Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders

ll

OPEN ACCESS

Article

Discovery of a Small Molecule Drug Candidate

for Selective NKCC1 Inhibition in Brain

Disorders

Annalisa Savardi, Marco

Borgogno, Roberto

Narducci, ..., Andrea

Contestabile, Marco De Vivo,

Laura Cancedda

marco.devivo@iit.it (M.D.V.)

laura.cancedda@iit.it (L.C.)

HIGHLIGHTS

NKCC1 is a promising target for

the treatment of brain disorders

The newly discovered ARN23746

presents selective NKCC1 versus

NKCC2 and KCC2 inhibition

ARN23746 restores altered

neuronal chloride homeostasis

in vitro

ARN23746 rescues core behaviors

in DS and ASD mice with no

diuretic effect or toxicity

Currently, therapeutic options for several neurological disorders are scant or not

highly effective. This is possibly due to the poor understanding of the mechanisms

underlying these conditions. Here, starting from former validation of the new

pharmacological target NKCC1 in brain disorders, we developed a novel, potent,

and safe NKCC1 inhibitor, able to restore core behaviors in Down syndrome and

autistic mouse models. This compound has the potential to become a solid drug

candidate for the treatment of several neurological conditions.

Savardi et al., Chem 6, 1–24

August 6, 2020 ª 2020 The Authors. Published

by Elsevier Inc.

https://doi.org/10.1016/j.chempr.2020.06.017

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

Discovery of a Small Molecule Drug Candidate

for Selective NKCC1 Inhibition in Brain Disorders

Annalisa Savardi,^1,2,8Marco Borgogno,^3,8Roberto Narducci,¹Giuseppina La Sala,³

Jose Antonio Ortega,³Maria Summa,⁴Andrea Armirotti,⁵Rosalia Bertorelli,⁴Andrea Contestabile,¹

1,6,7,9,

Marco De Vivo,^3,7,and Laura Cancedda

*

SUMMARY

Aberrant expression ratio of Cl^Àtransporters, NKCC1 and KCC2, is

implicated in several brain conditions. NKCC1 inhibition by the FDA-

approved diuretic drug, bumetanide, rescues core symptoms in ro-

dent models and/or clinical trials with patients. However, bumeta-

nide has a strong diuretic effect due to inhibition of the kidney Cl^À

transporter NKCC2, creating critical drug compliance issues and

health concerns. Here, we report the discovery of a new chemical

class of selective NKCC1 inhibitors and the lead drug candidate

ARN23746. ARN23746 restores the physiological intracellular Cl^À

in murine Down syndrome neuronal cultures, has excellent solubility

and metabolic stability, and displays no issues with off-target activ-

ity in vitro. ARN23746 recovers core symptoms in mouse models of

Down syndrome and autism, with no diuretic effect, nor overt

toxicity upon chronic treatment in adulthood. ARN23746 is ready

for advanced preclinical/manufacturing studies toward the ﬁrst sus-

tainable therapeutics for the neurological conditions characterized

by impaired Cl^Àhomeostasis.

The Bigger Picture

In the last few decades, drug

development for brain disorders

has struggled to deliver effective

small molecules as novel

breakthrough classes of drugs.

Discovery of effective chemical

compounds for brain disorders

has been greatly hampered by the

fact that the few currently clinically

used drugs were identiﬁed by

serendipity, and these drugs’

mechanism of action is often

poorly understood. Here, by

leveraging drug repurposing as a

means to quickly and safely

evaluate the new pharmacological

target NKCC1 and its implications

in brain disorders in animal

INTRODUCTION

models and patients, we report an

integrated strategy for the

In recent years, a large and growing body of literature has indicated that intracellular

chloride concentration ([Cl^À]_i) is defective in diverse neurological conditions,^1–3

including Down syndrome (DS)⁴and autism spectrum disorders (ASD).⁵In neurons,

[Cl^À]_iis mainly regulated by the opposing action of the sodium (Na⁺)- potassium

(K⁺)-Cl^Àimporter NKCC1 and the K⁺-Cl^Àexporter KCC2. In agreement with aber-

rant Cl^Àhomeostasis, the NKCC1/KCC2 expression ratio is altered in animal models

of DS,⁶ASD,⁷and several other brain disorders.^1–3In mature neurons, this leads to a

depolarizing (versus hyperpolarizing and inhibitory) GABA neurotransmitter action

through Cl^À-permeable GABA_Areceptors. Accordingly, NKCC1 inhibition by the

FDA-approved diuretic, bumetanide, rescues GABA transmission and behavioral

deﬁcits in the animal models characterized by aberrant Cl^Àhomeostasis.^2,8Notably,

the NKCC1/KCC2 expression ratio is also defective in individuals with DS,⁶and bu-

metanide reduced some of the diagnostic symptoms of ASD in seven independent

clinical studies,^9–15including a phase I and a phase II clinical trial. Moreover, bume-

tanide treatment has shown positive outcomes in clinical case studies of patients

with other neurological/psychiatric conditions.^16–23However, bumetanide has a

strong diuretic effect due to its inhibition of the kidney Cl^Àtransporter NKCC2.

This creates critical issues with drug compliance and health concerns for chronic

treatment,^11–13strongly jeopardizing bumetanide from becoming a solid therapy

for patients. Selective NKCC1 inhibition would be devoid of the diuretic effect,

thus solving these problems.

rational design and discovery of a

novel, selective, and safe NKCC1

inhibitor, active in vivo. This

compound has the potential to

become a clinical drug candidate

to treat several neurological

conditions in patients. Eventually,

this integrated drug-discovery

strategy has the prospective to

revive the appeal of drug-

discovery programs in the

challenging ﬁeld of neuroscience.

Chem 6, 1–24, August 6, 2020 ª 2020 The Authors. Published by Elsevier Inc.

1

This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

Here, we present a new chemical class that selectively inhibits NKCC1 over renal

NKCC2. We designed, synthesized, and tested these new molecular entities

in vitro, and optimized the initial hit compounds into a potent lead inhibitor

(ARN23746). ARN23746 is able to selectively block NKCC1 in a human cell line

and restore the physiological [Cl^À]_iin murine DS neurons in culture. ARN23746

has excellent drug-like properties, with great solubility, metabolic stability, and

target speciﬁcity, while also presenting no overt organ toxicity following in vivo

chronic treatment. Moreover, ARN23746 is characterized by an improved in vivo

pharmacokinetic proﬁle when compared with bumetanide. Importantly,

ARN23746 treatment recovers cognitive deﬁcits in a DS mouse model and rescues

behaviors related to ASD core symptoms in an ASD mouse model, with no diuretic

effect. These results show that selective NKCC1 inhibition devoid of diuretic effect is

able to rescue core diagnostic behaviors in DS and ASD mice. Thus, our study re-

ports the discovery of ARN23746 as a solid drug candidate, which can be developed

into a sustainable therapeutic strategy for DS, ASD, and the other brain disorders

characterized by increased [Cl^À]_iand depolarizing GABAergic transmission.

RESULTS

Design and Synthesis of NKCC1 Novel Inhibitors

To discover and develop selective NKCC1 inhibitors, we ﬁrst sought to identify and

isolate the structural features of bumetanide that could generate selective inhibition

of NKCC1 over NKCC2. NKCC1 inhibition is responsible for bumetanide’s beneﬁcial

central nervous system (CNS) effect. We wanted to distinguish the structural features

responsible for NKCC1 inhibition from the ones responsible for peripheral NKCC2

inhibition in the kidney, which causes the undesirable diuretic effect. We thus de-

signed, synthesized (Schemes S1 and S2; Supplemental Experimental Procedures),

and tested novel bumetanide analogs with diverse substituents in positions R1,

R3, and R5 in bumetanide’s core structure (Figures 1A and S1A).

To test our newly synthesized bumetanide analogs, we developed a functional

NKCC transporter assay (Cl^Àinﬂux assay), based on the detection of variations

of [Cl^À]_iby

a

Cl^À-sensitive, membrane-tagged yellow ﬂuorescent protein

(mbYFPQS).²⁴We performed the Cl^Àinﬂux assay in human HEK293 kidney cells

transfected with NKCC1, NKCC2 (or the corresponding empty plasmid as control-

mock; (Figure S1B), together with the mbYFPQS plasmid. mbYFPQS ﬂuorescence

is inversely dependent on [Cl^À]_i. In this assay, cells are kept in a Cl^À-free-hypotonic

solution and, upon application of NaCl to the assay well, Cl^Àions enter into cells by

NKCC1 or NKCC2 transport, bind mbYFPQS, and determine a decrease in

ﬂuorescence.

¹Brain Development and Disease Laboratory,

Istituto Italiano di Tecnologia, via Morego, 30,

16163 Genoa, Italy

²Universita` degli Studi di Genova, Via Balbi, 5,

16126 Genoa, Italy

³Molecular Modeling and Drug Discovery

Laboratory, Istituto Italiano di Tecnologia, via

Morego, 30, 16163 Genoa, Italy

First, we validated the assay by assessing NKCC1 or NKCC2 functionality upon

application of 74 mM NaCl in the well (Figures S1C and S1E). Then, we assessed

the inhibitory activity (upon NaCl application) of two known unselective inhibitors

of both NKCC1 and NKCC2 (i.e., bumetanide and furosemide; Figures S1D and

S1F), which we used as positive controls. Then, on the newly developed Cl^Àinﬂux

assay, we tested all our 15 synthesized bumetanide analogs. We found that the car-

boxylic acid group, although suboptimal in CNS drug design, in position R1 was

essential for the inhibition of NKCC1 (e.g., ARN21902, Figure S1A). Moreover, small

modiﬁcations of the linear (six to eight carbon atoms) alkyl chain in position R3 (e.g.,

ARN21878, ARN22351, and ARN22381, Figure S1A) and dimethylation of the sul-

fonamide in R5 (e.g., ARN23837, Figure S1A) decreased activity against NKCC2,

while retaining signiﬁcant activity against NKCC1 (Figure 1B). Nevertheless, all the

⁴In Vivo Pharmacology Facility, Istituto Italiano di

Tecnologia, via Morego, 30, 16163 Genoa, Italy

⁵Analytical Chemistry Facility, Istituto Italiano di

Tecnologia, via Morego, 30, 16163 Genoa, Italy

⁶Dulbecco Telethon Institute, Via Orus 2, 35129

Padova, Italy

⁷Senior author

⁸These authors contributed equally

⁹Lead Contact

*Correspondence: marco.devivo@iit.it (M.D.V.),

laura.cancedda@iit.it (L.C.)

https://doi.org/10.1016/j.chempr.2020.06.017

2

Chem 6, 1–24, August 6, 2020

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

Chem 6, 1–24, August 6, 2020

3

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

Figure 1. In Vitro Selection of the Selective NKCC1 Inhibitor ARN23746 as a Lead Compound

(A) Schematic representation of the intervention point in bumetanide’s structure for synthesizing novel bumetanide analogs.

(B) Quantiﬁcation of the inhibitory activity of bumetanide (Bume) and bumetanide analogs (10, 100 mM) in NKCC1-(left) or NKCC2-(right) transfected

HEK293 cells in the Cl^Àinﬂux assay. Data are presented as a percentage of the respective control DMSO. Data represent mean G standard error of the

mean (SEM) from 3–4 independent experiments (Kruskal-Wallis one way ANOVA on Ranks, NKCC1 10 mM: H = 84.898, DF = 6, p < 0.001; NKCC1 100 mM:

H = 86.799, DF=6, p < 0.001; NKCC2 10 mM: H = 40.700, DF = 6, p < 0.001; NKCC2 100 mM: H = 70.569, DF = 6, p < 0.001, Dunn’s post hoc test, *p < 0.05,

**p < 0.01, ***p < 0.001).

(C) Representation of the ligand-based computational strategy to discover novel molecular scaffolds that inhibit NKCC1. The obtained bumetanide

pharmacophore (1) consists of three H-bond acceptor (HBA) features (red spheres), three H-bond donor (HBD) interactions (blue spheres), one

lipophilic feature (green sphere), and one stacking feature (brown sphere) anchored around the central aromatic core. Ligand disposition was then

implemented by superimposing other known unspeciﬁc NKCC1 inhibitors (2), revealing shared features and different dihedral dispositions of

substituent around the central aromatic core. This model was used as a search ﬁlter for the virtual screening (3) of our internal chemical collection

(ꢀ15,000 compounds). Results generated from in vitro testing of the 165 initial hits (4) were then used to retrain the model (5) and perform a second more

speciﬁc screening of our chemical library and commercial chemical libraries (ꢀ135,000 compounds). This iterative computational cycle led to hit

compounds (6) ARN22393 and ARN22394. (D) Quantiﬁcation of the inhibitory activity of the indicated compounds (10, 100 mM) in NKCC1-transfected

HEK293 cells (Cl^Àinﬂux assay). Data are presented as a percentage of the respective control DMSO. Data represent mean G SEM from 3–4 independent

experiments (Kruskal-Wallis one way ANOVA on Ranks, 10 mM: H = 37.119, DF = 3, p < 0.001; 100 mM: H = 33.724, DF = 3, p < 0.001, Dunn’s post hoc test,

*p < 0.05, **p < 0.01, ***p < 0.001).

(E) Chemical structures of NKCC1 inhibitors with novel scaffold. (F) Left, example traces obtained in the Cl^Àinﬂux assay on NKCC1-transfected HEK293

cells for each compound (100 mM). The arrow indicates the addition of NaCl (74 mM) to initiate the NKCC1-mediated Cl^Àinﬂux. Right, quantiﬁcation of

the NKCC1 inhibitory activity of the indicated compounds (10, 100 mM) in experiments such as those on the right. Data are presented as a percentage of

the respective control DMSO. Data represent mean G SEM from 3–4 independent experiments (10 mM: one way ANOVA, F_(4,84)= 33.048, p < 0.001,

Dunnett’s post hoc test, *p < 0.05, ***p < 0.001; 100 mM: Kruskal-Wallis one way ANOVA on Ranks, H = 50.796, DF = 4, p < 0.001, Dunn’s post hoc test,

***p < 0.001.

(G) Left, example traces obtained in the Ca²⁺inﬂux assay on 3DIV hippocampal mouse neurons for each tested compound (100 mM). The arrows

indicates the addition of GABA (100 mM) and KCl (90 mM). Right, quantiﬁcation of the effect of the indicated compounds (10, 100 mM) in the Ca²⁺inﬂux

assay on 3DIV neurons. Data are presented as a percentage of the control DMSO. Data represent mean G SEM from three independent experiments.

10 mM: one-way ANOVA, F_(5,161)= 77.184, p < 0.001, Dunnett’s post hoc test **p < 0.01, ***p < 0.001; 100 mM: Kruskal-Wallis One ANOVA on Ranks, H =

134.681, DF = 5, p < 0.001, Dunn’s post hoc test, ***p < 0.001).

(H) Amplitude change average and single cell data points of GABA-evoked currents obtained by voltage-clamp recordings of 12–20 DIV hippocampal

mouse neurons before (gray example recordings in the inset above) and after (black example traces) bath application of the indicated drugs (10 mM).

Data are presented as mean G SEM (Paired t test, *p < 0.05, **p < 0.01). Scale bars: 250 pA, 250 ms. See also Figures S1A–S1D; Tables S1 and S2.

modiﬁcations that reduced NKCC2 inhibition also led to a signiﬁcant loss of potency

for NKCC1 inhibition (Figure 1B). Altogether, these data indicated a major difﬁculty

to develop new and potent bumetanide derivatives with signiﬁcant selectivity for

NKCC1 over NKCC2. This is in line with the literature on other existing bumetanide

analogs²⁵and prodrugs.²⁶We therefore sought for new molecular entities, structur-

ally unrelated to bumetanide, as selective NKCC1 inhibitors.

Because of the importance of the carboxylic group and the contribution to the selec-

tivity of a linear alkyl chain attached to an aromatic core, we applied a ligand-based

computational strategy to build a pharmacophore model based on bumetanide and

other known NKCC1 inhibitors (Figure 1C). We ﬁrst performed a force ﬁeld-based

conformational search on bumetanide’s structure. This approach returned the

preferred spatial arrangement of the pharmacophoric features of bumetanide,

such as H-bond donor and acceptor, lipophilic, and aromatic groups (Figure 1C).

We then implemented ligand disposition by superimposing this template pharma-

cophore with other structures of unselective NKCC1 inhibitors (e.g., diuretics furose-

mide, asozemide, piretanide, and chlorothiazide) to identify shared chemicophysical

properties and three-dimensional (3D) localization (Figure 1C). We used this phar-

macophore model as a ﬁlter for the virtual screening of our institution’s diverse

and non-redundant internal library of ꢀ20,000 molecules (Figure 1C). In a ﬁrst round

of experiments, 165 new compounds from the library emerged as initial hits and

were individually tested for their ability to inhibit NKCC1 in the Cl^Àinﬂux assay

(not shown). Of these compounds, 20% showed NKCC1 inhibition between

5%–10% at 10 mM. The other 80% showed no activity (not shown). Using the struc-

tural data acquired in the ﬁrst round, we reﬁned the pharmacophore model

4

Chem 6, 1–24, August 6, 2020

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

(Figure 1C). We then performed a second and more speciﬁc screening of the internal

library and other chemical libraries from commercial vendors for a total of ꢀ135,000

compounds (Figure 1C). In this second round, we selected the best 90 compounds

(structures have been deposited to Mendeley Data: https://doi.org/10.17632/

x9ttg84pzt.2) and tested them in the Cl^Àinﬂux assay. Two 2-amino-5-nitro-benzene-

sulfonamide derivatives showed signiﬁcant (although moderate) inhibitory activity

against NKCC1 (ARN22393 29.4 G 2.8% at 100 mM; ARN22394, 17.7 G 3.9% at

10 mM and of 28.6 G 4.3% at 100 mM; Figure 1D). Notably, both compounds had

the sulfonamide moiety and a nitro group (i.e., an isostere of a carboxylic acid),

linked to an aromatic core, although in a different disposition in respect to the other

substituents, as compared with bumetanide.

To improve the potency of the NKCC1 inhibition of the two 2-amino-5-nitro-benze-

nesulfonamide derivatives, we synthesized twelve 2-amino-5-nitro-benzene-sulfon-

amide and thirty-one 4-amino-3-sulfamoyl-benzoic acid new compounds (Scheme

S3; Supplemental Experimental Procedures), which shared the core chemical struc-

ture of the two 2-amino-5-nitro-benzenesulfonamide derivatives. The 2-amino-5-

nitro-benzene-sulfonamides did not exhibit enhanced potency. However, the

4-amino-3-sulfamoyl-benzoic acids showed promising activity against NKCC1 (not

shown). In particular, potent NKCC1 inhibition was displayed by compounds

bearing the carboxylic group in R1, a (methylated) sulfonamide in R3, and an

extended (eight carbon atom) alkyl chain in R4 (i.e., ARN22642 35.4 G 17.3% at

100 mM; ARN22430 70.6 G 9.1% at 100 mM; Figures 1E and 1F). Nevertheless,

ARN22642 and ARN22430 (Figure 1E) showed poor solubility and metabolic

stability (Table S1). To overcome these poor drug-like properties, we manipulated

the most eligible point of metabolism, i.e., the terminal methyl group on the n-octyl

carbon chain. We added a bioisosteric triﬂuoromethyl group (Schemes S3 and S4;

Supplemental Experimental Procedures), which is often used to improve the drug-

likeness of promising compounds. The resulting triﬂuoromethylated analog

(ARN23746; Figure 1E) showed not only a substantially improved solubility and

metabolic stability (Table S1), but also an increased potency (NKCC1 inhibition

31.8 G 5.4% at 10 mM, and 95.2 G 7.6% at 100 mM, Figure 1F).

Encouraged by these results, we next investigated the three most promising

compounds (ARN22642, ARN22430, and ARN23746; Figure 1E) for their NKCC1

inhibition in primary hippocampal mouse neurons. First, we used an indirect assay

(the Ca²⁺inﬂux assay). The Ca²⁺inﬂux assay takes advantage of the endogenous

high expression of NKCC1 in immature neurons, which causes depolarizing actions

of GABA to activate voltage-gated Ca²⁺channels.²⁷In immature neurons, a com-

pound that blocks NKCC1 is, therefore, predicted to inhibit Ca²⁺responses upon

GABA application. We thus used a ﬂuorescent, Ca²⁺-sensitive dye (Fluo4) to monitor

intracellular Ca²⁺variations upon application of GABA (100 mM) in the presence

of ARN22642, ARN22430, ARN2376, bumetanide, furosemide (as positive

controls), or vehicle, in immature neurons (Figure 1G). The tested compounds

signiﬁcantly reduced the Ca²⁺inﬂux upon GABA application in comparison to

DMSO-treated controls, without affecting the cell viability (measured as ﬂuores-

cence level upon KCl [90 mM] application and consequent neuronal depolarization)

(Figure 1G, left). Both bumetanide and ARN23746 displayed a level of alleged

NKCC1 inhibition that corresponded to the levels obtained for the Cl^Àinﬂux assay.

Conversely, furosemide, ARN22642, and ARN22430 showed higher potency

than in the Cl^Àinﬂux assay, at 10 and 100 mM (compare Figure 1G, right with Figures

1F, right and S1D, right). As described for furosemide,²⁸this was possibly due

to direct off-target inhibition of Cl^À-permeable GABA_Areceptors. Indeed,

Chem 6, 1–24, August 6, 2020

5

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

voltage-clamp patch-cell recordings in primary hippocampal mouse neurons

showed that, while furosemide, ARN22642, and ARN22430 exhibited variable inhib-

itory activity on GABA_Areceptors, this was not the case for bumetanide or

ARN23746 (Figure 1H). Based on NKCC1-inhibition potency, metabolic stability,

solubility, and lack of GABA_Ainhibition, we thus selected ARN23746 as a lead com-

pound for further characterization.

First, to directly measure the effect of ARN23746 on [Cl^À]_i, we used primary hippo-

campal cultures obtained from the Ts65Dn mouse model of DS, which is character-

ized by increased [Cl^À]_idue to NKCC1 upregulation.⁶Chloride imaging with the Cl^À-

sensitive ﬂuorescent dye MQAE⁶showed that bumetanide and ARN23746 (10 mM)

both restored aberrant [Cl^À]_ito physiological levels in mature Ts65Dn neurons,⁶

without signiﬁcantly affecting the [ClÀ]_iin wild-type (WT) neurons (Figure 2A).

Next, we used the Cl^Àinﬂux assay to test ARN23746’s NKCC2 inhibition potency

in HEK cells transfected with NKCC2. Notably, ARN23746 (10 mM) did not show sig-

niﬁcant NKCC2 inhibition (À6.9 G 7.8%, Figure 2B), in contrast to the potent NKCC2

inhibition of bumetanide (82.7 G 20.4%, Figure 2B). Finally, we tested ARN23746’s

potency against KCC2 (Figure 2C). To this aim, we used a thallium (Tl) inﬂux assay

based on a Tl-sensitive ﬂuorescent dye²⁹in HEK293 cells transfected with KCC2

or empty plasmid (as control-mock; Figure S1B, right). Cells maintained in Cl^À-

free-hypotonic solution were monitored upon sequential application of TlSO₄

(2 mM) followed by NaCl (74 mM). Upon activation of KCC2 by Cl^Àions, Tl also

enters the cells where it binds a Tl-sensitive dye, thus determining a ﬂuorescence

increase (Figure 2C, left). We validated the assay by assessing KCC2 functionality

(Figure S1G) by application of the known KCC2 inhibitor R-(+)-[(2-n-Butyl-6,7-di-

chloro-2-cyclopentyl-2,3-dihydro-1-oxo-1H-inden-5-yl)oxy]acetic

acid

(DIOA).

DIOA (10 mM) showed signiﬁcant KCC2 inhibition (37.6 G 6.7%, Figure 2C). Notably,

ARN23746 (10 mM) did not show KCC2 inhibition (4.6 G 6.9%, Figure 2C).

Efﬁcacy of the Lead Compound ARN23746

Since ARN23746 displayed an optimal absorption-distribution-metabolism-excre-

tion (ADME) proﬁle in vitro (Table S2), no signiﬁcant off-target activity as agonist

or antagonist for a panel of 47 classical pharmacological targets (Table S3), and

an improved in vivo pharmacokinetic proﬁle in comparison to bumetanide (Figures

2D and 2E), we proceeded to in vivo evaluations of our lead. First, we evaluated the

diuretic effect of ARN23746 and bumetanide (as positive control) in adult (postnatal

day P60–120) Ts65Dn mice and their WT littermates. We measured the urine volume

for 2 h following treatment (0.2 mg kg^À1IP⁶; Figure 3A). As expected, bumetanide

administration signiﬁcantly increased the urine volume in WT and Ts65Dn mice rela-

tive to vehicle-treated mice. Conversely, ARN23746 treatment had no signiﬁcant

diuretic effect in WT or Ts65Dn mice (Figure 3B). This is in agreement with the

lack of signiﬁcant NKCC2 inhibition in vitro. Next, we investigated ARN23746’s efﬁ-

cacy in rescuing cognitive impairment in Ts65Dn mice, as previously tested with bu-

metanide.⁶We evaluated the short-term working memory and the long-term hippo-

campus-dependent explicit memory after chronic (7–28 days) systemic treatment

with ARN23746 (0.2 mg kg^À1IP, daily; Figure 3C). ARN23746 administration

completely rescued the poor short-term working memory of Ts65Dn mice,³⁰which

we assessed in the T-maze task by analyzing the correct choice of the previously un-

explored arm (Figure 3D). Moreover, in the object-location (OL) test, ARN23746

treatment completely rescued the poor spatial memory of Ts65Dn mice to the level

of WT littermates (Figure 3E). Similarly, ARN23746 administration completely

rescued the poor novel-discrimination ability of Ts65Dn mice in the novel-object

recognition (NOR) test (Figure 3F). ARN23746’s effect in the OL and NOR tests

6

Chem 6, 1–24, August 6, 2020

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

Figure 2. ARN23746 Restores [Cl^À]_iin DS Neurons, Does Not Inhibit NKCC2 in HEK Cells and

Shows Improved Brain Penetration In Vivo in Comparison to Bumetanide

(A) Left, representative pseudo-color images (colored scale below) of the [Cl^À]_imeasured with the

MQAE Cl^À-sensitive dye, in WT and Ts65Dn hippocampal neurons at 15 DIV, after treatment with

control DMSO (0.1%) and the indicated compounds (10 mM). Scale bar: 70 mm. Right, quantiﬁcation

of the indicated compounds (10 mM) in modulating [Cl^À]_iin experiments such as those on the left.

Data represent mean G SEM from three independent experiments (two-way ANOVA, F_interaction

_(2,30)= 3.815, p = 0.033, Tukey’s post hoc test, *p < 0.05, **p < 0.01, ***p < 0.001).

(B) Left, example traces obtained in the Cl^Àinﬂux assay on NKCC2-transfected HEK293 cells for

each compound (10 mM). The arrow indicates the addition of NaCl (74 mM) to initiate the NKCC2-

mediated Cl^Àinﬂux. Right, quantiﬁcation of the NKCC2 inhibitory activity in experiments such as

those on the right. Data are presented as a percentage of the respective control DMSO. Data

represent mean G SEM from four independent experiments (Kruskal-Wallis one way ANOVA on

Ranks, H = 16,962, DF = 2, p < 0.001, Dunn’s post hoc test, ***p < 0.001).

(C) Left, example traces obtained in the Tl inﬂux assay on KCC2-transfected HEK293 cells for each

compound (10 mM). The arrows indicate the additions of TlSO₄and NaCl. Right, quantiﬁcation of

the KCC2 inhibitory activity in experiments such as those on the left. Data are presented as a

percentage of the respective control DMSO. Data represent mean G SEM from 4 independent

experiments (one way ANOVA, F

= 10.676, p < 0.001, Dunnett’s post hoc test ***p < 0.001).

(2,51)

(D) Quantiﬁcation of the level of bumetanide and ARN23746 in the brain at diverse time points (5,

15, 30, 60, 120 min) after injection in C57BL/6N mice (BioLASCO Taiwan). (Two-way ANOVA,

F_{treatment (1,50)}= 6.510, p = 0.014, Tukey’s post hoc test, *p < 0.05.).

(E) Quantiﬁcation of the ratio between brain and plasma concentration of bumetanide and

ARN2346 5 min after the injection (two-tailed t test, t = 7.915, p < 0.001). See also Figures S1E–S1G.

Chem 6, 1–24, August 6, 2020

7

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

Figure 3. ARN23746 Rescues Cognitive Impairment in the Ts65Dn Mouse Model of Down

Syndrome with no Diuretic Effect

(A) Schematic cartoon of the experimental protocol for the treatment of adult WT and Ts65Dn mice

to assess the diuretic effect.

(B) Quantiﬁcation of the mean G SEM and single animal cases of the urine volume collected for

120 min after mice were treated with the indicated drugs (two-way ANOVA, F_{treatment (2,69)}= 10.516,

p < 0.001,Tukey’s post hoc test, *p < 0.05, **p < 0.01).

(C) Schematic cartoon of the experimental protocol for the treatment of adult WT and Ts65Dn mice

with ARN23746 for in vivo efﬁcacy assessment of improved cognitive impairment in DS mice.

(D) Top, schematic representation of the T-maze test. Bottom, quantiﬁcation of the mean G SEM

and single animal cases of correct choices in mice treated with the indicated drugs (two-way

ANOVA, F_{interaction (1,46)}= 5.475, p = 0.024, Tukey’s post hoc test, **p < 0.01, ***p < 0.001).

(E) Top, schematic representation of the OL test. Bottom, quantiﬁcation of the mean G SEM and

single animal cases of discrimination index in mice treated with the indicated drugs (two-way

ANOVA, F_{interaction (1,45)}= 15.523, p < 0.001, Tukey’s post hoc test; ***p < 0.001).

8

Chem 6, 1–24, August 6, 2020

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

Figure 3. Continued

(F) Top, schematic representation of the NOR task. Bottom, quantiﬁcation of the mean G SEM and

single animal cases of the discrimination index in mice treated with the indicated drugs (two-way

ANOVA, F_{treatment (1,46)}= 7.154, p = 0.010, Tukey’s post hoc test, * p < 0.05, two-way).

(G) Top, schematic representation of the CFC test. Bottom, quantiﬁcation of the mean G SEM and

single animal cases of the freezing response in mice treated with the indicated drugs (two-way

ANOVA on Ranks, F_{treatment (1,45)}= 4.425, p = 0.041, Tukey’s post hoc test, * p < 0.05, **p < 0.01). See

also Figure S2; Table S4.

was not due to object preference (Table S4) or to alterations in total object explora-

tion (Table S4). Finally, ARN23746 treatment fully restored associative memory in

Ts65Dn mice in the contextual fear-conditioning (CFC) test, as demonstrated by

the rescue of the poor freezing response induced after re-exposure to the training

context 24 h after conditioning (Figure 3G). The rescue of the freezing response

was not induced by altered sensitivity to shock or by changes in non-associative

freezing (Table S4). Notably, 4 weeks of daily treatment with ARN23746 did not

affect the general health of the mice, as evaluated by daily eye assessment, body

weight measures (Figure S2A), and locomotor activity measures (Table S4). More-

over, there was no change in the weight of internal organs (kidneys, liver, spleen,

lungs, and brain; Figure S2B) or in the histopathological proﬁle of liver, kidney,

and spleen (Figure S2C). This indicates that the compound was well-tolerated during

chronic administration.

Finally, to further corroborate ARN23746 as a therapeutic for neurological disorders

with impaired Cl^Àhomeostasis and depolarizing GABA transmission,^1–3we assessed

its efﬁcacy compared with bumetanide in reducing behaviors related to ASD core

symptoms in the valproic acid (VPA) mouse model of ASD.⁷To the best of our knowl-

edge, the literature contains no preclinical data on the efﬁcacy of bumetanide treat-

ment in postnatal mouse models of ASD, although bumetanide has showed positive

outcomes in seven different clinical trials on children and young adults.^9–15As with

the DS mice, we ﬁrst evaluated the diuretic effect of ARN23746 and bumetanide

(as positive control) in young adult (P40–60) VPA mice and controls (CTR) by

measuring the urine volume for 2 h following treatment (0.2 mg kg^À1IP⁶; Figure 4A).

Similarly to DS mice, bumetanide signiﬁcantly increased the urine volume in control

and VPA mice relative to vehicle-treated mice, whereas ARN23746 treatment had no

signiﬁcant diuretic effect in control or VPA mice (Figure 4B).

Then, we evaluated the social interactions upon exposure to a stranger mouse and

the repetitive behaviors (grooming and marble burying) after a chronic (7–21 day)

systemic treatment with ARN23746 or bumetanide (0.2 mg kg^À1IP, daily; Figure 4C;

Figure S3A) in young adult (P40 to P60) VPA mice. ARN23746 completely rescued

the poor social interaction of VPA mice. This occurred for interaction with a never-

met intruder versus an object (expressed as sociability index) and for a novel mouse

versus an already-met mouse (expressed as social novelty index) in the three-cham-

ber test (Figure 4D). Bumetanide also fully rescued the poor social behavior of VPA

mice (Figure S3B). Moreover, ARN23746 rescued poor sociability during male-fe-

male interaction (Figure 4E), as did bumetanide (Figure S3C). Finally, ARN23746

reduced the repetitive behaviors in the marble burying test (Figure 4F) and in the

grooming task (Figure 4G), while bumetanide signiﬁcantly reduced repetitive behav-

iors in the marble burying test (Figure S3D) but showed only a non-signiﬁcant trend

toward WT levels in the grooming task (Figure S3E).

Finally, to increase our study’s translational value, we conducted a parallel investiga-

tion of the expression levels of NKCC1 and KCC2 in the cortex of adult VPA versus

Chem 6, 1–24, August 6, 2020

9

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

10

Chem 6, 1–24, August 6, 2020

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

Figure 4. ARN23746 Rescues Social Deﬁcits and Repetitive Behaviors in the Valproic Acid Mouse Model of Autism, with No Diuretic Effect

(A) Schematic cartoon of the experimental protocol for the treatment of control (CTR) and VPA mice with bumetanide or ARN23746 to the assess the

diuretic effect.

(B) Quantiﬁcation of mean G SEM and single animal cases of the urine volume collection for 120 min after mice were treated with the indicated drugs

(two-way ANOVA on Ranks, F_{treatment (2,63)}= 11.635, p < 0.001, Tukey’s post hoc test, *p < 0.05, **p < 0.01)

(C) Schematic cartoon of the experimental protocol for the treatment of young adult WT and VPA mice with ARN23746 for in vivo efﬁcacy assessment of

improved sociability and repetitive behaviors in ASD mice.

(D) Top, schematic representation of the three-chamber test. Bottom left, quantiﬁcation of the mean G SEM and single animal cases of sociability index

in mice treated with the indicated drugs (two-way ANOVA, F_{condition (1,42)}= 9.575, p = 0.003, Tukey’s post hoc test, **p < 0.01; ***p < 0.001). Bottom

right, quantiﬁcation of the mean G SEM and single animal cases of social novelty index in mice treated with the indicated drugs (two-way ANOVA,

F

_{condition (1,42)}= 8.195, p = 0.007, Tukey’s post hoc test, *p < 0.05, **p < 0.01).

(E) Top, schematic representation of the male-female interaction test. Bottom quantiﬁcation of the mean G SEM and single animal cases of the

interaction time in mice treated with the indicated drugs (two-way ANOVA, F_{treatment (1,42)}= 6.351, p = 0.016, Tukey’s post hoc test, **p < 0.01).

(F) Top, schematic representation of the marble burying test. Bottom, quantiﬁcation of the mean G SEM and single animal cases of the number of

marbles buried by mice treated with the indicated drugs (two-way ANOVA, F_{condition (1,56)}= 6.727, p = 0.012, Tukey’s post hoc test, *p < 0.05, **p < 0.01).

(G) Top, schematic representation of the grooming test. Bottom quantiﬁcation of the mean G SEM and single animal cases of the grooming time for

mice treated with the indicated drugs (two-way ANOVA on Ranks, F_{interaction (1,59)}= 5.019, p = 0.029, Tukey’s post hoc test, *p < 0.05, **p < 0.01).

(H) Top, representative immunoblots for NKCC1 and KCC2 on extracts of membrane-enriched protein fractions from prefrontal cortices of individuals

diagnosed with ASD and controls. Bottom left, quantiﬁcation of KCC2 in samples from individuals with ASD in comparison to age- and sex-matched

controls (two-tailed t test, t = 2.128, p = 0.0492). Bottom right, quantiﬁcation of the NKCC1/KCC2 expression ratio in samples from individuals with ASD

in comparison to age- and sex-matched controls (Mann-Whitney rank sum test, p = <0,001). See also Figures S3 and S4 and Tables S4–S6.

control mice and adult ASD individuals versus age/sex-matched controls (Table S5).

In particular, for human samples, we focused on the prefrontal cortex (PFC), a brain

region possibly involved in social and repetitive behaviors.³¹Compared to controls,

VPA mice had signiﬁcantly lower expression of KCC2 protein in the membrane-en-

riched fraction (Figure S3F). This agrees with studies on hippocampi from VPA

rats.⁷Moreover, we found no signiﬁcant differences in NKCC1 expression between

VPA and control mice (WT: 100 G 21.87; VPA: 104.8 G 27.44). The NKCC1/KCC2

expression ratio was thus signiﬁcantly increased in VPA versus control mice (Fig-

ure S3F). In agreement with these mice data, KCC2 expression in the membrane-en-

riched fraction of the PFC was signiﬁcantly lower in samples from ASD individuals

than from controls (Figure 4H). There were no signiﬁcant differences in NKCC1

expression between ASD or control samples (control: 100 G 28.97; ASD: 109.6 G

43.39). Again, the NKCC1/KCC2 expression ratio was thus signiﬁcantly increased

in ASD PFC samples versus controls (Figure 4H).

DISCUSSION

We discovered and developed a ﬁrst selective NKCC1 inhibitor prompted by recent

reports of the neurological activity of the diuretic bumetanide.²Bumetanide inhibits

NKCC1 and reduces core symptoms in several brain disorders characterized by

aberrant Cl^Àhomeostasis in animal models and in humans.²Nevertheless, bumeta-

nide also inhibits kidney NKCC2, exerting a strong diuretic effect that endangers

drug compliance and may lead to hypokalemia and general ionic imbalance.^12,13

In particular, the chronic administration of bumetanide in autistic subjects during a

phase II clinical trial (3 months treatment)¹²led to a high percentage of subjects pre-

senting several treatment emergent adverse events (TEAEs; e.g., hypokalemia in

30% of the subjects at 0.5 mg bis in die (b.i.d.) bumetanide dosage and in 76% of

the subjects at 2 mg b.i.d. bumetanide dosage). This caused a high percentage of

drop out from the study (33% at the 2 mg b.i.d. dosage). These drawbacks prevent

bumetanide from becoming a viable clinical option compatible with chronic treat-

ment over possibly several years.^6,12Moreover, in the younger population, bumeta-

nide can be dangerous because it can lead to dehydration and asthenia, loss of

appetite,¹³enuresis,¹²encephalopathy in children (due to the general ionic imbal-

ance),³²and potential ototoxicity in neonates (due to the inhibition of both

NKCC1 and NKCC2).^33–35To overcome bumetanide’s limitations as a neurological

Chem 6, 1–24, August 6, 2020

11

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

drug, we report here on the discovery of a new molecular entity, ARN23746, with an

NKCC1/NKCC2 inhibition ratio that is markedly higher than that of bumetanide.

ARN23746 shows no signiﬁcant NKCC2 inhibition in vitro and diuresis in vivo. It is,

therefore, predicted to avoid the general ionic imbalance and to reduce, at least

in part,³⁴the dangerous side effects and ototoxicity, possibly decreasing the dan-

gers of early treatment. Given that most brain defects occurring in DS, ASD, and

other neurodevelopmental disorders originate during development, an early thera-

peutic strategy could be indeed even more beneﬁcial than in adults. Although we

found here that cognitive impairment in DS, social and repetitive behavior in ASD

can be rescued by NKCC1 inhibition in adult animals, chronic treatments are

required.⁶Earlier treatments with ARN23746 during development may result in

ameliorated brain wiring and possibly in the requirement of no (or lower dosage)

treatment in adulthood. Moreover, other disease symptoms, such as increased sus-

ceptibility to seizures and hyperactivity in DS, are not rescued by NKCC1 inhibition in

adult animals.⁶Earlier NKCC1 inhibition by ARN23746 during development may

thus ameliorate brain wiring and possibly rescue these other symptoms too. Howev-

er, since depolarizing/excitatory GABA-mediated responses are essential during

early physiological brain development,^1,3it will be imperative to accurately deter-

mine when blockage of NKCC1 becomes detrimental and ensure that normal devel-

opment is not hindered by premature exposure to ARN23476. In this overall context,

ARN23746 may thus become a sustainable therapeutic for chronic treatment in

adults and possibly also in the younger population.

Another drawback of bumetanide is its suboptimal blood-brain barrier (BBB) pene-

tration. To enhance BBB penetration, some studies have evaluated lipophilic ana-

logs of bumetanide.^36,37However, these studies have no insights about the analogs’

diuretic effect relative to bumetanide or about the establishment of a potential as-

sociation between their activity and factual targeting of NKCC1.³⁶This has sug-

gested the possibility that bumetanide may inﬂuence brain-related behaviors not

by NKCC1 inhibition but by being a diuretic (i.e., through its well-known osmotic

regulation and/or by altering ionic balance).³⁸Here, our study shows that core symp-

toms in DS and ASD mice are rescued by selective pharmacological NKCC1 inhibi-

tion with a compound with a better brain/plasma ratio than bumetanide. This further

strengthens the idea that NKCC1 plays a major role in brain-related behaviors.

Moreover, our pharmacological study rules out the possibility that bumetanide’s

brain-related effects are due to osmotic regulation and/or ionic imbalance through

excessive NKCC2-mediated diuresis.

In addition to DS and ASD, bumetanide rescues core symptoms in mouse models of

several other brain disorders. Most importantly, bumetanide has shown positive out-

comes in patients during clinical trials and case studies of neurodevelopmental

(autism, fragile X, Asperger syndrome, schizophrenia), neurodegenerative (Parkin-

son disease), and neurological disorders (epilepsy).²In particular, in seven indepen-

dent clinical studies, including a phase II clinical trial, bumetanide reduced some of

the diagnostic signs of ASD in children and young adults.^9–15Nevertheless, NKCC1

and KCC2 expression had not been investigated in ASD human brain samples. Here,

we found a lower KCC2 expression in prefrontal cortices of adult human donors

diagnosed with non-genetic forms of ASD. Moreover, bumetanide treatment in ro-

dent models of ASD (i.e., rats exposed in utero to VPA and a mouse model of fragile

X) had only been investigated at the end of gestation, which is not indicative of the

current clinical protocols in patients.⁷Our data show that ARN23746 and bumeta-

nide both rescued the social deﬁcits and repetitive behaviors in young adult animals

of a mouse model of ASD. These ﬁndings offer preclinical evidence in support of the

12

Chem 6, 1–24, August 6, 2020

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

protocol performed in the several clinical trials with ADS individuals treated with bu-

metanide.^9–15We chose the VPA mouse model of ASD with the purpose of trying to

mimic, at least in part, the non-genetic condition of ASD human samples that we had

analyzed with biochemical approaches. In agreement with our ﬁndings in humans,

our biochemistry data in VPA mice showed dysregulation of KCC2 not NKCC1.

This indicates that, no matter which of the two transporters is dysregulated, it is

the NKCC1/KCC2 ratio imbalance that must be targeted to restore the physiological

Cl^Àhomeostasis. This highlights the potential of ARN23746 not only to treat pathol-

ogies or disorders characterized by increased NKCC1 expression (e.g., DS,⁶stroke,

cerebral edema,³⁹glioma⁴⁰) but also all the neurological disorders characterized by

KCC2 downregulation (e.g., autism,⁷Rett syndrome,^41–44Huntington’s disease⁴⁵).

In this respect, a combination therapy of ARN23746 with the newly discovered

KCC2 activator CLP257^46,47(but see ⁴⁸) or KCC2 expression-enhancing com-

pounds⁴⁹could synergize their beneﬁcial effects in improving the core symptoms

of pathologies or disorders characterized by impaired Cl^Àhomeostasis.

Conclusions

This is a time where neuroscience drug discovery is often dismissed in the pharma

industry. This is due to repeated failures in developing small molecules that typically

act on targets validated in animal models only, since target validation in humans is

extremely challenging. In this context, drug repurposing is often proposed as an

alternative strategy. However, this strategy often results in suboptimal care of the

patient, since the repurposed drug still exerts the effects associated with its initial

clinical indication. Here, we propose a new rational strategy for neuroscience drug

discovery that leverages drug repurposing as a tool to safely validate drug targets

shared between patients and animal models, and links it to the computer-aided

drug design of new molecules based on the structure and mechanism of action of

the repurposed drug. In doing so, we have discovered and developed a new com-

pound ARN23746. ARN23746 is a selective NKCC1 inhibitor devoid of diuretic ef-

fects in vivo, displays no issues with off-target activity, has excellent solubility and

metabolic stability in vitro, and does not induce toxicity following chronic treatment

in vivo. This makes it a solid compound candidate ready for advanced preclinical and

manufacturing studies toward development into a clinically relevant drug for un-

precedented sustainable therapeutics in DS, ASD, and possibly the several other

neurological conditions characterized by impaired Cl^Àhomeostasis.

EXPERIMENTAL PROCEDURES

Resource Availability

Lead Contact

Further information and requests for resources and reagents should be directed to

and will be fulﬁlled by the Lead Contact, Laura Cancedda (laura.cancedda@iit.it).

Materials Availability

This study did not generate new unique reagents.

Data and Code Availability

The original data generated during this study have been deposited to Mendeley

Data (https://doi.org/10.17632/x9ttg84pzt.2).

Pharmacophore Generation and Ligand-Based Virtual Screening (LBVS)

The ﬁrst pharmacophore model in this study was built starting from bumetanide and

other NKCC1 inhibitors (i.e., furosemide, azosemide, piretanide, bendroﬂumethia-

zide, benzthiazide, chlorothiazide, metolazone, quinethazone). The nine NKCC1

Chem 6, 1–24, August 6, 2020

13

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

inhibitors were designed with Maestro (Schrodinger suite version 2015-4) and pre-

¨

pared with LigPrep to retrieve the most probable protonation and tautomerization

states at physiological pH. Since the active conformation of NKCC1 inhibitors is un-

known, we decided to use the most energetically favored conformation (i.e., the en-

ergetic minimum) computed via MacroModel. This conformation was used for the

subsequent pharmacophore modeling. For each of the nine compounds, phase soft-

ware was used to obtain the pharmacophore features, such as H-bond donor and

acceptor, lipophilic, and aromatic groups. The ﬁrst pharmacophore model was ob-

tained by removing from bumetanide’s pharmacophore those features that were

shared by the other eight compounds. The ‘generate phase database’ utility in

Phase was used to prepare the internal and commercial molecular libraries for the

LBVS. Subsequently, the libraries were ﬁltered to retain only the molecules that

obey Lipinsky’s rule and that do not bear reactive functional groups. The LBVS

was conducted using Phase’s ‘ﬁnd matches to hypothesis’ utility. In the ﬁrst run

(i.e., screening of the internal database), we used default parameters. In the second

run (i.e., screening of the commercial library), we customized our settings by

˚

applying a strict tolerance value (i.e., < 1 A) for the features that strongly impacted

activity toward NKCC1 (e.g., sulfonamide moiety) and by increasing the tolerance

˚

values (i.e., >1 A) for the remaining features. After visually inspecting the resulting

matches, we selected the structures with the best scores for the in vitro tests.

Compound Synthesis

The detailed synthetic procedures of all compounds are available in Supplemental

Information.

Compound Preparation for In Vitro Testing

All the compounds tested in vitro were dissolved in the assay buffer to a ﬁnal concen-

tration of 10 or 100 mM from a DMSO stock solution of 10 mM. As control, we used

assay buffer supplemented with 0.1% DMSO or 1% DMSO.

HEK Cell Culture and Transfection

HEK293 cells were cultured in Dulbecco’s modiﬁed Eagle medium (DMEM) supple-

mented with 10% fetal bovine serum, 1% L-glutamine, 100 U/mL penicillin, and

100 mg/mL streptomycin, and maintained at 37^ꢁC in a 5% CO₂humidiﬁed atmo-

sphere. To assess NKCC1 and NKCC2 activity (Cl^Àinﬂux assay), 3 million HEK cells

were plated in a 10 cm cell-culture dish and transfected with a transfection mixture

comprising 5 mL of DMEM, 4 mL Opti-MEM, 8 mg of DNA plasmid coding for NKCC1

(PRK-NKCC1 obtained from Medical Research Council and the University of Dun-

dee), NKCC2 (OriGene plasmid #RC216145) subcloned in PRK5 plasmid, or mock

control (empty vector), together with 8 mg of a plasmid coding for the Cl^À-sensitive

variant of the mbYFPQS (Addgene plasmid #80742),²⁴and 32 mL of Lipofectamin

2000. To assess KCC2 activity (Tl inﬂux assay), cells were transfected with 16 mg of

KCC2⁵⁰subcloned in PRK5 plasmid or mock control (empty vector) and 32 mL of Lip-

ofectamin 2000. After 4 h, the cells were collected and plated in 96-well black-

walled, clear-bottomed plates at a density of 2.5 3 10⁴. After 48 h, cells were

used for the Cl^Àor Tl inﬂux assays. All reagents were purchased from Life Technol-

ogies, unless otherwise speciﬁed.

Cl^ÀInﬂux Assay in HEK Cells

Transfected cells were treated with 10 or 100 mM bumetanide and furosemide (as

positive controls), DMSO (see negative control), or with each of our compounds in

100 mL/well of a Cl^À-free-hypotonic solution (67.5 mM Na⁺Gluconate, 2.5mM K⁺

Gluconate, 15mM HEPES pH 7.4, 50 mM Glucose, 1mM Na₂HPO₄, 1 mM NaH₂PO₄,

14

Chem 6, 1–24, August 6, 2020

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

1 mM MgSO₄, 1 mM CaSO₄). After 30 min of incubation, plates were loaded into a

Victor 3V (Perkin Elmer) multi-plate reader equipped with an automatic liquid

injector system, and ﬂuorescence of Cl^À-sensitive mbYFPQS was recorded with exci-

tation at 485 nm and emission at 535 nm. For each well, ﬂuorescence was ﬁrst re-

corded for 20 s of baseline and for 60 s after delivery of a NaCl concentrated solution

(74 mM ﬁnal concentration in assay well). Florescence of Cl^À-sensitive mbYFPQS is

inversely correlated to the intracellular Cl^Àconcentration,²⁴therefore, chloride

inﬂux into the cells determined a decrease of mbYFPQS ﬂuorescence. To represent

the ﬂuorescent traces in time, we normalized the ﬂuorescence value for each time

point to the average of the ﬂuorescence value of the ﬁrst 20 s of baseline (DF/F0).

To quantify the average effects as represented by the bar plots, we expressed the

decrease in ﬂuorescence upon NaCl application as the average of the last 10 s of

DF/F0 normalized traces. Moreover, for each experiment, to account for the contri-

bution of Cl^Àchanges that were dependent on transporters/exchangers other than

NKCC1 or NKCC2, we subtracted the value of the last 10 s of DF/F0 normalized

traces obtained from mock-transfected cells (either control or treated) from the

respective DF/F0 value obtained from the cells transfected with NKCC1 or

NKCC2. We then presented in the ﬁgures all the data as a percentage of the ﬂuores-

cence decrease versus the value of the control DMSO.

Tl Inﬂux Assay in HEK Cells

The Tl inﬂux assay (FluxOR Potassium ion channel assay, Life Technologies) was

modiﬁed from previously published protocols.²⁹In this assay, the amount of Tl⁺

ions (as a substitute for K⁺) entering the cells by KCC2 transport is detected with a

Tl-sensitive ﬂuorogenic dye that increases ﬂuorescence upon Tl⁺binding. Cell

were loaded with a 1:1,000 dilution of Tl-sensitive ﬂuorogenic dye (Component A)

and 1:100 Probenecid (Component D) in 100 mL/well Cl^Àfree-hypotonic solution

(as above). After 1 h, cells were washed twice with 100 mL/well of hypotonic solution

and treated with DIOA (Sigma, positive control), DMSO (as negative control), and

ARN23746 diluted in 200 mL/well hypotonic solution and in the presence of

100 mM Ouabain to inhibit transport by the Na⁺/K⁺-ATPase pump. The plates

were loaded onto a Spark (Tecan) multi-plate reader equipped with a double auto-

matic liquid injector system. Fluorescence of Cl^À-sensitive dye was recorded with

excitation at 485 nm and emission at 535 nm. For each well, ﬂuorescence was ﬁrst

recorded for 20 s of baseline, for another 20 s after TlSO₄delivery (2 mM ﬁnal con-

centration in assay well), and for 60 s after delivery of a NaCl concentrated solution

(74 mM ﬁnal concentration in assay well). To represent the ﬂuorescent traces in time,

we normalized the ﬂuorescence value for each time point to the average of the ﬂuo-

rescence value of the ﬁrst 20 s of baseline (DF/F0). To quantify the average effects as

represented by the bar plots, we expressed the increase in ﬂuorescence upon NaCl

application as the average of the last 10 s of DF/F0 normalized traces after subtract-

ing the average of the last 10 s of DF/F0 normalized traces after TsSO₄injection.

Moreover, for each experiment, to account for the contribution of Tl⁺inﬂux that

was dependent on transporters/exchangers other than KCC2, we subtracted the

value of the last 10 s of DF/F0 normalized traces obtained from mock-transfected

cells (either control or treated) from the respective DF/F0 value obtained from the

cells transfected with KCC2. We then presented in the ﬁgures all the data as a per-

centage of the ﬂuorescence increase versus the value of the control DMSO.

Neuron Cultures

To perform the Ca²⁺inﬂux assay, primary neuronal cultures of dissociated hippo-

campal neurons were prepared from E18 C57BL/6J mice (Charles River), as previ-

ously described,⁵¹and plated in 96-well black-walled, clear-bottomed plates coated

Chem 6, 1–24, August 6, 2020

15

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

with poly-L-lysine (Sigma; 0.1 mg/mL in 100 mM borate buffer, pH 8.5) at a density of

30,000 cells per well. Neurons were maintained in Neurobasal medium supple-

mented with 2% B-27 supplement, 0.5 mM glutamine, 50 U/mL of penicillin, and

50 mg/mL of streptomycin (all from Gibco). The cells were incubated at 37^ꢁC and

5% CO₂until DIV 3 for the Ca²⁺inﬂux assay. To perform the MQAE experiment

and electrophysiology, primary neuronal cultures of dissociated hippocampal neu-

rons were prepared from WT and Ts65Dn pups at postnatal day 2 (P2). Cells were

plated on glass coverslips coated with poly-L-lysine at a density of 250–500 cells/

mm². Neurons were maintained in a culture medium consisting of Neurobasal-A

supplemented with 2% B-27, 1% GlutaMax, and 5 mg/mL gentamycin (all from

Gibco) at 37 ^ꢁC in humidiﬁed atmosphere (95% air, 5% CO₂) until DIV 15 for the

MQAE experiment and electrophysiology.

Ca²⁺Inﬂux Assay in Neuronal Cultures

At 3 DIVs, neurons were loaded with 2.5 mM of the calcium-sensitive day Fluo4-AM

(Invitrogen) in extracellular solution (145 mM NaCl, 5 mM KCl, 10 mM HEPES,

5.55 mM Glucose, 1 mM MgCl2, 2 mM CaCl2, pH 7.4). After 15 min, cells were

washed twice with extracellular solutions and treated with bumetanide or furose-

mide (as a positive control), DMSO (as negative control), or each of our compounds

at 10 or 100 mM, in extracellular solution for 15 min. Plates were then loaded into a

Victor 3V (Perkin Elmer) multi-plate reader equipped with an automatic liquid

injector system and Fluo4 ﬂuorescence was recorded with excitation at 485 nm

and emission at 535 nm. For each well, ﬂuorescence was ﬁrst recorded for 20 s of

baseline and for 20 s after delivery of GABA (100 mM). To evaluate neuronal viability,

ﬂuorescence was recorded for an additional 20 s after delivery of a depolarizing KCl

stimulus (90 mM ﬁnal concentration in wells). To represent the ﬂuorescent traces in

time, we normalized the ﬂuorescence value for each time point to the average of the

ﬂuorescence value for the ﬁrst 20 s of baseline (DF/F0). To quantify the average ef-

fects as represented by the bar plots, we measured the maximum peak increase in

DF/F0 ﬂuorescence upon GABA application normalized over the maximum peak in-

crease in DF/F0 ﬂuorescence upon KCl application. We then presented in the ﬁgures

all the data as a percentage of the ﬂuorescence increase versus the value of the con-

trol DMSO.

Electrophysiological Recordings in Neuronal Cultures

Hippocampal mouse neurons were recorded at DIV12-20 at room temperature

(22^ꢁC–24^ꢁC) in an extracellular solution containing in mM: 145 NaCl, 5 KCl, 10

HEPES, 5.55 Glucose, 1 MgCl2, 2 CaCl₂. Cells were visualized with an upright micro-

scope (Olympus BX51WI) with infrared differential interference contrast optics.

Patch pipettes (3–5 MU) were ﬁlled with an intracellular solution containing in mM:

130 K-Gluconate, 7 KCl, 10 HEPES, 4 MgATP, 0.6 EGTA, 0.3 NaGTP, 10 phospho-

creatine; GABA currents were evoked by keeping the cell’s membrane potential at

À65 mV and pufﬁng 20 mM GABA (10 psi, 20–50 ms) using a Picospritzer III (Parker

Instrumentation). Access resistance was monitored during voltage-clamp record-

ings, and cells with more than 25 MU were excluded. GABA-evoked currents were

obtained by averaging ﬁve sweeps for each experimental condition; we discarded

neurons with variations in GABA-current peaks greater than 20% compared with

the ﬁrst sweep recorded. Signals were sampled at 20 kHz, and low pass ﬁltered at

10 kHz with an Axon Multiclamp 700B (Molecular Devices).

MQAE Cl^ÀImaging in Neuronal Cultures

Cl^Àimaging was performed with the ﬂuorescent Cl^À-sensitive indicator MQAE

(N-(Ethoxycarbonylmethyl)-6-Methoxyquinolinium-Bromide; Molecular Probes)

16

Chem 6, 1–24, August 6, 2020

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

similarly to what was previously described.⁶Neurons at 15 DIV were loaded with

5 mM MQAE in the presence of bumetanide (as positive control), DMSO (as negative

control), or ARN23746 at 10 mM for 30 min at 37^ꢁC. Coverslips were then transferred

to a holding chamber and perfused (2 mL*min^À1) with extracellular solution (NaCl

145 mM, KCl 5 mM, CaCl₂2 mM, MgCl₂1 mM, Hepes 10 mM, D-glucose 5.5 mM,

pH 7.4) in the presence of the tested compound at 25^ꢁC for 5 min before imaging.

Images were acquired with a Nikon A1 scanning confocal microscope equipped with

a 20X air-objective (NA 0.75). MQAE was excited with a 405 nm diode laser, and ﬂuo-

rescence collected with a 525/50 nm band-pass emission ﬁlter. All excitation and

acquisition parameters (laser intensity, PMT offset and gain) were kept constant

throughout the experiments. Image analysis was performed with NIS-Elements soft-

ware (Nikon) by measuring the mean ﬂuorescent intensity of regions of interest

(ROIs) centered on the cell body of single neurons from six randomly selected ﬁelds

for each coverslip. For each experiment, the average ﬂuorescent intensity of all ROIs

from a coverslip was normalized to the average ﬂuorescent intensity of control sam-

ples (WT neurons treated with DMSO) in the same experiment. Pseudo-color images

were generated by ImageJ software (http://rsbweb.nih.gov/ij/).

Animals

All animal procedures were approved by IIT licensing in compliance with the Italian

Ministry of Health (D.Lgs 26/2014) and EU guidelines (Directive 2010/63/EU). A

veterinarian was employed to maintain the health and comfort of the animals.

Mice were housed in ﬁltered cages in a temperature-controlled room with a

12:12 h dark/light cycle and with ad libitum access to water and food. All efforts

were made to minimize animal suffering and use the lowest possible number of an-

imals required to produce statistical relevant results, according to the ‘‘3Rs

concept.’’ In this study, we used Ts65Dn mice maintained in their original genetic

background⁵²by crossing (more than 40 times) Ts65Dn female to C57BL/6JEi x

C3SnHeSnJ (B6EiC3) F1 males (Jackson Laboratories). Ts65Dn mice were geno-

typed by PCR as previously described.⁵³We obtained the VPA mouse model of

ASD by treating pregnant C57BL/6J (Charles River) dams at 12.5 days of pregnancy

with 600 mg/kg (i.p.) of VPA dissolved in PBS. VPA-treated dams give birth to

offspring that exhibit behaviors related to core signs of autism.⁷As control, we

used the offspring of C57BL/6J dams treated at E12.5 with PBS. Offspring aged be-

tween 4 and 16 weeks were used for experiments. For Ts65Dn and VPA mice and

their respective controls, both males (for Ts65Dn and controls n = 48; for VPA and

controls n = 41) and females (for Ts65Dn and controls n = 27; for VPA and controls

n = 28) were used for the analysis of diuresis; only males were used for behavioral

experiments. Both Ts65Dn and VPA mice with their respective controls, were

randomly assigned to vehicle groups (2% DMSO in saline), bumetanide (Sigma,

0.2 mg*Kg^À1body weight), or ARN23746 (0.2 mg*Kg^À1body weight), and treated

daily for 21–28 days by intraperitoneal injection (i.p.). On the day of behavioral

testing, injection was performed 1 h before the task began.

Compound Preparation for In Vivo Experiments

For the in vivo experiments, bumetanide and ARN23746 were dissolved in DMSO in

a stock solution of 1 mg/mL. On the day of the injection, the stock solution was dis-

solved in PBS at a concentration of 0.02 mg*mL^À1, and injected in a volume of

10 mL*g^À1to have a ﬁnal concentration of 0.2 mg*kg^À1

.

Pharmacokinetic Analysis

An external contractor performed the study. Additional information on the study

conditions can be obtained from the contractor’s website (https://www.

Chem 6, 1–24, August 6, 2020

17

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

pharmacologydiscoveryservices.com/). ARN23746 and bumetanide were adminis-

tered i.v. at 2 mg/kg in phosphate-buffered saline (PBS), pH 7.4, with 2% DMSO

to adult C57BL/6N mice (BioLASCO Taiwan). Blood collection was performed by

cardiac puncture at 0, 5, 15, 30, 60, and 120 min. Blood aliquots (300–400 mL)

were collected in tubes coated with lithium heparin, mixed gently, then kept on

ice and centrifuged at 2,500 g for 15 min at 4^ꢁC, within 1 h of collection. The plasma

was then harvested and stored frozen at (À70^ꢁC) until further processing. Immedi-

ately after the blood sampling, mice were decapitated, and the whole brains were

quickly removed, rinsed with cold saline (0.9% NaCl, g/mL), surface vasculature

ruptured, blotted with dry gauze, weighed, and kept on ice until further processing,

within 1 h of collection. Each brain was homogenized in 1.5 mL cold phosphate-buff-

ered saline (PBS), pH 7.4, for 10 s on ice. The brain homogenate from each brain was

then stored at À70^ꢁC until further processing. The plasma and brain samples were

then processed using acetonitrile precipitation and analyzed by LC-MS/MS. Plots

of plasma/brain concentration of the compound over time were constructed. The

fundamental pharmacokinetic parameters of each compound after dosing (AUClast,

AUCinf, T^1/2, Cl, Vz, Vss, Tmax, and C_max) were obtained from the non-compart-

mental analysis (NCA) of the plasma/brain data using WinNonlin.

Diuresis Analysis

Diuresis analysis was performed using mouse metabolic cages (Tecni-

plast,3600m021) equipped with a grid over a funnel and a plastic cone for the sepa-

rate collection of urine and feces. Immediately after i.p. treatment with vehicle, bu-

metanide, or ARN23746 at 0.2 mg*kg^À1, animals were placed inside the metabolic

cages (one animal per cage), where food and water were available ad libitum. After

2 h, mice were returned to their home cages, and the urine volume was measured.

Behavioral Testing

Ts65Dn male mice (8–16 weeks old) and VPA male mice (6–8 weeks old) were tested

after 1 week of treatment with vehicle, ARN23746, or bumetanide (0.2 mg*kg^À1i.p.).

The battery of tests was run over a total period of 21–28 days (four behavioral tests

for Ts65Dn and WT littermates in the following order: NOR, T-maze, NOL, CFC; four

behavioral tests for VPA mice and their controls with the order and modalities

described in Table S6). During the days of behavioral testing, animals were treated

daily with the drug, with tests beginning 1 h after injection. The tasks were video-re-

corded and then analyzed manually by a blind operator. After each trial or experi-

ment, the diverse apparatus and objects were cleaned with 70% ethanol.

T-Maze

The T-maze is a black opaque plastic apparatus with a starting arm and two perpen-

dicular goal arms, each equipped with a sliding door and evenly illuminated by over-

head red lighting (12–14 lux). The T-maze test (spontaneous alteration protocol, 11

trials) evaluates short-term memory by analyzing the correct choice of the unex-

plored arm. The test was performed in similar way to that previously conducted

on Ts65Dn mice.³⁰In each trial, a mouse was ﬁrst placed in the starting chamber

for 20 s. Then, the sliding door was removed, and the animal was free to explore

the apparatus. When the mouse entered (with all four limbs) one of the two goal

arms, the opposite arm was closed with the sliding door. When the mouse (free to

explore the remaining part of the apparatus) returned to the starting area, the pre-

viously closed goal arm was opened. The trial was repeated 11 times. Entry into a

goal arm opposite the one previously chosen was considered a correct choice, while

entry into the previously explored arm was considered an incorrect choice.

18

Chem 6, 1–24, August 6, 2020

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

Alternation score was calculated as the percentage of correct choices (i.e., left-right

or right-left) over the total number of the ten possible alternations.

Object Location Test (OL)

The test evaluates the spatial memory by measuring the ability of mice to recognize

the new location of a familiar object. The test was performed in a gray acrylic arena

(44 3 44 cm), evenly illuminated by overhead red lighting (12–14 lux). Mice were ﬁrst

habituated to the chamber for 15 min on day 1. On day 2, during the acquisition

phase, mice were exposed to two identical objects for 15 min. After 24 h, one of

the two objects was moved during the test session to a novel location, and the

mice were tested for 15 min for their ability to recognize the new location of the ob-

ject. The time spent exploring each object was deﬁned as the number of seconds

during which mice showed investigative behavior (i.e., head orientation, snifﬁng

occurring within < 1.0 cm) or clear contact between the nose and the object. A

discrimination index was calculated as the percentage of time spent investigating

the object in the new location minus the percentage of time spent investigating

the object in the old location [discrimination index = (new object location explora-

tion time/total exploration time 3 100) – (old object location exploration time/total

exploration time 3 100]. As a control, we monitored object preference during the

acquisition phase and the exploration time in the acquisition phase and trial phase.

Of note, during the trial phase, Ts65Dn mice treated with ARN23746 had signiﬁ-

cantly longer exploration times (Table S4), possibly indicative of greater explorative

interest for the objects. Nevertheless, given that the exploration of the single object

was normalized against the total exploration time for each group, the increased

exploration in Ts65Dn mice treated with ARN23746 did not affect the results of

the discrimination index in the test.

NOR Test

The test evaluates long-term object recognition memory by measuring the ability

of mice to recognize a new object with respect to familiar objects. The test was per-

formed in a gray acrylic arena (44 3 44 cm), evenly illuminated by overhead red

lighting (12–14 lux). On day 1, mice were habituated to the arena by freely

exploring the chamber for 15 min. On day 2, during the acquisition phase, mice

were free to explore three different objects (different in color, size, shape, material)

for 15 min. After 24 h, one object from the acquisition phase was replaced with a

novel object, and the mice were tested for 15 min for their ability to recognize the

new object. The time spent exploring each object was deﬁned as the number of

seconds during which mice showed investigative behavior (i.e., head orientation,

snifﬁng occurring within < 1.0 cm) or clear contact between the object and the

nose. The time spent exploring each object, expressed as a percentage of the total

exploration time, was measured for each trial. The discrimination index was calcu-

lated as the difference between the percentages of time spent investigating the

novel object and investigating the familiar objects: discrimination index = (novel-

object exploration time/total exploration time 3 100) À (familiar object exploration

time/total exploration time 3 100). As a control, we monitored object preference

during the acquisition phase and exploration time in the acquisition phase and trial

phase. Of note, both during the acquisition phase and trial phase, Ts65Dn mice

showed a signiﬁcant increase in exploration time (Table S4), possibly indicative

of their hyperactivity. Nevertheless, given that the exploration of the single object

was normalized on the total exploration time for each group, the increased explo-

ration in Ts65Dn did not affect the results on the discrimination index in the test.

Moreover, although there was a preference of ꢀ10% for object B within the

same group in all four groups, the preference was not signiﬁcantly different among

Chem 6, 1–24, August 6, 2020

19

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

the four groups (Table S4). Moreover, although there was a signiﬁcant general

preference for object C in WT versus TS65Dn mice, the preference was not signif-

icantly different among the four single groups as assessed with Tukey post hoc test

(Table S4).

CFC Test

The test evaluates the long-term associative memory by measuring the freezing time

of the animals placed in a location where they had received an adverse stimulus

(electric shock) 24 h earlier. The experiments were performed in a fear-conditioning

system (TSE), which is a transparent acrylic conditioning chamber (23 3 23 cm)

equipped with a stainless-steel grid ﬂoor. Mice were placed outside the experi-

mental room in their home cages before the test and individually transported to

the TSE apparatus in standard cages. Mice were placed in the conditioning chamber,

and they received one electric shock (2 s, 0.75mA constant electric current) through

the ﬂoor grid 3 min later. Mice were removed 15 s after the shock. After 24 h, mice

were placed in the same chamber for 3 min. After 2 h, they were moved to a new

context (black chamber with plastic gray ﬂoor and vanilla odor). The time spent

frozen was scored and expressed as percentage of the total time analyzed.

Three-Chamber Test

The test evaluates the social approach of the tested mouse versus a never-met

intruder in comparison to an object (sociability) or versus a novel never-met intruder

in comparison to the already-met intruder (social novelty). It was performed in a

similar way to that previously described for mouse models of ASD.⁵⁴The three-cham-

ber apparatus comprises a rectangle, three-chambered box of gray acrylic, evenly

illuminated by overhead red lighting (12–14 lux). The chambers are accessible by

rectangle openings with sliding doors. In the ﬁrst 10 min (habituation), the tested

mouse was free to explore the apparatus of two inverted stainless-steel wire pencil

cups (one in each of the two side chambers), with a weighted plastic cup on top to

prevent the mouse climbing on the top. Then, the tested mouse was brieﬂy conﬁned

in the center chamber, while a never-met intruder (previously habituated to the

apparatus) was placed in one of the side chambers, under the pencil cup. For the

following 10 min (sociability test), the tested mouse was allowed to explore all three

chambers. Then, the tested mouse was again brieﬂy conﬁned in the center chamber,

while a novel never-met intruder (previously habituated to the apparatus) was placed

in the other side chamber under the pencil cup. For the following 10 min (social nov-

elty test), the tested mouse was allowed to explore all the three chambers. The

time spent exploring the object or the intruder was calculated by measuring the num-

ber of seconds when the mice showed investigative behavior (i.e., head orientation,

snifﬁng occurring within < 1.0 cm). The sociability index was calculated as the differ-

ence between the time spent investigating the never-met intruder and the time spent

investigating the familiar object divided by the total exploration time: sociability in-

dex = (never-met intruder exploration time À object exploration time) / (never-met

intruder exploration time + object exploration time). The social novelty index was

calculated as the difference between the time spent investigating the never-met

intruder and the time spent investigating the already-met intruder divided by the

total exploration time: social novelty index = (never-met intruder exploration

time À already-met intruder exploration time) / (never-met intruder exploration

time + already-met intruder exploration time).

Male-Female Interaction Test

The test evaluates the social approach of the tested mouse versus a never-met fe-

male mouse. It was performed in a similar way to that previously described for mouse

20

Chem 6, 1–24, August 6, 2020

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

models of ASD.⁵⁵Brieﬂy, the tested mouse was placed alone in a cage

(26 3 48 3 20 cm). After 5 min of habituation, an unfamiliar C57BL/6J female mouse

was placed in the home cage of the isolated male mouse, and behavior was recorded

for 5 min. The time spent interacting with the female intruder was calculated by

measuring the seconds when the mice showed any of the following interacting

behavior: anogenital snifﬁng, body snifﬁng, head snifﬁng, following, mounting.

Marble Burying Test

The test evaluates the repetitive behavior as the tendency to dig and bury marbles in a

novel environment. It was performed as previously described in ASD mice.⁵⁴Brieﬂy,

mice were placed alone in a cage (26 3 48 3 20 cm) ﬁlled with 4 cm fresh mouse

bedding material, evenly illuminated by overhead red lighting (12–14 lux). After

30 min habituation, the mouse was brieﬂy removed, and bedding surface was leveled

to place 15 glass marbles equidistantly distributed in a 3 3 5 arrangement. Then, the

mouse was returned to the cage and allowed to explore for 30 min. Buried marbles

were counted as those marbles that were covered by at least two-thirds.

Grooming

The test evaluates repetitive behavior in mouse grooming of all body parts. It was per-

formed as previously described in ASD mice.⁵⁶Self-grooming was deﬁned as licking or

scratching the head or body parts with any of the forelimbs. Brieﬂy, mice were placed

individually into a clear Plexiglass cylinder (30 cm high, 10 cm wide). After 10 min habit-

uation, the mouse was scored for 5 min for the time spent self-grooming in all body re-

gions. For behavioral experiments, we adopted the following exclusion criteria inde-

pendent of genotype or treatment (before blind code was broken). In the T-maze

test, we excluded mice that did not conclude the 10 trials within 20 min of the test.

In the CFC test, we excluded mice showing very high non-associative freezing in the

new context. This was deﬁned as more than 30 s freezing during the 3-min test. In

the OL and NOR test, we excluded animals showing very low explorative behavior.

This was deﬁned as less than 10 s of direct object exploration during the 15-min

test. Following these criteria, a total of 5 mice among the NOR, NOL, and T-maze tests

were excluded. Original data used to calculate the discrimination index (for NOR and

OL), sociability index, and social novelty index (for three-chamber test) have been

deposited to Mendeley Data (https://doi.org/10.17632/x9ttg84pzt.2).

Human Brain Samples

PFC samples (Brodmann area 9) from adult humans with ASD and age- and sex-

matched controls were obtained from the NeuroBioBank (NIH). Information on the

samples is reported in Table S5. Samples were cryo-pulverized in dry ice (À78^ꢁC)

with a stainless-steel mortar. Aliquots of pulverized tissue were used for protein

extraction.

Biochemistry

Cells were lysated in RIPA buffer (1% NP40, 0.5% Deoxycholic acid, 0.1% SDS,

150 mM NaCl 1 mM EDTA, 50 mM Tris, pH 7.4) containing 1% (v/v) protease and

phosphatase inhibitor cocktails (Sigma). After 30 min incubation in ice, the samples

were clariﬁed by centrifugation at 20,000 3 g for 20 min at 4^ꢁC. Brain samples from

VPA and WT mice and from ASD people and controls were processed to obtain ex-

tracts of membrane-enriched fractions, as previously described.⁶Samples were ho-

mogenized in ice-cold homogenization buffer (320 mM sucrose, 1 mM EDTA, 10 mM

Tris, pH 7.4) supplemented with 1 mM PMSF, 10 mM NaF, 2 mM sodium orthovana-

date, and 1% (v/v) protease and phosphatase inhibitor cocktail (Sigma). Homoge-

nates were ﬁrst centrifuged at 800 g to remove nuclei and debris. The resulting

Chem 6, 1–24, August 6, 2020

21

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

supernatant (S1) was further centrifuged at 9,200 g. The resulting pellet (P2), repre-

senting the membrane-enriched fraction, was extracted in RIPA buffer (as above) and

used for western blot. The protein concentration of samples was determined with

BCA kit (Pierce). To perform immunoblot, protein extracts were diluted in lithium-

dodecyl-sulfate (LDS) sample buffer (ThermoFisher Scientiﬁc), with the addition of

50 mM DTT. To avoid NKCC1 and KCC2 protein precipitation or aggregation, all

samples were warmed at 40^ꢁC for 5 min.⁶Equivalent amounts of protein (20 mg)

were loaded on the 4%–12% Bis-Tris NuPAGE pre-cast gels (Invitrogen), and elec-

trophoresis were performed with MOPS buffer (Life Technologies). Next, gels

were transferred overnight at 4^ꢁC onto nitrocellulose membranes (GE Healthcare)

with Tris-Glycine transfer buffer (25 mM Tris-base, 192 mM glycine, 20% Methanol).

Equal amounts of protein loading was veriﬁed by staining with 0.1% Ponceau. Mem-

branes were blocked for 1 h in 5% milk in Tris-buffered saline (10 mM Tris, 150 mM

NaCl, pH 8.0) plus 0.1% Tween-20 and incubated overnight at 4^ꢁC with primary an-

tibodies: rabbit anti-actin (Sigma, catalog n^o: A2066; 1:10,000), mouse anti-NKCC1

(clone T4c, Developmental Studies Hybridoma Bank; 1:4,000), rabbit anti-KCC2

(Millipore catalog n^o: 07-432; 1:4000), sheep anti-NKCC2 (MRC-PPU, 1:2,000),

and mouse anti-Na+/K+-ATPase (clone C464.6, Millipore catalog no. 05-369;

1:2,000). Next, membranes were washed and incubated for 2 h at room temperature

with HRP-conjugated goat secondary antibodies (ThermoFisher Scientiﬁc;

1:10,000). Membranes were developed with SuperSignal West Pico chemilumines-

cent substrate (ThermoFisher Scientiﬁc). The chemiluminescent signals were ac-

quired on the LAS 4000 Mini imaging system (GE Healthcare). Bands were later

quantiﬁed by measuring the mean intensity of the band signal using ImageQuant

software (GE Healthcare). Uncropped blots can be found in Figure S4. For KCC2

analysis, we considered the monomeric form, both in human and in mouse samples.

Statistical Analysis

The results are presented as the means G SEM. The statistical analysis was per-

formed using SigmaPlot 13.0 (Systat) software. Where appropriate, the statistical

signiﬁcance was assessed using the following parametric test: Student’s t test,

one-way ANOVA followed by Dunnet post hoc test, two-way ANOVA followed by

all pairwise Tukey post hoc test. Where normal distribution or equal variance as-

sumptions were not valid, statistical signiﬁcance was evaluated using Mann-Whitney

Rank Sum Test, Kruskal-Wallis One Way ANOVA with Dunn’s post hoc test, or two-

way ANOVA on ranks followed by all pairwise Dunn’s post hoc test. p values < 0.05

were considered signiﬁcant. Outliers were excluded only from the ﬁnal pool of data

by a Grubb’s test run iteratively until no outliers were found.

SUPPLEMENTAL INFORMATION

Supplemental Information can be found online at https://doi.org/10.1016/j.chempr.

2020.06.017.

ACKNOWLEDGMENTS

This work was supported by Telethon (grant TCP15021 to L.C). This project received

partial funding from the European Research Council (ERC) under the European

Union’s Horizon 2020 Research And Innovation Program (grant agreement no

725563 to L.C.). Brain samples from individuals with ASD and controls were obtained

from the NeuroBioBank (NIH). We thank Marina Nanni (IIT, NBT) for technical sup-

port, the staff of the IIT animal facility, Silvia Venzano, Giuliana Ottonello, and

Sine Mandrup Bertozzi (IIT, D3) for technical assistance.

22

Chem 6, 1–24, August 6, 2020

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

AUTHOR CONTRIBUTIONS

A.S. designed and performed in vitro assays, diuresis experiment, behavioral exper-

iments, western blot, and analyzed data; M.B. and J.A.O. designed and synthesized

the compounds; R.N. collected and analyzed the electrophysiology data; G.L.S per-

formed the virtual library screening; M.S. and R.B. performed the histopathological

analysis; A.A. coordinated bioanalytical experiments; A.C. designed the in vitro as-

says and performed the MQAE assay; A.S., M.B., A.C., M.D.V., and L.C. designed

the experiments and wrote the manuscript; all authors read and revised the

manuscript.

DECLARATION OF INTERESTS

A.C. and L.C. are named as co-inventors on the following granted patent: US

9,822,368; EP 3083959; JP 6490077; A.C. and L.C. are named as co-inventors on

the patent application WO 2018/189225. A.S., M.B., J.A.O., A.C., M.D.V., and

L.C. are named as co-inventors on patent application IT 102019000004929.

Received: February 13, 2020

Revised: April 14, 2020

Accepted: June 11, 2020

Published: July 10, 2020

REFERENCES

˚

1. Deidda, G., Bozarth, I.F., and Cancedda, L.

(2014). Modulation of GABAergic transmission

in development and neurodevelopmental

disorders: investigating physiology and

pathology to gain therapeutic perspectives.

Front. Cell. Neurosci. 8 , 119.

9. Hadjikhani, N., Asberg Johnels, J., Lassalle, A.,

16. Grandgeorge, M., Lemonnier, E., Degrez, C.,

and Jallot, N. (2014). The effect of bumetanide

treatment on the sensory behaviours of a

young girl with Asperger syndrome. BMJ Case

Rep. 2014 .

Zurcher, N.R., Hippolyte, L., Gillberg, C.,

¨

Lemonnier, E., and Ben-Ari, Y. (2018).

Bumetanide for autism: more eye contact, less

amygdala activation. Sci. Rep. 8 , 3602.

10. Hadjikhani, N., Zurcher, N.R., Rogier, O., Ruest,

¨

17. Lemonnier, E., Lazartigues, A., and Ben-Ari, Y.

(2016). Treating schizophrenia with the diuretic

bumetanide: a case report. Clin.

T., Hippolyte, L., Ben-Ari, Y., and Lemonnier, E.

(2015). Improving emotional face perception in

autism with diuretic bumetanide: a proof-of-

concept behavioral and functional brain

imaging pilot study. Autism 19 , 149–157.

2. Ben-Ari, Y. (2017). NKCC1 chloride importer

antagonists attenuate many neurological and

psychiatric disorders. Trends Neurosci. 40 ,

536–554.

Neuropharmacol. 39 , 115–117.

18. Lemonnier, E., Robin, G., Degrez, C., Tyzio, R.,

Grandgeorge, M., and Ben-Ari, Y. (2013).

Treating fragile X syndrome with the diuretic

bumetanide: a case report. Acta Paediatr. 102 ,

e288–e290.

3. Schulte, J.T., Wierenga, C.J., and Bruining, H.

(2018). Chloride transporters and GABA

polarity in developmental, neurological and

psychiatric conditions. Neurosci. Biobehav.

Rev. 90 , 260–271.

11. Lemonnier, E., and Ben-Ari, Y. (2010). The

diuretic bumetanide decreases autistic

behaviour in ﬁve infants treated during

3 months with no side effects. Acta Paediatr.

99 , 1885–1888.

19. Damier, P., Hammond, C., and Ben-Ari, Y. (2016).

Bumetanide to treat Parkinson disease: a report

of 4 cases. Clin. Neuropharmacol. 39 , 57–59.

4. Contestabile, A., Magara, S., and Cancedda, L.

(2017). The GABAergic hypothesis for cognitive

disabilities in Down syndrome. Front. Cell.

Neurosci. 11 , 54.

12. Lemonnier, E., Degrez, C., Phelep, M., Tyzio, R.,

Josse, F., Grandgeorge, M., Hadjikhani, N.,

and Ben-Ari, Y. (2012). A randomised

20. Eftekhari, S., Mehvari Habibabadi, J., Najaﬁ

Ziarani, M., Hashemi Fesharaki, S.S.,

controlled trial of bumetanide in the treatment

of autism in children. Transl. Psychiatry 2 , e202.

Gharakhani, M., Mostafavi, H., Joghataei, M.T.,

Beladimoghadam, N., Rahimian, E., and

Hadjighassem, M.R. (2013). Bumetanide

reduces seizure frequency in patients with

temporal lobe epilepsy. Epilepsia 54 , e9–e12.

5. Cellot, G., and Cherubini, E. (2014). GABAergic

signaling as therapeutic target for autism

spectrum disorders. Front. Pediatr. 2 , 70.

13. Lemonnier, E., Villeneuve, N., Sonie, S., Serret,

S., Rosier, A., Roue, M., Brosset, P., Viellard, M.,

Bernoux, D., Rondeau, S., et al. (2017). Effects

of bumetanide on neurobehavioral function in

children and adolescents with autism spectrum

disorders. Transl. Psychiatry 7 , e1056.

6. Deidda, G., Parrini, M., Naskar, S., Bozarth, I.F.,

Contestabile, A., and Cancedda, L. (2015).

Reversing excitatory GABAAR signaling

restores synaptic plasticity and memory in a

mouse model of Down syndrome. Nat. Med.

21 , 318–326.

21. Rahmanzadeh, R., Eftekhari, S., Shahbazi, A.,

Khodaei Ardakani, M.R., Rahmanzade, R.,

Mehrabi, S., Barati, M., and Joghataei, M.T.

(2017). Effect of bumetanide, a selective NKCC1

inhibitor, on hallucinations of schizophrenic

patients; a double-blind randomized clinical

trial. Schizophr. Res. 184 , 145–146.

14. Du, L., Shan, L., Wang, B., Li, H., Xu, Z., Staal,

W.G., and Jia, F. (2015). A pilot study on the

combination of applied behavior analysis and

bumetanide treatment for children with autism.

J. Child Adolesc. Psychopharmacol. 25 ,

585–588.

7. Tyzio, R., Nardou, R., Ferrari, D.C., Tsintsadze,

T., Shahrokhi, A., Eftekhari, S., Khalilov, I.,

Tsintsadze, V., Brouchoud, C., Chazal, G., et al.

(2014). Oxytocin-mediated GABA inhibition

during delivery attenuates autism pathogenesis

in rodent offspring. Science 343 , 675–679.

22. Pressler, R.M., Boylan, G.B., Marlow, N.,

Blennow, M., Chiron, C., Cross, J.H., de Vries,

15. Bruining, H., Passtoors, L., Goriounova, N.,

Jansen, F., Hakvoort, B., de Jonge, M., and Poil,

S.S. (2015). Paradoxical benzodiazepine

response: a rationale for bumetanide in

neurodevelopmental disorders? Pediatrics

136 , e539–e543.

L.S., Hallberg, B., Hellstrom-Westas, L., Jullien,

¨

V., et al. (2015). Bumetanide for the treatment

of seizures in newborn babies with hypoxic

ischaemic encephalopathy (NEMO): an open-

label, dose ﬁnding, and feasibility phase 1/2

trial. Lancet Neurol. 14 , 469–477.

8. Kharod, S.C., Kang, S.K., and Kadam, S.D.

(2019). Off-label use of bumetanide for brain

disorders: an overview. Front. Neurosci. 13 , 310.

Chem 6, 1–24, August 6, 2020

23

Please cite this article in press as: Savardi et al., Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders,

Chem (2020), https://doi.org/10.1016/j.chempr.2020.06.017

ll

OPEN ACCESS

Article

23. Kahle, K.T., Barnett, S.M., Sassower, K.C., and

Staley, K.J. (2009). Decreased seizure activity in

a human neonate treated with bumetanide, an

inhibitor of the Na(+)-K(+)-2Cl(À ) cotransporter

NKCC1. J. Child Neurol. 24 , 572–576.

disorders but not newborn seizures. Front. Cell. 46. Gagnon, M., Bergeron, M.J., Lavertu, G.,

Neurosci. 10 , 90.

Castonguay, A., Tripathy, S., Bonin, R.P., Perez-

Sanchez, J., Boudreau, D., Wang, B., Dumas, L.,

et al. (2013). Chloride extrusion enhancers as

novel therapeutics for neurological diseases.

Nat. Med. 19 , 1524–1528.

36. Brandt, C., Seja, P., Tollner, K., Romermann, K.,

¨

Hampel, P., Kalesse, M., Kipper, A., Feit, P.W.,

Lykke, K., Toft-Bertelsen, T.L., et al. (2018).

Bumepamine, a brain-permeant benzylamine

derivative of bumetanide, does not inhibit

NKCC1 but is more potent to enhance

phenobarbital’s anti-seizure efﬁcacy.

24. Watts, S.D., Suchland, K.L., Amara, S.G., and

Ingram, S.L. (2012). A sensitive membrane-

targeted biosensor for monitoring changes in

intracellular chloride in neuronal processes.

PLoS One 7 , e35373.

47. Gagnon, M., Bergeron, M.J., Perez-Sanchez, J.,

Plasencia-Fern a ´n dez, I., Lorenzo, L.E., Godin,

A.G., Castonguay, A., Bonin, R.P., and De

Neuropharmacology 143 , 186–204.

Koninck, Y. (2017). Reply to the small molecule

CLP257 does not modify activity of th e K +-Cl- co-

transporter KCC2 but does potentiate GABAA

receptor activity. Nat. Med. 23 , 1396–1398.

37. Auer, T., Schreppel, P., Erker, T., and

Schwarzer, C. (2020). Functional

25. Lykke, K., Tollner, K., Feit, P.W., Erker, T.,

¨

MacAulay, N., and Loscher, W. (2016). The

¨

characterization of novel bumetanide

derivatives for epilepsy treatment.

Neuropharmacology 162 , 107754.

search for NKCC1-selective drugs for the

treatment of epilepsy: structure-function

relationship of bumetanide and various

bumetanide derivatives in inhibiting the human

cation-chloride cotransporter NKCC1A.

Epilepsy Behav. 59 , 42–49.

48. Cardarelli, R.A., Jones, K., Pisella, L.I., Wobst,

H.J., McWilliams, L.J., Sharpe, P.M., Burnham,

M.P., Baker, D.J., Chudotvorova, I., Guyot, J.,

et al. (2017). The small molecule CLP257 does

not modify activity of the K+-Cl- co-

38. Hochman, D.W. (2012). The extracellular space

and epileptic activity in the adult brain:

explaining th e a ntiepileptic effects of furosemide

and bumetanide. Epilepsia 53 , 18–25.

transporter KCC2 but does potentiate GABAA

receptor activity. Nat. Med. 23 , 1394–1396.

26. Tollner, K., Brandt, C., Topfer, M., Brunhofer, G.,

¨

Erker, T., Gabriel, M., Feit, P.W., Lindfors, J., Kaila,

K., and Loscher, W. (2014). A novel prodrug-

¨

based strategy to increase effects of bumetanide

in epilepsy. Ann. Neurol. 75 , 550–562.

39. Yan, Y., Dempsey, R.J., Flemmer, A., Forbush,

B., and Sun, D. (2003). Inhibition of Na(+)-K(+)-

Cl(À ) cotransporter during focal cerebral

ischemia decreases edema and neuronal

damage. Brain Res. 961 , 22–31.

49. Tang, X., Drotar, J., Li, K., Clairmont, C.D.,

Brumm, A.S., Sullins, A.J., Wu, H., Liu, X.S.,

Wang, J., Gray, N.S., et al. (2019).

Pharmacological enhancement of KCC2 gene

expression exerts therapeutic effects on human

Rett syndrome neurons and Mecp2 mutant

mice. Sci. Transl. Med. 11 .

27. Yuste, R., and Katz, L.C. (1991). Control of

postsynaptic Ca2+ inﬂux in developing

neocortex by excitatory and inhibitory

neurotransmitters. Neuron 6 , 333–344.

40. Schiapparelli, P., Guerrero-Cazares, H., Magana-

Maldonado, R., Hamilla, S.M., Ganaha, S.,

Goulin Lippi Fernandes, E., Huang, C.H.,

Aranda-Espinoza, H., Devreotes, P., and

Quinones-Hinojosa, A. (2017). NKCC1 regulates

migration ability of glioblastoma cells by

modulation of actin dynamics and interacting

with coﬁlin. EBioMedicine 21 , 94–103.

50. Cancedda, L., Fiumelli, H., Chen, K., and Poo,

M.M. (2007). Excitatory GABA action is essential

for morphological maturation of cortical

28. Korpi, E.R., and Luddens, H. (1997).

¨

Furosemide interactions with brain GABAA

receptors. Br. J. Pharmacol. 120 , 741–748.

neurons in vivo. J. Neurosci. 27 , 5224–5235.

29. Delpire, E., Days, E., Lewis, L.M., Mi, D., Kim, K.,

Lindsley, C.W., and Weaver, C.D. (2009). Small-

molecule screen identiﬁes inhibitors of the

neuronal K-Cl cotransporter KCC2. Proc. Natl.

Acad. Sci. USA 106 , 5383–5388.

51. Kaech, S., and Banker, G. (2006). Culturing

hippocampal neurons. Nat. Protoc. 1 , 2406–

2415.

41. Duarte, S.T., Armstrong, J., Roche, A., Ortez,

C., P e ´ rez, A., Mdel, O’Callaghan, Mdel, M.,

Pereira, A., Sanmartı, F., Ormaz a ´b al, A., Artuch,

´

52. Reeves, R.H., Irving, N.G., Moran, T.H., Wohn,

A., Kitt, C., Sisodia, S.S., Schmidt, C., Bronson,

R.T., and Davisson, M.T. (1995). A mouse model

for Down syndrome exhibits learning and

R., et al. (2013). Abnormal expression of

cerebrospinal ﬂuid cation chloride

cotransporters in patients with Rett syndrome.

PLoS One 8 , e68851.

30. Kleschevnikov, A.M., Belichenko, P.V., Faizi, M.,

Jacobs, L.F., Htun, K., Shamloo, M., and

behaviour deﬁcits. Nat. Genet. 11 , 177–184.

Mobley, W.C. (2012). Deﬁcits in cognition and

synaptic plasticity in a mouse model of Down

syndrome ameliorated by GABAB receptor

antagonists. J. Neurosci. 32 , 9217–9227.

42. Tang, X., Kim, J., Zhou, L., Wengert, E., Zhang,

L., Wu, Z., Carromeu, C., Muotri, A.R.,

Marchetto, M.C., Gage, F.H., and Chen, G.

(2016). KCC2 rescues functional deﬁcits in

human neurons derived from patients with Rett

syndrome. Proc. Natl. Acad. Sci. USA 113 ,

751–756.

53. Reinholdt, L.G., Ding, Y., Gilbert, G.J.,

Czechanski, A., Solzak, J.P., Roper, R.J.,

Johnson, M.T., Donahue, L.R., Lutz, C., and

Davisson, M.T. (2011). Molecular

31. Ha, S., Sohn, I.J., Kim, N., Sim, H.J., and Cheon,

K.A. (2015). Characteristics of brains in autism

spectrum disorder: structure, function and

connectivity across the lifespan. Exp.

Neurobiol. 24 , 273–284.

characterization of the translocation

breakpoints in the Down syndrome mouse

model Ts65Dn. Mamm. Genome 22 , 685–691.

43. Hinz, L., Torrella Barrufet, J., and Heine, V.M.

(2019). KCC2 expression levels are reduced in

post mortem brain tissue of Rett syndrome

patients. Acta neuropathol. Commun. 7 , 196.

54. Eissa, N., Jayaprakash, P., Azimullah, S., Ojha,

_

S.K., Al-Houqani, M., Jalal, F.Y., Łazewska, D.,

32. Danziger, J., and Zeidel, M.L. (2015). Osmotic

homeostasis. Clin. J. Am. Soc. Nephrol. 10 ,

852–862.

ꢀ

Kiec-Kononowicz, K., and Sadek, B. (2018). The

histamine H3R antagonist DL77 attenuates

autistic behaviors in a prenatal valproic acid-

induced mouse model of autism. Sci. Rep. 8 ,

13077, 13077.

44. Banerjee, A., Rikhye, R.V., Breton-Provencher,

V., Tang, X., Li, C., Li, K., Runyan, C.A., Fu, Z.,

Jaenisch, R., and Sur, M. (2016). Jointly reduced

inhibition and excitation underlies circuit-wide

changes in cortical processing in Rett

syndrome. Proc. Natl. Acad. Sci. USA 113 ,

E7287–E7296.

33. Delpire, E., Lu, J., England, R., Dull, C., and

Thorne, T. (1999). Deafness and imbalance

associated with inactivation of the secretory Na-

K-2Cl co-transporter. Nat. Genet. 22 , 192–195.

55. Drapeau, E., Riad, M., Kajiwara, Y., and

Buxbaum, J.D. (2018). Behavioral phenotyping

of an improved mouse model of Phelan-

McDermid syndrome with a complete deletion

of the Shank3 gene. E.Neuro. 5 .

34. Kakigi, A., Nishimura, M., Takeda, T., Taguchi,

D., and Nishioka, R. (2009). Expression of

aquaporin1, 3, and 4, NKCC1, and NKCC2 in

the human endolymphatic sac. Auris Nasus

Larynx 36 , 135–139.

45. Hsu, Y.T., Chang, Y.G., Chang, C.P., Siew, J.J.,

Chen, H.M., Tsai, C.H., and Chern, Y. (2017).

Altered behavioral responses to gamma-

aminobutyric acid pharmacological agents in a

mouse model of Huntington’s disease. Mov.

Disord. 32 , 1600–1609.

56. Campolongo, M., Kazlauskas, N., Falasco, G.,

Urrutia, L., Salgueiro, N., Hocht, C., and

¨

35. Ben-Ari, Y., Damier, P., and Lemonnier, E.

(2016). Failure of the Nemo trial: bumetanide is

a promising agent to treat many brain

Depino, A.M. (2018). Sociability deﬁcits after

prenatal exposure to valproic acid are rescued

by early social enrichment. Mol. Autism 9 , 36.

24

Chem 6, 1–24, August 6, 2020

Article Doi

Discovery of a Small Molecule Drug Candidate for Selective NKCC1 Inhibition in Brain Disorders

DOI: 10.1016/j.chempr.2020.06.017

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.chempr.2020.06.017

Full text of DOI:10.1016/j.chempr.2020.06.017

Products guided by the article

R&D Labs maybe for 28328-61-2

Relevant to this article

Hot Product