The identification of AF38469: An orally bioavailable inhibitor of the VPS10P family sorting receptor Sortilin

Article abstract of DOI:10.1016/j.bmcl.2013.11.046

The identification of the novel, selective, orally bioavailable Sortilin inhibitor AF38469 is described. Structure-activity relationships and syntheses are reported, along with an X-ray crystal structure of the sortilin-AF38469 protein-inhibitor complex.

Full text of DOI:10.1016/j.bmcl.2013.11.046

Bioorganic & Medicinal Chemistry Letters 24 (2014) 177–180
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
The identification of AF38469: An orally bioavailable inhibitor of the
VPS10P family sorting receptor Sortilin
Tenna Juul Schrøder a, Søren Christensen a, Samsa Lindberg a, Morten Langgård a, Laurent David a,
Philip J. Maltas a, Jørgen Eskildsen a, Jan Jacobsen a, Lena Tagmose a, Klaus Bæk Simonsen a,
Lars Christian Biilmann Rønn a, Inge E. M. de Jong a, Ibrahim J. Malik a, Jens-Jakob Karlsson a,
Christoffer Bundgaard a, Jan Egebjerg a, Jeffrey B. Stavenhagen a, Dorthe Strandbygård b, Søren Thirup b,
Jacob Lauwring Andersen b, Srinivas Uppalanchi c, Sridhar Pervaram c, Shiva Prasad Kasturi c,
Pradheep Eradi c, Durga Rao Sakumudi c, Stephen P. Watson a,
⇑
a Neuroscience Drug Discovery, H. Lundbeck A/S, Copenhagen, Denmark
b The Lundbeck Foundation Research Centre MIND, Department of Molecular Biology and Genetics, Aarhus University, Gustav Wieds Vej 10C, Aarhus C 8000, Denmark
c Medicinal Chemistry, GVK BioScience, Hyderabad, India
a r t i c l e i n f o
a b s t r a c t
Article history:
The identification of the novel, selective, orally bioavailable Sortilin inhibitor AF38469 is described.
Structure–activity relationships and syntheses are reported, along with an X-ray crystal structure of
the sortilin-AF38469 protein-inhibitor complex.
Received 16 October 2013
Revised 15 November 2013
Accepted 19 November 2013
Available online 27 November 2013
Ó 2013 Elsevier Ltd. All rights reserved.
Keywords:
Sortilin
Neurotensin
AF38469
AF40431
VPS10P
Sorting receptor
Sortilin crystal structure
Protein-inhibitor complex
Sortilin is a type I membrane receptor belonging to the vacuolar
protein sorting 10 protein (VPS10P) family of sorting receptors.1
Sortilin is widely expressed in both the central nervous system
and periphery. It mediates a number of important physiological
functions via trafficking of, and signalling with, a variety of differ-
ent protein partners. For example sortilin is involved in signalling
via the neurotrophins, nerve growth factor (NGF) and brain-
derived neurotrophic factor (BDNF). Indeed, in complex with the
protein p75, sortilin has been reported to form the receptor for
pro-neurotrophin-mediated apoptotic effects leading to degenera-
tion and cell death in cellular and animal models.1 Sortilin has also
been demonstrated to interact with apolipoprotein B100 in the
Golgi and facilitate the export of apoB100-containing lipoproteins,
thereby regulating plasma low-density lipoprotein (LDL)
function as a high affinity receptor for progranulin3, and to mediate
clearance of progranulin by binding followed by cellular uptake
and distribution to lysosomes.
The 13 amino acid neuropeptide Neurotensin (NTS) is a sortilin
ligand, indeed sortilin had been known as the neurotensin receptor
3 (NTR34). Moreover the structure of the Neurotensin–Sortilin
complex has been determined by X-ray crystallography.5
O
O
CF3
HO
OH
O
NH
HO
O
O
HN
N
cholesterol levels,
a key contributor to atherosclerosis and
ischemic heart disease.2 Recently, sortilin was also shown to
AF40431
Sortilin NTS IC50 4µM
AF38469
Sortilin NTS IC50 = 330nM
⇑
Corresponding author.
E-mail address: stwa@lundbeck.com (S.P. Watson).
0960-894X/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2013.11.046

178
T. J. Schrøder et al. / Bioorg. Med. Chem. Lett. 24 (2014) 177–180
W
Z
O
Thus due to its rich biology and association with pathological
W
Z
W
O
O
X
Y
X
Y
X
Y
ii
processes sortilin represents an interesting target. However there
is currently a dearth of molecular tools available to understand
and disease associate the biology of sortilin. We previously
reported6 the identification of the Sortilin inhibitor AF40431, the
first small molecule ligand for sortilin. Unfortunately, the very
low solubility and membrane permeability of AF40431 limited its
use as a molecular probe.
i
OH
O
O
N R
O
O
HN
Z
R
6
5
4
Scheme 2. Reagents and conditions: (i) ArNH2 in AcOH, 5 h, reflux, (ii) LiOH, THF/
Water, 1 h room temperature.
Herein we report the identification of a series of small molecule
sortilin inhibitors, optimisation of which led to the orally bioavail-
able inhibitor AF38469. This molecule could prove a valuable
molecular tool to further delineate and disease associate the biol-
ogy of sortilin.
O
O
O
tBu
Br
Br
Br
i
ii
OH
O
O
O
O
iii
O
HN
OH
A high throughput screen (HTS) of the Lundbeck collection was
performed using an NTS scintillation proximity assay (SPA) format
binding assay.7 A series of phthalimides, exemplified by compound
1 was one of the promising chemotypes that emerged from the
HTS. Whilst these pthalimides were fully characterised, demon-
strated good chemical stability in DMSO and DMSO/water solution
they displayed a perplexing structure–activity relationship (SAR)
which led to a more detailed investigation. LCMS analysis of assay
wells revealed that phthalimide (1) degraded under assay condi-
tions to afford the two corresponding phthalimide hydrolysis prod-
ucts, phthalamic acids 2 and 3, respectively (scheme 1). Synthesis
and testing of these two regioisomeric acids revealed 2 to be
responsible for the observed NTS binding inhibition (with an IC50
of 710 nM), whilst compound 3 was only weakly active. Compound
2 thus represented an interesting Sortilin inhibitor hit which be-
came the subject of an investigation/optimisation programme.
Analogues (4) of compound 2 were typically prepared via sim-
ple hydrolysis of the corresponding phthalimides (5) and separa-
tion of regioisomers via preparative HPLC. Phthalimides 5 were
in turn prepared via condensation-cyclo-dehydration from the cor-
responding phthalic anhydrides 6 (scheme 2) (overall this route
was more expedient than attempting a mono-coupling to the cor-
responding phthalic acids).
R
8
9
7
Scheme 3. Reagents and conditions: (i) tBuOH, DMPA, CH2Cl2, 2 h, then SFC, (ii)
ArNH2 HATU, CH2Cl2 18 h, room temperature, (iii) CH2Cl2/TFA 1:1, 4 h room
temperature.
Table 1
Sortilin binding affinity of compounds 10
W
O
6
5
X
Y
OH
O
4
3
HN
N
Z
10
Compound
W
X
Y
Z
NTS6 IC50 (nM)
10a
10b
2
H
Me
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Me
15%*
5%*
Br
Cl
710
10c
10d
10e
AF38469
10f
1400
6400
1100
330
Me
iPr
CF3
NO2
Ph
H
1700
10g
3
10h
10i
10j
10k
10l
H
2%*
Alternatively compounds 7, bearing just a 5-Bromo substituent
on the phthalamic acid ring, were selectively prepared from corre-
sponding phthalate mono-t-butyl ester 8 via coupling-deprotec-
tion (scheme 3). Intermediate 8 itself was prepared from the
reaction of phthalic anhydride 9 with tert-butanol and separation
of the regioisomeric mono esters by supercritical fluid chromatog-
raphy (SFC). In addition, palladium-catalysed coupling could be
used to replace the bromine of intermediate 8 (and also of com-
pounds 7) with alternative substituents.
Br
Cl
Me
CF3
Cl
Me
H
12000
69%*
57%*
9100
830
H
H
H
Cl
Me
H
3400
10m
43%*
*
% Inhibition at 50 lM for compounds too weak to fit a dose-response curve,
IC50 > 50 lM.
Table 1 shows the effect of modifications to the substituents of
the phthalamic acid ring of compounds of type 10 on sortilin
potency. Interestingly the compound with no such ring substitu-
ents (10a) is inactive. As seen with the initial hits 2 and 3 there
is a clear preference for substituents in the 5-position, affording
much higher potency compared to the 4-position, as is demon-
strated by Br, Cl, Me and CF3 substituents. Compounds 10b and
10m, bearing a methyl substituent at the 3- or 6-position respec-
tively are inactive, as is the 5-phenyl analogue (10g). Interestingly
the 4,5 di-chloro and 4,5-di-methyl analogues (10k and 10l)
show a slight increase in potency compared to their 5-chloro and
5-methyl mono-substituted counterparts (10c and 10d) The
trifluoromethyl group of AF38469 confers the highest potency of
the substituents explored.
Table 2 shows the effect of modifications to the amide substitu-
ent of compounds of type 11 on sortilin potency.
Simple alkyl amides were inactive e.g. 11a and 11b, as interest-
ingly, is phthalamic acid 11c with an unsubstituted phenyl group.
However, relative to the unsubstituted phenyl (11c), introduction
of a m-Me substituent or an o-pyridine nitrogen atom afforded
compounds of modest potency (11d and 11e) but a combination
of both of these features shows a synergistic effect in the potency
of the original hit compound 2. Some other pyridine nitrogen and
methyl substitutions afford active compounds but are all less
O
O
Br
O
OH
O
potent than the arrangement in compound 2 and AF38469
OH
O
Br
N
+
(e.g. 11g and 11j). Pyridine 6,5-di-substitution in di-methyl
compound 11n affords a modest increase in potency, and in the
tetrahydroquinoline 11o, with an IC50 of 100 nM, the most potent
compound prepared.
N
Br
HN
N
HN
N
O
1
2
3
NTS IC50 = 710nM
NTS IC50 = 12000uM
Methylation of the amide nitrogen of 2 in compound 12 com-
pletely ablates sortilin potency. Replacements of the carboxylic
Scheme 1. Degradation of phthalamide 1 under assay6 conditions.

T. J. Schrøder et al. / Bioorg. Med. Chem. Lett. 24 (2014) 177–180
179
Table 2
Sortilin binding affinity of compounds 11
O
X
OH
O
HN
R
11
Compound
X
R
NTS6 IC50 (nM)
11a
11b
11c
Br
Br
Br
sBu
iPr
Ph
*
30%*
35%*
69%*
11d
11e
2
Br
Br
Br
Br
5200
3600
710
N
N
N
*
*
*
Cl
11f
1600
N
Figure 1. Overlay of X-ray crystal structures of sortilin bound AF38469 (PDB code
4N7E) and Neurotensin (C-terminus) (PDB code 3F6K).
11g
Br
2400
*
*
N
11h
11i
Br
Br
Br
Br
64%*
70%*
1800
490
N
*
*
11j
N
N
N
N
N
N
N
*
*
*
*
*
11k
AF38469
11m
CF3
CF3
CF3
CF3
330
1300
170
100
OMe
11n
11o
*
% Inhibition at 50 lM for compounds too weak to fit a dose-response curve,
IC50 > 50 lM.
acid functionality also ablated activity; all compounds 13 were
inactive, indicative of an acidic interaction with the protein (as
was subsequently confirmed by x-ray crystallography of the
AF38469-Sortilin complex, Figures 1 and 2).
Figure 2. Overlay of X-ray crystal structures of sortilin bound AF38469 (PDB code
4N7E) and AF40431 (PDB code 4MSL)5.
O
X
Br
Br
OH
O
O
Table 3
HN
N
N
N
Lipohilic ligand efficiency (LLE) of compounds 11
Compound
clogP
LLE
12
13 X = SO2, NO2, H,F,Br, Me, OMe
2
11j
11k
11m
11n
11o
2.2
2.2
1.6
2.5
2.7
3.3
2.2
3.9
3.5
4.7
3.4
4.1
3.7
4.3
The more potent of compounds 11 typically demonstrated
moderate lipophilicity with some of the more polar examples
demonstrating promising Ligand Lipohilic Efficiency8 (LLE) above
4 (Table 3).
AF38469 was progressed for further studies. It should be noted
that Neurotensin itself exhibits a binding IC50 of 360 nM in the
3HNTS assay7, thus AF38469 is essentially equipotent with this sor-
tilin substrate.
AF38469
AF38469 showed no inhibition or stimulation of >50% at 10 lM
selected panel of targets known to bind acidic molecules (d-Opioid,
GPR40, PPARd, EP1, Angiotensin AT1, Endothelin ETA B,
MMP-12). Thus overall the selectivity profiling of AF38469
demonstrated a highly specific interaction with sortilin.
in a standard selectivity panel of ca. 70 targets run at CEREP.9
Importantly AF38469 showed no activity against the NTR1
receptor. In addition AF38469 showed no activity against a
&

180
T. J. Schrøder et al. / Bioorg. Med. Chem. Lett. 24 (2014) 177–180
O
Table 4
O
Cl
Cl
Cl
AF38469 pharmacokinetic parameters in rat
N
OH
O
N
In Vitro
Solubility 134 lg/ml, plasma free fraction
2% intrinsic clearance 0.4 l/h/kg
Cl
O
HN
N
Oral (2 mg/kg)
i.v. (1 mg/kg)
Cmax 12850 ng/ml
Clb 0.03 L/h/Kg, t1/2 1.2 h, Vss 0.02 L/kg
14
10k
Scheme 4. Symmetrical phthalamides as pro-drugs.
Table 4 describes key pharmacokinetic parameters for AF38469
in rat. The compound has a low volume of distribution (relative to
blood volume of 0.7 l/kg) and low clearance (relative to liver blood
flow of ca. 4.8 l/h/kg), and a half-life of 1.2 h. The initial exposure in
plasma is high, due in part to the low volume of distribution. The
oral bioavailability is 35%. Simple calculations based on Cmax, free
fraction and the sortilin potency suggest that a free plasma concen-
tration of AF38469 ca. 2 fold higher than its IC50 is obtained at
Cmax from an oral 2 mg/kg dose.
In principle a phthalimide could act as a pro-drug for an
AF38469 type compound, with hydrolysis in vivo affording the cor-
responding phthalamic acid by an analogous process to that which
lead to the liberation and identification of compound 2 itself
(Scheme 1). Whilst an asymmetrically substituted phthalimide
such as 1 would afford a mixture of regioisomers on hydrolysis
(viz. 2 and 3), a symmetrical phthalimide would afford a single
phthalamic acid (e.g. Scheme 4). To this end phthalimide 14 was
investigated as a potential oral prodrug for phthalamic acid 10k.
However, whilst oral administration of the phthalimide (14) did af-
ford systemic exposure of phthalamic acid 10k, the free exposure
of the acid was not an improvement on the exposure of AF38469
obtained via oral administration.
substituent of the former are virtually congruent with the pyran
ring and methyl substituent of the latter.
It is anticipated that the structure of the sortilin-AF38469 com-
plex will be used to further drive optimisation of the chemotype
through classical structure based drug design.
In summary we have identified a potent, selective and orally
bioavailable inhibitor for the VPS10P family sorting receptor Sorti-
lin. We hope and anticipate that AF38469 will serve as an impor-
tant tool to further delineate the biology of Sortilin, and to
facilitate evaluation of the therapeutic potential of this protein.
Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.bmcl.2013.11.046.
References and notes
1. Nykjaer, A.; Willnow, T. E. Trends Neurosci. 2012, 35, 261.
2. Kjolby, M.; Andersen, O. M.; Breiderhoff, T.; Fjorback, A. W.; Pedersen, K. M.;
Madsen, P.; Jansen, P.; Heeren, J.; Willnow, T. E.; Nykjaer, A. Cell Metab. 2010,
12, 213.
3. Hu, F.; Padukkavidana, T.; Vægter, C. B.; Zheng, Y.; Mackenzie, I. R.; Feldman, H.
H.; Nykjaer, A.; Strittmatter, S. M. Neuron 2010, 68, 654.
4. The Neurotensin receptors, NTR1 and NTR2 are G-protein coupled receptors
see Kitabgi, P. Curr. Opin. Drug Discovery Dev. 2002, 5, 764–768.
5. Quistgaard, E. M.; Madsen, P.; Grøftehauge, M. K.; Nissen, P.; Petersen, C. M.;
Thirup, S. S. Nature Struc. Mol. Biol. 2009, 9, 46.
6. Andersen, J.L.; Schrøder, T.J.; Christensen, S.; Strandbygaard, D.; Pallesen, L.T.;
García-Alai, M.M.; Lindberg, S.; Langgård,M.; Eskildsen, J.; David, L.; Tagmose,
L.; Simonsen, K.B.; Maltas, P.J.; Biilmann Rønn, L.C.; de Jong, I.M.E.; Malik, I.J.;
Egebjerg, J. Karlsson, J.J.; Uppalanchi, S.; Sakumudi, D.R., Eradi, P.; Watson,
S.P.; Thirup, S. Acta Cryst. 2014, D70, http://dx.doi.org/10.1107/S13990
04713030149.
A co-crystal of sortilin and AF38469 was successfully obtained
and its structure determined10 by X-Ray crystallography to a reso-
lution of 2.78 Å. Figure 1 shows an overlay of the crystal structure
of AF38469 with that of the structure of the C-terminal region of
sortilin bound Neurotensin.5 AF38469 makes similar interactions
with sortilin as does the C-terminal Leu residue of Neurotensin.
AF38469 makes a salt bridge to Arg292 via its carboxylic acid,
the CF3 group occupies the same hydrophobic binding pocket as
the iPr of the Leu side-chain, and the amide N–H makes a hydro-
gen-bond donor interaction with Tyr318. The pyridyl methyl group
7. Compound affinity was determined by measuring the displacement of 3H-
neurotensin binding to hSortilin in SPA format. Total volume of 40 ll in 50 mM
makes a hydrophobic p-interaction with Phe317, similar to hydro-
HEPES pH 7.4 assay buffer containing 100 mM NaCl, 2.0 mM CaCl2, 0.1% BSA
and 0.1% Tween-20. Compound pre-incubation for 30 min at RT with 150 nM of
6his-Sortilin before 5 nM [3H]-Neurotensin and Ni chelate imaging beads
(Perkin Elmer) were added, after 6 h plate was read on a ViewLux with 360 s
exposure time. Dose-response evaluation of compounds was performed with
10 concentrations of drugs (covering 3 decades). IC50 values were calculated by
nonlinear regression using the sigmoid concentration-response (variable slope)
phobic interaction of Ile12 of Neurotensin. The structure also
serves to rationalise many of the structure–activity relationships
outlined above, for example the key role of the pthalamic acid 5
substituent, the inactivity of the 5-Ph (10f) (due to size exclusion),
and the pivotal role of the carboxylic acid functionality.
using Xlfit
determination
8. Abad-Zapatero, C.; Metz, J. Drug Discovery Today 2005, 10, 464.
9. In earlier versions of the current CEREP ExpresSprofile, ADR CNS and blood
panels—see http://www.cerep.fr/.
4 (IDBS, UK). All values reported are average of at least 4
Figure 2 shows an overlay of the binding mode of AF38469
with the previously reported6 small molecule inhibitor AF40431.
AF38469 and AF40431 exhibit many common interactions,
including the salt bridge and leucine pocket binding features of
Neurotensin, but also interestingly the pyridine ring and methyl
10. PDB Code 4N7E See Supplementary Data.