Hippocampus

The identification and manipulation of genes (genetics) has influenced the study of most biological phenomena, including hippocampal learning and memory. The first studies to use genetics to manipulate plasticity and learning in mammals targeted the α -calcium calmodulin-dependent kinase II (α CaMKII) (Silva, Paylor, Wehner, and Tonegawa, 1992a; Silva, Stevens, Tonegawa, and Tang, 1992b) and the tyrosine kinase Fyn (Grant et al., 1992). The results showed that genetic manipulations that disrupt hippocampal synaptic plasticity result in hippocampal-dependent learning deficits. These studies were also influential in starting the new field of molecular and cellular cognition, which combines genetics, neurophysiology, and behavioral neuroscience. What have we learned so far about the hippocampus from this new field, and how successful has this general approach been?

The hippocampus plays a key role in learning and memory. Patients with hippocampal lesions are unable to store new conscious memories, with more severe impairments in recent memories than in remote ones. Hippocampal lesions spare other cognitive functions, including intelligence, attention, and emotion. Together with parallel animal experiments, studies of the hippocampus established that memory is separate from other cognitive abilities and that the hippocampus has a prominent role in early stages of memory consolidation (Eichenbaum, 2001).

Nearly twenty years after the publication of the first hippocampal lesion studies, Bliss, Lomo and colleagues discovered long-term potentiation or LTP in the hippocampus, a form of synaptic plasticity that strongly suggests a role in learning and memory. Moreover, computer simulations (parallel neuronet-works) implementing abstractions of LTP-like mechanisms process information in ways that are reminiscent of human and animal learning, suggesting that synaptic phenomena like LTP could be mediating learning and memory. But is LTP a necessary component of learning and memory mediated by the hippocampus?

Pharmacological inhibitors of the N-Methyl-DAspartate receptor (NMDAR) were used in the earliest attempts to answer this question. The NMDAR seems to be a coincidence detector. The receptor regulates a calcium channel that is normally blocked by magnesium; binding of glutamate and postsynaptic depolarization remove the magnesium block of the channel, allowing calcium to enter the synapse and induce LTP. NMDAR antagonists impair hippocampal LTP and cause severe deficits in hippocampal-dependent spatial learning tasks (Morris, Anderson, Lynch, and Baudry, 1986). However, NMDAR blockers have a number of other behavioral effects, such as sensimotor deficits, that complicated the interpretation of those studies. When the first NMDAR studies were published, there were few other pharmacological manipulations that could be used to investigate the possible role of LTP in learning. Fortunately, transgenetic manipulations (transgenics, knock outs, knock ins, and so on) provided the means to modify or delete any one of the molecular components known to be involved in LTP (i.e. CaMKII).

αCaMKII, TP, and Learning

α CaMKII is a member of a family of serine/threonine kinases activated by calmodulin loaded with calcium. The α CaMKII is expressed in postnatal forebrain structures such as the hippocampus and cortex. Studies with pharmacological inhibitors demonstrated that the CaMKII family of kinases was involved in the induction of LTP. These kinases, particularly the α and the β, can potentiate synaptic transmission by a variety of mechanisms, including the phosphorylation of glutamate receptors (Lisman and McIntyre, 2001). Transgenetic studies showed that both a null mutation of α CaMKII (Silva, Stevens, Tonegawa, and Tang, 1992b; Silva, Paylor, Wehner, and Tonegawa, 1992a) and a transgenic overexpression of a constitutively active form of this kinase (α CaMKIIT286D) (Bach et al., 1995; Mayford, Wang, Kandel, and O'Dell, 1995) altered hippocampal LTP and hippocampal-dependent learning. Later studies showed that a point-mutation that substituted threonine at position 286 for alanine (T286A) and prevented the autophosphorylation of α CaMKII at threonine 286 impaired hippocampal LTP and learning (Giese, Fedorov, Filipkowski, and Silva, 1998). α CaMKII, however, affects activity-dependent structural plasticity; hence the learning abnormalities described for these three mutants could be due to deficits in later stages of hippocampal development.

To address this problem, the tetracycline-controlled transactivator system (tTA) was used to regulate the expression of a constitutive active form of α CaMKII that disrupts LTP and learning (α CaMKIIT286D) (Mayford et al., 1996). With this system it was possible to repress the mutant α CaMKIIT286D transgene during development and simply lift the repression at appropriate experimental times. These inducible studies confirmed the importance of α C-aMKII function for hippocampal LTP and learning (Mayford et al., 1996). However, α CaMKII has a wide substrate specificity that is normally restricted by the localization of the enzyme. Thus, it is possible that the higher levels of constitutively active kinase in the inducible transgenic mice may have led to the phosphorylation of proteins that are not normally phosphorylated by this kinase. Consequently, the learning deficits in these tTA- α CaMKIIT286D mutants could be an artifact of the overexpression of this constitutively active kinase.

To circumvent this problem, researchers used a new approach that combines pharmacology and genetics to test the role of α CaMKII in LTP and learning (Ohno et al., 2001). This pharmacogenetic strategy takes advantage of synergisms between pharmacological and genetic manipulations. For example, unlike the homozygous T286A mutation of α CaMKII described above, the heterozygous mutation (T286A hets) does not affect hippocampal-dependent learning. Similarly, although 10 mg/kg of CPP (NMDAR blocker) injected intraperitoneally thirty minutes before training blocks contextual learning (hippocampal-dependent), 5 mg/kg of this drug does not. This same low dose, however, can reveal a contextual learning deficit in the T286A heterozygotes, thus making a compelling connection between kinase activity and contextual learning (Ohno et al., 2001). Although each of the experiments presented above suffers from specific technical limitations, taken together the results presented demonstrate that the activation of α CaMKII is critical for LTP and for hippocampal-dependent learning. This convergence of information is critical for every major finding described here.

Detecting Coincidences with the NMDAR

Learning is heavily dependent on the generation of associations between previously unrelated phenomena. As described above, the molecular properties of the NMDAR suggest that this receptor may have a role in these associations. A number of transgenetic NMDAR manipulations affect both hippocampal synaptic plasticity and learning. A deletion of the NMDAR epsilon subunit and a point mutation of the glycine site in the NMDAR1 subunit (glycine binding potentiates receptor function) disrupt hippocampal LTP and learning. Remarkably, a Cremediated deletion of the NMDAR1 subunit restricted to hippocampal CA1 pyramidal neurons also disrupts CA1 LTP and hippocampal-dependent learning (Tsien et al., 1996), indicating that NMDAR function in this hippocampal subregion is critical for learning. The bacterial phage Cre-recombinase can delete any genomic segment flanked by its recognition sites (LoxP). Thus, any gene flanked by LoxP sites can be deleted from any region or cell type expressing Crerecombinase. This strategy could be used to delete any gene from anywhere in the brain. Specific brain regions or cell types can even be targeted by using virus vectors expressing Cre-recombinase.

The Many Roads of Plasticity and Learning

Aside from NMDAR-dependent activation of α CaMKII, there are many other signaling pathways that are critical for both synaptic plasticity and learning. For example, the ERK signaling pathways are also involved in plasticity and learning (Sweatt, 2001)). Inhibitors of MEK, a kinase that activates ERK, disrupt LTP and learning; both the induction of LTP and training activate ERK. Mutations of genes thought to modulate ERK signaling, including Ras, the neurofibromin GTPase Activating Protein, the guanine-nucleotide-releasing-factor (GRF), and the B-Raf kinase disrupt hippocampal LTP and learning and memory.

The cAMP-signaling pathway is also involved in hippocampal learning and memory. A number of studies indicated that a balance between the activities of cAMP-dependent protein kinase A (PKA) and the phosphatases PP1 and calcineurin gate the stability of both synaptic changes and memory. For example, transgenic mice expressing R(AB), an inhibitory form of the regulatory subunit of PKA, show unstable LTP and memory (Abel et al., 1997). Transgenic expression of a constitutively active form of calcineurin, a Ca/CaM-dependent Ser/Thr phosphatase, also results in unstable LTP and memory. Importantly, repression of this calcineurin transgene, under the control of the tTA system, reverses the long-term memory deficits of the mutants, demonstrating that these effects are not due to the developmental expression of the transgene (Mansuy et al., 1998a).

Inducible overexpression of calcineurin with a modified tetracycline system (rtTa) also resulted in unstable LTP and memory (Mansuy et al., 1998b). The rtTA system was also used to overexpress a COOH-terminal autoinhibitory domain that represses calcineurin function (rtTA-CN inhibitor mice). Inducible overexpression of this inhibitory domain enhances early and late phases of CA1 LTP as well as short-and long-term memory (Malleret et al., 2001). The studies described above indicate that calcineurin and PKA play a critical role in one of the mechanisms that gate the stability of plasticity and memory. But what are the molecular mechanisms responsible for the stability of synaptic changes and memory?

Transcription, Translation, and Memory

A large body of work in a number of organisms and memory systems have demonstrated a universal requirement for transcription and translation in long-term memory (Davis and Squire, 1984)). Moreover, pioneering studies in Aplysia showed the involvement of the transcription factor cAMP Responsive Element Binding protein (CREB) in the stability of synaptic changes. This transcription factor is activated by a number of signaling pathways, including ERK and the cAMP cascades (see above). Interestingly, a null mutation of the CREB α and δ isoforms (CREB^αδ-), which disrupted the stability of hippocampal LTP, also impaired memory (but not learning) tested in a wide range of tasks, suggesting that CREB-dependent transcription was required for long-term memory. Injection of a herpes simplex virus carrying a CREB gene into the amygdala, a structure with a well-characterized role in fear conditioning, enhanced this form of learning. These and many other studies have shown that CREB has a universal role in memory (Silva, Kogan, Frankland, and Kida, 1998).

There is extensive evidence that cAMP-dependent signaling can activate CREB (Silva, Kogan, Frankland, and Kida, 1998). Indeed, deleting both calcium/calmodulin induced adenylate cyclases (1 and 8) blocks the maintenance but not the induction of CA1 LTP and disrupts hippocampal long-term memory. A drug that activates adenyl cyclases turns on CREB and rescues the LTP and memory deficits of this double mutant (Wong et al., 1999). Another protein that can activate CREB is calcium/calmodulin kinase IV (CaMKIV). Experiments with a dominant-negative form of CaMKIV (dnCaMKIV) expressed in the postnatal forebrain showed that this transgene affected late but not early stages of CA1 LTP. Similarly, behavioral experiments revealed specific long-term memory deficits in both spatial learning and contextual fear conditioning. CREB activation was impaired by the dnCaMKIV mutation, indicating that this kinase may activate CREB during learning (Kang et al., 2001).

CREB seems to regulate the expression of Zif268, a transcription factor whose expression is triggered by LTP and learning. Strikingly, studies of a Zif268 null-mutant mouse showed that this transcription factor is needed for the stability of hippocampal plasticity and memory (Jones et al., 2001). These results suggest that Zif268 is downstream of CREB and that this transcriptional cascade is critical for memory.

A key facet of the experiments summarized above is that short-term memory is normal in the mutant mice, demonstrating that the processes required for learning (sensory processing, motivation, and so on) are unaffected in these mice, thus simplifying the interpretation of the results.

Enhancing LTP and Learning

Remarkably, molecular manipulations that enhance LTP often also enhance learning and memory. For example, the mutation of the nociceptin receptor, which mediates the inhibition of adenylyl cyclase, facilitates hippocampal LTP and spatial learning and memory (Manabe et al., 1998). Presumably, during learning the nociceptin receptor mutants generate higher levels of cAMP than controls. This cAMP increase enhances LTP, resulting in faster learning.

A transgenic Tissue Plasminogen Activator (TPA) also enhances LTP and learning, whereas a null mutation of this gene impairs them. TPA is an extracellular protease that seems to be involved in synaptic remodeling triggered by plasticity and learning. Therefore, loss of TPA impairs this remodeling and disrupts hippocampal LTP and learning, whereas overexpressing TPA facilitates both synaptic and behavioral plasticity (Madani et al., 1999). Just like the overexpression of TPA, the transgenic overexpression of the NMDAR subunit 2B, which appears to lengthen the opening time of the NMDAR, also enhances LTP and learning (Tang et al., 1999).

The overexpression of the presynaptic Growth Associated Protein 43, the mutation of a telencephalon-specific cell adhesion molecule, the transgenic inhibition of Calcineurin, and the mutation of Rin 1 (a ras effector) all facilitate LTP and learning. Nevertheless, not all manipulations that enhance LTP improve learning or memory. For example, LTP enhancements that disrupt basic associativity mechanisms do not quicken learning.

Disconnections Between LTP and Learning

There are cases where hippocampal LTP seems to be disrupted with no apparent effect on learning and memory. For example, the deletion of the Glutamate Receptor 1 (GluR1) leads to deficits in CA1 LTP but not in spatial learning. Similarly, the deletion of the Thy-1 gene also spared spatial learning but disrupted LTP in the dentate gyrus. The interconnections between biological phenomena such as LTP and learning are almost always more complex than expected. For example, in vivo recordings revealed LTP in the Thy-1 mutants, and tests with more physiological LTP-inducing protocols demonstrated that the GluR1 mutants can express robust levels of LTP. Nevertheless, it is possible that glutamate receptors are not essential for the expression of all synaptic-specific changes underlying learning and memory. Perhaps other classes of channels (i.e., potassium channels) can also mediate these changes and therefore support learning in mutants with deficits in GluR LTP.

There have also been studies of the impact of some of the mutations described above in hippocampal circuit function. For example, hippocampal circuits represent spatial information. Recordings in behaving rats and mice have demonstrated that hippocampal pyramidal cells fire in a place-specific manner (place fields). Remarkably, a number of manipulations that disrupted hippocampal LTP did not block these spatial representations but impaired their stability. Thus, stable synaptic changes may be crucial for the stability of circuit representations of information in the brain. This possibility also implies that mechanisms other than synaptic plasticity are responsible for generating these representations. Undoubtedly, transgenetic approaches will also have a role in revealing the nature and function of these mechanisms.

The molecular and cellular cognitive studies summarized above provide compelling evidence that the mechanisms responsible for the induction and stability of synaptic changes have a critical role in the acquisition and storage of information in the hippocampus. Because many of the same molecular mechanisms are present throughout the brain, they might have a universal role in learning and memory in other structures. Nevertheless, there is also evidence for important differences between the molecular mechanisms underlying learning in different brain structures (e.g., the amygdala and hippocampus). The findings and ideas immerging from these studies could be the foundation stone for understanding the molecular and cellular processes that underlie our thoughts, fears, desires, and dreams.

See also:GLUTAMATE RECEPTORS AND THEIR CHARACTERIZATION; LONG-TERM POTENTIATION; PROTEIN SYNTHESIS IN LONG-TERM MEMORY IN VERTEBRATES; SECOND MESSENGER SYSTEMS

Bibliography

Abel, T., Nguyen, P. V., Barad, M., Deuel, T. A., Kandel, E. R., and Bourtchouladze, R. (1997). Genetic demonstration of a role for PKA in the late phase of LTP and in hippocampus-based long-term memory. Cell 88, 615-626.

Bach, M. E., Hawkins, R. D., Osman, M., Kandel, E. R., and Mayford, M. (1995). Impairment of spatial but not contextual memory in CaMKII mutant mice with a selective loss of hippocampal LTP in the range of the theta frequency. Cell 81, 905-915.

Davis, H. P., and Squire, L. R. (1984). Protein synthesis and memory. Psychology Bulletin 96, 518-559.

Eichenbaum, H. (2001). The hippocampus and declarative memory: Cognitive mechanisms and neural codes. Behavioral Brain Research 127, 199-207.

Giese, K. P., Fedorov, N. B., Filipkowski, R. K., and Silva, A. J. (1998). Autophosphorylation at Thr286 of the alpha calcium-calmodulin kinase II in LTP and learning. Science 279, 870-873.

Grant, S. G., O'Dell, T. J., Karl, K. A., Stein, P. L., Soriano, P., and Kandel, E. R. (1992). Impaired long-term potentiation, spatial learning, and hippocampal development in fyn mutant mice. Science 258, 1,903-1,910.

Jones, M. W., Errington, M. L., French, P. J., Fine, A., Bliss, T. V., Garel, S., Charnay, P., Bozon, B., Laroche, S., and Davis, S. (2001). A requirement for the immediate early gene Zif268 in the expression of late LTP and long-term memories. Nature Neuroscience 4, 289-296.

Kang, H., Sun, L. D., Atkins, C. M., Soderling, T. R., Wilson, M. A., and Tonegawa, S. (2001). An important role of neural activity-dependent CaMKIV signaling in the consolidation of long-term memory. Cell 106, 771-783.

Lisman, J. E., and McIntyre, C. C. (2001). Synaptic plasticity: A molecular memory switch. Current Biology 11, R788-791.

Madani, R., Hulo, S., Toni, N., Madani, H., Steimer, T., Muller, D., and Vassalli, J. (1999). Enhanced hippocampal long-term potentiation and learning by increased neuronal expression of tissue-type plasminogen activator in transgenic mice. EMBO Journal 18, 3,007-3,012.

Malleret, G., Haditsch, U., Genoux, D., Jones, M. W., Bliss, T. V., Vanhoose, A. M., Weitlauf, C., Kandel, E. R., Winder, D. G., and Mansuy, I. M. (2001). Inducible and reversible enhancement of learning, memory, and long-term potentiation by genetic inhibition of calcineurin. Cell 104, 675-686.

Manabe, T., Noda, Y., Mamiya, T., Katagiri, H., Houtani, T., Nishi, M., Noda, T., Takahashi, T., Sugimoto, T., Nabeshima, T., and Takeshima, H. (1998). Facilitation of long-term potentiation and memory in mice lacking nociceptin receptors. Nature 394, 577-581.

Mansuy, I. M., Mayford, M., Jacob, B., Kandel, E. R., and Bach, M. E. (1998a). Restricted and regulated overexpression reveals calcineurin as a key component in the transition from short-term to long-term memory. Cell 92, 39-49.

Mansuy, I. M., Winder, D. G., Moallem, T. M., Osman, M., Mayford, M., Hawkins, R. D., and Kandel, E. R. (1998b). Inducible and reversible gene expression with the rtTA system for the study of memory. Neuron 21, 257-265.

Mayford, M., Bach, M. E., Huang, Y. Y., Wang, L., Hawkins, R. D., and Kandel, E. R. (1996). Control of memory formation through regulated expression of a CaMKII transgene. Science 274, 1,678-1,683.

Mayford, M., Wang, J., Kandel, E. R., and O'Dell, T. J. (1995). CaMKII regulates the frequency-response function of hippocampal synapses for the production of both LTD and LTP. Cell 81, 891-904.

Morris, R. G. M., Anderson, E., Lynch, G. S., and Baudry, M. (1986). Selective impairment of learning and blockade of long-term potentiation by an N-methyl-D-asparate receptor antagonist, AP5. Nature 319, 774-776.

Ohno, M., Frankland, P. W., Chen, A. P., Costa, R. M., and Silva, A. J. (2001). Inducible, pharmacogenetic approaches to the study of learning and memory. Nature Neuroscience 4, 1,238-1,243.

Silva, A. J., Kogan, J. H., Frankland, P. W., and Kida, S. (1998). CREB and memory. Annual Review of Neuroscience 21, 127-148.

Silva, A. J., Paylor, R., Wehner, J. M., and Tonegawa, S. (1992a). Impaired spatial learning in alpha-calcium-calmodulin kinase II mutant mice. Science 25, 206-211.

Silva, A. J., Stevens, C. F., Tonegawa, S., and Wang, Y. (1992b). Deficient hippocampal long-term potentiation in alpha-calciumcalmodulin kinase II mutant mice. Science 257, 201-206.

Sweatt, J. D. (2001). The neuronal MAP kinase cascade: A biochemical signal integration system subserving synaptic plasticity and memory. Journal of Neurochemistry 76, 1-10.

Tang, Y. P., Shimizu, E., Dube, G. R., Rampon, C., Kerchner, G. A., Zhuo, M., Liu, G., and Tsien, J. Z. (1999). Genetic enhancement of learning and memory in mice. Nature 401, 63-69.

Tsien, J. Z., Chen, D. F., Gerber, D., Tom, C., Mercer, E. H., Anderson, D. J., Mayford, M., Kandel, E. R., and Tonegawa, S. (1996). Subregion-and cell type-restricted gene knockout in mouse brain. Cell 87, 1,317-1,326.

Wong, S. T., Athos, J., Figueroa, X. A., Pineda, V. V., Schaefer, M. L., Chavkin, C. C., Muglia, L. J., and Storm, D. R. (1999). Cal cium-stimulated adenylyl cyclase activity is critical for hippocampus-dependent long-term memory and late phase LTP. Neuron 23, 787-798.

Alcino J.Silva