Discovery of 6-Amino-2-{[(1S)-1-methylbutyl]oxy}-9-[5-(1-piperidinyl)pentyl]-7,9-dihydro-8H-purin-8-one (GSK2245035), a Highly Potent and Selective Intranasal Toll-Like Receptor 7 Agonist for the Treatment of Asthma

Article abstract of DOI:10.1021/acs.jmedchem.5b01647

Induction of IFNα in the upper airways via activation of TLR7 represents a novel immunomodulatory approach to the treatment of allergic asthma. Exploration of 8-oxoadenine derivatives bearing saturated oxygen or nitrogen heterocycles in the N-9 substituent has revealed a remarkable selective enhancement in IFNα inducing potency in the nitrogen series. Further potency enhancement was achieved with the novel (S)-pentyloxy substitution at C-2 leading to the selection of GSK2245035 (32) as an intranasal development candidate. In human cell cultures, compound 32 resulted in suppression of Th2 cytokine responses to allergens, while in vivo intranasal administration at very low doses led to local upregulation of TLR7-mediated cytokines (IP-10). Target engagement was confirmed in humans following single intranasal doses of 32 of ≥20 ng, and reproducible pharmacological response was demonstrated following repeat intranasal dosing at weekly intervals.

Full text of DOI:10.1021/acs.jmedchem.5b01647

This is an open access article published under an ACS AuthorChoice License, which permits
copying and redistribution of the article or any adaptations for non-commercial purposes.
Drug Annotation
pubs.acs.org/jmc
Discovery of 6‑Amino-2-{[(1S)‑1-methylbutyl]oxy}-9-[5-(1-
piperidinyl)pentyl]-7,9-dihydro‑8H‑purin-8-one (GSK2245035), a
Highly Potent and Selective Intranasal Toll-Like Receptor 7 Agonist
for the Treatment of Asthma
Keith Biggadike,† Mahbub Ahmed,† Doug I. Ball,† Diane M. Coe,*,† Deidre A. Dalmas Wilk,§
Chris D. Edwards,† Bob H. Gibbon,‡ Charlotte J. Hardy,† Stephen A. Hermitage,† Joanne O. Hessey,†
Aimee E. Hillegas,§ Stephen C. Hughes,‡ Linos Lazarides,† Xiao Q. Lewell,† Amanda Lucas,†
David N. Mallett,† Mark A. Price,‡ Fiona M. Priest,† Diana J. Quint,† Poonam Shah,† Anesh Sitaram,‡
Stephen A. Smith,† Richard Stocker,† Naimisha A. Trivedi,† Daphne C. Tsitoura,† and Victoria Weller†
†GlaxoSmithKline R&D, Medicines Research Centre, Gunnels Wood Road, Stevenage, Hertfordshire SG1 2NY, U.K.
‡GlaxoSmithKline R&D, David Jack Centre, Park Road, Ware, Hertfordshire SG12 ODP, U.K.
§GlaxoSmithKline R&D, UpperMerion, 709 Swedeland Road, King of Prussia, Pennsylvania 19406, United States
S
* Supporting Information
ABSTRACT: Induction of IFNα in the upper airways via activation of TLR7 represents a novel immunomodulatory approach
to the treatment of allergic asthma. Exploration of 8-oxoadenine derivatives bearing saturated oxygen or nitrogen heterocycles in
the N-9 substituent has revealed a remarkable selective enhancement in IFNα inducing potency in the nitrogen series. Further
potency enhancement was achieved with the novel (S)-pentyloxy substitution at C-2 leading to the selection of GSK2245035
(32) as an intranasal development candidate. In human cell cultures, compound 32 resulted in suppression of Th2 cytokine
responses to allergens, while in vivo intranasal administration at very low doses led to local upregulation of TLR7-mediated
cytokines (IP-10). Target engagement was confirmed in humans following single intranasal doses of 32 of ≥20 ng, and
reproducible pharmacological response was demonstrated following repeat intranasal dosing at weekly intervals.
otherwise innocuous environmental proteins (allergens). The
resulting overproduction of Th2 cytokines (e.g., IL-4, IL-5, IL-
9, and IL-13) drives the differentiation, survival, and functions
of the major effector cells, mast cells, basophils, and eosinophils
that are responsible for the manifestation of symptoms.2
Furthermore, Th2 cells play an essential role in the generation
of allergen-specific IgE by B cells, which mediates the type 1
hypersensitivity reactions, and in the orchestration of local
chronic allergic inflammation. Restoration of a more balanced
Th2/Th1 and regulatory T cell response toward respiratory
allergens could prevent the development of allergic symptoms
and potentially lead to long-term modification of disease
pathogenesis. Currently, allergen-specific immunotherapy
(SIT) is the only treatment that has been shown to modify
INTRODUCTION
■
Allergic rhinitis (AR) and asthma are two closely related and
very common chronic diseases of the airways. Their incidence
continues to increase with 10−30% of the world’s population
suffering from one or both conditions. Asthma is by far the
more serious of the two diseases and is responsible for
approximately 250,000 deaths annually worldwide.1 The
current standard of care for allergic asthma consists of inhaled
corticosteroids, to tackle the underlying inflammation, along
with short and long acting inhaled β-2 agonist bronchodilators
to relieve constriction of the airways. While such treatment
delivers reliable control of asthma symptoms and exacerbations
for the majority of patients, it requires continuous use of
medication while the underlying immune pathology remains
unaltered.
Allergic asthma is a complex disease which is characterized by
an aberrant type-2 T-helper cell (Th2) immune response to
Received: October 20, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.jmedchem.5b01647
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
this aberrant immune response to allergens, resulting in allergen
“desensitization” but usually requires long-term administration
(3−5 years) of the allergen(s) to which the patient is sensitive
and is accompanied by significant safety drawbacks.3 Con-
sequently, alternative approaches aiming to modulate the
immune milieu of the airways have begun to be explored. To
this end, scientific interest has recently been directed toward
the manipulation of Toll-like receptor (TLR) function.4
Human TLRs comprise a family of 10 transmembrane
glycoproteins that are widely expressed on immune and
epithelial cells and act as “danger sensors” by recognizing
conserved microbial components known as pathogen-associ-
ated molecular patterns (PAMPs). Activation of TLRs results in
mobilization of both innate and adaptive immune mechanisms
that eliminate “danger” and preserve immune homeostasis.5
Because of their critical role in immune regulation, TLRs have
been considered attractive targets for therapeutic intervention
in a wide variety of disease settings.6 In the context of allergic
respiratory diseases, local stimulation of TLRs in the airways
resulting in the production of Th1 directing cytokines, in
particular type 1 interferons (IFNs), has the potential to
modulate the aberant immune response to allergens. A
deficiency in this regulatory mechanism due to insufficient
microbial exposure in childhood is thought to be a key driver in
the increasing incidence of allergy, the so-called “Hygiene
Hypothesis”.7 The value of TLR activation in the treatment of
allergic diseases is now well recognized and is clearly illustrated
by the enhanced efficacy seen when TLR4 or TLR9 agonists
are included as adjuvants in SIT.8,9
Furthermore, studies in rodent models of allergic airway disease
have shown that local delivery of small molecule TLR7 agonists
results in inhibition of both acute inflammation and chronic
remodelling and in direct relaxation of airway smooth muscle
and reduction of airway hyperreactivity.16 More importantly,
comparison of the action of multiple classes of TLR ligands in
an experimental asthma model concluded that TLR7 agonists
have, in addition to the strongest antiallergic effect, the lowest
adverse pro-inflammatory potential.17 The powerful antiviral
potential of TLR7 agonists may also offer extra benefit in
asthma patients through the better control of virus-driven
asthma exacerbations.18 Finally, from a drug discovery
perspective, TLR7 has the distinct advantage of being amenable
to activation by small molecule ligands.19
TLR7 recognizes naturally single stranded viral RNA and,
along with the other TLRs which recognize microbial nucleic
acids (TLR3, 8 and 9), is activated within the endosomal
compartment of the cell. Upon ligand binding, TLR7 dimerizes
resulting in a signaling cascade via the adapter protein MyD88
culminating in the production of IFNα via an IRF-7 mediated
pathway and pro-inflammatory cytokines including TNFα via
an NFκ-B mediated pathway (Figure 1). TLR7 is expressed
Stand-alone application of TLR agonists to the upper
airways, the main site of exposure to allergens, is a novel
concept for the treatment of respiratory allergies that has only
recently begun to be evaluated in clinical trials.10,11 The
“United Airways Disease” hypothesis leads us to believe that
triggering an immunmodulatory response in the nose will
translate to benefit in both allergic rhinitis and asthma.12
Compared to SIT, this approach has the advantage of
theoretically providing protection from all aeroallergens to
which a patient is naturally exposed during treatment, thereby
eliminating the need for adjustment of treatment to each
patient’s sensitization. Administration of TLR agonists to the
airways has the potential to induce immune changes that will
restrict the existing allergen-specific Th2 responsiveness and
rebalance it toward a prophylactic Th1/Treg phenotype.13
Consequently, in addition to providing short-term symptom
control, this approach has the potential to deliver long-term
disease remission via allergen dependent T-cell and B-cell
memory responses. Evidence for short-term reduction in
allergic reactivity using this approach has been obtained
through early clinical evaluation of the intranasal TLR8 agonist,
2-amino-N,N-dipropyl-8-[4-(1-pyrrolidinylcarbonyl)phenyl]-
3H-1-benzazepine-4-carboxamide (VTX-1463) in allergic rhi-
nitis11 and with the intranasal plasma labile TLR7 agonist “ante-
drug” methyl [3-({[3-(6-amino-2-butoxy-8-oxo-7,8-dihydro-
9H-purin-9-yl)propyl][3-(4-morpholinyl)propyl]amino}-
methyl)phenyl]acetate (AZD8848)10a in both allergic rhini-
tis10b,c and asthma.10d However, duration of effect of longer
than 1 week has yet to be demonstrated following a course of
intranasal TLR treatment (up to 8 weeks).
Figure 1. TLR7 signaling pathways.
primarily on plasmacytoid dendritic cells (pDC) which on
activation produce large amounts of type-1 interferons, and it is
engagement of this cell type in the nasal mucosa that is believed
will drive the desired immunomodulatory response. Our
objective therefore was to identify a small molecule agonist of
TLR7, suitable for intranasal delivery, to explore this hypothesis
preclinically and, if successful, in human subjects.
SMALL MOLECULE TLR7 AGONISTS
■
Since the discovery in 1957 of type-1 interferon and its
potential antiviral and antitumor benefits,20 a number of small
molecule inducers of this key cytokine have been discovered.
Many of these have subsequently been shown to be agonists of
TLR7 and/or TLR8. Unsurprisingly, given that the natural
ligand for TLR7/8 is single stranded RNA, many of these small
molecule agonists resemble nucleosides or nucleoside bases
(Figure 2). Simple pyrimidine based IFN inducers, typified by
Among the TLRs, TLR7 is of particular importance in
respiratory allergic diseases. TLR7 polymorphism has been
strongly linked to atopy and asthma,14 while reduced function
of TLR7 has been demonstrated in asthmatic adolescents.15
B
DOI: 10.1021/acs.jmedchem.5b01647
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
chains of varying length as represented by the generic structures
2 and 3.
Whereas the reported modifications of the 8-oxoadenine
series had been directed toward the discovery of orally active
TLR7 agonists, we were looking to identify compounds that
would selectively engage TLR7 locally in the nose after
intranasal delivery with minimal systemic exposure. Con-
sequently, our target compound profile was very different.
We were aiming to identify a potent TLR7 agonist with high
clearance and low oral bioavailability to minimize the risk of
unwanted systemic cytokine induction from drug entering the
circulation via nasal absorption or via gut absorption of any
swallowed portion of the intranasal dose. This approach would
also serve to minimize any risk of autoimmunity that might
arise from systemic TLR7 activation.34 In addition, to minimize
the risk of a local inflammatory response, we ideally wanted the
compound to be selective for induction of IFNα over TNFα, a
pro-inflammatory cytokine implicated in many aspects of airway
pathology in asthma,35 although it was not clear at the outset
how this selectivity might be achieved. Finally, the compound,
or a suitable salt, needed to be sufficiently soluble and stable in
water to allow delivery as an intranasal solution formulation.
Figure 2. Representative small molecule TLR7 agonists.
bropirimine, were discovered by Upjohn in the 1970s,21 while
guanine nucleoside analogues, including isatoribine and
loxoribine, were discovered in the 1980s and shown to exert
their effects via TLR7 in 2003.22 The imidazoquinoline
analogues, also discovered in the 1980s, provided the first
small molecule IFN inducer to be introduced into clinical
practice with imiquimod launched in 1997 as a topical
treatment (Aldara) for a variety of skin diseases including
genital warts, actinic keratosis, and basal cell carcinoma.23
Resiquimod is a more potent imidazoquinoline with dual
TLR7/8 agonist activity which has progressed to clinical
evaluation for a variety of indications but which has yet to
deliver an approved medicine.24 In 2002, Sumitomo discovered
the 8-oxoadenine class of interferon inducers,25 and this series
was confirmed to act via TLR7 in 2006.26 We became
interested in TLR7 as an immunomodulatory mechanism in
2007 and conducted our own high throughput screen for IFN
inducers but failed to identify any novel leads. Faced with the
known scaffolds, we chose to investigate 8-oxoadenines as this
series offers versatility by allowing activity and physicochemical
properties to be manipulated by modification of both the C-2
and N-9 substituents. Subsequently, a variety of novel TLR7
agonist scaffolds retaining the key pharmacophore of 8-
oxoadenines have been discovered and evaluated in the search
for oral treatments for hepatitis B and C.27 These include
deazaadenine derivatives from Pfizer28 and 6,6-pteridinone
analogues from Gilead, an example of which (36, GS-9620) is
currently in the clinic for HBV.29
IN-VITRO PROFILING
■
A variety of screening strategies have been employed in the
search for novel IFNα inducing small molecules. Historically,
these involved directly monitoring IFNα induction or the
resulting antiviral activity after dosing compounds in vivo,
usually in mice.21 Subsequently, measuring IFNα induction in
vitro following incubation of compounds with mouse
splenocytes has been employed, for example, in the early
optimization of the 8-oxoadenine series.25 More recently, and
specifically for TLR7/8 agonists, reporter assays have been
employed giving readouts of activation of the NFκ-B arm of
TLR7/8 signaling, most commonly in a HEK293 cell line.36
Although reproducible and amenable to high throughput, we
chose not to employ such reporter assays as they provide a
readout only of activation of the undesired pathway leading to
TNFα induction. We chose instead to evaluate compounds
using a human peripheral blood mononuclear cell (PBMC)
assay to provide a direct readout of IFNα induction in a more
relevant cell population including the primary target
plasmacytoid dendritic cells. Incubation of test compounds
with cryopreserved human PBMC (method A) for 24 h
followed by MSD analysis for IFNα gave dose−response curves
from which pEC50 values were derived. This assay format was
used to guide SAR development during the early phase of the
program, and subsequently, to gain insight into selectivity of
TLR7 activation, the assay was adapted using freshly isolated
PBMC (method B) to give readouts of both IFNα and TNFα
induction.
MEDICINAL CHEMISTRY STRATEGY
■
Published SAR on the 8-oxoadenine series had demonstrated
considerable scope for modification of the C-2 substituent on
the 8-oxoadenine template with n-butyloxy identified as a
preferred substituent.30 There appeared to be less scope for
modification of the N-9 substituent with benzyl being
established as a particularly favored potency enhancing group.
Compound 1 (SA-2), combining these two activity enhancing
features, shows activity following intraperitoneal dosing in a
mouse model of Th2 mediated airway inflammation31 and, as
such, provided a useful benchmark for our novel analogues.
Exploration of the N-9 substituent had centered primarily on
substituted benzyl and pyridylmethyl analogues,32 and there
were no examples of saturated heterocycles having been
incorporated at this position.33 We therefore embarked on a
program of work to explore both 5 and 6-membered saturated
oxygen and nitrogen heterocycles at N-9 linked via methylene
OXYGEN HETEROCYCLES
Analogues first investigated contained saturated oxygen hetero-
cycles in the N-9 substituent and explored the impact of chain
■
C
DOI: 10.1021/acs.jmedchem.5b01647
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
a
length and position of attachment to the oxygen heterocycle in
the 2-n-butyloxy series. Compounds were prepared using the
methodology shown in Scheme 1 and were profiled using
cryopreserved human PBMC to give pEC50 values for IFNα
induction following 24 h of incubation.
Table 2. THP/THF Isomer SAR
a
Scheme 1. Synthetic Route to Oxygen Heterocycles
a
X = O, NH; Y = Br or methanesulfonate ester leaving group.
This initial data set revealed remarkable tolerance of a variety
of saturated oxygen heterocyclic substituents at the N-9
position (Tables 1 and 2). Encouraging activity was seen with
a
Table 1. 4′-THP Chain Length SAR
a
PBMC assay method A; pEC50 values are geometric means of at least
two independent measurements unless otherwise indicated.
Compound
n
IFNα pEC50
b
4
5
6
7
8
9
1
2
3
4
5
6
6.7
6.7
7.2
7.0
7.3
6.1
a
PBMC assay method A; pEC50 values (negative log of the molar
EC50) are geometric means of at least two independent measurements
b
unless otherwise indicated. n = 1.
4′-linked tetrahydropyran (THP) derivatives linked through
chains of 1 to 6 carbon atoms with activity superior to the
benchmark 9-benzyl derivative 1 (pEC50 6.9) seen with the n-
propyl, n-butyl, and n-pentyl-linked derivatives (6, 7, and 8)
(Table 1).
Examination of alternatively linked THP and tetrahydrofuran
(THF) analogues indicated that the ring size and position of
attachment also had little impact on activity (Table 2). Thus,
the 3-linked THP 10 showed comparable activity to the 4-
linked counterpart 7, whereas the 2-linked isomer 11 was
slightly less active. Comparable activity was seen with THF
analogues with the 3-linked isomer 12 again appearing more
active than the 2-linked analogue 13. Also, where investigated,
chirality at the point of attachment to the heterocycle appeared
to have little effect on activity as illustrated by the THF
enantiomers 14 and 15.
Encouraged by these initial results we explored the SAR of
THF and THP analogues more extensively using an array
approach which additionally incorporated investigation of the
impact of the C-2 substituent (Figure 3). Five different C-2
groups were investigated; R2 = n-butyloxy and n-butylamino as
standards and three novel C-2 substituents; R2 = cyclo-
propylethoxy, (R)-2-pentyloxy and (S)-2-pentyloxy (Figure 3).
Figure 3. Composition of the N-9 oxygen heterocycle array.
The latter pair of isomers was chosen based on earlier
encouraging data with this C-2 substituent in racemic form.
Data from this array of compounds (Figure 4) reinforced the
earlier observations that good IFNα potency could be achieved
with a wide variety of THF or THP containing N-9
substituents. Furthermore, some interesting trends were
observed regarding SAR of the 2-substituent. Consistent with
data in the literature, n-butyloxy was found to be superior to n-
butylamino, while the cyclopropylethoxy substituent provided
potency similar to that of n-butyloxy. More interesting was the
clear distinction between the novel (R) and (S)-pentyloxy
analogues with the latter isomer being clearly the more potent
and also more potent that the standard n-butyloxy substitution
at this center. This learning was taken forward into the N-9
D
DOI: 10.1021/acs.jmedchem.5b01647
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
a
Table 3. Nitrogen Heterocycle SAR
Figure 4. Spotfire plot of oxygen heterocycle array. PBMC assay
method A; pEC50 values are the geometric means of at least two
independent measurements unless otherwise indicated.
amino heterocyclic series where further exploration of the C-2
substituent was undertaken.
Examination of the cytokine induction profiles of the more
potent examples from the saturated oxygen heterocyclic series
revealed only modest selectivity for induction of IFNα over the
unwanted pro-inflammatory cytokine TNFα. For example, the
2-(S)-pentyloxy 4′-linked THP analogue 16 showed pEC50
values for IFNα and TNFα induction in human PBMC of 7.4
and 6.9, respectively.
NITROGEN HETEROCYCLES
■
Concerns that incorporation of a basic amine into the N-9
substituent might restrict permeability and prevent access to
TLR7 located in the endosome were quickly allayed when a
variety of nitrogen heterocycles were found to show enhanced
IFNα potency compared to that of their oxygen counterparts in
cellular assays. Surprisingly, this enhancement in IFNα potency
was not mirrored by an enhanced TNFα response, and
consequently, the amine derivatives displayed much improved
IFNα/TNFα selectivity (Table 3). As with the oxygen
heterocycles, there was found to be a wide tolerance for a
variety of nitrogen heterocycles at this position including N-
and C-linked piperidines (17−20) and N-linked piperazines
(21). Pyrrolidines were also active and selective, although the
N-linked derivative 22 was somewhat less potent than its 6-
membered counterpart 17, while expansion to the 7-membered
azepane 23 resulted in a slight enhancement in activity.
However, the morpholino derivative 24 was markedly less
potent and provided the first indication that the basic amine
was key to the enhanced IFNα potency in this series.
a
PBMC assay method B; pEC50 values are geometric means of at least
two independent measurements unless otherwise indicated.
there was a significant reduction in IFNα potency with linking
chains of less than 4 carbon atoms, but high levels of potency
and selectivity were seen with chains of 4−6 carbon atoms
(Table 4).
Final optimization of the C-2 substituent was undertaken in
the N-linked piperidine series linked by either the n-butyl or n-
pentyl chain (Table 5). The enhancement in potency seen
earlier in the oxygen series with the 2-(S)-pentyloxy substituent
was again observed with examples 29 and 32 being
approximately 10-fold more potent than their n-butyloxy
counterparts 17 and 27, respectively. The longer 2-(S)-hexyloxy
derivative 30 offered no advantage while shortening to the 2-
(S)-butyloxy derivative 33 resulted in a slight drop in potency.
Further truncation to give the isopropyloxy derivative 34 was
accompanied by a more marked reduction in IFNα potency
The N-linked piperidine series was selected for more detailed
examination beginning with a study of the effect of the length of
the linking chain. In contrast to the oxygen heterocyclic series,
E
DOI: 10.1021/acs.jmedchem.5b01647
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
Table 4. N-Linked Piperidine Chain Length SAR
N-9 MODIFICATION: STRUCTURAL
CONSIDERATIONS
■
The remarkable tolerance of a wide variety of substitutions at
the N-9 position of the 8-oxoadenine template has also been
observed by others, most notably by the Sumitomo/
AstraZeneca group who have used this position to incorporate
elaborate plasma labile ester “ante-drug” functionality. For
example, compound 35 (AZ12441970) is characterized by a
very bulky N-9 substituent incorporating not only a plasma
labile ester but also basic amine functionality.38 This tolerance
is further illustrated by the retention of TLR7 agonist activity
with 8-oxoadenines containing a fluorophore linked at N-9.39
Similar tolerance has been demonstrated at the corresponding
position on the imidazoquinoline scaffold.39 Activation of
TLR7 (and 8) by a variety of small molecule ligands is
remarkable given that the natural ligand for these key
components of our immune defense is the complex macro-
molecule ssRNA. A crystal structure of TLR7 has yet to be
reported, but the first TLR8 ectodomain crystal structure
revealed a site in which imidazoquinoline agonists, including
resiquimod, bind and cause a conformational reorganization of
the dimeric structure bringing the C-termini into closer
proximity.40 This crystal structure has been used to derive
TLR7 homology models41 which have been used to dock other
small molecule ligands. For example, the Gilead group have
shown that the pteridinone derivative 36 is able to bind into the
same ligand binding site and that this model could
accommodate the bulky amine containing substituent of this
compound.29
b
Compound
n
IFNα pEC50
TNFα pEC50
a
25
26
17
27
28
2
3
4
5
6
5.6
NT
NT
5.2
5.3
5.7
a
6.5
b
8.2
b
8.1
b
8.8
a
b
PBMC assay method A. PBMC assay method B; pEC50 values are
geometric means of at least two independent measurements unless
otherwise indicated. NT = not tested.
a
Table 5. C-2 Substituent SAR in N-Linked Piperidine Series
Very recently, a second TLR8 crystal structure from the
Tokyo University group has revealed a surprising aspect to
activation of TLR8 by ssRNA.42 This new X-ray structure
suggests that activation of TLR8 by the natural ligand does not
involve the binding of intact ssRNA but instead involves the
binding of two degradation products thereof. Uridine is shown
to bind at the same site at the imidazoquinoline ligands while a
short oligonucleotide binds to a second site on the receptor.
Occupation of both sites is required for activation of TLR8 by
ssRNA. It remains to be seen whether TLR7 activation also
relies on the binding of fragments of the natural ligand at more
than one binding site, and therefore, TLR7 homology modeling
based on the TLR8 crystal structure must be questionable.
a
N-9 MODIFICATION: ENHANCED IFNα
POTENCY/SELECTIVITY
PBMC assay method B; pEC50 values are geometric means of at least
■
two independent measurements unless otherwise indicated.
Enhancement of TLR7 agonist activity, and aqueous solubility,
of 8-oxoadenines by incorporation of a basic amine into the N-9
substituent has been reported by the Sumitomo/AZ group43
and also by the Pfizer group on a closely related 3-deaza
oxoadenine template.44 However, both of these groups assessed
TLR7 activity using an NFκ-B reporter assay and do not report
on the IFNα/TNFα selectivity of their analogues. In contrast,
Gilead have employed a screening strategy similar to our own
while cyclization to the cyclopentyloxy analogue 31 resulted in
a dramatic loss in activity of ca. 1000-fold.
Compound 32 (GSK2245035)37 was selected from this
series as a development candidate as it combines excellent
potency and selectivity with the desired pharmacokinetic and
pharmaceutical developability properties (vide infra).
F
DOI: 10.1021/acs.jmedchem.5b01647
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
one, measuring both IFNα and TNFα induction in human
PBMC in discovering the N-9 amino containing oral HBV asset
36, on the related 6,6-pteridinone template.29 In this series,
IFNα/TNFα selectivity up to 100-fold was reported (based on
minimum effective concentration (MEC) values), but selectiv-
ity was much less consistent than in our 8-oxoadenine series.
They ascribe TNFα induction to TLR8 activity and show, with
compound 36, a reasonable correlation of IFNα and TNFα
MEC values with TLR7 and TLR8 reporter assay EC values,
respectively.
Workers at Novartis used fluorescently tagged 8-oxoadenine
and imidazoquinolines to investigate intracellular localization in
human pDC and showed these compounds to localize in
endosomes.39 The Pfizer group postulate that the enhancement
in IFNα potency with their basic analogues may be the result of
concentration in the acidic endosomal compartment where
TLR7 activation occurs, and we also believe this to be the case,
but this alone would not explain the selective enhancement of
IFNα induction over TNFα induction if the latter is also driven
by endosomal TLR7 activation. However, recent studies have
shown that in murine bone marrow derived pDC incubated
with vaccinia or flu virus the TLR7/IRF7 pathway driving IFNα
induction and the TLR7/NF-κB pathway leading to TNFα
induction occur in different endosomal compartments.
Importantly, activation to give IFNα occurs in the lysosome
related organelle (LRO), a process reported to be dependent
on the trafficking of TLR7 by adaptor protein 3 (AP-3) (Figure
5).45 The highly acidic nature of the LRO (pH 4.5−5.5)46
This important endosomal compartmentalization of the
receptor is absent from TLR7 reporter assays which also
usually provide a readout only of signaling via the NF-κB
pathway. The extraordinary potency and selectivity of the
amino derivatives would therefore have been missed if the
TLR7 reporter had been used as a primary assay, as illustrated
by the weak activity seen with compound 32 in a TLR7
reporter assay in U20S cells (pEC50 = 5.7) compared to the
very high IFNα potency seen in human PBMC (pEC50 = 9.3).
Flow cytometry was used to identify the key cell types activated
by 32 in human whole blood cultures. These studies showed
IFNα to be produced exclusively by pDC while TNFα
originated in pDC, myeloid DC (mDC), and in particular
monocytes. As monocytes and mDC predominantly express
TLR8 and not TLR7,48 production of TNFα at higher drug
concentrations is likely to be due to weak TLR8 agonist activity
as demonstrated with compound 32 (vide infra).
SYNTHESIS
■
The synthetic route to compound 32 starting from 2-
fluoroadenine 37 is outlined in Scheme 2. Initial protection
of the N-9 position of 37 as the THP derivative 38 allowed the
2-(S)-pentyloxy side chain to be introduced by fluoro
displacement with 2-(S)-pentanol in the presence of sodium
tert-butoxide to give 39 in 95% yield. Bromination of 39 with
N-bromosuccinimide afforded the 8-bromo intermediate 40 in
99% yield. The masked 8-oxo functionality was then introduced
by treatment of 40 with sodium methoxide in methanol to give,
after removal of the THP protection with TFA, the 8-methoxy
core 42 in 79% yield for the two stages. Alkylation of 42 with 1-
bromo-5-chloropentane occurred predominantly at N-9, and a
small amount of N-7 alkylated material was observed, which
was removed by chromatography, to give the chloropentyl
intermediate 43 in 70% yield. Displacement of the chloro
functionality in 43 with piperidine yielded the 8-methoxy
precursor 44 (52%), which was finally deprotected under acidic
conditions to give the clinical candidate 32 as the crystalline
free base (73%). For safety assessment studies, and for clinical
use, compound 32 was converted in situ into the more soluble
and more stable maleate salt. Given the extraordinary potency
of this molecule (1 g ≡ 500 million 20 ng doses), unusually
small amounts of drug substances were required for safety
assessment studies and only tiny amounts for the provision of
clinical supplies. This route was successfully scaled up to
provide hundreds of grams of 32 with very strict containment
controls being applied to ensure safe handling of this highly
potent immunostimulatory agent.
PRECLINICAL PROFILING: IMMUNOLOGY
■
Cytokine Responses in Human PBMC and Whole
Blood. More detailed cytokine profiling of compound 32 was
conducted in human whole blood. Nanomolar IFNα potency
was retained in human whole blood, and IFNα/TNFα
selectivity was found to be even greater than that in PBMC
(Table 6). Similar selectivity was seen over IL-1β induction
giving a profile expected to minimize the potential for
pyrogenicity and inflammatory events associated with these
pro-inflammatory cytokines. In addition compound 32 was
shown to induce IL-12p70, IL-10, and IFNγ, cytokines known
to play a role in the reduction of Th2 responses, enhancement
of Th1 responses, and, in the case of IL-10, increase T
regulatory responses.
Figure 5. Compartmentalization of TLR7 signaling.
could result in selective concentration of the basic amine
derivatives in this compartment and therefore account for their
enhanced IFNα/TNFα selectivity. Supporting this hypothesis,
we have shown selective loss of TLR7 activated IFNα but not
NF-κB driven IL-12 production by splenic pDC from Pearl
mice which have an AP-3b1 mutation resulting in loss of
function of AP-3 complex47 (data not shown).
G
DOI: 10.1021/acs.jmedchem.5b01647
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
a
Scheme 2. Synthesis of Compound 32
a
Reagents and conditions: (a) N,O-bis(trimethylsilyl)acetamide, CH3CN, reflux then tetrahydropyran-2-yl acetate, TMS-triflate, CH3CN 0−10 °C,
49%; (b) 2(S)-pentanol, NaOBut 50 °C, 95%; (c) NBS, CHCl3, <5 °C, 99%; (d) NaOMe, MeOH, reflux, 97%; (e) TFA, MeOH, 20 °C, 81%; (f)
K2CO3, DMF, 50 °C then 1-bromo-5-chloropentane, 20 °C, 70%; (g) piperidine, NEt3, DMF, 70 °C, 52%; (h) HCl, dioxan, MeOH, 20 °C then
NaHCO3, 73%.
Table 6. Human in Vitro Cytokine Profile (20 h of
Incubation) for Compound 32
IFNα and TNFα responses in human and cynomolgus monkey
whole blood incubated with compound 32 showed the
potencies for IFNα to be very similar (pEC50 values 7.5 vs.
7.1 for human and monkey, respectively). The absolute human
IFNα potency in this assay was lower than that reported in
Table 6, possibly due to the use of an assay with different
antibody reagents from those used in the earlier studies. A
direct comparison of the human and monkey TNFα responses
showed comparable potencies in the two species (pEC50 values
5.3 vs 4.8 for humans and monkeys, respectively), showing that
the IFNα/ TNFα selectivity seen in humans was reflected in
cynomolgus monkey and giving confidence that in vivo studies
in the monkey are likely to be most predictive of responses in
humans.
TLR and Wider Selectivity Profiling. Selectivity for TLR7
over the other nucleic acid detecting TLRs (TLRs 3, 8, and 9)
was assessed in a human osteosarcoma (U20S) cell line
transfected with cDNA expression vectors for TLR7 or 8 and in
a human embryonic kidney (HEK) cell line transfected with
cDNA expression vectors for TLR3 or 9 using NF-κB-driven
luciferase reporter systems. Compound 32 showed only modest
activity (pEC50 5.9) in the TLR7 reporter system mirroring the
lower potency of 32 for the induction of TNFα via the NF-κB
pathway. Compound 32 showed very weak agonist activity
(pEC50 4.7) in the TLR8 reporter assay and was inactive when
tested up to 50 μM in HEK cells expressing TLR3 and TLR9.
The wider selectivity profile of 32 was evaluated against a panel
of 51 seven-transmembrane (7TM), enzyme, ion-channel,
kinase, transporter, and nuclear receptor targets. Weak activity
was detected against 12 targets, and these were evaluated in
secondary assays where pKi (affinity) values were determined,
and all were shown to be <6.
media
PBMC
cytokine
pEC50
n
IFNα
9.35
6.5
24
19
14
12
9
PBMC
TNFα
IFNα
whole blood
whole blood
whole blood
whole blood
whole blood
whole blood
9.05
5.0
TNFα
IL-1β
5.3
IFNγ
5.9
13
8
IL-10
6.6
IL-12p70
7.2
8
Cross-Species Pharmacology. The pattern of TLR
expression in hematopoietic cells and the cytokines produced
by DC subsets are different in rodents and humans.49 However,
murine PDC do produce IFNα in response to TLR7 activation,
and studies with compound 32 were undertaken to confirm
bioactivity in rodents prior to testing in vivo.
Mouse: potency of compound 32 for murine IFNα induction
was determined by flow cytometric analysis of intracellular
IFNα in pDC from murine splenocyte cultures (n = 11)
̈
derived from naıve BALB/c mice. The murine potency was ca.
10-fold lower than that determined in an equivalent assay in
human blood.
Rat: in vitro cultures of rat PBMC or whole blood with
compound 32 resulted in variable and minimal cytokine
responses (IL-1β, TNFα, IL-6, and IP-10) indicating the rat to
be a much poorer responder than the mouse. However, in vivo
dose-related increases in IFN bioactivity were demonstrated in
rats treated with 32 at 0.4 mg/kg and above, with the maximal
response seen 2 h after dosing.
In Vitro Immunomodulation. To gain evidence that the
immunomodulatory cytokines induced by compound 32 might
deliver benefit in the treatment of allergic diseases, the impact
Monkey: data in the literature indicates that TLR patterns of
expression in cynomolgus monkeys and their responses to TLR
activators are similar to those in humans.50 Comparison of
H
DOI: 10.1021/acs.jmedchem.5b01647
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
Figure 6. Impact of compound 32 on allergen driven cytokine responses in human PBMC.
Figure 7. Serum IFNα and IP-10 responses following i.n. dosing of compound 32 in BALB/c mice.
of the compound on the cytokine response to allergen using
PBMC from allergic individuals was investigated.51 Following
incubation for 6 days, 32 reduced levels of the Th2 cytokines
IL-5 and IL-13 released in response to Timothy grass (n = 9) or
house dust mite (n = 7) in PBMC cultures derived from
individuals allergic to these allergens, in a dose-dependent
manner (Figure 6). Furthermore, 32 gave rise to enhanced
production of IL-10 and IFNγ, indicating that, in the presence
of allergen, this candidate has the potential to modify the nature
of the T-cell responses.
In Vivo PD Studies. In vivo evaluation of the pharmacody-
namic (PD) response to compound 32 was first investigated in
the mouse. Female BALB/c mice were dosed intranasally (i.n.)
with 32 in 0.2% Tween 80/saline and IFNα and IP-10 (IFNγ-
inducible protein-10) levels monitored in serum and nasal
lavage, 2 and 6 h after dosing. IP-10 is a chemokine commonly
used as a downstream biomarker of IFN induction. Dose-
related increases in IFNα levels in serum were detected at doses
of 0.3 mg/kg and above at the 2 h time point which had
subsided at 6 h (Figure 7). IP-10 provided a more sensitive
biomarker confirming a clear serum response at 2 h at the 0.1
mg/kg dose level. IP-10, but not IFNα, could also be detected
in nasal lavage samples, with maximum levels being present at 6
h post-treatment, but these levels were ∼200-fold lower than
those seen in serum (data not shown) and were considered
insufficient to confirm target engagement locally in the nose.
Intranasal dosing of compound 32 in the cynomolgus
monkey provided similar evidence for TLR7 activation but at
very much lower doses than in the mouse. Plasma IP-10
measured 6 h after dosing provided the most sensitive
biomarker of target engagement with raised levels of this
chemokine detected at doses of 30 ng/kg and above (Figure 8).
No TNFα was detected, even at the 3000 ng/kg dose level,
reflecting the IFNα/TNFα selectivity seen in vitro.
Figure 8. Plasma IFNα responses 6 h after i.n. dosing of 32 in
cynomolgus monkeys.
scriptomic analysis. Thus, nasal scrapes were collected 6 and 24
h after intranasal dosing with 32 for TaqMan gene expression
analysis using a panel of 18 interferon stimulated genes (ISG).
Dose- and time-dependent increases in mRNA mean fold
expression were observed for ISG; Table 7 shows representa-
tive data from genes showing the largest fold increases, relative
to the vehicle mean. This data provided clear evidence for
engagement of TLR7 locally in the nose at a dose level of 3 ng/
kg. With the exception of minimal MCP-1 induction in a single
monkey, compound 32 did not induce any increase in plasma
cytokines at this dose level indicating that at low doses it is
possible to activate TLR7 in the nasal tissue without inducing a
systemic cytokine response.
Data collected from the cardiovascular and respiratory safety
study in the cynomolgus monkey showed that doses ≥30 ng/kg
were associated with mild, dose-dependent increases in body
temperature which returned to baseline within 24 h.52 This
transient fever response is entirely consistent with the TLR7
induced cytokine response and was subsequently also observed
during the clinical evaluation of 32 (vide infra). On the basis of
this in vitro and in vivo data, the cynomolgus monkey was
selected as the most relevant species from which to determine
In addition, target engagement in nasal tissue following
intranasal dosing in the monkey was investigated by tran-
I
DOI: 10.1021/acs.jmedchem.5b01647
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
and safety pharmacology studies have also been conducted.
Additional dermal and ocular irritancy studies were also
performed in CD-1 mice and cynomolgus monkeys,
respectively, using the intranasal formulation.
Table 7. TaqMan Gene Expression Analysis of IFN
Responsive Genes in Nasal Scrapes of Monkeys 24 h after
a
Intranasal Dosing with Vehicle or Compound 32
treatment group
Species selection was based on comparative in vitro
pharmacology studies looking at cytokine profiles on incubation
of 32 with PBMC from different species. As reported in the
literature, and shown above in the Preclinical Profiling:
Immunology section, the cynomolgus monkey was found to
most closely mirror the IFNα and TNFα responses seen in
human PBMC. The lack of suitable cytokine assays prevented
evaluation of the dog as a nonrodent species, and the mini-pig
was ruled out due to poor IFNα/TNFα selectivity (data not
shown). The rabbit is known not to respond to TLR7
agonists.23 The rat and monkey were therefore selected as the
most appropriate preclinical species for toxicity testing; the rat
was selected because as a poor responder, it would allow higher
systemic exposures to be achieved without initiating potentially
dose limiting pharmacologically mediated effects.
IFN responsive gene
vehicle
3 ng/kg
30 ng/kg
CCL8
GBP1
IFIT2
CXCL10
OAS1
OAS2
OAS3
1.0
2.4
2.1
1.6
1.2
1.0
0.9
3,972
68.5
19,315
97.6
108.5
230
123.8
638
38.9
53.7
128.2
86.3
179.9
101.2
a
Values represent mean fold change of gene expression at 24 h (n = 3/
group) vs average vehicle treated expression. CCL, chemokine (C−C
motif) ligand; GBP, interferon-induced guanylate-binding protein;
IFIT, interferon-induced protein with tetratricopeptide repeats;
CXCL, chemokine (C-X-C motif) ligand; OAS, 2′-5′-oligoadenylate
synthetase.
The main challenges faced during the toxicological assess-
ment of compound 32 were ensuring sufficient containment of
the extraordinarily potent immunostimulatory agonist and the
known tolerization of the response when repeat doses were
administered too frequently. High containment practices were
implemented to protect operators dispensing and delivering the
intranasal spray formulation. Tolerization was investigated in
the early dose ranging studies and was determined by
comparison of once, twice, or three-times weekly dosing with
daily dosing in the mouse and comparison of once daily and
weekly dosing in the monkey. The optimal dosing regimen
required to maintain pharmacological activity (cytokine
response) was weekly dosing because tolerization was rapid
with daily dosing resulting in a loss of cytokine response after
only 3 days of administration.
the minimal anticipated biological effect level (MABEL). This
was set at 3 ng/kg based on the detection of minimal systemic
levels of the most sensitive and reproducible PD biomarker IP-
10, and this figure was in turn used to calculate the starting
intranasal dose of 2 ng for the first time in human (FTIH)
study.
PRECLINICAL PROFILING: DMPK
■
Compound 32 shows desirable DMPK characteristics for a
topically active intranasal drug with high systemic clearance
(mouse 107, rat 96 mL/min/kg) and negligible oral
bioavailability (rat, 1%). Very low levels of 32 were detected
in the hepatic portal vein following oral administration to rats,
indicating negligible passage of intact parent across the
gastrointestinal tract and probably resulting largely from
complete ionization of the basic piperidine amine (pKa 9.7)
in the gut. In contrast, bioavailability following intranasal
administration of a solution of compound 32 is estimated to be
42% in the mouse, 12% in the rat, and predicted to be 100% in
the cynomolgus monkey. Furthermore, an early Tmax of ≤5 min
(first sampling occasion) after intranasal dosing in rats and
cynomolgus monkeys indicates rapid absorption across nasal
tissue where the target plasmacytoid dendritic cells are located.
Compound 32 was highly turned over by mouse, rat, monkey,
and human liver microsomes with a good in vitro/in vivo
correlation for mouse and rat, indicating that in vivo clearance is
likely to be due mainly to metabolism. Following incubation in
vitro with hepatocytes from rats, humans, and monkeys, no
major qualitative differences were observed in the routes of
metabolism of 32 between the species. The metabolism was
complex with routes identified including hydroxylation,
dehydrogenation, and a combination of hydroxylation and/or
dehydrogenation of the pentane side chain and piperidine
moiety, N,N-depentylation, N-depiperinidation with subse-
quent oxidation of the hexyl chain, glucuronidation, oxidations,
and N-acetylation.
Genotoxicity assessments indicated that 32 does not present
a hazard to humans, and there were no safety pharmacology
findings of concern. This toxicological assessment is predom-
inantly based on the results of the definitive 8 week intranasal
toxicity studies in the rat and monkey.
The principal local effect noted in rats and monkeys after
repeat intranasal dosing was nasal irritancy, believed to be
mechanism related and driven by local cytokine induction; this
was only apparent at very high dose levels/concentrations in
the rat and was the primary finding in the determination of the
no observed adverse effect level (NOAEL) in both species.
These findings comprised slight to marked inflammation and
minimal to moderate exudate, erosion/ulceration of squamous
and respiratory epithelium, squamous epithelium regeneration/
hyperplasia with and without keratinization, squamous
metaplasia of respiratory epithelium, and regeneration/hyper-
plasia of respiratory epithelium in the monkey. The NOAEL in
the rat was 453/772 μg/kg/day for 14 days in male and female,
respectively, and 406/698 μg/kg/week for 8 weeks in male and
female, respectively. The NOAEL in the monkey was very
much lower at 35.9 ng/kg/week for 8 weeks reflecting the
much greater sensitivity of the monkey compared to that in the
rat. These NOAELs provided adequate cover with regard to
nasal exposure over the whole range of proposed doses in the
clinic.
PRECLINICAL PROFILING: SAFETY ASSESSMENT
■
Compound 32 (maleate salt) has been evaluated in nonclinical
studies for up to 8 weeks of weekly intranasal dosing in
Sprague−Dawley rats and cynomolgus monkeys and up to 14
days of daily oral dosing in CD-1 mice, Sprague−Dawley rats,
and New Zealand White rabbits. Standard genotoxicity assays
The principal systemic effect was increased levels of specific
cytokines (IFNα, IP-10, IL-6, IL-IR antagonist, and MCP-1, G-
CSF, and CRP); all other effects were considered to be
secondary to this primary response (i.e., increased body
J
DOI: 10.1021/acs.jmedchem.5b01647
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
temperature, increased heart rate, alterations in splenic and
lymph node cellularity, and minor changes in some clinical
chemistry parameters). With the exception of the splenic and
lymph node effects, all of the observed changes would be
readily monitored in humans. The effects that were seen in the
spleen and lymph node were of low severity and were shown to
be fully reversible on withdrawal from treatment in the monkey.
The observed cytokine response is consistent with the known
TLR7 pharmacology of compound 32.
Similar pharmacologically mediated effects were seen after
oral administration in rats and mice; these mainly comprised
altered cellularity in selected lymphoid tissues (spleen,
mesenteric and mandibular lymph nodes, and thymus). TLR7
agonism leads to changes in lymphoid cell populations, either
directly or due to downstream effects of cytokine production.
There was no evidence of dermal irritancy in the mouse after
a single 4 h exposure to the clinical intranasal formulation at
concentrations up to 1000 μg/mL; the clinical intranasal
formulation is therefore classified as nonirritating according to
the Draize scheme. The mouse has previously been shown to
be predictive for adverse skin reactions resulting from exposure
to other TLR7 agonists.53 The risk of adverse skin reactions in
the clinic from accidental skin exposure to relevant
concentrations of the clinical intranasal formulation was
therefore considered to be minimal.
There were a wide variety of adverse ocular effects in the
monkey after a single exposure to the clinical intranasal
formulation at a concentration of 1000 μg/mL. These included
swollen and red skin (peri-orbital region), corneal opacity/
edema/staining, pupil dilation/pupil miotic, conjunctival
congestion/swelling/discharge, and hyperaemic retina. The
onset of the majority of these effects were delayed until at least
48 h after dosing, although some changes could be detected by
ophthalmic examination at 6 h after dosing (aqueous cells and
aqueous flare). A lower dose of 10 μg/mL was well tolerated.
There is, therefore, a theoretical possibility of concentration-
dependent ocular effects after accidental exposure in humans;
however, any ocular symptoms can be easily reported and
monitored in the clinic.
which can be enhanced to give sufficient solubility (200 μg/mL
at pH 7) to cover clinical doses by inclusion of the surfactant
Tween 80. However, solutions of the free base were found to
show unacceptable instability, believed to be due to N-oxide
formation. This problem was overcome by in situ formation of
the maleate salt for early studies. Subsequently, preparation of
32 as the maleate salt revealed this version to be a highly
soluble compound with sufficient solubility (11 mg/mL at
pH7) to cover the highest doses in preclinical safety
assessment. Furthermore, formulations prepared with the
maleate salt have delivered a pharmaceutical product displaying
good stability (≥18 months at temperatures up to 30 °C) and
which gives very reproducible delivery across the dose range
from a device fitted with a Valois VP7 Pump. The very low
strength formulations presented significant analytical chal-
lenges, but these were overcome by the development of a
suitably sensitive assay (LLQ ca.1 ng/mL) for 32, which has
allowed the stability of even the lowest strength formulations to
be successfully monitored.
CLINICAL STUDIES
■
The FTIH study was a randomized, double-blind, placebo-
controlled study exploring the safety, tolerability, and
pharmacokinetics of single escalating intranasal doses of 32 in
healthy volunteers (HVT) and patients with active pollen-
driven AR (NCT01480271).52 In addition, target engagement
was explored by monitoring the induction of TLR7-mediated
biomarkers both locally in the nose (nasal scrapes and/or nasal
lavage) and also in peripheral blood. The starting dose for this
study was set at 2 ng using a MABEL approach based on
extrapolation from safety and PD findings in the monkey and
applying a safety factor of 10.52 Doses of 2, 20, 40, 60, and 100
ng were explored in HVT and doses of 2, 6, 20, and 40 ng in
AR patients (further doses were not evaluated due to the end of
the pollen season). No serious adverse events (AEs) occurred
in either group. Mild to moderate and transient “flu-
like”symptoms potentially related to TLR7-mediated cytokine
release syndrome (CRS) were seen in HVT at doses of ≥40 ng
with a dose-dependent increase incidence. Doses higher than
100 ng were not tested as the CRS-related AEs induced at this
level were not considered acceptable for the target population
and met the predefined stopping criteria.52 In AR patients, only
few sporadic mild CRS-related AEs were observed without a
clear dose relationship. Activation of the TLR7 pathway was
confirmed in both populations at doses ≥20 ng, by a dose-
dependent increase of serum and nasal lavage IP-10 levels and
upregulation of IFNα-stimulated genes in nasal scrapes. IP-10
monitoring in nasal lavage provides a very convenient and
sensitive method of confirming TLR7 engagement locally in the
nose and has been successfully employed to assess target
engagement in subsequent studies.
There were therefore no effects to preclude the intranasal
administration of 32 to humans in proposed clinical trials with
appropriate monitoring of nasal inflammation and/or dis-
comfort.
PRECLINICAL PROFILING: PHARMACEUTICAL
FORMULATION
■
While suspension formulations can be successfully delivered
using conventional nasal spray devices, an aqueous solution
formulation is very much preferred. This requires the
compound to be sufficiently soluble and stable in water (with
potential excipients to aid solubility or stability) in order to
provide a pharmaceutical product with a suitable shelf life. In
addition, consistent delivery of the drug from the device on
repeated actuation and over time is required, which can be
compromised by any adsorption of the drug to components of
the device; this becomes a more significant issue for a very low
dose product. The early clinical evaluation of 32 required an
extraordinarily low strength formulation of 0.01 μg/mL (100
μL per nostril) to deliver the starting dose of 2 ng, while the
highest dose studied in the monkey intranasal safety study (320
μg/kg) required a formulation of 10 mg/mL. The free base,
compound 32, is an off-white crystalline solid (melting point
207 °C) with a relatively low aqueous solubility (50 μg/mL),
The systemic cytokine response is believed to result from
spillover of locally produced cytokines since, with the exception
of one sample, no quantifiable systemic exposure to 32 was
detected despite a very sensitive assay (LLQ 2 pg/mL). The
observed “flu-like” symptoms are therefore also considered to
be driven by the local TLR7 engagement, which is consistent
with the observation by AstraZeneca of similar AEs with their
intranasal TLR7 agonist “ante-drug” designed to be rapidly
inactivated in plasma.10b
Having identified a well-tolerated single dose of compound
32 that activates the TLR7 cascade in the nose, the safety and
pharmacodynamics of repeat intranasal administration was next
K
DOI: 10.1021/acs.jmedchem.5b01647
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
Figure 9. 4-Week repeat dose study: IP-10 levels in nasal lavage pre (0 h)- and 24 h post-intranasal administration of 40 ng of compound 32 on
dosing visits 1−4.
investigated. Repeated stimulation of TLRs is known to be
associated with tachyphylaxis, and our preclinical evaluation had
shown that dosing with 32 more frequently than once-weekly
resulted in a marked reduction of the pharmacodynamic
response. In contrast, the opposite phenomenon was observed
in a clinical study with the oral TLR7 agonist N-(4-(4-amino-2-
e t h y l - 1 H - i m i d a z o [ 4 , 5 - c ] q u i n o l i n - 1 - y l ) b u t y l ) -
methanesulphonamide (PF-4878691), which was terminated
due to the dangerous amplification of cytokine production
following twice weekly repeat dosing.54 To evaluate the effects
of repeat dosing with compound 32, a randomized, double-
blind, placebo-controlled study was conducted in patients with
symptomatic AR and mild asthma during the pollen season
(NCT01607372).52 Compound 32 was administered intra-
nasally at weekly intervals for 4 weeks. Doses of 40 and 80 ng
were studied, and both dose levels were shown to be well
tolerated with only mild “flu-like” symptoms seen in some
subjects (67%, 14%, and 20% of subjects in 40 ng, 80 ng, and
placebo groups, respectively).52 There was no evidence of nasal
irritancy in this study in any of the treatment groups. The
serum and nasal lavage IP-10 levels very clearly increased in a
consistent manner after each weekly dosing with no evidence
for amplification or tolerization of the response. No dose-
dependent effect on IP-10 induction was observed; nasal lavage
data for the placebo and 40 ng cohorts is shown in Figure 9.
Compound 32 has now progressed into an 8 week clinical
study treating subjects with respiratory allergies and investigat-
ing the safety, pharmacodynamics, and efffect on allergic
reactivity in response to allergen challenge 1 and 3 weeks, plus
1 year after treatment (NCT01788813, NCT02446613).
TLR7, and our favored hypothesis is that this IFNα potency is
driven by the concentration of these basic derivatives in the
highly acidic endosomal compartment where TLR7 induced
IRF7 activation to generate IFNα takes place. Incorporation of
PD biomarkers into the preclinical safety studies with
compound 32 proved invaluable in setting the starting dose
and dosing regimen for the clinic as well as establishing
methodologies to assess target engagement following intranasal
dosing in humans. The extreme potency of 32 presented
additional challenges, but the compound shows excellent
pharmaceutical properties that allow reliable delivery of doses
as low as 1 ng per actuation of an intranasal spray formulation.
Early clinical evaluation of compound 32 in healthy volunteers
and patients with allergic rhinitis by the intranasal route has
confirmed local activation of TLR7 at very low, well tolerated,
doses, and a reproducible pharmacological response has been
confirmed following repeat dosing at weekly intervals. An 8
week clinical study with compound 32 is nearing completion,
which will provide the first indications of the impact of repeated
TLR7 activation in the nose on allergic reactivity and duration
of effect. The results of this study will be presented in due
course.
EXPERIMENTAL SECTION
■
TLR7 Agonists. The highly potent immunostimulatory TLR7
agonists described herein should be handled with appropriate
containment controls particularly when manipulating powder samples
which may become easily airborne.
Biology. The PBMC assay A measured IFNα induction following
incubation of test compounds for 24 h with cryopreserved human
PBMC. The PBMC assay B measured both IFNα and TNFα
induction following incubation of test compounds for 24 h with freshly
prepared human PBMC. Full details of these primary screening assays
along with methods used for additional in vitro and in vivo immunology
profiling of compound 32 can be found in the Supporting Information.
All in vivo studies were conducted in accordance with the GSK Policy
on the Care, Welfare and Treatment of Laboratory Animals and were
reviewed by the Institutional Animal Care and Use Committee at
GSK, and/or by the ethical review process at the institution where the
work was performed.
Chemistry. TLC was performed on Merck 0.25 mm Kieselgel 60
F254 plates using UV light and/or staining with aqueous KMnO4
solution for visualization. LCMS analysis of reactions and products was
conducted using one of three methods (A−C) which are described in
detail in the Supporting Information along with conditions used for
mass-directed autopreparative HPLC (MDAP). Chromatographic
CONCLUSIONS
■
Optimization of both the C-2 and N-9 substituents on the 8-
oxoadenine scaffold had provided the intranasal clinical
candidate 32, which displays extraordinary IFNα inducing
potency and IFNα/TNFα selectivity. 2-(S)-Pentyloxy was
identified as a novel potency enhancing 2-substituent, but the
major breakthrough came with the introduction of basic amine
functionality in the N-9 substituent. Thus, switching from
saturated oxygen heterocycles to saturated nitrogen hetero-
cycles in the N-9 side chain was accompanied by a dramatic and
selective enhancement of IFNα potency. SAR of the N-9
substituent strongly suggests that the origin of this potency
enhancement is not the result of specific interactions with
L
DOI: 10.1021/acs.jmedchem.5b01647
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
purification was typically performed using prepacked silica gel
cartridges on a flash chromatography system such as the Flashmaster
II system (Argonaut Technologies Ltd.). Solvent removal using a
stream of nitrogen was performed at 30−40 °C on a GreenHouse
Blowdown system available from Radleys Discovery Technologies.
with DCM (50 mL). The combined organic layers were dried through
a hydrophobic frit, and the solvent removed in vacuo to yield the title
compound as an orange foam (18.5 g, 99%). LCMS (System C): tRET
1
= 3.06 min; MH+ 384/386. H NMR δ (chloroform-d) 5.61 (dd, J =
11.0, 2.5 Hz, 1H), 5.36 (br s, 2H), 5.07−5.17 (m, 1H), 4.12−4.20 (m,
1H), 3.65−3.74 (m, 1H), 2.95−3.10 (m, 1H), 2.05−2.14 (m, 1H),
1.58 (s, 8H+ (integral obscured by water), 1.32−1.37 (m, 3H), 0.91−
0.97 (m, 3H) [mixture of diastereoisomers; signals are grouped
together and treated as a single component for integration].
1
Unless otherwise stated, H NMR spectra were recorded at 400 MHz
in either chloroform-d or DMSO-d6 on either a Bruker DPX 400 or
Bruker Avance DRX or Varian Unity 400 spectrometer. Spectra at 600
MHz were recorded using a Bruker AVII+ spectrometer. The internal
standard used was either tetramethylsilane or the residual protonated
solvent at 7.25 ppm for chloroform-d or 2.50 ppm for DMSO-d6. The
purity of all compounds screened in biological assays was assessed by
LCMS and NMR analysis and was ≥95% as unless otherwise specified.
Synthesis of Compound 32. 2-Fluoro-9-(tetrahydro-2H-pyran-
2-yl)-9H-purin-6-amine (38). N,O-Bis(trimethylsilyl)acetamide (975
mL, 3.988 mol) was added to a stirred suspension of 2-fluoro-9H-
purin-6-amine (37) (200 g, 1.306 mol) in anhydrous acetonitrile (4 L)
in a 10 L controlled laboratory reactor and the resulting mixture
heated to reflux and maintained at that temperature for 2 h. The
circulator was then reprogrammed and the reaction mixture cooled to
0 °C. A solution of tetrahydropyran-2-yl acetate (preparation
described in Tetrahedron Lett. 2006, 47, 4741−4741) (282 g, 1.959
mol) in anhydrous acetonitrile (500 mL) was then added slowly via a
dropping funnel followed by trimethylsilyl trifluoromethanesulfonate
(283 mL, 1.567 mol) dropwise via a dropping funnel. No significant
exotherm was observed. The circulator temperature was readjusted to
10 °C and stirring maintained for a further 1 h. The mixture was then
quenched by the addition of 1 M sodium carbonate (4 L). A solid
precipitate was observed and the pH checked to be basic. Additional
water was added to the suspension (1 L), and on standing, the layers
separated with the aqueous layer containing significant solid
inorganics. The majority of the aqueous and inorganic solid was
separated. The organic layer still contained significant solid and was
cooled to 0 °C with stirring to encourage further precipitation. This
solid was then collected by filtration, and the pad was washed very well
with water then dried in vacuo at 40 °C overnight to give the title
compound as a cream colored solid (152.8 g, 49.4%). LCMS (System
2-{[(1S)-1-Methylbutyl]oxy}-8-(methyloxy)-9-(tetrahydro-2H-
pyran-2-yl)-9H-purin-6-amine (41). Compound 40 (7.1 g, 18.48
mmol) was dissolved in anhydrous methanol (70 mL), and a solution
of sodium methoxide (25%) in methanol (8 mL) was added dropwise
under an atmosphere of nitrogen. The solution was heated to reflux at
90 °C for 4 h under an atmosphere of nitrogen. Additional sodium
methoxide in methanol (25% solution, 3 mL) was added, and the
reaction was stirred at 60 °C for a further 16 h. An additional portion
of sodium methoxide in methanol (25% solution, 5 mL) was added,
and the reaction was stirred at 90 °C for a further 7 h. The solvent was
removed on a rotary evaporator, and the crude product was partitioned
between ethyl acetate (75 mL) and saturated ammonium chloride
solution (75 mL). The organic layer was washed with brine (75 mL).
The solvent was removed on a rotary evaporator to yield the title
compound as a pale orange foam (6 g, 97%). LCMS (System B): tRET
1
= 1.14 min; MH+ 336, 337. H NMR δ (chloroform-d) 5.50 (dd, J =
11.3, 2.3 Hz, 1H), 5.06−5.18 (m, 3H), 4.07−4.15 (m, 4H), 3.64−3.73
(m, 1H), 2.68−2.83 (m, 1H), 1.98−2.08 (m, 1H), 1.64 (br s, 8H+
(integral obscured by water)), 1.30−1.36 (m, 3H), 0.89−0.97 (m, 3H)
[mixture of diastereoisomers; signals are grouped together and treated
as a single component for integration].
2-{[(1S)-1-Methylbutyl]oxy}-8-(methyloxy)-9H-purin-6-amine Tri-
fluoroacetate Salt (42). Compound 41 (6 g, 17.89 mmol) was
dissolved in methanol (50 mL). Trifluoroacetic acid (20.67 mL, 268
mmol) was added dropwise and the mixture stirred at 20 °C for 72 h
under an atmosphere of nitrogen. The solvent was removed in vacuo,
and the resulting solid was washed with ethyl acetate and filtered. The
filtrate was stripped and the residue washed with ethyl acetate. The
combined solid residues were dried in a vacuum oven for 2 h to give
the title compound as an off white solid (5.3 g, 81%). LCMS (System
B): tRET = 0.76 min; MH+ 252, 253. 1H NMR δ (DMSO-d6) 7.63 (br
s, 2H), 5.07−5.15 (m, 1H), 4.05 (s, 3H), 1.51−1.75 (m, 2H), 1.32−
1.47 (m, 2H), 1.30 (d, J = 6.0 Hz, 3H), 0.90 (t, J = 7.5 Hz, 3H).
9-(5-Chloropentyl)-2-{[(1S)-1-methylbutyl]oxy}-8-(methyloxy)-
9H-purin-6-amine (43). Compound 42 (4.430 g, 12.13 mmol) was
dissolved in DMF (70 mL). Potassium carbonate (4.30 g, 31.1 mmol)
was added, and the reaction mixture was heated and stirred at 50 °C
for 2 h and then cooled to room temperature. 1-Bromo-5-
chloropentane (1.598 mL, 12.13 mmol) was added and the mixture
stirred at room temperature for 48 h. The solvent was then removed,
and the residue was partitioned between DCM (100 mL) and 6%
NaHCO3 solution (100 mL), and the layers were separated. The
aqueous layer was washed with additional DCM (2 × 50 mL), and the
organic extracts were combined, dried by passage through a
hydrophobic frit, and the solvent removed. The residue (ca. 6 g)
was loaded in dichloromethane and purified by aminopropyl SPE (70
g,three cartridges) using a 0−100% ethyl acetate/cyclohexane gradient.
The appropriate fractions were combined and evaporated in vacuo to
give the title compound as a yellow oil (3.037 g, 70.4%). LCMS
(System A): tRET = 2.58 min; MH+ = 356, 358. 1H NMR δ
(chloroform-d) 5.06−5.21 (m, 3H), 4.11 (s, 3H), 3.93 (t, J = 7.0 Hz,
2H), 3.51 (t, J = 7.0 Hz, 2H), 1.73−1.86 (m, 5H), 1.37−1.61 (m, 5H),
1.32 (d, J = 6.0 Hz, 3H), 0.93 (t, J = 7.5 Hz, 3H).
1
C): tRET = 1.71 min; MH+ = 238. H NMR δ (chloroform-d) 7.98 (s,
1H), 5.90 (br s, 2H), 5.59−5.66 (m, 1H), 4.13−4.21 (m, 1H), 3.71−
3.81 (m, 1H), 1.91−2.17 (m, 3H), 1.60−1.84 (m, 3H+ (integral
obscured by solvent)).
2-{[(1S)-1-Methylbutyl]oxy}-9-(tetrahydro-2H-pyran-2-yl)-9H-
purin-6-amine (39). Sodium t-butoxide (206 g, 2.144 mol) was added
to (S)-2-pentanol (720 mL, 6.58 mol) in a 2 L round bottomed flask.
The mixture was stirred at 50 °C until all of the sodium t-butoxide had
dissolved. Compound 38 (130 g, 548 mmol) was then added in
portions over 5 min. After 3 h, LCMS analysis indicated complete
consumption of the starting material, and the mixture was poured into
ice/water (3 L) and then extracted with methyl t-butyl ether. This
resulted in emulsion formation, the mixture was filtered through
Celite, and the organic phase was separated. The aqueous layer was
then treated with solid NaCl and then re-extracted with methyl t-butyl
ether. The organic extracts were combined and washed with brine,
dried over anhydrous magnesium sulfate, filtered, and then evaporated
to yield the title compound as a pale brown gum (158.59 g, 95%).
LCMS (System C): tRET = 2.65 min; MH+ 306. 1H NMR δ
(chloroform-d) 7.84 (s, 1H), 5.61−5.66 (m, 1H), 5.55 (br s, 2H), 5.18
(sxt, J = 6.2 Hz, 1H), 4.10−4.17 (m, 1H), 3.69−3.78 (m, 1H), 1.37−
2.13 (m, 10H+ (integral obscured by impurities)), 1.31−1.36 (m, 3H),
0.94 (t, J = 7.3 Hz, 3H) [mixture of diastereoisomers; signals grouped
together and treated as a single component for integration].
8-Bromo-2-{[(1S)-1-methylbutyl]oxy}-9-(tetrahydro-2H-pyran-2-
yl)-9H-purin-6-amine (40). N-Bromosuccinimide (12.16 g, 68.3
mmol) was added portionwise over 5 min to a stirred solution of
compound 39 (14.9 g, 48.8 mmol) in chloroform (80 mL) at <5 °C
under an atmosphere of nitrogen. The reaction mixture was stirred at
<5 °C for 5 h then washed with saturated sodium hydrogen carbonate
solution (80 mL) then water (80 mL) and evaporated. The resulting
foam was dissolved in DCM (50 mL) and washed with water (50 mL)
and then brine (50 mL). The combined aqueous phases were washed
2-{[(1S)-1-Methylbutyl]oxy}-8-(methyloxy)-9-[5-(1-piperidinyl)-
pentyl]-9H-purin-6-amine (44). Compound 43 (80 mg, 0.225 mmol),
triethylamine (0.031 mL, 0.225 mmol), and piperidine (0.045 mL,
0.45 mmol) were suspended in DMF (3 mL) and the mixture heated
to 70 °C for 18 h. The solvent was removed and the residue
partitioned between DCM (4 mL) and saturated sodium bicarbonate
(4 mL). The aqueous phase was re-extracted with further DCM, and
the combined organic extracts were concentrated and the residue
M
DOI: 10.1021/acs.jmedchem.5b01647
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
dissolved in 1:1 MeOH/DMSO (1 mL) and purified by MDAP. The
product containing fractions were combined and evaporated under a
stream of nitrogen to give the title compound (47.2 mg, 51.9%).
LCMS (System C): tRET = 3.11 min; MH+ = 405. 1H NMR δ
(chloroform-d) 5.08−5.17 (m, 3H), 4.10 (s, 3H), 3.91 (t, J = 7.3 Hz,
2H), 2.28−2.40 (m, 4H), 2.20−2.27 (m, 2H), 1.71−1.81 (m, 4H),
1.37−1.61 (m, 10H), 1.32 (d, J = 6.0 Hz, 3H), 1.24−1.31 (m, 2H),
0.93 (t, J = 7.3 Hz, 3H).
SMILES data (CSV)
AUTHOR INFORMATION
Corresponding Author
*Phone: +44 1438 763473. Fax: +44 1438 762807. E-mail:
diane.m.coe@gsk.com.
■
Notes
6-Amino-2-{[(1S)-1-methylbutyl]oxy}-9-[5-(1-piperidinyl)pentyl]-
7,9-dihydro-8H-purin-8-one (32). Compound 44 (2.5 g, 6.18 mmol)
was dissolved in methanol (25 mL). 4 M Hydrogen chloride in 1,4-
dioxane (38.6 mL, 154 mmol) was added, and the mixture was stirred
at room temperature for 4 h. The solvent was evaporated in vacuo to
leave an off-white solid (2.555 g). Then, 2.455g of this material was
partitioned between DCM (50 mL) and saturated NaHCO3 solution
(50 mL). The layers were separated, and the aqueous phase was
washed with additional DCM (3 × 25 mL). The organic extracts were
combined, dried (hydrophobic frit), and the solvent evaporated in
vacuo to give the title compound as an off-white solid (1.756 g, 72.8%).
LCMS (System A): tRET = 1.42 min; MH+ = 391. 1.7 g of this material
was recrystallized from ethyl acetate (ca. 50 mL). The crystals were
collected, washed with ice-cold ethyl acetate (15 mL), and dried in
vacuo at 50 °C for 3 h to give the title compound as a cream crystalline
solid (1.33 g). Melting point onset (DSC): 207.4 °C. LCMS (System
The authors declare the following competing financial
interest(s): All authors were employees of GSK at the time
the work was carried out. Keith Biggadike, Diane Coe, Bob
Gibbon, Stephen Hermitage, Amanda Lucas, Charlotte Hardy,
Stephen Smith, Naimisha Trivedi, and Xiao Lewell are co-
inventors on patents covering compound 32 or its maleate salt.
ACKNOWLEDGMENTS
■
We thank the following: Heather Barnett, Ian Campbell,
Sebastian Campos, Robin Daly, Jordi Juarez, Stefano Livia,
Andy Mason, Simon Partridge, Alastair Roberts, Colin
Theobald, and Elisabet Viana-Gaza for their synthetic
contributions; Lewis Entwistle for conducting in vivo studies;
Philip Whitehead, Maggie Lynch, and Siwan Oldham for
Project Management; and Alex Povey from Patents. Com-
pound 32 is owned by GSK, and the research outlined in this
article was funded by GSK.
1
A): tRET = 1.42 min; MH+ = 391. H NMR δ (DMSO-d6, 600 MHz)
9.80 (s, 1H), 6.33 (br s, 2H), 4.99 (sxt, J = 6.2 Hz, 1H), 3.64 (t, J = 7.0
Hz, 2H), 2.23 (br s, 4H), 2.15 (t, J = 7.3 Hz, 2H), 1.57−1.68 (m, 3H),
1.27−1.52 (m, 11H), 1.19−1.25 (m, 5H), 0.88 (t, J = 7.3 Hz, 3H). 13C
NMR δ (DMSO-d6, 151 MHz) 159.7, 152.3, 149.4, 147.6, 97.9, 71.1,
58.4, 54.0, 38.9, 37.8, 27.6, 25.8, 25.5, 24.2, 23.9, 19.8, 18.2, 13.9.
Images of the DSC trace and also of the XRPD can be found in
WO2010/018133.37a
ABBREVIATIONS USED
■
AP-3, adaptor protein-3; AR, allergic rhinitis; CRS, cytokine
release syndrome; FTIH, first time in human; HVT, healthy
volunteer; IFNα, interferon-α; IP-10, IFNγ-inducible protein-
10; IRF-7, interferon regulatory factor-7; ISG, interferon
stimulated genes; LLQ, lower limit of quantification; LRO,
lysosome related organelle; MABEL, minimum anticipated
biological effect level; MSD, meso scale discovery; MyD88,
myeloid differentiation primary response protein 88; PAMP,
pathogen-associated molecular pattern; PBMC, peripheral
blood mononuclear cell; pDC, plasmacytoid dendritic cell;
SIT, specific immunotherapy; Th1, type-1 T-helper cell; Th2,
type-2 T-helper cell; ssRNA, single stranded RNA
6-Amino-2-{[(1S)-1-methylbutyl]oxy}-9-[5-(1-piperidinyl)-
pentyl]-7,9-dihydro-8H-purin-8-one, Maleate Salt (32 Maleate
Salt). Preparation 1. Compound 32 (0.384 g, 0.98 mmol) was
dissolved in isopropyl alcohol (4.6 mL) and heated to 40 °C. Maleic
acid (0.114 g, 0.98 mmol) was added. A clear solution was obtained.
During cooling to room temperature, precipitation occurred. The
slurry was filtered, washing with isopropyl alcohol (5 mL), and dried
under reduced pressure at 40 °C to constant weight to give the title
1
compound as a white solid (0.305 g, 61%). H NMR confirms a 1:1
ratio of maleic acid: 32 1H NMR δ (DMSO-d6) 9.85 (s, 1H), 8.85 (br
s,1H), 6.39 (s, 2H), 6.02 (s, 2H), 5.00 (m, J = 6.2 Hz, 1H), 3.68 (t, J =
6.8 Hz, 2H), 3.40 (m, 2H), 2.98 (m, J = 8.1 Hz, 2H), 2.82 (br s, 2H),
1.85−1.24 (m, 16H), 1.21 (d, J = 6.1 Hz, 3H), 0.89 (t, J = 7.3 Hz, 3H).
Preparation 2. A solution of compound 32 (1.46 g, 3.74 mmol) in
isopropyl alcohol (14.6 mL) was clarified (filtered at room
temperature through a BondElut cartridge) and then heated to
approximately 50 °C. A solution of maleic acid (0.434 g, 3.74 mmol)
in isopropyl alcohol (2.9 mL) was added. The resulting solution was
then seeded and cooled to 45 °C. Further seed was added. The
resulting slurry was cooled to room temperature and held overnight
(approximately 16 h), then cooled in an ice/water bath for 30 min.
The slurry was filtered, washing with isopropyl alcohol (4.5 mL and
then 3 mL). The product was dried under reduced pressure at 40 °C
to constant weight to give the title compound (1.305 g, 69%). Analysis
by XRPD indicated this sample to be crystalline. Images of the XRPD
can be found in WO2011/098452.37b
REFERENCES
■
(1) Croisant, S. Epidemiology of asthma: prevalence and burden of
disease. Adv. Exp. Med. Biol. 2014, 795, 17−29.
(2) Holgate, S. T. Asthma: a simple concept but in reality a complex
disease. Eur. J. Clin. Invest. 2011, 41, 1339−1352.
(3) Frew, A. J. Allergen immunotherapy. J. Allergy Clin. Immunol.
2010, 125, S306−313.
(4) Aryan, Z.; Holgate, S. T.; Radzioch, D.; Rezaei, N. A new era of
targeting the ancient gatekeepers of the immune system: Toll-like
agonists in the treatment of allergic rhinitis and asthma. Int. Arch.
Allergy Immunol. 2014, 164, 46−63.
(5) Kawai, T.; Akira, S. The role of pattern-recognition receptors in
innate immunity: update on Toll-like receptors. Nat. Immunol. 2010,
11, 373−384.
ASSOCIATED CONTENT
(6) O’Neill, L. A. J.; Golenbock, D.; Bowie, A. G. The history of Toll-
like receptors − redefining innate immunity. Nat. Rev. Immunol. 2013,
13, 453−460.
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jmed-
chem.5b01647.
(7) Okada, H.; Kuhn, C.; Feillet, H.; Bach, J.-F. The ‘hygiene
hypothesis’ for autoimmune and allergic diseases: an update. Clin. Exp.
Immunol. 2010, 160, 1−9.
(8) Rosewich, M.; Lee, D.; Zielen, S. Pollinex Quatro. An innovative
four injections immunotherapy in allergic rhinitis. Hum. Vaccines
Immunother. 2013, 9, 1523−1531.
Synthesis and characterizing data for compounds 4−31,
33, and 34, LCMS and MDAP methods, details of
human PBMC screening assays, and preclinical in vitro
and in vivo profiling of compound 32 (PDF)
(9) Creticos, P. S.; Schroeder, J. T.; Hamilton, R. G.; Balcer-Whaley,
S. L.; Khattignavong, A. P.; Linblad, R.; Li, H.; Coffman, R.; Seyfert,
N
DOI: 10.1021/acs.jmedchem.5b01647
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
V.; Eiden, J. J.; Brodie, D.; the Immune Tolerance Network Group..
Immunotherapy with a Ragweed-Toll-like receptor 9 agonist vaccine
for allergic rhinitis. N. Engl. J. Med. 2006, 355, 1445−1455.
(26) Lee, J.; Wu, C. C. N.; Lee, K. J.; Chuang, T.-H.; Katakura, K.;
Liu, Y.-T.; Chan, M.; Tawatao, R.; Chung, M.; Shen, C.; Cottam, H.
B.; Lai, M. M. C.; Raz, E.; Carson, D. A. Activation of anti-hepatitis C
virus responses via Toll-like receptor 7. Proc. Natl. Acad. Sci. U. S. A.
2006, 103, 1828−1833.
(10) (a) Structure from National Center for Biotechnology
Information. PubChem Compound Database; CID = 11592228.
https://pubchem.ncbi.nlm.nih.gov/compound/11592228 (accessed
(27) Horscroft, N. J.; Pryde, D. C.; Bright, H. Antiviral applications of
Toll-like receptor agonists. J. Antimicrob. Chemother. 2012, 67, 789−
801.
Dec 30, 2015). (b) Greiff, L.; Cervin, A.; Ahlstrom-Emanuelsson,
̈
C.; Almqvist, G.; Andersson, M.; Dolata, J.; Eriksson, L.; Hogestatt, E.;
̈
̈
(28) Pryde, D. C.; Tran, T.-D.; Jones, P.; Parsons, G. C.; Bish, G.;
Adam, F. M.; Smith, M. C.; Middleton, D. S.; Smith, N. N.; Calo, F.;
Hay, D.; Paradowski, M.; Procter, K. J. W.; Parkinson, T.; Laxton, C.;
Fox, D. N. A.; Horscroft, N. J.; Ciaramella, G.; Jones, H. M.;
Duckworth, J.; Benson, N.; Harrison, A.; Webster, R. The discovery of
a novel prototype small molecule TLR7 agonist for the treatment of
hepatitis C infection. MedChemComm 2011, 2, 185−189.
(29) Roethle, P. A.; McFadden, R. M.; Yang, H.; Hrvatin, P.; Hui, H.;
Graupe, M.; Gallagher, B.; Chao, J.; Hesselgesser, J.; Duatschek, P.;
Zheng, J.; Lu, B.; Tumas, D. B.; Perry, J.; Halcomb, R. L. Identification
and optimization of pteridone Toll-like receptor 7 (TLR7) agonists for
the oral treatment of viral hepatitis. J. Med. Chem. 2013, 56, 7324−
7333.
(30) Kurimoto, A.; Ogino, T.; Ichii, S.; Isobe, Y.; Tobe, M.; Ogita,
H.; Takaku, H.; Sajiki, H.; Hirota, K.; Kawakami, H. Synthesis and
evaluation of 2-substituted 8-hydroxyadenines as potent interferon
inducers with improved oral bioavailabilities. Bioorg. Med. Chem. 2004,
12, 1091−1099.
(31) Vultaggio, A.; Nencini, F.; Fitch, P. M.; Fili, L.; Maggi, L.; Fanti,
P.; deVries, A.; Beccastrini, E.; Palandri, F.; Manuelli, C.; Bani, D.;
Giudizi, M. G.; Guarna, A.; Annunziato, F.; Romagnani, S.; Maggi, E.;
Howie, S. E. M.; Parronchi, P. Modified adenine (9-benzyl-2-butoxy-8-
hydroxyadenine) redirects Th2-mediated murine lung inflammation by
triggering TLR7. J. Immunol. 2009, 182, 880−889.
(32) Isobe, Y.; Kurimoto, A.; Tobe, M.; Hashimoto, K.; Nakamura,
T.; Norimura, K.; Ogita, H.; Takaku, H. Synthesis and biological
evaluation of novel 9-substituted-8-hydroxyadenine derivatives as
potent interferon inducers. J. Med. Chem. 2006, 49, 2088−2095.
(33) During the preparation of this manuscript, investigation of a
homologous series of 4-piperidinylalkyl-8-oxoadenines was reported.
Bazin, H. G.; Li, Y.; Khalaf, J. K.; Mwakwari, S.; Livesay, M. T.; Evans,
J. T.; Johnson, D. A. Structural requirements for TLR7-selective
signaling by 9-(4-piperidinylalkyl)-8-oxoadenine derivatives. Bioorg.
Med. Chem. Lett. 2015, 25, 1318−1323.
(34) Krieg, A. M.; Vollmer, J. Toll-like receptors 7, 8 and 9: linking
innate immunity to autoimmunity. Immunol. Rev. 2007, 220, 251−269.
(35) Brightling, C.; Berry, M.; Amrani, Y. Targeting TNF-α; a novel
therapeutic approach for asthma. J. Allergy Clin. Immunol. 2008, 121,
5−10.
(36) Kurimoto, A.; Hashimoto, K.; Nakamura, T.; Norimura, K.;
Ogita, H.; Takaku, H.; Bonnert, R.; McInally, T.; Wada, H.; Isobe, T.
Synthesis and biological evaluation of 8-oxoadenine derivatives as Toll-
like receptor agonists introducing the antedrug concept. J. Med. Chem.
2010, 53, 2964−2972.
(37) (a) Biggadike, K.; Coe, D. M.; Lewell, X. Q.; Mitchell, C. J.;
Smith, S. A.; Trivedi, N. Purine Derivatives for the Use in the
Treatment of Allergic, Inflammatory and Infectious Diseases. PCT
WO2010/018133, Feb 18, 2010. (b) Gibbon, R. H.; Lucas, A.;
Hermitage, S. A. 6-Amino-2-{[(1S)-1-methylbutyl]oxy}-9-[5-(1-
piperidinyl)pentyl]-7,9-dihydro-8H-purin-8-one Maleate. PCT
WO2011/098452, Aug 18, 2011.
Kallen
́
, A.; Norlen
TLR7 stimulation reduces allergen responsiveness in allergic rhinitis.
Respir. Res. 2012, 13, 53. (c) Greiff, L.; Cervin, A.; Ahlstrom-
́
, P.; Sjolin, I.-L.; Widegren, H. Repeated intranasal
̈
̈
̈
Emanuelsson, C.; Allenas, M.; Almqvist, G.; Andersson, M.; Cervin,
A.; Dolata, J.; Lindgren, S.; Martensson, A.; Young, B.; Widegren, H.
Biological effects and clinical efficacy of a topical Toll-like receptor 7
agonist in seasonal allergic rhinitis: a parallel group controlled phase
IIa study. Inflammation Res. 2015, 64, 903. (d) Leaker, B. R.; Singh, D.;
̈
̊
́
Lindgren, S.; Kallen, A.; Almqvist, G.; Young, B.; O’Connor, B. J. The
̈
Effects of the Novel Toll-Like Receptor 7 (TLR7) Agonist AZD8848
on Allergen-Induced Responses in Patients with Mild Asthma. Am. J.
Respir. Crit. Care Med. 2012, 185, A4184.
(11) (a) Horak, F. VTX-1463, a novel TLR8 agonist for the
treatment of allergic rhinitis. Expert Opin. Invest. Drugs 2011, 20, 981−
986. (b) Hershberg, R. Methods for the Treatment of Allergic
Diseases. PCT WO 2012/045089, Apr 5, 2012.
(12) Ciprandi, G.; Caimmi, D.; Miraglia del Giudice, M.; La Rosa,
M.; Salpietro, C.; Marseglia, G. L. Recent developments in United
Airways Disease. Allergy, Asthma Immunol. Res. 2012, 4, 171−177.
(13) Horner, A. A. Toll-like receptor ligands and atopy: a coin with at
least two sides. J. Allergy Clin. Immunol. 2006, 117, 1133−1140.
(14) Møller-Larsen, S.; Nyegaard, M.; Haagerup, A.; Vestbo, J.;
Kruse, T. A.; Børglum, A. D. Association analysis identifies TLR7 and
TLR8 as novel risk genes in asthma and related disorders. Thorax
2008, 63, 1064−1069.
(15) Roponen, M.; Yerkovich, S. T.; Hollams, E.; Sly, P. D.; Holt, P.
G.; Upham, J. W. Toll-like receptor 7 is reduced in adolescents with
asthma. Eur. Respir. J. 2010, 35, 64−71.
(16) Drake, M. G.; Kaufman, E. H.; Fryer, A. D.; Jacoby, D. B. The
therapeutic potential of Toll-like receptor 7 stimulation in asthma.
Inflammation Allergy: Drug Targets 2012, 11, 484−491.
(17) Duechs, M. J.; Hahn, C.; Benediktus, E.; Werner-Klein, M.;
Braun, A.; Gerd Hoymann, H.; Gantner, F.; Erb, K. J. TLR agonist
mediated suppression of allergic responses is associated with increased
innate inflammation in the airways. Pulm. Pharmacol. Ther. 2011, 24,
203−214.
(18) Miller, R. L.; Meng, T.-C.; Tomai, M. A. The antiviral activity of
Toll-like receptor 7 and 7/8 agonists. Drug News Perspect. 2008, 21,
69−87.
(19) Czarniecki, M. Small molecule modulators of Toll-like receptors.
J. Med. Chem. 2008, 51, 6621−6626.
(20) Isaacs, A.; Lindenmann, J. Virus interference. I. The interferon.
Proc. R. Soc. London, Ser. B 1957, 147, 258−267.
(21) Nichol, F. R.; Weed, S. D.; Underwood, G. E. Stimulation of
murine interferon by a substituted pyrimidine. Antimicrob. Agents
Chemother. 1976, 9, 433−439.
(22) Lee, J.; Chuang, T.-H.; Redecke, V.; She, L.; Pitha, P. M.;
Carson, D. A.; Raz, E.; Cottam, H. B. Molecular basis for the
immunostimulatory activity of guanine nucleoside analogues:
activation of Toll-like receptor 7. Proc. Natl. Acad. Sci. U. S. A. 2003,
100, 6646−6651.
(23) Miller, R. L.; Gerster, J. F.; Owens, M. L.; Slade, H. B.; Tomai,
M. A. Imiquimod applied topically: a novel immune response modifier
and new class of drug. Int. J. Immunopharmacol. 1999, 21, 1−14.
(24) Meyer, T.; Surber, C.; French, L. E.; Stockfleth, E. Resiquimod,
a topical drug for viral skin lesions and skin cancer. Expert Opin. Invest.
Drugs 2013, 22, 149−159.
(25) Hirota, K.; Kazaoka, K.; Niimoto, I.; Kumihara, H.; Sajiki, H.;
Isobe, Y.; Takaku, H.; Tobe, M.; Ogita, H.; Ogino, T.; Ichii, S.;
Kurimoto, A.; Kawakami, H. Discovery of 8-hydroxyadenines as a
novel type of interferon inducer. J. Med. Chem. 2002, 45, 5419−5422.
(38) Biffen, M.; Matsui, H.; Edwards, S.; Leishman, A. J.; Eiho, K.;
Holness, E.; Satterthwaite, G.; Doyle, I.; Wada, H.; Fraser, N. J.;
Hawkins, S. L.; Aoki, M.; Tomizawa, H.; Benjamin, A. D.; Takaku, H.;
McInally, T.; Murray, C. M. Biological characterisation of a novel class
of Toll-like receptor 7 agonists designed to have reduced systemic
activity. Br. J. Pharmacol. 2012, 166, 573−586.
(39) Russo, C.; Cornella-Taracido, I.; Galli-Stampino, L.; Jain, R.;
Harrington, E.; Isome, Y.; Tavarini, S.; Sammicheli, C.; Nuti, S.;
Mbow, M. L.; Valiante, N. M.; Tallarico, J.; De Gregorio, E.; Soldaini,
E. Small molecule Toll-like receptor 7 agonists localize to the MHC
O
DOI: 10.1021/acs.jmedchem.5b01647
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Drug Annotation
class II loading compartment of human plasmacytoid dendritic cells.
Blood 2011, 117, 5683−5691.
(40) Tanji, H.; Ohto, U.; Shibata, T.; Miyake, K.; Shimuzu, T.
Structural reorganisation of the Toll-like receptor 8 dimer induced by
agonistic ligands. Science 2013, 339, 1426−1429.
(41) (a) Gentile, F.; Deriu, M. A.; Licandro, G.; Prunotto, A.;
Danani, A.; Tuszynski, J. A. Structure based modelling of small
molecules binding to the TLR7 by atomistic simulations. Molecules
2015, 20, 8316−8340. (b) Gupta, C. L.; Akhtar, S.; Sayyed, U.; Pathak,
N.; Bajpai, P. In silico analysis of human Toll-like receptor 7 ligand
binding domain. Biotechnol. Appl. Biochem. 2015, DOI: 10.1002/
bab.1377.
(42) Tanji, H.; Ohto, U.; Shibata, T.; Taoka, M.; Yamauchi, Y.; Isobe,
T.; Miyake, K.; Shimizu, T. Toll-like receptor 8 senses degradation
products of single-stranded RNA. Nat. Struct. Mol. Biol. 2015, 22,
109−115.
(43) Nakamura, T.; Wada, H.; Kurebayashi, H.; McInally, T.;
Bonnert, R. Synthesis and evaluation of 8-oxoadenine derivatives as
potent Toll-like receptor 7 agonists with high water solubility. Bioorg.
Med. Chem. Lett. 2013, 23, 669−672.
(44) Jones, P.; Pryde, D. C.; Tran, T.-D.; Adam, F. M.; Bish, G.; Calo,
F.; Ciaramella, G.; Dixon, R.; Duckworth, J.; Fox, D. N. A.; Hay, D. A.;
Hitchin, J.; Horscroft, N.; Howard, M.; Laxton, C.; Parkinson, T.;
Parsons, G.; Proctor, K.; Smith, M. C.; Smith, N.; Thomas, A.
Discovery of a highly potent series of TLR7 agonists. Bioorg. Med.
Chem. Lett. 2011, 21, 5939−5943.
(45) Sasai, M.; Linehan, M. M.; Iwasaki, A. Bifurcation of Toll-like
receptor 9 signaling by adaptor protein 3. Science 2010, 329, 1530−
1534.
(46) Sorkin, A.; von Zastrow, M. Signal transduction and
endocytosis: close encounters of many kinds. Nat. Rev. Mol. Cell
Biol. 2002, 3, 600−614.
(47) Blasius, A. L.; Arnold, C. N.; Georgel, P.; Rutschmann, S.; Xia,
Y.; Lin, P.; Ross, C.; Li, X.; Smart, N. G.; Beutler, B. Slc15a4, AP-3,
and Hermansky-Pudlak syndrome proteins are required for Toll-like
receptor signaling in plasmacytoid dendritic cells. Proc. Natl. Acad. Sci.
U. S. A. 2010, 107, 19973−19978.
(48) Iwasaki, A.; Medzhitov, R. Toll-like receptor control of the
adaptive immune responses. Nat. Immunol. 2004, 5, 987−995.
(49) Barchet, W.; Wimmenauer, V.; Schlee, M.; Hartmann, G.
Accessing the therapeutic potential of immunostimulatory nucleic
acids. Curr. Opin. Immunol. 2008, 20, 389−395.
(50) (a) Ketloy, C.; Engering, A.; Srichairatanakul, U.; Limsalakpetch,
A.; Yongvanitchit, K.; Pichyangkul, S.; Pichyangkul, S.; Ruxrungtham,
K. Expression and function of Toll-like receptors on dendritic cells and
other antigen presenting cells from non-human primates. Vet.
Immunol. Immunopathol. 2008, 125, 18−30. (b) Wagner, T. L.;
Horton, V. L.; Carlson, G. L.; Myhre, P. E.; Gibson, S. J.; Imbertson, L.
M.; Tomai, M. A. Induction of cytokines in cynomolgus monkeys by
the immune response modifiers, Imiquimod, S-27609 and S-28463.
Cytokine+ 1997, 9, 837−845.
(51) Puggioni, F.; Durham, S. R.; Francis, J. N. Monophosphoryl
lipid A (MPL®) promotes allergen-induced immune deviation in
favour of Th1 responses. Allergy 2005, 60, 678−684.
(52) Tsitoura, D.; Ambery, C.; Price, M.; Powley, W.; Garthside, S.;
Biggadike, K.; Quint, D. Early clinical evaluation of the intranasal
TLR7 agonist GSK2245035: use of translational biomarkers to guide
dosing and confirm target engagement. Clin. Pharmacol. Ther. 2015,
98, 369−380.
(53) Richwald, G. A. Imiquimod. Drugs Today 1999, 35, 497−511.
(54) Fidock, M. D.; Souberbielle, B. E.; Laxton, C.; Rawal, J.;
Delpeuch-Adams, O.; Corey, T. P.; Colman, P.; Kumar, V.; Cheng, J.
B.; Wright, K.; Srinivasan, S.; Rana, K.; Craig, C.; Horscroft, N.;
Perros, M.; Westby, M.; Webster, R.; van der Ryst, E. The innate
immune response, clinical outcomes and ex vivo HCV antiviral efficacy
of a TLR7 agonist (PF-4878691). Clin. Pharmacol. Ther. 2011, 89,
821−829.
P
DOI: 10.1021/acs.jmedchem.5b01647
J. Med. Chem. XXXX, XXX, XXX−XXX