Journal of Medicinal Chemistry - ACS Publications - American

Subscriber access provided by University of Otago Library

Article

Discovery of N-[bis(4-methoxyphenyl)methyl]-4-hydroxy-2(pyridazin-3-yl)pyrimidine-5-carboxamide (MK-8617) an Orally Active Pan-Inhibitor of Hypoxia-Inducible Factor Prolyl Hydroxylase 1-3 (HIF PHD1-3) for the Treatment of Anemia. John S Debenham, Christina B. Madsen-Duggan, Matthew J. Clements, Thomas F. Walsh, Jeffrey T Kuethe, Mikhail Reibarkh, Scott P. Salowe, Lisa M. Sonatore, Richard Hajdu, James A. Milligan, Denise M. Visco, Dan Zhou, Russell B. Lingham, Dominique Stickens, Julie A. DeMartino, Xinchun Tong, Michael Wolff, Jianmei Pang, Randy R. Miller, Edward C. Sherer, and Jeffrey J. Hale J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b01242 • Publication Date (Web): 09 Nov 2016 Downloaded from http://pubs.acs.org on November 10, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Medicinal Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

Discovery of N-[bis(4-methoxyphenyl)methyl]-4-hydroxy-2-(pyridazin-3yl)pyrimidine-5-carboxamide (MK-8617) an Orally Active Pan-Inhibitor of Hypoxia-Inducible Factor Prolyl Hydroxylase 1−3 (HIF PHD1−3) for the Treatment of Anemia.

John S. Debenham*, Christina Madsen-Duggan, Matthew J. Clements, Thomas F. Walsh, Jeffrey T. Kuethe, Mikhail Reibarkh, Scott P. Salowe, Lisa M. Sonatore, Richard Hajdu, James A. Milligan, Denise M. Visco, Dan Zhou, Russell B. Lingham, Dominique Stickens, Julie A. DeMartino, Xinchun Tong, Michael Wolff, Jianmei Pang, Randy R. Miller, Edward C. Sherer, and Jeffrey J. Hale.

Merck Research Laboratories, Merck & Co., Inc., PO Box 2000, Rahway, NJ 07065, United States. ABSTRACT: The discovery of novel 4-hydroxy-2-(heterocyclic)pyrimidine-5-carboxamide inhibitors of hypoxia-inducible factor (HIF) prolyl hydroxylases (PHD) are described. These are potent, selective, orally bioavailable across several species and are active in stimulating erythropoiesis. Mouse and rat studies showed hematological changes with elevations of plasma EPO and circulating reticulocytes following single oral dose administration while 4 week QD po administration in rat elevated hemoglobin levels. A major focus of the optimization process was to decrease the long half-life observed in higher species with early compounds. These efforts led to the identification of 28 (MK-8617) which has advanced to human clinical trials for anemia. Key Words: PHD inhibitors, Prolyl hydroxylase inhibitors, HIF, hypoxia inducible factor, erythropoietin, epo, erythropoietin stimulating agents, reticulocytes, anemia.

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 36

INTRODUCTION Anemia is a condition of insufficient red blood cells (RBCs) or hemoglobin (Hb) levels that result in reduced functional capability, fatigue, and shortness of breath. The incidence of anemia increases with the severity of renal insufficiency progressing to most patients with stage 5 chronic kidney disease (CKD).1 Cancer patients also suffer high prevalence of anemia at >40% with this approaching 90% in patients on chemotherapy.2 The current standards of care include RBC/blood transfusions for immediate interventions and the parenteral administration of recombinant human erythropoietin (rhEPO) agents such as epoetin alfa or darbepoetin alfa that stimulates the division, differentiation and maturation of erythroid progenitors in bone marrow over time.3 A small molecule stimulator of endogenous erythropoietin production would be less costly and obviate the complexity of parenteral administration of biologics. Hypoxia-inducible factor (HIF) is a α/β heterodimeric gene transcription factor involved in erythropoiesis, angiogenesis, glycolytic energy metabolism and apoptosis.4 Levels of HIF are regulated metabolically via hydroxylation of proline residues on the α-subunit by a family of hydroxylases known as prolyl hydroxylases (HIF-PHDs).

HIF is short-lived under conditions of normoxia due to oxidative

degradation of the HIFα subunit via the action of HIF PHD which require oxygen as a substrate. Hypoxia or inhibitors of the three isoforms of HIF-PHDs (1-3) stabilize HIF and stimulate the production of red blood cells (RBC's) through the modulation of erythropoietin (EPO), the EPO receptor, and proteins responsible for iron handling and transport.5 The foundational work of William G. Kaelin, Jr., Peter J. Ratcliffe and Gregg L. Semenza for unraveling the molecular basis of oxygen sensing and regulation through the HIF pathway was recently recognized with the 2016 Albert Lasker Basic Medical Research Award.6 The use of inhibitors of HIF PHDs to stimulate erythropoiesis in both animal and clinical studies has been previously described7 and reviewed.8 Several compounds are in clinical development with FibroGen’s PHD inhibitor FG-4592 (roxadustat) currently in phase III (Figure 1).9,10


Page 3 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Our research operating plan, a hierarchy of biological assays utilized for SAR optimization was detailed previously.11 After assessment of compounds in assays for PHD catalytic activity, compounds of interest were selected for pharmacodynamic (PD) evaluation in mouse pharmacodynamic erythropoietin determination assay (MoPED) to assess for serum EPO elevation 4 hr post iv dose.11 Downstream of EPO generation, reticulocyte elevations could be measured 72 hours post po dose and RBC increases enumerated 1-2 weeks following po dose. In general the compounds described herein displayed little selectivity for PHD subtypes (usually within 20 fold or less) and discussion will therefore highlight PHD2 activity. PHD1 & 3 IC50s are available in the supporting information.

Figure 1. Clinical compound 1 and uHTS screening hit 2. RESULTS AND DISCUSSION Our efforts started with a screen of the Merck compound collection identifying 2 as a fairly potent PHD2 inhibitor at 110 nM IC50 and MW of 324 (Figure 1). The core 4-hydroxy-2-(heterocyclic) pyrimidine-5-carboxamide scaffold 3 was represented by many members with a range of potency building confidence in the quality of the hit. Small molecule inhibitors of PHD2 (such as PDB complex 2G19 which contains an inhibitor similar to 1)15 form directed interactions to protein side chains of the active site and chelate the bound Fe. The Fe is hexa-coordinated to two histidines (H313 & H374), one aspartic acid (D315), and one water molecule which leaves open two coordination sites for heteroaryl interactions. Crystal structures of small bicyclic inhibitors demonstrate a tight salt bridge formed between the acid tail of the inhibitor and Arg383 at the base of a narrow and deep binding pocket. Noticeably absent in the chemical series described here is the acid tail which eliminates this salt bridge to arginine. ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 36

Molecular modeling (details in Methods) of the screening hit led us to develop a hypothesis that the hetero biaryl chelated Fe with the pyridine ring filling the space normally occupied by a carboxylic acid substituent of Fibrogen-like PHD inhibitors. The benzyl tail of the hit series can adopt three possible conformations interacting with the outer regions of the binding site (Figure 2). A water molecule forms a hydrogen bonded bridge between the oxyanion of the hydroxypryimidine and the phenolic hydrogen of Tyr303. A stacking interaction between Tyr310 and the terminal heteroaryl of the inhibitor is present. Amide substitution can display mono- or di-substituted branched aromatic rings which project out of the primary binding pocket and interact with both solvent and several aromatic residues: Trp258, Trp389, Arg322. Taken together, these observations supported further optimization of the hit to recruit additional directed interactions and the salt-bridge to Arg383. When comparing rotamers of the branched biaryl, docking scores for the three rotamers of the fluorophenol were roughly equivalent and the ligand would be expected to sample multiple conformations when bound. This observation led us to consider di- and tri-substitution on the methylene linker.


Page 5 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 2: Molecular models of the core scaffold to human PHD2 crystal structure 2G1915. Where multiple rotamers are possible for the branched aromatics, only one pose per structure has been displayed. Methylation of the methylene linker of 2 improved potency 4-fold for 10 (Table 1) . The substitution does not pick up directed interactions other than van der Waals contact to the protein so it is hypothesized that the gem-dimethyl preferentially adopts a conformation of the phenyl ring shown in Figure 2 to pick up an edge to face stacking interaction with Trp258. Building from 10, the affinity again improves when the terminal heteroaryl is changed from pyridine to pyrazole 7. There is a slight steric interaction between the top of the pyridine ring and Tyr303 which may influence an improvement in affinity when moving to smaller 5-membered heterocycles. While variation in chelation strength was expected for 5and 6-membered biaryls, quantum mechanical calculations quantifying variation in Fe chelation energetics did not improve interpretation of SAR and are not described further. When the methylene linkage is disubstituted with phenyl rings affinity is improved as evidenced by 25, 26, and 28 (Tables 3-4). Docking models for these ligands always favor one aryl ring stacked over Arg322 owing to cation-pi interactions; however, substitution of the terminal phenyls can lead to functionality which more favorably picks up directed interactions to other residues influencing rotameric variation in binding of the branched aromatics. With the observation that substitution of the terminal phenyl rings could provide additional directed interactions, we suggest that methoxy substitution 26 specifically led to hydrogen bonds to each methoxy oxygen, one from the indole NH of Trp258, and one from the Nδ of Asn318. Further aromatic substitution of 10 is unfavorable, as the trajectory of the last open position of the methylene linkage is directed at a suboptimal angle towards the Tyr389 and a third aromatic ring would clash with the protein. On the carboxamide portion of scaffold 3 wide structural diversity was tolerated in the hit class. The hits were primarily represented by large amides with significant structural flexibility. The early optimi-



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 36

zation strategy was to identify a reasonably efficient right hand side amide and then further refine the war head interaction with the heterocycle at R1 (Table 1). This could then be followed by another round of potency tuning for amide R2 (Table 3). After initial benchmarking of the hit, confidence continued to grow in the scaffold as 2 showed >30 µM off-target hERG activity which contrasts sharply with our laboratory’s previous HIF PHD inhibitor scaffold.11 Rat pharmacokinetic (PK) assessment showed a reasonable starting point with %F 67, Clp 2.0 ml/min/kg, Vdss 0.35 L/kg and t½ 3.3 hr. In terms of mouse PD, 2 showed modest activity with a minimum effective dose (MED) of 100 mg/kg (mpk) in MoPED resulting in an EPO concentration of 1200 pg/mL. Table 1. Heterocyclic 2-position SAR

PHD2 IC50 (nM)

MoPED pg/mL (mpk)

Rat PK %F, t1/2

4

24

110 (100 iv)

Nd

5

16

2100 (100 iv)

2.9 hr

6

210

Nd

Nd

7

3.7

460 (5 iv) 5000 (15 iv) 24000 (50 iv)

4.2 hr

8

2.1

1900 (15 iv)

2.1 hr

9

730

Nd

Nd

10

28

Nd

Nd

11

>2700

Nd

Nd

12

6.0

220 (15 iv)

0.23 hr

13

390

Nd

Nd

Compd

R1

61

99

3

5


Page 7 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


14

12

Nd

Nd

A quick survey of amines (data not shown) demonstrated that the cumyl amide 10 provided a 4-fold improvement in potency vs the screening hit to 28 nM PHD2 IC50. This enabled a starting point to anchor the carboxamide thereby allowing examination of the 2-position SAR. A general synthetic approach to the assembly of these compounds is exemplified for 28 in Scheme 1. Both 5 and 6 membered heterocycles were tolerated at this position. Modeling had suggested the presence of an acid on this ring system would make a strong interaction with Arg383. This indeed improved potency with both pyrazole 8 and pyridine 12 analogs showing improvement as low as 2 nM PHD2 IC50, but unfortunately was accompanied by very low rat bioavailability for both compounds. However as MoPED is IV dosed, improvement was apparent in 12 with the MED dropping to 15 mpk. For the 6 membered ring system, CF3 substitution of the pyridine ring 11 resulted in a large drop in activity. Diaza rings showed a wide range of activity. Pyrimidine 13 was particularly ill favored, but pyrazole 7 and pyridazine 14 delivered large potency improvements to 3.7 and 12 nM PHD2 IC50 respectively. The methylated pyrazole isomer 6 lost over 50-fold of activity. The thiazole analogs 4 and 5 had improved potency to 2, but still required the maximum feasible dose of 100 mpk to show an EPO elevation in MoPED. In contrast to the pyrazole acid, the unsubstituted pyrazole 7 displayed high bioavailability and a 30-fold improvement in potency relative to the screening lead 2. This resulted in a greatly improved PD response with MoPED MED dropping to 5 mpk. Table 2. Pharmacokinetic profile of 7 mouse* Rat* dog** rhesus** F (%) 98 99 61 71 cmax (µM) 30 49 11 22 Cl (ml/min/kg) 1.0 0.2 0.004 0.001 Vdss (L/kg) 0.3 0.06 0.1 0.13 t1/2 (hr) 3.7 4.2 >100 95 PPB (% unbound ) 9.3 1.2 2.7 0.4 *2 mg/kg po, 1 mg/kg iv; **1 mg/kg po, 0.5 mg/kg iv. PPB is from 100% serum; Mouse = Clb, rat, dog, rhesus Clp. Mouse: IV and P.O.: DMSO: PEG400: water (5:40:55, v/v/v). Rats: IV and P.O.: DMSO: PEG400: water (10:50:40, v/v/v). Dog and Monkey: IV: PEG200:12%HPCD: water (30:12:58 v/v/v); P.O.: Imwitor:Tween (1:1 w/w).



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 36

PK studies with 7 demonstrated robust oral absorption in all preclinical species tested (Table 2). However, in higher species, an exceptionally long t1/2 of >95 hours was observed in monkeys and dogs. In vitro incubations with both liver microsomes and hepatocytes from rat, dog, monkey and human displayed very little turnover. These data suggested the compound could have a very long human t1/2 which was not in keeping with our desire for a QD dosing paradigm.

It was therefore decided that

compounds would need to have shorter residence times in higher species in order to advance. Since 7 was low molecular weight (323) and not readily metabolized (as evident by little turnover in both microsome and hepatocyte incubations) the team anticipated that by increasing MW and adding structural features amenable to phase I or II metabolism would result in compounds with more appropriate ADME properties. As a first in vivo read out of shortened residence time in higher species, compounds of interest were screened for PK via dog cassette IV dosing. As the 2-pyrazole of 7 provided both potency and PD, attention focused towards 5-carboxamide optimization (Table 3). In order to get a better sense of the minimum pharmacophore the benzamide was replaced with methyl cyclopropyl amide 15 which resulted in a 40-fold loss of PHD2 activity. Returning to benzamides, 3-6 nM PHD2 IC50s could be achieved without the need for alpha methyl groups (16 or 17). Adding back one methyl group generated optically active amides 18-21 that showed a slight preference in potency for the S-isomer. While the potency difference was small there was a noted drop in PD activity for the R isomers 19 and 21. Regardless of the reason for decreased in vivo activity, the compounds, represented by 20, had a very long t1/2 in dog of 60 hours. Building in a metabolic soft-spot with the 4-methoxy group of 23 provided further improved potency and efficacy but no improvement in reducing dog t1/2. While adding in an extra phenyl ring as in biphenyl 24 and benzhydryl 25 amides provided 1-2 nM PHD2 potency, the dog t t1/2 moved in the wrong direction as exemplified by 24 with a 143hr t1/2. In an attempt to really leverage phase I metabolism the dimethoxybenzhydryl 26 was prepared. This compound had the lowest MED tested with significant EPO stimulation at a 1.5 mpk dose


Page 9 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


in mice. Off-target hERG counter-screening revealed that 26 was active at 900 nM IC50. This was a departure from the smaller amides that were generally inactive on hERG. Table 3. 5-Carboxamide SAR

PHD2 IC50 (nM)

MoPED pg/mL (mpk)

Dog PK t1/2 hr

15

150

Nd

Nd

16

2.9

1100 (15 iv)

Nd

17

5.6

180 (15 iv) 8900 (50 iv)

Nd

18

5.0

690 (5 iv) 10000 (15 iv)

Nd

19

8.1

322 (15 iv)

Nd

20

1.9

16000 (15 iv)

60

21

10

350 (15 iv)

Nd

22

4000

Nd

Nd

23

1.4

5300 (15 iv)

94

24

1.3

Nd

143

25

2.0

2900 (15 iv)

Nd

26

1.1

190 (1.5 iv) 740 (5 iv) 31000 (15 iv)

Nd

Compd

R2

Table 4. Pyridazine warhead leading to 28.

Compd

R3

PHD2 IC50

MoPED pg/mL (mpk)


hERG IC50


Page 10 of 36

(nM)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

380 (5 iv) >30 µM 1600 (15 iv) 360 (1.5 iv) 6200 (5 iv) 28* OCH3 1.0 12000 (15 iv) >30 µM 18000 (50 iv) 27000 (100 iv) * 28 PHD1 and 3 IC50 are 1.0 and 14 nM respectively. 27

H

2.5

To evaluate if the choice of warhead could impact hERG activity and PK, the ring sizes were changed from 5 to 6 with an expansion of pyrazole to pyridazine (Table 4). There was no change in PHD2 potency, but the hERG activity was eliminated in both 27 and 28. The robust PD response noted for 26 was further enhanced with 28 utilizing the pyridazine warhead resulting in a MoPED MED of 1.5 mpk for EPO elevation.

Even with maximal EPO responses (EPO concentration >10,000 pg/mL), no

changes in plasma VEGF were observed. Off target activity of 28 was evaluated to establish specificity. It was not a significant inhibitor of the cytochrome p450 enzymes in vitro (IC50): CYP1A2, 3A4, 2B6, 2C9, 2C19, or 2D6 >60 µM and was a moderate reversible inhibitor of CYP2C8 at 1.6 µM in vitro.

Probing more broadly, 28 was inactive

when screened at 10 µM against a general panel of 171 radioligand binding and enzymatic assays.12 Additionally in terms of a related enzyme in the same pathway, the IC50 of 28 was determined for factor inhibiting HIF (FIH) to be 18 uM suggesting good selectivity of 28 for the target. Also of note was that, in mouse, there were no elevations in plasma ALT levels below 200 mg/kg which had been observed in our previously disclosed scaffold.11 Tritiated 28 exhibited minimal metabolic turnover in liver microsomes (+NADPH) from rat, dog and monkey (

Journal of Medicinal Chemistry - ACS Publications - American

Recommend Documents