Discovery and Optimization of Pyrrolopyrimidine Inhibitors of Interleukin-1 Receptor Associated Kinase 4 (IRAK4) for the Treatment of Mutant MYD88L265P Diffuse Large B-Cell Lymphoma

Article abstract of DOI:10.1021/acs.jmedchem.7b01290

Herein we report the optimization of a series of pyrrolopyrimidine inhibitors of interleukin-1 receptor associated kinase 4 (IRAK4) using X-ray crystal structures and structure based design to identify and optimize our scaffold. Compound 28 demonstrated a favorable physicochemical and kinase selectivity profile and was identified as a promising in vivo tool with which to explore the role of IRAK4 inhibition in the treatment of mutant MYD88^L265P diffuse large B-cell lymphoma (DLBCL). Compound 28 was shown to be capable of demonstrating inhibition of NF-κB activation and growth of the ABC subtype of DLBCL cell lines in vitro at high concentrations but showed greater effects in combination with a BTK inhibitor at lower concentrations. In vivo, the combination of compound 28 and ibrutinib led to tumor regression in an ABC-DLBCL mouse model.

Full text of DOI:10.1021/acs.jmedchem.7b01290

Subscriber access provided by READING UNIV
Article
Discovery and Optimisation of Pyrrolopyrimidine Inhibitors
of Interleukin-1 Receptor Associated Kinase 4 (IRAK4) for the
L265P
Treatment of Mutant MYD88 Diffuse Large B-Cell Lymphoma.
James S. Scott, Sébastien L. Degorce, Rana Anjum, Janet Culshaw, Robert M. D. Davies,
Nichola L Davies, Keith S. Dillman, James E Dowling, Lisa Drew, Andrew D. Ferguson,
Sam D Groombridge, Christopher T. Halsall, Julian A. Hudson, Scott Lamont, Nicola A.
Lindsay, Stacey K. Marden, Michele F. Mayo, J. Elizabeth Pease, David R. Perkins, Jennifer
H. Pink, Graeme R. Robb, Alan Rosen, Minhui Shen, Claire McWhirter, and Dedong Wu
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b01290 • Publication Date (Web): 27 Nov 2017
Downloaded from http://pubs.acs.org on November 30, 2017
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 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery and Optimisation of Pyrrolopyrimidine
Inhibitors of Interleukinꢀ1 Receptor Associated
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Kinase 4 (IRAK4) for the Treatment of Mutant
L265P
MYD88
Diffuse Large BꢀCell Lymphoma.
James S. Scott,♣ Sébastien L. Degorce, Rana Anjum, Janet Culshaw, Robert D. M.
*
♣
♠
♦
♦
♣
♠
♠
♠
Davies, Nichola L. Davies, Keith S. Dillman, James E. Dowling, Lisa Drew, Andrew D.
Ferguson,♠ Sam D. Groombridge,♦ Christopher T. Halsall,♦ Julian A. Hudson,♦ Scott
♣
♣
♥
♠
♣
Lamont, Nicola A. Lindsay, Stacey K. Marden, Michele F. Mayo, J. Elizabeth Pease,
♦
♦
♣
♠
♠
David R. Perkins, Jennifer H. Pink, Graeme R. Robb, Alan Rosen, Minhui Shen, Claire
McWhirter, Dedong Wu.♥
♣
♣Oncology, IMED Biotech Unit, AstraZeneca, Cambridge, United Kingdom.
♠Oncology, IMED Biotech Unit, AstraZeneca, Boston, United States.
♦Oncology, IMED Biotech Unit, AstraZeneca, Macclesfield, United Kingdom.
♥Pharmaceutical Sciences, IMED Biotech Unit, AstraZeneca, Boston, United States.
ABSTRACT
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Herein we report the optimization of a series of pyrrolopyrimidine inhibitors of interleukinꢀ1
receptor associated kinase 4 (IRAK4) using Xꢀray crystal structures and structure based
design to identify and optimize our scaffold. Compound 28 demonstrated a favorable
physicochemical and kinase selectivity profile and was identified as a promising in vivo tool
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
L265P
with which to explore the role of IRAK4 inhibition in the treatment of mutant MYD88
diffuse large Bꢀcell lymphoma (DLBCL). Compound 28 was shown to be capable of
demonstrating inhibition of NFꢀκB activation and growth of the ABC subtype of DLBCL cell
lines in vitro at high concentrations but showed greater effects in combination with a BTK
inhibitor at lower concentrations. In vivo, combination of compound 28 and ibrutinib led to
tumor regression in an ABCꢀDLBCL mouse model.
INTRODUCTION
Diffuse large Bꢀcell lymphoma (DLBCL) is the most common Bꢀcell nonꢀHodgkin
1
lymphoma (NHL) comprising 30–35% of all NHLs. More than half of DLBCL patients can
be cured with immunochemotherapy, however approximately 30 to 40 % of patients will
develop relapsed/refractory disease that remains a major cause of morbidity and mortality due
2
to the limited therapeutic options. The molecularly distinct subtype ‘activated BꢀCell like’
(ABC) DLBCL has particularly poor clinical outcomes with this standard treatment regimen
when compared to the ‘germinal centre Bꢀlike’ (GCB) DLBCL which has favorable
3
prognosis. In ABCꢀDLBCL patients, recurrent gain of function mutations in the essential
adaptor protein MYD88 is seen in 39% of patients, with one dominant mutation of MYD88,
4
L265P, occurring in 29% of cases. Occurrence of L265P mutation in DLBCL patients has
5
been linked to poor prognosis on standard immunochemotherapy regimen.
ACS Paragon Plus Environment

Page 3 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
MYD88 is an adaptor protein of the Tollꢀlike receptors (TLRs) and interleukinꢀ1 receptors
(ILꢀ1Rs) that participates in the innate immune response and plays a crucial role in the
6
homeostasis of human Bꢀcells. After activation of TLRs, MYD88 is phosphorylated and
subsequently recruits IRAKs and other downstream proteins such as TRAF6 resulting in NFꢀ
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
,8
κB, JAK kinase / STAT3 activation and secretion of IL6, IL10 and interferonꢀβ. In ABCꢀ
DLBCL cells, an L265P mutation of MYD88 leads to a spontaneous assembly of a stable
signaling complex involving IRAK4, IRAK1 and TRAF6 that leads to an increase in kinase
activity of IRAK4 resulting in constitutive NFꢀκB activation, promoting cell proliferation and
4
survival. MYD88 has also been found to be mutated at different frequencies in several
lymphoid neoplasms including Waldenström’s macroglobulinemia (90%), primary cutaneous
leg type DLBCL (69%), central nervous system lymphomas (38ꢀ50%) and primary testicular
9
ꢀ13
lymphomas (68%).
IRAK4ꢀmediated signalling is also important for inflammation and
bridges innate and adaptive immunity. IRAK4 or MYD88 deficiency has been linked to
susceptibility to bacterial infections in children, particularly recurrent pyogenic bacterial
14,15
infections.
Given the critical role of IRAK4 in inflammatory processes and MYD88 driven cell
proliferation, the inhibition of IRAK4 activity presents an attractive therapeutic approach for
the treatment of inflammatory diseases and cancer. A number of groups have reported
16
chemical equity capable of inhibiting this target. This area has recently been reviewed with
17
18
notable contributions published in the literature from groups including Amgen, UCB,
19
20
21
22
Astellas, Merck, Pfizer and Nimbus. In the context of oncology, the contribution made
by Nimbus was particularly significant in that it showed IRAK4 inhibition, together with
inhibition of Bꢀcell receptor (BCR) signalling could have synergistic effects on ABC DLBCL
2
2
killing. This was also demonstrated in vivo, where a modest tumour growth inhibition effect
in OCIꢀLy10 xenografts with an IRAK4 inhibitor was significantly enhanced by coꢀtreatment
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
2
with ibrutinib, an inhibitor of Bruton's tyrosine kinase (BTK). As a starting point for our
program, we were intrigued by the tricyclic scaffold reported by Nimbus and profiled a
representative example (1) in terms of biological activity, selectivity and physicochemical
23
properties. This compound showed potency against IRAK4 (Table 1) and a good level of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
3
selectivity against a panel of kinases. A crystal structure (Figure 1) showed that this
compound binds at the ATP site, forming interactions with the kinase ‘hinge’ (Val263 and
Met265) and a salt bridge with Asp272. The hingeꢀbinding is unusual in that it involves an
interaction between the sulfur atom and the carbonyl of Val263 which can be understood as
the donation of electron density from the carbonyl to a chargeꢀdepleted region in the valence
24
shell of the aromatic sulfur.
Figure 1. Xꢀray crystal structure of tricylic inhibitor 1 in complex with IRAK4 (5K75).
Green dotted lines indicate intermolecular interactions involved in binding.
We therefore embarked on a medicinal chemistry program to investigate and understand the
structureꢀactivity relationship (SAR) surrounding this scaffold and to develop and optimize
novel chemical equity in order to validate IRAK4 as an oncology target.
RESULTS AND DISCUSSION
ACS Paragon Plus Environment

Page 5 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The synthesis of compounds is described below (Schemes 1ꢀ6). Bicyclic compounds 2–5 and
N
8–10 were prepared through an S Ar reaction with transꢀN,Nꢀdimethylcyclohexaneꢀ1,4ꢀ
diamine and the corresponding chloroheterocycle under microwave conditions (Scheme 1).
a
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
N
Scheme 1. Core variation with S Ar displacement
(a)
X=S
X=NH
1
2
1
2
2
R =Me; R =Me 8 R =Me; R =H
1
2
1
2
3 R =Me; R =H 9 R =H; R =Me
1
2
1
2
4 R =H; R =Me
10 R =H; R =H
1
2
5
R =H; R =H
a
o
Reagents and Conditions: (a) transꢀH
7ꢀ61%.
2 2
NCyNMe , DIPEA, iPrOH, MW, 85ꢀ150 C, 4ꢀ14 h,
1
In cases where the requisite chloroheterocycle was not available, these were constructed as
shown in Scheme 2. Tricyclic compound 6 was constructed through formation of cyclopentyl
hydrazone 6a followed by a Fischer indole synthesis to 6b which proceeded in good yield
N
(79% over 2 steps). Chlorination to 6c and an S Ar reaction with transꢀN,Nꢀ
dimethylcyclohexaneꢀ1,4ꢀdiamine gave 6. In the case of dimethyl substituted 7, the required
chloroheterocycle was prepared through assembly of benzyl protected pyrrole 7a, followed
by construction of the pyrimidine ring 7b using formic acid. Chlorination gave 7c which was
displaced with transꢀdimethylcyclohexaneꢀ1,4ꢀdiamine to give 7d. Removal of the benzyl
protecting group using aluminium trichloride gave 7.
a
Scheme 2. Core variation through heterocycle construction
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
a)
(b)
(d)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6a
X=OH 6b
X=Cl 6c
6
(c)
(
e)
(f)
(h)
2
H NBn
7a
X=OH 7b
Y=Bn 7d
Y=H 7
(
g)
(i)
X=Cl 7c
a
o
o
Reagents and Conditions: (a) cyclopentanone, EtOH, 80 C, 4 h, 89%; (b) Ph
2
O, 250 C, 6 h,
o
o
8
9%; (c) POCl3, 100 C, 45 min, 85%; (d) transꢀH
2
NCyNMe
2
, DIPEA, iPrOH, MW, 160 C,
9
h, 39%; (e) HCl, PhMe,
ꢀ
, 24 h, 36%; (f) HCOOH,
ꢀ
2
, 7 h, 90%; (g) POCl3, PhNMe , ꢀ,
o
3
.5 h, 100%; (h) transꢀH NCyNMe , DIPEA, iPrOH, MW, 150 C, 10 h, 52%; (i) AlCl ,
2
2
3
PhMe, ꢀ, 135 min, 32%.
Variation of the 5ꢀposition was achieved via the 5ꢀbromo substituted key intermediate 11a,
prepared by S Ar reaction on the 4ꢀchloro group of SEM protected 4ꢀchloroꢀ5ꢀbromoꢀ
N
pyrrolopyrimidine with transꢀdimethylcyclohexaneꢀ1,4ꢀdiamine (Scheme 3). The alkyl group
of compounds 11, 15–17 could then be installed by Negishi reaction with organozincates
followed by deprotection of the SEM group (steps b,c) or alternatively, in the case of
compounds 12, 18–21, through Suzuki reaction with alkenyl organoboronates followed by
2
5
hydrogenation and removal of the SEM protection (steps d,e,c). The 5ꢀtrifluoromethyl
2
6
group was introduced using a trifluoromethylsulfonium salt from the protected iodo
intermediate to give 14a, followed by S Ar reaction and removal of the SEM group to give
4. Attempts to introduce the analogous 5–tertꢀbutyl substituent directly from the 5ꢀbromo
N
1
precursor 11a were unsuccessful, therefore this analogue was synthesised by condensation of
ACS Paragon Plus Environment

Page 7 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
6ꢀaminopyrimidineꢀ2,4ꢀdiol with 2ꢀbromoꢀ3,3ꢀdimethylbutanal to construct the pyrrole ring
of 13a. Bisꢀchlorination with phosphorous oxychloride gave 13b, which underwent a
N
regioselective S Ar reaction in which the 4ꢀchloro group was preferentially displaced to give
13c. Hydrogenation to remove the 2ꢀchloro substituent gave 13.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Scheme 3. Variation of the 5-position of the pyrrolopyrimidine
(b,c)
(d,e,c)
i
1
1 R=Et
12 R=Pr
(b,c)
c
1
1
1
5 R= Pr
18 R=2-THF
19 R=3-THF
20 R=3-THP
(a)
or
c
6 R= Bu
(d,e,c)
c
7 R= Pn
2
1 R=4-THP
11a
(g,c)
(f)
14a
14
(h)
(j)
X=OH 13a
X=Cl 13b
Y=Cl 13c
Y=H 13
(i)
(k)
a
o
Reagents and Conditions: (a) transꢀH
b) R Zn or RZnBr, Pd(dppf)Cl .CH Cl
C, 16ꢀ72 h, 9ꢀ35% or TFA, CH
2
NCyNMe
2
, DIPEA, iPrOH, MW, 130 C, 5 h, 70%;
o
(
2
2
2
2
, THF, 80 C, 3ꢀ16 h, 55ꢀ100%; (c) TBAF, THF, 80
o
o
2
Cl2, 20 C, 3ꢀ24 h then NH
3
/MeOH, 1ꢀ24 h, 44ꢀ81%; (d)
o
RBpin, Pd(PPh ) or Pd(dtbpf)Cl , Na CO or K CO , DME/H O, 85ꢀ90 C, 5ꢀ16 h, 41ꢀ67%;
3
4
2
2
3
2
3
2
2 4 2
(e) Pd/C, H , EtOH/THF or MeOH, r.t., 3ꢀ16 h, 28ꢀ100% (alkenes) or Pd/C, NH HCO ,
o
MeOH,
ꢀ, 4ꢀ8h, 57ꢀ82% (furyls); (f) Cu, Ph
2
SCF
3
.CF
3
SO
2
, DMF, 80 C, 16 h, 44%; (g)
o
o
transꢀH
2
NCyNMe
2
, K
2
CO
3
, MeCN, 80 C, 4 h, 51%; (h) NaOAc, H
2
O, 20 C, 3 d, 72%; (i)
o
POCl
3
, DIPEA,
ꢀ
, 16 h, 100%; (j) transꢀH
2
NCyNMe
2
, K
2
CO
3
, MeCN, MW, 140 C, 16 h,
o
8
4%; (k) Pd/C, H , K CO , MeOH, 20 C, 16 h, 16%.
2
2
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Variation of the linker atom was carried out using a similar approach to that described above,
N
by S Ar displacement of the 4ꢀchloro group of SEM protected 4ꢀchloroꢀ5ꢀbromoꢀ
pyrrolopyrimidine (Scheme 4). In the case of the NMe linked compound 22, the conditions
were similar to those used for the NH linker. For the oxygen linked analogue 23, the transꢀ4ꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(dimethylamino)cyclohexanol was treated with sodium hydride before reaction with the
chloroheterocycle, using DMF as solvent. The 5ꢀTHP substituent was introduced in each case
by Suzuki coupling with cyclohexenyl boronate and hydrogenation, followed by removal of
the SEM protection.
a
2
Scheme 4. Variation of the 4-linkage of the pyrrolopyrimidine (NMe )
(a)
(b,c,d)
22 X=NMe
2
3 X=O
a
o
Reagents and Conditions: (a) transꢀHXCyNMe
2
, (DIPEA, iPrOH, 90 C, 12 h, 75% for 22 or
o
NaH, DMF, 20 C, 24 h, 63% for 23); (b) DihydropyranylBpin, Pd(PPh
3
)
4
or Pd(dtbpf)Cl
, MeOH or
2 2 3
Cl , r.t., 1ꢀ3 h then NH /MeOH, 1ꢀ24 h, 32ꢀ
2
,
o
K
2
CO
3
, 1,4ꢀdioxane or DME, H
2
O, 85ꢀ100 C, 1ꢀ6 h, 60ꢀ85%; (c) Pd/C, H
2
EtOH/THF, r.t., 3ꢀ16 h, 86ꢀ93%; (d) TFA, CH
7%.
6
For variation of the amine capping group, two complementary strategies were employed
Scheme 5). The first was to install the entire aminocyclohexanamine fragment with defined
stereochemistry through S Ar reaction followed by Suzuki coupling with cyclohexenyl
(
N
boronate, hydrogenation and removal of the SEM group (Route I; 25–29, 32). The alternative
approach was to use the S Ar to introduce an aminocyclohexanone fragment at C4 followed
N
by a Suzuki reaction to install a cyclohexenyl group at C5. Reductive amination could then be
ACS Paragon Plus Environment

Page 9 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
employed to install the desired amine capping group, although the cis/trans stereochemistry
was poorly controlled and so separation of the isomers was required. Hydrogenation and
SEM deprotection gave the desired products (Route II; 24, 30, 31).
a
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 5. Variation of the 4-substituent of the pyrrolopyrimidine
Route I
(b,c,e)
(+d for 29)
(a)
(f,b)
Route II
(g,c,e)
Route I
Route II
Y=
Y=
24
2
9
2
7
28
32
2
5
26
3
0
31
a
Reagents and Conditions: (a) H NCyY or H NCyX, where X=N[(CH ) ] NBz for 25/26 and
2
2
2 2 2
i
n
o
N[(CH
2
) ]
2 2
CHCO
2
Me for 29, DIPEA, PrOH or BuOH, MW, 80ꢀ150 C, 3ꢀ18 h, 53ꢀ100%;
(b) DihydropyranylBpin, Pd(PPh
3
)
4
or Pd(dtbpf)Cl
2
, K
3
PO
, MeOH or EtOAc or EtOH/THF, r.t.,
O, THF, r.t., 2 h, 86%; (ii) Me NH, HATU,
4 2 3
or K CO , THF or 1,4ꢀdioxane or
o
DME, H
2
O, 80ꢀ100 C, 3ꢀ16 h, 52ꢀ100%; (c) Pd/C, H
2
11ꢀ18 h, 29ꢀ100%; (d) for 29 (i) LiOH, H
2
2
DIPEA, DMF, r.t., 2 h, 24%; (e) TFA, CH Cl , r.t., 1ꢀ3 h then NH /MeOH, 1ꢀ24 h, or
2
2
3
Me
2
NCOCl for 25 / MeSO
2
Cl for 26, DIPEA, CH
2
2 2 2
Cl , r.t., 1 h then TFA, CH Cl , r.t., 1 h,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
i
o
NH
3
/MeOH, 1ꢀ24 h, 39ꢀ74%; (f) 4ꢀH
2
Ncyclohexanone, DIPEA, PrOH, MW, 120 C, 36 h,
o
34%; (g) HY, NaBH(OAc)
3
, AcOH, CH Cl or MeCN, 15ꢀ25 C, 16ꢀ18 h, 52ꢀ100%.
2 2
For the oxygen linked ether 33, an analogous approach to that described above was used
Scheme 6). For the thioether 34, the 4ꢀchloro group was first converted to the thiol 34a,
which then used in an S 2 reaction with the mesylate of cisꢀ4ꢀmorpholinocyclohexanol,
proceeding with inversion of stereochemistry to set up the desired transꢀstereochemistry of
4d. The 5ꢀTHP group was introduced under standard conditions in the presence of the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
N
3
sulfur, however the reduction of the alkene required Wilkinson’s catalyst at elevated
temperature. Removal of the SEM group under acidic conditions gave 34.
a
Scheme 6. Variation of the 4-linkage of the pyrrolopyrimidine (morpholine)
(a)
(b,c,d)
3
3
3 (X=O)
4 (X=S)
3
3a
(b,g,d)
(e)
(f)
34a
34d
a
o
Reagents and Conditions: (a) transꢀ4ꢀmorpholinocyclohexanol, NaH, DMF, 20 C, 24 h,
o
61%; (b) DihydropyranylBpin, Pd(dtbpf)Cl
2
, K
2
CO
3
, DME or 1,4ꢀdioxane, H
2
O, 80ꢀ85 C, 2ꢀ
6
h, 73ꢀ100%; (c) Pd/C, H , EtOH/THF, r.t., 3 h, 97%; (d) TFA, CH Cl , r.t., 1ꢀ3 h then
2
2
2
NH
3
/MeOH, 1ꢀ24 h, 25ꢀ56%; (e) Na
2
S, DMSO, r.t.,
2
h, 78%; (f) cisꢀ4ꢀ
2
Cl, H ,
o
morpholinocyclohexanol mesylate, K
2
CO
3
, MeCN, 80 C, 16 h, 92%; (g) Rh(PPh
3
)
3
o
THF, 50 C, 7 d, 30%.
ACS Paragon Plus Environment

Page 11 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
For variations of the 6ꢀmembered ring of the core (Scheme 7), the corresponding pyridine
was made by a Suzuki reaction on the unprotected azaindole core followed by SEM
protection of the NH to give 35b. A BuchwaldꢀHartwig amination of the 4ꢀchloro group with
rd
27
transꢀ4ꢀmorpholinocyclohexanamine using a 3 generation Brettphos precatalyst proceeded
in moderate yield (56%), then hydrogenation and removal of the SEM protection gave 35.
The 3ꢀcyano derivative 36 was synthesised by bromination of the available 3ꢀcyano
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
N
pyrrolopyridine, SEM protection and an S Ar reaction to install the C4 amine of 36c. A
subsequent Suzuki reaction, hydrogenation and deprotection gave 36.
a
Scheme 7. Variation of the 6-membered ring
(a)
(c,d,e)
35
X=H 35a
X=SEM 35b
(b)
Y=
(f,b,g)
(a,d,e)
36c
36
a
Reagents and Conditions: (a) DihydropyranylBpin, Pd(dtbpf)Cl
2
, K
2
3 2
CO , 1,4ꢀdioxane/H O,
o
8
0 C, 1ꢀ2 h, 76% for 35a, 100% for 36a; (b) SEMCl, NaH, THF, r.t., 4ꢀ16 h, 48% for 35b,
o
6
3% for 36b; (c) LHMDS, transꢀH
2
NCyMorph, BrettPhos Pd G3, THF, 80 C, 16 h, 56%;
(d) Pd/C, H (2 atm), MeOH, r.t., 2ꢀ3 h, 84% for 35, 100% for 36; (e) TFA, CH Cl , r.t., 3ꢀ4 h
2
2
2
then NH
3
2 2
/MeOH, 3ꢀ4 h, 61% for 35, 70% for 36; (f) NBS, CH Cl , r.t., 16 h, 100% (g) transꢀ
o
H
2
NCyMorph, DIPEA, iBuOH, MW, 180 C, 5 h, 81%.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Examination of the SAR, through deconstruction of the cyclopentyl ring of 1, revealed that
opening the ring to 5,6ꢀdimethyl substituted 2 reduced potency
∼10x (Table 1). Further
truncation of the 5ꢀsubstituent 3, 6ꢀsubstituent 4 or both groups 5 reduced potencies in line
with lipophilicity with ligand lipophilic efficiency (LLE) scores remaining around 5. We
hypothesised that replacement of the thiophene sulfur with a pyrrole NH could make a more
traditional, and potentially stronger, hydrogen bond with the carbonyl and introduce polarity
to the hinge binding motif. The fully analogous tricycle 6 had reduced potency, albeit with
lower lipophilicity and similar LLE. Opening of the ring to 5,6ꢀdimethyl 7 reduced potency
and LLE further and the 6ꢀmethyl analogue 8 had even lower activity. By contrast however,
the 5ꢀmethyl bicycle 9 had improved enzyme potency relative to the original tricycle start
point 1 and at significantly lower lipophilicity (logD7.4 0.1) with consequently higher cellular
LLE (6.6) suggesting the 5ꢀsubstituted pyrrolopyrimidine core as a potentially interesting
template for further optimization. Removal of all substitution 10 reduced potency but the
unsubstituted core remained at higher LLE than the original tricyclic core. The divergent
SAR between the S and N cores is shown graphically in Figure 2.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 1. Effect of substitution on thienopyrimidine and pyrrolopyrimidine cores.
1
2
c
Cpd
A
R
R
IRAK4
IRAK4
logD7.4
IRAK4
Cell
Enz IC50
Cell IC50
a
b
d
(
ꢁM)
0.14
1.5
(ꢁM)
LLE
1
2
S
S
ꢀ(CH
2
)
3
ꢀ
0.49
2.5
1.2
0.7
5.2
4.9
Me
Me
ACS Paragon Plus Environment

Page 13 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
3
4
5
6
7
8
9
S
Me
H
H
Me
H
1.6
2.4
3.6
4.3
0.6
0.3
0.1
0.5
0.6
ꢀ0.2
0.1
ꢀ0.6
4.8
5.1
5.0
5.3
4.0
ꢀ
S
f
S
H
4.3
8.4
f
NH
NH
NH
NH
NH
ꢀ(CH
2
)
3
ꢀ
0.41
9.0
1.7
f
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Me
Me
H
Me
H
23
f
>10
0.005
0.64
>30
e
Me
H
0.14
6.6
6.5
f
1
0
H
1.4
a
b
Enzyme inhibition based on n ≥ 3 with SEM within 0.3 units unless otherwise stated; Cell inhibition based on
c
d
n ≥ 2 with SEM within 0.3 units unless otherwise stated; logD7.4 determined by shake flask method; cell LLE
e
f
calculated from pIC ꢀlogD; SEM=0.28; n=1.
5
0
Figure 2. Plot of IRAK4 enzyme pIC50 (A); cell pIC50 (B); cell LLE (C) against scaffolds
with A = S or N showing divergent SAR between the cores. X=Out of range value
28
Exploration of the 5ꢀsubstitution with larger alkyl groups (11, 12) showed cell potency
increasing in line with lipophilicity, unfortunately metabolism in rat hepatocytes was also
seen to increase (Table 2). Tertiary butyl (13) and trifluoromethyl (14) substituents had
reduced potency and LLE values. Cyclic secondary alkyl substituents (15ꢀ17) showed a
similar trend of potency increasing in line with lipophilicity as the ring size increased from 3
to 5, but LLE remained constant and rat hepatocyte metabolism increased. In an attempt to
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
increase polarity, cyclic ethers were investigated. In the tetrahydrofurans, both 2ꢀsubstituted
stereoisomers (18a/b) showed similar potency but with the 3ꢀsubstituted isomers (19a/b) a
29
clear preference for one isomer was observed. Unfortunately, despite having low logD7.4
values, the compounds were rapidly turned over by rat hepatocytes. Expansion to the
tetrahydropyrans showed an increase in LLE for both 3ꢀ and 4ꢀsubstituted isomers (20 and
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
21) with the 4ꢀisomer (21) having an advantage of increased stability in rat hepatocytes. The
4ꢀTHP substituent was considered to offer an attractive balance of cellular potency / low
lipophilicity / low turnover in rat hepatocytes and a crystal structure of 21 in IRAK4 was
obtained (Figure 3). Comparison of the crystal structures of 21 and 1 showed a clear
difference in positioning of the two scaffolds with the pyrrolopyrimidine core being held
closer into the hinge by virtue of the tighter hydrogen bond between the pyrrolo NH and the
carbonyl of Val262 (OꢀꢀN distance is 2.9 Å, c.f. OꢀꢀS distance for the tricyclic compound at
3.7 Å). Closer inspection of the structures suggested that 6ꢀsubstituents on the
pyrrolopyrimidine ring may clash with the protein surface, potentially explaining the reduced
potency observed with 7 and 8, and that the enhancement in potency seen with 5ꢀsubsitution
may be rationalised by the placement of lipophilicity in a similar region occupied by the
cyclopentyl ring of the tricyclic core. The 4ꢀTHP of 21 shows a lipophilic stacking interaction
with Tyr262 as well as a hydrogenꢀbond to Lys213.
ACS Paragon Plus Environment

Page 15 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 3. Xꢀray crystal structure of pyrrolopyrimidine inhibitor 21 (5K72) in complex with
IRAK4 and overlaid with tricyclic inhibitor 1 (5K75).
Table 2. Exploration of 5-substituent.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Cpd
R
IRAK4
IRAK4
Cell IC50
(ꢁM)
0.14
logD7.4
IRAK4
Cell
Rat heps
c
Enz IC50
Clint
(
ꢁM)
LLE
d
9
Me
Et
0.005
0.1
0.6
1.0
1.7
1.5
0.9
1.5
1.7
0.4
0.3
0.1
0.1
0.3
0.2
6.6
6.6
6.5
4.4
5.1
6.2
5.9
5.9
5.5
5.4
6.2
7.5
6.9
7.1
9
29
40
25
66
15
52
71
59
36
120
73
30
14
11
12
0.013
0.005
0.14
0.063
0.029
0.81
iPr
tBu
13
14
15
CF
0.099
0.011
0.003
0.25
3
cPr
cBu
0.075
0.042
0.023
1.2
16
17
e
cPn
0.006
0.27
a
a
a
a
b
1
8a
2ꢀTHF
2ꢀTHF
3ꢀTHF
3ꢀTHF
3ꢀTHP
18b
0.28
2.1
1
9a
0.12
0.54
19b
20
0.008
0.016
0.006
0.034
0.066
0.059
2
1
4ꢀTHP
a
b
c
6
Separated enantiomers of unknown stereochemistry; Racemic mixture; Rate of metabolism (µl/min/10 cells)
6
d
determined from DMSO stock solution in isolated rat hepatocytes diluted to 1x10 cells/mL; SEM=0.28;
d
SEM=0.33.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The effect of changes to the atom linking the 4ꢀposition of the pyrrolopyrimidine to the
cyclohexyl motif was examined (Table 3). Methylation of the nitrogen (22) resulted in a
reduction in enzyme and cell potency relative to the unsubstituted amine (21) and an increase
in logD7.4 (+0.4). Switching to an ether linkage (23) also reduced potency, despite a sizeable
increase in logD7.4 (+1.0), and showed increased turnover in rat hepatocytes. On balance the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4ꢀNH linker was viewed as optimal and variations in the amine group were subsequently
investigated.
Table 3. Variation of 4-linker.
Cpd
X
IRAK4
IRAK4
Cell IC50
(ꢁM)
logD7.4
IRAK4
Cell
Rat heps
Clint
Enz IC50
(
ꢁM)
LLE
21
22
NH
NMe
O
0.006
0.19
0.059
0.40
0.2
0.6
1.2
7.1
5.8
5.4
14
8.2
100
23
0.044
0.28
ꢀ6
More detailed profiling of compound 21 highlighted poor permeability (Papp <1 x10 cm/s
as measured in a Cacoꢀ2 assay) and a high degree of efflux (ER=32) as potential concerns.
We attributed this to the strength of the basic nitrogen (cpKA(BH+) = 10.6) and overall polarity
(logD7.4 = 0.2). In order to address this, we designed analogues with increased lipophilicity
and attenuated basicity (Table 4). Capping the basic nitrogen with various groups such as
amide (24), urea (25) and sulphonamide (26) increased the logD7.4 and gave modest gains in
ACS Paragon Plus Environment

Page 17 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
permeability, although efflux remained an issue. In contrast, capping with a methyl
carbamate (27) gave significantly enhanced permeability and removed the efflux liability
observed with previous compounds, although the potency was slightly reduced compared to
21. Introduction of a morpholine was also observed to enhance permeability and reduce
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
efflux with 28 exhibiting high potency in both enzyme and cellular assays and good stability
in rat hepatocytes. Piperidines with amides in the 4ꢀposition (29) were also found to have
activity and attempts were made to conformationally rigidify and increase lipophilicity with
lactams such as 30 and 31. Intriguingly, both enantiomers of the lactam were very potent in
both enzyme and cell although low permeability and high efflux ratios precluded further
progression of these compounds. Our interpretation was that these groups project towards
bulk solvent and hence, in the absence of a preference from the protein, both enantiomers are
accommodated equally well.
Table 4. Variation of amine capping group.
21
Y=
24
25
26
27
2
8
29
3
0
31
Cpd
IRAK4
IRAK4
Cell IC50
(ꢁM)
logD7.4
IRAK4
Cell
Rat heps
Caco2
Enz IC50
Clint
P
app
a
(
ꢁM)
LLE
(ER)
2
1
4
0.006
0.059
0.2
1.3
1.5
7.1
6.0
6.0
14
4.6
13
<1 (32)
3 (38)
2 (16)
b
2
0.002
0.058
25
0.003
0.034
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
2
2
2
6
7
8
9
0.14
0.003
0.005
0.12
0.082
0.052
1.5
2
5.4
4.9
5.8
6.7
6.4
6.6
16
17
15
2
5 (9)
20 (2)
1.5
ꢀ0.1
1
14 (3)
c
f
0.009
0.27
<1 (8)
<1 (>50)
d
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
30
0.002
0.042
0.038
7.3
7.3
31
0.002
0.8
<1 (>50)
a
ꢀ6
Compounds were incubated at 10 ꢁM in cultured Cacoꢀ2 cells. Permeability (x10 cm/s) was measured in both
the apical to basolateral (A to B) and basolateral to apical (B to A) directions with efflux ratio (ER) calculated
b
c
d
f
from (BtoA)/(AtoB); SEM=0.40; SEM=0.37; SEM=0.31; n=1.
The crystal structure of 28 revealed a near identical binding mode to compound 21 and a
highly complementary shape to the pocket (Figure 4). The basic morpholine moiety is larger
than the dimethylamine of 21 and the ring is sandwiched between Asp272 (making a saltꢀ
bridge) and the backbone of Met192 (a strong polar interaction).
Figure 4. Xꢀray crystal structure of inhibitor 28 (5K76) in complex with IRAK4.
Morpholine 28 offered an attractive balance of potency and permeability and prompted us to
explore close analogues varying the linker and hinge binding group (Table 5). Synthesis of
the cisꢀstereoisomer (32) confirmed the stereochemical preference for the transꢀ
stereochemistry across the cyclohexyl ring. In line with previously observed SAR, the ether
ACS Paragon Plus Environment

Page 19 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
linkage (33) reduced potency, raised lipophilicity and increased instability in rat hepatocytes.
The thioether (34) offered similar potency and rat hepatocyte metabolism but at higher
lipophilicity (ꢀlogD +1.5) and with a consequent reduction in aqueous solubility (5 ꢁM
compared with 130 ꢁM for 28 at pH 7.4). Replacement of the pyrrolopyrimidine with a
pyrrolopyridine (35) and incorporation of a 3ꢀcyano group (36) did not offer any advantage.
On the overall balance of properties, compound 28 was selected for further profiling.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 5. Linker and hinge binder analogues of 28.
Cpd
X
Y
stereo
IRAK4
IRAK4
Cell IC50
(ꢁM)
logD7.4
IRAK4
Cell
Rat heps
Enz IC50
Clint
(
ꢁM)
LLE
2
8
NH
NH
O
N
N
trans
cis
0.005
0.57
0.052
1.5
1.7
2.4
3.0
1.7
2.3
5.8
3.8
4.5
4.5
4.8
4.2
15
58
110
15
7
a
32
3.1
33
34
35
36
N
trans
trans
trans
trans
0.033
0.004
0.022
0.056
0.16
S
N
0.035
a
NH
NH
CH
CCN
0.34
0.300
22
a
n=1.
In terms of solid state profile, compound 28 presents as a stable crystalline anhydrate with a
melting point >300 °C. The compound converts to a dihydrate during aqueous solubility
measurements but with drying and heating the crystalline anhydrate can be reꢀformed. The
compound is a weak base with a calculated pKa of 7.84 for the morpholineꢀN. Solubility of
the free form is pH dependent with increasing solubility at lower pH. Solubility in water at 25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
°C is 84 µM, and aqueous solubility in buffered pH 7.4 at 25 °C is 130 µM. Compound 28
shows good hydrolytic stability across the pH range 1ꢀ10.
In order to better understand the potential to further optimize this scaffold, we were keen to
understand the energy penalty required for the molecules to adopt the bioactive conformation.
The structure of the anhydrous form of 28 was determined by singleꢀcrystal Xꢀray diffraction.
In the molecule, all three saturated 6ꢀmembered rings adopt “chair” conformations, and the
two groups attached to the cyclohexane ring sit in the equatorial positions of the “chair”. A
centroꢀsymmetric dimeric structure forms via intermolecular hydrogen bonding between the
pyrrolopyrimidine units. Additional but weaker CHꢀX hydrogen bonds form between the
adjacent polarized CH bond and the THP group of neighbouring molecules (Figure 5).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. A) Single crystal Xꢀray structure of 28; B) Crystal packing in the unit cell showing
a centroꢀsymmetric dimeric structure linked by intermolecular hydrogen bonding between the
pyrrolopyrimidine units and flanked by two additional units bound through the THP moiety.
In addition, 2DꢀNMR was used to look at the conformation populations of 28 in solution.
Fitted results indicate that 90% of the population exists as a single conformer (including
indistinguishable, symmetrical conformations). Evolution of the proton coupling constants
indicate all saturated rings exist in chair conformations with equatorial substitution. The
overall conformation of 28 bound to IRAK4 is very similar to those observed in the ligand
singleꢀcrystal structure and to the NMRꢀderived solution conformation (Figure 6). There is
ACS Paragon Plus Environment

Page 21 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
essentially a perfect overlap of all atoms, with the exception of the terminal morpholine
group. Differences in the orientation of this group between the three structures implies
several nearꢀdegenerate rotational conformations are possible here. This implies that the
molecule is largely preꢀorganized for binding and is in an unstrained conformation upon
formation of the protein ligand complex.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 6. Overlay between small molecule Xꢀray structure (grey carbons), major solution
conformation as determined by NMR (orange carbons) and observed conformation of 28
bound to IRAK4 protein (purple carbons). Overlay based on pyrrolopyrimidine unit only.
Plasma protein binding showed high amounts of free drug across species (mouse 77% free;
rat 59% free; dog 86% free; human 76% free). No inhibition was observed (IC values > 30
5
0
µM) of five cytochrome P450 enzymes (CYP1A2, CYP2C9, CYP2C19, CYP2D6 and
CYP3A4) in a high throughput human liver microsome assay. The compound showed no
discernable activity (IC50 > 100 µM) against the hERG ion channel and was inactive (IC50
>
30 µM) against a wider selection (NaV1.5, Ito Kv4.3, Iks) of ion channels. Selectivity
against other kinase targets was investigated by profiling 28 in a panel (Millipore kinase
selectivity panel) of 275 kinases at a concentration of 0.1 ꢁM (Figure 7). The profile was
generally clean (IRAK4 + 4/275 >75%inh; 8/275 >50%inh) but showed inhibition of the
CLK family (x3) and haspin kinase. These hits were followed up with concentration response
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
data (Table 6) and confirmed inhibition of CLK1, 2 & 4 and haspin together with other
®
30
d
members of the IRAK family. We also generated K data (DiscoverX KdELECT ) for the
IRAK isoforms which showed >30x selectivity for IRAK4 over IRAK1 and >100x selectivity
over IRAK2 and 3.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 7. Millipore kinome selectivity for 28 (%inhibition @ 0.1 ꢁM)
Table 6. Inhibitory concentrations & selected Kd values for 28 in kinases identified in
the kinome selectivity profile as having %inhibition>75%
Kinase IC50 (ꢁM) Selectivity K (ꢁM) Selectivity
d
IRAK4
IRAK3
IRAK2
IRAK1
CLK1
CLK2
CLK3
CLK4
haspin
0.005
ꢀ
1.0x
ꢀ
0.0007
1.0x
0.093
>130x
ꢀ
ꢀ
0.170
>200x
0.023
0.050
0.005
>1
4.6x
10.0x
1.0x
>200x
1.6x
0.8x
0.026
37x
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
0.008
0.004
ACS Paragon Plus Environment

Page 23 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Plasma pharmacokinetic parameters for compound 28 were determined in Han Wistar rats
(male; intravenous and oral) and beagle dogs (male; intravenous only). The compound was
administered orally (PO) in rats as a suspension in 0.5% hydroxypropylmethylcellulose, 0.1%
tween in water at a dose of 5 mg/kg, and intravenously (IV) as a solution in 5% DMSO:95%
hydroxylpropyl βꢀcyclodextrin (30% w/v) in water, at doses of 1 and 2 mg/kg in dog and rat
respectively, as shown in Table 7. The compound was characterized by high clearance (Cl) in
rat (75 mL/min/kg) despite moderate predictions based on hepatocyte data (Clint 15
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6
µl/min/10 cells, predicted clearance 42 mL/min/kg) with low bioavailability consistent with
a high first pass effect. In dog the clearance was lower (29 mL/min/kg), and in better
6
accordance with the in vitro predictions based on hepatocyte data (Clint 5 µl/min/10 cells,
predicted clearance 24 mL/min/kg). No discernible turnover was observed in human
6
hepatocytes (Clint <1.1 µl/min/10 cells; n=8), translating through to a low to moderate
predicted clearance in vivo (<6 mL/min/kg). A rat bile duct cannulated study (2 mg/kg IV,
using the same formulation as the rat IV study) was carried out to investigate the under
prediction observed when scaling clearance from rat hepatocytes. A significant proportion of
compound was excreted as parent in the urine (~27%) with low levels of parent eliminated in
the bile (~2%). The amount eliminated in the urine was in excess of what would be expected
from passive diffusion alone, suggesting active secretion was occurring. By considering the
active renal secretion component, clearance due to metabolism was calculated as 50
mL/min/kg and in much closer agreement with the prediction from hepatocytes (42
mL/min/kg). This data suggested that the amount of active renal secretion occurring in the
dog was low, given the good prediction of clearance when scaling from dog hepatocytes.
a
Table 7. Pharmacokinetic parameters for compound 28
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Species
Rat
Dose (mg/kg)
Cl (mL/min/kg) Vss
L/kg)
PO halfꢀ IV halfꢀ
F
abs
F (%)
(
life (h)
life (h)
0.8
(%)
100
ꢀ
2, 5
1
75
29
2.1
2.0
ꢀ
28
ꢀ
Dog
3.0
3.3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Compounds were dosed intravenously at either 1 (dog) or 2 mg/kg (rat) in 5% DMSO:95%
hydroxylpropyl βꢀcyclodextrin (30% w/v) in water and orally at 5 mg/kg (rat) using a 0.5%
hydroxypropylmethylcellulose, 0.1% tween suspension in water, at a volume of 2 mL/kg and
4
mL/kg (rat, intravenous and orally respectively) and 1 mL/kg (dog).
Oncogenic activation of NFꢀκB pathway via MYD88 involves IκB kinase β (IKKβ)–
4
mediated phosphorylation and degradation of IκB proteins. Treatment of OCIꢀLY10 cells, an
ABCꢀDLBCL cell line harboring a MYD88 mutation, with compound 28 suppressed IKK
activation, as indicated by decreased phosphorylation of IκBꢀα (Figure 8A). The inhibition of
NFꢀkB signaling by compound 28 also inhibited growth of OCIꢀLY10 cells in a doseꢀ
dependent manner, whereas SUDHL2, a GCBꢀcell line was not sensitive to IRAK4 inhibitor
(Figure 8B). Similar data was obtained in other ABC and GCBꢀDLBCL cell lines (data not
shown).
Figure 8. Compound 28 inhibits NFꢀκB signaling and growth of ABCꢀDLBCL cell line. A)
Doseꢀdependent inhibition of IκBα phosphorylation in OCIꢀLY10 cells upon treatment of
ACS Paragon Plus Environment

Page 25 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
cells with compound 28 for 14 h. B) OCIꢀLY10 and SUDHL2 were monitored for growth in
presence of various concentrations of compound 28 in a 3ꢀday growth assay.
While compound 28 inhibits phosphorylation of IκBα and shows a concentration dose
response, we did not observe complete inhibition of IκBα phosphorylation at concentrations
up to 3.3 ꢁM. In addition to MYD88, the OCIꢀLY10 cell line has a mutation in CD79A.
ABCꢀDLBCL cell lines, like OCIꢀLY10, are characterized by genetic aberrations in both the
BCR signalling (CD79A or B, CARD11) and the MYD88 pathways, suggesting that dual
inhibition of these two pathways could be required for complete inhibition of NFꢀκB
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
31
signaling. Bruton's tyrosine kinase (BTK) is a crucial kinase in the BCR signaling pathway
that maintains NFꢀκB signaling and it has been shown to be a promising therapeutic target in
32
33
ABCꢀDLBCL. Addition of 3 nM of the BTK inhibitor ibrutinib to compound 28 inhibits
NFꢀκB signalling (Figure 9A) at much lower concentrations than either compound 28 or
ibrutinib alone and leads to complete inhibition of IκBα phosphorylation at concentrations
above 0.37 ꢁM (compare figure Figure 8A with 9A). Inhibition of NFꢀκB activity translates
to synergistic effects on the loss of cell viability in cells treated with a combination of
compound 28 and ibrutinib (Figure 9B). There is an increase in cell death in OCIꢀLY10 cells
upon increasing concentrations of compound 28 and BTK ibrutinib (Figure 9B, 9C). In
contrast, no increased cell killing was observed for SUDHL2, a GCBꢀDLBCL cell line that is
not dependent on MYD88 or BCR signaling (Figure 9B). Induction of cell death upon
combination of compound 28 and ibrutinib in OCIꢀLY10 cells was also confirmed by
detection of cleaved caspase 3 in cellular lysates which is a known marker of apoptosis
(Figure 9D).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 9. Compound 28 combines synergistically with ibrutinib to inhibit growth and
NF-κB signaling in vitro in OCI-LY10 cells. A) Effects of compound 28 upon combination
with ibrutinib (3 nM) on IκBα phosphorylation in OCIꢀLY10 cells. B) Percent growth
inhibition of compound 28 on proliferation of OCIꢀLY10 or SUDHL2 cells upon
combination with ibrutinib in a 5x5 dose matrix in a 3ꢀday growth assay. C) Schematic to
show the typical scale in a combination growth assay where % growth inhibition is measured
from 0ꢀ200 and 100 is regarded as zero growth or stasis. Increasing numbers from 100ꢀ200
exemplifies increase in cell death. D) Western blots showing induction of apoptosis, as
measured by cleavage of caspase 3 upon treatment of OCIꢀLY10 cells with a combination of
10 nM ibrutinib and either 0.3 or 3 µM of compound 28.
We subsequently investigated the antiꢀtumor efficacy of compound 28 in combination with
ibrutinib in OCIꢀLY10 xenograft studies in mouse. As shown in Figure 10 and Table 8, both
compound 28 (12.5 mg/kg) and ibrutinib (12 mg/kg) dosed orally (qd) had modest antiꢀtumor
activity as single agents but a combination of both agents led to tumor regression. All three
ACS Paragon Plus Environment

Page 27 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
regimens were well tolerated throughout treatment (Figure 10B). The in vivo efficacy data is
consistent with the in vitro data where either of the compounds alone (compound 28 or
ibrutinib) leads to cellular stasis but a combination of both leads to cell death and synergistic
inhibition of NFꢀκB signaling. We considered whether any of the offꢀtargets identified
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
4
(CLK1,2,4 or haspin; Table 6) could be making a contribution to the observed efficacy.
Whilst it is not possible to rule this out, we were reassured by the subsequent in vivo profiling
of additional inꢀhouse derived IRAK4 chemical equity with improved selectivity against
these targets. These compounds produced similar findings to 28 both as single agent and in
combination with ibrutinib and we aim to publish these findings in due course.
Pharmacokinetic analysis (Figure 11) of compound 28 indicated sufficient exposure to give
coverage over the cellular IRAK4 IC for up to 6 hours daily with or without ibrutinib. No
5
0
apparent drugꢀdrug interactions were observed in terms of the concentrations of compound 28
when coꢀdosed with ibrutinib. Compound 28 as a single agent leads to modest tumor growth
inhibition and this may be due to the partial coverage over IRAK4 IC50 for 6 hours and
presence of lower levels of the compound in circulation at later timepoints. But in
combination with ibrutinib, these levels of IRAK4 inhibition appear sufficient to inhibit NFꢀ
kB signaling resulting in inhibition of cell growth and survival.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 10. Anti-tumor activity of compound 28 in combination with ibrutinib in OCI-
LY10, an ABC-subtype DLBCL, xenograft. A) Daily oral dosing of Vehicle (x) ibrutinib,
1
2 mg/kg (■), compound 28 12.5 mg/kg (●), and combination (▲) in female SCID mice (n
=
10 per data point) with blue colouring representing duration of dosing. B) Body weight data
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
for the four dose groups.
Table 8. Anti-Tumor Activity of compound 28, Ibrutinib and Combination in the OCI-
LY10 Xenograft Model
Dose
mg/kg)
ꢀ
Tumor Growth
pꢀvalue
Treatment
Schedule
(
Inhibition (Day 51) (Day 51)
Vehicle
ibrutinib
qd (po)
qd (po)
qd (po)
qd (po)
qd (po)
ꢀ
ꢀ
12
56%
60%
0.0079
0.0009
compound 28
ibrutinib
12.5
12
>
100%
<0.0001
compound 28
12.5
ACS Paragon Plus Environment

Page 29 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 11. Steady-state plasma concentration with time in mice bearing OCI-LY10
xenografts. Female SCID mice were administered seven daily oral doses and plasma taken
postꢀseventh dose of compound 28 12.5 mpk alone (●) and in combination with ibrutinib (▲)
(n =3 per data point).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
CONCLUSIONS
In conclusion, we have identified and optimised a series of pyrrolopyrimidine inhibitors of
IRAK4 using Xꢀray crystal structures to identify key interactions and optimize our scaffold.
This led to the identification of compound 28 as a potent inhibitor of IRAK4. Comparison of
small molecule Xꢀray structure and solution phase NMR data of 28 with the Xꢀray structure
of the protein ligand complex confirmed the ligand was conformationally preꢀorganized for
binding. Compound 28 showed a favorable physicochemical and kinase selectivity profile
and was identified as a promising in vivo tool with which to explore the potential role of
L265P
IRAK4 inhibition in the treatment of mutant MYD88
diffuse large Bꢀcell lymphoma.
Compound 28 inhibited NFꢀκB activation and growth of ABCꢀDLBCL cell lines in a doseꢀ
dependent manner. In combination with a BTK inhibitor, compound 28 completely inhibited
NFꢀκB signaling and induced cell death at lower concentration in vitro in OCIꢀLY10 cells. In
vivo, combination of compound 28 and ibrutinib led to tumor regression in an ABCꢀDLBCL
mouse model (OCIꢀLY10).This is consistent with OCIꢀLY10 being driven by both CD79 and
MYD88 mutations wherein inhibition of both the BCR signaling and MYD88 signaling
would be synergistic in inhibiting the NFꢀκB pathway. This data suggests that
pyrrolopyrimidine inhibitors of IRAK4 in combination with a BTK inhibitor has therapeutic
potential for the treatment of DLBCL.
ACS Paragon Plus Environment