Substituted 4-Carboxymethylpyroglutamic Acid Diamides as Potent and Selective Inhibitors
`of Fibroblast Activation Protein
`
`Ting-Yueh Tsai,† Teng-Kuang Yeh,† Xin Chen,† Tsu Hsu, Yu-Chen Jao, Chih-Hsiang Huang, Jen-Shin Song, Yu-Chen Huang,
`Chia-Hui Chien, Jing-Huai Chiu, Shih-Chieh Yen, Hung-Kuan Tang, Yu-Sheng Chao, and Weir-Torn Jiaang*
`
`Division of Biotechnology and Pharmaceutical Research, National Health Research Institutes, No. 35, Keyan Road, Zhunan Town, Miaoli
`Country 350, Taiwan, R.O.C. †These authors contributed equally to this work.
`
`Received February 26, 2010
`
`Fibroblast activation protein (FAP) belongs to the prolyl peptidase family. FAP inhibition is expected
`to become a new antitumor target. Most known FAP inhibitors often resemble the dipeptide cleavage
`products, with a boroproline at the P1 site; however, these inhibitors also inhibit DPP-IV, DPP-II,
`DPP8, and DPP9. Potent and selective FAP inhibitor is needed in evaluating that FAP as a therapeutic
`target. Therefore, it is important to develop selective FAP inhibitors for the use of target validation. To
`achieve this, optimization of the nonselective DPP-IV inhibitor 8 led to the discovery of a new class of
`substituted 4-carboxymethylpyroglutamic acid diamides as FAP inhibitors. SAR studies resulted in a
`number of FAP inhibitors having IC50 of <100 nM with excellent selectivity over DPP-IV, DPP-II,
`DPP8, and DPP9 (IC50 > 100 μM). Compounds 18a, 18b, and 19 are the only known potent and selec-
`tive FAP inhibitors, which prompts us to further study the physiological role of FAP.
`
`Introduction
`Fibroblast activation protein (FAPa) is a type II transmem-
`brane serine protease belonging to the prolyl peptidase family,
`which comprises serine proteases that cleave bioactive pep-
`tides preferentially after proline residues. Members of this
`family include dipeptidyl peptidase-IV (DPP-IV), DPP-II
`(DPP7), DPP8, and DPP9, and this family has been impli-
`cated in several diseases, including diabetes, cancer, and mood
`disorder.1,2 FAP is expressed on reactive stromal fibroblasts in
`over 90% of common human epithelial cancers,3 in granula-
`tion tissue of healing wounds, and in bone and soft tissue
`sarcomas.4,5 It has been suggested that FAP promotes tumori-
`genesis and that FAP inhibition may attenuate tumor
`growth.6-8 The development of potent and specific inhibitors
`for each of these DPP enzymes can be an important tool to
`study the physiological function and to validate their potential
`as a therapeutic target. Specific inhibitors for DPP-IV,9-11
`DPP-II,12,13 and DPP8/914 have been identified, and investi-
`gations of selective inhibitors for FAP have only started
`recently.15-17
`Dipeptidyl peptidase IV (DPP-IV, also known as CD26)
`(EC 3.4.14.5) is a drug target for type II diabetes. The active
`form of glucagon-like peptide-1 (GLP-1) stimulates insulin
`secretion, inhibits glucagons release,18,19 and slows gastric empty-
`ing,20-22 each a benefit in the control of glucose homeostasis
`
`*To whom correspondence should be addressed. Phone: 886-37-
`246166, extension 35712. Fax: 886-37-586456. E-mail: wtjiaang@nhri.
`org.tw.
`a Abbreviations: FAP, fibroblast activation protein; DPP, dipeptidyl
`peptidase; GLP-1, glucagon-like peptide-1; LiHMDS, lithium hexam-
`ethyldisilazide; DBU, 1,8-diazabicyclo[5.4.0]undec-7-ene; EDC, 1-ethyl-3-
`(3-dimethylaminopropyl)carbodiimide; TFA, trifluoroacetic acid; LC-MS,
`liquid chromatography coupled mass spectrometry; SAR, structure-
`-activity relationship; HOBt, N-hydroxybenzotrizole.
`
`in patients with type 2 diabetes.23,24 Therefore, inhibition of
`DPP-IV prolongs the action of GLP-1, which offers a new
`strategy for treating type 2 diabetes. Antidiabetic efficacy has
`been demonstrated clinically with DPP-IV inhibitors; sita-
`gliptin 1 (MK-0431) was approved by the U.S. Food and
`Drug Administration for the treatment of type 2 diabetes,10
`and vildagliptin 2 (LAF237) was approved for use in the
`European market (Figure 1).9
`Dipeptidyl peptidase II (DPP-II, EC 3.4.14.2) is believed to
`be involved in the physiological breakdown of some proline-
`containing neuropeptides and in the degradation of collagen
`and substance P.25,26 Using 1-[(S)-2,4-diaminobutanoyl]pipe-
`ridine as lead compound, Senten et al. developed a series of
`potent and selective DPP-II inhibitors.12,13 The representative
`DPP-II inhibitor 3 (Figure 1) showed IC50 = 0.23 nM and a
`high selectivity toward DPP-IV (IC50 = 345 μM, Table 1).13
`Our studies found that compound 3 was inactive toward FAP,
`DPP8, and DPP9 (Table 1). These selective DPP-II inhibi-
`tors are outstanding tools to determine the physiological
`function of DPP-II and the therapeutic potential of DPP-II
`inhibitors.
`Dipeptidyl peptidase 8/9 (DPP8/9) are cytoplasmic pro-
`teases with a 51% homology in amino acid level with DPP-
`IV.27 Previously, the administration of selective DPP8/9 inhi-
`bitor 4 (Figure 1) may be associated with profound toxicities
`in preclinical species, which included alopecia, thrombocyto-
`penia, anemia, enlarged spleen, multiple histological patho-
`logies, and animal mortality shown in rats.28 Highly specific
`and potent DPP8/9 inhibitors were also developed by Jiaang
`et al. The representative DPP8/9 inhibitor 5 (also called 1G-
`244, Figure 1) had IC50 values of 14 and 53 nM against DPP8
`and DPP9, respectively (Table 1).14 It did not inhibit DPP-IV,
`FAP, or DPP-II, with IC50 values greater than 100 μM. Re-
`cently, by using this potent and selective DPP8/9 inhibitor,
`
`pubs.acs.org/jmc
`
`Published on Web 08/18/2010
`
`r 2010 American Chemical Society
`
`6572 J. Med. Chem. 2010, 53, 6572–6583
`DOI: 10.1021/jm1002556
`
`Downloaded via GEORGE WASHINGTON UNIV on March 18, 2025 at 23:38:57 (UTC).
`
`See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
`
`Petitioner GE Healthcare – Ex. 1047, p. 6572
`
`

`

`Wu et al. show that the inhibition is not associated with any
`animal toxicity.29
`
`F
`
`F1?VNH2 0
`
`I ~ ~NT~ 'N
`
`F
`
`l~N-z
`
`1
`
`CF3
`
`Cl
`
`1Crt1 0
`
`H2N Y
`
`0
`
`OH
`
`){;i ~-·,l~N
`
`0
`
`2
`
`0
`
`v):\{J
`
`3
`
`4
`
`Ft}.,t::J•'
`
`(NJ
`
`X r\
`
`H2N yN--{
`0
`/B'OH
`HO
`
`6
`
`:::)r,£(
`
`O
`
`CN
`
`r)J~NTT/Q
`
`\
`
`~-----
`
`B,
`0 HO OH
`
`Previously, N- and RC-substituted Gly-boro-Pro deriva-
`tives have been developed, and these compounds strongly
`inhibited DPP-IV, DPP7 (DPP-II), and FAP. In this class, a
`representative compound is Val-boroPro 6 (Figure 1), having
`IC50 of <40 nM against all the three DPPs (Table 1).15 On the
`basis of FAP’s preference for Gly-Pro-based endopeptidase
`substrates, Tran et al. developed a series of N-acyl-Gly-,
`N-acyl-Sar- (sarcosine), and N-blocked-boroPro derivatives.
`Representative compound in this class was inhibitor 7
`(Figure 1), which preferentially inhibited FAP versus DPP-
`IV, but the selectivity against DPP-II, DPP8, and DPP9 was
`not reported (Table 1).16 In our studies, this compound had
`weak inhibitory activity against DPP8 and DPP9 (11 and
`6.5 μM, respectively) and was inactive toward DPP-II (Table 1).
`In this paper, we report a systematic search for potent and
`selective FAP inhibitors. A structure-activity relationship
`was investigated starting from L-homoglutamic acid 8 (Figure 2),
`a dual DPP-V and FAP inhibitor (IC50 = 10 and 54 nM, re-
`spectively), as a lead compound (Table 1).30 To develop selec-
`tive FAP inhibitors, introduction of ring constraint in the P2
`portion of lead 8 and modification of the P2 site secondary
`amine to amide were done as depicted in Figure 2. The replace-
`ment of secondary amine with amide is based on the fact that
`FAP acts as both prolin-specific endopeptidase and dipeptidyl
`peptidase, but DPP-IV prefers to display the latter activ-
`ity.8,16,31 Further exploration of the 2-position pyrrolidine
`derivatives (P1 site) and the 4-position amine substituents at
`the carbonylmethyl group (P2 site) led to the discovery of
`potent pyroglutamic acid-based FAP inhibitors with a high
`selectivity for FAP over DPP-IV, DPP-II, DPP8, and DPP9.
`
`5
`
`7
`
`Chemistry
`
`Figure 1. DPP-IV inhibitors 1 and 2, DPP-II inhibitor 3, DPP8/9
`inhibitors 4 and 5, nonselective inhibitor 6, and FAP inhibitor 7.
`Table 1. Potency and Selectivity of Compounds 1-8
`
`Reference compounds N-acyl-Gly-boroPro 9 and N-acyl-
`Gly-cyanoPro 10 were prepared according to the literature
`
`IC50 (μM)a
`
`compd
`
`name
`
`FAP
`
`DPP8
`
`DPP9
`
`DPP-II
`
`DPP-IV
`
`referenceb
`
`this study
`this study
`13
`this study
`this study
`this study
`15
`16
`this study
`this study
`
`1
`2
`3
`
`4
`5
`6
`7
`
`sitagliptin
`vildagliptin
`DPP-II selective
`
`DPP8/9 selective
`1G-244
`nonselective
`FAP selective
`
`0.023
`>100
`>100
`>100
`>100
`0.056
`>50
`1.2
`14
`>100
`345
`0.00023
`ND
`ND
`ND
`>100
`0.005
`>100
`>100
`>100
`>100
`>100
`0.24
`0.15
`>100
`>100
`>100
`0.053
`0.014
`>100
`0.0003
`0.038
`ND
`ND
`0.011
`23
`ND
`ND
`ND
`0.008
`8.0
`>100
`6.5
`11
`0.12
`0.010
`2.6
`0.089
`0.36
`0.054
`1G-409
`8
`a Mean values of at least three experiments; standard deviations are (20%. ND, no data. b “This study” means in-house data.
`
`S2 pocket
`
`DPPN
`
`SI pocket
`
`Design
`=====)
`
`P2
`
`Pl
`
`Lead compound 8
`
`Ring constrained analogue
`
`Figure 2. Design of 4-carboxymethylpyroglutamic acid diamides as FAP inhibitors.
`
`S2 pocket
`
`SAR Study
`
`Sl pocket
`
`w = 1° or 2° amine
`R 1 = Pyrrolidine analogues
`4-Carboxymethylpyroglutamic acid diamides
`
`Article
`
`Journal of Medicinal Chemistry, 2010, Vol. 53, No. 18
`
`6573
`
`Petitioner GE Healthcare – Ex. 1047, p. 6573
`
`

`

`Scheme 1. Synthesis of C(4)-Substituted Pyroglutamate Ester Urethanesa
`
`0
`r'\_(OH ~ ~ _(\- p Bn
`OBn
`b
`l_)
`0-:J-- / ' b + Br
`o?-- r/ \b - ----- 0-)-_/---<o + Br~
`H
`H
`Boe
`11
`12
`
`13
`
`14
`
`0
`
`'<
`
`y O,
`/ \ ~-
`0
`
`C ------
`
`_,,,.
`, __ \
`
`>',,,0 '-,,--0 --,,J ____ H
`pBn / \ II
`/ ~ \
`cF -- ~
`'o
`o ;?--- ~
`Boe
`Boe
`cis-15
`trans-15
`
`+
`
`~OBn
`
`o
`
`d
`a Reagents: (a) DIEPA, CH2Cl2; (b) (Boc)2O, 4-DMAP, CH2Cl2; (c) LHMDS, THF, -78 °C; (d) DBU, CH2Cl2, 0 °C to room temp.
`
`Scheme 2. Synthesis of FAP Inhibitors 17-22, 24, 27, and 30a
`
`R1 ,R2
`
`X0T',,r\.-'.c(
`o,;J--{b N
`Boe
`18
`
`d, b
`
`------
`
`17 R1, R2= H
`18 R1 = F, R2 • H
`19 R1 , R2= F
`
`20 R1,R2=H
`21 R1 = F, R2 • H
`22 R1, R2= F
`
`0
`/ \ y
`
`oBn
`0
`
`~
`0
`
`~
`Boe
`15
`
`X =S orCH2
`
`23
`
`d,b
`
`w - 1° or 2° amine
`w
`r
`
`~
`
`(" x
`N_j
`
`O
`
`~
`
`0
`
`24a X=S
`24b X= CH2
`
`w - 1° or 2° amine
`
`~oZ K
`
`N
`
`__ d,_b_► ((>JL K
`
`OJ_ ~~
`OJ_ /~,O
`I
`Boe
`30
`29
`a Reagents: (a) H2, Pd/C, MeOH; (b) EDC,HOBt, CH2Cl2/1,4-dioxane, various amines; (c) POCl3, imidazole, pyridine; (d) TFA, CH3CN, 0 °C to
`room temp; (e) BCl3.
`
`HN :o
`
`N
`
`procedures with some modifications.9,16 (2S)-Cyanopyrroli-
`dine analogues 17-19, 24, 27, 30, and 31 were prepared as des-
`cribed in Schemes 1 and 2 and are listed in Tables 2 and 3. The
`synthesis (Scheme 1) proceeded in good yield following the
`procedure of Young and co-workers to afford the fully protec-
`ted pyroglutamate 13.32 The stereoselective alkylation using
`lithium hexamethyldisilazide (LiHMDS) and tert-butyl bro-
`
`moacetate 14 gave compound 15 in 70% yield with a cis/trans
`ratio of 3:1. The cis isomer 15 was treated with 1,8-diaza-
`bicyclo[5.4.0]undec-7-ene (DBU) in methylene chloride for 24 h
`to give inverted product in a 1:3 ratio and high yield. The
`assignment of the cis/trans configuration was based on the results
`reported in the literature.33,34 By use of the building block of C(4)-
`substituted pyroglutamate ester urethanes cis- and trans-15,
`
`6574 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 18
`
`Tsai et al.
`
`Petitioner GE Healthcare – Ex. 1047, p. 6574
`
`

`

`Table 2. Inhibition of FAP, DPP-IV, DPP8, DPP9, and DPP-II by Compounds 9, 10, 17, and 31
`
`10
`
`ICso (µM)"
`
`FAP
`0.019
`
`0.61
`
`DPP-IV
`>20
`
`>100
`
`DPP8
`3.7
`
`>100
`
`DPP9
`2.7
`
`>100
`
`DPP-11
`>100
`
`>100
`
`0.079
`
`>100
`
`>100
`
`>50
`
`>100
`
`0.21
`
`>100
`
`>100
`
`>100
`
`>100
`
`Ft;C N-
`
`/
`
`0.066
`
`>100
`
`>100
`
`>50
`
`>100
`
`0.059
`
`>100
`
`>100
`
`>100
`
`>100
`
`0.11
`
`>100
`
`>100
`
`>50
`
`>100
`
`0.088
`
`>100
`
`>100
`
`>100
`
`>100
`
`Compd
`
`9
`
`10
`
`17a
`
`31
`
`17b
`
`17c
`
`17d
`
`17e
`
`17f
`
`17g
`
`17h
`
`17i
`
`17j
`
`17k
`
`171
`
`17m
`
`0.25
`
`>100
`
`>100
`
`>20
`
`>100
`
`0.65
`
`>100
`
`>100
`
`>20
`
`>100
`
`15
`
`>100
`
`>100
`
`>20
`
`>100
`
`0.51
`
`>100
`
`>100
`
`>100
`
`>100
`
`3.2
`
`>100
`
`>100
`
`>100
`
`>100
`
`C l -0 -CN -
`
`0.20
`
`>100
`
`>100
`
`>100
`
`>100
`
`0.073
`
`>100
`
`>100
`
`>100
`
`>100
`
`0.73
`
`>100
`
`>100
`
`>100
`
`>100
`
`>20
`>100
`17n
`a Mean values of at least three experiments; standard deviations are (20%.
`
`>100
`
`>100
`
`>100
`
`Article
`
`Journal of Medicinal Chemistry, 2010, Vol. 53, No. 18
`
`6575
`
`Petitioner GE Healthcare – Ex. 1047, p. 6575
`
`

`

`Table 3. Inhibition of FAP, DPP-IV, DPP8, DPP9, and DPP-II by Compounds 17-20, 24, 27, 30, and 32
`
`R2
`
`\ { -\
`v··1
`\_.-N~o
`/
`4
`", -,
`I~ ' R1
`0~
`N 2
`H
`
`0
`
`R2
`
`H
`
`H
`
`Cl
`
`H
`
`H
`
`H
`
`H
`
`Compd
`
`R,
`
`17a
`
`18a
`
`18b
`
`19
`
`NC
`
`-~-] r
`,-NC
`
`.1··.rF
`-N
`
`,, •. (F
`-N
`\ ~-
`/
`NC
`F
`F·-/
`''F
`-N
`.
`
`j ,-NC
`
`24a
`
`//----s
`-N\_J
`
`24b
`
`/~--,
`-N,'--_J
`
`-✓-]
`r
`
`Ho·B,
`OH
`
`27
`
`20
`
`32
`
`0-,
`
`\.-N./o
`-, (?
`/
`\ N~
`0~ ~ \
`-0
`'
`,
`N
`H
`O HN.
`R
`20 R = t-butyl
`32R=H
`
`Ch---1°
`~~~-/ h 0~
`/
`'
`N
`o~~N~ "\'-
`b
`N
`H
`30
`
`ICso (µM)"
`
`FAP
`
`DPP-IV
`
`DPP8
`
`DPP9
`
`DPP-11
`
`0.079
`
`>100
`
`>100
`
`>50
`
`>100
`
`0.020
`
`>100
`
`>100
`
`>50
`
`>100
`
`0.063
`
`>100
`
`>100
`
`>100
`
`>100
`
`0.022
`
`>100
`
`>100
`
`>100
`
`>100
`
`>50
`
`>100
`
`>100
`
`>100
`
`>100
`
`>100
`
`>100
`
`>100
`
`>100
`
`>100
`
`0.031
`
`0.57
`
`0.50
`
`0.52
`
`0.40
`
`>100
`
`>100
`
`>100
`
`>100
`
`>100
`
`>100
`
`>100
`
`>100
`
`>100
`
`>100
`
`>100
`
`>50
`
`>100
`
`>100
`30
`>20
`a Mean values of at least three experiments; standard deviations are (20%.
`ability to suppress the formation of 20-22 was not effective.
`The more promising result was observed when acetonitrile
`(CH3CN) was applied as cosolvent with trifluoroacetic acid
`(TFA), and thus, the nitrile group of CH3CN participated in
`the reaction by acting as acceptor of the leaving tert-butyl
`cation. As expected, this approach slashed the yield of 20-22
`and produced another byproduct N-tert-butylacetamide
`(Mw = 115, detected by LC-MS). Compound 30 was pre-
`pared in five steps from the known building block 28, which
`was synthesized according to literature report (cis or trans
`configuration is not assigned).36
`
`pyroglutamic acid-based FAP inhibitors were synthesized. As
`shown in Scheme 2, the fully protected pyroglutamate trans-
`15 was deprotected by standard hydrogenation condition and
`was 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC)-
`coupled with pyrrolidine derivatives to give tert-butyl esters 23
`and 2517 or followed by dehydration of the amides to give
`16.9,35 Removal of the tert-butyl protecting group of 16, 23,
`and 25 with trifluoroacetic acid followed by EDC coupling
`with primary or secondary amines provided the desired (2S)-
`cyanopyrrolidine analogues 17-19, thiazolidine 24, and boro-
`nic ester 26, respectively. The free boronic acid 27 was ob-
`tained through acid catalyzed transesterification of boronic
`ester 26 with boron trichloride.17 The synthesis of cis-31 from
`cis-15 was identical to that of trans-17 from trans-15. The
`proposed mechanism for the formation of byproducts 20-22
`is that the nitrile group of 17-19 trapped the tert-butyl car-
`bocation followed by hydrolysis to produce amide com-
`pounds 20-22, respectively. Therefore, the search for scaven-
`gers of alkylating agents is required. When the thioanisole
`was used as a scavenger for the intermediate carbocation, its
`
`Results and Discussion
`
`To establish an optimized P2 site for FAP inhibition, C(4)-
`substituted pyroglutamic acid based inhibitors with various
`amines were explored, including bicyclic ring system (17a-h),
`piperazine ring system (17i,j), monocyclic system (17l,m), and
`phenylamine (17n). These derivatives described above were tested
`for inhibition of FAP, DPP-IV, DPP8, DPP9, and DPP-II, and
`the data are summarized in Table 2. Patent search found that
`
`6576 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 18
`
`Tsai et al.
`
`Petitioner GE Healthcare – Ex. 1047, p. 6576
`
`

`

`N-acyl-Gly-boroPro 9 is a potent FAP inhibitor with an IC50
`value of 1.8 nM.37 In our study, the reference compound 9 has
`an IC50 of 19 nM for the inhibition of FAP and is inactive
`toward DPP-IV and -II (IC50 greater than 20 and 100 μM,
`respectively), but this compound shows moderate inhibition
`against DPP8 and DPP9 (3.7 and 2.7 μM, respectively).
`A structure-activity relationship (SAR) work was carried
`out by replacing the boronic acid moiety of 9 with a cyano
`group. This work has led to the identification of N-acyl-
`Gly-cyanoPro 10 (FAP, IC50 = 0.61 μM), which is 30-fold
`less potent than 9 and inactive toward the other related en-
`zymes. In the pyroglutamic acid series of compounds, we star-
`ted initially with the bicyclic system using the isoquinoline
`ring. Compound 17a (trans configuration) inhibited FAP with
`an IC50 value of 79 nM and inactive toward DPP-IV, DPP-II,
`DPP8, and DPP9 (IC50 > 100 μM). The cis epimer 31 was
`around 3-fold less potent than trans-17a. Therefore, a focused
`structure-activity profiling effort was initiated by modifying
`the P2 site amine of trans isomer 17a. The introduction of
`chloro (17b), fluoro (17c), or trifluoromethyl (17d) substituent
`at the 5-position of the isoindoline gave similar potency (IC50
`≈ 66-110 nM) compared to 17a. When the isoindoline moiety
`of 17a was replaced with isoquinoline, compound 17e showed
`a similar level of FAP inhibition as seen for 17a. Further
`modification carried out with heterobicyclic building blocks,
`such as 4,5,6,7-tetrahydrothieno[2,3-c]pyridine 17f, imidazo-
`pyrazine 17g, and triazolopyrazine derivative 17h, led to at
`least a 3-fold decrease in FAP inhibitory potency compared
`to lead compound 17a. Compound 17h is the least potent
`FAP inhibitor in the bicyclic ring system with an IC50 value
`of 15 μM.
`Replacement of the bicyclic ring system with piperazine
`derivatives, such as 1-(4-fluorophenyl)piperazine 17i and
`1-pyridin-4-ylpiperazine 17j, led to no improvement in FAP
`potency. Among them, the potency of more polar 1-pyridin-4-
`ylpiperazine 17j gave a relatively low IC50 of 3.2 μM. On the
`basis of the results of 17h (bicyclic system) and 17j (piperazine
`system), it seemed that polar substituents at the P2 site would
`be detrimental to potency; compound 17h had weak inhibition
`(IC50 > 10 μM), and compound 17j exhibited low micromolar
`activity against FAP. Compound 17k with a 4-chlorophenyl
`substituent at the 4-position of 1,2,3,6-tetrahydropyridine is a
`close analogue of 17i. This compound showed 2.5-fold more
`potency than 17i. As for the effects of monocyclic system,
`pyrrolidine 17l (IC50 = 73 nM) is equipotent with bicyclic 17a
`as FAP inhibitors. However, replacement of the pyrrolidine
`ring of 17l with the piperidine (17m, IC50 = 730 nM) led to a
`marked 10-fold decrease in potency. Next, when we brought
`the nitrogen out of the ring system, the aniline 17n dramati-
`cally decreased the inhibitory activity against FAP (IC50 > 20
`μM). Therefore, the nitrogen out of the ring system was not
`further modified. In general, these active FAP inhibitors
`17a-n were inactive toward DPP-IV, DPP8, DPP9, and
`DPP-II (IC50 > 20 μM).
`After optimizing the P2 portion, we turned our attention to
`the P1 portion of the molecule. The requirements of the S1
`pocket for FAP inhibition were explored by keeping isoindo-
`line as the 4-position substituent and varying the 2-position
`substituent. Their inhibitory properties are shown in Table 3.
`Incorporation of the (4S)-fluoro substituent at the (2S)-2-
`cyanopyrrolidine ring (18a, IC50 = 20 nM) led to a 4-fold
`increase in FAP inhibitory activity compared to that of un-
`substituted analogue (17a, IC50 = 79 nM). However, intro-
`duction of a chloro substituent at the 5-position of isoindoline
`
`0 "
`
`0
`~ ---t~
`,}--N"'\b N
`0
`H
`
`-\ ~~
`
`_): N
`0
`H
`
`O
`
`N
`
`33
`
`34
`
`Figure 3. Inactive 4-alkyl substituted pyroglutamic acid amides.
`
`ring (18b) showed a decrease in activity (IC50 = 63 nM).
`Compound 19 with a 4,4-difluoro substituent at the 2-cyano-
`pyrrolidine ring exhibited very similar potency to monofluoro
`substituted 18a. Changing the 2-position substituent of pyro-
`glutamic acid from 2-cyanopyrrolidine 18a to thiazolidine
`24a or pyrrolidine 24b lacking a nitrile moiety showed no
`significant inhibition toward FAP (IC50 > 50 μM). All of the
`active FAP inhibitors shown in Table 3 were inactive toward
`the other related enzymes except boroproline 27. Even though
`27 showed a potent FAP inhibition (IC50 = 31 nM), it also
`exhibited strong inhibitory activity against the other DPP
`enzymes (IC50 values ranged from 400 to 600 nM). So far,
`from our studies and literature reports, we found that inhibi-
`tors of DPPs containing proline boronic acid (boro-Pro) in the
`P1 site cause poor selectivity of the inhibitors. Compound 20 is
`a byproduct in the synthesis of compound 17a. We also in-
`vestigated the effect of carboxylic acid tert-butylamide at the
`2-position of pyrrolidine. The bulky amide 20 was inactive
`against FAP and the other DPP enzymes (IC50 > 100 μM). As
`for no substituent amide 32 was also inactive toward all the
`DPP enzymes. It means that dipeptides containing prolina-
`mide in the P1 site cannot be well tolerated by FAP. In com-
`parison with 17a, 4-carboxypyroglutamic acid diamide 30,
`shortened by one carbon at the 4-position of 17a, caused no
`significant inhibition against FAP (IC50 > 20 μM) and the
`other DPP enzymes (IC50 > 50 μM). In addition, we intro-
`duced less polar alkyl groups, such as benzyl (33) and 3-methyl-
`but-2-enyl (34) groups, to replace the 2-(1,3-dihydroisoindol-
`2-yl)-2-oxoethyl moiety of the lead 17a at the 4-position of
`pyroglutamic acid diamide, but none of them showed signifi-
`cant inhibition against FAP (Figure 3, see Supporting Infor-
`mation for details).
`Next, we measured the inhibition constant Ki of potent
`compound 19 against FAP. Assessment of reaction progress
`curves in the presence of various concentrations of compound
`19 revealed a clear, time-dependent approach toward steady
`state, which is characteristic of slow-binding inhibition kine-
`tics, and the Ki value of 19 is 1.3 nM. In addition, the kinetic
`mechanism of the inhibition of FAP by the reference com-
`pound N-acyl-Gly-boroPro 9 was also determined. It showed
`a competitive inhibition pattern well-fitted to a Lineweaver-Burk
`plot, and its Ki value was estimated to be 7.9 nM (Figure 4).29
`Inhibitors that possessed excellent potency and selectivity
`profiles were selected for plasma pharmacokinetic screening
`in mice, and the data are summarized in Table 4. 4-Fluoropyr-
`rolidine derivative 18a demonstrated short-to-moderate oral
`half-life (1.8 h), low oral bioavailability (12%), and relatively
`high clearance (94 (mL/min)/kg). Introducing a 3-chloro sub-
`stituent at the isoindoline gave compound 18b, which showed
`reduction of total body clearance (55 (mL/min)/kg) and pro-
`longed oral half-life (t1/2 = 3.6 h) compared to 18a. On the
`other hand, although compound 18b had a very low oral bio-
`availability (F = 0.7%), it had a very long half-life (t1/2 (iv) =
`13.3 h) and a favorable drug exposure (AUC) after iv dosing.
`Therefore, compound 18b can be selected as an iv route in
`
`Article
`
`Journal of Medicinal Chemistry, 2010, Vol. 53, No. 18
`
`6577
`
`Petitioner GE Healthcare – Ex. 1047, p. 6577
`
`

`

`(A)
`
`60
`
`50
`
`I a 0
`
`(B)
`
`ov
`0
`
`0.07
`
`0.06
`
`0.05
`
`0.04
`
`0.03
`
`0.02
`
`0.01
`
`0.00
`
`0
`
`30 nM
`
`• 20 nM
`" 40 nM
`
`/:,
`50 nM
`■ 60 nM
`□ 70 nM
`
`-0.001
`
`0.000
`
`0.001
`
`0.002
`
`0.003
`
`0.004
`
`0
`
`5
`
`10
`
`15
`
`20
`
`25
`
`30
`
`1/[Substrate] (uM)
`
`Time (min)
`
`Figure 4. Determination of the Ki values for compounds 9 and 19 against human recombinant FAP (hFAP): (A) competitive inhibition of
`hFAP by compound 9; (B) slow-binding kinetics for the inhibition of hFAP by compound 19. One representative experiment is shown in both
`panels A and B.
`
`Table 4. Pharmacokinetic Properties of Selected FAP Inhibitors in Mice
`
`compd
`
`18a
`18b
`19
`
`CLP (mL/min/kg)
`
`T1/2 (h)
`
`intravenous (dose, 2 mg/kg)
`AUC (ng/(mL 3 h))
`460
`593
`495
`
`94
`58
`87
`
`2
`13.3
`0.8
`
`Cmax (ng/mL)
`
`210
`21
`108
`
`T1/2 (h)
`
`oral (dose, 20 mg/kg)
`AUC ng/(mL 3 h)
`408
`43
`336
`
`1.8
`3.6
`3.9
`
`F (%)
`
`12
`0.7
`10
`
`animal models for therapeutic target validation of FAP. The
`4,4-difluoropyrrolidine analogue 19 exhibited similar clear-
`ance, po AUC, and oral bioavailability to 4-fluoropyrrolidine
`18a. On the contrary, compound 19 showed 2-fold longer oral
`half-life (3.9 h) and lower Cmax than 18a (1.8 h). An analysis of
`the mouse pharmacokinetic data from the selected FAP
`inhibitors indicated a strong trend where a more hydrophobic
`chloro substituent at the isoindoline (18b) has a considerably
`lower plasma clearance and longer half-life after iv dosing and
`led to a compound with low Cmax and oral bioavailability after
`oral dosing.
`
`Conclusions
`
`By introduction of ring-constraint in the lead compound 8
`and carrying out SAR studies in a series of substituted 4-car-
`boxymethylpyroglutamic acid diamides, some potent and
`selective FAP inhibitors have been discovered. Notable among
`these are compounds 18a and 19 with a 4-fluoro substituent at
`the P1 site pyrrolidine ring, which are IC50 = 20 nM FAP
`inhibitors with excellent selectivity profile over DPP-IV, DPP-
`II, DPP8, and DPP9 (IC50 > 50 μM). The SAR suggests that
`the polar substituents at the P2 site (17h and 17j) would be
`detrimental to potency and that shortening by one carbon at
`the 4-position of pyroglutamic acid diamide (30) results in a
`loss in FAP inhibitory potency (IC50 > 20 μM). Our kinetic
`analyses demonstrated that compound 19 is a slow-binding
`inhibitor of FAP with Ki value of 1.3 nM. A survey of mouse
`pharmacokinetic parameters showed that selected compounds
`18a,b and 19 showed poor oral availability in mice. However,
`18b has a longer half-life, lower total body clearance, and
`favorable drug exposure (AUC) after iv treatment. Thus,
`compound 18b can be dosed as iv route to evaluate the
`potential therapeutic effects of FAP inhibitors in vivo.
`
`Experimental Section
`
`All commercial chemicals and solvents are reagent grade and
`were used without further treatment unless otherwise noted. 1H
`
`NMR spectra were obtained with a Varian Mercury-300 spectro-
`meter operating at 300 or 400 MHz. Chemical shifts were recorded
`in parts per million (ppm, δ) and were reported relative to the
`solvent peak or TMS. High-resolution mass spectra (HRMS)
`were measured with a Finnigan (MAT-95XL) electron impact
`(EI) mass spectrometer. LC-MS data were measured on an Agi-
`lent MSD-1100 ESI-MS/MS System. Microanalyses were carried
`out on a Heraeus VarioIII-NCH analyzer. Flash column chro-
`matography was done using silica gel (Merck Kieselgel 60, no.
`9385, 230-400 mesh ASTM). Reactions were monitored by TLC
`using Merck 60 F254 silica gel glass backed plates (5 cm  10 cm).
`
`Zones were detected visually under ultraviolet irradiation (254
`nm) or by spraying with phosphomolybdic acid reagent (Aldrich)
`followed by heating at 80 °C. All starting materials and amines
`were commercially available unless otherwise indicated. Purity of
`target compounds were over 95% based on reversed-phase HPLC
`analyses (Chromolith performance RP-18e 4.6 mm 100 mm
`column) under the two elution conditions. Condition A was 40:60
`MeOH-H2O, and condition B was 70:30 MeOH-H2O. The flow
`rate was 0.2 mL/min, and the injection volume was 5 μL. The
`system operated at 25 °C. Peaks were detected at λ = 220/254 nm.
`Benzyl (2S,4S)-N-tert-Butoxylcarbonyl-4-tert-butoxycarbony-
`lmethypyrglutamate (trans-15) and Benzyl (2S,4R)-N-tert-Butoxy-
`lcarbonyl-4-tert-butoxycarbonylmethypyrglutamate (cis-15). Benzyl
`(2S)-N-tert-butoxylcarbonylpyrglutamate 1332 (628 mg, 2 mmol)
`was dissolved in tetrahydrofuran (THF, 10 mL) and cooled to
`-78 °C, with stirring under nitrogen. LiHMDS (1.0 M in THF,
`2.2 mL, 4.4 mmol) was added dropwise over 5 min, and stirring was
`continued for 1 h. tert-Butyl bromoacetate 14 (0.32 mL, 4.0 mmol)
`was added dropwise over 3 min, and stirring was continued at
`-78 °C for 2 h. The reaction was quenched with saturated aqueous
`ammonium chloride and extracted into ethyl acetate. The organic
`layer was washed with water and saturated aqueous sodium
`chloride and dried (MgSO4). The solvent was removed in vacuo
`to give a brown oil, which was purified by column chromatography
`on silica gel (eluted with hexane/CH2Cl2/EA = 4/5/1) to yield the
`desired compounds trans-15 (130 mg, 15%) and cis-15 (390 mg,
`45%) as white crystalline solids. trans-15: mp 89-91 °C. [R]D
`24 -24.0
`(c 0.2, CH2Cl2). 1H NMR (300 MHz, CDCl3): δ 7.35 (m, 5H), 5.22
`and 5.17 (AB quartet, J = 12.0 Hz, 2H), 4.62 (dd, J = 9.6, 1.2 Hz,
`
`6578 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 18
`
`Tsai et al.
`
`Petitioner GE Healthcare – Ex. 1047, p. 6578
`
`

`

`1H), 3.10-2.90 (m, 1H), 2.78 (dd, J = 17.0, 3.9 Hz, 1H), 2.39 (dd,
`J = 17.0, 8.4 Hz, 1H), 2.32 (d, J = 9.0 Hz, 1H), 2.16-2.02 (m, 1H),
`þ
`) m/z calcd for C23H31NO7:
`1.42 (s, 9H), 1.40 (s, 9H). MS (ES
`433.49. Found: 456.1 (Mþ Na
`þ
`). cis-15: mp 70-72 °C. [R]D
`24 -15.3
`(c 0.2, CH2Cl2). 1H NMR (400 MHz, CDCl3): δ 7.36 (m, 5H), 5.22
`and 5.17 (AB quartet, J = 12.0 Hz, 2H), 4.55 (t, J = 7.8 Hz, 1H),
`3.05-2.90 (m, 1H), 2.84 (dd, J = 17.0, 3.9 Hz, 1H), 2.60-2.70 (m,
`1H), 2.34 (dd, J = 16.7, 10.0 Hz, 1H), 1.74-1.64 (m, 1H), 1.43 (m,
`þ
`) m/z calcd for C23H31NO7: 433.49. Found: 456.2
`18H). MS (ES
`(M þ Na
`þ
`).
`Isomerization from cis-15 to trans-15.34 The cis-15 (433 mg,
`1 mmol) was dissolved in CH2Cl2 and cooled to 0 °C. DBU (0.44
`mL, 3 mmol) was added dropwise and the mixture stirred at 0 °C
`for 30 min and then at room temperature for 24 h. The mixture
`was diluted with CH2Cl2 and washed once each with 1 N HCl
`and water. The organic layer was dried over MgSO4, filtered,
`and concentrated in vacuo to give the crude product, which was
`purified by column chromatography as described above to yield
`trans-15 (292 mg, 68%) and recovered cis-15 (100 mg, 23%).
`General Procedure for the Synthesis of Compounds 16, 23, 25,
`and 29. To a solution of compound 15 or 28 (1.0 mmol) in
`MeOH (10 mL) was added 5% palladium on carbon under an
`atmosphere of nitrogen. The mixture was stirred vigorously
`under an atmosphere of hydrogen for 5 h at room temperature.
`The mixture was filtered through Celite and concentrated in
`vacuo to give (2S,4S)-N-tert-butoxylcarbonyl-4-tert-butoxycar-
`bonylmethypyrglutamic acid or (2S,4S)-N-tert-butoxylcarbo-
`nyl-4-tert-butoxycarbonylpyrglutamic acid as a pale yellow oil.
`The crude acid was used for the next reaction without further
`purification. To a solution containing this acid (343 mg, 1.0
`mmol) and N-hydroxybenzotrizole (HOBt, 168 mg, 1.1 mmol)
`in 1,4-dioxane (5 mL) was added a solution of EDC (211 mg, 1.1
`mmol) in CH2Cl2 (5 mL). The mixture was stirred for 10 min at
`room temperature. To the resulting solution was added the
`(2S,4S)-4-
`appropriate proline derivatives
`(L-prolinamide,
`fluoroprolinamide,35 (2S)-4,4-difluoroprolinamide,35 thiazolid-
`ine or prolineboronic ester,17 1.2 equiv) in CH2Cl2 (4 mL), with
`stirring. After 2 h, the reaction mixture was diluted with CH2Cl2
`and washed with saturated aqueous NaHCO3 solution (10 mL),
`1 N aqueous citric acid solution (10 mL), and brine (10 mL). The
`combined organic layers were dried over MgSO4, filtered, and
`concentrated. The crude prolinamides 16 and 29 and thiazoli-
`dine 23 were used for the next reaction without further purifica-
`tion. The crude boronic ester 25 was purified over silica gel using
`hexane/EA (7:3) as an eluant to yield the desired compound 25
`(92%, two steps) as a colorless oil.
`To a mixture of the prolinamide 16 or 29 (1.6 mmol) and
`imidazole (0.12 g, 1.8 mmol) in pyridine (5 mL) cooled to -20 °C
`was added phosphoryl chloride (0.64 g, 4.2 mmol) under nitro-
`gen. After being stirred for 30 min at -20 °C, the mixture was
`evaporated to dryness in vacuo. The resulting browm solid was
`dissolved in CH2Cl2 (40 mL) and washed with 1.0 N aqueous
`citric acid (40 mL). The organic phase was dried over magne-
`sium sulfate, filtered, and concentrated under reduced pressure
`to yield the crude material as a viscous oil. The crude material
`was purified by chromatography on silica gel (eluted with
`hexane/EA = 2/1) to give the corresponding nitrile 16 or 29
`(>90%) as a white powder. Only the data of representative
`compounds are shown.
`(2S,4S)-3-tert-Butoxycarbonylmethyl-5-((2S)-2-cyanopyrroli-
`dine-1-carbonyl)-2-oxopyrrolidine-1-carboxylic Acid tert-Butyl
`Ester (16a). Mp 160-162 °C. 1H NMR (300 MHz, CDCl3): δ
`4.83 (dd, J = 6.9, 2.1 Hz, 1H), 4.71 (dd, J = 9.2, 1.8 Hz, 1H),
`3.79-3.58 (m, 2H), 3.22-3.06 (m, 1H), 2.74 (dd, J = 17.1, 3.9 Hz,
`1H), 2.50 (dd, J = 17.0, 8.1 Hz, 1H), 2.04-2.40 (m, 6H), 1.49 (s,
`þ
`) m/z calcd for C21H31N3O6: 421.49.
`9H), 1.44 (s, 9H). MS (ES
`), 266.0 (M - (Boc)-(tert-butyl) þ H
`Found: 444.1 (M þ H
`þ
`