Discovery of Dihydropyrrolo[1,2- a]pyrazin-3(4 H)-one-Based Second-Generation GluN2C- And GluN2D-Selective Positive Allosteric Modulators (PAMs) of the N-Methyl- d -Aspartate (NMDA) Receptor

Article abstract of DOI:10.1021/acs.jmedchem.9b01733

The N-methyl-d-aspartate receptor (NMDAR) is an ion channel that mediates the slow, Ca2+-permeable component of glutamatergic synaptic transmission in the central nervous system (CNS). NMDARs are known to play a significant role in basic neurological functions, and their dysfunction has been implicated in several CNS disorders. Herein, we report the discovery of second-generation GluN2C/D-selective NMDAR-positive allosteric modulators (PAMs) with a dihydropyrrolo[1,2-a]pyrazin-3(4H)-one core. The prototype, R-(+)-EU-1180-453, exhibits log unit improvements in the concentration needed to double receptor response, lipophilic efficiency, and aqueous solubility, and lowers cLogP by one log unit compared to the first-generation prototype CIQ. Additionally, R-(+)-EU-1180-453 was found to increase glutamate potency 2-fold, increase the response to maximally effective concentration of agonist 4-fold, and the racemate is brain-penetrant. These compounds are useful second-generation in vitro tools and a promising step toward in vivo tools for the study of positive modulation of GluN2C- and GluN2D-containing NMDA receptors.

Full text of DOI:10.1021/acs.jmedchem.9b01733

Subscriber access provided by Uppsala universitetsbibliotek
Article
The Discovery of Dihydropyrrolo[1,2-a]pyrazin-3(4H)-one-Based
Second-Generation GluN2C- and GluN2D-Selective Positive Allosteric
Modulators (PAMs) of the N-Methyl-D-Aspartate (NMDA) Receptor
Matthew Epplin, Ayush Mohan, Lynnea Harris, Zongjian Zhu, Katie L. Strong, John
Bacsa, Phuong Le, David S. Menaldino, Stephen F. Traynelis, and Dennis C. Liotta
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b01733 • Publication Date (Web): 15 Jun 2020
Downloaded from pubs.acs.org on June 15, 2020
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 116
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The Discovery of Dihydropyrrolo[1,2-a]pyrazin-3(4H)-one-Based Second-Generation
GluN2C- and GluN2D-Selective Positive Allosteric Modulators (PAMs) of the N-Methyl-D-
Aspartate (NMDA) Receptor
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
†
†
†
‡
†
Matthew P. Epplin, Ayush Mohan, Lynnea D. Harris, Zongjian Zhu, Katie L. Strong, John
†
‡
†
*,‡
*,†
Bacsa, Phuong Le, David S. Menaldino, Stephen F. Traynelis, Dennis C. Liotta
†Emory University, Department of Chemistry, 1515 Dickey Drive, Atlanta, GA 30322, USA
‡Emory University, Department of Pharmacology, 1510 Clifton Road, Atlanta, GA 30322, USA
*Co-corresponding Authors
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 116
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Abstract
The N-methyl-D-aspartate receptor (NMDAR) is an ion channel that mediates the slow,
2
+
Ca -permeable component of glutamatergic synaptic transmission in the central nervous system
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
CNS). NMDARs are known to play a significant role in basic neurological functions, and their
dysfunction has been implicated in several CNS disorders. Herein, we report the discovery of
second-generation GluN2C/D-selective NMDAR positive allosteric modulators (PAMs) with a
dihydropyrrolo[1,2-a]pyrazin-3(4H)-one core. The prototype, R-(+)-EU-1180-453, exhibits log
unit improvements in the concentration needed to double receptor response, lipophilic efficiency,
aqueous solubility, and lowers cLogP by one log unit compared to the first-generation prototype
CIQ. Additionally, R-(+)-EU-1180-453 was found to increase glutamate potency 2-fold, increase
the response to maximally effective concentration of agonist 4-fold, and the racemate is brain
penetrant. These compounds are useful second-generation in vitro tools and a promising step
towards in vivo tools for the study of positive modulation of GluN2C- and GluN2D-containing
NMDA receptors.
ACS Paragon Plus Environment

Page 3 of 116
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Introduction
The N-methyl-D-aspartate receptor (NMDAR) belongs to the family of mammalian ligand-
gated excitatory ionotropic glutamate receptors (iGluRs) that includes the 2-amino-3-(5-methyl-
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
-oxo-1,2-oxazol-4-yl)propanoic acid receptors (AMPARs) and (2S,3S,4S)-3-(carboxymethyl)-4-
1
prop-1-en-2-ylpyrrolidine-2-carboxylic acid (kainate) receptors. NMDARs are known to play a
role in learning,2, 3 memory, brain development, and synaptic plasticity
4
5
6, 7
and have been
implicated in numerous neurological conditions including schizophrenia,8, 9 Alzheimer’s disease,10
Parkinson’s disease,11 Huntington’s chorea,12 epilepsy,13 neuropathic pain,14 ischemic brain
injury,15-17 and depression.18, 19 This has driven interest in the development of both positive and
negative modulators of the NMDAR for potential therapeutic gain.20, 21
Structurally, NMDARs are heterotetrameric assemblies of two GluN1 and two GluN2
protein subunits, which bind co-agonists glycine and glutamate, respectively. Each subunit
contains four semiautonomous domains: an amino-terminal domain (ATD), an agonist-binding
1
domain (ABD), a transmembrane domain (TMD), and a carboxyl terminal domain (CTD). Upon
2
+
22, 23
co-agonist binding at the ABD, membrane depolarization occurs which removes Mg block
2
+
+
from the pore. This allows Ca and Na ions to flow into the cell, which contribute to post-synaptic
2
4
signal transmission. The four isoforms of the GluN2 subunit (GluN2A-D) are expressed with
different spatiotemporal patterns in the brain25-27 and endow the receptor with unique
pharmacological properties including open probability,28-31 agonist potency, and deactivation
32
3
3
time. For example, GluN2C- and GluN2D-containing NMDA receptors are expressed more
prominently in the cerebellum, basal ganglia, thalamus, and cortical and hippocampal
interneurons,25, 26, 34 as compared to GluN2A- and GluN2B-containing receptors. Functionally,
GluN2C/GluN2D-containing NMDA receptors show increased glutamate and glycine
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 116
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
sensitivity32, 35 and reduced open probability,29, 30, 36 Mg sensitivity, Ca permeability, and
single-channel conductance when compared to GluN2A/GluN2B-containing receptors.37
Because these subunit-specific expression patterns and GluN2-specific functional
properties allow NMDARs to serve unique roles at different synapses, subunit-selective
modulators provide an opportunity to target the aforementioned diseases through selective
2+
25
2+
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
8
modification of specific circuits. However, FDA-approved drugs that target NMDARs non-
selectively block the highly-conserved NMDAR pore, irrespective of subunit composition. This
results in the inhibition of all NMDAR subtypes, potentially leading to the neurological side effects
that include altered cardiovascular activity, hallucinations, delusions, and impaired motor
function.39
Subunit-selective NMDAR modulators have therefore become a focus of the research
community, given their potential to reduce side effects by targeting one subunit in specific areas
of the brain for therapeutic gain. Specifically, NMDAR positive allosteric modulators (PAMs)
could be useful in conditions associated with reduced NMDAR activity such as schizophrenia,40
4
2
1, 42
autism spectrum disorders,43, 44 anti-NMDAR encephalitis,45-47 and age-related memory loss.
, 38, 48, 49
Further, there is evidence suggesting that positive therapeutic outcomes could be achieved
with the development of GluN2C- and/or GluN2D-subtype selective PAMs. One hypothesis is that
5
0
the hypofunction of GluN2C-containing receptors plays a role in schizophrenic symptoms and,
thus, GluN2C-selective PAMs could have therapeutically relevant actions. Additionally, post-
mortem examinations of schizophrenic patients with a specific NRG1 gene polymorphism revealed
51
reduced GluN2C expression in the right cerebellum. GluN2D-containg receptors may also play
a role in schizophrenia.51-54 For example, there is evidence to suggest decreased inhibitory drive
can lead to hyperactivity of principal cells and enhanced dopamine release.55 Expression of
ACS Paragon Plus Environment

Page 5 of 116
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
GluN2D in interneurons25, 56, 57 suggests that selective enhancement of GluN2D-containing
receptor function could potentially rectify this hypothesized circuit imbalance. Finally, (3-
chlorophenyl)(6,7-dimethoxy-1-((4-methoxyphenoxy)methyl)-3,4-dihydroisoquinolin-2(1H)-
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
yl)methanone (CIQ, 1), a GluN2C/D-selective PAM previously developed in our lab (Figure 1),58,
5
9
has shown promising results in behavioral mouse models. This includes the ability recover
6
0
striatal synaptic plasticity deficit in a Parkinson’s model, facilitate the retention of fear and
6
1
extinction learning, and prevent MK-801 (a noncompetitive antagonist of NMDARs) induced
6
2
prepulse inhibition deficit, suggesting the clinical potential of potent positive modulators of these
subtypes.
OMe
O
OEt
O
O
H
O
O
N
N
N
N
S
MeO
MeO
Cl
F
n-Pr
O
Me
O
N
Me
N
N
H
Me
O
CIQ (1)
GluN2C/D-selective PAM
2
GNE-9278 (3)
Non-selective PAM
GluN2B-preferring PAM
Me
Me
Me
OH
CN
O
O
O
Me
Me
Me
Me
OH
O
N
CF3
N
Cl
H
N
N
Me
H
MeO2C
Me
NH
Me
H
N
O
O
H
H
Cl
H
H
S
HO
O
HO
PS (4)
Non-selective PAM
24(S)-hydroxycholesterol (5)
GNE-5729 (6)
GluN2A-selective PAM
PYD-106 (7)
GluN2C-selective PAM
Non-selective PAM
Figure 1. PAMs of the NMDAR.
There have been numerous other non-selective and subunit-selective NMDAR PAMs
identified in addition to CIQ (Figure 1). For example, we have previously identified a GluN2B-
preferring class of isopropoxy-substituted tetrahydroisoquinolines (TIQs) based off CIQ
(prototype compound 2).63 Wang et al. published GNE-9278 (3), a non-selective
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 116
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
benzenesulfonamide-based PAM that acts on each of the GluN1/GluN2 subtypes in the low-
6
4
micromolar (3-16 μM) range. Structural determinants suggest a possible binding site near the
GluN1 pre-M1 helix and it was shown to potentiate the receptor by increasing co-agonist affinity,
increasing receptor response at saturating co-agonist concentrations, and slowing deactivation after
glutamate removal. Other non-selective PAMs include the neurosteroid pregnenolone sulfate (PS,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6
5
66
4
) and 24(S)-hydroxycholesterol (5). Genentech also published a GluN2A-selective series of
thiazole derivatives that bind in a similar pocket to the N-(4-(2-benzoylhydrazine-1-
carbonyl)benzyl)-3-chloro-4-fluorobenzenesulfonamide (TCN) class of negative allosteric
modulators (NAMs). These compounds potentiate via stabilization of the agonist-bound state.67, 68
Significant optimization resulted in GNE-5729 (6), a PAM of GluN2A-containing receptors (EC50
=
37 nM) showing improved AMPAR selectivity and unbound clearance and an excellent
secondary pharmacology profile.69 The only known series of compounds able to distinguish
between GluN2C- and GluN2D-containing receptors is the pyrrolidinone-based (prototype PYD-
1
06 (7)) GluN2C-selective PAMs.70, 71 These compounds appear to increase mean open time and
open probability via actions on the ATD and S1 domains.
Although CIQ has made a significant impact as a first-in-class tool compound, it exhibits
several liabilities limiting its use both in vitro and in vivo. These include modest potency, modest
potentiation, low aqueous solubility, and high lipophilicity.58, 59 Herein, we report the discovery of
second-generation dihydropyrrolo[1,2-a]pyrazin-3(4H)-one-based GluN2C/D-selective PAMs
that are the result of successive core changes from CIQ. First, we modified the TIQ scaffold to a
1
,4-dihydroisoquinolin-3(2H)-one core (Figure 2, left), which resulted in improved concentrations
needed to double receptor response (doubling concentration, 0.5 log unit improvement on average)
and lipophilic ligand efficiency (LLE, defined here as LLE = ― log (doubling conc.) – cLogP,
1
0
ACS Paragon Plus Environment

Page 7 of 116
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
0
.4 log unit improvement on average). We have chosen to report the doubling concentration
alongside the EC throughout this manuscript as it captures both the potency and potentiation
5
0
contributions of a PAM’s biological activity in one directly comparable number. Since it considers
both contributions, we believe it to be more useful than EC in describing the biological activity
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
0
of these PAMs, as the EC describes only the potency without information on maximal achievable
5
0
response. We have also chosen to define LLE using the doubling concentration (compared to the
standard definition using EC ) due to its increased relevance. Finally, we have chosen two-fold
5
0
as the cutoff for several reasons: (1) it is currently unknown what percent potentiation of
GluN2C/D-containing receptors will lead to pharmacologically relevant actions, (2) this percent
potentiation was achieved by most compounds studied, (3) it is a convenient and memorable
63
cutoff, and (4) to be consistent with previous reports. Building on this core modification resulted
in a dihydropyrrolo[1,2-a]pyrazin-3(4H)-one series (Figure 2, right), with derivatives showing
roughly log unit improvements in doubling concentration, aqueous solubility, and LLE, and a log
unit decrease in cLogP compared to CIQ. The selectivity for GluN2C and GluN2D is maintained
at sub-saturating concentrations of co-agonists for both new cores. The enantiomers of promising
derivatives were separated via chiral HPLC and show stereo-dependent activity, with one
enantiomer inactive against all examined NMDAR subtypes. The active enantiomers exhibit
improved potency over the racemic mixture in each case. Finally, a crystal structure was solved
for prototypical compound R-(+)-EU-1180-453 that allowed for assignment of the absolute
configuration of the active enantiomers. Therefore, the improvements in activity and
physicochemical properties observed in these two new structural modifications compared to the
parent CIQ have created improved in vitro tool compounds for GluN2C- and GluN2D-containing
NMDA receptors.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 116
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
OMe
OMe
B
B
O
O
R2
R1
R3
R1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
N
N
C
A
C
A
N
O
O
Dihydroisoquinolinone Core
Pyrrolopyrazinone Core
Figure 2. Labeled 1,4-dihydroisoquinolin-3(2H)-one (left) and dihydropyrrolo[1,2-a]pyrazin-
(4H)-one (right) to distinguish between the A-, B-, and C-rings.
3
Two-Electrode Voltage-Clamp (TEVC) Recordings
All compounds synthesized were evaluated via TEVC recordings of Xenopus laevis
oocytes expressing four recombinant NMDAR subtypes: GluN1/GluN2A, GluN1/GluN2B,
GluN1/GluN2C, and GluN1/GluN2D. Each compound was co-applied at 30 M with saturating
concentrations of glutamate (100 µM) and glycine (30 µM). When the mean potentiated response
exceeded 130% of control, concentration-effect curves were generated by co-applying increasing
concentrations of compound with saturating co-agonists. We determined the pEC value and
5
0
maximal degree of potentiation by fitting the concentration-effect curves to eq 1 and eq 2 (see
Experimental Procedures). Each concentration-effect curve was generated with data from 6-19
oocytes from 2-3 frogs; all inactive compounds were tested in 4-10 oocytes from 1-2 frogs. Tables
1
-7 summarize the mean pEC50 value with 95% confidence interval and average maximum
potentiation (max) for each active compound at GluN1/GluN2C and GluN1/GluN2D receptors.
The negative log of the doubling concentration (eq 3 and eq 4, see Experimental Procedures),
aqueous solubility (see Supporting Information for Methods), and LLE (eq 5, see Experimental
Procedures) are also reported when appropriate. Table 1 compares the doubling concentration of
ACS Paragon Plus Environment

Page 9 of 116
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
select compounds reported here to previously reported compounds. For compounds in which a
pEC value was not determined (ND), the maximum potentiation is reported as the average
5
0
potentiation at 30 µM drug as a percent of control. The average percent potentiation is defined as
the mean ratio of current upon application of drug at the stated concentration to the current
response in its absence. The doubling concentration is defined as the concentration of test
compound at which the response to maximally effective concentration of coapplied glutamate and
glycine is increased 2-fold over the response to glutamate and glycine alone. All compounds were
also tested at GluN1/GluN2A and GluN1/GluN2B receptors at 10 or 30 M, and were typically
found to have no effect (4-17 oocytes from 1-3 frogs). Only two compounds (12c-d) exhibited
potentiation above 130% at GluN2A- or GluN2B-containing receptors and neither exceeded 2-
fold potentiation (Supporting Information, Table S1).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Results and Discussion
Design of NMDAR Potentiators
Previous computational studies on a CIQ analogue suggested that potency could be
improved if the A-ring could be directed to adopt an s-trans conformation within the binding site,
with s-cis/s-trans defined by the orientation of the amide carbonyl with respect to the
6
3
stereocenter. To investigate this hypothesis, a novel 1,4-dihydroisoquinolin-3(2H)-one scaffold
was designed and synthesized (Scheme 1). We predicted that translocation of the amide carbonyl
of CIQ into the TIQ ring would provide enough steric influence to favor the s-trans conformation,
and therefore increase the fraction of drug in the predicted more active pose (Figure 3). This
phenomenon can be visualized via a crystal structure of prototype R-(+)-EU-1180-453. The
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 116
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
crystallographic data and methods can be viewed in Table S2 and the Crystal Structure Methods
portion of the Supporting Information, respectively.
OMe
OMe
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Cl
O
O
H
H
H
MeO
MeO
Cl
MeO
MeO
N
N
H
O
O
s-cis conformation of
compound 12m
s-trans conformation of
compound 12m
R-(+)-EU-1180-453 crystal structure
Figure 3. Model of the steric encumbrance upon translocation of the carbonyl in the 1,4-
dihydroisoquinolin-3(2H)-one series. The s-cis conformation was hypothesized to create steric
clash between the carbonyl and A-ring, creating preference for the more active s-trans
conformation (left). This is visualized on this scaffold through a crystal structure of R-(+)-EU-
1
180-453 (right).
Chemistry
The synthesis of 1,4-dihydroisoquinolin-3(2H)-one analogues is shown in Scheme 1.
6
3
Methyl esters 8a and 8b were prepared according to modified literature procedures while all
other similar intermediates were commercially available. Alpha-chloro ketones 9a-d were
prepared via Friedel-Crafts acylation of the appropriate methyl esters with chloroacetyl chloride
and tin(IV) chloride. The resulting α-chloroamide was then reacted with 4-methoxyphenol and
potassium iodide to give ketones 10a-d, which were then cyclized in a microwave reactor using
the appropriate amine, titanium(IV) isopropoxide, and sodium triacetoxyborohydride to afford the
penultimate eneamide intermediates 11a-n as inconsequential mixtures of their E and Z isomers.
Each eneamide could then either be reduced via hydrogenation at 50 psi using palladium on carbon
ACS Paragon Plus Environment

Page 11 of 116
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
or in cases with aryl chlorides, hydrogenated with platinum oxide at 1 atm to give the final 1,4-
dihydroisoquinolin-3(2H)-one compounds (12a-m).
Scheme 1. Synthesis of Dihydroisoquinolinonesa
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
OMe
OMe
OMe
Cl
R2
O
R2
O
O
a
b
d
O
R1
R2
R1
c
O
R1
R2
R1
R2
R1
O
OMe
N
R3
N
R3
O
OMe
O
O
8
a, R =OiPr, R =H
9a, R =OiPr, R =H, 47%
O
OMe
1
2
1
2
8b, R =OiBu, R =H
9b, R =OiBu, R =H, 48%
11a-n
12a-m
1
2
1
2
8
c, R =OMe, R =H
9c, R =OMe, R =H, 46%
9d, R =OMe, R =OMe, 33%
1 2
10a, R1=OiPr, R2=H, 67%
10b, R1=OiBu, R2=H, 52%
10c, R =OMe, R =H, 53%
1
2
1
2
8d, R =OMe, R =OMe
1
2
1
1a/12a, R =OiPr, R =H, R =m-ClPh, 71% then 10%
1 2 3
1
2
11b/12b, R =OiPr, R =H, R =m-CF Ph, 45% then 71%
1
0d, R =OMe, R =OMe, 36%
1
2
3
3
1
2
11c/12c, R =OiPr, R =H, R =m-FPh, 58% then 64%
1
2
3
11d/12d, R =OiBu, R =H, R =m-ClPh, 73% then 54%
1 2 3
1
1e/12e, R =OMe, R =H, R =m-ClPh, 59% then 56%
1
2
3
11f/12f, R =OMe, R =H, R =CH CH CF , 34% then 23%
1
2
3
2
3
3
11g/12g, R =OMe, R =H, R =CF , 19% then 27%
1 2 3 3
1
1h/12h, R =OMe, R =H, R =Et, 40% then 48%
1
2
3
11i/12i, R =OMe, R =H, R =methyl 1H-indol-3-yl, 40% then 45%
1
2
3
11j/12j, R =OMe, R =H, R =CH OCH CH OCH , 19% then 83%
1 2 3 2 2 2 3
1
1k/12k, R =OMe, R =H, R =CH CH NMe , 23% then 83%
1
2
3
2
2
2
11l, R =OMe, R =H, R =
C
CH, 80%
1
2
3
e, 29%
1
1
1m/12l, R =OMe, R =H, R =triazole, 68% for step d
1
2
3
1n/12m, R =OMe, R =OMe, R =m-ClPh, 59% then 22%
1
2
3
a
Reagents and conditions: (a) chloroacetyl chloride, SnCl , DCE, reflux, 90 m; (b) 4-
4
methoxyphenol, KI, K CO , reflux, 4 h; (c) substituted primary amines, Ti(OiPr) , NaBH(OAc)
3,
2
3
4
o
DCE, 120 C, mw, 20 m; (d) (i) PtO , H (balloon), EtOH, rt, 10 h, (ii) Pd/C, H (50 psi), EtOH,
2
2
2
rt, 24 h; (e) NaN , CH O (37% in H O), AcOH, cat. sodium ascorbate, cat. CuSO , THF, rt, 24 h.
3
2
2
4
The procedure for the synthesis of B-ring derivatives is outlined in Scheme 2. The
synthesis began with commercially available 3,4-dimethoxyphenethylamine (13), which was
acylated with chloroacetyl chloride to give α-chloro amide 14. This which was then reacted with
sodium hydride and 4-methoxyphenol to give α-hydroxy amides 15a-d. Intermediates 15a and 15d
were then cyclized to their respective imine intermediate using modified Bischler-Napieralski
conditions and subsequently reduced via sodium borohydride or hydrogenation to yield their
respective TIQ amines (16a in the case of 15a). These TIQ amines were then acylated using the
appropriate benzoyl chloride, deprotected to their respective primary alcohols 17a-b, and further
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 116
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
functionalized using coupling or S Ar conditions to produce the final compounds (18a-c).
N
Alternatively, in the case where the final B-ring substitution was installed in step (b), such as with
15b-c, the respective secondary amine (16b-c) obtained via Bischler-Napieralski cyclization was
directly acylated with 3-(trifluoromethyl)benzoyl chloride to give the final compounds (18d-e).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 2. Synthesis of B-Ring Derivativesa
MeO
MeO
MeO
MeO
MeO
O
O
a
b
Cl
O
MeO
NH2
N
H
N
R1
74%
H
13
14
15a, R1 = Bn, 70%
15b, R1 = 4-tetrahydropyranyl, 75%
1
5c, R1 = 3-pyridinyl, 60%
c, d, e
15d, R1 = 4-methyoxybenzyl, 74%
R1
O
R1
HO
O
O
O
MeO
d, f
MeO
MeO
R2
g
c
NH
N
MeO
MeO
R2
N
MeO
16a, R1 = Bn, 62%
17a: R2 = Cl, 44% (2 steps from 16a)
1
1
6b, R1 = 4-tetrahydropyranyl, 16%
6c, R1 = 3-pyridinyl, 60%
17b: R2 = CF , 8% (3 steps from 15d)
3
18a, R1 = 5-methoxy-2-thiophenyl, R2 = Cl, 21%
18b, R1 = 2-pyridinyl, R2 = CF , 27%
3
1
8c, R1 = 2-pyrimidinyl, R2 = CF , 47%
d
3
18d, R1 = 4-tetrahydropyranyl, R2 = CF , 100% (from 16b)
3
1
8e, R1 = 3-pyridinyl, R2 = CF , 44% (from 16c)
3
a
Reagents and conditions: (a) chloroacetyl chloride, Et N, DCM, 0 °C, 30 m; (b) R OH, NaH or
3
1
Cs CO ; (c) POCl , ACN, reflux then NaBH , MeOH, 0 °C, 10 m; (d) substituted benzoyl chloride,
2
3
3
4
Et N, DCM, rt, 1 h; (e) Pd/C, H (50 psi), MeOH, o/n; (f) methane sulfonic acid, DCM, rt, 6h; (g)
3
2
heteroaryl fluoride, NaH, THF, reflux, o/n, (ii) heteroaryl iodide, 1,10-Phen, CuI, Cs CO , PhMe,
2
3
80 ℃, o/n.
Heterocyclic derivatives of the C-ring were synthesized utilizing a modified synthesis of
5
9
CIQ described in Scheme 3. Ethylamine starting materials 21a-e were synthesized from the
corresponding aldehyde via a Henry-type nitroalkenylation and subsequent reduction via in situ-
generated alane (Scheme 3, step a and step b). 2-Methoxy-5-formylthiophene was synthesized
ACS Paragon Plus Environment

Page 13 of 116
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
72
according to literature precedent. The remaining synthesis proceeded as described with only
slight modifications.59
Finally, pyrrolopyrazinone derivatives were synthesized as described in Scheme 4. This
synthesis is a modified version of the synthesis described above in Scheme 1, where the pyrrole
starting material was made through a Clauson-Kaas reaction starting with glycine methyl ester
hydrochloride.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 3. Synthesis of Heterocyclic C-Ringsa
O
Het
Het
a
Het
Het
b
c
d
O
Cl
NH2
N
NO2
H
19a-e
21a-g
2
2
2
2
2
0a, 79%
22a, 51% (2 steps)
22b, 61% (2 steps)
22c, 5% (2 steps)
22d, 43% (2 steps)
22e, 41% (2 steps)
0b, 73% (2 steps)
0c, 31%
0d, 100%
0e, 62%
22f, 82%
22g, 100%
OMe
O
OMe
2
26
5h
g, 18%
O
Het
O
e
f
N
H
O
OMe
O
R
23a, 84%
23b, 79%
23c, 93%
23d, 87%
23e, 84%
23f, 86%
23g, 89%
N
NH
Het
Het
25a, R = m-ClPh, 33%
2
2
4a, 11%
4b, 19%
24c, 9%
4d, 59%
4e, 38%
24f, 39%
4g, 74%
25b, R = 5-Cl-thiopheneyl, 26%
25c, R =
25d, R = m-ClPh, 36%
25e
CF , 33%
3
2
2
,
R = m-ClPh, 93%
5f, R = m-ClPh, 44%
25g, R = m-ClPh, 96%
2
2
2
2
2
5h, R = m-ClPh, 65%
5i, R = m-ClPh, 87%
5j, R = m-CF Ph, 62%
3
S
S
S
N
O
Het
=
N
MeO
Me
N
N
S
19b-24b,
25d
19c-24c,
25e
19d-24d,
25f
19e-24e,
25g
21f-24f,
25h, 26
21g-24g,
25i-j
1
9a-24a,
25a-c
a
Reagents and conditions: (a) cat. butylamine, cat. AcOH, 4 Å MS, MeNO , reflux, 30 m; (b)
2
LiAlH /H SO , THF, 0 °C to reflux, 5 m; (c) chloroacetyl chloride, Et N, DCM, 0 °C, 30 m; (d)
4
2
4
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 116
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
4
1
-methoxyphenol, Cs CO , o/n, 79-93%; (e) POCl , ACN, reflux, o/n then NaBH , MeOH, 0 °C,
2
3
3
4
0 m; (f) (i) acid chloride, Et N, DCM, rt, 1 h, (ii) carboxylic acid, EDCI, DMAP, rt, o/n; (g)
3
Lawesson’s reagent, PhMe, 150 °C, mw, 2 h.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 4. Synthesis of Pyrrolopyrazinonesa
O
O
Cl
O
ClH3N
a
N
O
b
c
OH
N
O
N
O
O
OMe
68%
83%
100%
OMe
OMe
OMe
26
27
28
29
OMe
OMe
OMe
d
e
f
O
N
O
N
O
N
3
2%
R
R
CF3
N
N
N
O
O
S
3
0a, R = CF , 32%
31a, R = CF , 36%
3
EU-1180-453, R = F, 58%
3
32
30b, R = F, 38%
aReagents and conditions: (a) 2,5-dimethoxytetrahydrofuran, NaOAc, 2:1 v/v AcOH:water, reflux,
h; (b) chloroacetyl chloride, DBN, PhMe, 115 ℃, 4 h; (c) 4-methoxyphenol, KI, K CO , acetone,
4
2
3
rt, o/n; (d) substituted primary amines, Ti(OiPr) , STAB, DCE, 120 ℃, mw, 2 h; (e) Pd/C, H (1
4
2
atm), EtOAc, rt, o/n; (f) Lawesson’s reagent, PhMe, 150 ℃, mw, 2 h.
1
,4-Dihydroisoquinolin-3(2H)-one Compounds Show Improved Doubling Concentrations Over
TIQ Counterparts
As predicted, compound 12m showed improved potentiation (270-340%) of GluN2C/D-
containing NMDAR responses to saturating co-agonists compared to CIQ (220-230%) (Table 1).
Compound 12m also exhibited similar potency (pEC = 5.0-5.3 at GluN2C/D) and subunit-
50
selectivity (no activity at GluN2A/B) to CIQ, resulting in improved doubling concentrations
(p(doubling conc.) = 5.1). Selectivity for GluN2C and GluN2D was also maintained upon
ACS Paragon Plus Environment

Page 15 of 116
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
application of 12m and sub-saturating levels of co-agonists glutamate and glycine (Table S3).
Based on this promising result, we designed and synthesized a targeted library of
dihydroisoquinolinones to demonstrate this trend is maintained irrespective of substitution.
We calculated the doubling concentration using eq 3 for all synthesized 1,4-
dihydroisoquinolin-3(2H)-ones with corresponding TIQs59, 63 to demonstrate the trend of improved
activity for 1,4-dihydroisoquinolin-3(2H)-one analogues (Table 1). The doubling concentration at
GluN2C or GluN2D was improved in the 1,4-dihydroisoquinolin-3(2H)-one scaffold in 92%
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(11/12) of cases. The only case in which it decreased (compare 12d and 88 at GluN2C) resulted in
only a 0.2 log unit drop. The average doubling concentration improvement was 0.5 log units, which
led to a 0.4 log unit improvement in LLE due to the similar cLogP values of the isomeric scaffolds.
This data demonstrates the ability of the 1,4-dihydroisoquinolin-3(2H)-one core to improve
activity over the TIQ core.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 116
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 1. 1,4-Dihydroisoquinolin-3(2H)-one Compounds and TIQ Counterparts
O
B
0
1
2
3
4
5
6
7
8
9
O
X
R2
R1
R3
N
C
A
Y
pEC50 [95% CI] (max)
p(doubling conc.)b
Δp(doubling conc.)
LLE
GluN2D
ΔLLE
a
(%)
#
R
1
R
2
R
3
X
Y
cLogP
GluN2C
GluN2D
GluN2C GluN2D GluN2C GluN2D
GluN2C
GluN2C GluN2D
c
0CIQ
O
-
-
O
-
5.3 (230)
5.3 [5.1-
5.3 (220)
4.9
5.1
5.5
5.6
5.3
6.6
6.4
6.2
6.0
6.3
5.4
6.0
4.5
5.1
5.3
5.7
5.2
6.6
6.3
6.4
5.8
6.3
5.5
5.8
5.6
5.7
5.9
5.9
6.7
6.8
7.3
7.4
6.9
6.9
6.1
6.2
-0.7
-1.1
-0.6
-0.6
-0.2
-1.5
-0.2
-1.0
-1.0
-1.1
-0.6
-0.6
-0.4
OMe
OMe
OiPr
OiBu
OiPr
OiPr
OMe
H
Cl
Cl
Cl
Cl
0.2
0.1
1.3
-0.2
0.3
0.6
0.6
0.4
1.4
0.1
0.5
0.3
0.1
0.1
1.2
-0.3
0.3
0.5
0.5
0.4
1.3
0
1
5.0 [4.9-
5.1] (340)
1
2m
-0.6
-0.4
-0.3
-1.4
-0.2
-0.9
-1.2
-0.9
-0.6
-0.7
-0.2
5
.4] (270)
2
c
3117
O
-
5.9 (240)
6.0 (220)
4
5
62*
7
8
988
5.7 [5.6-
5.6 [5.5-
5.8] (330)
1
2e
O
-
5
.8] (270)
5.7 (240)
6.4 [6.3-
c,d
O
-
5.5 (250)
H
6.3 [6.1-
6.6] (370)
1
2a
O
-
6
.5] (350)
6.0 (450)
6.4 [6.3-
c
O
-
5.9 (520)
H
0
1
2127
6.2 [6.1-
6.3] (370)
1
2d
O
-
6
.4] (280)
5.8 (380)
6.4 [6.2-
c
O
-
5.6 (360)
H
CF
3
0.5
0.2
3
4
5114
6.2 [6.2-
6.3] (310)
1
2b
O
-
6
.6] (280)
5.2 (370)
6.0 [5.9-
c
O
-
5.1 (430)
H
F
6
5.7 [5.6-
5.8] (330)
1
2c
O
6
.0] (320)
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Fitted pEC values are shown to two significant figures when potentiation at 30 µM exceeded
5
0
1
30% of control; values in brackets are the 95% confidence interval for the corresponding fitted
pEC mean value, and values in parentheses are the fitted maximum response as a percentage of
5
0
the initial glutamate (100 µM) and glycine (30 µM) current. Data are from between 8 and 19
b
oocytes from 2-3 frogs for each compound and receptor. p(doubling conc.) is the negative log of
c
59, 63
the doubling concentration. Previously published
data for compounds CIQ, 117, 2*, 88, 127,
and 114 were included for comparison. The numbering corresponds to the compound’s numbering
d
59
in its respective publication. Compound 2* is compound 2 from Santangelo et al.
ACS Paragon Plus Environment

Page 17 of 116
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The modification to a dihydroisoquinolinone core improved doubling concentrations
substantially over the previous TIQ series, however, there were still opportunities to improve the
physicochemical properties of the series (e.g. its high lipophilicity). Prototypical compounds (e.g.
CIQ, 12m, etc.) exhibited cLogP values near 6 (and >8 for the most potent analogues) and aqueous
solubilities in the single-digit micromolar range (<10 µM for CIQ). The highly lipophilic nature
of CIQ is also likely the cause of its low free fraction that limits its concentration and receptor
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6
3
occupancy in vivo, making optimization of this property vital in medicinal chemistry efforts.
Highly lipophilic compounds also typically exhibit liabilities including high clearance,73, 74 low
membrane permeability,75 and low oral bioavailability.76 There have been scattered previous
attempts at decreasing the lipophilicity of the series,59 but in nearly each case this led to a
substantial or complete loss in potency. With this goal in mind, we developed a series of
hydrophilic-focused changes to the three phenyl rings of CIQ, which resulted in the discovery of
a brain penetrant pyrrolopyrazinone-based scaffold with dramatically improved lipophilic
efficiency for GluN2C- and GluN2D-containing NMDARs. Prototypical compounds also display
an approximate order of magnitude improvement in aqueous solubility, a relatively clean off-target
profile, and improved ease of analysis due to elimination of rotamers observed with CIQ. From
this point forward, we assume the SAR from the TIQ series is transferrable to the
dihydroisoquinolinone series. A subset of analogues is provided to demonstrate the ability to
extrapolate between scaffolds (Table S4).
Evaluation of A-Ring Replacements
First, we explored replacements of the A-ring on the dihydroisoquinolinone scaffold aimed
at lowering cLogP using the synthesis described in Scheme 1. Of the alternative substituents
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 116
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
examined, only alkyl trifluoromethyl derivative 12f (p(doubling conc.) = 5.0-5.1) (Table 2) was
tolerated. Despite lowering cLogP values by nearly two log units, 12f still displayed doubling
concentrations in the single-digit micromolar range. Although this result is below desired potency
ranges, this substitution may prove to be useful for future promising analogues to fine-tune ADME
properties due to the relatively favorable change in LLE (LLE = 0.9-1.0). Other A-ring
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
substitutions, such as aliphatic chains (12g, pEC < 4.5 or 12h, max = 94-97%), heterocycles (12l,
50
max = 98-110% or 12i, max = 85%), and heterolinear derivatives (12j, max = 98-100% or 12k,
max = 96-110%) were not tolerated. These results provide evidence hydrophobic effects dominate
receptor-ligand interactions in this portion of the molecule. Therefore, based on previous
optimization,59, 63 meta-Cl or meta-CF substituted benzyl derivatives were synthesized from this
3
point forward almost exclusively since they provided the most promising combination of activity
and lipophilic efficiency.
ACS Paragon Plus Environment

Page 19 of 116
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 2. Hydrophilic A-Ring Substitutionsd
O
B
O
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
X
Y
R1
O
N
R2
C
pEC50 [95% CI] (max)
p(doubling
conc.)
(GluN2C/2D)
a
LLE
(GluN2C/2D)
(%)
#
R
1
R
2
X
Y
cLogP
b
GluN2C
GluN2D
Cl
Cl
CIQc OMe
O
-
-
5.3 (230)
5.3 (220)
4.9/4.5
5.5/5.3
5.6
5.9
-0.7/-1.1
-0.4/-0.6
117c
H
O
5.9 (240)
6.0 (220)
5.1 [4.9-5.3] 4.7 [4.5-4.9]
1
2f
12g
2h
H
H
CF3
-
-
O
O
5.0/5.1
4.1
4.1
0.9/1.0
-
(290)
(450)
CF3
Me
<4.5 (330)
<4.5 (160)
<4.5/ID
1
H
H
-
-
O
O
ND (97)
ND (94)
ND (98)
-
4.0
2.9
-
12l
NH
N
ND (110)
-
-
N
NH
1
2i
H
-
O
ND (85)
ND (85)
-
5.1
-
O
1
2j
H
H
-
-
O
O
ND (98)
ND (100)
ND (96)
-
-
3.1
3.6
-
-
OMe
Me
N
12k
ND (110)
Me
a
Fitted pEC values are shown to two significant figures when potentiation at 30 µM exceeded
5
0
1
30% of control; values in brackets are the 95% confidence interval for the corresponding fitted
pEC mean value, and values in parentheses are the fitted maximum response as a percentage of
50
the initial glutamate (100 µM) and glycine (30 µM) current. For compounds in which a pEC was
50
not determined, max potentiation is reported as percent potentiation at 30 µM drug. Data are from
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 116
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
b
between 6 and 11 oocytes from 2 frogs for each compound and receptor. p(doubling conc.) is the
c
negative log of the doubling concentration. Previously published data for CIQ and 117 were
included for comparison. Single methoxy substitution on the C-ring maintains similar or better
activity to dimethoxy substitution.59 dND = not determined.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Evaluation of Hydrophilic B-Ring Replacements
Next, we examined hydrophilic replacements of the B-ring via the divergent synthesis
described in Scheme 2. Upon TEVC evaluation, none of the derivatives synthesized were tolerated
at any GluN2 subtype. Nitrogen heterocycles such as the 2-pyridinyl (18b), 3-pyridinyl (18e), and
2
9
-pyrimidinyl (18c) derivatives all showed responses when co-applied with agonist that were 88-
7% of control for GluN2C- and GluN2D-containing receptors. Incorporation of saturation (18d),
bioisosteric replacement with thiophene (18a), and removal of the B-ring completely (17b) were
also not tolerated. Para-methoxy phenyl substitution was therefore maintained for subsequent
analogues based on previous optimization.59, 63
ACS Paragon Plus Environment

Page 21 of 116
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 3. Hydrophilic B-Ring Substitutions
R1
O
O
O
O
CF3
N
C
A
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
a
pEC (max) (%)
pEC (max) (%)
50
50
R1
R1
#
cLogP
#
cLogP
GluN2C GluN2D
GluN2C GluN2D
OMe
198b
5.6 (200) 5.6 (220)
5.8
18c
ND (97)
ND (95)
4.3
N
N
O
1
8d
ND (100) ND (95)
ND (100) ND (110)
3.4
1
8b
N
N
ND (96)
ND (90)
ND (97)
ND (88)
5.2
4.8
O
1
8a
5.6
3.2
S
1
8e
1
7b
H
ND (98)
ND (95)
a
Fitted pEC values are shown to two significant figures when potentiation at 30 µM exceeded
5
0
1
30% of control; values in brackets are the 95% confidence interval for the corresponding fitted
pEC mean value, and values in parentheses are the fitted maximum response as a percentage of
50
the initial glutamate (100 µM) and glycine (30 µM) current. For compounds in which a pEC was
50
not determined, max potentiation is reported as percent potentiation at 30 µM drug. Data are from
b
4
oocytes from 1 frog. Previously published data for compound 198 were included for
comparison.59
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 116
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Effect of Bioisosteric Replacement of the C-Ring
Although derivatization of the three rings of CIQ has led to useful tool compounds,59, 63
major changes to the C-ring and TIQ core itself have not been thoroughly explored. We viewed
this as an underdeveloped area of the structure-activity relationship (SAR) in which opportunities
for decreasing lipophilicity and improving affinity for the target protein could emerge, assuming
appropriate choice of a hydrophilic core.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Due to thiophene’s known use as a classical phenyl isostere, tetrahydrothienopyridine 25a
was synthesized (Scheme 3) and its activity against each of the GluN2 subunits (Table 4) was
examined. Compound 25a showed improved maximum potentiation (280-350%) while
maintaining similar potency (pEC = 5.4) and selectivity (no activity at GluN2A/B) to CIQ. Due
5
0
to the improved potentiation, 25a showed improved doubling concentrations and LLE (p(doubling
conc.) = 5.3-5.6 and LLE = -0.3-0.0) in comparison to CIQ (p(doubling conc.) = 4.5-4.9 and LLE
= -0.7‒-1.1). A regioisomer of 25a, compound 25h, also showed improved LLE over CIQ (ΔLLE
=
0.6-0.9) due to improved potentiation (max = 330-340%). Using these results as a starting point,
several tetrahydrothienopyridine-based derivatives were subsequently explored (Table 4, Scheme
). First, bioisosteric 2-chlorothiophene replacement of the A-ring provided compound 25b, which
3
was similarly tolerated (LLE = -0.1-0.2) to 25a (LLE = -0.3-0.0). The trifluorocrotonic acid
derivative 25c suffered a loss in potency (pEC = 4.8-4.9) compared to CIQ, but resulted in a
5
0
large improvement in LLE (LLE = 1.2-1.3) due to its notable decrease in cLogP (cLogP = 3.6).
Similar to 12f, this may be an attractive substitution for fine-tuning the ADME properties of more
potent future derivatives but rendered this compound outside the range of useful potency. Finally,
methoxy substitution on the thiophene (25d) was examined due to previous promising SAR
trends,59, 63 but this compound showed no advantages in terms of LLE over CIQ. However, each
ACS Paragon Plus Environment

Page 23 of 116
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
of these derivatives exemplified changes to the TIQ core were tolerated and we hypothesized other
more hydrophilic substitution may follow a similar trend.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 4. Optimization of A- and C-Ring Functionality for Thiophene Core
O
B
O
O
N
R2
R1
C
pEC50 [95% CI] (max)
p(doubling
conc.)
(GluN2C/2D)
sol.
µM)
LLE
(GluN2C/2D)
a
#
R
1
C-ring
R
2
(%)
cLogP
(
b
GluN2C
GluN2D
Cl
Cl
Cl
CIQc
TIQ core)
5.3 (230)
5.3 (220)
4.9/4.5
8
46
-
5.6
5.6
5.6
5.4
3.6
5.7
-0.7/-1.1
-0.3/0.0
-0.1/-0.2
-0.1/0.2
1.2/1.3
(
S
5.4 [5.3-5.5] 5.4 [5.3-5.5]
(280) (350)
2
5a
H
H
H
H
5.3/5.6
5.5/5.4
5.3/5.6
4.8/4.9
4.6/4.5
5
.4 [5.3-5.4] 5.3 [5.2-5.3]
330) (340)
2
2
5h
5b
(
S
S
S
5
.3 [5.3-5.3] 5.3 [5.2-5.4]
290) (400)
Cl
-
(
S
S
4.9 [4.8-5.0] 4.8 [4.6-4.9]
290) (330)
2
5c
-
CF3
Cl
(
5
.3 [5.2-5.4] 5.2 [5.1-5.2]
220) (220)
25d OMe
-
-1.1/-1.2
(
a
Fitted pEC values are shown to two significant figures when potentiation at 30 µM exceeded
5
0
1
30% of control; values in brackets are the 95% confidence interval for the corresponding log
fitted pEC value and values in parentheses are the fitted maximum response as a percentage of
5
0
the initial glutamate (100 µM) and glycine (30 µM) current. Data are from between 10 and 12
b
oocytes from 2 frogs for each compound and receptor. p(doubling conc.) is the negative log of
c
the doubling concentration. Previously published data for CIQ were included for comparison.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 116
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
SAR of Additional TIQ Core Changes
Several hydrophilic heterocycles were designed and synthesized to test this hypothesis
(Table 5, Scheme 3). The furan derivative, 25f, was the first compound in the series to show a
simultaneous improvement in doubling concentration (p(doubling conc.) = 5.4-5.5) and decrease
in cLogP (cLogP = 5.1) compared to CIQ. This led to a dramatic improvement in LLE (LLE =
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
0
.3-0.4). Even more impressive was pyrrolopyrazine 25i, which exhibited excellent doubling
concentrations (p(doubling conc.) = 5.8-5.9) while also lowering cLogP by 0.7 (cLogP = 4.9)
compared to CIQ. This resulted in an approximate two log unit improvement in LLE (LLE = 1.0-
1
.1). It also maintained the aqueous solubility gains observed with 25a, displaying a 4.5-fold
improvement over CIQ at 36 µM. However, other more hydrophilic derivatives such as the 2-
methylthiazole 25e (max = 100%) and N-methylpyrazole 25g (pEC50 < 4.5) were both not
tolerated. Unsubstituted phenyl derivative 36 (see Supporting Information for synthesis) was also
only weakly active (p(doubling conc.) = 5.0, LLE = -1.0). Replacement of the A-ring amide linker
with a thioamide linker has previously been shown to improve potency >10-fold in certain cases.63
However, this modification was not tolerated in analogue 26 (max = 94-97%), which contains the
tetrahydrothienopyridine scaffold. Nonetheless, compound 25i was an exciting discovery in
identifying compounds with improved lipophilic efficiencies and other derivatives with this same
pyrrolopyrazine core were subsequently developed.
ACS Paragon Plus Environment

Page 25 of 116
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 5. Optimization of Heterocyclic C-Ring Derivativesd
O
B
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
X
Cl
N
C
A
pEC50 [95% CI] (max)
p(doubling
sol.
µM)
LLE
(GluN2C/2D)
a
#
C-ring
X
(%)
conc.)
cLogP
(
b
GluN2C
GluN2D
(GluN2C/2D)
O
CIQc
25f
O
O
O
5.3 (230)
5.3 (220)
4.9/4.5
5.4/5.5
5.9/5.8
8
-
5.6
5.1
4.9
-0.7/-1.1
0.3/0.4
1.0/1.1
O
O
5
.2 [5.1-
5.1 [5.0-5.3]
(470)
5
.3] (340)
5
.6 [5.5-
5.5 [5.4-5.7]
(370)
2
5i
36
N
5
.7] (370)
S
2
5e
O
O
O
S
ND (100)
ND (100)
-
-
-
-
-
4.8
4.2
6.0
6.0
-
N
N
2
5g
< 4.5 (210)
< 4.5 (220)
-
5.0/5.0
-
-
N
5
.2 [5.1-
5.1 [4.9-5.2]
(300)
36
-1.0/-1.0
-
5
.3] (260)
2
6
ND (97)
ND (94)
S
a
Fitted pEC values are shown to two significant figures when potentiation at 30 µM exceeded
5
0
1
30% of control; values in brackets are the 95% confidence interval for the corresponding log
fitted pEC value and values in parentheses are the fitted maximum response as a percentage of
5
0
the initial glutamate (100 µM) and glycine (30 µM) current. For compounds in which a pEC was
50
not determined, max potentiation is reported as percent potentiation at 30 µM drug. Data are from
between 7 and 15 oocytes from 2 frogs for each compound and receptor for which an EC value
5
0
b
was determined and between 4 and 9 oocytes from 1-2 frogs for all other analogues. p(doubling
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 116
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
c
conc.) is the negative log of the doubling concentration. Previously published data for CIQ were
d
included for comparison. ND = not determined.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Pyrrolopyrazine- and Pyrrolopyrazinone-Based Analogues
Based on previously established SAR within the series, a targeted sample of derivatives
was designed and synthesized, keeping the pyrrole core consistent throughout (Table 6). First,
based on previously described SAR,59 the chlorine of compound 25i was substituted for a
trifluoromethyl substituent to give compound 25j (Scheme 3). This compound improved the
doubling concentrations (p(doubling conc.) = 5.6-6.1) into the nanomolar range while also
displaying the highest aqueous solubility (58 µM) of any derivative in the series tested thus far.
To examine whether the LLE boosts seen individually with the pyrrolopyrazine core and
dihydroisoquinolinone core would synergize, a hybridized pyrrolopyrazinone core was designed
and synthesized as described in Scheme 4. The resulting pyrrolopyrazinone (31a) displayed
similar doubling concentrations (p(doubling conc.) = 5.9-6.1) and aqueous solubility (57 µM) to
25j. Fluorine was also well tolerated (EU-1180-453) in place of the trifluoromethyl group.
Although a slight potency drop was observed (pEC = 5.5), it maintained the improvements in
5
0
LLE (LLE = 1.2) and doubling concentrations (p(doubling conc.) = 5.7) seen with 25i, while
improving aqueous solubility nearly an order of magnitude (74 µM) compared to CIQ. EU-1180-
453 is a free-flowing, white solid that is stable as a powder or a DMSO stock for at least 30 days
(
longest time tested) at room temperature under regular atmosphere. Finally, the amide of 31a was
converted to the thioamide to give 32, but this modification resulted in a complete loss of activity
on the pyrrolopyrazinone core, similar as observed for the tetrahydrothienopyridine core above.
ACS Paragon Plus Environment

Page 27 of 116
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 6. Optimization of Potency and Aqueous Solubility for Pyrrolopyrazine- and
Pyrrolopyrazinone-Based Derivativesd
O
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
B
O
X
R
N
C
A
N
Y
pEC50 [95% CI] (max)
p(doubling
conc.)
(GluN2C/2D)
sol.
µM)
LLE
(GluN2C/2D)
a
#
X
Y
R
(%)
cLogP
(
b
GluN2C
GluN2D
CIQc
TIQ core)
Cl
5.3 (230)
5.3 (220)
4.9/4.5
6.1/5.6
5.9/6.1
5.7/5.7
-
8
58
57
74
-
5.6
5.1
5.2
4.5
5.4
-0.7/-1.1
1.0/0.5
0.7/0.9
1.2/1.2
-
(
5
.8 [5.7-5.9] 5.7 [5.5-5.9]
380) (290)
.9 [5.7-6.0] 5.8 [5.6-6.0]
300) (400)
5.5 [5.4-5.6] 5.5 [5.4-5.6]
2
5j
O
-
-
CF
CF
F
3
(
5
3
1a
O
3
3
(
EU-1180-
-
O
S
4
53
(350)
(350)
32
-
CF
ND (120)
ND (110)
a
Fitted pEC values are shown to two significant figures when potentiation at 30 µM exceeded
5
0
130% of control; values in brackets are the 95% confidence interval for the corresponding log
fitted pEC value and values in parentheses are the fitted maximum response as a percentage of
5
0
the initial glutamate (100 µM) and glycine (30 µM) current. For compounds in which a pEC was
50
not determined, max potentiation is reported as percent potentiation at 30 µM drug. Data are from
b
between 7 and 15 oocytes from 2 frogs for each compound and receptor. p(doubling conc.) is the
c
negative log of the doubling concentration. Previously published data for CIQ were included for
d
comparison. ND = not determined.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 116
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Enantiomers
Four compounds with promising profiles (25i, 25j, 31a, and EU-1180-453) were selected
for separation via chiral HPLC to examine the behavior of the individual enantiomers. Upon
separation and isolation, X-ray crystallography studies on (+)-EU-1180-453 determined its
absolute configuration to be the R-(+)-enantiomer and allowed for absolute stereochemical
assignment of the enantiomers.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Consistent with CIQ and its derivatives, the R-(+)-enantiomers were active while the S-
(‒)-enantiomers were not. We selected compound R-(+)-EU-1180-453 for further study because
of its attractive doubling concentrations (p(doubling conc.) = 5.8-5.9) and improved aqueous
solubility (74 µM), exhibiting a >50:1 aqueous solubility to doubling concentration ratio compared
to <1:1 for CIQ. In addition, R-(+)-EU-1180-453 had virtually no effect on GluN1/GluN2A and
GluN1/GluN2B, or a variety of other receptors (Tables S5 and S6). Together, these properties
make this a more attractive and useful tool compound than CIQ.
ACS Paragon Plus Environment

Page 29 of 116
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 7. Stereodependence of Select Pyrrolopyrazine Corese
O
B
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
X
*
R
N
C
A
N
Y
pEC50 [95% CI] (max)
p(doubling
conc.)
sol.
µM)
LLE
(GluN2C/2D)
a
#
X
Y
R
(%)
cLogP
c
(
b
GluN2C
GluN2D
(GluN2C/2D)
CIQd
TIQ core)
Cl
5.3 (230)
5.3 (220)
4.9/4.5
8
5.6
5.2
5.2
4.9
4.9
5.1
5.1
-0.7/-1.1
(
S-(─)-31a
R-(+)-31a
S-(─)-25i
R-(+)-25i
S-(─)-25j
R-(+)-25j
-
O
O
-
CF
CF
3
ND (84)
ND (88)
6.1 [6.0-
-
57
57
36
36
58
58
-
6
.1 [6.0-6.3]
(280)
-
3
6.0/6.2
-
0.8/1.0
-
6.2] (330)
O
O
O
O
Cl
Cl
ND (98)
ND (99)
5.7 [5.7-5.8]
410)
5.7 [5.7-
5.8] (450)
-
6.1/6.2
-
1.2/1.3
-
(
-
CF
3
ND (99)
ND (98)
5
.9 [5.8-6.0]
5.9 [5.8-
5.9] (540)
-
CF
3
6.3/6.4
1.2/1.3
(460)
S-(─)-
EU-1180-
-
-
O
O
F
F
ND (100)
ND (88)
-
74
74
4.5
4.5
-
4
53
R-(+)-
EU-1180-
5.5 [5.4-5.6]
(410)
5.5 [5.4-
5.6] (390)
5.9/5.8
1.4/1.3
4
53
a
Fitted pEC values are shown to two significant figures when potentiation at 30 µM exceeded
50
130% of control; values in brackets are the 95% confidence interval for the corresponding log
fitted pEC value and values in parentheses are the fitted maximum response as a percentage of
5
0
the initial glutamate (100 µM) and glycine (30 µM) current. For compounds in which a pEC was
50
not determined, max potentiation is reported as percent potentiation at 30 µM drug. Data are from
b
between 6 and 13 oocytes from 2 frogs for each compound and receptor. p(doubling conc.) is the
c
negative log of the doubling concentration. Aqueous solubilities reported from racemic mixtures.
d
e
Previously published data for CIQ were included for comparison. ND = not determined.
ACS Paragon Plus Environment