J. Med. Chem. 2006, 49, 7843-7853
`
`7843
`
`Synthesis and Structure-Activity Relationships of 3,8-Diazabicyclo[4.2.0]octane Ligands, Potent
`Nicotinic Acetylcholine Receptor Agonists
`
`Jennifer M. Frost (ne´e Pace),*,† William H. Bunnelle,† Karin R. Tietje,† David J. Anderson,† Lynne E. Rueter,† Peter Curzon,†
`Carol S. Surowy,† Jianquo Ji,† Jerome F. Daanen,† Kathy L. Kohlhaas,† Michael J. Buckley,† Rodger F. Henry,† Tino Dyhring,‡
`Philip K. Ahring,‡ and Michael D. Meyer†
`Neuroscience Research, Abbott Laboratories, Abbott Park, Illinois 60064, and NeuroSearch A/S, 93 PederstrupVej,
`DK-2750 Ballerup, Denmark
`
`ReceiVed July 19, 2006
`
`A series of potent neuronal nicotinic acetylcholine receptor (nAChR) ligands based on a 3,8-diazabicyclo-
`[4.2.0]octane core have been synthesized and evaluated for affinity and agonist efficacy at the human high
`affinity nicotine recognition site (hR4(cid:226)2) and in a rat model of persistent nociceptive pain (formalin model).
`Numerous analogs in this series exhibit picomolar affinity in radioligand binding assays and nanomolar
`agonist potency in functional assays, placing them among the most potent nAChR ligands known for the
`hR4(cid:226)2 receptor. Several of the compounds reported in this study (i.e., 24, 25, 28, 30, 32, and 47) exhibit
`equivalent or greater affinity for the hR4(cid:226)2 receptor relative to epibatidine, and like epibatidine, many
`exhibit robust analgesic efficacy in the rat formalin model of persistent pain.
`
`Introduction
`
`Neuronal nicotinic acetylcholine receptors (nAChRs) have
`become exciting new targets for medicinal research. One of the
`most studied areas of interest for this receptor family recently
`has been that of analgesia, but the nAChRs have also been
`investigated as potential
`targets to treat Alzheimer’s and
`Parkinson’s diseases, schizophrenia, and depression.1 The
`analgesic effects of nicotine (1) were first reported in 1932.2
`The much later discovery that the extremely potent antino-
`ciceptive activity of epibatidine (2), isolated from the skin of a
`poisonous Ecuadorian tree frog, was the result of the interaction
`of epibatidine with nAChRs renewed interest in selectively
`targeting these receptors for the treatment of pain.3 While
`epibatidine was found to be an extremely potent analgesic, its
`poor side effect profile at or near effective analgesic doses
`(paralysis, seizures, death, etc.) precluded its development for
`clinical use.3 These side effects of epibatidine are thought to
`stem in large part from the activity of epibatidine at
`the
`ganglionic and neuromuscular nAChRs.4 Analogs of epibatidine
`and nicotine with improved side effect profiles have been sought.
`One such compound discovered in our labs is ABT-594 (3;
`Figure 1).5
`Compound 3 is a potent nAChR agonist that is active in a
`broad range of preclinical models of nociceptive and neuropathic
`pain.5,6 Compound 3 is more selective for neuronal nAChRs
`(vs ganglionic and neuromuscular nAChRs) than is epibatidine
`and does have an improved therapeutic profile in in vivo models
`relative to epibatidine. However, 3 exhibits only modest
`selectivity among the neuronal nAChR subtypes (Table 1).7a
`Generally, nAChR subtype selectivity is thought to be key to
`improving the therapeutic margin between analgesia and side
`effects (i.e., gastrointestinal and cardiovascular effects).8,16
`Specifically, it is thought that agonist activity at the R4(cid:226)2
`receptor subtype, found mainly in the CNS, is responsible for
`the observed analgesic activity,9 while activity at the R3(cid:226)4
`
`* To whom correspondence should be addressed. Tel.: 847-937-0721.
`Fax: 847-937-9195. E-mail:
`jennifer.frost@abbott.com.
`† Abbott Laboratories.
`‡ NeuroSearch A/S.
`
`Figure 1.
`
`subtype, abundantly expressed in the peripheral nervous system,
`is the main cause of side effects.10
`Compound 3 (ABT-594) has been the starting point for a
`variety of more rigid structural variants making use of an NCCX
`structural motif (bolded in Structure A) where X has been N,
`O, C, and S (Scheme 1).11 One such series is the 3,8-
`diazabicyclo[4.2.0]octane series that includes regioisomers B
`(3-N regioisomeric series) and C (8-N regioisomeric series). The
`synthesis and pharmacological profile of this series will be
`discussed herein.
`
`Chemistry
`All of the R4(cid:226)2 agonists described in this paper were
`generated by the Buchwald-Hartwig coupling12 of the ap-
`propriate enantiomerically pure diamine with a halopyridine
`(Scheme 2). The diamine syntheses and the general coupling
`procedures are discussed below.
`The synthesis of the chiral diamines 4 and 5 and their
`enantiomers (6 and 7, respectively) began as shown in Scheme
`3. The commercially available ethyl 1-benzyl-3-oxo-4-piperidine
`carboxylate hydrochloride (8) was first converted into the
`corresponding tert-butyl carbamate 9 via removal of the
`N-benzyl protecting group, followed by reaction of the resulting
`free amine with di-tert-butyl dicarbonate. A solution of 9 and
`(R)-R-methylbenzylamine in toluene was refluxed under a
`Dean-Stark trap to generate the enamine 10. Reduction of the
`double bond with sodium triacetoxyborohydride13 yielded an
`approximately 1.5:1 mixture of the (3S,4S)-cis-isomer, 11, and
`the (3R,4R)-cis-isomer, 12. At this point, the stereochemistry
`of the products was assumed based on literature precedence for
`the predominance of the cis-isomer. The cis-configuration was
`confirmed by X-ray analysis of the enantiomer of the reduced
`product, 14 (ent-14, see Experimental Section). This mixture
`
`10.1021/jm060846z CCC: $33.50 © 2006 American Chemical Society
`Published on Web 12/21/2006
`
`Downloaded via UNIV OF PENNSYLVANIA on January 29, 2021 at 03:35:16 (UTC).
`
`See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
`
`Liquidia's Exhibit 1022
`IPR2020-00770
`Page 1
`
`

`

`7844 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26
`
`Frost (ne´e Pace) et al.
`
`[3H]-cytisine
`pKi ( SEM
`9.02 ( 0.07
`10.38 ( 0.04
`10.32 ( 0.09
`
`[3H]-cytisine
`Ki (nM)
`0.96
`0.042
`0.048
`
`hR4(cid:226)2 EC50 (nM)
`(SEM range)
`6600 (5900-7400)
`54 (48-61)
`23 (20-26)
`
`hR4(cid:226)2 max
`101 ( 3%
`133 ( 13%
`115 ( 3%
`
`hR3(cid:226)4 EC50 (nM)
`(SEM range)
`8000 (7400-8600)
`12 (9.9-14)
`188 (160-221)
`
`hR3(cid:226)4 max
`89 ( 3%
`103 ( 15%
`169 ( 12%
`
`Table 1.
`
`nicotine (1)
`epibatidine (2)
`ABT-594 (3)
`
`Scheme 1
`
`Scheme 2
`
`Scheme 3
`
`Scheme 5
`
`Scheme 4
`
`was then reduced with lithium aluminum hydride to afford the
`corresponding alcohols, 13 and 14.
`Initially, the two cis-isomers, 13 and 14, were separated by
`arduous flash column chromatography and were carried on
`separately. In Scheme 4, this initial route is elaborated using
`isomer ent-13, which was formed as described above, but using
`(S)-R-methylbenzylamine. It should be noted that when (S)-R-
`methylbenzylamine was used, the (3R,4R)-isomer predominated,
`whereas use of (R)-R-methylbenzylamine resulted in larger
`amounts of the (3S,4S)-isomer. As shown in Scheme 4, the
`R-methylbenzyl group was cleaved via hydrogenolysis to give
`amine 15. Reaction with 2-nitrobenzenesulfonyl chloride gave
`the intermediate bisnosylate, which upon treatment with 5%
`
`aqueous NaOH in ethanol, cyclized to sulfonamide 16.14 At this
`point,
`the nitrobenzenesulfonyl group was removed with
`14 to give diamine core 4.
`thiophenol and K2CO3
`Alternatively, to obtain isomer 5, which allowed coupling at
`N3, the t-butylcarbamate was cleaved, and the resulting amine
`was protected as the benzyl carbamate 17. The N-nitrobenzene-
`sulfonyl group was then removed14 to give the free amine, which
`was reprotected in situ to give the corresponding t-butylcar-
`bamate. Hydrogenation then resulted in the free amine isomer
`5.
`
`While this route was effective in generating initial quantities
`of core diamines 4 and 5, as well as their enantiomers (6 and
`7), the separation of diastereomers 13 and 14 (and ent-13 and
`ent-14) was difficult, and the protecting group manipulations
`were cumbersome. It was thought that if alcohols 13 and 14
`could be cyclized directly, prior to removal of the chiral
`auxiliary, the presence of said auxiliary on the rigid, cyclized
`product might facilitate separation of the two cis-isomers.
`Furthermore, this route would allow for fewer protecting group
`manipulations.
`Three primary questions needed to be addressed with the
`proposed route. First, could the alcohol oxygen be selectively
`activated in the presence of the basic, acyclic nitrogen? Second,
`if the selective O-activation was possible, would the resulting
`intermediate cyclize in the presence of the large R-methylbenzyl
`substituent? Finally, would this route facilitate the separation
`of the two cis-isomers?
`To address the first question, single isomer 13 was treated
`with methanesulfonyl chloride and pyridine in CH2Cl2 (Scheme
`5). Mesylation occurred exclusively on oxygen, and furthermore,
`attempted purification of the monomesylate via silica gel
`chromatography resulted in isolation of both the mesylate 19
`and the cyclized product 18. Ultimately, it was found that the
`cyclized product 18 could be obtained directly by addition of
`Cs2CO3 to the mesylation reaction mixture.
`
`Liquidia's Exhibit 1022
`IPR2020-00770
`Page 2
`
`

`

`3,8-Diazabicyclo[4.2.0]octane Ligands
`
`Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26 7845
`
`Scheme 6
`
`Scheme 7
`
`Scheme 8
`
`Scheme 9
`
`As it had now been demonstrated that both selective activation
`and the subsequent cyclization were favorable, the only remain-
`ing question was that of improving the separation of the
`diastereomers. To that end, a mixture of isomers 13 and 14 was
`treated with methanesulfonyl chloride and Et3N in tetrahydro-
`furan (Scheme 6). After the mixture had stirred for 1 h at
`ambient temperature, Cs2CO3 was added and the temperature
`was raised to 60 (cid:176)C. After stirring at 60 (cid:176)C for 18 h, it was
`found that the two products 18 and 20 had readily formed and
`that they were easily separable via flash column chromatogra-
`phy. The individual isomers such as 18 could then be converted
`to either the 3-N accessible (7) or the 8-N accessible cores (6),
`as shown in Scheme 7. Diamine 6 was obtained directly by
`hydrogenolysis of 18, while 7 was isolated after a protecting
`group shuffle to give the trifluoroacetamide 21 followed by
`hydrogenation, protection of the resulting free amine as the
`t-butylcarbamate, and removal of the trifluoroacetyl group. Thus,
`this new route allowed the more efficient generation of large
`quantities of the diamine cores (4-7) than was possible with
`the initial synthesis.
`
`Figure 2.
`
`As discussed above, the diamine cores were coupled to
`numerous halopyridines via the Buchwald-Hartwig coupling
`(Scheme 8).12 Coupling of the diamine core was also carried
`out using 5-bromopyrimidine and 3,6-dichloropyridazine (Scheme
`9).
`
`Results and Discussion
`
`Two main structural series were investigated and are de-
`scribed herein:
`that in which the pyridine is attached to the
`diamine at the 8-position (8-N-substituted; structure C, Figure
`2) and that in which the pyridine is attached to the diamine at
`the 3-position (3-N-substituted; structure B, Figure 2). All
`compounds shown in Tables 2-4 were evaluated for affinity
`to the R4(cid:226)2 binding site in the cytisine binding assay and for
`functional activity at the hR4(cid:226)2 and hR3(cid:226)4 subtypes in the
`FLIPR cellular assays. Selected compounds were evaluated for
`analgesic activity in phase 2 of the formalin flinch assay.
`Background SAR. The structure-activity relationship (SAR)
`of the 3,8-diazabicyclo[4.2.0]octane series was primarily ex-
`plored by varying the small substituents on the 5 and 6 positions
`of the pyridine ring (R1 and R2 in structures B and C,
`respectively). Previous work by ourselves and others had
`suggested that small groups at these positions of the pyridine
`were optimal for R4(cid:226)2 ligands.15,16 Larger groups at
`the
`5-position led to ligands with high affinity but reduced functional
`activity.16,17 Ligands with large (i.e., an aromatic group)
`substitution at the 6-position have reduced cytisine binding and
`R4(cid:226)2 functional activity.15 Substitutions at the 2 or 4 positions
`generally resulted in reduced binding and functional activi-
`ties.15,16,18 Attachment of the diamine moiety at the 3-pyridyl
`position has also been found to be optimal.19 In general, the
`pyridine ring was preferred, but the 5-pyrimidinyl and 3-pyrid-
`azinyl were also investigated in this and other series.16,20 Finally,
`in most cases studied, an unsubstituted nitrogen on the diamine
`moiety (RdH) was required for activity at the R4(cid:226)2 receptor.16,21
`SAR of the 3,8-Diazabicyclo[4.2.0]octane Ligands. As a
`group, the 3,8-diazabicyclo[4.2.0]octane ligands are among the
`most potent and efficacious compounds reported for the
`nAChRs. Several compounds in this series show nanomolar
`potency in the hR4(cid:226)2 functional assay (i.e., 28, 30, 32, and 47)
`and most have subnanomolar potency in the cytisine binding
`assay. Many of these ligands exhibit supramaximal efficacy in
`the functional assays with responses on the order of 200% that
`of the maximal nicotine response. Several compounds (i.e., 24,
`25, 28, 30, 32, and 47) reported here are as potent or more potent
`than epibatidine (2) in these assays. The specific effects on in
`vitro and in vivo activity with changes in pyridine substitution
`as well as diamine regio- and stereoselectivity are discussed
`below.
`The most potent compounds reported in this study are from
`the 3-N-substituted isomeric series (Table 2). For example, the
`(1R,6S)-5,6-dibromo analog, 30, has a Ki of 0.014 nM in the
`rat cytisine binding assay and EC50 values of 7.8 nM (209%)
`and 8.2 nM (117%) in the hR4(cid:226)2 and hR3(cid:226)4 functional assays,
`respectively. As shown in Table 3, substrates in the 8-N-
`
`Liquidia's Exhibit 1022
`IPR2020-00770
`Page 3
`
`

`

`7846
`
`Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26
`
`Frost (ne´e Pace) et al.
`
`Table 2. In Vitro Biological Activity of 3-N-Substituted nAChRs
`
`cmpd
`22
`
`23
`
`24
`
`25
`
`26
`
`27
`
`stereoisomer
`
`R1
`
`1R,6S
`
`1S,6R
`
`1R,6S
`
`1S,6R
`
`1R,6S
`
`1S,6R
`
`H
`
`H
`
`H
`
`H
`
`Br
`
`Br
`
`R2
`H
`
`H
`
`Cl
`
`Cl
`
`H
`
`H
`
`[3H]-cytisine
`pKi ( SEM
`10.46 ( 0.08
`
`10.23 ( 0.07
`
`10.72 ( 0.08
`
`10.50 ( 0.08
`
`10.61 ( 0.05
`
`10.28 ( 0.03
`
`hR4(cid:226)2 EC50
`(nM)
`(SEM range)
`
`[3H]-cytisine
`Ki (nM)
`0.035
`
`0.059
`
`0.019
`
`0.032
`
`0.024
`
`0.053
`
`hR3(cid:226)4 EC50
`(nM)
`(SEM range)
`
`hR4(cid:226)2 max
`171 ( 8%
`
`112 ( 5%
`
`114 ( 2%
`
`130 ( 2%
`
`148 ( 7%
`
`58 ( 2%
`
`hR3(cid:226)4 max
`121 ( 4%
`
`118 ( 2%
`
`151 ( 10%
`
`125 ( 2%
`
`122 ( 5%
`
`91 ( 4%
`
`132 ( 7%
`
`28
`
`29
`
`30
`
`31
`
`32
`
`33
`
`34
`
`1R,6S
`
`1S,6R
`
`1R,6S
`
`1S,6R
`
`1R,6S
`
`1S,6R
`
`1R,6S
`
`Cl
`
`Cl
`
`Br
`
`Br
`
`CH3
`
`CH3
`
`CN
`
`Cl
`
`Cl
`
`Br
`
`Br
`
`Cl
`
`Cl
`
`H
`
`10.61 ( 0.06
`
`10.57 ( 0.04
`
`10.86 ( 0.03
`
`10.62 ( 0.08
`
`10.51 ( 0.03
`
`10.44 ( 0.07
`
`9.92 ( 0.12
`
`0.025
`
`0.027
`
`0.014
`
`0.024
`
`0.031
`
`0.037
`
`0.12
`
`221 ( 15%
`
`79 ( 5%
`
`209 ( 14%
`
`86 ( 3%
`
`207 ( 15%
`
`90 ( 6%
`
`112 ( 6%
`
`76 ( 6%
`
`111 ( 6%
`
`117 ( 2%
`
`100 ( 3%
`
`123 ( 2%
`
`103 ( 3%
`
`127 ( 8%
`
`116 ( 2%
`
`71
`(61-82)
`330
`(300-370)
`13
`(11-15)
`24
`(22-26)
`150
`(130-170)
`1000
`(850-1200)
`12
`(9.7-14)
`82
`(67-100)
`7.8
`(6.1-9.9)
`76
`(67-87)
`7.2
`(5.4-9.6)
`47
`(38-58)
`1900
`(1200-2900)
`1900
`(1500-2400)
`130
`(120-160)
`2700
`(2000-3600)
`2200
`(1900-2600)
`4980
`(4330-5730)
`620
`(570-680)
`1350
`(300-6200)
`650
`(500-830)
`100
`(94-110)
`110
`(94-120)
`21
`(18-25)
`360
`(260-480)
`6.1
`(5.0-7.5)
`180
`(160-210)
`
`54 ( 4%
`
`196 ( 7%
`
`114 ( 6%
`
`217 ( 12%
`
`40 ( 3%
`
`215 ( 19%
`
`194 ( 19%
`
`48
`(44-52)
`260
`(250-280)
`29
`(26-32)
`25
`(24-27)
`76
`(70-82)
`950
`(910-990)
`9.9
`(8.3-12)
`110
`(100-120)
`8.2
`(7.4-9.1)
`110
`(100-120)
`7.2
`(6.5-7.9)
`71
`(68-74)
`1000
`(590-1700)
`490
`(460-520)
`180
`(170-190)
`4500
`(4400-4700)
`1100
`(1040-1140)
`2400
`(2300-2450)
`124
`(117-131)
`2800
`(2700-2900)
`1400
`(1300-1500)
`390
`(360-410)
`830
`(780-880)
`8.0
`(7.3-8.8)
`160
`(150-180)
`4.4
`(3.5-5.6)
`5700
`(5300-6100)
`
`119 ( 3%
`
`98 ( 4%
`
`120 ( 2%
`
`108 ( 3%
`
`133 ( 3%
`
`99 ( 3%
`
`43
`
`44
`
`45
`
`46
`
`47
`
`48
`
`1R,6S
`
`1S,6R
`
`1R,6S
`
`1S,6R
`
`1R,6S
`
`1S,6R
`
`CH3
`
`H
`
`H
`
`OMe
`
`OMe
`
`CN
`
`H
`
`NO2
`
`NO2
`
`Br
`
`Br
`
`Br
`
`H
`
`10.19 ( 0.07
`
`8.83 ( 0.07
`
`8.61 ( 0.08
`
`10.57 ( 0.02
`
`10.16 ( 0.04
`
`10.57 ( 0.05
`
`10.21 ( 0.07
`
`0.064
`
`1.5
`
`2.4
`
`0.027
`
`0.069
`
`0.027
`
`0.062
`
`substituted series are also relatively potent but generally have
`lower affinities at the tested receptors than the 3-N-substituted
`regioisomers. In all cases where direct comparisons between
`the two regioisomeric series can be made (i.e., 22 vs 49, 25 vs
`52, etc.), the 3-N-substituted regioisomers are the more potent
`in hR4(cid:226)2 binding and functional assays.
`Within the 3-N-substituted regioisomeric series (Table 3),
`there is a pronounced stereochemical effect, with the 1R,6S
`absolute stereochemistry generally imparting higher binding
`
`affinity and greater functional potency than the 1S,6R enan-
`tiomers, as exemplified by compounds 22 and 23, 30 and 31,
`and 36 and 37. In addition to greater potency, compounds with
`the 1R,6S stereochemistry are characterized by significantly
`greater agonist efficacy in the hR4(cid:226)2 functional assay, with
`many analogs exhibiting 200% of the maximal response elicited
`by nicotine (see 28, 30, 32, 43, 47). By contrast, compounds of
`the 1S,6R stereochemistry generally exhibit maximal agonist
`efficacies comparable to nicotine.
`
`35
`
`36
`
`37
`
`38
`
`39
`
`40
`
`41
`
`42
`
`1S,6R
`
`1R,6S
`
`1S,6R
`
`1R,6S
`
`1S,6R
`
`1R,6S
`
`1S,6R
`
`1S,6R
`
`CN
`
`OMe
`
`OMe
`
`H
`
`H
`
`OEt
`
`OEt
`
`H
`
`H
`
`H
`
`OMe
`
`OMe
`
`H
`
`H
`
`10.19 ( 0.08
`
`10.42 ( 0.11
`
`9.74 ( 0.04
`
`9.01 ( 0.08
`
`8.57 ( 0.058
`
`10.28 ( 0.06
`
`9.80 ( 0.01
`
`0.065
`
`0.038
`
`0.18
`
`0.98
`
`2.7
`
`0.052
`
`0.16
`
`178 ( 17%
`
`35 ( 4%
`
`96 ( 6%
`
`54 ( 1%
`
`116 ( 3%
`
`18 ( 4%
`
`103 ( 3%
`
`86 ( 2%
`
`88 ( 4%
`
`93 ( 2%
`
`103 ( 1%
`
`86 ( 2%
`
`87 ( 2%
`
`Liquidia's Exhibit 1022
`IPR2020-00770
`Page 4
`
`

`

`3,8-Diazabicyclo[4.2.0]octane Ligands
`
`Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26 7847
`
`Table 3. In Vitro Biological Activity of 8-N-Substituted nAChRs
`
`cmpd
`49
`
`50
`
`51
`
`52
`
`53
`
`54
`
`55
`
`56
`
`stereoisomer
`
`R1
`
`1R,6S
`
`1S,6R
`
`1R,6S
`
`1S,6R
`
`1R,6S
`
`1S,6R
`
`1R,6S
`
`1S,6R
`
`1R,6S
`
`H
`
`H
`
`H
`
`H
`
`Cl
`
`Cl
`
`CH3
`
`CH3
`
`OMe
`
`[3H]-cytisine
`pKi ( SEM
`9.91 ( 0.10
`
`8.95 ( 0.02
`
`9.92 ( 0.08
`
`9.10 ( 0.04
`
`9.93 ( 0.08
`
`9.52 ( 0.06
`
`10.37 ( 0.08
`
`9.62 ( 0.03
`
`8.85 ( 0.16
`
`R2
`H
`
`H
`
`Cl
`
`Cl
`
`Cl
`
`Cl
`
`Cl
`
`Cl
`
`Br
`
`hR4(cid:226)2 EC50
`(nM)
`(SEM range)
`
`[3H]-cytisine
`Ki (nM)
`0.12
`
`1.1
`
`0.12
`
`0.80
`
`0.12
`
`0.3
`
`0.043
`
`0.24
`
`1.4
`
`hR4(cid:226)2 max
`105 ( 2%
`
`84 ( 3%
`
`94 ( 6%
`
`102 ( 6%
`
`60 ( 2%
`
`127 ( 3%
`
`103 ( 6%
`
`142 ( 9%
`
`65 ( 4%
`
`hR3(cid:226)4 EC50
`(nM)
`(SEM range)
`
`hR3(cid:226)4 max
`98 ( 4%
`
`97 ( 3%
`
`129 ( 3%
`
`93 ( 9%
`
`69 ( 2%
`
`103 ( 2%
`
`91 ( 3%
`
`103 ( 4%
`
`84 ( 3%
`
`410
`(380-440)
`1400
`(1300-1500)
`37.4
`(33.6-41.5)
`216
`(181-258)
`271
`(252-291)
`176
`(164-190)
`76.1
`(67.3-86.0)
`134
`(116-156)
`1690
`(1500-1900)
`200
`(187-214)
`2190
`(1890-2540)
`1930
`(1590-2350)
`109
`(96.2-124)
`172
`(155-192)
`1490
`(1320-1680)
`340
`(317-365)
`7690
`(6010-9580)
`1105
`(890-1370)
`
`120 ( 2%
`
`64 ( 4%
`
`83 ( 3%
`
`104 ( 4%
`
`157 ( 5%
`
`75 ( 4%
`
`133 ( 6%
`
`44 ( 5%
`
`104 ( 4%
`
`1700
`(1600-1700)
`1540
`(1460-1630)
`1390
`(1270-1530)
`1520
`(1290-1790)
`4110
`(3930-4290)
`339
`(316-364)
`1910
`(1790-2030)
`298
`(276-322)
`2980
`(2810-3150)
`202
`(184-220)
`2560
`(2180-3000)
`5560
`(5220-5920)
`794
`(757-832)
`238
`(226-252)
`23300
`(22500-24200)
`7460
`(7170-7770)
`63500
`(47400-85200)
`4360
`(4140-4580)
`
`98 ( 4%
`
`91 ( 8%
`
`71 ( 4%
`
`88 ( 2%
`
`100 ( 3%
`
`73 ( 2%
`
`90 ( 1%
`
`47 ( 1%
`
`88 ( 2%
`
`57
`
`58
`
`59
`
`60
`
`61
`
`62
`
`63
`
`64
`
`65
`
`66
`
`1S,6R
`
`1R,6S
`
`1S,6R
`
`1R,6S
`
`1S,6R
`
`1R,6S
`
`1S,6R
`
`1R,6S
`
`1S,6R
`
`OMe
`
`CN
`
`CN
`
`CN
`
`CN
`
`C(O)NH2
`
`C(O)NH2
`
`OMe
`
`OMe
`
`Br
`
`H
`
`H
`
`Br
`
`Br
`
`Br
`
`Br
`
`H
`
`H
`
`10.07 ( 0.060
`
`9.30 ( 0.05
`
`8.58 ( 0.15
`
`10.15 ( 0.01
`
`9.52 ( 0.01
`
`8.55 ( 0.04
`
`8.08 ( 0.02
`
`8.36 ( 0.08
`
`9.25 ( 0.02
`
`0.085
`
`0.50
`
`2.6
`
`0.070
`
`0.3
`
`2.8
`
`8.4
`
`4.4
`
`0.56
`
`In the 8-N-substituted series (Table 4), many of the com-
`pounds investigated showed little difference in affinity between
`the 1R,6S and 1S,6R isomers. However, in the cases of the most
`potent compounds, which includes the unsubstituted compounds
`(49 vs 50) and many 6-halogenated compounds (51 vs 52, 55
`vs 56, and 61 vs 62), the 1R,6S isomers exhibited higher
`affinities than their 1S,6R counterparts. Of the compounds
`investigated, only the 1S,6R isomers with 5-methoxy groups
`showed increased affinity relative to the corresponding 1R,6S
`isomers (57 vs 58 and 65 vs 66).
`For the sake of clarity, the discussion of the effects of pyridine
`substitution will be limited to the most potent stereochemical
`and regiochemical series, specifically, the (1R,6S)-3-N-substi-
`tuted series. The SAR trends exhibited in this series are generally
`consistent among the other regio- and stereochemical series.
`As discussed above, previous work had suggested that for
`R4(cid:226)2 nAChR activity, small substituents on the 5 and 6 positions
`of the pyridine ring (R1 and R2, respectively) were optimal and
`larger substitutions at the 2 or 4 positions were detrimental.15-17
`The unsubstituted pyridine, 22, is a very potent compound at
`both R4(cid:226)2 (Ki ) 0.035 nM, EC50 ) 70.7 nM) and R3(cid:226)4 (EC50
`) 47.9 nM). However, as has been previously reported, halogens
`in the 6-pyridyl position lead to increased potency at nAChRs,16
`and this trend was found to be true in the case of the
`3,8-diazabicyclo[4.2.0]octane ligands as well. The analogs with
`
`6-halogen substitution (24, 28, 30, 32, 45, and 47) all exhibit
`subnanomolar binding affinities and low nanomolar potencies
`in the functional assays. None of these compounds is selective
`for R4(cid:226)2 over R3(cid:226)4 nAChRs, with the exception of 24, which
`shows only a 2-fold preference for R4(cid:226)2 over R3(cid:226)4. Other
`substitutions in the 6-pyridyl position include the methoxy (38)
`and nitro (43) moieties, both of which lead to decreased activity
`in the binding and functional assays relative to 22.
`Ligand 26 is the only case in the stereo- and regiochemical
`series being discussed that has 5-halo substitution without
`accompanying 6-halo substitution. Although the binding affini-
`ties are comparable between 5-H derivative 22 and the 5-bromo
`derivative 26, compound 22 is a more potent agonist in the R4(cid:226)2
`and R3(cid:226)4 nAChR functional assays. A comparison of the 5-H,
`6-Cl ligand 24 with 28 and 30, the 5,6-diCl and 5,6-diBr ligands,
`respectively, demonstrates that the presence of halogens in the
`5 and 6 positions of the pyridine is not detrimental to R4(cid:226)2 or
`R3(cid:226)4 activity.
`Substitution at the 5-pyridyl position without a concomitant
`6-halo group (34, 36, 40) led to reduced R4(cid:226)2 and R3(cid:226)4
`activities relative to the 5,6-unsubstituted ligand, 22. However,
`with 6-halo substitution, several groups were well tolerated in
`the 5-pyridyl position (32, 45, 47), with R4(cid:226)2 and R3(cid:226)4
`activities equal to or better than the unsubstituted ligand, 22.
`Overall, substitutions on the pyridine ring could have a large
`
`Liquidia's Exhibit 1022
`IPR2020-00770
`Page 5
`
`

`

`7848
`
`Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26
`
`Frost (ne´e Pace) et al.
`
`Table 4. In Vitro Biological Activity of Pyrimidine and Pyridazine nAChRs
`
`cmpd
`67
`
`stereoisomer
`
`[3H]-cytisine
`pKi ( SEM
`8.81 ( 0.07
`
`[3H]-cytisine
`Ki (nM)
`1.6
`
`hR4(cid:226)2 EC50
`(nM)
`(SEM range)
`
`8420
`(7440-9530)
`
`hR4(cid:226)2 max
`104 ( 4%
`
`hR3(cid:226)4 EC50
`(nM)
`(SEM range)
`
`9940
`(9360-10600)
`
`hR3(cid:226)4 max
`93 ( 2%
`
`68
`
`69
`
`70
`
`71
`
`1R,6S
`
`1S,6R
`
`1R,6S
`
`1S,6R
`
`10.39 ( 0.05
`
`9.71 ( 0.07
`
`0.040
`
`0.20
`
`350
`(293-418)
`4390
`(3620-5330)
`
`181 ( 10%
`
`79 ( 3%
`
`108
`(102-114)
`1020
`(984-1070)
`
`121 ( 2%
`
`106 ( 2%
`
`9.83 ( 0.06
`
`8.57 ( 0.18
`
`0.15
`
`2.7
`
`78.3
`(65.8-93.1)
`359
`(311-414)
`
`222 ( 17%
`
`129 ( 10%
`
`64.8
`(57.6-73.0)
`394
`(369-420)
`
`128 ( 6%
`
`113 ( 4%
`
`Figure 3. Dose responses of 28 and 36 in the rat formalin model. *The asterisks denote statistical significance.
`
`effect on potencies in the in vitro assays but, in this series, they
`do not appear to have a substantial effect on selectivity between
`R4(cid:226)2 and R3(cid:226)4 subtypes. These same general trends are also
`apparent in the (1S,6R)-3-N-substituted series and the 8-N-
`substituted series.
`As shown in Table 4, heterocycles other than pyridine were
`also briefly investigated. The two 3-N-substituted isomers (68
`and 69) and one 8-N-substituted isomer (67) of the 3,8-
`diazabicyclo[4.2.0]octane pyrimidines studied had reduced
`potency in the functional assays relative to their pyridine analogs
`(67 vs 50, 68 vs 22, and 69 vs 23). These compounds were
`also not subtype-selective (hR4(cid:226)2 vs hR3(cid:226)4) nor were they
`active in the rat formalin screen. Because of these results, the
`pyrimidines were not investigated further. The two stereoisomers
`of
`the 3-N-substituted diazabicyclo[4.2.0]octane-6-chloro-
`pyridazine (70 and 71) were made.
`In vitro,
`these two
`compounds were less potent than their 6-chloropyridine analogs,
`24 and 25. The (1R,6S)-isomer, 70, was still relatively potent
`and did show some activity in vivo, but overall, offered no
`advantages over the pyridine analogs.
`In Vivo Results. The 3-N-substituted series exhibits a good
`correlation between activity in the in vitro assays and activity
`
`in the rat formalin model, although it should be noted that while
`our functional assays use the human R4(cid:226)2 and R3(cid:226)4 receptors,
`our in vivo studies were performed in rats. For example, the
`1R,6S enantiomer 28 has good activity in the R4(cid:226)2 FLIPR assay
`(EC50 ) 11.8 nM, 221% response) and, as shown in Figure 3,
`has a good dose-response curve in the formalin model while
`its enantiomer 29, which was weaker in the in vitro assays (R4(cid:226)2
`EC50 ) 82.1 nM, 79% response) exhibited only modest activity
`in the in vivo model (55% reduction in flinches at 19 (cid:237)mol/
`Kg, data not shown). The same is true for ligands 30 (99%
`reduction in flinches at 6.2 (cid:237)mol/Kg) and 31 (26% reduction
`in flinches at 19 (cid:237)mol/Kg), as well as 34 (80% reduction in
`flinches at 6.2 (cid:237)mol/Kg) and 35 (no reduction in flinches at 19
`(cid:237)mol/Kg). The relatively less potent analog 36 (R4(cid:226)2 EC50 )
`134 nM, 178% response) also elicited a good dose-response
`in the formalin model (Figure 3), while its enantiomer 37 (R4(cid:226)2
`EC50 ) 2670 nM, 35% response) was inactive in the in vivo
`model.
`Overall, in the 8-N-substituted series there appears to be less
`of a correlation between in vitro and in vivo results. The (1R,6S)-
`6-chloro analog 51 is more potent in the R4(cid:226)2 FLIPR assay
`(EC50 ) 37.4 nM, 94% response) than its enantiomer 52 (R4(cid:226)2
`
`Liquidia's Exhibit 1022
`IPR2020-00770
`Page 6
`
`

`

`3,8-Diazabicyclo[4.2.0]octane Ligands
`
`Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26 7849
`
`Experimental Section
`
`Biological Assays. Rat Cytisine Binding Assay. Binding to a
`desensitized state of nAChRs (predominantly R4(cid:226)2) was evaluated
`by measuring displacement of [3H]-cytisine from rat brain homo-
`genate (n g 3).23
`Functional Assays. HEK cell lines expressing the R4(cid:226)2 and
`R3(cid:226)4 subunit combinations were used in the determination of
`functional nAChR agonist activity by measuring intracellular
`calcium changes using the fluorometric imaging plate reader
`(FLIPR) system (Molecular Devices, Sunnydale, CA). The cell line
`was obtained from NeuroSearch (Ballerup, Denmark). Cells were
`plated at densities of 25000-50000 cells/well in DMEM (GIBCO),
`supplemented with 10% FBS (GIBCO) in 96-well, clear-bottom,
`black-walled plates (Corning Costar) manually precoated with poly-
`D-lysine (Sigma, 75 (cid:237)L/well of 0.01 g/L solution g 30 min) and
`allowed to incubate for 24-48 h at 37 (cid:176) C in 5% CO2 in a humidified
`environment. After aspirating off the media, the cell lines were
`incubated in the dark at room temperature for (cid:24)0.75-1 h with
`2-4 (cid:237)M Fluo-4 AM calcium indicator dye (Molecular Probes,
`Eugene, OR) dissolved in 0.1 to 0.2% v/v of DMSO (Sigma, U.K.)
`in NMDG ringer buffer (in mM: 140 NMDG, 5 KCl, 1 MgCl2, 10
`HEPES, 10 CaCl2, pH ) 7.4). Cells were placed in the FLIPR and
`50 (cid:237)L of 3(cid:2) stock concentrations of test compounds or buffer only
`prepared in the same NMDG ringer buffer were added. Raw
`fluorescence data were corrected by subtracting fluorescence values
`from wells that received buffer only additions. Peak fluorescent
`values were determined over the range of drug exposure using
`FLIPR software and expressed as a percentage of the reference
`peak response for the positive control of 100 (cid:237)M nicotine and
`exported for analysis using Microsoft Excel and GraphPad Prism
`(San Diego, CA). Data were fitted using a single sigmoidal function
`in GraphPad determining EC50 and maximum responses and
`expressed as means ( SEM(n), SEM is the standard error of the
`means and an n ) 6 constitutes two replicates per plate across three
`plates.
`Rat Formalin Model of Persistent Pain. Following a 30 min
`habituation period to the testing room and cages, rats were injected
`i.p. (1 mL/kg) with either the test compound or its vehicle control.
`Five minutes later, 50 (cid:237)L of a 5% formalin solution was injected
`subcutaneously into the dorsal aspect of one of the hind paws.
`Immediately after formalin injection, the cages were placed on a
`suspended rack with mirrors positioned below permitting the
`experimenter to observe the rat from all angles. From 30 to 50
`min after the injection of formalin, there is a marked increased in
`nocifensive behaviors such as flinching, licking, and biting of the
`injected paw, an increase which has been termed phase 2 or the
`persistent pain phase of the model. During this 20 min period, rats
`were observed for the occurrence of nocifensive behaviors. Four
`rats were run simultaneously, and the experimenter observed each
`rat for one 15 s observation period during each 1 min interval
`throughout the 20 min of phase 2. The nocifensive responses were
`recorded and summed for statistical analyses.8
`Chemistry. Proton NMR spectra were obtained on a General
`Electric QE 300 or QZ 300 MHz instrument with chemical shifts
`((cid:228)) reported relative to tetramethylsilane as an internal standard.
`Elemental analyses were performed by Robertson Microlit Labo-
`ratories. Column chromatography was carried out on silica gel 60
`(230-400 mesh). Thin-layer chromatography was performed using
`250 mm silica gel 60 glass-backed plates with F254 as indicator.
`The X-ray crystal structures were obtained on a Bruker SMART
`system. All materials were commercially available and were
`obtained from Aldrich unless otherwise specified.
`3-Oxo-piperidine-1,4-dicarboxylic Acid 1-tert-Butyl Ester
`4-Ethyl Ester (9). A mixture of commercially available ethyl-N-
`benzyl-3-oxo-4-piperidinecarboxylate hydrochloride (8; 75.4 g, 0.25
`mol), di-t-butyl dicarbonate (58.5 g, 0.27 mol), Et3N (36 mL, 0.26
`mol), and Pd(OH)2/C (7.5 g, 50% in H2O) in 660 mL of EtOH
`was put under 60 psi of H2 and was shaken for 25 min. The mixture
`was then filtered, and the filtrate was concentrated under reduced
`pressure to provide the title compound, which was used in the next
`
`Figure 4. Dose response of 52 in the rat formalin model. *The asterisks
`denote statistical significance.
`
`EC50 ) 216 nM, 102% response). Likewise, 51 exhibits 50%
`reduction in flinches in the rat formalin model at low doses
`(0.19 (cid:237)mol/Kg), while 52 is less potent (Figure 4). In contrast,
`although 55 is more potent at R4(cid:226)2 (EC50 ) 76.1 nM, 103%
`response) than its enantiomer 56 (R4(cid:226)2 EC50 ) 134 nM, 142%
`response), 55 is inactive in the rat formalin model, while 56
`causes a 75% reduction in the number of flinches at 19 (cid:237)mol/
`Kg (data not shown). Overall, the 8-N-substituted series is not
`as active in the in vivo model as the 3-N-substituted series,
`which is consistent with the reduced in vitro activity of the 8-N-
`substituted series relative to the 3-N-substituted series.
`It should be noted that while many of the tested compounds
`were shown to be active in the rat formalin model, most active
`compounds also exhibited side effects, which included prostra-
`tion, seizures, ataxia, and dyspnea. While the behavioral side
`effects of some compounds likely impacted the ability of the
`animals to flinch, they cannot solely account for the observed
`analgesic effects. For example, multiple compounds showed
`mild and/or transient side effects that were not apparent at the
`time of analgesia testing, while others continued to show
`analgesic efficacy when tested at lower doses that did not induce
`behavioral side effects (i.e., 30, 33, 36, and 42). Interestingly,
`the opposite could also be demonstrated, that is, administration
`of several of the reported compounds resulted in side effects
`without any accompanying analgesia at
`the tested doses
`(maximum of 19 (cid:237)mol/Kg; i.e., 44, 49, 53, 55, 61, and 71).
`Thus, analgesic effects and behavioral side effects were disso-
`ciable for these analogs, as has been previously described for
`epibatidine, which also exhibits significan