Conditioned food aversion long-term memory formation and amnesia after memory reconsolidation impairment
Untrained snails had average consummatory reactions times of 20 – 35 s to raw and boiled carrots (Fig. 1B; Test 1). We observed the superiority in CS reactions time compared to the same in DS (p<0.05, Wilcoxon test) and in CS recorded the day before (p<0.05, Wilcoxon test). Consummatory responses to CS (raw carrots) were significantly slower than those to DS (boiled carrots; p<0.0001; Wilcoxon test) and to raw carrots presented before training (p<0.0001; Wilcoxon test). Five to six presentations of unconditioned and conditioned stimuli were required for long-term memory formation in three days.
Two days after training, snails of the control group (n=20) were injected with saline and presented with a reminder. Animals tests 10 and 45 days after injections paired with the reminder showed (Fig. 1, B, T3, and T4) the superiority in CS reactions time compared to the same in DS (after 10 and 45 days z>5.1, p<0.0001; Wilcoxon test) and in raw carrots prior to training procedure (after 10 and 45 days: z>5.03 p<0.0001; Wilcoxon test).
To snails of another group (n=16), we injected MK-801 2 days after training and performed a reminder procedure (Fig. 1, A). Testing snails 10 or 45 days after pairing of MK-801 and reminder revealed the amnesia development (Fig. 1, B, T3, and T4): in both time intervals, responses to CS were significantly slower than those to DS (for 10 and 45 days: z<0.84, p>0.22; Wilcoxon test) and faster than the those to CS in control snails (z>4.9, p<0.0001; Mann-Whitney test).
Thus, not only long-term conditioned memory, but also amnesia occurred after disruption of memory reconsolidation persisted in snails for at least 1.5 months. We obtained similar results earlier [19, 21, 45].
Ten or forty-five days after induction of amnesia caused by pairings of NMDA glutamate receptors antagonist+reminder, no long-term memory was revealed
Two days after training, we administrated two groups of snails with MK-801+reminder. Animals of one group (n=24) after ten days, and the other group after 45 days, underwent repeated aversion training for same CS (raw carrots) used during the first training session (Fig. 2, A). We injected the snails with a DMSO solution 1 hour before each training session. It was found the superiority in CS reactions time compared to those in previous day during each of the three days of repeated training and to the same parameter in DS (Fig. 2, D) (p<0.05, Wilcoxon test). However, on each next day, responses times to CS were the same as those to DS (p>0.2, Wilcoxon test). Thus, memory of short-term type still was untouched in the amnestic snails. At the same time, one day after repeated training we showed (Fig. 2, B and C, T3) that there was no memory of long-term type – it was a reduction in CS reactions time compared to the same in control animals (for 10 and 45 days: z=4.9, p<0.0001; Mann-Whitney test) and no distinction to the same parameter in DS (z=0.91, p>0.16; Wilcoxon test). During repeated training ten days after initial training, the number of CS and US pairings was more significant compared to the first training session: 9.7±0.7 and 6.1±0.4, respectively (z=2.9; p=0.0019, Wilcoxon test); after 45 days, the corresponding values were: 10.1±0.8 and 6.3±0.5 (z=2.8; p=0.008, Wilcoxon test).
Thus, repeated training at different time intervals after the induction of NMDA-dependent amnesia led to formation of short-term memory not long-term memory transforming. We have previously shown that impairment of conditioned food aversion memory by another NMDA glutamate receptor antagonist or protein synthesis inhibitor also led to the development of anterograde amnesia [19, 20, 21]. Amnesia was specific since amnestic snails were able to be trained for aversion to a new type of food for them [19, 21]. According to the indicated features of amnesia, we characterized it as specific anterograde amnesia.
Ten or forty-five days after amnesia induction injections of DNMT inhibitors before repeated training-induced memory formation
Two groups of snails ten days (n=24) or 45 days (n=12) after amnesia induction (MK-801+reminder) were injected with RG108 and repeatedly trained after one h (Fig. 3, A). In both time intervals, we found the progressive growth in CS reactions time during the first day of training (Fig. 3, B and C). By the fourth pairing of CS+US, consummatory responses to CS rose up to 110 – 120 s and were slower than those to DS (for 10 or 45 days: z>3.6; p<0.0001; Wilcoxon test). The CS and US stimuli pairings as much as 2.5±0.12 were required for the formation of long-term memory in repeated training after ten days and 2.3±0.21 in repeated training after 45 days. On days 2 and 3, repeated training was not performed since all animals trained during the first day retained aversive responses to CS: during testing 1 and 10 days after repeated training (Fig. 3, B and C, T3 and T4) there was no distinction in CS reactions time compared to the same parameter in control trained snails (for 10 or 45 days: z <0.11; p> 0.42; Mann-Whitney test) and there was superiority in CS reactions time compared to the same in DS (z>3.6; p<0.0001; Wilcoxon test).
Snails of the other two groups, as after induction of amnesia ten days passed (MK-801+reminder), got zebularine (n=10) or 5-AZA (n=11) injections and were repeatedly trained (Fig. 4, A). As in the experiments described above, both snails groups showed aversion to the CS as the first day of training finished. After 1 and 10 days after this shortened repeated training, we found (Fig. 4, B and C, T3 and T4) the consummatory reactions to CS were the same as those to CS in control trained animals (for zebularine and 5-AZA, after 1 and 10 days: z <1.8; p> 0.14; Mann-Whitney test) and were slower than those to DS (z>2.75; p<0.0019; Wilcoxon test).
One group of snails (n=8), as after induction of amnesia (MK-801+reminder) ten days passed, was injected with RG108, and, one h later, one pairing of CS+US was presented (Fig. 5, A). After 24 h, we found (Fig. 5, B) the reduction in CS reaction time compared to control trained animals (z=3.2, p=0.002; Mann-Whitney test) but also the superiority in CS reaction time compared to the same parameter in DS (z=2.45, p=0.007; Wilcoxon test). However, after ten days, CS reactions were the same as in DS (z=0.59, p=0.57; Wilcoxon test).
Thus, repeated training for aversion of raw carrots under the action of a DNMT inhibitor led to the rapid long-term memory formation within one day of training, and snails retained this memory for at least ten days. For the memory formation during repeated training, four pairings of CS+US are required, a single pairing of CS+US induced memory expression, which was testable after 24 h, but not after ten days.
Interestingly the described DNMT inhibitors effects are not related to the peculiarities of the taste and smell of the food used. In an additional series of experiments using a banana as the СS, we found that injections of the DNMT inhibitor RG108 before repeated training of the amnestic animals caused rapid long-term memory formation, similar in dynamics to those described for the raw carrots CS. In addition, it was found that if it passed ten days as amnesia started because of APV and reminder pairing, injections of RG108 before repeated training also caused a rapid long-term memory formation. Given the fundamental similarity of the results obtained, we do not provide specific data on these experiments.
Injection of DNMT inhibitors before the unpaired CS and US presentation brings no memory recovery (unpaired control)
As after induction of amnesia (MK-801+reminder) ten days passed, snails (n=11) got the RG108 injections and unpaired CS, DS, and US after one h with an interval of 7 – 14 min. We presented each stimulus four times a day (Fig. 5, A). 24 h after administration of RG108 before unpaired control, it was found (Fig. 5, C) the reduction in CS reactions time compared to the same parameter in control trained animals (n=16) (z=5.21, p<0.0001; Mann-Whitney test) and no distinction in CS reactions time compared to the same parameter in DS (z=0.56, p=0.33; Wilcoxon test).
Thus, injections of the DNMT inhibitor prior to the unpaired presentation of CS or US did not result in formation of long-term memory.
Dose-dependent effects of DNMT inhibitors during repeated training
In most experiments, we used RG108 as a DNMT inhibitor with an inhibitor dose of 3 mg/kg. We also conducted experiments to determine the dose dependence of the RG108 effects (Fig. 6, A). As after induction of amnesia (MK-801+reminder) ten days passed, we injected snails (n=8) with RG108 6 mg/kg and did a repeated training. We found (Fig. 6, B) that the consummatory reactions to CS and DS became slower on the day of training. On the next day, as repeated training passed (Fig. 6, B, T3), there was no distinction in CS and DS reactions times compared to each other (z=0.54; p=0.51; Wilcoxon test) or to CS reaction time in control trained snails (z=62; p=0.33; Mann-Whitney test). In the next ten days, consummatory reactions to DS became as slow as those in control animals (Fig. 6, B, T4) (z=1.17; p=0.15; Mann-Whitney test). At the same time, consummatory reactions to CS were still the same as those in control snails (z=0.62; p=0.21; Mann-Whitney test). Thus, the administration of RG108 high dose before retraining led to impairment of the snail's discriminatory ability to distinguish between CS and DS, which persisted for several days. It is interesting to note that similar results were obtained in studies on bees, in which the use of a DNMT inhibitor during olfactory learning caused impairment of odor discrimination on the CS and odor on new food [32].
Another group of snails (n=10), as after induction of amnesia (MK-801+reminder) ten days passed, got 1 mg/kg RG108 injections and were subjected to repeated training one h later (Fig. 6, A). 24 hours and ten days after repeated training, it was found (Fig. 6, C, T3, T4) the reduction in CS reactions time compared to the same parameter in control trained animals (z>4.7, p<0.0001; Mann-Whitney test) and no distinction compared to the same parameter in DS (z<2.5, p>0.7; Wilcoxon test). Thus, using RG108 at a 1 mg/kg dose prior to repeated training did not cause memory formation.
Dependence of the DNMT inhibitors effects on protein and RNA synthesis
Many experiments have shown that protein and RNA synthesis inhibitors can impair long-term memory consolidation [46]. We have studied the conditioned food aversion memory formation dependence on the translation and transcription processes induced by the administration of DNMT + repeated training in amnestic snails. In this series of experiments, six groups of animals were studied (Fig. 8, A). All snails were injected with RG108 10 days after amnesia induction 1 hour before repeated training. The RG108 combined with cycloheximide (n=22) or actinomycin D (n=12) were injected to animals of the two groups. Snails of the other two groups got cycloheximide (n=14) or actinomycin D (n=12) 15 min after repeated training. We injected the snails of the remaining two groups with cycloheximide (n=12) or actinomycin D (n=15) 4 hours after repeated training.
In snails that received RG108+cycloheximide before repeated training or cycloheximide 15 min after administration of RG108 + repeated training, there was a fast growth of CS reactions time followed by aversive responses to CS by the animals as day of training fininshed (Fig. 8, B and C). However, as after induction of amnesia one day passed (Fig. 8, B and C, T3, T4) there was the absence of long-term memory: consummatory reactions to CS were the same as those to DS (z<0.24; p>0.41; Wilcoxon test) and consummatory reactions to CS were slower than those to CS in control snails (z>3.6; p<0.0001; Mann-Whitney test). At the same time, the injections of cycloheximide four h after the RG108 and training pairing could not prevent formation of long-term memory: testing after 1 and 10 days revealed (Fig. 8, D, T3, T4) no distinction in CS reaction time compared to the same parameter in control snails (z<1.8; p>0.14; Mann-Whitney test) and superiority in CS reaction time compared to the same parameter in DS (z>3.0; p<0.0001; Wilcoxon test).
In animals that were administrated with RG108+actinomycin D before repeated training, no impairments in formation of long-term memory were found: consummatory reactions to CS became slower during repeated training and were the same as those to CS in control snails during testing 1 and 10 days after training (Fig. 8, E, T3, T4) (z<1.06; p>0.47; Mann-Whitney test) and consummatory reactions to CS were slower than those to DS (z>3.02; p<0.0001; Wilcoxon test).
In snails that received actinomycin D injections 15 min or four h after RG108 + repeated training administration, memory formation impairment was found (Fig. 8, F and G, T3, T4): testing after 1 and 10 days showed no distinction in CS reaction time compared to the same parameter in DS (for both tests: z<1.6; p>0.09; Wilcoxon test) and inferiority in CS reaction time compared to the same parameter in control animals (z>4.8; p<0.0001; Mann-Whitney test).
Thus, the protein synthesis inhibitor completely suppressed the formation of long-term memory induced by RG108 upon injections of the inhibitor at two-time intervals – before training and 15 min after it. At the same time, the administration of RNA synthesis inhibitor before training was not effective, whereas its injections 15 min or four h after training suppressed long-term memory formation. Interestingly, both inhibitors did not affect the growth in CS reaction times during training. This fact indicates that during repeated training during the action of the DNMT inhibitor, the expression of aversive responses to CS did not depend on the proteins and RNA synthesis. According to these features, we can characterize the facilitation of aversive reactions on the day of training as short-term memory.