Reversing Excitatory GABAAR Signaling Restores Synaptic Plasticity and Memory in A Mouse Model of Down Syndrome

Articles
Reversing excitatory GABAAR signaling restores synaptic

plasticity and memory in a mouse model of Down syndrome
npg
2015 Nature America, Inc. All rights reserved.
Gabriele Deidda1,2, Martina Parrini1,2, Shovan Naskar1,2, Ignacio F Bozarth1, Andrea Contestabile1,3 &
Laura Cancedda1,3
Down syndrome (DS) is the most frequent genetic cause of intellectual disability, and altered GABAergic transmission through Cl permeable GABAA receptors (GABAARs) contributes considerably to learning and memory deficits in DS mouse models. However,
the efficacy of GABAergic transmission has never been directly assessed in DS. Here GABA AR signaling was found to be excitatory
rather than inhibitory, and the reversal potential for GABAAR-driven Cl currents (ECl) was shifted toward more positive potentials
in the hippocampi of adult DS mice. Accordingly, hippocampal expression of the cation Cl cotransporter NKCC1 was increased
in both trisomic mice and individuals with DS. Notably, NKCC1 inhibition by the FDA-approved drug bumetanide restored ECl,
synaptic plasticity and hippocampus-dependent memory in adult DS mice. Our findings demonstrate that GABA is excitatory in
adult DS mice and identify a new therapeutic approach for the potential rescue of cognitive disabilities in individuals with DS.
Down syndrome (DS) is the most frequent genetic cause of intellectual disability in children and adults. Intellectual disability is the
most striking clinical feature of individuals with DS, who display low
intelligence quotients, learning deficits and memory impairment, particularly in hippocampus-related functions1. The Ts65Dn mouse is the
best characterized animal model of DS and carries an extra copy of the
distal segment of mouse chromosome 16, which is syntenic to the long
arm of human chromosome 21 (refs. 1,2). This trisomic mouse model
of DS replicates the essential hippocampal cognitive deficits of the
human syndrome, such as impaired synaptic plasticity (for example,
long-term potentiation3 (LTP)) and learning and memory deficits2,4,5.
These impairments have been ascribed to neurodevelopmental alterations observed in the Ts65Dn mouse model of DS, including increased
generation of forebrain GABAergic interneurons6. This is believed to
lead to excessive inhibition68, which affects synaptic plasticity and
cognition3,4 in DS mice as a result of an imbalance in excitatory and
inhibitory neurotransmission9. Both LTP and cognitive impairments
can be rescued in Ts65Dn mice by treatment with GABAA receptor
(GABAAR) antagonists to reduce the magnitude of GABAergic signaling. This evidence suggests that alterations in GABAAR signaling
in the hippocampus are causally linked to LTP abnormalities and
cognitive disabilities in DS mice4,10. However, not all studies on DS
mice have reported enhanced GABAAR-mediated inhibition1114.
Notably, no study has directly assessed the efficacy of GABA AR sig
naling in mouse models of DS. Here we found that GABAAR signaling
was excitatory rather than inhibitory in Ts65Dn mice. This excitatory activity was accompanied by (i) a shift in the reversal potential
for GABAAR-driven Cl currents (ECl) toward more positive potentials and (ii) increased hippocampal expression of the cation Cl
cotransporter NKCC1 in both Ts65Dn mice and individuals with DS.
The treatment of Ts65Dn mice with the FDA-approved NKCC1

inhibitor bumetanide restored ECl to potentials seen in wild-type
(WT) mice and rescued both synaptic plasticity and hippocampusdependent memory in adult DS mice.
RESULTS
Dysregulation of NKCC1 drives excitatory GABAAR signaling in
Ts65Dn mice
To investigate the efficacy of GABAAR signaling in DS, we obtained
cell-attached patch-clamp recordings of hippocampal CA1 pyramidal
neurons in acute slices from adult Ts65Dn mice and WT littermates
during bath application of increasing concentrations of exogenous
GABA (1200 M). Consistent with the known inhibitory action
of GABA in the adult brain, we found that 100 and 200 M GABA
elicited a decrease in spike frequency in neurons from WT mice.
Remarkably, GABA application at the same concentrations elicited
a strong increase in spike frequency in neurons from Ts65Dn mice
that was abolished by application of the GABAAR antagonist bicuculline (10 M; Fig. 1a,b and Supplementary Fig. 1a,b). We therefore
tested whether endogenous GABAAR signaling also elicited excitatory
actions in adult Ts65Dn mice. Recordings from a large number of CA1
pyramidal neurons revealed that the average spontaneous spiking
frequency of neurons from Ts65Dn mice was more than double that
of cells from WT mice (P = 0.035, MannWhitney test) and that
the cumulative distribution of spike frequencies was shifted toward
higher levels (Fig. 1c). Then, in a subset of recordings, we examined the effect of bath application of bicuculline. As expected, spike
frequency increased in neurons from WT mice after the removal of
GABAAR-mediated inhibition. Conversely, spike frequency in neurons
from Ts65Dn mice was decreased by bicuculline application (Fig. 1d),
1Neuroscience and Brain Technologies Department, Istituto Italiano di Tecnologia, Genova, Italy. 2These authors contributed equally to this work. 3These authors jointly
directed this study. Correspondence should be addressed to L.C. (laura.cancedda@iit.it) or A.C. (andrea.contestabile@iit.it).
Received 8 August 2014; accepted 19 February 2015; published online 16 March 2015; doi:10.1038/nm.3827
318
VOLUME 21 | NUMBER 4 | APRIL 2015 nature medicine
articles
a
50 pA
Spike frequency (Hz)
Figure 1 CA1 hippocampal pyramidal neurons

8
exhibit excitatory responses to exogenous
WT
Ts65Dn
**
Baseline
and endogenous GABA in adult Ts65Dn
mice. (a) Example traces of spontaneous
WT
6
Ts65Dn
spiking activity in cell-attached patch-clamp
5s
GABA 50 M
configuration in acute slices derived from
Ts65Dn mice and WT littermates at 1012
4
weeks of age before (baseline) and during bath
GABA 100 M
application of GABA at different concentrations
2
(50, 100 and 200 M). (b) Quantification of
the average data (s.e.m.) summarizing the
GABA 200 M
effect of GABA application on spontaneous
0
spiking activity in neurons from Ts65Dn mice
Baseline GABA GABA GABA GABA GABA
(n = 14 cells) and WT mice (n = 12 cells).
1 M 5 M 50 M 100 M 200 M
Significant differences were determined by
WT
Ts65Dn
WT
Ts65Dn
comparison to baseline frequencies using
Vehicle
Bicuculline Vehicle
Bicuculline
10
two-way repeated-measure analysis of variance
(ANOVA) on ranked data (**P = 0.002,
5s
5s
Ts65Dn; P = 0.01, WT; post hoc HolmSidak
test). (c) Cumulative-distribution curve for
100
*
WT - vehicle
the frequency of spontaneous spiking activity
WT - bicuculline
80
in neurons from Ts65Dn mice (n = 66 cells)
Ts65Dn - vehicle
5
Ts65Dn - bicuculline
versus that in neurons from WT mice (n = 33
WT
60
cells). Significant differences were determined
Ts65Dn
by KolmogorovSmirnov Test (P = 0.013).
40
*
Insets show sample traces of spontaneous
20
spiking activity in neurons from WT and
Ts65Dn mice. (d) Spike frequency before
0
0
and after bath application of the GABAAR
0
1
2
3
4
5
6
WT
Ts65Dn
antagonist bicuculline (10 M) in neurons
from Ts65Dn (n = 6 cells) and WT (n = 5 cells)
mice (*P = 0.028, Ts65Dn; P = 0.034, WT; paired t-test). Insets show sample traces of spiking activity in neurons from WT and Ts65Dn mice before
and after bath application of bicuculline. Open circles indicate single data points, and their averages (s.e.m.) are reported by colored bars to the side.
20 pA
npg
consistent with excitatory signaling of endogenous GABA through

GABAARs in the hippocampus of Ts65Dn mice.
Previous studies have demonstrated increased GABABR responses
in Ts65Dn mice8,12, owing to triplication of the DS gene Kcnj6
(Girk2 or Kir3.2), which encodes an inward-rectifying K+ channel
that couples to GABABRs15. To examine the effect of GABABR
signaling on spontaneous spiking, we performed cell-attached
recordings in the presence of the GABABR antagonist CGP55845
(10 M). Consistent with previous reports 8,12, CGP55845 caused
an increase in the spiking frequency of neurons from Ts65Dn mice
without significantly altering spiking in neurons from WT mice
(Supplementary Fig. 1c).
We next directly investigated the strength, direction and ECl of
GABAAR Cl currents using gramicidin-perforated whole-cell
recordings, a technique that preserves the endogenous intracellular
Cl concentration [Cl]i (ref. 16). Local puffing of GABA at the cell
body revealed that neurons from WT mice exhibited an ECl (66.0
1.46 mV) that was more negative than the resting membrane potential
(62.4 2.07 mV), predictive of inward Cl currents mediating hyperpolarization upon GABAAR activation (Cl driving force = 3.64 mV;
Fig. 2a,b). In neurons from Ts65Dn mice, the average ECl was ~8 mV
less negative (58.3 1.33 mV) than that in cells from WT mice,
and it was above the resting membrane potential (64.4 0.50 mV),
predictive of outward Cl currents upon GABAAR activation and in
agreement with the excitatory action of GABA described above (Cl
driving force = +6.07 mV; Fig. 2a,b). These results were confirmed
in experiments using two-photon imaging with the chloride-sensitive
dye MQAE, which showed an increased resting [Cl]i in hippocampal
CA1 neurons from Ts65Dn mice compared to neurons from WT
littermates (Supplementary Fig. 1d).
nature medicine VOLUME 21 | NUMBER 4 | APRIL 2015
Cumulative distribution (%)
20 pA
ECl is mainly determined by [Cl]i, which in turn depends on the

opposite actions of the Cl importer NKCC1 and exporter KCC2.
High NKCC1 expression leads to a high [Cl]i, whereas high KCC2
expression leads to a low [Cl]i (ref. 17). To address whether altered
expression of these transporters correlated with the shift of GABAAR
signaling from inhibitory to excitatory in Ts65Dn mice, we examined
NKCC1 and KCC2 protein expression by western blotting. We found
that NKCC1 expression was increased both in the entire hippocampus
and in the CA3CA1 subregion in Ts65Dn mice compared to WT littermates (Fig. 3a,b). To establish a parallel with the human condition,
we also examined expression in the hippocampi of adult individuals
with DS. We found that NKCC1 expression was increased in samples
from humans with DS (Fig. 3c). Given that the function of NKCC1
and KCC2 depends on their localization at the cell membrane, we performed subcellular-fractionation experiments and found that NKCC1
was overexpressed in the synaptosomal membrane fraction of tissue
from both Ts65Dn mice and human subjects with DS (Fig. 3d,e).
We also executed pulldown experiments on biotinylated membrane
proteins from acute brain slices of the CA1 hippocampal region from
Ts65Dn and WT mice. We found that NKCC1 surface expression was
increased in samples from Ts65Dn mice (Fig. 3f). Conversely, KCC2
protein expression was unaffected in Ts65Dn mice and was not significantly different between hippocampus samples from humans with
DS and controls (Supplementary Fig. 2). We also used quantitative
RT-PCR (qRT-PCR) to measure NKCC1 and KCC2 mRNA expression
in samples from both Ts65Dn mice and humans with DS. Interestingly,
whereas the expression of KCC2 mRNA was not significantly altered
in either Ts65Dn mice or human subjects with DS (Supplementary
Fig. 3ac), NKCC1 mRNA was specifically increased in the
human samples (Supplementary Fig. 3b), but not in Ts65Dn mice.
319
Articles
a
npg
Notably, we were able to detect the expected overexpression of the

triplicated gene APP in samples from both Ts65Dn mice and humans
with DS (Supplementary Fig. 3ad).
Next, to investigate whether excitatory GABAergic signaling also
occurs in other brain regions, we performed electrophysiological and
biochemical experiments in the neocortex in Ts65Dn mice. We found
excitatory responses to exogenous GABA application and an increase
in the specific expression of NKCC1 in the cortices of Ts65Dn mice
relative to that in WT mice (Supplementary Fig. 4a,b). Conversely,
we observed no changes in the protein or mRNA expression of KCC2
in the neocortices of Ts65Dn mice compared to that in WT mice
(Supplementary Fig. 4c,d).
We assumed that if the more positive values of ECl and GABAinduced spiking activity were indeed due to the increased expression of NKCC1 in Ts65Dn mice, then acute application of the
FDA-approved NKCC1 inhibitor bumetanide (a widely used loop
diuretic)18,19 would restore normal ECl values and spiking activity.
We found that the application of bumetanide (10 M) rescued ECl,
decreased spontaneous spiking and suppressed GABA-induced spike
frequency in hippocampal neurons of Ts65Dn mice (Fig. 2ce and
Supplementary Fig. 1e). Conversely, we found no significant effect
of bumetanide application on ECl, spontaneous spiking or GABAinduced spiking in neurons of WT mice. These results indicate that
the shift in ECl was responsible for excitatory GABAAR signaling
in the adult Ts65Dn mice.
Bumetanide treatment rescues LTP in adult Ts65Dn mice
As impaired synaptic plasticity in Ts65Dn mice has been linked to
excess GABAAR signaling3, we next addressed whether more positive ECl values could account for the deficits in hippocampal LTP.
320
ECl (mV)
50 pA
ECl (mV)
IGABA amplitude (pA)
50 pA
50 pA
WT
Ts65Dn
Figure 2 Hippocampal CA1 neurons exhibit
WT
Ts65Dn
a less negative ECl in adult Ts65Dn mice
40
60
than in adult WT mice. (a) Examples of
5s
40
5s
currentvoltage relationships of GABA
50
currents (IGABA) elicited by puffing GABA at
20
the neuronal cell body and recorded by
gramicidin-perforated patch-clamp in slices
0
60
from Ts65Dn and WT mice. Vertical dotted
20
lines indicate the value of ECl on the x-axis.
Insets show sample traces of GABA-induced
40
70
currents at different holding potentials.
(b) ECl values for neurons from Ts65Dn
60
(n = 14 cells) and WT (n = 9 cells) mice
***
80
80
recorded as in a (***P < 0.001 by
100 80
60
40
20
100 80 60 40 20
Students t-test). Dotted horizontal lines
Holding potential (mV)
Holding potential (mV)
indicate average resting potentials determined
by whole-cell patch-clamp from independent
WT
Ts65Dn
Vehicle Bumetanide Vehicle Bumetanide
sets of neurons (Ts65Dn, n = 6 cells; WT,
WT
Ts65Dn
WT
40
6
12
n = 7 cells). (c) ECl values before and 20 min
Ts65Dn
*
after bath application of NKCC1-inhibitor
WT - vehicle
5s
bumetanide (10 M) to a subset of the
WT - bumetanide
*
**
Ts65Dn - vehicle
same cells studied in b (*P = 0.012, Ts65Dn;
4
Ts65Dn - bumetanide
P > 0.05, WT; paired t-test). (d) Spike
*
frequency before and after bath application
65
6
of bumetanide in neurons from WT and
Ts65Dn mice (*P = 0.037, Ts65Dn; P = 0.548,
2
WT; paired t-test). Insets show sample traces
WT - vehicle
from neurons of WT and Ts65Dn mice.
WT - bumetanide
Ts65Dn - vehicle
(e) Average frequency (s.e.m.) of spiking
Ts65Dn - bumetanide
90
activity before and during bath application
0
0
Baseline
GABA
GABA
WT
Ts65Dn
of either GABA (100 M) or GABA with
100 M 100 M +
bumetanide
bumetanide (10 M) in neurons from
WT (n = 5 cells) and Ts65Dn mice (n = 6 cells). *P = 0.017, GABA; **P = 0.0026, GABA + bumetanide; post hoc HolmSidak test after two-way
repeated-measure ANOVA. In b, c and d, circles indicate single data points, and their averages (s.e.m.) are reported by colored bars to the side.
LTP elicited by theta-burst stimulation (TBS) at Schaffer collateral

CA1 synapses (CA3CA1 LTP) was decreased in vehicle-treated
hippocampal slices from Ts65Dn mice in comparison to slices from
WT littermates (Fig. 4ac). Strikingly, bath application of bumetanide
(10 M) completely rescued LTP in Ts65Dn mice, returning it to the
level observed in WT mice (Fig. 4b,c). Conversely, we found no effect
of bumetanide on LTP in WT mice. The effect in Ts65Dn mice was not
due to an alteration of basal synaptic transmission or short-term plasticity (Fig. 4df), and it was dependent on the rescue of ECl. Indeed,
inhibition of KCC2, which shifts ECl toward more positive levels20,
by the specific inhibitor VU 0240551 (ref. 21) did not significantly
affect LTP (Supplementary Fig. 5ac). We investigated the effect of
bath-applied GABA (100 M) on LTP and found no significant effect
on LTP in either Ts65Dn or WT mice (Supplementary Fig. 5d,e).
Thus the reestablishment of hyperpolarizing ECl leads to the complete
recovery of synaptic plasticity in Ts65Dn mice.
Bumetanide treatment rescues memory deficits in adult
Ts65Dn mice
We next investigated a possible contribution of excitatory GABA to
cognitive impairment in subjects with DS by assessing the performance of Ts65Dn mice in various behavioral tests. In particular, we
evaluated long-term hippocampus-dependent explicit memory after
either subchronic (1 week) or chronic (4 weeks) systemic treatment
with bumetanide (0.2 mg kg1 IP daily). As previously described5,
vehicle-treated Ts65Dn mice showed impaired memory in the contextual fear-conditioning (CFC) test, as demonstrated by a strong reduction of the freezing response elicited upon re-exposure to the training
context 24 h after conditioning (Fig. 5a,b). Bumetanide treatment
fully restored associative memory in Ts65Dn mice, without inducing
articles
a
WT
Ts65Dn
DS
Mouse hippocampus
WT
Mouse CA3CA1
Ts65Dn
WT
Ts65Dn
Control
Mouse synaptosomes
WT
DS
NKCC1
NKCC1
NKCC1
Actin
Actin
Actin
NaK-ATPase
100
80
60
40
140
120
100
80
60
40
20
20
WT
Ts65Dn
WT
Ts65Dn
180
160
NKCC1/NaK-ATPase
(% of WT)
120
160
200
NKCC1/actin (% of control)
140
180
NKCC1/actin (% of WT)
NKCC1/actin (% of WT)
160
**
200
180
140
120
100
80
60
40
20
0
Control
DS
220
200
180
160
140
120
100
80
60
40
20
0
Ts65Dn
WT
Ts65Dn
Mouse CA1
changes in non-associative freezing (Fig. 5b) or altering sensitivity

to shock (Supplementary Fig. 6a,b). Furthermore, both subchronic
and chronic bumetanide administration restored the performance of
Ts65Dn mice to the level of WT mice in the object-location (OL) test,
indicating a full recovery of spatial-memory performance (Fig. 5c).
Similarly, we found that the poor novelty-discrimination capability
of Ts65Dn mice in the novel-object recognition (NOR) test was completely rescued by bumetanide administration (Fig. 5d). We also
found a slight promnesiant effect of bumetanide on WT mice in the
NOR test only after chronic drug treatment (4 weeks). This implies
a long-term effect in WT mice that specifically targets recognition
memory in the NOR test. The effect of bumetanide in both OL and
NOR tests was not due to alterations in total object exploration or
object preference (Supplementary Figs. 7 and 8). However, bumetanide treatment (24 h) did not rescue the increased locomotor activity
of Ts65Dn mice2 (Supplementary Fig. 9ac).
To assess the effects of bumetanide treatment on the GABAAR 1,
3 and subunits22, we examined the expression of these proteins by
western blotting of tissue from Ts65Dn and WT mice that were systemically treated for 1 month with bumetanide or vehicle. We found
no difference in 1 or 3 subunit expression between Ts65Dn and WT
mice. Moreover, although expression of the subunit was increased in
NKCC1 expression
(% of WT)
Surface
Figure 3 NKCC1 protein expression is increased
Human synaptosomes
Total (input)
(pulldown)
in the hippocampi of Ts65Dn mice and in samples
WT Ts65Dn
WT
Ts65Dn
Control
DS
from human subjects with DS. (ac) Top,
NKCC1
representative immunoblots for NKCC1 in protein
NKCC1
NaK-ATPase
extracts from samples of (a) mouse whole
NaK-ATPase
Actin
hippocampus (WT, 25 mice; Ts65Dn, 25 mice),
*
(b) mouse CA3CA1 hippocampal subregion
200
Total
Surface
300
180
(Ts65Dn, 15 mice; WT, 16 mice) and (c) human
*
*
160
postmortem whole hippocampus (five samples
250
140
each). Bottom, quantification of NKCC1 in
200
120
samples from Ts65Dn mice and humans with
100
150
DS in comparison to WT mice and age- and
80
sex-matched non-trisomic controls, respectively
100
60
(P = 0.025, 0.003 and 0.045 in a, b and c, respectively,
40
50
by Students t-test). Actin was used as an internal
20
0
0
standard. (d,e) Top, representative immunoblots for NKCC1
Control
DS
WT
Ts65Dn
WT
Ts65Dn
in protein extracts from (d) mouse (Ts65Dn, 16 mice;
WT, 16 mice) and (e) human (five samples each) synaptosomal fractions of the whole hippocampus. Bottom, quantification of NKCC1 in Ts65Dn mice
and samples from human subjects with DS in comparison to WT and non-trisomic controls, respectively (P = 0.043 and 0.023 in d and e, respectively,
by Students t-test). The membrane protein Na +/K+-ATPase was used as an internal standard. (f) Top, representative immunoblots for NKCC1 on protein
extracts from total or surface biotinylation experiments with samples from WT (n = 4 mice) and Ts65Dn (n = 4 mice) mice. Bottom, quantification of
NKCC1 in total and surface fractions (P = 0.013 and 0.047, respectively, by Students t-test). Actin and Na+/K+-ATPase were used as internal standards
for total and surface quantification, respectively. For all panels, circles indicate values from single samples, and their averages (s.e.m.) are reported by
colored bars to the right. Full blots are presented in Supplementary Figure 13. *P < 0.05, **P < 0.01.
NKCC1/NaK-ATPase
(% of control)
Human hippocampus
NKCC1
200
npg
WT mice22, we did not observe a significant increase in Ts65Dn mice

(Supplementary Fig. 10). Thus, the effect of bumetanide on memory
in Ts65Dn mice could not be ascribed to changes in the expression of
GABAAR subunits 1, 3 and .
Together, these data indicate that restoring the inhibitory action
of GABA by correcting neuronal [Cl]i rescues cognitive disabilities
in a DS mouse model.
Effects of bumetanide on memory are independent of neuronalcircuit rearrangement
We next tested whether the rescue of cognitive function in Ts65Dn
mice by bumetanide required rearrangement at the level of hippo
campal neural circuits. First, we treated Ts65Dn and WT mice with
bumetanide or vehicle for 1 month and assessed LTP in acute hippocampal slices in the absence of a bumetanide bath (Fig. 6a). As the
slice-cutting procedure guaranteed complete washout of previously
administered bumetanide, this experiment tested whether bumetanide induced plastic changes in neuronal circuits that were able to
sustain the rescue of LTP in the absence of the drug. We found that
a 1-month in vivo treatment did not rescue LTP in acute slices from
bumetanide-treated Ts65Dn mice in comparison to slices from WT
mice (Fig. 6b,c).
321
npg
Recording electrode
(CA1 stratum radiatum)
Stimulating electrode
(CA3 Schaffer collateral)
Theta-burst stimulation protocol

30 ms
WT - vehicle
WT - bumetanide
Ts65Dn - vehicle
Ts65Dn - bumetanide
1 mV
b
10 ms
WT - vehicle
Ts65Dn - vehicle
WT - bumetanide
Ts65Dn - bumetanide
170
TBS
250
fEPSP slope (% of baseline)
260
10
10
20
30
40
50
**
200
150
100
50
80
60
0
Vehicle
Bumetanide
Time (min)
e
WT
Ts65Dn
140
Bumetanide 10 M
300
175
WT - vehicle
Ts65Dn - vehicle
WT - bumetanide
Ts65Dn - bumetanide
200
fEPSP slope (mV ms )
fEPSP slope (% facilitation)
Figure 4 Bath application of the FDA-approved

NKCC1 inhibitor bumetanide rescues the deficit
in hippocampal CA3CA1 LTP in Ts65Dn mice.
(a) Sketch of the in vitro recording configuration
(left) and stimulation protocol (right). (b) The
average time course (s.e.m.) of the change in
the slope of the field excitatory postsynaptic
potential (fEPSP) in hippocampal slices during
bath application of either vehicle (Ts65Dn,
n = 12 slices; WT, n = 17 slices) or bumetanide
(10 M; Ts65Dn, n = 12 slices; WT, n = 12
slices). Traces (top) represent the average of 10
fEPSPs recorded for each experimental group
before (continuous lines) and 60 min after
theta-burst stimulation (TBS; dashed lines).
Stimulus artifacts have been deleted from
traces for clarity. (c) Average LTP (s.e.m.)
and single-slice cases (open circles) of the
last 5 min of recordings in b. Significant
differences were determined by two-way ANOVA
followed by Tukeys post hoc test (*P = 0.038,
**P = 0.005). (d) Basal synaptic transmission
in slices from WT (n = 4 slices) and Ts65Dn
(n = 5 slices) mice (P > 0.05, Students t-test).
(e) Paired-pulse facilitation in vehicle-treated
(WT, n = 6; Ts65Dn, n = 6) and bumetanidetreated slices (WT, n = 6; Ts65Dn, n = 4;
P > 0.05, ANOVA on ranks). (f) Stimulation
inputfEPSP output response curve for
vehicle-treated (WT, n = 6; Ts65Dn, n = 6)
and bumetanide-treated (WT, n = 5; Ts65Dn,
n = 4) slices (P > 0.05, two-way ANOVA).
Articles
WT - vehicle
Ts65Dn - vehicle
WT - bumetanide
Ts65Dn - bumetanide
Second, we examined memory in Ts65Dn

1
mice after 1 month of bumetanide treatment
followed by 1 week of drug withdrawal (Fig. 6a).
We found that Ts65Dn mice performed as
0
80
50
poorly as the vehicle group after bumetanide
50
100
200
5
0
5
10
15
50
100
0
withdrawal in the CFC, OL and NOR tests
Time (min)
Interpulse interval (ms)
Stimulation (%)
(Fig. 6d,e), which suggests that the beneficial
effect of bumetanide relied on direct cotransporter antagonism and ineffective in preventing seizures, bumetanide treatment does not
was quickly lost upon drug washout.
show any pro-convulsion effect.
Third, we proposed that if neuronal-circuit rearrangements were not
necessary to rescue cognitive function after chronic bumetanide treat- DISCUSSION
ment, then acute treatment with bumetanide should also rescue the Cognitive impairment in Ts65Dn mice has been attributed to an
memory of Ts65Dn mice. Indeed, we found that acute administration increased number of GABAergic interneurons in the cortex and
of bumetanide was sufficient to completely rescue the cognitive per- hippocampus, which results in enhanced GABA signaling 4,68,12.
formance of Ts65Dn mice in the CFC, OL and NOR tests (Fig. 6a,d,f). However, other studies have not found increased GABAAR signaling
Conversely, we observed no major changes in non-associative in Ts65Dn mice1114. Nevertheless, both LTP and cognitive impairfreezing in the CFC task (Supplementary Fig. 6ce). Additionally, ments can be rescued by reducing the magnitude of GABA-mediated
no significant differences were observed in total object explo- signaling through treatment with GABAAR antagonists3,4,10. Our
ration or object preference in the OL (Supplementary Fig. 11) results on excitatory signaling by GABAAR activation in Ts65Dn
and NOR tests (Supplementary Fig. 12). Together, these results indi- mice provide an interpretative framework for the existing literature. In
cate that the effect of bumetanide in rescuing memory deficits in fact, the rescue of hippocampal deficits in Ts65Dn mice by GABAAR
Ts65Dn mice depends on the direct antagonism of NKCC1, and not antagonists could be explained by the fact that aberrant (excitatory)
on neuronal-circuit rearrangements.
GABAA signaling in Ts65Dn mice is reduced by GABAAR antagoLastly, as treatments targeting the GABAergic system have been nists independent of whether GABAAR signaling is increased in DS.
questioned because of the increased seizure susceptibility of individu- However, the lack of effect of bath application of GABA on LTP could
als with DS1,2325, we investigated the effects of bumetanide treatment reflect the high (possibly saturating) level of endogenous GABAergic
on audiogenic seizure (AGS) induction in WT and Ts65Dn mice. signaling in the hippocampus or desensitization of GABAARs upon
Although vehicle-treated Ts65Dn mice showed increased seizure prolonged activation, as previously described26,27.
susceptibility compared to controls25, we found that bumetanide
Moreover, by extending the previous literature on GABAergic transtreatment did not substantially change AGS susceptibility in Ts65Dn mission in mouse models of DS3,4,68,1114, our results on excitatory
mice (Supplementary Fig. 9d). This result indicates that, although signaling by GABAAR activation in Ts65Dn mice indicate that the
322
articles
Bumetanide 0.2 mg kg
34-month-old
WT and Ts65Dn mice
0
(IP)
Contextual fear conditioning
Shock
Context test
Trained context
Weeks of treatment
100
Discrimination index
50
***
***
75
***
40
50
25
WT - vehicle
Ts65Dn - vehicle
WT - bumetanide
Ts65Dn - bumetanide
30
20
10
New context
Untrained context
Veh
Bum 1W
Object location
WT - vehicle
Ts65Dn - vehicle
WT - bumetanide
Ts65Dn - bumetanide
24 h
Bum 4W
Veh
***
80
40
40
80
Veh
GABAergic system is altered at multiple levels in DS17. In particular,

we found that the [Cl]i was higher and ECl for GABAARs was more
positive in excitatory neurons of adult Ts65Dn mice compared to values
for WT littermates. The results for ECl are seemingly in contrast to the
results of a previous study8, in which no difference was reported in ECl
values for dentate granule neurons between Ts65Dn and WT mice. This
is possibly due to the fact that the ECl was assessed using whole-cell
patch-clamp recordings in the previous study, in contrast to the gramicidin-perforated patch-clamp recordings used here, which allowed for
preservation of the [Cl]i. The higher [Cl]i and the more positive ECl in
pyramidal neurons of adult Ts65Dn mice in our study resulted in overall excitatory GABAAR signaling. Indeed, inhibition of GABAARs in
acute slices from Ts65Dn mice decreased spontaneous spiking, and the
application of exogenous GABA increased spiking activity. Consistently,
the restoration of ECl by bumetanide treatment decreased spontaneous
spiking and inhibited exogenous GABA-induced spiking in Ts65Dn
mice. Together, these results are in line with the lower threshold for
action-potential generation described for Ts65Dn mice28.
If GABAAR signaling triggers an increase in action-potential
frequency, what supports inhibitory signaling in Ts65Dn mice? We
found increased responses to the GABABR antagonist CGP55845
in Ts65Dn mice in comparison to WT mice, indicating that the
excitatory actions of GABAAR signaling can at least in part be compensated for by increased GABABR inhibition. Moreover, as we have not
investigated possible differences between synaptic and extrasynaptic
GABAAR signaling in Ts65Dn mice, it is possible that shunting mechanisms could still provide inhibition in Ts65Dn mice at specific cell
Bum 1W
Bum 4W
Bum 4W
Novel object recognition
***
80
Bum 1W
24 h
Figure 5 Systemic treatment with bumetanide restores

cognitive function in behavioral tasks in Ts65Dn mice.
(a) Schematic cartoon of the experimental protocol
for the different experimental groups. (b) Top, schematic
representation of the contextual fear-conditioning
test. Bottom, quantification of the freezing response
in mice treated with vehicle (Veh; Ts65Dn, n = 20;
WT, n = 44) or bumetanide (Bum) for either 1 week
(1W; Ts65Dn, n = 9; WT, n = 16) or 4 weeks (4W;
Ts65Dn, n = 16; WT, n = 27; ***P < 0.001 by
Tukeys post hoc test following two-way ANOVA)
upon exposure to a trained (left) or untrained (right)
context. (c) Top, schematic representation of the
object-location test. Bottom, quantification of the
discrimination index in mice treated with vehicle
(Ts65Dn, n = 15; WT, n = 31) or bumetanide for
either 1 week (Ts65Dn, n = 10; WT, n = 18) or
4 weeks (Ts65Dn, n = 9; WT, n = 14; ***P < 0.001
by Tukeys post hoc test following two-way ANOVA).
(d) Top, schematic representation of the novel-object
recognition task. Bottom, quantification of the
discrimination index in mice treated with vehicle
(Ts65Dn, n = 15; WT, n = 26) or bumetanide for
either 1 week (Ts65Dn, n = 8; WT, n = 14) or
4 weeks (Ts65Dn, n = 9; WT, n = 17; *P < 0.05,
***P < 0.001; Tukeys post hoc test following two-way
ANOVA). For all panels, circles indicate values from
single animals, and their averages (s.e.m.) are reported
by colored bars to the right.
Time freezing (%)
Behavioral testing
npg
2h
24 h
Time freezing (%)
***
***
40
40
80
Veh
Bum 1W
Bum 4W
compartments29. Indeed, because we recorded the GABA-induced

increase in spiking activity with cell-attached patch-clamp recordings
from the cell bodies of Ts65Dn neurons, we addressed changes only
in their global excitability, and we cannot exclude the possibility that
specific cell compartments (for example, single dendritic branches)
may still be inhibited by local shunting inhibition 30. Furthermore,
we demonstrated that GABAAR signaling is excitatory both in the
hippocampus and in the neocortex in Ts65Dn mice; nevertheless,
GABA might be hyperpolarizing in other brain regions. Lastly,
although our study focused on excitatory neurons, GABAAR signaling may also be affected in interneurons31 in Ts65Dn mice.
Bumetanide treatment also rescued hippocampal synaptic plasticity and memory in Ts65Dn mice, consistent with published reports
of rescued plasticity and cognition after GABAAR inhibition in DS
animal models3,4,10. Similar evidence for a specific impairment in
GABAAR signaling in individuals with DS was lacking. We found a
specific dysregulation in the expression of the Cl importer NKCC1
in the hippocampus of both Ts65Dn mice and human subjects with
DS. Whether these expression changes are a primary event due to the
triplication of genes related to DS or a secondary response to other
modifications is yet to be established. Notably, NKCC1 mRNA was
specifically increased in samples from humans with DS, but not in
Ts65Dn mice. The apparent discrepancy between NKCC1 mRNA and
protein expression levels in Ts65Dn mice might be due to changes at
the translational level and/or in protein stability, or it might reflect a
compensatory effect at the transcriptional level. Additionally, Ts65Dn
mice do not carry the full triplication of all the genes orthologous
323
Bumetanide 0.2 mg kg
34-month-old
WT and Ts65Dn mice
d1 (IP)
Washout
2
3
Treatment duration (weeks)
Acute
behavioral
testing
170
TBS
80
10
Washout
behavioral
testing
10
20
30
Time (min)
40
50
200
150
100
50
0
Vehicle
60
Bumetanide
Contextual fear conditioning
Object location
Novel-object recognition
24 h
24 h
24 h
Context test
WT - vehicle
WT - bumetanide
Ts65Dn - vehicle
Ts65Dn - bumetanide
Chronic treatment - washout
60
40
20
0
Vehicle
100
80
***
100
***
50
25
0
25
50
75
100
Bumetanide
***
75
***
100
Vehicle
50
25
0
25
50
75
100
Bumetanide
***
***
75
Vehicle
Bumetanide
Acute treatment
80
60
***
100
75
50
25
0
***
***
100
***
100
LTP
recordings
250
Ts65Dn - bumetanide
Shock
WT - vehicle
WT - bumetanide
Ts65Dn - vehicle
Ts65Dn - bumetanide
10 ms
WT - vehicle
Ts65Dn - vehicle
WT - bumetanide
260
c
1 mV
Time freezing (%)
Figure 6 Effect of bumetanide on memory does

not depend on changes in neuronal connectivity.
(a) Representation of the experiments.
(b) Average time course (s.e.m.) of fEPSP
slope change in slices from animals treated
systemically for 4 weeks with vehicle
(WT, n = 10 slices; Ts65Dn, n = 7 slices) or
bumetanide (WT, n = 9 slices; Ts65Dn, n = 7
slices). Traces are the average of 10 fEPSPs
before (continuous lines) and 60 min after
TBS (dashed lines). (c) Average LTP in the last
5 min of recordings in b (*P = 0.03, two-way
ANOVA followed by Tukeys post hoc test).
(d) Representation of the behavioral tasks.
(e) Quantification of the behavioral data after
a 1-week washout period following 4 weeks of
treatment with bumetanide. Left, quantification
of the freezing response in vehicle-treated
(Ts65Dn, n = 10; WT, n = 17) or bumetanidetreated (Ts65Dn, n = 9; WT, n = 16) mice
in the CFC test. Middle, quantification of the
discrimination index in vehicle-treated (Ts65Dn,
n = 10; WT, n = 11) or bumetanide-treated
(Ts65Dn, n = 9; WT, n = 11) mice in the OL
test. Right, quantification of the discrimination
index in vehicle-treated (Ts65Dn, n = 9; WT,
n = 15) or bumetanide-treated (Ts65Dn, n = 10;
WT, n = 14) mice in the NOR test (***P < 0.001,
two-way ANOVA followed by Tukeys post hoc
test). The data point of one outlier was excluded
from single-value presentation in the graph for
freezing response (left), but it was included
in calculations of the average and s.e.m.
(f) Quantification of the behavioral data after
acute bumetanide treatment. Left, freezing
response in vehicle-treated (Ts65Dn, n = 11;
WT, n = 14) or bumetanide-treated (Ts65Dn,
n = 10; WT, n = 14) mice in the CFC test.
Middle, discrimination index in vehicle-treated
(Ts65Dn, n = 8; WT, n = 18) or bumetanidetreated (Ts65Dn, n = 8; WT, n = 13) mice in
the OL test. Right, discrimination index in
vehicle-treated (Ts65Dn, n = 10; WT, n = 16)
or bumetanide-treated (Ts65Dn, n = 10;
WT, n = 15) mice in the NOR test (***P < 0.001,
two-way ANOVA followed by Tukeys post hoc
test). In c, e and f, circles indicate data from
single animals, and their averages (s.e.m.)
are reported by colored bars to the right.
Time freezing (%)
npg
Articles
***
***
75
50
25
0
to human chromosome 21 and express three

40
25
25
copies of a small number of genes not tripli50
50
20
cated in humans with DS32. Thus the discrep75
75
0
100
100
ancy between NKCC1 mRNA and protein
Vehicle
Bumetanide
Vehicle
Bumetanide
Vehicle
Bumetanide
expression in Ts65Dn mice and humans might
be due to potential differences in mRNA and/or protein expression is an FDA-approved drug, the pharmacological treatment that we
between humans with DS and Ts65Dn mice. In contrast, KCC2 expres- propose has high potential to be rapidly translated as a safe therasion did not significantly vary in hippocampi from Ts65Dn mice or peutic approach for clinical trials in subjects with DS. Indeed, in
humans with DS in comparison to controls. Nevertheless, as KCC2 contrast to other, newly developed molecules10 requiring extensive
expression showed a high degree of variability among human samples evaluations of safety, tolerability and pharmacokinetics (e.g., trials
and the number of available samples was relatively small, we cannot NCT01436955, NCT01684891 and NCT01667367 listed at http://
confidently exclude potential differences between individuals with DS clinicaltrials.gov/), the drug we propose has a long clinical history
and healthy subjects.
as a diuretic in both acute and chronic (lifelong) treatments 18,19.
The overexpression of NKCC1 observed in subjects with DS sug- Importantly, this drug has no major side effects as indicated by
gests that cation Cl cotransporters are promising targets for thera- a clinical trial in children with autism (trial NCT01078714, http://
peutic interventions aimed at alleviating the cognitive disabilities of clinicaltrials.gov/)33. Furthermore, two major clinical trials on the use
individuals with DS by modulating GABAAR signaling. As bumetanide of bumetanide for neonatal seizures are ongoing in Europe and the
324
npg
articles
United States (NCT01434225 and NCT00830531, http://clinicaltrials.
gov/). Among the reported minor side effects is mild hypokalemia,
which can be compensated for by dietary supplementation 33,34.
Moreover, our results on the acute effects of bumetanide in Ts65Dn
mice will allow for preliminary testing of this pharmacological approach
in subjects with DS after just a short period of drug exposure.
Although brain penetration by bumetanide may not be optimal35,
several studies have shown that it does reach the brain3638. Our
results, together with evidence of the systemic effects of bumetanide
in the treatment of other brain-related diseases33,3947, confirm these
data. The dose of 0.2 mg kg1 used here was chosen on the basis of previous studies on rodents38,39,42,4850. This dose is higher than the usual
therapeutic range for humans (total daily dose, 0.52 mg; maximum
daily dose, 10 mg)51, but it was justified by the rapid elimination rate of
bumetanide in rodents5254. The use of this dosage was further justified
by the encouraging results of parallel studies on the use of bumetanide
for the treatment of autism performed with dosages similar to those
described above for patients and rodent models33,44. Moreover, an
ongoing phase 1 clinical trial is currently testing the efficacy and safety
of bumetanide at 0.1, 0.2 and 0.3 mg kg1 for the treatment of epilepsy
in newborns (NCT00830531, http://clinicaltrials.gov/).
In conclusion, our data describe an innovative and safe pharmacological approach with the FDA-approved drug bumetanide that can
be readily translatable into clinical trials directly on subjects with DS.
This is in keeping with the strategy of drug repurposing55 intended
to reduce the high costs and long time periods between the testing of
new therapeutic approaches in animal models and the commercialization of newly developed drugs. Moreover, clinical33,44,45 and recent
experimental46,56 evidence indicates that GABAAR signaling may also
be excitatory in young patients and in mouse models of autism and
fragile X syndrome. Our results, obtained in mature animals of a reliable mouse model of DS, suggest that a general mechanism involving
excitatory GABAAR signaling may be common to several neurodevelopmental disorders and persist through development into adulthood.
Lastly, our results support the possibility that cognitive symptoms that
result from neurodevelopmental disorders may still be rescued by late
pharmacological intervention in adulthood4,5760.
Methods
Methods and any associated references are available in the online
version of the paper.
Note: Any Supplementary Information and Source Data files are available in the
online version of the paper.
Acknowledgments
This work was supported by Compagnia di San Paolo (grant 2008.1267 to L.C.)
and the Jerome Lejeune Foundation (grants 995-CA2012A, to A.C., and 1266_
CL2014A, to L.C.). Human Down syndrome and control samples were obtained
from the Brain and Tissue Bank for Developmental Disorders at the University of
Maryland, Baltimore. We thank F. Benfenati (Istituto Italiano di Tecnologia
(IIT)) for financial support, J. Assad (IIT) for critical reading of the manuscript,
M. Pesce (IIT NBT imaging facility) for technical assistance with two-photon
microscopy and the staff of the IIT animal facility and genotyping service for their
valuable work. We also thank K. Kaila (University of Helsinki, Helsinki, Finland)
for providing brain tissue from NKCC1-deficient and KCC2-deficient mice.
AUTHOR CONTRIBUTIONS
G.D. and S.N. collected and analyzed the electrophysiology data. M.P. collected
and analyzed the behavioral data. G.D. prepared the figures. A.C. collected
and analyzed the biochemical data. I.F.B. performed animal treatments and
collaborated with M.P. on AGS experiments. L.C. and A.C. designed the
experiments and wrote the manuscript. All authors read and revised the
manuscript.
COMPETING FINANCIAL INTERESTS

The authors declare competing financial interests; details are available in the
online version of the paper.
Reprints and permissions information is available online at http://www.nature.com/
reprints/index.html.
1. Dierssen, M. Down syndrome: the brain in trisomic mode. Nat. Rev. Neurosci. 13,
844858 (2012).
2. Reeves, R.H. et al. A mouse model for Down syndrome exhibits learning and
behaviour deficits. Nat. Genet. 11, 177184 (1995).
3. Costa, A.C. & Grybko, M.J. Deficits in hippocampal CA1 LTP induced by TBS but
not HFS in the Ts65Dn mouse: a model of Down syndrome. Neurosci. Lett. 382,
317322 (2005).
4. Fernandez, F. et al. Pharmacotherapy for cognitive impairment in a mouse model
of Down syndrome. Nat. Neurosci. 10, 411413 (2007).
5. Costa, A.C., Scott-McKean, J.J. & Stasko, M.R. Acute injections of the NMDA
receptor antagonist memantine rescue performance deficits of the Ts65Dn mouse
model of Down syndrome on a fear conditioning test. Neuropsychopharmacology
33, 16241632 (2008).
6. Chakrabarti, L. et al. Olig1 and Olig2 triplication causes developmental brain defects
in Down syndrome. Nat. Neurosci. 13, 927934 (2010).
7. Kleschevnikov, A.M. et al. Hippocampal long-term potentiation suppressed
by increased inhibition in the Ts65Dn mouse, a genetic model of Down syndrome.
J. Neurosci. 24, 81538160 (2004).
8. Kleschevnikov, A.M. et al. Increased efficiency of the GABAA and GABAB receptormediated neurotransmission in the Ts65Dn mouse model of Down syndrome.
Neurobiol. Dis. 45, 683691 (2012).
9. Costa, A.C. & Scott-McKean, J.J. Prospects for improving brain function in
individuals with Down syndrome. CNS Drugs 27, 679702 (2013).
10. Braudeau, J. et al. Specific targeting of the GABA-A receptor alpha5 subtype by a
selective inverse agonist restores cognitive deficits in Down syndrome mice.
J. Psychopharmacol. 25, 10301042 (2011).
11. Szemes, M., Davies, R.L., Garden, C.L. & Usowicz, M.M. Weaker control of the
electrical properties of cerebellar granule cells by tonically active GABAA receptors
in the Ts65Dn mouse model of Downs syndrome. Mol. Brain 6, 33 (2013).
12. Best, T.K., Cramer, N.P., Chakrabarti, L., Haydar, T.F. & Galdzicki, Z. Dysfunctional
hippocampal inhibition in the Ts65Dn mouse model of Down syndrome. Exp. Neurol.
233, 749757 (2012).
13. Mitra, A., Blank, M. & Madison, D.V. Developmentally altered inhibition in Ts65Dn,
a mouse model of Down syndrome. Brain Res. 1440, 18 (2012).
14. Hanson, J.E., Blank, M., Valenzuela, R.A., Garner, C.C. & Madison, D.V.
The functional nature of synaptic circuitry is altered in area CA3 of the hippocampus
in a mouse model of Downs syndrome. J. Physiol. (Lond.) 579, 5367 (2007).
15. Harashima, C. et al. Abnormal expression of the G-protein-activated inwardly
rectifying potassium channel 2 (GIRK2) in hippocampus, frontal cortex, and
substantia nigra of Ts65Dn mouse: a model of Down syndrome. J. Comp. Neurol.
494, 815833 (2006).
16. Cancedda, L., Fiumelli, H., Chen, K. & Poo, M.M. Excitatory GABA action is essential
for morphological maturation of cortical neurons in vivo. J. Neurosci. 27,
52245235 (2007).
17. Deidda, G., Bozarth, I.F. & Cancedda, L. Modulation of GABAergic transmission in
development and neurodevelopmental disorders: investigating physiology and
pathology to gain therapeutic perspectives. Front. Cell. Neurosci. 8, 119 (2014).
18. Ward, O.C. & Lam, L.K. Bumetanide in heart failure in infancy. Arch. Dis. Child.
52, 877882 (1977).
19. Maa, E.H., Kahle, K.T., Walcott, B.P., Spitz, M.C. & Staley, K.J. Diuretics and
epilepsy: will the past and present meet? Epilepsia 52, 15591569 (2011).
20. Fiumelli, H., Cancedda, L. & Poo, M.M. Modulation of GABAergic transmission by
activity via postsynaptic Ca2+-dependent regulation of KCC2 function. Neuron 48,
773786 (2005).
21. Doyon, N. et al. Efficacy of synaptic inhibition depends on multiple, dynamically
interacting mechanisms implicated in chloride homeostasis. PLoS Comput. Biol.
7, e1002149 (2011).
22. Succol, F., Fiumelli, H., Benfenati, F., Cancedda, L. & Barberis, A. Intracellular
chloride concentration influences the GABAA receptor subunit composition.
Nat. Commun. 3, 738 (2012).
23. Rissman, R.A. & Mobley, W.C. Implications for treatment: GABAA receptors in
aging, Down syndrome and Alzheimers disease. J. Neurochem. 117, 613622
(2011).
24. Pueschel, S.M., Louis, S. & McKnight, P. Seizure disorders in Down syndrome.
Arch. Neurol. 48, 318320 (1991).
25. Westmark, C.J., Westmark, P.R. & Malter, J.S. Alzheimers disease and Down
syndrome rodent models exhibit audiogenic seizures. J. Alzheimers Dis. 20,
10091013 (2010).
26. Danglot, L., Triller, A. & Marty, S. The development of hippocampal interneurons
in rodents. Hippocampus 16, 10321060 (2006).
27. Alger, B.E. Gating of GABAergic inhibition in hippocampal pyramidal cells.
Ann. NY Acad. Sci. 627, 249263 (1991).
28. Usowicz, M.M. & Garden, C.L. Increased excitability and altered action potential
waveform in cerebellar granule neurons of the Ts65Dn mouse model of Down
syndrome. Brain Res. 1465, 1017 (2012).
325
29. Wlodarczyk, A.I. et al. Tonic GABAA conductance decreases membrane time
constant and increases EPSP-spike precision in hippocampal pyramidal neurons.
Front. Neural Circuits 7, 205 (2013).
30. Jean-Xavier, C., Mentis, G.Z., ODonovan, M.J., Cattaert, D. & Vinay, L.
Dual personality of GABA/glycine-mediated depolarizations in immature spinal cord.
Proc. Natl. Acad. Sci. USA 104, 1147711482 (2007).
31. Song, I., Savtchenko, L. & Semyanov, A. Tonic excitation or inhibition is set by
GABAA conductance in hippocampal interneurons. Nat. Commun. 2, 376 (2011).
32. Duchon, A. et al. Identification of the translocation breakpoints in the Ts65Dn and
Ts1Cje mouse lines: relevance for modeling Down syndrome. Mamm. Genome 22,
674684 (2011).
33. Lemonnier, E. et al. A randomised controlled trial of bumetanide in the treatment
of autism in children. Transl. Psychiatry 2, e202 (2012).
34. Ward, A. & Heel, R.C. Bumetanide. A review of its pharmacodynamic and
pharmacokinetic properties and therapeutic use. Drugs 28, 426464 (1984).
35. Puskarjov, M., Kahle, K.T., Ruusuvuori, E. & Kaila, K. Pharmacotherapeutic targeting
of cation-chloride cotransporters in neonatal seizures. Epilepsia 55, 806818
(2014).
36. Li, Y. et al. Sensitive isotope dilution liquid chromatography/tandem mass spectrometry
method for quantitative analysis of bumetanide in serum and brain tissue.
J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 879, 9981002 (2011).
37. Cleary, R.T. et al. Bumetanide enhances phenobarbital efficacy in a rat model of
hypoxic neonatal seizures. PLoS One 8, e57148 (2013).
38. Deidda, G. et al. Early depolarizing GABA controls critical-period plasticity in the
rat visual cortex. Nat. Neurosci. 18, 8796 (2015).
39. Dzhala, V.I. et al. NKCC1 transporter facilitates seizures in the developing brain.
Nat. Med. 11, 12051213 (2005).
40. Sipil, S.T., Schuchmann, S., Voipio, J., Yamada, J. & Kaila, K. The cation-chloride
cotransporter NKCC1 promotes sharp waves in the neonatal rat hippocampus.
J. Physiol. 573, 765773 (2006).
41. Kahle, K.T., Barnett, S.M., Sassower, K.C. & Staley, K.J. Decreased seizure activity
in a human neonate treated with bumetanide, an inhibitor of the Na(+)-K(+)-2Cl(-)
cotransporter NKCC1. J. Child Neurol. 24, 572576 (2009).
42. Mazarati, A., Shin, D. & Sankar, R. Bumetanide inhibits rapid kindling in
neonatal rats. Epilepsia 50, 21172122 (2009).
43. Edwards, D.A. et al. Bumetanide alleviates epileptogenic and neurotoxic effects of
sevoflurane in neonatal rat brain. Anesthesiology 112, 567575 (2010).
44. Lemonnier, E. & Ben-Ari, Y. The diuretic bumetanide decreases autistic behaviour
in five infants treated during 3 months with no side effects. Acta Paediatr. 99,
18851888 (2010).
45. Lemonnier, E. et al. Treating Fragile X syndrome with the diuretic bumetanide: a
case report. Acta Paediatr. 102, e288e290 (2013).
46. Tyzio, R. et al. Oxytocin-mediated GABA inhibition during delivery attenuates autism
pathogenesis in rodent offspring. Science 343, 675679 (2014).
47. Hadjikhani, N. et al. Improving emotional face perception in autism with diuretic
bumetanide: a proof-of-concept behavioral and functional brain imaging pilot study.
Autism 19, 149157 (2013).
48. Mares, P. Age- and dose-specific anticonvulsant action of bumetanide in immature
rats. Physiol. Res. 58, 927930 (2009).
49. Brandt, C., Nozadze, M., Heuchert, N., Rattka, M. & Loscher, W. Diseasemodifying effects of phenobarbital and the NKCC1 inhibitor bumetanide in the
pilocarpine model of temporal lobe epilepsy. J. Neurosci. 30, 86028612
(2010).
50. Wang, D.D. & Kriegstein, A.R. Blocking early GABA depolarization with bumetanide
results in permanent alterations in cortical circuits and sensorimotor gating deficits.
Cereb. Cortex 21, 574587 (2011).
51. Validus Pharmaceuticals Bumex: brand of bumetanide tablets. Drugs@FDA: FDA
Approved Drug Products http://www.accessdata.fda.gov/drugsatfda_docs/label/2010/
018225s024lbl.pdf (2008).
52. Pentikinen, P.J., Penttil, A., Neuvonen, P. & Gothoni, G. Fate of [14C]-bumetanide
in man. Br. J. Clin. Pharmacol. 4, 3944 (1977).
53. Ostergaard, E.H., Magnussen, M.P., Nielsen, C.K., Eilertsen, E. & Frey, H.H.
Pharmacological properties of 3-n-butylamino-4-phenoxy-5-sulfamylbenzoic
acid (Bumetanide), a new potent diuretic. Arzneimittelforschung 22, 6672
(1972).
54. Lscher, W., Puskarjov, M. & Kaila, K. Cation-chloride cotransporters NKCC1 and
KCC2 as potential targets for novel antiepileptic and antiepileptogenic treatments.
Neuropharmacology 69, 6274 (2013).
55. Strittmatter, S.M. Overcoming drug development bottlenecks with repurposing: old
drugs learn new tricks. Nat. Med. 20, 590591 (2014).
56. He, Q., Nomura, T., Xu, J. & Contractor, A. The developmental switch in GABA
polarity is delayed in fragile X mice. J. Neurosci. 34, 446450 (2014).
57. Ehninger, D. et al. Reversal of learning deficits in a Tsc2+/ mouse model of tuberous
sclerosis. Nat. Med. 14, 843848 (2008).
58. Han, S. et al. Autistic-like behaviour in Scn1a+/ mice and rescue by enhanced
GABA-mediated neurotransmission. Nature 489, 385390 (2012).
59. Castrn, E., Elgersma, Y., Maffei, L. & Hagerman, R. Treatment of neurodevelopmental
disorders in adulthood. J. Neurosci. 32, 1407414079 (2012).
60. Contestabile, A. et al. Lithium rescues synaptic plasticity and memory in Down
syndrome mice. J. Clin. Invest. 123, 348361 (2013).
npg
Articles
326
ONLINE METHODS
npg
Animals and treatment. 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 with ad libitum
access to water and food. All animal experiments were performed in
accordance with the guidelines established by European Community Council
Directive 2010/63/EU (September 22, 2010) and were approved by the Italian
Ministry of Health.
Ts65Dn and WT mice were generated by repeated backcrossing of Ts65Dn
females (obtained from the Jackson Laboratory) to C57BL/6JEi C3SnHeSnJ
(B6EiC3) F1 males. Animals were genotyped by PCR as previously described32.
Animals aged between 10 and 16 weeks were used for experiments. Both males
and females were used for biochemistry and electrophysiology experiments; only
males were used for behavioral experiments. Ts65Dn and WT littermates were
randomly assigned to bumetanide (Sigma; 0.2 mg/kg body weight) or vehicle
groups (2% DMSO in saline) and treated daily by intraperitoneal (IP) injection.
On the day of behavioral testing, injections were given at least 1 h before the
beginning of the task.
In vitro electrophysiology. Cell-attached and perforated patch-clamp recordings.
To obtain acute hippocampal or cortical slices, we anesthetized mice with isoflurane and transcardially perfused them with an ice-cold cutting solution with
the following composition: 115 mM NaCl, 4 mM MgCl2, 3.5 mM KCl, 1.2 mM
NaH2PO4, 0.5 mM CaCl2, 25 mM NaHCO3 and 25 mM D-glucose (~300 mOsm,
pH 7.4, oxygenated with 95% O2 and 5% CO2). Animals were killed, and brains
were removed and immersed in the cutting solution. Coronal slices (350 m
thick, cut with a VT1000S Leica Microsystems vibratome) were incubated at 35 C
for 10 min in an N-methyl-d-glucamine (NMDG)-HEPES recovery solution61
with the following composition: 93 mM NMDG, 2.5 mM KCl, 1.2 mM
NaH2PO4, 10 mM MgSO4, 0.5 mM CaCl2, 30 mM NaHCO3, 20 mM HEPES,
5 mM sodium ascorbate, 3 mM sodium pyruvate, 2 mM thiourea and 25 mM
d-glucose (~300 mOsm, pH 7.4, oxygenated with 95% O2 and 5% CO2). Slices
were then maintained in an artificial cerebrospinal fluid (ACSF) solution with
the following composition: 115 mM NaCl, 3.5 mM KCl, 1.2 mM NaH2PO4,
2 mM MgCl2, 1.5 mM CaCl2, 25 mM NaHCO3 and 25 mM glucose (~310 mOsm,
pH 7.4, oxygenated with 95% O2 and 5% CO2). Recordings were made from CA1
hippocampal and cortical layer II/III pyramidal neurons at 3032 C.
For cell-attached recordings, glass micropipettes (resistance, 25 M) were
filled with ACSF, and slices were bath perfused with the same solution. After
a giga-ohm seal had been obtained, each cell was recorded under different
conditions depending on the experiments for 4 min each: baseline in ACSF,
GABA (1, 5, 50, 100 and 200 M), GABA 200 M + tetrodotoxin (TTX) 1 M
(Tocris), CGP55845 10 M, GABA 100 M + bumetanide 10 M, and GABA
200 M + bicuculline 10 M (the vehicle solution consisted of ACSF for GABA,
TTX and bicuculline; for CGP55845 10 M and bumetanide it consisted of
DMSO 0.1% or 0.001% in ACSF) . All cell-attached recordings were performed
in the voltage-clamp mode, and the recording pipette was kept at 0 mV for all
such recordings. Data from the whole 4 min of recording were analyzed for
quantification of spiking activity. The dependence of the spiking activity on
voltage-gated Na+ channels was confirmed by TTX application at the end of a
subset of the recordings, which abolished all spikes. Recordings were terminated
and data were discarded if the seal resistance was 1 G. Data, filtered at 0.1 Hz
and 5 kHz and sampled at 25 kHz, were acquired with a patch-clamp amplifier
(Multiclamp 700B, Molecular Devices) and analyzed using pClamp 10.2 software
(Molecular Devices).
For gramicidin-perforated patch-clamp recordings, glass micropipettes
(resistance, 35 M) were tip-filled with an internal solution of 150 mM KCl and
10 mM HEPES (300 mOsm, pH 7.4) and back-filled with the same solution containing 20 g/mL gramicidin A. Slices were bath perfused with ACSF. To determine the ECl in recorded cells, we varied the starting holding potential (90 mV)
in increasing steps of 10 mV and measured the amplitude of inhibitory postsynaptic currents in response to exogenous puffs of GABA (100 M; 2050 ms at
15-ms intervals). In a subset of cells, exogenous GABA puffing was performed
before and during bath perfusion of bumetanide (10 M). Linear regression
was used to plot a line of best fit for the voltage dependence of GABA-induced
currents, and the interpolated intercept of this line with the x-axis was taken as
doi:10.1038/nm.3827
the ECl value. For each experiment, ECl measurements were performed in triplicate at 3-min intervals. For all cells recorded, TTX (1 M; Tocris) was added
to the extracellular solution to block voltage-gated Na+ channels. Recordings
were terminated and data were discarded if the series resistance varied by >15%
during the course of the experiment or if the perforated patch ruptured into a
whole-cell configuration, as evidenced by an ECl of ~0 mV (ref. 16). For the
measurement of resting membrane potential, a whole-cell patch-clamp configuration was achieved, the amplifier was set to I = 0, and the corresponding
potential was measured. Data, filtered at 0.1 Hz and 5 kHz and sampled at
10 kHz, were acquired with a patch-clamp amplifier (Multiclamp 700B, Molecular
Devices) and analyzed using pClamp 10.2 software (Molecular Devices).
All chemicals were purchased from Sigma, unless otherwise specified.
LTP field recordings. To obtain acute hippocampal slices, we anesthetized
mice with isoflurane and transcardially perfused them with an ice-cold ACSF
solution with the following composition: 120 mM NaCl, 3.5 mM KCl, 1.25 mM
NaH2PO4, 1.3 mM MgSO4, 2.5 mM CaCl2, 26 mM NaHCO3, and 10 mM
D-glucose3 (~310 mOsm, pH 7.4, oxygenated with 95% O2 and 5% CO2).
The animal was killed, and the brain was removed and immersed in ACSF. Sagittal
slices (400 m thick, cut with a VT1000S Leica Microsystems vibratome) were
incubated at 35 C. After 2 h of recovery, slices were transferred to a recording
chamber and perfused with ACSF (32 C, 1.7 mL/min) or ACSF containing
bumetanide 10 M, VU0240551 5 M (Tocris), GABA 100 M or vehicle
(DMSO 0.1%0.001% in ACSF for bumetanide and VU0240551). Electrical
stimulation (100-s duration) of Shaffer collateral was achieved with a bipolar
tungsten stimulating electrode (WPI) placed in the stratum radiatum. Field
potentials in CA1 stratum radiatum were recorded by a micropipette (13 M)
filled with ACSF. Baseline responses were obtained every 30 s with a stimulation
intensity that yielded a half-maximal slope (mV/ms) response. After a baseline
had been maintained as stable for 10 min, a theta-burst stimulation (TBS) was
delivered. The protocol for TBS called for one train of TBS, which consisted
of five bursts at 5 Hz. Each burst consisted of four pulses at 100 Hz (ref. 3).
To study basal synaptic transmission, we obtained inputoutput relationships
and paired-pulse ratios after 20 min of stable recording at a stimulation intensity
that yielded a half-maximal response. For the paired-pulse experiment, two
pulses spaced at determined time intervals (50, 100 and 200 ms) were delivered.
The paired-pulse ratio was calculated by dividing the fEPSP response to the
second pulse by the response to the first pulse and multiplying by 100. For each
animal, we recorded slices treated with vehicle and bumetanide, VU0240551
or GABA. Data, filtered at 0.1 Hz and 600 Hz and sampled at 25 kHz, were
acquired with a patch-clamp amplifier (Multiclamp 700B, Molecular Devices)
and analyzed using pClamp 10.2 software (Molecular Devices). All chemicals
were purchased from Sigma, unless otherwise specified.
Two-photon Cl imaging. Imaging of neuronal intracellular Cl in acute
hippocampal slices was performed with the fluorescent indicator MQAE
[N-(ethoxycarbonylmethyl)-6-methoxyquinolinium bromide]. MQAE dye
detects Cl ions via diffusion-limited collisional quenching, resulting in a
concentration-dependent decrease in fluorescence emission after an increase in
Cl concentration62. Brain slices (prepared as described above for patch-clamp
experiments) were incubated at 37 C in the dark for 30 min in oxygenated ACSF
containing 5 mM MQAE (Molecular Probes). Slices were then transferred to a
holding chamber and perfused with oxygenated ACSF at 30 C for 5 min before
imaging. Images were taken with a Leica SP5 (Leica Microsystems) upright
scanning confocal microscope coupled to a Ti:sapphire laser (Chameleon Ultra,
Coherent) and equipped with a water-immersion objective (25, numerical
aperture (NA) 0.95) and a 460/50-nm band-pass emission filter. MQAE was
two-photon excited at 750 nm. All excitation and acquisition parameters were
kept constant throughout the experiments. Image analysis was performed with
LAS-AF software (Leica) by measuring the mean fluorescent intensity of regions
of interest centered on the cell bodies of hippocampal CA1 neurons. Pseudocolor images shown in Supplementary Figure 1d were generated with ImageJ
software (http://rsbweb.nih.gov/ij/).
Behavioral testing. Different cohorts of mice were used for the diverse
behavioral testing. Mice were tested after acute drug injection (1 h before the
beginning of the session), after 1 week of treatment (acquisition on day 6 and trial
nature medicine
npg
on day 7) or during the fourth week of treatment. For washout experiments,

mice were treated daily for 4 weeks and were subsequently left untreated
for 1 additional week before the start of behavioral testing. Vehicle- and
bumetanide-treated mice were always evaluated in parallel and with the same
time schedule. In order to avoid any confounding effects, tests were administrated only once to individual mice. A detailed outline of the order of the tests
for the different experimental cohorts is reported in Supplementary Table 1.
Animal behavior was video-recorded throughout the experimental sessions
and subsequently analyzed by a trained operator blinded to the experimental
groups. Images of the objects used in the OL and NOR tests are reported in
Supplementary Figures 7 and 8.
Behavioral experiments were performed as previously described60.
OL test. The OL test evaluates spatial memory by measuring the ability of mice
to recognize the new location of a familiar object on the basis of the available
extra-maze cues. The test was conducted in a gray acrylic arena (44 44 cm).
Mice were habituated to the chamber for 15 min on the day before testing. The
next day, mice were exposed to two identical objects for 15 min during the
acquisition phase. Object preference was evaluated during this session. Testing
occurred 24 h later in the same arena. During the trial session, one of the objects
was moved to a novel location, and mice were allowed to explore the objects
for 15 min. The objects and the arena were cleaned with 70% ethanol after each
trial. The time mice spent exploring each object was measured. Exploration
was defined as any investigative behavior toward (i.e., head orientation, sniffing
occurring within <1.0 cm) or deliberate contact with an 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 Exploration
Time/Total Exploration Time) 100) ((Old Object Location Exploration
Time/Total Exploration Time) 100).
NOR test. The NOR test was conducted in a gray acrylic arena (44 44 cm).
On the day before the NOR test, mice were allowed to become habituated to
the apparatus by freely exploring the open arena for 15 min. NOR is based on
the preference of mice for a novel object versus a familiar object when they
are allowed to explore freely. The objects used were different in shape, color,
size and material. During the acquisition sessions, three different objects were
placed into the arena, and the mice were allowed to explore for 15 min. Object
preference was evaluated during these sessions. Testing occurred 24 h later in
the same arena. In the test, one of the objects used in the acquisition session was
replaced by a novel object, and mice were allowed to explore freely for 15 min.
The objects were counterbalanced between the sessions and were cleaned with
70% ethanol after each trial. Exploratory behavior toward an object was defined
as direct contact with the object by the animals mouth, nose or paws or as an
instance when the animal approached the object so that its nose was within
1 cm of the object. Any indirect or accidental contact with the objects was not
included in the scoring. The time spent exploring each object, expressed as
a percentage of the total exploration time, was measured for each trial. The
discrimination index was calculated as the difference between the percentage
of time spent investigating the novel object and that of time spent investigating
the familiar objects: Discrimination Index = ((Novel Object Exploration Time/
Total Exploration Time) 100) ((Familiar Object Exploration Time/Total
Exploration Time) 100).
CFC test. The mice were placed in a fear-conditioning system (TSE Systems)
consisting of a transparent acrylic conditioning chamber (20 10 cm) equipped
with a stainless-steel grid floor. Mice were held outside the experimental room
in their home cages before testing and were individually transported to the
conditioning apparatus in standard cages. The chamber was cleaned with 70%
ethanol after each trial. Mice were placed in the conditioning chamber and
received one electric shock (constant electric current; 2 s, 0.75 mA) through
the floor grid 3 min later. Mice were removed 15 s after the shock. Twenty-four
hours later, mice were placed in the same chamber for 3 min (context test), and
2 h later they were moved to a new context (black chamber with gray plastic
floor and vanilla odor). The freezing behavior was scored by a trained operator
blinded to the experimental groups.
Locomotor activity. Spontaneous locomotor activity was evaluated as previously described2 over a 24-h period (12/12 h dark/light cycle) during the
fourth week of treatment. Horizontal activity and stereotyped movements were
nature medicine
automatically evaluated using a VersaMax apparatus (AccuScan Instruments)

equipped with an array of photocell beams. On the day of testing, mice were
administered either vehicle or bumetanide just before the onset of the light phase
(8 a.m.) and dark phase (8 p.m.) of the day.
Audiogenic seizures. Assessment of AGS sensitivity was conducted essentially as
previously described25,63 during the fourth week of treatment. On the day of testing, mice were injected with bumetanide or vehicle 30 min before being introduced
to a gridded cylindrical box (15-cm diameter) located in a sound-attenuating
cubicle equipped with two loudspeakers and a video camera (TSE Systems).
Mice were then exposed for 3 min to a 120-dB white noise and monitored for
seizure induction. AGS severity was scored from 0 to 3 as previously described64:
0, no response; 1, wild running; 2, clonic seizure; and 3, tonic seizure.
In behavioral experiments, we adopted the following exclusion criteria independently of genotype or treatment (before blind code break). In the CFC test,
we excluded mice showing very high non-associative freezing in the new context:
(i.e., more than 30 s of freezing during the 3-min test (2 out of 235 mice)). In the
OL and NOR tests, we excluded animals showing very low explorative behavior
(i.e., less than 10 s of direct object exploration during the 15-min test (7 out of
191 mice for the OL test, and 6 out of 192 mice for the NOR test)).
Biochemistry. Protein extraction. For total protein extraction, hippocampal and
cortical samples were homogenized 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 mM PMSF, 10 mM NaF, 2 mM sodium orthovanadate and 1% (v/v) protease
and phosphatase inhibitor cocktail (Sigma). The samples were clarified by centrifugation at 20,000g, and the protein concentration was determined using a
Bicinchoninic Acid Assay (BCA) kit (Pierce).
Synaptosomal membrane preparation. Subcellular fractionation was performed
as previously described65. Samples were homogenized in buffer A (320 mM
sucrose, 1 mM EDTA, 10 mM Tris, pH 7.4) supplemented with inhibitors as
described above. Homogenates were centrifuged at 800g to remove debris, and
the resulting supernatant (S1) was further centrifuged at 9,200g. The resulting
pellet (P2) was resuspended in buffer B (35.6 mM sucrose, 1 mM EDTA,
10 mM Tris, pH 7.4) and centrifuged at 25,000g. Finally, the resulting pellet (LP1), representing the synaptosomal membrane-enriched fraction, was
solubilized in RIPA buffer.
Biotinylation of cell-surface proteins. Pulldown of biotinylated cell-surface proteins from acute hippocampal slices was performed as previously described66.
Mouse sagittal brain slices (as prepared for LTP field recordings) were first
cooled in oxygenated ASCF at 4 C for 10 min and then incubated at 4 C in
the same solution containing 100 M membrane-impermeable Sulfo-NHSLC-Biotin (ThermoScientific). After 1 h, slices were transferred to oxygenated
ASCF containing 100 M lysine and incubated at 4 C for 10 min to quench
excess biotin. Finally, slices were washed in oxygenated ASCF at 4 C, and the
CA1 hippocampal region was rapidly dissected under a stereomicroscope. To
control for specific pulldown of biotinylated proteins, we processed some slices
in parallel, omitting Sulfo-NHS-LC-Biotin. For each sample, three CA1 regions
obtained from adjacent slices from the same mice were pooled together and lysed
in RIPA buffer (as above). To pull down biotinylated surface proteins, we diluted
80 g of protein lysate for each sample with pulldown buffer (0.1 RIPA in PBS)
and incubated it at 4 C for 3 h on a rotating wheel with NeutrAvidin beads
(ThermoScientific). The beads were then washed three times in cold pulldown
buffer, collected by centrifugation and resuspended in SDSPAGE sample buffer.
Surface biotinylated proteins were analyzed by western blot in parallel with 10%
of the corresponding total input lysate.
Western blotting. For immunoblot analysis, equal amounts of protein were run
on 412% Bis-Tris NuPAGE (Invitrogen) or Criterion-XT (Bio-Rad) gels and
transferred onto nitrocellulose membranes (GE Healthcare). Membranes were
probed with mouse anti-NKCC1 (clone T4, Developmental Studies Hybridoma
Bank; 1:4,000), rabbit anti-KCC2 (Millipore, catalog no. 07-432; 1:4,000), rabbit
anti -actin (Sigma, catalog no. A2066; 1:10,000), mouse anti-Na+/K+-ATPase
(clone C464.6, Millipore, catalog no. 05-369; 1:2,000), mouse anti-APP (clone
22C11, Millipore, catalog no. MAB248; 1:2,000), rabbit anti-GABAAR subunits
1 and 3 (Alomone Labs, catalog nos. AGA-001 and AGA-003, respectively;
1:2,000 and 1:1,000), and rabbit anti-GABAAR subunit (Millipore, catalog
no. AB9752; 1:1,000), followed by HRP-conjugated secondary antibodies goat
doi:10.1038/nm.3827
npg
anti-rabbit and goat anti-mouse (Thermo Scientific, catalog nos. 31460 and
31430, respectively; 1:5,000). Chemiluminescence signals were digitally acquired
on an LAS 4000 Mini Imaging System (GE Healthcare), and the band intensities
were quantified using ImageQuant software (GE Healthcare). In some experiments, membranes were stripped and reprobed with a second antibody. The
specificity of antibodies to NKCC1 and KCC2 was verified on brain samples
from NKCC1- and KCC2-deficient mice, respectively (kind gift of Dr. K. Kaila,
University of Helsinki; Supplementary Fig. 2a,b).
Quantitative RT-PCR. qRT-PCR was performed in accordance with MIQE
guidelines67. RNA was extracted with QIAzol reagent and purified on RNeasy
spin columns (Qiagen). RNA samples were quantified at 260 nm with an
ND1000 Nanodrop spectrophotometer (Thermo Scientific). RNA purity was
also determined by absorbance at 280 and 230 nm. All samples showed A260/280
and A260/230 ratios greater than 1.9. RNA quality and integrity were verified by
microfluidic assay with an Experion RNA Analysis System (Bio-Rad). The RNA
quality indicator (RQI) for mouse samples ranged between 8.2 and 9.5 (RQI
scale: 110). The RQI for human samples ranged between 5.7 and 8.5. Reverse
transcription was performed according to the manufacturers recommendations on 1 g of RNA with the QuantiTect Reverse Transcription Kit (Qiagen),
which includes a genomic DNAremoval step. SYBR green qRT-PCR was performed in triplicate with 10 ng of template cDNA using QuantiTect Master Mix
(Qiagen) on a 7900-HT Fast Real-time System (Applied Biosystems) as previously described68 and using the following universal conditions: 5 min at 95 C,
40 cycles of denaturation at 95 C for 15 s, and annealing/extension at 60 C
for 30 s. Product specificity and occurrence of primer dimers were verified by
melting-curve analysis. Primers were designed with Beacon Designer software
(Premier Biosoft) to avoid template secondary structure and significant crosshomology with other genes by BLAST search. For each target gene, primers were
designed to target all possible transcript variants annotated in the RefSeq database (http://www.ncbi.nlm.nih.gov/refseq). In each experiment, no-template
controls and RT-minus controls were run in parallel to the experimental samples.
The PCR reaction efficiency for each primer pair was calculated via the standard
curve method with four serial-dilution points for cDNA (32, 8, 2 and 0.5 ng). The
PCR efficiency calculated for each primer set was used for subsequent analysis.
All experimental samples were detected within the linear range of the assay.
Gene-expression data were normalized via the multiple-internal-control-gene
method69. To determine an accurate normalization factor for data analysis, we
evaluated the expression stability of different control genes with the GeNorm
algorithm69 available in qBasePlus software (Biogazelle). The tested control genes
were GAPDH (glyceraldehyde-3-phosphate dehydrogenase), PPIA (peptidylprolyl isomerase A), ACTB (-actin) and HPRT1 (hypoxanthine phosphoribosyl
transferase 1). On the basis of the relative expression stability of the control genes
calculated via GeNorm analysis, expression data for the different samples were
normalized as follows: Gapdh and Ppia (mouse cortex and CA1CA3 region),
Hprt and Actb (mouse whole hippocampus), and GAPDH and PPIA (human
hippocampus). However, we noted that the expression stability of the control
doi:10.1038/nm.3827
genes was generally similar; therefore, comparable expression results were also
obtained by normalization with the other control genes.
SYBR Green primer sequences are reported in Supplementary Table 2.
Calibration curve parameters, PCR reaction efficiency and amplicon information are listed in Supplementary Table 3.
Human brain samples. Hippocampal samples from adult humans with Down
syndrome and age- and sex-matched controls were obtained from the Brain
and Tissue Bank for Developmental Disorders at the University of Maryland,
Baltimore, MD. Information on the samples is reported in Supplementary Table 4.
Samples were cryo-pulverized in dry ice (78 C) with a stainless steel mortar.
Aliquots of pulverized tissue were used for RNA or protein extraction.
Statistical analysis. The results are presented as mean sem. Statistical analyses were performed using SigmaPlot (Systat), GraphPad (Prism) or OriginPro
(OriginLab) software. Where appropriate, the statistical significance was
assessed using two-tailed paired or unpaired t-test, two-way ANOVA or
repeated-measure ANOVA followed by all pairwise Tukey or HolmSidak
post hoc testing. If normal-distribution or equal-variance assumptions were not
valid, statistical significance was evaluated using the MannWhitney test, the
Wilcoxon signed rank test, ANOVA or repeated-measure ANOVA on ranks
(nonparametric) followed by all pairwise HolmSidak, StudentNewmanKeuls
or Dunn post hoc testing. AGS data were analyzed by 2 test with Sidak adjustment for multiple comparisons. Frequency-distribution data were analyzed by
KolmogorovSmirnov test. P < 0.05 was considered significant.
The number of animals used for each experiment; the degrees of freedom;
and the number of recorded cells, recorded slices or samples used for statistical
analysis are reported in Supplementary Table 5.
61. Pea, J. et al. Shank3 mutant mice display autistic-like behaviours and striatal
dysfunction. Nature 472, 437442 (2011).
62. Verkman, A.S., Sellers, M.C., Chao, A.C., Leung, T. & Ketcham, R. Synthesis and
characterization of improved chloride-sensitive fluorescent indicators for biological
applications. Anal. Biochem. 178, 355361 (1989).
63. Westmark, C.J. et al. Reversal of fragile X phenotypes by manipulation of AbetaPP/
Abeta levels in Fmr1KO mice. PLoS ONE 6, e26549 (2011).
64. Peterson, S.L. & Albertson, T.E. Neuropharmacology Methods in Epilepsy Research
(CRC Press, 1998).
65. Hallett, P.J., Collins, T.L., Standaert, D.G. & Dunah, A.W. Biochemical fractionation
of brain tissue for studies of receptor distribution and trafficking. Curr. Protoc.
Neurosci. Chapter 1, Unit 1.16 (2008).
66. Thomas-Crusells, J., Vieira, A., Saarma, M. & Rivera, C. A novel method for
monitoring surface membrane trafficking on hippocampal acute slice preparation.
J. Neurosci. Methods 125, 159166 (2003).
67. Bustin, S.A. et al. The MIQE guidelines: minimum information for publication of
quantitative real-time PCR experiments. Clin. Chem. 55, 611622 (2009).
68. Pozzi, D. et al. REST/NRSF-mediated intrinsic homeostasis protects neuronal
networks from hyperexcitability. EMBO J. 32, 29943007 (2013).
69. Vandesompele, J. et al. Accurate normalization of real-time quantitative RT-PCR
data by geometric averaging of multiple internal control genes. Genome Biol. 3,
RESEARCH0034 (2002).
nature medicine

Reversing Excitatory GABAAR Signaling Restores Synaptic Plasticity and Memory in A Mouse Model of Down Syndrome

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Reversing Excitatory GABAAR Signaling Restores Synaptic Plasticity and Memory in A Mouse Model of Down Syndrome

Încărcat de

Drepturi de autor:

Formate disponibile

Articles

Reversing excitatory GABAAR signaling restores synaptic

2015 Nature America, Inc. All rights reserved.

The treatment of Ts65Dn mice with the FDA-approved NKCC1

VOLUME 21 | NUMBER 4 | APRIL 2015 nature medicine

Spike frequency (Hz)

Figure 1 CA1 hippocampal pyramidal neurons

consistent with excitatory signaling of endogenous GABA through

nature medicine VOLUME 21 | NUMBER 4 | APRIL 2015

Spike frequency (Hz)

Cumulative distribution (%)

2015 Nature America, Inc. All rights reserved.

ECl is mainly determined by [Cl]i, which in turn depends on the

Notably, we were able to detect the expected overexpression of the

Spike frequency (Hz)

Spike frequency (Hz)

2015 Nature America, Inc. All rights reserved.

IGABA amplitude (pA)

LTP elicited by theta-burst stimulation (TBS) at Schaffer collateral

VOLUME 21 | NUMBER 4 | APRIL 2015 nature medicine

changes in non-associative freezing (Fig. 5b) or altering sensitivity

nature medicine VOLUME 21 | NUMBER 4 | APRIL 2015

2015 Nature America, Inc. All rights reserved.

WT mice22, we did not observe a significant increase in Ts65Dn mice

Theta-burst stimulation protocol

fEPSP slope (mV ms )

fEPSP slope (% facilitation)

Figure 4 Bath application of the FDA-approved

fEPSP slope (% of baseline)

2015 Nature America, Inc. All rights reserved.

Second, we examined memory in Ts65Dn

VOLUME 21 | NUMBER 4 | APRIL 2015 nature medicine

Contextual fear conditioning

2015 Nature America, Inc. All rights reserved.

GABAergic system is altered at multiple levels in DS17. In particular,

nature medicine VOLUME 21 | NUMBER 4 | APRIL 2015

Novel object recognition

Figure 5 Systemic treatment with bumetanide restores

Time freezing (%)

Time freezing (%)

compartments29. Indeed, because we recorded the GABA-induced

fEPSP slope (% of baseline)

Contextual fear conditioning

Chronic treatment - washout

fEPSP slope (% of baseline)

Time freezing (%)

Figure 6 Effect of bumetanide on memory does

Time freezing (%)

2015 Nature America, Inc. All rights reserved.

to human chromosome 21 and express three

VOLUME 21 | NUMBER 4 | APRIL 2015 nature medicine

2015 Nature America, Inc. All rights reserved.

nature medicine VOLUME 21 | NUMBER 4 | APRIL 2015

COMPETING FINANCIAL INTERESTS

2015 Nature America, Inc. All rights reserved.

VOLUME 21 | NUMBER 4 | APRIL 2015 nature medicine

2015 Nature America, Inc. All rights reserved.

Animals and treatment. A veterinarian was employed to maintain the

2015 Nature America, Inc. All rights reserved.

on day 7) or during the fourth week of treatment. For washout experiments,

automatically evaluated using a VersaMax apparatus (AccuScan Instruments)

2015 Nature America, Inc. All rights reserved.

S-ar putea să vă placă și