ACS Medicinal Chemistry Letters
Letter
A general synthesis of isoindoline anti-AβO compounds was
developed (Scheme 2). Ready access to the gem-dimethyl-
substituted isoindolines was realized via palladium-mediated
coupling46 of aryl iodides 14 and 1545 to 2-methyl-3-butyn-2-
amine to provide the gem-dimethyl-substituted propargyl-
amines 16 and 17 (74% yield for 17). These were
subsequently hydrogenated to the intermediate amines 19
and 20 (100% yield for 20). The available chiral alpha-methyl
amines 12 and 1837−39 and the gem-dimethyl-substituted
amines 19 and 20 were condensed with the corresponding bis-
benzylbromides 21−23 to provide the isoindoline anti-AβO
compounds 4−7 (63% yield for 7).45
dependent learning and memory compared to vehicle-treated
animals in the fear conditioning assay (Table 5).
Microelectrodes coated with antibodies for total Aβ or AβOs
were inserted into the brains of live transgenic mice and
measured soluble Aβ concentration by the minute.24,49
Intravenous administration of compound 7 increased release
of AβO oligomers and increased AβO concentration in the
interstitial brain fluid without causing a measurable increase in
Aβ monomer concentration. Remarkably, a concomitant
increase in the concentration of AβO in the cerebrospinal
fluid (CSF) was also observed.24 This strongly suggests that
AβO displacement by compound 7 in the brain facilitates
clearance of interstitial AβO to the CSF.
Our early lead compounds 2 and 3 had exhibited promising
neuronal activity and performed well in animal behavioral
models but were exceedingly lipophilic as reflected in their low
PSA (∼12). These compounds exhibited problematic off-target
activity profiles and hERG activity. Certain trends in
physiochemical properties including a substantial increase in
total polar surface area (PSA) (Table 2) contributed to
progress in the program and enabled the nomination of
suitable development candidates. It became clear that
increased PSA addressed off-target and hERG activities. The
reintroduction of phenol and alkoxy substituents as in
compounds 5 and 6 informed the design of next generation
compounds. Incorporation of the aryl methyl sulfone increased
PSA, minimized off-target activities and hERG activity, and
enabled the nomination of our clinical candidate 7. Compound
7 exhibits good absorption and a particularly high brain-to-
plasma ratio of 5.7 at 24 h postdose (Kp,uu = 6.75). The free
fraction of compound 7 in the brain is 13% and in the plasma
is 5%. Compound 7 is not a Pgp substrate and has a CACO-2
Papp(B-A)/Papp(A-B) = 0.95.
This limited structure−activity relationship represents a
medicinal chemistry effort that produced hundreds of
structural analogues. Our effort to optimize anti-AβO activity
in our neuronal assay was complicated by issues of off-target
activities and pharmacokinetics that are beyond the scope of
this Letter. However, the structural features required of the
anti-AβO pharmacophore are well represented here. In
summary, the simple lipophilic benzylic amines 2 and 3
exhibited promising anti-AβO activity and acceptable meta-
bolic stability. They were efficacious in animal behavioral
testing22,24 but suffered from hERG and other off-target
activities due, apparently, to their relatively low PSA. The
isoindolines 4−7 exhibited promising submicromolar neuronal
activity. The isoindoline scaffold tolerated polar substitution
and allowed dramatic increases in PSA leading to highly active
compounds with minimal hERG and off-target activities. Our
advanced candidates 6 and 7 benefitted as well from gem-
dimethyl substitution alpha to nitrogen that imparts metabolic
stability and optimal conformational biases for anti-AβO
activity.
In primary hippocampal and cortical neuronal cultures AβOs
were found to bind specifically and to saturate a single receptor
site on some but not all neurons.22,23 The anti-AβO
compounds reported here prevented and restored trafficking
deficits caused by AβOs in neurons. The offset between sigma-
2 binding affinity and potency in the in vitro membrane vesicle
trafficking assay likely results from several factors such as (1)
lipophilic compound adsorption to microtiter plate plastic that
lowers the effective concentration of the compounds, (2)
greater than 80% of sigma-2 receptors need to be occupied by
compound to achieve an effect on binding and function, or (3)
the high concentration of the low potency synthetic Aβ
preparations needed to achieve adequate testing windows in
the in vitro assays. Clinical candidate 7 prevented and reversed
trafficking deficits caused by AβOs but had no effect in the
absence of AβOs (Figure 1). Compound 7 also prevented
binding AβO to neuronal receptors, displaced prebound AβO,
and was determined by a one-site ELISA assay to have no
effect on AβO assembly or AβO dissociation.24,47
The anti-AβO compounds 2, 3, 6, and 7 restored cognitive
function in transgenic hAPP Swe/Ldn mice (Tables
3−5).22,24,48 The drug-treated transgenic mice performed in
the Morris water maze task significantly better than did the
transgenic vehicle-treated mice (Table 3). Treatment with
compound 7 does not affect nontransgenic animal perform-
ance. Transgenic mice treated with compound 7 remembered
previous arms entered in the Y-maze task significantly better
than chance but vehicle-treated transgenic animals did not
(Table 4). Transgenic mice treated with compound 7
demonstrated significant improvements in spatial and cue-
The CNS druglike small molecules reported here are
potentially first-in-class therapeutics for the treatment and
prevention of early cognitive decline and neurodegeneration in
MCI and AD patients. Our first-in-class clinical candidate
compound 7 (CT1812)45 is an AβO-displacing compound
and a potent and highly selective antagonist of the sigma-2
receptor. Compound 7 has been demonstrated to prevent AβO
binding to neurons and also to displace bound AβO from
neuronal receptors.22−24 It has been determined to have no
effect on AβO assembly or AβO dissociation in a biochemical
ELISA assay.24,47 It is metabolically stable and exhibits good
pharmacokinetics and robust brain exposure and restores
synapse number and cognitive function in transgenic mouse
models.24,48 Remarkably, compound 7, upon AβO-displace-
ment in the rodent brain, facilitates the clearance of AβO from
brain interstitial fluid to CSF as demonstrated in a mouse
microimmunoelectrode study.24 Unlike other Aβ-targeted
therapeutics, by displacing Aβ oligomer binding, CT1812
lowers Aβ oligomer affinity for its receptor, the same
phenotype observed with the protective Icelandic mutation
that confers 4-fold lower incidence of Alzheimer’s disease on
carriers.50
In phase I clinical studies, compound 7 was determined to
be safe and well tolerated at single doses up to 1120 mg and at
multiple doses up to 560 mg in healthy elderly volunteers.51,52
Adverse events, most commonly headache and gastrointestinal
symptoms, were mild to moderate in severity. Plasma
concentrations of drug were found to be dose proportional,
and CSF concentrations upon multiple doses exceeded the
expected minimum target concentrations required to improve
D
ACS Med. Chem. Lett. XXXX, XXX, XXX−XXX