A Fragment-Derived Clinical Candidate for Antagonism of X-Linked and Cellular Inhibitor of Apoptosis Proteins: 1-(6-[(4-Fluorophenyl)methyl]-5-(hydroxymethyl)-3,3-dimethyl-1 H,2 H,3 H-pyrrolo[3,2- b]pyridin-1-yl)-2-[(2 R,5 R)-5-methyl-2-([(3R)-3-methylmor

Article

pubs.acs.org/jmc

Cite This: J. Med. Chem. XXXX, XXX, XXX −XXX

A Fragment-Derived Clinical Candidate for Antagonism of X‑Linked

and Cellular Inhibitor of Apoptosis Proteins: 1‑(6-[(4-

Fluorophenyl)methyl]-5-(hydroxymethyl)-3,3-

dimethyl‑1H,2H,3H‑pyrrolo[3,2‑b]pyridin-1-yl)-2-[(2R,5R)‑5-methyl-

2-([(3R)-3-methylmorpholin-4-yl]methyl)piperazin-1-yl]ethan-1-one

(ASTX660)

Christopher N. Johnson,* Jong Sook Ahn, Ildiko M. Buck, Elisabetta Chiarparin, James E. H. Day,

Anna Hopkins, Steven Howard, Edward J. Lewis, Vanessa Martins, Alessia Millemaggi, Joanne M. Munck,

Lee W. Page, Torren Peakman, Michael Reader, Sharna J. Rich, Gordon Saxty, Tomoko Smyth,

Neil T. Thompson, George A. Ward, Pamela A. Williams, Nicola E. Wilsher, and Gianni Chessari

Astex Pharmaceuticals, 436 Cambridge Science Park, Milton Road, Cambridge CB4 0QA, United Kingdom

S

* Supporting Information

ABSTRACT: Inhibitor of apoptosis proteins (IAPs) are

promising anticancer targets, given their roles in the evasion

of apoptosis. Several peptidomimetic IAP antagonists, with

inherent selectivity for cellular IAP (cIAP) over X-linked IAP

(XIAP), have been tested in the clinic. A fragment screening

approach followed by structure-based optimization has

previously been reported that resulted in a low-nanomolar

cIAP1 and XIAP antagonist lead molecule with a more

balanced cIAP−XIAP proﬁle. We now report the further

structure-guided optimization of the lead, with a view to

improving the metabolic stability and cardiac safety proﬁle, to

give the nonpeptidomimetic antagonist clinical candidate 27

(ASTX660), currently being tested in a phase 1/2 clinical trial

(NCT02503423).

antiapoptotic activity is mediated by direct binding to and

inactivation of caspases via BIR domains.⁸IAP antagonists

such as the endogenous second mitochondria derived activator

of caspases (SMAC) bind competitively to IAP BIR domains,

thus disrupting interactions such as those between XIAP and

caspase 9.⁹On binding to cIAP1 and cIAP2, SMAC activates

E3 ligase function leading to autoubiquitination and

proteasomal degradation.¹⁰Loss of cIAP, together with the

release of the XIAP-mediated caspase block, leads to a

sustained pro-apoptotic eﬀect in the presence of TNF-α via

the extrinsic apoptosis pathway. Hence, tumors with suﬃcient

levels of TNF-α in their environment may be highly

susceptible to IAP antagonism.¹¹Furthermore, antagonism of

the XIAP-mediated caspase block promotes apoptosis induced

by stimulation of the intrinsic apoptosis pathway by agents

INTRODUCTION

■

One of the hallmarks of cancer is evasion of apoptosis,¹

achievable through overexpression of antiapoptotic proteins.

Inhibitor of apoptosis proteins (IAPs), such as cellular IAP1

and 2 (cIAP1 and cIAP2 respectively) and X-linked IAP

(XIAP), are key regulators of antiapoptotic and pro-survival

signaling. XIAP directly inhibits caspases, cysteine proteases

with an essential role in the apoptotic cascade. In contrast,

cIAPs prevent the formation of pro-apoptotic signaling

complexes such as complex-II, composed of Fas-associated

protein with death domain (FADD), procaspase-8, and

receptor-interacting serine/threonine-protein kinase 1

(RIPK1). These eﬀects cause suppression of apoptosis through

both extrinsic and intrinsic apoptosis pathways.²⁻⁴IAP

dysregulation occurs in various cancers and is associated with

tumor growth and poor prognosis, rendering IAPs attractive

targets for anticancer therapy.⁵IAPs are characterized by one

or more baculovirus IAP repeat (BIR) domains which mediate

protein−protein interactions. Some family members (e.g., cIAP

and XIAP) also possess Really Interesting New Gene (RING)

zinc ﬁnger domains with E3 ubiquitin ligase activity.^6,7XIAP

such as chemotherapeutics or DNA damaging agents.¹²

A

cIAP/XIAP antagonist might therefore be used to enhance

apoptosis through both extrinsic and intrinsic pathways.

Received: June 6, 2018

A

DOI: 10.1021/acs.jmedchem.8b00900

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Previously, IAPs have been targeted by antisense oligonu-

cleotides and antagonist small molecules.⁴Pioneering research

has led to ﬁrst generation SMAC-mimetic IAP antagonists

such as 1−4, with designs based on AVPI, the N-terminal

sequence of SMAC, and these have shown promising

preclinical activity.¹³⁻¹⁶However, all of these compounds are

selective for cIAP.

Table 1. Cross-Species Pharmacokinetic Parameters for

Compound 5

CL_p

V_ss

C_max(PO)

a

species

mouse

mL min⁻¹kg⁻¹

L kg⁻¹

ng mL⁻¹

F%

41

26

17

6.4

12

120

47

22

24

2

rat

cynomolgus

monkey

3.7

22

a

Plasma clearance (CL_p), volume of distribution at steady state (V_ss),

maximum concentration in plasma (C_max) and oral bioavailability (F)

were determined for each species using an intravenous (IV) dose of 1

mg/kg and an oral (PO) dose of 5 mg/kg. For assay details see

Experimental Section.

liver blood ﬂow), but oral bioavailability at 5 mg/kg was only

2%. The low oral bioavailability in NHP indicated a potential

for suboptimal exposure in humans. Furthermore, the low peak

concentration obtained in plasma (C_max) suggested that a key

goal would be to obtain a compound capable of attaining

plasma drug levels in the NHP species equivalent to or greater

than those producing a pharmacodynamic response in

mouse.¹⁸In addition, compound 5 was assessed for potential

cardiac toxicity in a I_Krpotassium channel (hERG) population

patch clamp (PPC) assay, and it showed inhibition with IC₅₀9

μM. Therefore, a further round of optimization was under-

taken starting from 5, with the objective of improving PK

proﬁle in the NHP and reducing hERG inhibition.

In order to understand the oral exposure in cynomolgus

monkey, the permeability and eﬄux potential of 5 were

evaluated in a Caco2 monolayer assay (eﬄux ratio A-B/B-A =

36 and permeability = 1.1 × 10⁻⁶cm s⁻¹), suggesting that

metabolism may explain the in vivo observations better than

permeability. Compound 5 had signiﬁcant in vitro intrinsic

clearance in microsomes and hepatocytes across species, with

the exception of rat (Supporting Information Table S2), and

the low stability in monkey microsomes was notable.

Compound 5 also showed signiﬁcant in vivo clearance across

species, so the route of metabolism was investigated in mouse,

rat, NHP and human liver hepatocytes (Figure 1). Assuming

the mass spectrometric response of the parent compound and

proposed metabolites is similar, the main metabolic route was

identiﬁed as the opening of the morpholine ring to give

metabolite M1 (<7% response compared to parent in mouse

and human, 20% compared to parent in NHP), followed by

conversion to the primary amine derivative (metabolite M2).

Piperazine ring-opened amine M4 was identiﬁed as a minor

metabolite (<1% compared to parent in mouse and human, 4%

compared to parent in NHP). A further minor oxidation

metabolite M3 of compound 5 was identiﬁed in mouse, NHP

and human hepatocytes, though the site of oxidation was not

elucidated. No metabolites were observed in rat hepatocyte

preparations due to the low metabolic turnover of 5 in this

system.

We previously reported a fragment-based drug design

(FBDD) approach to the discovery of a nonpeptidomimetic,

low nanomolar, orally bioavailable antagonist of cIAP1 and

XIAP, designated AT-IAP 5,^17,18starting from the millimolar-

aﬃnity fragment 6. Compound 5 had particularly high aﬃnity

for XIAP and a more balanced cIAP1/XIAP proﬁle (∼16-fold)

compared to AVPI-based antagonists 1−4, where selectivity

for cIAP1 versus XIAP ranged from 39- to 227-fold.^18,19This

selectivity observation was structurally well understood in

terms of the interactions made between cIAP1 or XIAP and

the piperazine moiety of 5, compared with those made by the

terminal N-methylated alanine of the ﬁrst generation

peptidomimetic compounds. Here we describe further

medicinal chemistry optimization of advanced lead 5 to give

ASTX660 27,¹⁹which is now being evaluated in a phase 1/2

clinical trial in cancer patients.

RESULTS AND DISCUSSION

■

Advanced lead 5 was a useful chemical probe, having oral

bioavailability in the mouse (22%) and antitumor growth

activity in mouse MDA-MB-231 and A375 xenograft models.¹⁸

The compound also had moderate plasma clearance (26 mL

min⁻¹kg⁻¹) and oral bioavailability in the rat (24%) (see Table

1). It was important to characterize the pharmacokinetic (PK)

proﬁle in a nonrodent species, in consideration of the need to

perform toxicological evaluation on the compound as a

potential drug candidate. It has been previously reported that

dogs are very sensitive to NF-κB pathway eﬀects caused by

antagonism of cIAP1,²⁰and therefore a nonhuman primate

(NHP) may be a more appropriate nonrodent species for

toxicology studies. We therefore determined the PK proﬁle of

5 in cynomolgus monkey (Table 1) via the oral and

intravenous route. Plasma clearance was moderate (<50%

Cytochrome P450 (CYP) phenotyping was also carried out

in human supersomes in order to determine the enzymes

primarily responsible for compound turnover. Importantly, the

observed metabolism was found to be mediated by CYP3A4.

Cynomolgus monkeys are known to highly express CYP3A8 in

the gut,²¹which can lead to low exposure of compounds

metabolized by CYP3A enzymes. We therefore hypothesized

that extensive gut metabolism in NHP was an important

contributor to the low oral bioavailability of 5 in this species,

B

DOI: 10.1021/acs.jmedchem.8b00900

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Figure 1. Proposed route of metabolism of compound 5 in mouse, NHP, and human hepatocytes.

Figure 2. (A) X-ray crystal structure of compound 5 (PDB code 5M6L) bound to XIAP with pockets P1−P4 labeled. Connolly surface of protein is

colored by electrostatic potential (blue = positive, red = negative), hydrogen bonds are displayed as dashed red lines, and crystallographic water is

shown as a red sphere. (B) Aronov hERG pharmacophore overlaid on bound conformation of compound 5. Pharmacophore points are shown as

color-coded mesh spheres according to the legend, with distances between features marked by dashed red lines.

and thus incorporated a CYP3A4 turnover assay into the

screening cascade for lead optimization.

ﬁlling the P1 pocket (Figure 2A), so the optimization strategy

required that this part of the molecule should not be changed.

Examination of the binding mode shows that the metabolically

labile methylmorpholine group occupies the P2 pocket and

that space exists for further variation in this part of the

molecule. The azaindoline bicycle stacks on a ﬂat part of the

protein and also occupies P3 by virtue of the gem-dimethyl

substitution, while the ﬂuorobenzyl ﬁlls P4. The structure

suggests that there may be a limited scope for increasing the

size of the P4 substituent, but that some modiﬁcations may be

tolerated in the bicycle, subject to satisfying the previously

described requirement to match the electrostatic potential of

the aryl portion to the presence of an electronegative patch on

the protein surface.¹⁷

To assess hERG activity, we screened using the PPC assay²²

to identify compounds with lower risk of QT interval

prolongation in vivo. We assessed percentage inhibition at

three concentrations (10, 30, and 100 μM) and determined

PPC IC₅₀values for key compounds of interest. For simplicity,

the subsequent data tables show only data obtained at 30 μM;

compound 5 itself gave 70% inhibition at 30 μM.

During the fragment to lead campaign that led to compound

5, crystal structures had been used primarily to drive potency

increases by growing eﬃciently into pockets. However, the

principal objectives were now diﬀerent: to improve the

metabolism and cardiac proﬁle. Structural understanding was

nonetheless essential in identifying parts of the molecule where

changes might be tolerated as opposed to those that would

likely result in potential reduction of binding aﬃnity. For

example, the piperazine amide core of 5 is known to make

several important polar interactions to the protein as well as

In order to attenuate the hERG activity, we considered how

compound 5 might map to a number of proposed hERG

pharmacophore models. In the model proposed by Aronov,²³

a

number of known hERG ligands have an H-bond acceptor

positioned between positively charged nitrogen and hydro-

C

DOI: 10.1021/acs.jmedchem.8b00900

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Table 2. Initial SAR Exploration around Compound 5

a

b

c

Cellular XIAP-caspase 9 immunoprecipitation assay. Intracellular degradation of cIAP1 in MDA-MB-231 cells. MDA-MB-231 cell proliferation

d

assay (cIAP1 sensitive). Not tested. For all assay details see Experimental Section.

phobic centroid. The three-dimensional structure of 5 in its

bound conformation (Figure 2B) has all three elements in a

similar geometric relationship. Our design strategy then

considered how one or more of these pharmacophore elements

could be perturbed in order to attenuate binding to hERG.

The binding mode of 5 in XIAP (Figure 2A) suggests that

judicious placement of polarity in the region of the

hydrophobic centroid could be a viable option to decrease

hERG channel inhibition while retaining target potency.

We thus had two design goals: the ﬁrst was to reduce

metabolism by making changes to the P2 morpholine moiety,

and the second was to introduce polarity in the region of the

P4 group and azaindoline heterobicycle. Table 2 shows

modiﬁcations to the P2 group aimed at improving metabolic

stability. The eﬀect of these changes on cellular XIAP and

cIAP1 potency, together with antiproliferative eﬀect in the

MDA-MB-231 cell line (sensitive to cIAP1 inhibition) is

shown, alongside compound half-life on incubation with

CYP3A4 supersomes, together with hERG PPC inhibition

data. Cell-based assays for cIAP1 and XIAP were as previously

reported:¹⁸for XIAP, a XIAP-caspase 9 immunoprecipitation

(IP) assay was used in a HEK293 cell line overexpressing XIAP

and caspase 9, while for the cIAP1 assay, intracellular cIAP1

degradation in MDA-MB-231 cells was measured as a

surrogate for cIAP1 aﬃnity.

Examination of the P2 pocket environment in the X-ray

crystal structure of 5 bound to XIAP indicated that attempts to

block metabolism in the morpholine ring by adding increased

steric bulk should be tolerated in terms of binding aﬃnity.

Therefore morpholine derivatives bearing additional methyl

substituents were synthesized 7−9. Of these, 8 had exceptional

potency at both XIAP and cIAP1 and had improved potency in

the MDA-MB-231 cell proliferation assay compared to 5.

Previously, promising activity had been seen with a P2 2-

pyrrolidinone substituent,¹⁸where a folded conformation

necessary for binding was favored by the complementary

electrostatic potentials of the amide region of the piperazine-

azaindoline scaﬀold and the lactam carbonyl. We therefore

introduced a carbonyl into the morpholine group to give 10.

Although this compound had relatively modest XIAP potency

and MDA-MB-231 antiproliferative activity, both CYP3A4

turnover and hERG activity were reduced. Further exploration

of lactams was undertaken, aiming to ﬁll additional space in the

P2 region while continuing to favor a folded conformation. It

was thought that extending the π-system of the lactam

carbonyl might promote hydrophobic collapse-induced folding,

so the unsaturated lactam 11 and benzolactam 12 were

prepared. Both had improved CYP3A4 stability compared to 5,

but the eﬀect on hERG was marginal. Benzolactam 12 indeed

adopted the desired folded conformation in the XIAP X-ray

D

DOI: 10.1021/acs.jmedchem.8b00900

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Figure 3. (A) X-ray crystal structure of benzolactam 12 bound to XIAP. (B) X-ray crystal structure of compound 21 bound to XIAP with ligand

electron density map (F_o− F_cmaps contoured at 1σ and clipped around the ligand) shown as magenta mesh. (C) Protein structure of cIAP1

overlaid with XIAP − compound 21 complex in same frame of reference. For clarity, XIAP protein is not displayed. The salt bridge between

Arg308 and Asp297 is shown as dashed red lines. The potential clash between Arg 308 and ligand benzylic OH is shown as a dashed cyan line. For

(A), (B), and (C), the displayed protein Connolly surface is colored by electrostatic potential (blue = positive, red = negative).

crystal structure (Figure 3A), but target potency was slightly

decreased. However, we reasoned that the benzolactam P2

moiety might oﬀer an opportunity to introduce a group

capable of disrupting the hERG pharmacophore, especially if it

could be established that a folded conformation predominates

guide the optimization more eﬀectively. Chromatographic

logD (ChrlogD) measurements were therefore performed at

pH 7.4 on the same compound set (see Experimental Section),

and the results were plotted against hERG inhibition in the

PPC assay and half-life in the CYP3A4 turnover assay (Figure

4, panels A and B respectively). Here the correlation was much

1

in solution. In the H NMR spectrum of 12 in CD₃OD, the

chemical shifts of the methyl groups on the azaindoline 5-

membered ring were signiﬁcantly diﬀerent (3H, δ1.34 and 3H,

δ1.48), suggesting a perturbation of one resonance due to

proximity of an aromatic group. Such a situation could arise if

the benzolactam group was folded in the manner previously

seen for the simple pyrrolidinone and the morpholine of

compound 5. NMR spectroscopy on 12 dissolved in 50 mM

pH 7.4 phosphate buﬀer conﬁrmed that observations in

solution are consistent with the X-ray conformation (see

Supporting Information, Figure S1).

Having identiﬁed a bicyclic P2 substituent of potential

interest, we investigated the eﬀect of adding an additional

nitrogen to the azaindoline moiety as a means of disrupting the

hERG pharmacophore. Diazaindoline 13 encouragingly had

good CYP3A4 stability and reduced hERG inhibition, and

replacement of the P4 4-ﬂuorobenzyl with 2,4-diﬂuorobenzyl

14 restored good target potency without signiﬁcant detriment

to the CYP3A4 and hERG proﬁle. The PK proﬁle of

compound 14 was characterized in mouse and cynomolgus

monkey. Oral bioavailability in mouse was low (2% at 10 mg/

kg), but in monkey was signiﬁcantly improved compared with

original lead 5 (18% at 5 mg/kg). In the PPC hERG assay, 14

had low inhibition at 30 μM, suggesting an improved cardiac

proﬁle and respectable exposure in NHP were achievable

concurrently within the overall chemotype. Nevertheless,

further optimization was required in order to address the

low oral exposure in mouse and the relatively unbalanced IAP

antagonism proﬁle of 14 (>30 fold selectivity for cIAP1 over

XIAP) compared to lead compound 5.

At this stage of the optimization, a total of 53 compounds

had been analyzed for hERG and 32 compounds for CYP3A4.

With respect to physicochemical properties, a broad trend

could be observed with respect to C logP, where highly

lipophilic compounds had the least favorable proﬁles.

However, the overall correlation was weak, and it was diﬃcult

to set a target value for C logP that could usefully guide further

optimization. Metabolic stability and hERG inhibitory liability

are often inﬂuenced by lipophilicity,²⁴and therefore we were

keen to identify an alternative lipophilicity measure that would

Figure 4. (A) Percentage hERG inhibition at 100 μM in the

population patch clamp (PPC) assay plotted for diﬀerent ranges of

chromatographic LogD (ChrlogD). Total numbers of compounds per

inhibition category are displayed in each segment. (B) Half-life (T_1/2

)

in the CYP3A4 turnover assay plotted for the same ChrlogD ranges,

with total numbers per T_1/2category displayed in each segment (C).

Scatter plot of C logP versus ChrlogD colored by P2 side chain

subtype.

improved and, for both hERG inhibition and CYP3A4 stability,

the data suggested that compounds with ChrlogD ≤ 3 should

be targeted in order to maximize the chance of success. In

comparison, compound 5 had a ChrlogD value of 3.7. Figure

4C shows the relationship between C logP and ChrlogD for

compounds bearing morpholine and lactam P2 side chains. It

can be seen that morpholine containing compounds cluster to

E

DOI: 10.1021/acs.jmedchem.8b00900

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Table 3. Benzolactam Subseries SAR

cellular IC₅₀(nM)

a

b

c

d

Cpd

R

ChrlogD

XIAP

cIAP1

MDA-MB-231

CYP3A4 t_1/2(min)

hERG PPC (%inh. at 30 μM)

14

15

16

17

18

19

H

4-F

5-F

6-F

5-CN

6-CN

2.4

2.7

2.6

3.1

3.0

2.9

12

0.34

0.17

0.20

0.74

1.0

5.4

0.59

1.9

2.9

5.2

>70

37

41

42

27

28

25

18

19

18

4.3

3.5

7.5

16

5.8

1.6

14

34

a

b

c

Chromatographic LogD measured at pH 7.4. Cellular XIAP-caspase 9 immunoprecipitation assay. Intracellular degradation of cIAP1 in MDA-

d

MB-231 cells. MDA-MB-231 cell proliferation assay (cIAP1 sensitive). For all assay details see Experimental Section.

Table 4. Morpholine Subseries SAR

a

b

c

Chromatographic LogD measured at pH 7.4. Cellular XIAP-caspase 9 immunoprecipitation assay. Intracellular degradation of cIAP1 in MDA-

d

e

MB-231 cells. MDA-MB-231 cell proliferation assay (cIAP1 sensitive). Not tested. For all assay details see Experimental Section.

the left of the line of unity, whereas the lactams predominately

cluster to the right. In general, the morpholines are less

lipophilic than expected from the C logP value, possibly due to

partial protonation of the morpholine moiety, whereas the

lactams are more lipophilic than C logP would suggest.

Subsequent optimization focused on the P2 benzolactam

and P2 morpholine subseries. For the benzolactams, the

challenge was to identify changes that improved XIAP aﬃnity

while preserving the favorable CYP3A4 and hERG proﬁles.

Conversely, the morpholine subseries already had excellent

XIAP and cIAP1 proﬁles, so we adopted the strategy of

addition of polar functionality in order to ascertain whether the

morpholine ring metabolism and hERG activity in this

subseries could be resolved by ﬁne-tuning physicochemical

properties. Table 3 illustrates modiﬁcations made in the

benzolactam subseries.

Examination of the X-ray crystal structure of 12 bound to

XIAP (Figure 3A) suggested that addition of a substituent in

the benzene ring of the isoindolinone ring system would be

tolerated at C4, C5, and C6. A ﬂuorine scan was carried out,

F

DOI: 10.1021/acs.jmedchem.8b00900

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

resulting in compounds 15−17 with improved XIAP aﬃnity

and satisfactory proﬁles in the CYP3A4 and hERG assays

(Table 3). Of these, 17 had the most balanced XIAP/cIAP1

proﬁle, comparable to 5, and was progressed to in vivo PK

evaluation. Oral bioavailability in mouse was 3%, comparable

with 14, so the compound was not progressed further.

Substitution with the more polar cyano group at C5 and C6,

18 and 19 respectively, gave satisfactory CYP3A4 and hERG

proﬁles and was reasonably well tolerated with respect to XIAP

aﬃnity. However, these changes resulted in lower activity in

both cIAP1 degradation assay and MDA-MB-231 antiprolifer-

ation assays compared with the corresponding ﬂuoro

derivatives.

aﬃnity in the cellular IP assay (IC₅₀5.4 nM). However, this

change also resulted in exceptionally high cIAP1 potency in the

degradation assay (IC₅₀130 pM) and, consequently, an

unbalanced XIAP/cIAP1 proﬁle. There was, encouragingly, an

improvement in metabolic stability in the CYP3A4 turnover

assay and reduced hERG inhibition. N-Methylation of the

pyridone 24 surprisingly gave a much more balanced proﬁle,

largely due to a signiﬁcant increase in XIAP aﬃnity, without

increasing hERG inhibition, though stability in the CYP3A4

assay was suboptimal. These observations for CYP3A4 are

consistent with an increase in lipophilicity for 24, the ChrlogD

value of 3.2 being slightly above the limit suggested by the

model described above. The absence of a similar deleterious

eﬀect on hERG may be due to eﬀective disruption of the

hERG pharmacophore by the presence of the pyridone

carbonyl group. The isomeric pyridone 25 was also

investigated, where the pyridone nitrogen constitutes the

growth point for the P4 side chain, but this change proved

ineﬀective for XIAP potency.

Having seen some encouraging eﬀects on metabolic

turnover and hERG inhibition from addition of an oxygen

substituent, we examined incorporation of hydroxymethyl

substitution at azaindoline C5. The prototype compound 26

resulted in a modest increase in potency compared to 8.

However, CYP3A4 stability and hERG inhibition were not

optimized, ascribed to the relatively high lipophilicity of the

molecule (ChrlogD 3.5), beyond the upper limit of our

targeted range. The exceptional potency of 26 allowed some

headroom for sacriﬁce of primary activity in the pursuit of

better physicochemical properties. Removal of a methyl group

from the morpholine gave 27,¹⁹with XIAP cellular IC₅₀2.8

nM (LE 0.30), cIAP1 degradation IC₅₀0.22 nM, and potent

antiproliferative eﬀect in the MDA-MB-231 assay (IC₅₀1.8

nM). The ratio of cIAP1 to XIAP potency for 27 (13-fold) was

similar to the lead 5 and therefore met our target for a

relatively balanced IAP antagonist proﬁle. Compound 27 has

also been shown to cause degradation of both cIAP1 and

cIAP2 in the diﬀuse large B-cell lymphoma cell line, WSU-

DLCL2, which has a high basal level of cIAP2.¹⁹In the hERG

PPC assay, 27 gave 30% inhibition at 30 μM and IC₅₀69 μM.

Compound 27 was subsequently proﬁled in a cardiovascular

study in conscious cynomolgus monkeys and showed no eﬀect

on QT interval (corrected for heart rate) at all doses up to a

maximum tolerated single oral dose of 100 mg/kg. In the

CYP3A4 turnover assay, 27 had improved metabolic stability

compared to 5, despite retaining the morpholine P2

substituent. Therefore, it was gratifying that, in addition to

having acceptable oral bioavailability in mouse (34% at 10 mg/

kg), 27 had signiﬁcantly improved oral exposure in

cynomolgus monkey compared to 5 (F 12% at 5 mg/kg rising

to 50% at 60 mg/kg). Table 5 shows the cross-species

pharmacokinetic data for 27. Compound 27 was subsequently

selected as a clinical drug candidate, having a balanced IAP

antagonist proﬁle, improved oral bioavailability in NHP with

respect to 5 and acceptable preclinical cardiac safety proﬁle.

Antagonism of cIAP1 and XIAP by 27 along with antitumor

activity in cell lines and xenograft models has been recently

described.¹⁹For example, signiﬁcant cIAP1 and XIAP

antagonism was observed for up to 72 h in the MDA-MB-

231 xenograft model following a single 20 mg/kg oral dose of

27 (Figure 5A). These pharmacodynamic eﬀects correlated

with compound levels detected in the tumor throughout the

duration of the experiment. Inhibition of tumor growth was

Table 4 illustrates changes made in the P2 morpholine

subseries. The optimization strategy adopted here involved

performing a polar substituent scan in positions that were (i)

capable of disrupting the hERG pharmacophore and (ii)

considered to be most accommodating with respect to the X-

ray crystal structure of 5 bound to XIAP. With a view to

minimizing added molecular weight, priority was given to

addition of a hydroxy substituent and introduction of a

carbonyl group in the pyridyl ring to form a pyridone.

Dimethylmorpholine 8, having the highest XIAP and cIAP

potency, was chosen as the baseline compound for this work,

and data are shown in Table 4. The main areas of opportunity

were on the periphery of the P4 pocket and the aromatic

portion of the azaindole bicycle. A diastereomeric pair of α-

hydroxylated P4 4-ﬂuorobenzyl derivatives was prepared 20

and 21, with the latter having the higher aﬃnity in the XIAP

cell-based IP assay. The additional hydroxyl group also had a

beneﬁcial eﬀect on CYP3A4 stability, and this translated into

improved bioavailability in mouse (31% at 10 mg/kg) and,

importantly, in cynomolgus monkey (13% at 5 mg/kg).

Compound 21 also emphatically reduced hERG activity, with

minimal inhibition in the PPC assay. However, the

antiproliferative eﬀect in MDA-MB-231 cells was weaker

than 5, reﬂecting a loss of potency in the cIAP1 degradation

assay. Figure 3B shows the X-ray crystal structure of 21 bound

to XIAP, from which the ligand electron density was consistent

with the assigned absolute stereochemistry of the ligand at the

benzylic carbon. The ligand adopts a binding mode similar to

that of 5, with the benzyl moiety occupying the P4 pocket and

the hydroxyl group well accommodated. However, when the

cIAP1 protein structure is overlaid with the 21−XIAP complex

(Figure 3C), a clash is apparent with the side chain of cIAP1

Arg308. This arginine residue is presumed to be quite rigid,

being involved in a salt bridge with the carboxy side chain of

Asp297. Binding of 21 to cIAP1 may therefore require

disruption of the salt bridge and thus incur an energetic

penalty. In XIAP by contrast, the residue corresponding to

Arg308 is much smaller, namely, Thr308. Hence the

diﬀerential eﬀects on XIAP and cIAP1 activity can be

rationalized in terms of a residue diﬀerence in close proximity

to the P4 pocket. Incorporation of hydroxyl in the aromatic

portion of the P4 substituent 22 is well-tolerated with respect

to cIAP1 but not XIAP, resulting in an unbalanced proﬁle.

Hence addition of polarity to the P4 substituent did not

optimally achieve the target proﬁle, and we turned our

attention to the azaindoline scaﬀold.

Eﬀorts were focused primarily on the 4-azaindoline core,

with C5 oﬀering the opportunity to add polar substitution.

Formal hydroxylation of the azaindoline scaﬀold at C5 to give

the pyridone 23 was well tolerated with respect to XIAP

G

DOI: 10.1021/acs.jmedchem.8b00900

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Table 5. Cross-Species Pharmacokinetic Parameters for

Compound 27

CL_p

V_ss

C_max(PO)

a

species

mouse

rat

cynomolgus

monkey

mL min⁻¹kg⁻¹

L kg⁻¹

ng mL⁻¹

F%

41

38

16

2.7

9.3

6.0

433

117

137

34

29

12

a

Plasma clearance (CL_p), volume of distribution at steady state (V_ss),

maximum concentration in plasma (C_max), and oral bioavailability (F)

were determined using an intravenous (IV) dose of either 1 mg/kg

(cynomolgus monkey) or 5 mg/kg (mouse, rat) and an oral (PO)

dose of either 5 mg/kg (cynomolgus monkey) or 10 mg/kg (mouse,

rat). For assay details, see Experimental Section.

Figure 6. X-ray crystal structure of 27 (cyan) bound to XIAP (PDB

5OQW) overlaid in same frame of reference as the XIAP complex

with fragment hit 6. Compound 6 is displayed as orange balls and

sticks. Hydrogen bonds are shown as dashed red lines and the

displayed protein Connolly surface is colored by electrostatic

potential (blue = positive, red = negative).

also demonstrated in the MDA-MB-231 model following oral

daily dosing of 27 as the hydrochloride and ^L-lactate salt

(Figure 5B).

The binding mode of 27 in XIAP (PDB code 5OQW) has

been described,¹⁹and it is instructive to compare the

interactions made by the clinical candidate with those of the

original fragment hit 6¹⁷(Figure 6). Notably all polar

interactions made by the fragment are conserved in the

binding mode of 27 and all additional protein−ligand contacts

are hydrophobic in nature. In particular the P2 2-

methylmorpholine only makes contact with the protein via

the methyl group. Protein−protein interaction surfaces tend to

be highly lipophilic and it can be challenging to design

antagonists with favorable physicochemical properties. The

morpholine moiety here makes an important contribution to

controlling lipophilicity, although the primary importance of

this group lies in (i) its ability to stabilize the active ligand

conformation and (ii) shield the key polar interaction of the

amide carbonyl with the backbone NH of Thr308, as has

previously been discussed.¹⁸A surprising aspect of the

optimization from 5 to 27 is that success was achieved in

terms of metabolic stability despite leaving the metabolically

most labile group intact; attempts to add blocking groups or

replace with alternative P2 substituents were less successful

with respect to overall proﬁle. The key enabler in the

morpholine subseries with respect to CYP3A4 stability and

hERG inhibition was research utilizing crystal structure

information to help identify positions where polar substitution

would be most optimally tolerated. The positioning of a

hydroxymethyl group on the 4-azaindoline core, proximal to

the hydrophobe feature in the Aaronov pharmacophore (green

sphere in Figure 2B), oﬀered the best balance between desired

XIAP and cIAP1 target proﬁles, as well as providing optimized

metabolic and cardiac proﬁles.

Several synthetic challenges needed to be addressed in order

to access the growing vectors for optimal pocket ﬁlling. First,

for the P1 and P2 region, a stereospeciﬁc synthesis of

substituted piperazines was required. Second, for the P3 and

P4 region, a reliable method for constructing a range of

Figure 5. Pharmacodynamic and tumor growth inhibitory eﬀects of 27 (ASTX660) in MDA-MB-231 xenograft mouse model. (A) Antagonism of

XIAP (gray bars) and cIAP1 (red bars) over a seven-day time course following a single 20 mg/kg oral dose of 27 (free base equivalent,

hydrochloride). Quantiﬁcation of XIAP:SMAC complex and cIAP1 levels in the tumor lysate was performed using a Mesoscale Discovery (MSD)

platform. The black triangles denote concentration of 27 in tumor during the time course of the same experiment. (B) Tumor growth inhibitory

activity of 27 in MDA-MB-231 xenograft model following daily (qd) oral (PO) dosing with 27 at 16 mg/kg (free base equivalent, ^L-lactate salt) and

20 mg/kg (free base equivalent, hydrochloride salt). Test article was dosed on days 1−21 inclusive.

H

DOI: 10.1021/acs.jmedchem.8b00900

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

a

Scheme 1

a

Reagents and conditions: (a) K₂CO₃, KI, MeCN, 90 °C, 31−64%; (b) HCl, dioxane, EtOAc, RT, 59−86%.

a

Scheme 2

a

Reagents and conditions: (a) ClCH₂COCl, MeCN, ∼ 5 °C, used without puriﬁcation; (b) K₂CO₃, KI, MeCN, 0 °C - RT, 55−98%; (c) HCl,

dioxane, EtOAc, RT, 58−100%.

a

Scheme 3

a

Reagents and conditions: (a) R¹H, K₂CO₃, KI, MeCN, 70 °C, 94% (for 32a, 32b); (b) R¹H/NaH, DMF, RT then 80 °C, 20−69% (for 32c−32h,

34); (c) H₂, Pd/C, EtOH, HOAc, RT, 72−100%; (d) Zn(CN)₂, Pd₂(dba)₃, Pd(PPh₃)₄, dppf, DMF, water, 100 °C, 75%; (e) TFA, RT then

(Boc)₂O, Et₃N, MeOH, 0 °C − RT, 41%.

heterobicyclic scaﬀolds had to be developed, together with

appropriate C−C bond formation and functionalization

methods. Some of this work has been described previously.^18,25

Compounds 7−27 in Tables 2−4 were synthesized as shown

in Schemes 1−6 using a strategy similar to that described for

5.¹⁸Full experimental details are provided in Supporting

Information. For 7−9, the substituted morpholine P2

substituent was introduced at a late stage in the synthesis by

reaction of previously reported chloro derivative 28¹⁸with the

appropriate amine 29 (Scheme 1). Subsequent acid-mediated

Boc deprotection gave desired targets 7−9 as hydrochloride

salts. Preparation of all other compounds followed the more

convergent route shown in Scheme 2 involving a two stage,

three-component coupling, whereby a heterobicyclic precursor

30 was chloroacetylated and the resulting chloroamide 31

aminated with a protected piperazine 32 in which the desired

P2 substituent had already been installed. Removal of the Boc

protection as described above aﬀorded the desired targets 10−

27 as hydrochloride salts.

Substituted piperazine coupling partners 32 were prepared

according to Scheme 3, by reaction of orthogonally protected

chloropiperazine 33¹⁸with the appropriate cyclic amine or

I

DOI: 10.1021/acs.jmedchem.8b00900

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

amide under basic conditions and subsequent hydrogenolysis

of the benzyl moiety. For the cyano substituted isoindolinones

32i and 32j, the synthesis began with 2,4-dimethoxybenzyl

(Dmob) protected piperazine 33a (prepared in an analogous

method to that for 33) and involved reaction with the sodium

salt of the appropriate bromoisoindolinone, followed by

palladium mediated cyanation of adducts 34, global depro-

tection, then reattachment of the Boc group to give 32i and

32j.

Heterobicyclic coupling precursors 30 were either prepared

in the manner previously disclosed¹⁸(e.g., those for

compounds 10−12) or using the methods shown in Schemes

4−6. For fused pyridazine containing compounds such as 13,

tribromide in dichloromethane to give hydroxy substituted

coupling partner 40.

Previously reported azaindolines 41 and 43¹⁸proved to be

useful starting points for the preparation of the corresponding

pyridone precursors 42 and 45 (Scheme 5). Treatment of 41

with 3-chloroperbenzoic acid resulted in N-oxidation, and this

product was rearranged to the pyridone using acetic anhydride

followed by base hydrolysis. Acid treatment then removed the

Boc group to give desired coupling precursor 42. For the N-

methylated pyridone 45, the more stable acetyl protecting

group was required. Acetylation of 43 proceeded in high yield,

and an oxidation−rearrangement sequence as described above

gave pyridone 44. Methylation was performed using iodo-

methane in the presence of base, albeit in modest yield, and the

acetyl protecting group was removed using acid treatment at

elevated temperature. Building block 43 also allowed access to

hydroxymethyl derivative 47 (Scheme 5), by virtue of the

relatively high reactivity of the azaindoline C5 position toward

electrophiles. Thus, treatment with N-bromosuccinimide gave

bromo compound 46 in high yield, and then two-stage

lithiation using methyllithium (which removed the proton

from N1) followed by tert-butyllithium and subsequent quench

at low temperature with N,N-dimethylformamide gave the C5

substituted aldehyde. This product, used without further

puriﬁcation, was reduced with sodium borohydride to give

coupling partner 47. This three-stage procedure proceeded in

good overall yield and has proved highly suitable for large scale

synthesis (>1 kg).

a

Scheme 4

In order to access the isomeric pyridone coupling partner 51

required for 25, a de novo ring synthesis was required (Scheme

6). Commercially available substituted pyridine 48 was Boc

protected then treated with an excess of sec-butyllithium to

form the C-lithio species. Quenching with carbon dioxide gas

then gave acid 49. Lactam formation under standard amide

bond forming conditions was followed by treatment with

iodomethane under basic conditions to give the 3,3-dimethyl

derivative 50. Quaternization of the pyridine nitrogen with 4-

ﬂuorobenzyl bromide followed by in situ demethylation

introduced the pyridone functionality, and the remaining

steps involved Boc deprotection and selective reduction of the

lactam carbonyl to furnish desired intermediate 51.

a

Reagents and conditions: (a) ArZnCl, LiBr, PEPPSI-IPr, NMP,

THF, 20 °C, used without puriﬁcation; (b) HCl_aq, MeOH, 20−50 °C,

90−100% over 2 steps; (c) BuLi, hexanes, THF, −78 °C, 0.5 h then

appropriate substituted benzaldehyde, −78 °C, 0.5 h, 82−93%; (d)

chiral HPLC separation; (e) TMSCl, NaI, MeCN, 60 °C, 6 h,44%; (f)

BBr₃, CH₂Cl₂, RT, 66%.

CONCLUSIONS

■

Boc protected chloro precursor 35²⁵(Scheme 4) was subjected

to Negishi coupling with the appropriately substituted

benzylzinc chloride, and the resulting product was treated

with acid to give the desired coupling partners 36 in high yield.

For compounds bearing a hydroxy group in the P4 substituent,

bromo-bicycle 37 proved to be a useful starting point (Scheme

4). Lithium halogen exchange followed by quench with 4-

ﬂuorobenzaldehyde gave an enantiomeric mixture of secondary

alcohols. Boc deprotection gave the racemic alcohol precursor

38, which was used in the initial small-scale preparation of 20

and 21, where the epimeric mixture of ﬁnal compounds was

separated by chiral HPLC. Alternatively, for preparation of

larger quantities of 21, chiral separation was performed on the

Boc protected 4-ﬂuorobenzaldehyde adduct, and this was

followed by Boc deprotection of the desired (S)-enantiomer to

aﬀord precursor 38a. Similarly, the anion derived from 37

could be quenched with 2-methoxy-4-ﬂuorobenzaldehyde, and

the resulting racemic mixture of alcohols was treated with

chlorotrimethylsilane and sodium iodide²⁶to give methoxy

derivative 39. Demethylation was then eﬀected using boron

Starting from a millimolar aﬃnity fragment 6, compound 27

has been identiﬁed having a balanced antagonist proﬁle for

cIAP1 and XIAP and suitable proﬁle for progression into the

clinic. Multiple structure-based techniques were used, such as

X-ray crystallography, computational modeling, and NMR-

guided conformational analysis. By employing a fragment-

based as opposed to a peptidomimetic strategy, and focusing

the FBDD optimization around the XIAP crystal structure, the

intrinsic high cIAP1 selectivity resulting from approaches

based on the AVPI tetrapeptide has been avoided. The value of

experimental research and development focusing on crystal

structures and biological data can be valuable in guiding aﬃnity

optimization in FBDD. However, as shown here, structural

information can also be helpful in directing multiple changes

which result in various compounds having desirable potency,

while improving hERG and metabolism proﬁles. After

extensive testing, the successful optimization ultimately

involved making important but relatively small structural

changes to 5, including substituent scanning with polar groups

which carried with them remarkable impact in terms of

J

DOI: 10.1021/acs.jmedchem.8b00900

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

a

Scheme 5

a

Reagents and conditions: (a) 3-chloroperbenzoic acid, CH₂Cl₂, RT, 65%; (b) 1. Ac₂O, 105−140 °C; 2. NaOH_aq, MeOH, RT, 100%; (c) HCl_aq,

MeOH, reﬂux, 59%; (d) Ac₂O, MeCN, RT, 100%; (e) MeI, K₂CO₃, DMF, 0 °C − RT, 29%; (f) HCl_aq, EtOH, 95 °C, used without puriﬁcation;

(g) N-bromosuccinimide, DMF, 5 °C − RT, 95%; (h) 1. MeLi, Et₂O, THF, −78 °C; 2. tert-BuLi, hexane; 3. DMF, −78 °C, used without

puriﬁcation; (i) NaBH₄, MeOH, ∼ 5 °C, 76% over 2 steps.

a

Scheme 6

a

Reagents and conditions: (a) (Boc)₂O, Na₂CO₃, H₂O, THF, RT, 100%; (b) sec-butyllithium, cyclohexane, THF, −78 °C then CO₂(g), 100%; (c)

EDC, HOAT, iPr₂NEt, CH₂Cl₂, RT, 70%; (d) MeI, K₂CO₃, acetone, reﬂux, 68%; (e) 4-ﬂuorobenzyl bromide, NaI, MeCN, reﬂux, 11 h, 48%; (f)

TFA, CH₂Cl₂, RT, 90%; (g) BH₃.Me₂S, THF, Reﬂux, 99%.

MWD SLUV detector. LC retention times, molecular ion (m/z) and

LC purity (by UV) were based on the method below. Purity of

compounds (as measured by peak area ratio) was >95%, as

determined by the LC method described below. High resolution

mass spectrometry was performed on an Agilent 6550 Quadrupole

Time-of-Flight mass spectrometer.

LC Method (BASIC)

Eluent A: 95:5 10 mM NH₄HCO₃+ NH₄OH/CH₃CN (pH = 9.2)

Eluent B: CH₃CN

Gradient: 5−95% eluent B over 1.1 min

Flow: 0.9 mL/min

Column: Waters Acquity UPLC BEH C18; 1.7 μ; 2.1 × 50 mm

Column T: 50 °C

Preparation of Compound 27 (Schemes 2 and 5). 1-(6-[(4-

Fluorophenyl)methyl]-5-(hydroxymethyl)-3,3-dimethyl-1H,2H,3H-

pyrrolo[3,2-b]pyridin-1-yl)-2-[(2R,5R)-5-methyl-2-([(3R)-3-methyl-

morpholin-4-yl]methyl)piperazin-1-yl]ethan-1-one dihydrochloride.

Step 1: 5-Bromo-6-(4-ﬂuorobenzyl)-3,3-dimethyl-2,3-dihydro-

1H-pyrrolo[3,2-b]pyridine 46. A solution of 6-(4-ﬂuorobenzyl)-3,3-

dimethyl-2,3-dihydro-1H-pyrrolo[3,2-b]pyridine 43¹⁸(88.5 g, 0.345

mol) in DMF (1.67 L) was cooled to −5 °C. Solid N-

bromosuccinimide (61.5 g, 0.345 mol) was added in portions with

exotherm. The mixture was stirred for 1 h warming to room

temperature. Water (2.66 L) was added with exotherm and the

resulting mixture was stirred for 18 h at room temperature. The solids

were ﬁltered and cake washed with water (270 mL). The ﬁlter cake

was dissolved in THF (1.5 L), dried over MgSO₄, ﬁltered and

optimizing hERG and metabolism. For example, the structural

diﬀerence between 5 and 27 is quite small, namely, the

addition of a single two heavy atom substituent, but this

change had a profoundly beneﬁcial eﬀect on the metabolic and

hERG parameters, as well as resulting in a small but signiﬁcant

improvement in IAP antagonist activity. Throughout the

rigorous testing, the combination of structural knowledge and

control of physicochemical properties proved critical to the

identiﬁcation of the clinical candidate. Compound 27

(ASTX660) is now being evaluated in a phase 1/2 study in

subjects with advanced solid tumors and lymphomas

(NCT02503423). Additional clinical studies with 27 will

further explore the therapeutic potential of this novel IAP

antagonist, both as a single agent and in combination.

EXPERIMENTAL SECTION

■

General Chemistry. All solvents employed were commercially

available anhydrous grade, and reagents were used as received unless

otherwise noted. Hereafter, petrol denotes the petroleum ether

fraction boiling at 40−60 °C. Flash column chromatography was

performed on a Biotage SP1 system (32−63 μm particle size, KP-Sil,

60 Å pore size). NMR spectra were recorded on a Bruker AV400

(Avance 400 MHz) spectrometer. Analytical LC−MS was conducted

using an Agilent 1200 series with Mass Spec Detector coupled with an

Agilent 6140 single quadrupole mass detector and an Agilent 1200

K

DOI: 10.1021/acs.jmedchem.8b00900

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

concentrated in vacuo to give the title compound (109.7 g, 95%) as a

layer was extracted with further EtOAc (1 × 75 mL). The combined

EtOAc layers were washed with 10% aqueous KH₂PO₄(4 × 100 mL)

and then brine (70 mL). The organic layer was dried (MgSO₄) and

1

yellow solid. H NMR (DMSO-d₆): 7.29−7.20 (2H, m), 7.20−7.03

(2H, m), 6.64 (1H, s), 5.88 (1H, s), 3.89 (2H, s), 3.26 (2H, d), 1.20

1

(6H, s). LCMS: [M + H]⁺= 335, 337.

evaporated to give the product as a colorless solid (25.8 g, 98%). H

Step 2: [6-(4-Fluoro-benzyl)-3,3-dimethyl-2,3-dihydro-1H-

pyrrolo[3,2-b]pyridin-5-yl]-methanol 47. To 5-bromo-6-(4-ﬂuoro-

benzyl)-3,3-dimethyl-2,3-dihydro-1H-pyrrolo[3,2-b] pyridine 46

(22.8 g, 68.2 mmol) in THF (300 mL), cooled to −78 °C, was

added methyllithium (1.6 M in Et₂O; 51.1 mL, 91.8 mmol) over 15

min tert-butyllithium (1.7 M in hexane; 96 mL,164 mmol) was then

added over 30 min. After 15 min, DMF (26 mL) was added, and the

mixture was stirred at −78 °C for a further 50 min. Saturated aqueous

NH₄Cl (450 mL) was added, and the mixture was stirred for 10 min

at RT. The organic layer was isolated, and the aqueous layer was

extracted with EtOAc (2 × 150 mL). The combined organic fractions

were washed with brine (200 mL), dried (MgSO₄) and evaporated to

give 6-(4-ﬂuoro-benzyl)-3,3-dimethyl-2,3-dihydro-1H-pyrrolo[3,2-b]-

pyridine-5-carbaldehyde as a yellow solid which was used without any

further puriﬁcation. LCMS: [M + H⁺]⁺= 285. This product (∼68

mmol) was suspended in MeOH (250 mL) and cooled in an ice bath.

NaBH₄(3.4 g, 81.8 mmol) was added portionwise over 5 min.

Cooling was removed and the mixture was stirred for a further 20

min. The mixture was cooled in an ice bath followed by careful

addition of 10% aqueous KHSO₄over 10 min (care: eﬀervescence).

After stirring for 5 min at RT, the mixture was recooled using an ice

bath. The mixture was basiﬁed by addition of 50% aqueous NaOH

(∼18 mL) and then concentrated in vacuo to approximately one-third

volume. The resulting aqueous mixture was extracted with CH₂Cl₂(1

× 200 mL, 2 × 100 mL), and the combined CH₂Cl₂layers were dried

(MgSO₄). The CH₂Cl₂solution was concentrated in vacuo to ∼30 mL

and then diluted with toluene (70 mL) to initiate crystallization of the

product. Collection by ﬁltration gave the product as a colorless

crystalline solid (10.6 g). A second crop (2.1 g) was collected from

the ﬁltrate. The ﬁltrate was concentrated, and the remaining material

was puriﬁed by SiO₂chromatography (eluting with 25−50% EtOAc/

hexanes) to give a third batch of material (2.1 g), giving the title

compound in an overall yield of 14.8 g (76% over 2 steps). ¹H NMR

(400 MHz, CDCl₃): 7.13−7.05 (2H, m), 7.05−6.89 (2H, m), 6.60

(1H, s), 4.79 (1H, t), 4.61 (2H, d), 3.78 (2H, s), 3.65 (1H, s), 3.41

(2H, d), 1.36 (6H, s). LCMS: [M + H]⁺= 287.

NMR (400 MHz, Me-d₃-OD): 8.17 (1H, s), 7.21 (2H, dd), 7.03 (2H,

t), 4.71 (2H, s), 4.23−4.15 (1H, m), 4.11−3.96 (4H, m), 3.80−3.51

(5H, m), 3.30−3.21 (2H, m), 3.01−2.63 (4H, m), 2.61−2.43 (2H,

m), 2.37 (1H, t), 2.31−2.17 (1H, m), 1.47 (9H, s), 1.43 (6H, s),

1.25−1.22 (3H, m), 1.01 (3H, d). LCMS: [M + H]⁺= 640.

Step 5: 1-(6-[(4-Fluorophenyl)methyl]-5-(hydroxymethyl)-3,3-

dimethyl-1H,2H,3H-pyrrolo[3,2-b]pyridin-1-yl)-2-[(2R,5R)-5-methyl-

2-([(3R)-3-methylmorpholin-4-yl]methyl)piperazin-1-yl]ethan-1-

one dihydrochloride 27. A mixture of tert-butyl (2R,5S)-4-(2-(6-[(4-

ﬂuorophenyl)methyl]-5-(hydroxymethyl)-3,3-dimethyl-1H,2H,3H-

pyrrolo[3,2-b]pyridin-1-yl)-2-oxoethyl)-2-methyl-5-([(3R)-3-methyl-

morpholin-4-yl]methyl)piperazine-1-carboxylate (0.47 g), ethyl ac-

etate (10 mL) and HCl−dioxane (4 M; 10 mL) was stirred at 20 °C

for 18 h, and the resulting solid was collected by ﬁltration to give the

title compound (0.43 g, 95%) as a colorless solid. [α]²⁰_D10° (c 1.00,

1

EtOH), H NMR (400 MHz, Me-d₃-OD): 8.56 (1H, s), 7.34 (2H,

dd), 7.15 (2H, dd), 5.00 (2H, s), 4.33−4.19 (4H, m), 4.18−3.92 (5H,

m), 3.83−3.38 (8H, m), 3.16 (4H, m), 1.65 (6H, s), 1.46 (3H, d),

1.38−1.25 (3H, m). ¹³C NMR (101 MHz, DMSO-d₆) δ 170.43,

160.73 (d, J = 241.7 Hz), 156.09, 152.12, 136.43 (d, J = 3.1 Hz),

135.06, 133.41, 130.59 (d, J = 8.0 Hz), 123.91, 115.09 (d, J = 21.1

Hz), 72.06, 66.25, 63.26, 59.92, 58.26, 56.12, 55.62 (broad), 55.34,

51.70, 51.60, 50.52, 48.97, 40.47, 35.37, 27.22, 26.90, 17.54, 13.43

(broad) [¹³C NMR data acquired on free base due to broad signals

seen for dihydrochloride]. HRMS (ESI-QTOF): m/z [M + H]⁺

Calcd for C₃₀H₄₂FN₅O₃540.3344; Found 540.3345. Δ = 0.19 ppm.

LCMS: [M + H]⁺= 540.

Derivation of Chromatographic LogD (ChrlogD ≡ Chrom-

LogD_7.4). Chromatographic hydrophobicity index (CHI) values were

measured using reversed phase HPLC column (PolymerX RP-1,

Phenomenex, UK) with a fast acetonitrile gradient at starting mobile

phase of pH = 7.4. CHI values are derived directly from the gradient

retention times by using a calibration line obtained for standard

compounds. The CHI value approximates to the volume % organic

concentration when the compound elutes. CHI is linearly transformed

into ChromLogD_7.4(abbreviated to ChrlogD in main text of article)

by least-squares ﬁtting of experimental CHI values to calculated C

logP values for over 20 000 research compounds using the following

formula: ChromLogD_7.4= 0.0857CHI-2.00. The average error of the

assay is 3 CHI unit or 0.25 ChromLogD_7.4.

Crystallography. XIAP-BIR3 250−354 was crystallized using a

1:1 ratio of 10 mg/mL protein and 0.1 M Hepes-NaOH pH 8.0, 3.0−

3.9 M NaCl. Crystals appeared over the course of a few days at 4 °C.

Crystals were soaked in fragments using 2.5 μL of compound in 100%

DMSO, 47.5 μL of 0.1 M Hepes-NaOH pH 8.0, 4 M NaCl to give a

ﬁnal concentration of fragment in the range of 50−100 mM. The pH

of the soaking solution was adjusted if necessary, and crystals were left

at 4 °C for 24−72 h. Crystals were cryoprotected using 0.05 M

Hepes-NaOH pH 8.0, 4 M NaCl, 15% ethylene glycol. The crystals

had cell dimensions of approximately 70 Å, 70 Å, 105 Å, and belong

to space group P4122. The diﬀraction observed ranged from 1.7 to

3.0 Å.

Cell Line Proliferation Assay. Inhibition of cell growth was

measured using the Alamar Blue assay.²⁷For each proliferation assay

cells were plated onto 96 well plates and allowed to recover for 16 h

prior to the addition of inhibitor compounds (in 0.1% DMSO v/v)

for a further 72 h. At the end of the incubation period 10% (v/v)

Alamar Blue (Bio-Rad AbD Serotec, Oxford, UK) was added and

incubated for a further 6 h prior to determination of ﬂuorescent

product at 535 nM excitation/590 nM emission.

Step 3: 2-Chloro-1-(6-[(4-ﬂuorophenyl)methyl]-5-(hydroxymeth-

yl)-3,3-dimethyl-1H,2H,3H-pyrrolo[3,2-b]pyridin-1-yl)ethan-1-one.

To a cooled (∼5 °C) suspension of [6-(4-ﬂuoro-benzyl)-3,3-

dimethyl-2,3-dihydro-1H-pyrrolo[3,2-b]pyridin-5-yl]-methanol 47

(11.8 g, 41.3 mmol) in MeCN (175 mL) was added chloroacetyl

chloride (6.9 mL, 86.7 mmol). Cooling was removed and the mixture

was stirred for 30 min at RT. The mixture was then evaporated in

vacuo and dissolved in MeOH (200 mL). K₂CO₃solution (12 g in

100 mL H₂O) was added, and the mixture was stirred at RT for 20

min after which the mixture was concentrated in vacuo to

approximately one-quarter volume. The aqueous mixture was

extracted with CH₂Cl₂(1 × 100 mL, 2 × 30 mL), and the combined

CH₂Cl₂layers were dried (MgSO₄). Evaporation in vacuo gave the

1

title compound as colorless crystalline solid (12.1 g, 99%). H NMR

(400 MHz, Me-d₃-OD): 8.18 (1H, s), 7.21 (2H, dd), 7.08−6.98 (2H,

m), 4.70 (2H, s), 4.39 (2H, s), 4.11 (2H, s), 4.03 (2H, s), 1.43 (6H,

s). LCMS: [M + H]⁺= 363.

Step 4: tert-Butyl (2R,5S)-4-(2-(6-[(4-ﬂuorophenyl)methyl]-5-

(hydroxymethyl)-3,3-dimethyl-1H,2H,3H-pyrrolo[3,2-b]pyridin-1-

yl)-2-oxoethyl)-2-methyl-5-([(3R)-3-methylmorpholin-4-yl]methyl)-

piperazine-1-carboxylate. (2R,5S)-2-Methyl-5-((R)-3-methyl-mor-

pholin-4-ylmethyl)-piperazine-1-carboxylic acid tert-butyl ester 32b¹⁸

(15.5 g, 46.4 mmol), KI (12.8 g, 77.4 mmol), and K₂CO₃(21.4 g, 155

mmol) were stirred in MeCN (70 mL) and cooled in an ice bath. 2-

Chloro-1-{6-[(4-ﬂuorophenyl)methyl]-5-(hydroxymethyl)-3,3-di-

methyl-1H,2H,3H-pyrrolo[3,2-b]pyridin-1-yl}ethan-1-one (14.0 g,

38.7 mmol) was then added as a solution in MeCN (100 mL). The

mixture was stirred at RT for 2 h and then concentrated in vacuo to

approximately one-quarter volume. The mixture was partitioned

between EtOAc (150 mL) and H₂O (150 mL), and then the aqueous

The antiproliferative activity of test compounds was determined by

measuring the ability of the compounds to inhibit growth in two

cancer cell lines: (1) MDA-MB-231 (human breast carcinoma)

[ECACC, Salisbury, UK], (2) HCT116 (human colon carcinoma)

[ECACC, Salisbury, UK] - insensitive cell line used as a control for

nonspeciﬁc cytotoxicity.

L

DOI: 10.1021/acs.jmedchem.8b00900

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

The MDA-MB-231 proliferation data reported in the paper are the

result of two duplicates. Compounds 2, 5, and 27 were tested multiple

times in independent experiments and statistical limits are calculated

for those compounds:

2: IC₅₀= 7.8 nM, SD = 3.4, n = 6.

5: IC₅₀= 4.4 nM, SD = 1.9, n = 21.

27: IC₅₀= 1.8 nM, SD = 1.0, n = 4.

For HCT116 negative control cell proliferation data on compounds

5 and 7−27; see Supporting Information Table S1.

Population Patch Clamp Assay Protocol. Inhibition of the hERG

channel was measured by automated patch clamp assay in CHO K1

cells, stably transfected with the hERG ion channel. PPC measure-

ments were performed using an IonWorks Quattro instrument

(Molecular Devices Corporation, Union City, CA) using a 384 well

PatchPlate (Molecular Devices Corporation) with 64 apertures per

well. Each concentration of test compound was tested in duplicate

wells. Amphotericin B was used to obtain electrical access to the cell

̀

interior at a ﬁnal concentration of 200 μg/mL. Human ether-a-gogo

XIAP Antagonism Immunoprecipitation Assay. An engi-

neered HEK293 cell line was generated by transfecting the

HEK293 cell line (ECACC, Salisbury, UK) with a full-length

FLAG-tagged XIAP expression construct [Origene Technologies

Inc., Rockville, USA] and a full-length untagged caspase-9 construct

[Origene Technologies, Inc., Rockville, USA]. Stable cotransfectants

were selected after culture in selection medium containing Geneticin

(Life Technologies, Paisley, UK).

related gene (hERG) currents were measured with a prepulse to +40

mV (2 s) from the holding potential of −80 mV, followed by a step to

−50 mV (2 s) to elicit the deactivating tail currents, before returning

to the holding potential for 1 s. Compounds were incubated for 600 s

between the pre- and postcompound reads. The external recording

solution used was 130 mM Na gluconate, 20 mM NaCl, 4 mM KCl, 1

mM MgCl₂, 1.8 mM CaCl₂, 10 mM Hepes, 5 mM glucose, pH to 7.3

with NaOH. All data were ﬁltered for seal quality, seal drop, and

current amplitude. The maximum current amplitude of the third pulse

tail current was calculated before (Pre) and after (Post) compound

addition and the amount of block assessed by dividing the

Postcompound current amplitude by the Precompound current

amplitude.

CYP Phenotyping. Test compounds were incubated in duplicate

at 1 μM in human supersomes expressing CYP3A4 at 37 °C for 40

min at protein concentration of 20 pmol/mL, in the presence of

NADPH. Positive control reference compounds (midazolam,

bufuralol, diclofenac, ethoxyresoruﬁn, efavirenz, omeprazole) were

included. Negative control incubations, in the absence of NADPH,

were performed. Incubations were sampled (50 μL) at 0, 5, 10, 20,

and 40 min, except for the negative controls which were sampled at 0

and 40 min.

Metabolite Identiﬁcation. Hepatocyte suspensions with pooled

cryopreserved hepatocytes at cell densities of 0.5 million cells/mL for

male SD rat, male cynomolgus NHP, and mixed gender human and

0.25 million cells/mL for male CD1 mouse, were performed on

compound 5 (10 μM) at 37 °C. Incubations were sampled (100 μL)

at 0, 10, 20, 45, and 90 min. Hepatocyte samples were full scanned

over a range from m/z 50−1000 using LC/MS in positive ionization

mode. Potential metabolite peaks were identiﬁed by comparison with

negative control samples and conﬁrmed by MS/MS fragmentation

pattern scans compared to that of parent compound.

Stable HEK293-XIAP-Caspase-9 cells were plated out into 96-well

plates and left overnight at 37 °C to recover. Compounds were added

to duplicate wells in 0.1% DMSO for 2 h at 37 °C. Cells were lysed in

50 μL lysis buﬀer (1% Triton X-100 in 20 mM Tris.Cl (pH 7.6), 150

mM NaCl, including protease inhibitors (Roche Diagnostics Ltd.,

Burgess Hill, UK) for 20 min rocking at room temperature.

Streptavidin-coated high-bind MSD plates (Meso Scale Discovery,

Gaithersburg, USA) were coated with biotinylated anti-FLAG M2

antibody (Sigma, Poole, UK) and then blocked with 3% BSA in TBST

(20 mM Tris.Cl (pH 7.6), 150 mM NaCl, 0.1% Tween-20). Cell

lysate was added to the 96-well anti-FLAG coated MSD plate and

placed on a shaker overnight at 4 °C. After washing 3 times with

TBST, rabbit anti-Caspase-9 (Cell Signaling Technology Inc.,

Danvers, USA) was added for 2 h at room temperature, with shaking.

After washing three times with TBST, antirabbit-sulfo tag (Meso Scale

Discovery, Gaithersburg, USA) was added for 2 h at RT. Plates were

washed three times with TBST and then read buﬀer was added,

before reading the plate on a MESO QuickPlex SQ 120 (Meso Scale

Discovery, Gaithersburg, USA). The data reported in the paper are

the result of two duplicates. Compounds 2, 5, and 27 were tested

multiple times in independent experiments, and statistical limits were

calculated for those compounds:

2: XIAP IP EC₅₀= 35 nM, SD = 12, n = 6.

5: XIAP IP EC₅₀= 5.1 nM, SD = 2.3, n = 43.

27: XIAP IP EC₅₀= 2.8 nM, SD = 1.5, n = 5.

In Vitro Permeability. Permeability of 5 was assessed using the

CacoReady system (ReadyCell, Barcelona, Spain). 5 and control

compounds (propranolol, antipyrine, vinblastine) were incubated at a

ﬁnal concentration of 10 μM in duplicate to either the apical (180 μL)

of the monolayer to measure apical to basolateral transport (A > B)

across the cell barrier or to the basolateral side (750 μL) to measure

the basolateral to apical transport (B > A). For A > B spiking, 5 and

control compounds were diluted from 10 mM DMSO stocks to 10

μM in HBSS buﬀer with 100 mM CaCl₂·2H₂O, 50 mM MgCl₂·5H₂O

and 0.5 mg/mL lucifer yellow made to volume with sterile water

(lucifer yellow was used to determine the integrity of the Caco-2

monolayer). For B > A compound spiking, 5 and control compounds

were prepared as for the A > B solutions without the addition of

lucifer yellow. 5 and control compounds were incubated for 1 h at 37

°C in a highly humidiﬁed atmosphere of 95% air and 5% CO_2.

Animal Husbandry and Handling. Rodent studies were

performed according to the Animal (Scientiﬁc Procedures) Act

(1986) law. Mouse studies were performed with mice allowed access

to food and water ad libitum. Rat studies were performed by

Pharmaron (Rushden, UK). NHP studies were performed at WuXi

AppTec (Suzhou, China) in accordance with the Wuxi IACUC

standard animal procedures along with the IACUC guidelines that are

in compliance with the Animal Welfare Act, the Guide for the Care

and Use of Laboratory Animals, the Oﬃce of Laboratory Animal

Welfare (OLAW) and following local ethical reviews. NHPs were

housed individually in stainless steel mesh cages in a controlled

environment (temperature 18 to 29 °C, relative humidity 30 to 70%)

and fed twice daily.

Cellular cIAP1 Antagonism Assay. MDA-MB-231 cells were

plated out into 96-well plates and left overnight at 37 °C to recover.

Compounds were added to duplicate wells in 0.1% DMSO for 2 h at

37 °C. Cells were lysed in 50 μL of lysis buﬀer (1% Triton X-100 in

20 mM Tris·Cl (pH 7.6), 150 mM NaCl, including protease inhibitors

(Roche Diagnostics Ltd., Burgess Hill, UK) for 20 min rocking at

room temperature. Streptavidin-coated high-bind MSD plates (Meso

Scale Discovery, Gaithersburg, USA) were coated with biotinylated

anti-cIAP1 antibody (R&D Systems, Abingdon, UK) and then

blocked with 3% BSA in TBST (20 mM Tris.Cl (pH 7.6), 150 mM

NaCl, 0.1% Tween-20). Cell lysate was added to the 96-well anti-

cIAP1 coated MSD plate and placed on a shaker for 2 h at room

temperature. After washing three times with TBST, anti-cIAP1

antibody (R&D Systems, Abingdon, UK) which had been tagged with

sulfo-tag (Meso Scale Discovery, Gaithersburg, USA) was added for 2

h at room temperature, with shaking. Plates were washed three times

with TBST, and then read buﬀer was added, before reading the plate

on a MESO QuickPlex SQ 120 (Meso Scale Discovery, Gaithersburg,

USA). The data reported in the paper are the result of two duplicates.

Compounds 2, 5, and 27 were tested multiple times in independent

experiments and statistical limits are calculated for those compounds:

2: cIAP1 degradation EC₅₀= 0.40 nM, SD = 0.13, n = 7.

5: cIAP1 degradation EC₅₀= 0.32 nM, SD = 0.45, n = 30.

27: cIAP1 degradation EC₅₀= 0.22 nM, SD = 0.21, n = 7.

hERG Inhibition. IC₅₀and % inhibition data were determined in

the population patch clamp (PPC) assay conducted by Millipore

(now Euroﬁns Panlabs, St. Charles, MO, USA).

M

DOI: 10.1021/acs.jmedchem.8b00900

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

In Vivo Pharmacokinetic Studies. For intravenous admin-

istrations, test compounds were formulated in 100% saline for all

species. For oral administrations test compounds were formulated

either in 100% water for mice, 0.5% methocel (aqueous) for rats, or

100% saline for NHPs. In rodent species, intravenous dosing was

administered via the lateral tail vein at dose volumes of 1−5 mL/kg.

In NHPs, test compounds were administered intravenously as a slow

injection via the cephalic vein at a dose volume of 1 mL/kg. For all

species, test compounds were administered orally by nasogastric

gavage at dose volumes of 2−10 mL/kg. All doses were calculated as

freebase equivalent per kg of bodyweight. Pharmacokinetic studies

were performed in male Balb/c mice, obtained from Harlan

Laboratories Inc. (Shardlow, UK), male SD rats (Pharmaron,

Rushden, UK) and male cynomolgus NHPs (WuXi AppTec, Suzhou,

China). Following dosing in mice, blood samples (0.2 mL) were

drawn in tubes containing potassium EDTA, via either saphenous vein

bleeding or cardiac puncture at various time points over 30 h using

sparse sampling (n = 3 per time point), prior to centrifugation (2000

g at 4 °C, 10 min). The resultant plasma was separated from the

erythrocyte pellets for analysis and stored at −20 °C. In rat, blood

samples (0.3 mL) were serially sampled in parallel (n = 3 per route)

over 24 h via the lateral tail vein and collected into EDTA containing

tubes. Terminal samples were collected via cardiac puncture and

centrifuged (10 000 g at 4 °C, 2 min) and the resultant plasma stored

at −20 °C. Following dosing in NHPs (crossover design, n = 3),

blood samples (5 mL) were drawn in tubes containing potassium

EDTA at scheduled time points over 24 h and centrifuged (3000 g at

2−8 °C, 10 min). The resultant plasma was stored at −70 °C.

Noncompartmental pharmacokinetic (PK) analyses were performed

using Phoenix 6.3.0.395 (Certara USA, Inc.) software. Calculated

parameters included clearance, volume of distribution (V_ss), time of

maximum observed concentration (T_max), maximum concentration

(C_max), terminal half-life, area under the curve (AUC) from the time

of dosing to the last measurable concentration (AUC_last) and

extrapolated to inﬁnity (AUC_0‑∞). Noncompartmental PK ﬁtting to

sparse sampling data allowed the calculation of standard errors on

AUC_lastand C_maxfor mouse PK studies. For all other studies where a

full-time proﬁle was available noncompartmental PK ﬁtting was

applied to each individual animal (rat and NHP)

Bioanalysis. All in vitro and in vivo samples were extracted by

protein precipitation with acetonitrile containing internal standard

(1:3 v/v). For quantitative studies, calibration standards and quality

controls were prepared in blank matrix and extracted under the same

conditions. All samples were centrifuged at 3700 rpm at 4 °C for 20

min. Test compound bioanalysis of all in vitro and in vivo samples was

performed using high performance liquid chromatography mass

spectrometry with either a Quattro Ultima (Waters, UK) tandem

mass spectrometer coupled to a CTC HTS PAL autosampler (CTC

Analytics AG, Switzerland) and Acquity UHPLC pump (Waters, UK)

or a Qtof 6150 time-of-ﬂight mass spectrometer coupled to a 1200

UHPLC pump (Agilent, UK). Test compound and the internal

standard were ionized using positive mode (ESI+) electrospray

ionization. Analytes were detected using full scan or multiple reaction

monitoring (MRM). Compounds typically ran on a gradient HPLC

method over 5 min with a 0.1% formic acid and acetonitrile mobile

phase at a ﬂow rate of 0.5 mL/min. Separation was typically achieved

using an Acquity HSS T3 1.8 μM 50 × 2.1 mm column maintained at

40 °C.

dissolved in water then adjusted to pH 5.5 with sodium hydroxide).

Control animals received water. During the treatment period, tumors

were measured at least twice a week and the eﬀect on body weight

recorded daily where possible. Statistical analyses were performed

using GraphPad Prism version 6. The eﬀects of treatments were

compared using one-way ANOVA and two-way ANOVA with

Dunnett’s multiple comparisons test against vehicle control. Diﬀer-

ences were deemed statistically signiﬁcant when P < 0.05.

Analysis of Tumor Sample Pharmacodynamic Markers. Xeno-

graft tumor lysates were prepared by grinding the frozen tissue to a

ﬁne powder with a mortar/pestle under liquid nitrogen, and then

adding ice-cold lysis buﬀer (1% Triton X-100, 150 mM NaCl, 20 mM

Tris·HCl pH 7.5, plus protease inhibitors (Roche, Burgess Hill, UK),

50 mM NaF and 1 mM Na₃V0₄), to the ground-up tumor powder.

Samples were vortexed and left on ice for 30 min. Lysates were

cleared by centrifugation and samples of the supernatant removed for

protein determination by BCA assay (Pierce). Meso Scale Discovery

(MSD) plate-based assays were used to quantify levels of cIAP1

(cIAP1 antagonism) or levels of SMAC binding to XIAP (XIAP

antagonism) in MDA-MB-231 tumor lysates as described previ-

ously.^18,19

Pharmacokinetic Analysis. Compound levels in plasma and tumor

samples were measured and pharmacokinetic parameters calculated as

described previously with the exception of sample bioanalysis which

was undertaken using reverse-phase liquid chromatography−mass

spectrometry (MS), using a Qtrap 4000 MS (AB Sciex, UK), coupled

to an Agilent 1200 HPLC system (Agilent, UK) or a Quattro Premier

MS coupled to an Acquity UPLC system (Waters, UK).¹⁸

ASSOCIATED CONTENT

■

S

* Supporting Information

The Supporting Information is available free of charge on the

ACS Publications website at DOI: 10.1021/acs.jmed-

chem.8b00900.

NMR solution conformation studies on 12; protein

production for bioassay and crystallography; HCT116

negative control cell line proliferation data; cross-species

in vitro intrinsic clearance values for compound 5;

synthesis of compounds 7−26, NMR and LCMS plots

for compound 27 (PDF)

Molecular formula strings (CSV)

Accession Codes

Coordinates for XIAP-BIR3 complexes with compounds 12

and 21 have been deposited in the Protein Data Bank (PDB)

under accession codes 6h6r and 6h6q, respectively.

AUTHOR INFORMATION

■

Corresponding Author

*Phone: +44 1223 226247. Fax: +44 1223 226201. E-mail:

chris.johnson@astx.com.

ORCID

Xenograft Studies. MDA-MB-231 xenografts were prepared by

subcutaneously injecting 5 × 10⁶cells, suspended in 100 μL of serum-

free medium, into the right hind ﬂank of male severe combined

immunodeﬁcient (SCID, BALB/cJHanHsd-Prkdcscid) mice. Tumors

were measured using digital calipers and volumes calculated by

applying the formula for ellipsoid. For tumor growth inhibition

studies, tumor-bearing animals were randomized into groups of 7 to 8

with the average tumor volume of 100 mm³(approximately 30 days

after MDA-MB-231 cell injection). Mice were randomized and

compound 27 oral treatment started on Day 1. Drug-treated animals

received 27 at 16 mg/kg (free base equivalent, L-lactate salt dissolved

in water) and 20 mg/kg (free base equivalent, hydrochloride salt

Christopher N. Johnson: 0000-0001-8618-9729

Notes

The authors declare no competing ﬁnancial interest.

ACKNOWLEDGMENTS

■

We acknowledge valuable scientiﬁc discussions with colleagues,

including Tom Heightman, Chris Murray, and David Rees.

The authors also thank Stuart Whibley for provision of high

resolution mass spectrometry data, and Anne Cleasby for

crystallographic support.

N

DOI: 10.1021/acs.jmedchem.8b00900

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

tumors and hematological malignancies. J. Med. Chem. 2014, 57 (9),

3666−3677.

ABBREVIATIONS USED

■

BIR, baculovirus IAP repeat; ChrlogD, chromatographic

LogD; cIAP, cellular inhibitor of apoptosis protein; CYP,

cytochrome P450; DMF, N,N-dimethylformamide; Dmob, 2,4-

dimethoxybenzyl; EDC, (3-[[(ethylimino)methylidene]-

amino]propyl)dimethylamine hydrochloride; FADD, Fas-

associated protein with death domain; hERG, human ether-

(15) Flygare, J. A.; Beresini, M.; Budha, N.; Chan, H.; Chan, I. T.;

Cheeti, S.; Cohen, F.; Deshayes, K.; Doerner, K.; Eckhardt, S. G.;

Elliott, L. O.; Feng, B.; Franklin, M. C.; Reisner, S. F.; Gazzard, L.;

Halladay, J.; Hymowitz, S. G.; La, H.; LoRusso, P.; Maurer, B.;

Murray, L.; Plise, E.; Quan, C.; Stephan, J. P.; Young, S. G.; Tom, J.;

Tsui, V.; Um, J.; Varfolomeev, E.; Vucic, D.; Wagner, A. J.;

Wallweber, H. J.; Wang, L.; Ware, J.; Wen, Z.; Wong, H.; Wong, J.

M.; Wong, M.; Wong, S.; Yu, R.; Zobel, K.; Fairbrother, W. J.

Discovery of a potent small-molecule antagonist of inhibitor of

apoptosis (IAP) proteins and clinical candidate for the treatment of

cancer (GDC-0152). J. Med. Chem. 2012, 55 (9), 4101−4113.

(16) Hennessy, E. J.; Adam, A.; Aquila, B. M.; Castriotta, L. M.;

Cook, D.; Hattersley, M.; Hird, A. W.; Huntington, C.; Kamhi, V. M.;

Laing, N. M.; Li, D.; MacIntyre, T.; Omer, C. A.; Oza, V.; Patterson,

T.; Repik, G.; Rooney, M. T.; Saeh, J. C.; Sha, L.; Vasbinder, M. M.;

Wang, H.; Whitston, D. Discovery of a novel class of dimeric smac

mimetics as potent IAP antagonists resulting in a clinical candidate for

the treatment of cancer (AZD5582). J. Med. Chem. 2013, 56 (24),

9897−9919.

̀

a-go-go related gene (I_Krpotassium channel); HOAT, 1-

hydroxy-7-azabenzotriazole; IAP, inhibitor of apoptosis

protein; IP, immunoprecipitation; MPC, manual patch

clamp; NHP, nonhuman primate; NT, not tested; PPC,

population patch clamp; RING, really interesting new gene;

RIPK1, receptor-interacting serine/threonine-protein kinase 1;

RT, room temperature; SMAC, second mitochondria derived

activator of caspases; THF, tetrahydrofuran; TFA, triﬂuoro-

acetic acid; TNF, tumor necrosis factor; XIAP, X-linked

inhibitor of apoptosis protein.

REFERENCES

■

(1) Hanahan, D.; Weinberg, R. A. Hallmarks of cancer: The next

generation. Cell 2011, 144 (5), 646−674.

(2) Fulda, S.; Vucic, D. Targeting IAP proteins for therapeutic

intervention in cancer. Nat. Rev. Drug Discovery 2012, 11 (2), 109−

124.

(3) Dubrez, L.; Berthelet, J.; Glorian, V. IAP proteins as targets for

drug development in oncology. OncoTargets Ther. 2013, 9, 1285−

1304.

(4) Bai, L.; Smith, D. C.; Wang, S. Small-molecule SMAC mimetics

as new cancer therapeutics. Pharmacol. Ther. 2014, 144 (1), 82−95.

(5) Hunter, A. M.; LaCasse, E. C.; Korneluk, R. G. The inhibitors of

apoptosis (IAPs) as cancer targets. Apoptosis 2007, 12 (9), 1543−

1568.

(17) Chessari, G.; Buck, I. M.; Day, J. E.; Day, P. J.; Iqbal, A.;

Johnson, C. N.; Lewis, E. J.; Martins, V.; Miller, D.; Reader, M.; Rees,

D. C.; Rich, S. J.; Tamanini, E.; Vitorino, M.; Ward, G. A.; Williams,

P. A.; Williams, G.; Wilsher, N. E.; Woolford, A. J. Fragment-based

drug discovery targeting inhibitor of apoptosis proteins: Discovery of

a non-alanine lead series with dual activity against cIAP1 and XIAP. J.

Med. Chem. 2015, 58 (16), 6574−6588.

(18) Tamanini, E.; Buck, I. M.; Chessari, G.; Chiarparin, E.; Day, J.

E. H.; Frederickson, M.; Griffiths-Jones, C. M.; Hearn, K.;

Heightman, T. D.; Iqbal, A.; Johnson, C. N.; Lewis, E. J.; Martins,

V.; Peakman, T.; Reader, M.; Rich, S. J.; Ward, G. A.; Williams, P. A.;

Wilsher, N. E. Discovery of a potent nonpeptidomimetic, small-

molecule antagonist of cellular inhibitor of apoptosis protein 1

(cIAP1) and x-linked inhibitor of apoptosis protein (XIAP). J. Med.

Chem. 2017, 60 (11), 4611−4625.

(19) Ward, G. A.; Lewis, E. J.; Ahn, J. S.; Johnson, C. N.; Lyons, J.

F.; Martins, V.; Munck, J. M.; Rich, S. J.; Smyth, T.; Thompson, N.

T.; Williams, P. A.; Wilsher, N. E.; Wallis, N. G.; Chessari, G.

ASTX660, a novel non-peptidomimetic antagonist of cIAP1/2 and

XIAP, potently induces TNFalpha-dependent apoptosis in cancer cell

lines and inhibits tumor growth. Mol. Cancer Ther. 2018, 17 (7),

1381−1391.

(20) Erickson, R. I.; Tarrant, J.; Cain, G.; Lewin-Koh, S. C.; Dybdal,

N.; Wong, H.; Blackwood, E.; West, K.; Steigerwalt, R.; Mamounas,

M.; Flygare, J. A.; Amemiya, K.; Dambach, D.; Fairbrother, W. J.;

Diaz, D. Toxicity profile of small-molecule IAP antagonist GDC-0152

is linked to TNF-alpha pharmacology. Toxicol. Sci. 2013, 131 (1),

247−258.

(21) Komura, H.; Iwaki, M. In vitro and in vivo small intestinal

metabolism of CYP3A and UGT substrates in preclinical animals

species and humans: Species differences. Drug Metab. Rev. 2011, 43

(4), 476−498.

(22) Guo, L.; Guthrie, H. Automated electrophysiology in the

preclinical evaluation of drugs for potential QT prolongation. J.

Pharmacol. Toxicol. Methods 2005, 52 (1), 123−135.

(23) Aronov, A. M.; Goldman, B. B. A model for identifying hERG

K+ channel blockers. Bioorg. Med. Chem. 2004, 12 (9), 2307−2315.

(24) Waring, M. J. Lipophilicity in drug discovery. Expert Opin. Drug

Discovery 2010, 5 (3), 235−248.

(25) Day, J. E. H.; Frederickson, M.; Hogg, C.; Johnson, C. N.;

Meek, A.; Northern, J.; Reader, M.; Reid, G. The synthesis of 3,3-

dimethyl aza- and diazaindolines using a palladium-catalysed intra-

molecular reductive cyclisation. Synlett 2015, 26 (18), 2570−2574.

(26) Wiggins, J. M. A convenient procedure for the reduction of

diarylmethanols with dichlorodimethylsilane/sodium iodide. Synth.

Commun. 1988, 18 (7), 741−749.

(6) Deveraux, Q. L.; Reed, J. C. IAP family proteins−suppressors of

apoptosis. Genes Dev. 1999, 13 (3), 239−252.

(7) Vaux, D. L.; Silke, J. IAPs, rings and ubiquitylation. Nat. Rev.

Mol. Cell Biol. 2005, 6 (4), 287−297.

(8) Eckelman, B. P.; Salvesen, G. S.; Scott, F. L. Human inhibitor of

apoptosis proteins: Why XIAP is the black sheep of the family. EMBO

Rep. 2006, 7 (10), 988−994.

(9) Du, C.; Fang, M.; Li, Y.; Li, L.; Wang, X. SMAC, a mitochondrial

protein that promotes cytochrome c-dependent caspase activation by

eliminating IAP inhibition. Cell 2000, 102 (1), 33−42.

(10) Feltham, R.; Bettjeman, B.; Budhidarmo, R.; Mace, P. D.;

Shirley, S.; Condon, S. M.; Chunduru, S. K.; McKinlay, M. A.; Vaux,

D. L.; Silke, J.; Day, C. L. SMAC mimetics activate the E3 ligase

activity of cIAP1 protein by promoting ring domain dimerization. J.

Biol. Chem. 2011, 286 (19), 17015−17028.

(11) Petersen, S. L.; Wang, L.; Yalcin-Chin, A.; Li, L.; Peyton, M.;

Minna, J.; Harran, P.; Wang, X. Autocrine TNFalpha signaling renders

human cancer cells susceptible to SMAC-mimetic-induced apoptosis.

Cancer Cell 2007, 12 (5), 445−456.

(12) Obexer, P.; Ausserlechner, M. J. X-linked inhibitor of apoptosis

protein - a critical death resistance regulator and therapeutic target for

personalized cancer therapy. Front. Oncol. 2014, 4, 197.

(13) Cai, Q.; Sun, H.; Peng, Y.; Lu, J.; Nikolovska-Coleska, Z.;

McEachern, D.; Liu, L.; Qiu, S.; Yang, C. Y.; Miller, R.; Yi, H.; Zhang,

T.; Sun, D.; Kang, S.; Guo, M.; Leopold, L.; Yang, D.; Wang, S. A

potent and orally active antagonist (SM-406/AT-406) of multiple

inhibitor of apoptosis proteins (IAPs) in clinical development for

cancer treatment. J. Med. Chem. 2011, 54 (8), 2714−2726.

(14) Condon, S. M.; Mitsuuchi, Y.; Deng, Y.; LaPorte, M. G.;

Rippin, S. R.; Haimowitz, T.; Alexander, M. D.; Kumar, P. T.; Hendi,

M. S.; Lee, Y. H.; Benetatos, C. A.; Yu, G.; Kapoor, G. S.; Neiman, E.;

Seipel, M. E.; Burns, J. M.; Graham, M. A.; McKinlay, M. A.; Li, X.;

Wang, J.; Shi, Y.; Feltham, R.; Bettjeman, B.; Cumming, M. H.; Vince,

J. E.; Khan, N.; Silke, J.; Day, C. L.; Chunduru, S. K. Birinapant, a

SMAC-mimetic with improved tolerability for the treatment of solid

O

DOI: 10.1021/acs.jmedchem.8b00900

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

(27) Nociari, M. M.; Shalev, A.; Benias, P.; Russo, C. A novel one-

step, highly sensitive fluorometric assay to evaluate cell-mediated

cytotoxicity. J. Immunol. Methods 1998, 213 (2), 157−167.

P

DOI: 10.1021/acs.jmedchem.8b00900

J. Med. Chem. XXXX, XXX, XXX−XXX

Article Doi

A Fragment-Derived Clinical Candidate for Antagonism of X-Linked and Cellular Inhibitor of Apoptosis Proteins: 1-(6-[(4-Fluorophenyl)methyl]-5-(hydroxymethyl)-3,3-dimethyl-1 H,2 H,3 H-pyrrolo[3,2- b]pyridin-1-yl)-2-[(2 R,5 R)-5-methyl-2-([(3R)-3-methylmor

DOI: 10.1021/acs.jmedchem.8b00900

Source and publish data:

Authors:

Article abstract of DOI:10.1021/acs.jmedchem.8b00900

Full text of DOI:10.1021/acs.jmedchem.8b00900

Products guided by the article

R&D Labs maybe for 1605312-93-3

Relevant to this article

Hot Product