`REPORTS IN
`MEDICINAL
`CHEMISTRY
`
`Sponsored by the Division of Medicinal Chemistry
`of the American Chemical Society
`
`Editor-in-Chief: JOHN E. MACOR
`
`BRISTOL-MYERS SQUIBB, R&D
`WALLINGFORD, CT, USA
`
`Boehringer Ex. 2012
`Mylan v. Boehringer Ingelheim
`IPR2016-01564
`Page 1
`
`
`
`Annual Reports in
`MEDICINAL CHEMISTRY
`
`VOLlJME 42
`
`Sponsored by the Division of Medicinal Chemistry
`of the American Chemical Society
`
`Editor-in-Chief
`JOHN E. MACOR
`Neuroscience Discovery Chemistry
`Bristol-Myers Squibb
`Wallingford, CT, United States
`
`Section Editors
`
`ROBICHAUD • STAMFORD • BARRISH • MYLES •
`LOWE • DESAI
`
`ElSEVIER
`
`Amsterdam • Boston • Heidelberg • London • New York • Oxford
`Paris • San Diego • San Francisco • Singapore • Sydney • Tokyo
`Academic Press is an imprint of Elsevier
`
`ACADEMIC
`PRESS
`
`Boehringer Ex. 2012
`Mylan v. Boehringer Ingelheim
`IPR2016-01564
`Page 2
`
`
`
`Academic Press is an imprint of Elsevier
`84 Theobald's Road, London WC1X 8RR, UK
`Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands
`Linacre House, Jordan Hill, Oxford OX2 8DP, UK
`30 Corporate Drive, Suite 400, Burlington, MA 01803, USA
`525 B Street, Suite 1900, San Diego, CA 92101-4495, USA
`
`First edition 2007
`Copyright © 2007 Elsevier Inc. All rights reserved
`
`No part of this publication may be reproduced, stored in a retrieval system or
`transmitted in any form or by any means electronic, mechanical, photocop(cid:173)
`ying, recording or otherwise without the prior written permission of the
`publisher
`
`Permissions may be sought directly from Elsevier's Science & Technology
`Rights Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0)
`1865 853333; email: permissions@elsevier.com. Alternatively you can
`submit your request online by visiting the Elsevier web site at http:/ I
`www.elsevier.com/locate/permissions, and selecting Obtaining permission to
`use Elsevier material
`
`Notice
`No responsibility is assumed by the publisher for any injury and/or damage
`to persons or property as a matter of products liability, negligence or other(cid:173)
`wise, or from any use or operation of any methods, products, instructions or
`ideas contained in the material herein. Because of rapid advances in the
`medical sciences, in particular, independent verification of diagnoses and drug
`dosages should be made
`
`ISBN: 978-0-12-373912-4
`ISSN: 0065-7743
`
`For information on all Academic Press publications
`visit our website at books.elsevier.corn
`
`Printed and bound in USA
`
`07 08 09 10 11 10 9 8 7 6 5 4 3 2 1
`
`Working together to grow
`libraries in developing countries
`www.elsevier.com I www.bookaid.org I www.sabre.org
`~~?..~ .. ~~?, Sabre foundation
`
`ELSEVIER
`
`Cor
`
`CHAP
`
`CHAf
`
`CHAF
`
`Boehringer Ex. 2012
`Mylan v. Boehringer Ingelheim
`IPR2016-01564
`Page 3
`
`
`
`Contents
`
`Case History: JANUVIA TM
`(Sitagliptin), a Selective Dipeptidyl
`Peptidase IV Inhibitor for the
`Treatment of Type 2 Diabetes
`
`Ann E. Weber* and Nancy Thornberry**
`
`1.
`Introduction
`2. Pathogenesis of Type 2 Diabetes
`3. Rationale for the Use of DPP-4 Inhibitors to Treat Type 2 Diabetes
`Inhibitor Program: Threo-
`4. MRL's DPP-4
`and Allo-lsoleucyl
`Thiazolidides
`5. Medicinal Chemistry Efforts Leading to Sitagliptin
`5.1 Program objectives
`5.2 o:-Amino acid derived DPP-4 inhibitors
`5.3 High throughput screening hits
`5.4 SAR in the p-aminoacyl amide series
`5.5 SAR in the piperazine series
`6. Properties of Analog 27 and Sitagliptin
`7. Clinical Studies of Sitagliptin
`8. Conclusion
`References
`
`95
`96
`97
`
`98
`100
`100
`100
`101
`102
`103
`105
`106
`107
`107
`
`1. INTRODUCTION
`
`Diabetes is a global epidemic affecting more than 240 million people worldwide.
`The incidence of this disease is growing at an alarming rate, with 380 million
`cases predicted by 2025. Each year over 3.8 million people die from complications
`of diabetes, including heart disease, stroke and kidney failure. The vast majority
`
`* Department of Medicinal Chemistry, Merck Research Laboratories, P.O. Box 2000, Rahway, NJ 07065, USA
`** Department of Metabolic Disorders, Merck Research Laboratories, P.O. Box 2000, Rahway, NJ 07065, USA
`© 2007 Elsevier Inc.
`Annual Reports in Medicinal Chemistry, Volume 42
`ISSN 0065-7743, DOl 10.1016/S0065-7743(07)42007-3
`All rights reserved.
`
`95
`
`Boehringer Ex. 2012
`Mylan v. Boehringer Ingelheim
`IPR2016-01564
`Page 4
`
`
`
`96
`
`A.E. Weber and N. Thornberry
`
`Figure 1 JANUVIA TM (sitagliptin).
`
`(90-95%) of cases are type 2 diabetes, largely resulting from the increasing prev(cid:173)
`alence of obesity and sedentary lifestyles [1].
`Despite the availability of a range of agents to treat type 2 diabetes, glucose
`control remains suboptimal, with less than 50% of patients achieving stated
`glycemic goals. In addition, current therapies have limited durability and/ or are
`associated with significant side effects such as GI intolerance, hypoglycemia,
`weight gain, lactic acidosis and edema [2]. Thus, significant unmet medical needs
`remain. In particular, safer, better tolerated medications which frrovide increased
`(sitagliptin, 1,
`efficacy and long-term durability are desired. JANUVIAT
`Figure 1), a dipeptidyl peptidase IV (DPP-4) inhibitor, represents a promising
`new approach to the treatment of this disease.
`
`2. PATHOGENESIS OF TYPE 2 DIABETES
`
`The pathogenesis of type 2 diabetes involves a set of three primary defects:
`insulin resistance, insulin secretory dysfunction, and hepatic glucose overpro(cid:173)
`duction. Insulin resistance is a common predisposing defect, and is believed to
`occur as a consequence of obesity in most individuals. As long as an individual
`maintains insulin secretion adequate to compensate for insulin resistance, plasma
`glucose levels remain normal; however, if ~-cell function declines, and the pan(cid:173)
`creas is no longer able to produce adequate amounts of insulin to compensate for
`the insulin resistance, hyperglycemia - and subsequently, diabetes mellitus -
`results. Not only does this ~-cell defect lead to hyperglycemia and the onset of
`diabetes, the progressive decline in ~-cell function during the course of diabetes
`leads to the need for more and more complex treatment regimens to manage
`glucose control in diabetic patients, and ultimately, to the need for insulin. As
`expected from the pathogenesis of type 2 diabetes, therapies that increase the
`circulating concentrations of insulin have proven therapeutically beneficial in
`the treatment of type 2 diabetes [2]. Indeed, sulfonylureas and related insulin
`secretagogues currently represent 42% of the total worldwide oral market, with
`sales in excess of $1.7 billion, notwithstanding mechanism-based side effects
`of hypoglycemia and weight gain. In addition, current insulin secretagogues
`commonly fail to maintain adequate glycemic control, and may contribute to
`the progressive decline in ~-cell function. Thus, current unmet medical needs
`in the treatment of type 2 diabetes include insulin secretagogues which are
`
`glucos
`to wei;
`
`3. RA
`DIJ
`
`Inhibit
`pro ad
`part, <
`agents
`inhibit
`via st<
`has a'
`tion. C
`logs b
`fastin!
`resultE
`stanti<
`agoni~
`speciE
`tane01
`opme
`the cc
`Dl
`avail<
`incre<
`are b
`hypo:
`led b
`throu
`or at
`ficial
`and E
`E<
`inhib
`inclu
`or "k
`hum<
`tions
`[8,9].
`comr
`culat
`anim
`In 1S
`indeJ
`pour
`
`Boehringer Ex. 2012
`Mylan v. Boehringer Ingelheim
`IPR2016-01564
`Page 5
`
`
`
`r
`I
`l
`
`prev-
`
`lucose
`stated
`or are
`cemia,
`needs
`reased
`tin, 1,
`mising
`
`lefects:
`rerpro(cid:173)
`~ved to
`ividual
`plasma
`1e pan-
`3ate for
`llitus -
`mset of
`liabetes
`nan age
`tlin. As
`ase the
`ficial in
`insulin
`et, with
`~ effects
`tgogues
`ibute to
`tl needs
`lich are
`
`Case History
`
`97
`
`glucose-dependent, decreasing the risk for hypoglycemia, and which do not lead
`to weight gain.
`
`3. RATIONALE FOR THE USE OF DPP-4 INHIBITORS TO TREAT TYPE 2
`DIABETES
`
`Inhibitors of DPP-4, a proline selective serine protease, are a new therapeutic ap(cid:173)
`proach to the treatment of type 2 diabetes [3,4]. DPP-4 inhibitors function, at least in
`part, as indirect stimulators of glucose-dependent insulin secretion, and these
`agents may address a number of the unmet medical needs noted above. The DPP-4
`inhibitor induced increase in insulin secretion is believed to be mediated primarily
`via stabilization of the incretin hormone glucagon-like peptide-1 (GLP-1), which
`has a clearly established role in glucose-dependent insulin biosynthesis and secre(cid:173)
`tion. Continuous infusion of GLP-1 or subcutaneous administration of GLP-1 ana(cid:173)
`logs to diabetic humans has resulted in normalization of both postprandial and
`fasting glucose [5]. For example, sub-chronic (6 week) continuous infusion of GLP-1
`resulted in profound and significant decreases in fasting plasma glucose, and sub(cid:173)
`stantial improvement in HbA1CI a marker of overall glycemic control [6]. A GLP-1
`agonist, BYETTA® (exenatide), originally identified as a salivary protein in a lizard
`species, has been approved for use in patients with type 2 diabetes as a subcu(cid:173)
`taneously administered peptide, and multiple other GLP-1 analogs are in devel(cid:173)
`opment. An alternate oral strategy involves the use of DPP-4 inhibitors to increase
`the concentrations of endogenously released GLP-1.
`DPP-4 inhibitors have at least three potential advantages over currently
`available oral insulin secretagogues: first, because the incretin peptide GLP-1
`increases insulin in a strictly glucose-dependent manner (i.e., when glucose levels
`are below normal, no stimulation of insulin secretion occurs), a low risk of
`hypoglycemia would be expected. Second, since the use of GLP-1 analogues has
`led to decreased appetite and weight reduction, a DPP-4 inhibitor, that acts
`through augmentation of GLP-1, would be expected to provide either weight loss
`or at least no gain in weight. Finally, DPP-4 inhibitors may have long-term bene(cid:173)
`ficial effects on ~-cell function in that GLP-1 stimulates both insulin biosynthesis
`and secretion, and is suggested to have a role in regulation of ~-cell mass [7].
`Early in the Merck Research Laboratories (MRL) program, evidence that DPP-4
`inhibitors may be useful for treatment of type 2 diabetes mellitus was published,
`including studies showing that Dppr1- mice (i.e., genetically deficient in DPP-4,
`or "knock-out" mice) have improved glucoregulation, and studies conducted in
`humans with DPP-4 inhibitors showing lowering of plasma glucose concentra(cid:173)
`tions. Dpp4-l- mice are healthy, fertile, and have improved metabolic function
`[8,9]. Specifically, these animals have improved glucose tolerance, which is ac(cid:173)
`companied by increased levels of insulin and active GLP-1, and decreased cir(cid:173)
`culating glucagon concentrations. Similar effects have been observed in several
`animal models of diabetes with structurally distinct DPP-4 inhibitors [10-12].
`In 1999, Novartis (with DPP728, 2, Figure 2) and Probiodrug (with P32/98, 3)
`independently launched Phase I clinical trials. In single dose studies, both com(cid:173)
`pounds were well tolerated, increased active GLP-1, and reduced glycemic
`
`Boehringer Ex. 2012
`Mylan v. Boehringer Ingelheim
`IPR2016-01564
`Page 6
`
`
`
`98
`
`A.E. Weber and N. Thornberry
`
`MeJ:/
`NH2 l/
`
`3
`
`Figure 2 Structures of DPP-4 inhibitors.
`
`excursion following food or glucose intake in normal volunteers [13-15]. Subse(cid:173)
`quently, Novartis reported the results of a 4-week phase II study with DPP728 in
`93 patients at 100mg T.I.D. or 150mg B.I.D.: significant decreases were observed
`in maximal glucose excursion, fasting plasma glucose, and 24 h glucose [16].
`Development of this compound was discontinued in favor of LAF237 (4), now
`known as GALVUS™ (vildagliptin) [17].
`As described above, the reductions in glucose levels that are observed with
`DPP-4 inhibitors are believed to be mediated primarily through stabilization of
`the incretin GLP-1 [7-36] amide, a ~3 kDa peptide hormone that is intimately
`involved in the regulation of glucose homeostasis. GLP-1 is efficiently hydrolyzed
`in vitro (kcat!Km ~1 x 106 M/s) by DPP-4 to generate an inactive product, GLP-1
`[9-36] amide [18]. The most compelling evidence that DPP-4 is primarily respon(cid:173)
`sible for the rapid regulation of GLP-1 in vivo (t112 ~1 min) is provided by studies
`with specific DPP-4 inhibitors, which produce increased circulating concentrations
`of GLP-1 in both rodents and humans [13,14,19], and by the finding that DPP-4-
`deficient mice have increased ( ~3-fold) circulating levels of intact GLP-1 [9].
`Although GLP-1 is believed to be the primary mechanism by which DPP-4
`inhibitors lower glucose, other substrates may also be important. Several mem(cid:173)
`bers of the glucagon peptide family are cleaved by DPP-4 in vitro, and the incretin
`glucose-dependent insulinotropic peptide (GIP), in particular, is clearly regulated
`by DPP-4 in vivo in both rodents and humans. Evidence that both GLP-1 and GIP
`are important for improved glucose AUC in rodents has been provided in studies
`with a DPP-4 inhibitor in GLP-1 and GIP receptor knockout mice, and mice
`which are deficient in both receptors [8,20]. The importance of GIP stabilization to
`improved glucose control in diabetic humans, who have a diminished response
`to exogenous GIP, has not been established.
`·
`
`4. MRL'S DPP-4 INHIBITOR PROGRAM: THREO- AND
`ALLO-ISOLEUCYL THIAZOLIDIDES
`
`In order to jump-start internal efforts on the DPP-4 inhibitor program, L-fhreo(cid:173)
`(25,35)-isoleucyl thiazolidide 3 (Figure 2) and its allo (25,3R) stereoisomer were
`
`licensed fro
`discontinue.
`dogs [21]. 1
`ated, but rn
`related prol
`itory activit
`toxicity pro
`dogs, the al
`compared c
`toxicity pro
`ized by blo<
`and multipl
`In view
`dose-respor
`were not rn(
`one or more
`tissue extra
`from the ki.
`found to o
`detected us
`eptidase act
`that measur
`was differer
`and 86nM,
`against this
`peptidases l
`In an ef
`screened for
`inhibitory a'
`activity aga
`differed by c:
`were simila
`served diffe
`esized that i
`compounds.
`macokinetic
`and QPP (q1
`exploratory
`DPP-4 inhib
`The resu
`produced h
`Both compo
`and both a
`>30mpk, a
`selective inl
`lOOmpk dm
`
`Boehringer Ex. 2012
`Mylan v. Boehringer Ingelheim
`IPR2016-01564
`Page 7
`
`
`
`Case History
`
`99
`
`licensed from Probiodrug in late 2000. Development of both compounds was
`discontinued in February 2001 due to unacceptable toxicity profiles in rats and
`dogs [21]. Available evidence suggested that the toxicity was not DPP-4 medi(cid:173)
`ated, but more likely due to an off-target activity, in particular DPP-8, a closely
`related proline-specific protease. The threo and allo isomers had identical inhib(cid:173)
`itory activity against DPP-4, both in vitro and in vivo; however, although the
`toxicity profiles of these compounds are qualitatively similar in both rats and
`dogs, the allo isomer is significantly more toxic (approximately 10-fold) when
`compared on either a milligram per kilogram or plasma exposure basis. The
`toxicity profiles of these compounds include gastrointestinal toxicity character(cid:173)
`ized by bloody diarrhea and tenesmus in dogs, thrombocytopenia and anemia,
`and multiple organ pathology in both species.
`In view of the comparable pharmacodynamic activity, the differences in the
`dose-response curves for the various toxic effects suggested that these toxicities
`were not mechanism-based. Evidence that they might be due to the inhibition of
`one or more proline-specific dipeptidyl peptidases was provided by studies with
`tissue extracts from DPP-4-deficient mice [21]. Detergent-solubilized extracts
`from the kidneys, liver, lung, and gastrointestinal tract of these animals were
`found to contain low levels of a Pro-specific dipeptidyl peptidase activity,
`detected using the fluorogenic substrate Gly-Pro-AMC. The Pro-selective dip(cid:173)
`eptidase activity in tissues isolated from Dpp4-l- mice was 10-25 fold lower than
`that measured in corresponding tissues of wild-type animals, and unlike DPP-4,
`was differentially inhibited by threo- and allo-isoleucyl thiazolidide (IC50 = 726
`and 86 nM, respectively). The 8.5-fold greater potency of the allo diastereomer
`against this activity suggested that off-target inhibition of one or more DPP-4-like
`peptidases by this inhibitor could be responsible for preclinical toxicity.
`In an effort to evaluate this hypothesis, the allo and threo isomers were
`screened for activity against a panel of related dipeptidases. A comparison of the
`inhibitory activities of the isomers revealed that although they had comparable
`activity against DPP-4, their activities against the related dipeptidase DPP-8
`differed by about 10-fold. Activities against five other available related peptidases
`were similar. Since the differences against DPP-8 were consistent with the ob(cid:173)
`served differences in dose necessary to produce toxicity, it was further hypoth(cid:173)
`esized that inhibition of DPP-8 was responsible for the observed toxicities of these
`compounds. To evaluate this theory, a series of inhibitors with similar phar(cid:173)
`macokinetic profiles in rats but with differing activities against DPP-4, DPP-8,
`and QPP (quiescent cell proline peptidase, aka DPP-11 and DPP-7) were tested in
`exploratory 2-week rat oral toxicity studies. In addition, the potent and selective
`DPP-4 inhibitor desfluorositagliptin 27 (Figure 9) was also tested in parallel.
`The results of these studies showed a remarkable similarity in the effects
`produced by the selective DPP-8 inhibitor and allo-isoleucyl thiazolidide [21].
`Both compounds produced mortality and alopecia at the highest doses tested,
`and both also produced thrombocytopenia of similar magnitude at doses
`> 30 mpk, and enlarged spleens and lymph nodes at all doses tested. The QPP
`selective inhibitor produced significant reductions in reticulocyte counts at the
`100 mpk dose. No other changes were noted. In contrast to the above compounds,
`
`bse-
`8 in
`ved
`[16].
`llOW
`
`with
`n of
`1.tely
`{zed
`LP-1
`pon(cid:173)
`tdies
`:ions
`>P-4-
`
`PP-4
`lem(cid:173)
`retin
`lated
`GIP
`1dies
`mice
`on to
`1onse
`
`fhreo(cid:173)
`were
`
`Boehringer Ex. 2012
`Mylan v. Boehringer Ingelheim
`IPR2016-01564
`Page 8
`
`
`
`100
`
`A.E. Weber and N. Thornberry
`
`the selective DPP-4 inhibitor did not produce any changes in physical appear(cid:173)
`ance, body weight, hematology, clinical chemistry, and urinalysis, and histopa(cid:173)
`thology was clean. Therefore, these results strongly support the hypothesis that
`the toxicities observed in rats with thiazolidide 3 and its allo (2S,3R) stereoisomer
`are not DPP-4 related but are consistent with inhibition of DPP-8.
`This conclusion is also supported by studies in dogs. Acute oral and intra(cid:173)
`venous administration of both the allo and threo compounds resulted in bloody
`diarrhea, emesis, and tenesmus. These effects were reproduced in dogs treated
`with 10 mg/kg of the DPP-8 selective inhibitor; however, no acute effects were
`observed in dogs given desfluorositagliptin [21]. Since no toxicity was observed
`in dogs given single oral doses of the QPP selective inhibitor, these results also
`support the conclusion that the dog acute toxic effects are mediated by DPP-8.
`During the course of these studies, another proline selective peptidase, DPP-9,
`closely related to DPP-8, was described [22]. The DPP-8 selective inhibitor was
`then found to be a dual DPP-8/9 inhibitor. Thus inhibition of one or both of these
`enzymes, or possibly another closely related enzyme, may be responsible for the
`observed toxicity. Efforts to identify potent and selective inhibitors of the indi(cid:173)
`vidual enzymes in order to further deduce the mechanism of toxicity have thus
`far not succeeded.
`
`5. MEDICINAL CHEMISTRY EFFORTS LEADING TO SITAGLIPTIN
`
`5.1 Program objectives
`
`Following the completion of the studies described above, the objective of the
`internal program was to identify a potent DPP-4 inhibitor for development with
`> 1000-fold selectivity over related proline peptidases, especially DPP-8 and
`DPP-9. A half-life suitable for once daily dosing was preferred, though a com(cid:173)
`pound with twice daily dosing would still provide an advantage over DPP728,
`the only DPP-4 inhibitor known to be in development for diabetes at the time. To
`achieve a greater duration of action, we elected from the beginning to only con(cid:173)
`sider structures lacking a reactive electrophile. Most of the known DPP-4 inhib(cid:173)
`itors contained such an electrophile, typically a nitrile, which forms a covalent
`(though reversible) bond with the alcohol of the active site serine [23]. Because
`these inhibitors also require a free amine five atoms away, the potential for intra(cid:173)
`molecular cyclization to form a six-membered ring exists. Chemical instability is
`seen with many of these inhibitors in vitro [23], and this may contribute to the
`short half-life often observed with these DPP-4 inhibitors in vivo.
`
`5.2 tX-Amino acid derived DPP-4 inhibitors
`
`When the medicinal chemistry program began in late 1999, nearly all of the DPP-
`4 inhibitors known in the literature were derived from a-amino acids, and those
`lacking an electrophlle such as the isoleucyl thiazolidides were considerably less
`potent than those containing a nitrile or boronic acid such as DPP728. One report
`suggested that cyclohexylglycine derived inhibitors showed improved potency
`
`51
`6(
`
`Figure 3
`
`a
`
`[24]. Inde<
`assay. In <
`the cycloh
`substituer
`potent to
`oxidation
`culminate,
`had good
`species [2:
`able inhib
`halted dw
`then focus
`Later, '
`resumed.,
`~-methyl F
`selectivity
`provided iJ
`hERG pota
`IC50 = 8.8r
`> 10,000-fo
`enzymes, a
`acterized b
`oral bioava
`as an "ins-c
`
`5.3 Hight
`
`High throu
`only two,~
`thy of exter
`prepared ir
`inhibited tr
`exploratory
`logs of 9 ha
`center was
`part of apr
`
`Boehringer Ex. 2012
`Mylan v. Boehringer Ingelheim
`IPR2016-01564
`Page 9
`
`
`
`Case History
`
`101
`
`~Q A=<N~~~$
`"RN
`- R N~
`~H2
`
`F=
`5 (R = H,X = S)
`7 (R = H,R' = Me,R" = 4-F-Ph)
`6 (R = 4-CF30-PhS02NH-, X= C 1111F)
`8 (R = F,R' = CONMe2,R" =A)
`Figure 3 Cl-Amino acid derived DPP-4 inhibitors.
`
`[24]. Indeed, cyclohexylglycyl thiazolidide 5 (Figure 3) has an IC50 of 89 nM in our
`assay. In order to develop a proprietary position in this series, substitutions on
`the cyclohexyl ring were explored [25,26]. In addition, SAR studies of the amine
`substituent indicated that the fluoropyrrolidine derivatives were nearly equi(cid:173)
`potent to the thiazolidine analogs [27]. Since the thiazolidine ring is prone to
`oxidation on sulfur, the former analogs are more metabolically stable. This work
`culminated in 'the identification of sulfonamide 6 (DPP-4 IC50 = 36 nM) which
`had good to excellent oral bioavailability and half-life of 4-12 h in preclinical
`species [27]. Once the toxicity of the isoleucyl thiazolidides was traced to prob(cid:173)
`able inhibition of DPP-8 and/ or DPP-9, development of this compound was
`halted due to its low micromolar affinity for these two enzymes. Efforts were
`then focused on two promising leads that emerged from screening.
`Later, after the identification of sitagliptin, work in the a-amino acid series
`resumed. A new approach based on the threo isoleucyl thiazolidide lead gave a
`P-methyl phenylalanine series, typified by 7 (DPP-4 IC50 = 64nM), with increased
`selectivity [28]. Substitution of the P-methyl group with a P-dimethylamido moiety
`provided increased selectivity against off-target activity, in particular binding to the
`hERG potassium channel [29]. The 4-heteroarylphenylalanine derivative 8 (DPP-4
`IC50 = 8.8 nM) is among the most selective analogs made in this series, with
`> 10,000-fold selectivity against other dipeptidyl peptidases, cytochrome P450
`enzymes, and ion channels [30]. In addition, its pharmacokinetic profile was char(cid:173)
`acterized by low clearance (1-SmL/min/kg), good half-life (2-7h), and excellent
`oral bioavailability (56--100%) across species. This compound was brought forward
`as an "insurance back-up", to be developed should sitagliptin have faltered.
`
`5.3 High throughput screening hits
`
`High throughput screening led to the identification of surprisingly few hits, and
`only two, P-aminoacyl amide 9 and piperazine 10 (Figure 4), were deemed wor(cid:173)
`thy of extensive follow-up. Amide 9, with a DPP-4 IC50 of 1.9 11M, was originally
`prepared in-house for the MRL thrombin inhibitor program. As such, amide 9
`inhibited thrombin with an IC50 of 52 nM. SAR evident from screening and initial
`exploratory chemistry is shown in Figure 4. Both the benzyl and phenethyl ana(cid:173)
`logs of 9 had DPP-4 inhibitory activity, and the (R) stereochemistry at the amino
`center was preferred. The second hit 10 (DPP-4 IC50 = 11 11M) was prepared as
`part of a proprietary screening library. A variety of groups was tolerated on the
`
`·ear(cid:173)
`)pa(cid:173)
`that
`·mer
`
`1tra-
`1ody
`a ted
`were
`rved
`also
`P-8.
`)P-9,
`was
`hese
`r the
`indi(cid:173)
`thus
`
`f the
`with
`and
`com(cid:173)
`P728,
`te. To
`con(cid:173)
`nhib(cid:173)
`·alent
`cause
`intra(cid:173)
`lity is
`:o the
`
`DPP(cid:173)
`those
`y less
`:eport
`1tency
`
`Boehringer Ex. 2012
`Mylan v. Boehringer Ingelheim
`IPR2016-01564
`Page 10
`
`
`
`102
`
`A.E. Weber and N. Thornberry
`
`R stereochemistry
`
`H
`
`pmre~:: t 0 ~x o~c:'<j
`l ) 0 Nl?""'
`
`~ /
`
`phenethyl analo:
`active
`
`g
`
`..7
`
`primary amine
`,::?'
`1
`/
`required ~
`~ no stereochemical preference
`
`n
`()~ t ~NJN~NHS02Me
`
`H2N~
`
`amides
`inactive
`
`H
`'----...:_:___--... ,.----__/
`N-phenyl and N-benzyl acetamides,
`indolylethylamines active
`unsubstituted derivatives > 100 f.lM
`
`Cl
`
`10
`
`Figure 4 HTS hits ~-aminoacyl amide 9 and piperazine 10.
`
`"right hand side" of the piperazine moiety, although unsubstituted derivatives
`were substantially less potent. The primary amine was strictly required, and
`conversion of the reduced phenylalanine to a phenylalanine amide led to com(cid:173)
`plete loss of potency. Both of these hits were developed simultaneously, even(cid:173)
`tually merging into one lead series.
`
`5.4 SAR in the ~-aminoacyl amide series
`
`Replacement of the ~-aminoacyl moiety with an a-aminoacid derivative such as
`isoleucyl or cyclohexylglycylled to a 2- to 4-fold decrease in potency. This was
`the first indication that SAR between this series and the a-amino acid series was
`distinct. Early on it was discovered that the "right hand side" amide could be
`replaced with an ester or acid moiety. This result led to a more systematic
`exploration of acid substitutions. Ortho-, meta- and para-substituted phenylacetic
`acid derivatives were prepared, and the latter analog (11, Figure 5) proved to be
`the first submicromolar inhibitor prepared in this series (IC50 = 510 nM). Grati(cid:173)
`fyingly, 11 was devoid of thrombin inhibitory activity [31].
`Because the entire proline amide moiety could be replaced with a thiazolidine
`without significant loss of activity, much of the initial SAR studies on the
`~-aminoacyl portion of the molecule was done in the thiazolidide series [32].
`Shortening or lengthening the distance between the primary amine and the phe(cid:173)
`nyl ring led to decreased activity. Replacing the phenyl group with a heterocycle
`or saturated ring led to derivatives that were also much less potent. Substitutions
`on the phenyl ring were somewhat tolerated. The 2-fluorophenyl analog (12) had
`an IC50 of 931 nM, and the addition of a second and third fluorine led to further
`increases in potency. In particular, the 2,5-difluorophenyl and 2,4,5-trifluorophenyl
`analogs 13 and 14 inhibited DPP-4 with IC50s of 270 and 119 nM, respectively.
`Substitution with a 2-fluoro group in the fully elaborated phenylacetic acid
`series gave analog 15 (IC50 = 54 nM), with an even more dramatic boost in potency
`[31]. The acetic acid was readily replaced with (25)-lactic acid to provide inhibitor
`16 which had an IC50 of 12 nM (Figure 6). More lipophilic groups at the carbon a to
`the acid are preferred, with the (S)-isopropyl substitution optimal. SAR of 2,5-
`difluoro analog 17, a subnanomolar DPP-4 inhibitor, is summarized in Figure 6.
`
`Figure 5
`
`Figure 6
`
`Ami
`(DPP-4
`DPP-9).
`sented,
`com pot
`marily
`high (1!
`SAR frc
`a hybri<
`
`5.5 SAl
`
`Initial ~
`nitroge1
`ration o
`in pate
`pharma
`creases
`substitu
`eties bo
`major b
`hand si
`increasE
`
`Boehringer Ex. 2012
`Mylan v. Boehringer Ingelheim
`IPR2016-01564
`Page 11
`
`
`
`ce
`
`com(cid:173)
`even-
`
`Case History
`
`103
`
`~COH
`R'~ 0~
`~"' B = l
`""'-
`~ <.fV'v
`""'-I
`I
`N
`Lx
`
`R"
`
`,;:?'
`
`R
`
`NH2 0
`
`2
`
`1
`
`11(R=R'=R"=H,R"'=B,X CH2)
`12 (R = F, R' = R" = R"' = H, X= S)
`13 (R = R" = F, R' = R'" = H, x = S)
`14 (R = R' = R" = F, R"' = H, X= S)
`15 (R = F,R' = R" = H, R'" = B, X= CH 2)
`
`Figure 5 Analogs of p-aminoacyl amide hit 9.
`
`fluorine substitution optimal
`
`short tethers preferred
`acid/acid mimic allowed
`substitution
`not well tolerated ~
`
`~~F ~. ~OrC02H
`
`y ~'--r-----1
`0
`
`0
`
`:::
`
`F
`
`lipophilic groups preferred
`(S) more potent than (R)
`
`16 (R =Me) ~ s also allowed
`17 (R = CHMe2)
`Figure 6 SAR of p-aminoacyl amide lead.
`
`Aminoacyllead 17 had several desirable properties. It was exquisitely potent
`(DPP-4 IC50 = 0.48 nM) and highly selective (IC50s > 100 ~-tM at QPP, DPP-8 and
`DPP-9). In addition, in an increasingly competitive area of research, it repre(cid:173)
`sented a unique structural class. Unfortunately, oral bioavailability in rats of this
`compound and many others in this series was very low (~1% for 17) due pri(cid:173)
`marily to poor absorption, and clearance after IV administration was generally
`high (150 mL/min/kg for 17). This lead was not progressed further, but rather
`SAR from this series was eventually incorporated into screening hit 10 to provide
`a hybrid lead series (vide infra).
`
`5.5 SAR in the piperazine series
`
`Initial SAR derived from screening suggested that substitution on the terminal
`nitrogen of the piperazine was tolerated; however, extensive systematic explo(cid:173)
`ration of SAR in this region of the molecule did not lead to substantial increases
`in potency. Recognizing that both leads shared a common phenethylamine
`pharmacophore, substitution of the phenyl ring by fluorine, which led to in(cid:173)
`creases in potency of the amide hit, was examined. In this series as well, fluorine
`substitution gave compounds with increased potency suggesting that these moi(cid:173)
`eties bound in the same site of the enzyme. With this information in hand, a
`major breakthrough was achieved by replacing the reduced phenylalanine "left
`hand side" with the corresponding homophenylalanine, leading to a 100-fold
`increase in potency [33].
`
`Boehringer Ex. 2012
`Mylan v. Boehringer Ingelheim
`IPR2016-01564
`Page 12
`
`
`
`104
`
`A.E. Weber and N. Thornberry
`
`In this new ~-aminoacyl amide series, the (R) benzyl isomer is ~50-fold more
`potent than the (S) isomer, and 2-fluoro substitution on the phenyl ring provides
`inhibitor 18 which had an IC50 of 14 nM (Figure 7). Truncation of the right-hand
`side to provide monosubstituted piperazine 19 (IC50 = 140 nM) led to only a
`10-fold loss in potency. Morpholine analog 20 had similar potency, though the
`piperidine amide was much less active (IC50 = 1040 nM). Addition of fluorines to
`the phenyl ring gave a further boost in activity, with the 2,4,5-trifluoro derivative
`(21, IC50 = 19 nM) being the most potent analog in this series.
`Despite the low molecular weight of many of these derivatives oral bioavail(cid:173)
`ability in rats was low and variable. This was traced in part· to extensive meta(cid:173)
`bolism, in particular, on the piperazine ring. In order to stabilize this ring to
`oxidation, bicyclic derivatives were prepared [34,35]. A wide variety of hetero(cid:173)
`bicyclic amides was evaluated. The initial compounds in this series included
`imidazolopiperazine 23 (IC50 = 640 nM) and triazolopiperazine 24 (IC50 = 460 nM).
`A comparison of these bicyclic analogs with the unsubstituted piperazine 22
`(IC50 = 3100 nM) showed that cyclization led to an increase in potency. Initial work
`was done in the 3,4-difluorophenyl series as shown in Figure 8; the requisite
`~-amino acid starting material is commercially available, and much of the SAR of
`fluorine substitution was developed in parallel.
`Substitution of hydrogen for ethyl in the triazolopiperazine lead gave analog
`25 (IC50 = 230nM), which was extensively profiled. This analog had weak
`activity at both DPP-8 and DPP-9 (IC50s = 45 and 100 f..lM, respectively). It was