Impaired fear memory, altered object memory and modified hippocampal synaptic plasticity in split-
By: Peter MacPherson, Ruth McGaffigan, Douglas Wahlsten, Peter V. Nguyen
MacPherson, P., McGaffigan, R., Wahlsten, D., and Nguyen, P.V. (2008) Impaired fear memory, altered object
memory and modified hippocampal synaptic plasticity in split-brain mice. Brain Research, 1210: 179-188.
Made available courtesy of Elsevier: http://www.elsevier.com
***Reprinted with permission. No further reproduction is authorized without written permission from
Elsevier. This version of the document is not the version of record. Figures and/or pictures may be
missing from this format of the document.***
The hippocampus is critical for memory formation. However, the contributions of the hippocampal commissure
(HC) and the corpus callosum (CC) are less clear. To elucidate the role of the forebrain commissures in learning
and memory, we performed a behavioural and electrophysiological characterization of an inbred mouse strain
that displays agenesis of the CC and congenitally reduced HC (BTBR T+ tf/J; „BTBR‟). Compared to a control
strain, BTBR mice have severely impaired contextual fear memory, with normal object recognition memory.
Interestingly, continuous environmental “enrichment” significantly increased object recognition in BTBR, but
not in control C57BL/6 („BL/6‟) mice. In area CA1 of hippocampal slices, BTBR displayed intact expression of
long-term potentiation (LTP), paired-pulse facilitation (PPF) and basal synaptic transmission, compared to BL/6
mice. However, BTBR hippocampal slices show an increased susceptibility to depotentiation (DPT), an
activity-induced reversal of LTP. We conclude that the HC and CC are critical for some forms of hippocampal
memory and for synaptic resistance to DPT. Agenesis of the CC and HC may unmask some latent ability to
encode, store or retrieve certain forms of recognition memory. We suggest that the increased susceptibility to
DPT in BTBR may underlie the memory phenotype reported here.
Keywords: Hippocampal commissure, Corpus callosum, Memory, Synaptic plasticity, LTP, Mouse strain
Systematic surveys of multiple inbred strains have identified extreme phenotypes that can enhance our
understanding of brain function. Unique among all mouse strains studied to date (Wahlsten et al., 2003b), the
BTBR T+ tf/J („BTBR‟) strain always lacks a corpus callosum (CC) and has severely reduced hippocampal
commissure (HC). Two X chromosome regions contribute to the anatomical defect of the BTBR forebrain
(Kusek et al., 2006). Although there is a significant but small increase in the number of unmyelinated axons in
the anterior commis-sure when the CC is absent (Livy et al., 1997), this compensatory increase is dwarfed by
the massive loss of connectivity between the hemispheres in BTBR mice.
Here we report the first electrophysiological study of the hippocampus in these animals and find that most
measures are remarkably normal. On two very different tests of mouse memory, BTBR mice show substantial
deficits in contextual fear conditioning but not object recognition memory. These findings indicate that the
commissural system is not important for most aspects of normal ipsilateral function of the hippocampus but
may nonetheless be involved in the formation of certain kinds of long-term memories in these mice.
Clinically, CC agenesis in humans is associated with mental retardation, developmental delay, cerebral palsy,
and schizophrenia (Serur et al., 1988; Motomura et al., 2002; Chinnasamy et al., 2006). CC size in humans
show both X-linked and autosomal patterns of inheritance, and CC development depends on a number of
cellular and molecular mechanisms (Richards et al., 2004). The hippocampal commissure (HC) connects the
two hippocampi, structures that are critically involved in memory (Milner et al., 1998; Andersen et al., 2007).
Somewhat surprisingly, patients with transection of the CC display relatively normal memory function (Ledoux
et al., 1977; for review, see Clark and Geffen (1989)). It has been reported that patients whose corpora callosum
and hippocampal commis-sure have been surgically sectioned show deficits in short-term memory (Zaidel and
Sperry, 1974) as well as in recall memory (Phelps et al., 1991). Additionally, individuals with callosal agenesis
show some specific memory deficits (Finlay et al., 2000). Therefore, the integrity of the CC and HC seem to be
required for normal memory function in humans. Interpretation of the human data is difficult, however, because
studies of callosal agenesis often use subjects with other serious brain abnormalities or comorbid conditions
(Bloom and Hynd, 2005). Furthermore, electrophysiological study of hippocampal function in humans is
currently not feasible. Hence, a mouse model may help to elucidate the role of the CC and HC in memory.
No BTBR T+ tf/J brain had any CC present at midplane (Fig. 1B) and only 1 of 58 brains had a HC of normal
adult size. Most BTBR brains had either no detectable HC or a commissure that was greatly reduced in size to
less than 25% of the normal cross-sectional area.
2.2.1. BTBR shows impaired fear memory at all intervals
Are callosal and commissural inputs required for fear learning? Contextual fear conditioning depends on the
hippocampus (Kim and Fanselow, 1992; Chen et al., 1996) and the amygdala (Phillips and Ledoux, 1992; Kim
et al., 1993).
The contextual fear data are summarized in Table 1 and discussed below. In the present study, BTBR mice
displayed deficient short-term and long-term memories for contextual fear (Fig. 1C). In the fear conditioning
chamber, BL/6 froze slightly, but significantly, more than BTBR both before the tone (Table 1, Pre-CS; p <
0.05) and after the shock (Post-US; p < 0.05) when all data from Table 1 were pooled. BTBR froze significantly
less than BL/6 when re-exposed to the training chamber at 1 h for a short-term memory test (p <0.01). Long-
term fear memory was also impaired in BTBR at 24 h (p < 0.001) as well as at 48 h (p < 0.001). Whereas the
BL/6 displayed robust contextual fear learning, BTBR mice displayed only very modest levels of freezing at
these time points.
The difference in Pre-CS values suggests that there might be a slight difference in activity level or anxiety
behaviour between the two strains. In order to compensate for this potential confound, we subtracted the Pre-CS
freezing values for each strain from subsequent intervals as in Schimanski et al. (2002). This correction is
shown in Fig. 1C. Following this correction, BL/6 still froze significantly more than BTBR at 1, 24 and 48 h,
but there was no significant difference between the Pre-CS and Post-US values. In fact, the significance levels
of the testing intervals shown in Fig. 1C were unaltered by the correction. Additionally, Wahlsten et al. (2006)
reported nearly identical open field activity levels for BL/6 and BTBR. Thus, the freezing deficits observed
cannot be attributed to a difference in activity levels. BL/6 and BTBR also display similar behaviour in the
elevated plus maze and do not differ significantly in thigmotactic behaviour in an open field test (D. Wahlsten,
Mouse Phenome Database, www.jax.org/phenome). This suggests that there are no significant differences in
anxiety levels between BTBR and BL6. We therefore conclude that BTBR shows marked deficits in contextual
fear memory compared to BL/6.
2.2.2. Object recognition memory is spared in BTBR and continuous environmental "enrichment"
enhances object recognition in BTBR but not in BL/6
The novel object recognition task is a one-trial memory test in rodents. In this test, memory of a familiar object
is manifested as preferential exploration of novel objects. This task quantifies a naturalistic rodent behaviour in
a non-stressful environment without primary reinforcing stimuli and is similar to visual recognition tests used in
non-human primates (Ennaceur and Delacour, 1988). The neural substrates of the object recognition task have
been reviewed elsewhere (see Dere et al. (2007)).
The object recognition results are summarized in Tables 2 and 3 and discussed below. One-way analysis of
variance (ANOVA) showed that the four groups of mice presented in Fig. 1D and E differed significantly in
terms of exploratory preference and exploratory time (p < 0.05 for both comparisons).
What is the role of the CC and HC in object recognition memory? We found that BTBR and BL/6 displayed
comparable but rather weak evidence of object recognition memory at 24 h in the absence of enrichment (p >
0.4; Fig. 1D). Therefore, intact HC and CC are not required for object recognition memory in mice. Compared
to chance performance (hypothetical mean of 50%), BL/6 displayed significant object recognition memory at 24
h. In BTBR, there was a tendency towards exploration of the novel object at 24 h, but this did not reach
significance (p = 0.18). Thus, both strains seemed to preferentially explore the novel object at 24 h, but only
BL/6 performed significantly above chance.
Wild-type mice placed in a large, “enriched” environment for several hours each day display enhanced
performance in the object recognition task (Tang et al., 2001). In contrast, we examined the effect of continuous
environmental “enrichment” (EE) presented in the home cage for 8 days prior to training.
In BL/6, continuous environmental “enrichment” did not significantly alter performance in the object
recognition task (p>0.7). The discrepancy between this result and other reports of increased object recognition
memory following “enrichment” (e.g. Tang et al., 2001) is likely attributable to the marked differences in
“enrichment” protocols used. The “enrichment” protocol used here presented mice with much less stimulation
than many reports in the literature, over a longer period. Interestingly, BTBR displayed a marked enhancement
of object recognition memory after “enrichment” (Fig. 1D). After “enrichment”, BTBR had a significantly
higher exploratory preference than the “enriched” BL/6 group (p<0.05). Continuous “enrichment” also
significantly improved the object memory of BTBR mice from the non-“enriched” level (p<0.01). Additionally,
“enriched” BTBR mice performed significantly above chance (p<0.001). In comparison, BL/6 mice seemed to
preferentially explore the novel object, but this tendency did not reach statistical significance (p=0.06). Our
environmental “enrichment” paradigm therefore selectively increased object recognition memory in BTBR, but
not in BL/6.
Without continuous “enrichment”, BTBR and BL/6 displayed similar exploratory times at 24 h (p>0.8). After 8
days of enrichment, both strains seemed to display an attenuation of exploratory time (Fig. 1E). The exploratory
time of the “enriched” BL/6 group was significantly lower than their non-“enriched” counterparts (p <0.05).
The tendency towards an attenuation of exploratory time in the “enriched” group was not significant in BTBR
(p>0.1). Continuous exposure to novelty therefore seems to diminish the time spent exploring objects.
2.3. Electrophysiological analysis
Synaptic plasticity, including LTP, can regulate some forms of learning and memory in the mammalian brain
(for reviews, see Bliss and Collingridge (1993), and Martin et al. (2000)). LTP is an activity-dependent
enhancement of excitatory synaptic strength that is induced by high-frequency electrical stimulation (Bliss and
Lømo, 1973). Recently, learning was demonstrated to induce LTP in vivo (Whitlock et al., 2006). We used
various electrical stimulation protocols in the Schaeffer collateral (SC)-CA1 pathway of the hippocampal slices
of BTBR mice using BL/6 slices as a control.
2.3.1. Three forms of LTP are intact in BTBR
In hippocampal slices, one train of HFS (100 Hz for 1 s duration) results in early-LTP (E-LTP). E-LTP decays
to baseline within 1–2 h of induction, and requires activation of NMDA receptors (Collingridge et al., 1983) and
certain protein kinases (Malinow et al., 1989; Silva et al., 1992), but not protein synthesis (Huang and Kandel,
1994). One-train HFS elicited very similar mean fEPSP slope values at 50 min after HFS in the two strains
(118± 8%, for 7 C57BL/6 mice, 115±7% for 7 BTBR, p>0.7; see Fig. 2A).