The team created mice lacking two inhibitory phosphorylation sites on αCaMKII. Wild-type αCaMKII is displaced from synapses in part by the phosphorylation, but the mutant protein has an enhanced affinity for synapses. This lowers the threshold for establishing long-term potentiation (LTP)—a synapse-strengthening event associated with learning—probably because it now takes less calcium rushing into the synapse to reach the necessary level of αCaMKII activity.
Initial water-maze learning by the mice was normal. But the mutant mice, unlike wild type, did not improve on their early learning. Silva believes that the mice solidify all early information—both correct and incorrect—as permanent memories that are difficult to erase.
“A weak signal to these animals is a strong signal,” says Silva. “Initially that may be good.” But later it may be more difficult to filter out the earlier mistakes. “The problem with bringing in a lot of garbage,” says Silva, “is that it is really hard to fine-tune.”
Consistent with this view, retraining the mutant mice does not work well. The fixated mice continue to respond to the original training, and ignore the new maze target.
Fine detail is also missing in another task. The mutants can learn to associate an upcoming shock with a particular cage. But when presented with two similar but different cages, the mutant mice freeze in fear in both cages, whereas wild-type mice freeze only in the cage that is identical to the cage where the shock was originally administered.
Thus, as a memory booster, this method backfires. In contrast, overexpression of CREB has been shown to enhance learning. But in these experiments the animals were in a uniform environment and had to learn specific, isolated tasks. Silva thinks that CREB boosters would not work in the complexities of a human environment because the brain would overload. “They may work for sporadic use [such as against Alzheimer's disease], but they won't work in the long term,” he says. “More is not more is not more. This is not about getting more plasticity; it's about getting circuits that store more information.” ▪