Pharmacotherapy for Cognitive Impairment in a Mouse Model of Down
Fabian Fernandez, Wade Morishita, Elizabeth Zuniga, James Ngyuen, Martina
Blank, Robert C. Malenka and Craig C. Garner
Department of Psychiatry and Behavioral Sciences, Nancy Pritzker Laboratory
Stanford University, Palo Alto, CA 94304-5485
To whom correspondence should be addressed: Craig C. Garner, Ph.D.,
Department of Psychiatry and Behavioral Sciences, Nancy Pritzker Laboratory,
Stanford University, 1201 Welch Rd., Palo Alto, CA 94304-5485; Tel. 650-723-
4942; Fax. 650-498-7761; E-Mail: email@example.com
Recent neuroanatomical and electrophysiological findings from a
mouse model of Down syndrome (DS), Ts65Dn, suggest that there is
excessive inhibition in the dentate gyrus, a brain region important for
learning and memory. This circuit abnormality is predicted to compromise
normal mechanisms of synaptic plasticity, and perhaps mnemonic
processing. Here, we show that chronic systemic administration of non-
competitive GABAA antagonists, at non – epileptic doses, leads to a
persistent, post drug, recovery of cognition in Ts65Dn mice, as well as
recovery of deficits in long – term potentiation (LTP). These data suggest
that excessive GABAergic inhibition of specific brain circuits is a potential
cause of mental retardation in DS, and that GABAA antagonists may be
useful therapeutic tools to facilitate functional changes that can ameliorate
cognitive impairment in children and young adults with the disorder.
Ts65Dn mice, like patients with DS, exhibit comprehensive deficits in
declarative learning and memory1. Interestingly, these cognitive deficits are likely
not due to gross abnormalities in Ts65Dn neuroanatomy2, but rather, appear to
derive from selective decreases in the number of excitatory synapses in the
brain3 and corresponding changes in synaptic connectivity4, 5. These findings are
supported by in vitro studies demonstrating that synapses in the Ts65Dn
hippocampus can express normal LTP, but that excessive GABA – mediated
inhibition impairs its induction6, 7. Assuming that triplicated hC21 – related genes
in Ts65Dn mice shift the optimal balance of excitation and inhibition in the
dentate gyrus (and perhaps other brain regions) to a state where excessive
inhibition obscures otherwise normal learning and memory, we hypothesized that
subtly reducing the inhibitory load in the Ts65Dn brain with GABAA receptor
antagonists might rescue defective cognition.
After establishing that the pattern of cognitive impairments in Ts65Dn mice
(3 – 4 months of age) matched those observed in children and young adults with
DS (Supplementary Figure 1 online)8, we proceeded to assess whether a non –
epileptic dose of the non-competitive GABAA antagonist picrotoxin (PTX; i.p., 1.0
mg/kg, a dose used extensively in classic rodent studies on memory
consolidation)9, could improve Ts65Dn object recognition memory. While pilot
studies indicated that a single dose, one day before testing, could not rescue
Ts65Dn object recognition performance, a chronic 2 – week daily regimen
showed obvious beneficial effects (Supplementary Figure 2 online). We
therefore initiated a 4 – week longitudinal, cross – over study. Here, WT and
Ts65Dn mice were randomly assigned to groups receiving daily i.p. injections of
saline or PTX (1.0 mg/kg), and were submitted to four weekly repetitions of
object recognition testing, in which the animals were serially presented with 4
different object sets. At the 2 – week midpoint of this experimental period, WT
and Ts65Dn mice that had been receiving saline were randomly segregated into
groups that continued to receive daily saline injections, or into groups that began
daily injections of PTX. Conversely, WT and Ts65Dn mice that had been
chronically administered PTX in the first fortnight of testing, were now switched
onto a saline regimen. Alongside saline and PTX, we also evaluated the efficacy
of bilobalide (BB; i.p., 5.0 mg/kg)10, a PTX – like compound that could be safely
administered for the whole, 4 – week experimental period.
Not surprisingly, Ts65Dn mice injected with saline during the first 2 – week
period of novel object recognition testing, or those receiving saline over the
course of the whole experimental period, did not exhibit novelty discriminations
significantly above chance (Figs. 1a – 1d). In marked contrast, Ts65Dn mice
treated with PTX during the first or second fortnight, showed normalized object
recognition performance, as did those receiving BB throughout the study (Figs.
1a – 1d). Unexpectedly, Ts65Dn mice that had undergone chronic PTX
administration during the first 2 – week period of novel object recognition testing,
maintained their improved performance when evaluated one and two weeks later
(Figs. 1a – 1d, and Supplementary Table 1 online). Importantly, WT and
Ts65Dn mice did not differ in total object exploration time, spending invariably ~
25% of their experimental sessions investigating objects (Supplementary Table
To extend these findings, we next evaluated the effects of
pentylenetetrazole (PTZ), a non – competitive GABAA antagonist with a long
history of medical use11, on declarative memory in the novel object recognition
test and in a modified spontaneous alternation task. To mimic the most typical
route of drug administration in humans, WT and Ts65Dn were administered PTZ
(3.0 mg/kg in milk; a non – epileptic dose that can be safely given to rodents for
up to a year)12, via voluntary oral feeding (see Methods online). In total, WT and
Ts65Dn mice received 17 daily doses of milk or a milk – PTZ cocktail, and were
submitted to two repetitions of novel object recognition testing, or to three daily T
– maze sessions at the tail end of the treatment regimen. In agreement with
previous results, milk – fed Ts65Dn mice showed an inability to discriminate
object novelty in the object recognition task. PTZ – treated Ts65Dn mice, on the
other hand, exhibited discrimination indices on par with those of WT mice (Fig.
2a and Supplementary Table 1 online). In the spontaneous alternation task,
milk – fed Ts65Dn mice also exhibited a similar pattern of impairment as
untreated Ts65Dn. However, those orally receiving PTZ showed normal levels of
alternation, approximately 70%13 (Fig. 2c – 2d and Supplementary Table 3
online). Notably, WT and Ts65Dn mice or those on PTZ did not differ in object
exploration time in the object recognition task, or exhibit arm biases in the
spontaneous alternation task (Supplementary Tables 2 and 4 online).
To better define the longevity of Ts65Dn cognitive improvement after
GABAA antagonist administration, we subsequently evaluated Ts65Dn mice in
the novel object recognition task, exactly 2 months after the termination of a 17 –
day oral PTZ regimen. Consistent with the post – drug recovery in cognition
observed with PTX, Ts65Dn mice having underwent PTZ administration showed
normal object recognition performance at this time point (Fig. 2b).
The ability of animals to learn and remember is thought to be encoded at
the synaptic level, and involves the ability of synapses to undergo long-term
changes in synaptic strength. Indeed, recent work has provided compelling
evidence that LTP in the hippocampus occurs during learning14, and is required
for memory15. Accordingly, we assessed LTP in the dentate gyrus, a structure
that shows the most pronounced inhibition – related pathology in the Ts65Dn
brain4. Specifically, we determined whether LTP deficits at perforant path
synapses in Ts65Dn mice were rescued by chronic oral PTZ administration at 3 –
4 weeks post – drug treatment, a time window congruent with performance
improvement by Ts65Dn mice in the novel object recognition task post – PTX
treatment. In agreement with these behavioral findings, we found that PTZ –
treated Ts65Dn mice exhibited normalized LTP in the dentate gyrus
approximately 1 month after the cessation of drug administration (Figs. 3a – 3d).
We then assessed the relative permanency of this LTP rescue in Ts65Dn mice,
and found that the Ts65Dn dentate gyrus continued to show greater LTP in PTZ
– treated mice versus milk – fed ones for up to 3 months post – drug regimen
(albeit diminished relative to that of WT mice) (Figs. 3e – 3h), in keeping with
Ts65Dn behavioral improvement 2 months after PTZ administration.
In summary, we have demonstrated that chronic administration of non-
competitive GABAA antagonists (at non – epileptic doses) ameliorates cognitive
deficits in Ts65Dn mice for a period of months extending beyond the window of
drug treatment. Likewise, we have shown that drug – mediated improvements in
Ts65Dn learning and memory are accompanied by rescue of impaired LTP, the
most prominent synaptic correlate of learning and memory in the hippocampus.
These results point to overinhibition, in at least some brain regions, as one
possible mechanism that reduces cognitive performance in a mouse model of DS
(see Supplementary Discussion online), though further experimentation will be
necessary in order to more directly test this mechanism, and to elucidate the
neuroadaptations that are orchestrated in response to repetitive GABAA
antagonist administration. They also highlight the potential clinical utility of non –
competitive GABAA antagonists in DS (including bilobalide and
pentylenetetrazole), providing one window into how cognitive impairment in DS
may be pharmacologically mitigated over time (see Supplementary Discussion
1. Seregaza, Z., Roubertoux, P.L., Jamon, M., and Soumireu-Mourat, B. Behav.
Genet. 36, 387-404 (2006).
2. Holtzman, D.M. et al. Proc. Natl. Acad. Sci. U.S.A. 93, 13,333-13,338 (1996).
3. Kurt, M.A., Davies, D.C., Kidd, M., Dierssen, M., and Florez, J. Brain Res. 858,
4. Belichenko, P.V. et al. J. Comp. Neurol. 480, 281-298 (2004).
5. Hanson, J.E., Blank, M., Valenzuela, R.A., Garner, C.C., and Madison, D.V. J.
Physiol. (in press, 2006).
6. Kleschevnikov, A.M. et al. J. Neurosci. 24, 8153-8160 (2004).
7. Costa, A.C.S., and Grybko, M.J. Neurosci. Lett. 382, 317-322 (2005).
8. Nadel, L. Genes Brain Behav. 2, 156-166 (2003).
9. McGaugh, J.L. Science 153, 1351-1358 (1966).
10. Ivic, L. et al. J. Biol. Chem. 278, 49279-49285 (2003).
11. Levy, S. J. Am. Med. Assoc. 153, 1260-1265 (1953).
12. Landfield, P.W., Baskin, R.K., and Pitler, T.A. Science 214, 581-584 (1981).
13. Gerlai, R. Behav Brain Res. 95, 91-101 (1998).
14. Whitlock, J.R., Heynen, A.J., Shuler, M.G., and Bear, M.F. Science 313,
15. Pastalkova, E. et al. Science 313, 1141-1144 (2006).
Supplementary Information is linked to the online version of the paper at
We thank the Down syndrome Research and Treatment Foundation (DSRTF),
the Hillblom Foundation, the National Science Foundation (NSF), the National
Institute of Health (NIH), Jax West Laboratories, as well as the Stanford Down
syndrome Center, and W.C. Mobley for their support.
Figure 1. Picrotoxin (PTX) and bilobalide (BB) rescue Ts65Dn performance
in the novel object recognition task.
Discrimination Indices (DI’s) of WT and Ts65Dn mice involved in a 4 week
cross – over study. (a) Animals in the first two weeks were either left untreated,
or received daily injections of saline, PTX or BB. While untreated and saline
treated Ts65Dn mice do not exhibit a novel object preference during the first
fortnight (i.e., DI > 0; t17 = 0.7737, P > 0.4497; and t16 = 0.8169, P > 0.4260), PTX
and BB treated Ts65Dn mice show an ability to discriminate object novelty (t9 =
4.083, P < 0.003; and t15 = 4.390, P < 0.001). (b) Saline treated Ts65Dn mice
given PTX during the second period of object recognition testing (Sal → PTX),
start out with the same deficits as those continuing to receive saline in the
second fortnight (Sal → Sal), suggesting that there is no sampling bias for
animals in subsequent treatment groups. (c) During the second two weeks,
saline treated Ts65Dn mice switched to PTX discriminate novel objects similarly
to WT mice (t8 = 3.756, P < 0.006). Surprisingly, PTX treated Ts65Dn mice
switched to saline in the second fortnight also maintain their ability to discriminate
novel objects versus familiar ones (t6 = 3.250, P < 0.02), suggesting a persistent
change in brain function. (d) Compilation of WT and Ts65Dn mouse novelty
discrimination scores with no treatment or treatment with saline, PTX or BB
demonstrating that PTX and BB normalize Ts65Dn object recognition memory
(F5, 187 = 5.204, P < 0.0002; all post hoc comparisons with Ts65Dn control, P’s <
0.05; all other post hoc comparisons, P’s > 0.05). Note that control observations
are pooled from untreated and saline treated (PTX naïve) mice, while PTX
observations are pooled from mice treated in either the first or second fortnight.
Discrimination scores for each condition are tabulated and defined in
Supplementary Data Table 1 online. Error bars represent SEM. All experimental
procedures were approved by the Stanford University Institutional Animal Care
and Use Committee (IACUC), and were conducted in compliance with the NIH
Guide for the Care and Use of Laboratory Animals. See Supplementary
Methods online for experimental details.
(weeks 1 - 2)
Saline - Saline
Saline - PTX
PTX - Saline
BB - BB
(weeks 3 - 4)
(weeks 1 - 2)
Figure 2. Pentylenetetrazole (PTZ) elicits long-lasting cognitive
improvement in Ts65Dn mice.
Novelty discrimination scores (DI’s) of WT and Ts65Dn mice directly
following a ~ 2 week treatment with PTZ (a), or two months post treatment (b).
(a) Although Ts65Dn mice on milk do not exhibit a net novelty preference (t17 =
1.099, P > 0.28), those receiving PTZ perform as well as WT mice receiving
either milk or PTZ (F3, 71 = 3.356, P < 0.03; all post hoc comparisons with Ts65Dn
on milk, P’s < 0.05; all other post hoc comparisons, P’s > 0.05). (b) The
normalized object recognition memory shown by Ts65Dn mice immediately post
treatment, is sustained at two months post drug treatment (F3, 38 = 5.134, P <
0.005; all post hoc comparisons with Ts65Dn previously on milk, P’s < 0.05; all
other post hoc comparisons, P’s > 0.05). Discrimination scores are tabulated in
Supplementary Data Table 1 online. (c & d) Alternation scores (%’s) of WT and
Ts65Dn mice across 3 – 6 sessions of testing in the spontaneous alternation task.
In contrast to WT mice, which exhibit optimal alternation percentages (~ 70%),
untreated or milk treated Ts65Dn mice exhibit significantly lower alternation
percentages (t83 = 5.051, P < 0.0001). However, PTZ normalizes Ts65Dn
alternation scores to WT levels (F3, 272 = 5.998, P < 0.0006; all post hoc
comparisons with Ts65Dn control, P’s < 0.05; all other post hoc comparisons, P’s
> 0.05). Alternation scores for each condition are tabulated and defined in
Supplementary Data Table 3 online. Error bars represent SEM.
Short - term PTZ evaluation
Long - term PTZ evaluation
Short - term PTZ evaluation
Figure 3. PTZ rescues LTP at medial perforant path-granule cell synapses
in Ts65Dn mice.
(a & b) Averaged data for LTP induced in WT (a) or Ts65Dn mice (b)
treated with milk (WT Milk LTP: 115 ± 4.3%, filled circles, 2 mice, n = 7 slices;
Ts65Dn Milk LTP: 104 ± 3.2%, filled squares, 2 mice, n = 7 slices ) or PTZ (WT
PTZ LTP: 110 ± 2.9%, open circles, 3 mice, n = 9 slices; Ts65Dn PTZ LTP: 113 ±
2.1%, open squares, 3 mice, n = 9 slices), evaluated ~ 1 month after the
cessation of drug administration (For comparison, data from milk treated WT
mice are also shown in b; circles). (c & d) Cumulative probability plots of LTP
observed in WT (c) or Ts65Dn mice (d) fed milk (black line), or PTZ (gray line). (e
& f) Averaged LTP graph for WT (e) or Ts65Dn mice (f), 2 – 3 months after
discontinuation of milk (WT Milk LTP: 117 ± 3.6%, filled circles, 5 mice, n = 14
slices; Ts65Dn Milk LTP: 108 ± 2.1%, filled squares, 3 mice, n = 12 slices) or
PTZ treatment (WT PTZ LTP: 115 ± 2.9%, open circles, 5 mice, n = 16 slices;
Ts65Dn PTZ LTP: 113 ± 2.1 %, open squares, 4 mice , n = 13 slices; For
comparison, data from milk treated WT mice are also shown in f; circles). (g & h)
Cumulative probability plots of average LTP for WT (g) or Ts65Dn mice (h)
previously fed milk (black line) or PTZ (grey line). Sample traces in graphs a, b, e,
and f are averaged from 10 consecutive fEPSPs taken at the indicated time
points. Accompanying scale bars are 1 mV, 5 ms. Values are expressed as
mean ± SEM.
Supplementary Information Fernandez et al.
Supplementary Data Figure 1: Ts65Dn mice exhibit declarative memory
problems, but intact procedural learning.
Supplementary Data Figure 2: Chronic but not acute administration of
picrotoxin (PTX) normalizes Ts65Dn performance in the novel object
Supplementary Data Table 1: Tabulated discrimination indices (DI’s) in the
novel object recognition task.
Supplementary Data Table 2: Picrotoxin (PTX), bilobalide (BB) and
pentylenetetrazole (PTZ) do not affect exploration times in the novel object
Supplementary Data Table 3: Tabulated alternation percentages (%) in the
spontaneous alternation task.
Supplementary Data Table 4: Pentylenetetrazole (PTZ) does not affect arm
entries in the spontaneous alternation task.
Supplementary Discussion: Potential utility of non-competitive GABAA
receptor antagonists for the treatment of cognitive impairment in Down
Supplementary Reference list