Discussion

Previous studies have indicated the significance of variations in the daily schedule
of cocaine within the conditioning phase, rather than the dose of cocaine, on the
development of drug-associated memory (Itzhak & Anderson, 2012; Conrad et al., 2013).
While conditioning by fixed daily doses of cocaine resulted in relatively low magnitude
of place preference and rapid extinction, conditioning by escalating doses of cocaine
resulted in higher magnitude of place preference, and resistance to extinction by reexposure
to nonreinforced context (Itzhak & Anderson, 2012; Liddie et al., 2012). Given
these observations, the present study was undertaken to investigate whether Fix-C and
Esc-C memory engaged different neural pathways in the formation and reconsolidation of
cocaine-associated memory.
Since we were interested in the effect of cocaine-context-associated learning
following Fix-C and Esc-C schedules, we focused on the expression levels of NMDAR
subunits. The NMDAR is thought to mediate synaptic plasticity and learning and
memory (Collingridge, 1987) and different subunits are coupled to specific downstream
signaling molecules (Chen et al., 2007). Therefore, the investigation of NMDAR subunits
expression could elucidate differences between Fix-C and Esc-C memory (Itzhak &
Anderson, 2012). We found that conditioning by Esc-C results in increased hippocampal
expression of both Grin2b mRNA and NR2B protein, suggesting an induction of both
transcription and translation of NR2B subunits of the NMDAR. However, in the Fix-C
group we did not detect an increase in Grin2b mRNA and only an increase in NR2B
protein was observed. The lack of increase in Grin2b mRNA in the Fix-C group is
unclear but studies have shown varied correlation between mRNA and protein levels; a
result of biological factors which influence transcription and translation processes (Nie et
al., 2006). Other studies have shown NR2B protein up-regulation in rat nucleus
accumbens following repeated cocaine administration (Huang et al., 2009) and in rat
hippocampus following morphine-induced CPP (Ma et al., 2006). Interestingly, however,
we observed that the increase in expression levels of NR2B in mice conditioned by the
Esc-C schedule was higher than in mice conditioned by the Fix-C schedule. This finding
may be relevant to the strength of Esc-C memory compared to Fix-C memory.
Evidence suggests that the NR2B subunit of NMDAR has potential to carry
greater calcium current per unit charge (Sobcyk et al., 2005) which may confer a greater
influence on downstream signaling cascades that affect synaptic plasticity and learning
and memory. For example, rats over-expressing NR2B in the cortex and hippocampus
showed improved performance in a number of learning and memory tasks (Wang et al.,
2009), while pharmacological or genetic blockade of the NR2B subunit in the cingulate
cortex of mice impaired the formation of contextual fear memory (Zhao et al., 2005).
Additionally, recent reports suggest that repeated cocaine administration generates silent
synapses (Huang et al., 2009). Silent synapses contain higher levels of NR2B-containing
NMDARs compared to neighboring synapses and are capable of undergoing rapid
metaplasticity to strengthen synapses (Lee et al., 2010). Hence, it is plausible that the
increased expression of NR2B in the present study may have contributed to the
development of a more ‘stable’ Esc-C memory compared to Fix-C. A trend toward a
reduction in levels of NR2A (Fig 2.1D) may indicate a switch in NMDAR subunit
composition in Esc-C mice where NR2B replaces NR2A.
Because of the changes in expression of NR2B, we investigated whether selective
antagonism of NR2B-containing NMDAR could disrupt cocaine-associated memory. The
NR2B antagonist ifenprodil a) attenuated acquisition and b) reduced subsequent place
preference expression when administered following memory retrieval in both Fix-C and
Esc-C groups. The former result is in accordance with others who showed that ifenprodil
prevented the development of cocaine (Kiraly et al., 2011) and morphine (Ma et al.,
2006) CPP following fixed administration schedules. The latter result was likely due to
disruption of memory reconsolidation since administration of ifenprodil either a)
following retrieval but outside the reconsolidation window or b) in the absence of
memory retrieval, had no effect on subsequent CPP expression (Fig. 2.4D). The absence
of place preference in the acquisition and reconsolidation experiments was not due to an
aversive influence of ifenprodil since a dose of 10mg/kg ifenprodil is neither rewarding
nor aversive (Suzuki et al., 1999). While we cannot completely rule out extinction
learning, it is feasible to assume that ifenprodil disrupted memory reconsolidation since
we have previously shown that mice conditioned by Esc-C do not exhibit extinction with
few unreinforced exposures to the CPP context (Itzhak & Anderson, 2012). We
corroborated our findings by demonstrating that another NR2B antagonist traxoprodil
was similarly effective at disrupting the reconsolidation of both Fix-C and Esc-C
memory. Taken together, we demonstrate that NR2B-containing NMDARs play a
behaviorally significant role in the development and reconsolidation of cocaineassociated
memory independent of the schedule of conditioning.
However, the NMDAR antagonist MK-801 had differential effects on
reconsolidation of Fix-C and Esc-C memory and it appears that the dependence of
NMDAR in the process of memory reconsolidation is temporally mediated. First, postretrieval
administration of low dose MK-801 (0.1mg/kg) disrupted Fix-C memory
reconsolidation while a higher dose (0.3mg/kg) of the NMDAR antagonist was required
to disrupt Esc-C memory reconsolidation. The increased expression of NR2B subunit in
the Esc-C group, relative to the Fix-C group, may be associated with facilitated calcium
entry (Sobcyk et al., 2005) and therefore a higher dose of the NMDAR antagonist was
required to inhibit downstream signaling molecules involved in memory reconsolidation.
Second, pre-retrieval administration of MK-801 a) prevented extinction of Fix-C CPP
since saline-treated mice showed reduced place preference while CPP was maintained in
MK-801-treated mice and b) had no effect on Esc-C CPP (Fig. 3). This finding implies
that the Fix-C group, but not the Esc-C group, was undergoing extinction learning
(reduction in preference for the cocaine-paired compartment), which was prevented by
MK-801. This premise is support by our previous observations that Fix-C memory is
more susceptible to extinction compared to Esc-C memory (Itzhak & Anderson, 2012).
Similar to the current observations, a) post-retrieval antagonism of the NMDAR
disrupted reconsolidation of object recognition (Akirav & Mamoun, 2006) and odorreward
memory (Torras-Garcia et al., 2005) and b) pre-retrieval antagonism of NMDAR
prevented extinction of conditioned freezing response (Ben Mamou et al., 2006). Other
studies demonstrated that pre- but not post-retrieval systemic administration of MK-801
prevented memory reconsolidation in a conditioned reinforcement behavioral task
(Milton et al., 2008). This discrepancy in temporal-dependence of MK-801 may be due to
differences in behavioral paradigms which investigate instrumental (self-administration)
versus non-operant memory tasks.
Further differences between Fix-C and Esc-C memory are associated with
NMDAR downstream signaling molecules. Since the activity of nNOS is coupled to
calcium influx through the NMDAR (Christopherson et al., 1999; Sattler et al., 1999) we
investigated whether the observed differences in behavioral phenotype between Fix-C
and Esc-C was a function of nNOS involvement. First, the nNOS inhibitor 7-NI
prevented the formation of Fix-C memory but not Esc-C memory (Fig 2). Second,
reconsolidation of Fix-C but not Esc-C memory was disrupted by the nNOS inhibitor.
The latter confirm our previous studies on the dependency of Fix-C memory on NO
signaling (Itzhak & Anderson, 2007) but reveals now a NO-independent signaling
mechanism for Esc-C memory. This differential effect of nNOS inhibition may be due to
the recruitment of additional signaling pathways in response to the more salient
conditioning schedule (Esc-C). Hence, although nNOS could potentially be activated in
response to training by Esc-C, the activation of other signaling pathways may
overshadow the involvement of nNOS.
Calcium influx through NMDARs has the potential to activate several calciumdependent
signaling pathways; for instance, the NMDAR-RasGRF1-MEK-ERK
pathway. ERK is an important regulator of neuronal plasticity, long-term potentiation
(LTP) and LTM formation (Krapivinsky et al., 2003). Since RasGRF1 specifically binds
the NR2B subunit of the NMDAR, it couples the activity of ERK with NR2B-containing
NMDARs (Krapvinisky et al., 2003). Studies have shown that associative learning leads
to an up-regulation of ERK in the hippocampus and inhibition of the ERK kinase MEK,
disrupted the formation of fear memory (Atkins et al., 1998). Furthermore, it has been
suggested that NR2A- and NR2B-containing NMDARs are coupled to different signaling
pathways; activation of NR2A increased brain derived neurotrophic factor (BDNF)
expression while activation of NR2B-containing NMDARs led to phosphorylation of
ERK (Chen et al., 2007).
Because the CPP paradigm involves associative learning and we observed
differential expression of NR2B subunits, we investigated whether reduced ERK
activation via MEK inhibition would disrupt memory reconsolidation of Fix-C and Esc-
C. We found that MEK inhibition had differential effects on Fix-C and Esc-C memory
where reconsolidation of Esc-C but not Fix-C memory was disrupted by SL327. Because
NMDAR-dependent ERK phosphorylation is coupled to NR2B, inhibition of MEK is
expected to have a greater effect on NMDAR-dependent ERK signaling where levels of
NR2B are substantially elevated; that is, in Esc-C group. The modest increase in NR2B in
the Fix-C group may have been sufficient to allow sensitivity to ifenprodil treatment but
insufficient to promote a robust NMDAR-dependent activation of ERK. Contrary to our
findings where MEK inhibition had no effect on Fix-C memory reconsolidation, bilateral
injections of U0126, another MEK inhibitor, into nucleus accumbens core disrupted
expression and reconsolidation of Fix-C CPP (Miller & Marshall, 2005). This apparent
discrepancy may be due to differences in the route of inhibitor administration
(intracerebral vs intraperitoneal) and duration of conditioning (9 days vs 4 days). Others
have reported that systemic administration of SL327 in a fixed-dose cocaine CPP
paradigm effectively disrupted memory reconsolidation only when memory retrieval
occurred by pairing the context with cocaine (Valjent et al., 2006). This finding suggests
that the conditioned stimulus alone (CPP context) was insufficient to engage MEK
signaling in Fix-C but instead it required the much stronger unconditioned stimulus
(cocaine). Thus in the present study, SL327 may have been ineffective at disrupting Fix-
C memory reconsolidation because subjects were exposed to the CPP context alone.
However, the MEK inhibitor was effective in disrupting reconsolidation of Esc-C
memory. This finding suggests that context re-exposure of Esc-C (but not Fix-C) mice
was sufficient to invoke ‘strong memory’ of drug-context association which engaged
MEK-signaling.
While the nNOS signaling pathway may also be activated in response to training
by Esc-C, it appears that other signaling pathways including NMDAR-MEK-ERK
signaling plays a more behaviorally significant role in the development of Esc-C CPP.
Additionally, though both NO-cGMP-PKG and MEK signaling pathways converge at the
level of ERK (Ota et al., 2008) we posit that the contribution of each pathway to drug
memory is dependent on cocaine conditioning schedule.
We focused on identifying molecular changes in the hippocampus because of its
role in contextual memory. While our results lend credence to the involvement of the
hippocampus with respect to changes in NR2B receptor subunit expression, downstream
NMDAR signaling molecules in other brain such as nucleus accumbens, amygdala and
prefrontal cortex may also be involved in the behavioral effects we observed.
In summary, we show that the salience of cocaine reward influences memory
strength by engaging different neural pathways. The NR2B subunit of NMDARs and
MEK-associated signaling appears to have a major role in drug memory acquired by
escalating dose of cocaine, while NMDAR and NO-associated signaling appear to be
involved in drug memory encoded by the Fix-C schedule. Given that drug addiction is
associated with escalation in drug use, we posit that different schedules of drug exposure
may result in differential strength in drug memory which could be relevant to the severity
of addiction. Our data opens the door to further investigations into the effects of drug
memory strength and its differential susceptibility to pharmacological manipulation.


Leave a Reply