLE 03-ADRENERGIC RECEPTOR AGONISTS
`
`B. HU AND L.L. JENNINGS
`
`181
`
`OCHF2
`
`(17) R = H BMS-146085
`(18) R= Cl BMS-2I0285
`
`An indole derivative (19) was identified as a lead (Ea EC50 = 12 nM, a full
`agonist) by workers at Dainippon using the human B3-AR expressed in a stable
`CHO cell line to assay derivatives of ethanolamine [67]. Modification of the
`indole ring led to the orally active 133-AR agonist (20), AD-9677 (or /0-9677 or
`TAK-677). It is a potent and selective human 133 agonist (133 EC50 = 0.062 nM,
`IA = 1.2; I3 EC50 = 6.4 nM, IA = 0.26; Pa EC50 = 13 nM, IA = 0.26).
`Chronic treatment of genetically obese diabetic ICK-Ay mice and db/db mice
`(0.01-1.0 mg/kg/day p.o.) with AD-9677 for 14 days significantly reduced
`plasma glucose, non-esterified fatty acid (NEFA), and TG levels. AD-9677
`corrected hyperinsulinemia and inhibited weight gain in KK-Ay mice at doses
`over 0.1 mg/kg. In db/db mice, however, AD-9677 had no effect on body weight
`at a lower dose. This indicates that the anti-diabetic effect of AD-9677 is not a
`result of body weight reduction. In an oral glucose tolerance test, AD-9677 also
`improved glucose utilization and insulin response in KK-Ay mice [68]. The
`improved insulin response may be caused by a decrease in protein levels of
`TNF-ce and by an increase in FFA production associated with increased UCP1
`expression and lipolysis [68]. Based on its anti-diabetic effects at low doses and
`anti-obesity effects at relatively higher doses, AD-9677 might be a new class of
`agent for obesity and diabetes.
`
`A patent from Asald claimed tricyclic compounds particularly suitable for
`oral administration in the treatment of obesity and diabetes [69]. Carbazole and
`dibenzothiophene derivatives (21) and (22) were reported to be highly potent,
`subnanomolar 133 agonists.
`
`HO
`
`HO
`
`In the early 1990s, Sanofi-Midy (now Sanofi-Synthélabo) identified a series
`of phenylethanolaminotetralins (PEATs) as selective ft3 agonists in rodents.
`SR58611A (23) was found to be the most potent and selective 113 agonist in a
`series of close analogues and stereoisomers [70-72]. Although SR58611A is
`about 6-fold less potent than isoproterenol, it is a full agonist and it demonstrates
`the expected lipolytic and thermogenic responses in animals, hi an acute glucose
`tolerance test (0.25 mg/kg i.p.) in normal lean and spontaneously obese/diabetic
`CBA/Ca mice, SR58611A was effective in reducing the blood glucose response.
`Chronic treatment (0.25 mg/kg i.p. for 15 days) improved insulin response, but
`had no effect on body weight or food intake. At higher doses in obese mice,
`lipogenesis in both brown and WAT was increased.
`
`SAWAI EX. 1025
`Page 8 of 15
`
`(cid:9)
`(cid:9)
`

`

`182 (cid:9)
`
`ORALLY BIOAVAILABLE f3q-ADRENERG1C RECEPTOR AGONISTS
`
`B. HU AND LL. JENNINGS (cid:9)
`
`183
`
`intra-bladder pressure (ED30 = 31 p.g/kg i.v.) without increasing heart rat. The
`phenoxy acetic acid analogue (27), KUL-7211, was reported to be orally
`bioavailable and to have a long half-life in dogs [77-79]. It showed ED50 of
`0.72 nM in male ferret bladder and dose dependently suppressed KCI-induced
`contraction in isolated human uteri. Recently, it has been shown that 133-ARs
`exist in the human detrusor and it has been suggested its relaxation is mainly
`under p3-AR controL Consequently, KUL-7211 may provide a new approach for
`the treatment of urinary bladder dysfunction [81].
`OH
`
`HO
`
`•
`
`1
`Me
`(26) R -= Me
`(27) R. H, KUL-7211
`
`R4—CO2H
`
`Compound (28) (N-5984), a benzodioxan carboxylic acid analogue of CL-
`316243, was discovered by the Kyorin group [82]. The selectivity, potency, and
`efficacy of N-5984 on human p-adrenergic receptors were examined using both
`binding and cAMP CHO cell functional assays. The binding affinity of N-5984
`for (33-AR is 0.03 p.M, which is 220-fold more potent than that of CL-316243.
`Ki values for 131- and 02-adrenergic receptors were 2.2 and 0.6 p.M, respectively.
`In the CHO cell assay, it had an EC30 value of 1.7 nM for the 133-AIL Low
`intrinsic activities (<10%) were observed for both the (cid:9)
`and f32-ARs. These
`results indicate that N-5984 is a potent and selective P3-AR agonist, although it
`is only a partial agonist (IA = 0.52).
`
`CI
`
`(23)SR-58611A
`
`Researchers at Pfizer have disclosed a series of aminopyridylethanolamines
`as 133-adrenergic agonists. The indole analogue (24) (CP-331679) was identified
`as a potent human 133 agonist (EC.50 = 30011M based on maximal response
`induced by isoprotemnol, which itself has an EC50 of 2 1.r.M) [73]. It is a full
`agonist and is more than 100-fold selective over human 131 and 132 receptors.
`However, subsequent studies showed CP-331679 and its ethyl ester to have poor
`oral bioavailability in rats. Replacements to the indole ring system were
`investigated to improve oral bioavailability and the phenyl acetic acid analogue
`(25) (CP-331684) was identified as a novel, orally active 133 agonist. It is a
`moderately potent (ECso = 400 nM) and selective agonist (>25-fold over PI
`and 132 receptors) for the human (33 receptor. However, it is only a partial agonist
`(IA = 0.8) [74]. More pm-clinical evaluation of CP-331684 is currently ongoing
`[75, 76].
`
`HiN
`
`OH
`
`\ 0
`
`Me
`(24) CP-33I679
`
`OH
`
`Me
`(25) CP-331684
`
`00311
`
`Investigators at Kissei have disclosed a series of phenoxyacetic acid
`derivatives as selective stimulants of the (33-AR [77-80]. In vitro, a
`representative compound (26) was found to be a selective 133 agonist in
`functional assays using the ferret detrusor (133-AR EC30= 7.8 nM), rat uterus
`(132-AR Ms° = 7.3 liM) and rat atrium (Pi-AR EC50 = 23 1.1.1v1) [80]. In an in
`vivo study, i.v. administration of compound (26) in the anaesthetized rat lowered
`
`Trecadrine (29), which was originally developed as an anti-ulcer agent, has
`also been revealed to be a 133 agonist [83-86]. It induces relaxation of rat
`oesophageal muscular mucosae, a tissue used for characterizing 133 agonist
`activity. Additionally, trecadrine does not appear to have 131 (in guinea-pig
`isolated auricle) or 132 (in guinea-pig isolated trachea) adrenergic receptor
`activity. The oral administration of trecadrine decreased glucose and TG levels
`and increased adipose tissue oxygen consumption [84, 85]. These results
`
`SAWAI EX. 1025
`Page 9 of 15
`
`

`

`184 (cid:9)
`
`ORALLY BIOAVAILABLE133-ADRENERGIC RECEPTOR AGON1STS
`
`B. 1111 AND L.L. JENNINGS (cid:9)
`
`185
`
`strengthen the case for the investigation of trocadrine as an anti-obesity and anti-
`diabetic agent in humans.
`
`Ph
`
`Me
`
`I
`Me
`(29) Trecadrine
`
`ARYLOXYPROPANOLAMINES
`
`As a dais, aryloxypropanolamines are antagonists or partial agonists at [31 and
`132 receptors and these properties, therefore, determine the selectivity of their
`in vivo biological effects. Researchers at Lilly have reported a series of amides as
`anti-obesity/