Choose a form of plasticity that occurs at a particular synapse. Describe its properties. Give arguments for and against this synaptic plasticity representing the cellular substrate of particular types of learning. How would you design an experiment

Choose a form of plasticity that occurs at a particular synapse. Describe its properties. Give arguments for and against this synaptic plasticity representing the cellular substrate of particular types of learning. How would you design an experiment to demonstrate the causal link between synaptic plasticity and learning and memory?

 

 

Synaptic plasticity describes the activity-dependent changes in synaptic efficacy that can be observed in multiple brain regions. Long-term potentiation (LTP) was first described by Bliss and Lomo (1973), and enables synaptic transmission to be increased. The vast majority of experimental work on LTP has been performed at excitatory synapses between the Schaffer-collateral and commissural axons and apical dendrites of pyramidal cells in the CA1 region of hippocampus (Hc), though the LTP observed at CA1 synapses appears to be identical (or at least very similar) to LTP observed at glutamatergic synapses throughout the mammalian brain. Indeed, the observation that LTP can be most reliably generated in brain regions involved in learning and memory (Neves et al 2008), in conjunction with the fact that LTP can be rapidly induced, long-lasting and display properties of input specificity and associativity, has made it a prime candidate as a cellular correlate of learning and memory. The synaptic plasticity and memory (SPM) hypothesis, defined by Martin et al (2000) states that ‘activity-dependent synaptic plasticity is induced at appropriate synapses during memory formation, and is both necessary and sufficient for the information storage underlying the type of memory mediated by the brain area in which that plasticity is observed’. In this essay, I will discuss the properties of the LTP observed in Hc CA1 (focussing on its postsynaptic expression mechanisms), the evidence for and against its involvement in Hc-dependent types of memory, and what further work must be done to provide evidence for the causal link between LTP and memory.

 

LPT can be induced by either high-frequency stimulation (tetanic stimulation) of the presynaptic neuron (resulting in strong temporal summation of EPSPs in the postsynaptic spine), low-frequency stimulation of the axon held at a strongly depolarised membrane potential (typically -10mV, called the pairing protocol), or precisely timed stimulation of the presynaptic neuron followed by the postsynaptic neuron (within 10ms). Any of these three protocols results in strong postsynaptic depolarisation, maximal activation of NMDA receptors (which are blocked by Mg²⁺ at resting membrane potential, acting as coincidence detectors (Nowak et al (1984)), and thus maximal NMDAR-dependent influx of Ca²⁺. A number of studies have shown that LTP is reliant on this NMDAR-dependent postsynaptic Ca²⁺ influx. In studies on rat CA1 slices, Collingridge et al (1983) showed that application of AP5 inhibits the induction of LTP without altering the performance of the synapse, whilst Malenka et al (1988) used Nitr-5 (a photolabile calcium chelator) to show that postsynaptic Ca²⁺ is not only necessary but also sufficient to induce LTP. Microflourometric measurements in individual CA1 pyramidal cells during LTP induction showed that high-frequency stimulus trains produce transient components of postsynaptic Ca²⁺ accumulation that is blocked by AP5, indicating that LTP-induction protocols induce an NMDAR-mediated increase in intracellular Ca²⁺ (Regehr & Tank, 1990). This Ca²⁺ then interacts with a number of enzymes to bring about the induction and expression of LTP.

 

In 1989, whilst trying to determine how both LTP and LTD (long-term depression) could rely on a Ca²⁺ signal, Lisman proposed that LTP is a consequence of a shift towards protein kinase activity that occurs at higher [Ca²⁺], and there is much evidence in support of this. Malenka et al (1989) found that intracellular injection of H-7 (a general protein kinase inhibitor) into CA1 pyramidal cells blocks LTP induction, indicating the necessity of kinase activity (and thus phosphorylation) for LTP. Further experiments by Malinow et al (1989) showed that postsynaptic CaMKII and PKC are required for LTP induction, and Lledo et al (1995) showed that the injection of a truncated, constitutively active form of CaMKII into CA1 pyramidal cells is sufficient for synaptic strength augmentation. Evidence shows that CaMKII is able to directly phorphorylate AMPAR GluR1 at Ser⁸³¹ in situ (Barria et al, 1997), thus increasing the AMPAR single-channel conductance (Benke et al, 1998; Derkach et al, 1999). In addition, it is thought that CaMKII may act to bring about the insertion of AMPARs at the synapse. Other protein kinases have also been implicated in LTP induction. Inhibition of PKC or PKA blocks LTP induction (Malinow et al, 1989; Frey et al, 1993). High [Ca²⁺] activates adenylyl cyclase, increasing intracellular cAMP levels, and resulting in the activation of PKA, which goes on to inhibit PP1, thus inhibiting the dephosphorylation pathway implicated in LTD.

 

However, whilst CaMKII and PKC appear necessary for the induction of LTP, there remains debate about how LTP can be maintained beyond the Ca²⁺ signal. Though CaMKII is able to be autophosphorylated on Thr²⁸⁶ upon activation by Ca²⁺ (thus rendering its activity no longer dependent on Ca²⁺ and enabling its activity to continue beyond the transient Ca²⁺ signal), postsynaptic H-7 application after the induction of LTP has no effect on its maintenance, indicating that CaMKII and PKC are not necessary for LTP’s maintenance. Instead, it has been suggested that PKM, which is constitutively active, is both necessary and sufficient for LTP maintenance. Ling et al (2002) showed that chelerythrine and ZIP (PKM inhibitor) inhibits the maintenance of established LTP, whilst the diffusion of PKM into cells enhanced the EPSC amplitudes within minutes. However, Volk et al (2013) used KO mice to show that mice lacking PKM exhibited normal synaptic transmission and LTP at schaffer collateral-CA1 synapses. Thus, the debate about the involvement of PKM continues on.

 

As alluded to above, a number of the properties of LTP make it an ideal candidate as the cellular substrate of learning and memory. In addition to being rapidly induced and long-lasting, LTP is also highly input specific, as shown by Andersen et al (1980) in their studies on CA1 regions of guinea pig Hc sliced maintained in vitro. This property of LTP is accounted for by the local source of Ca²⁺ within dendritic spines, and greatly increases the computational possibilities and storage capacities of each neuron. Furthermore, the associative induction of LTP described by Barrionuevo and Brown (1983) has been argued to be a cellular analogue of associative or classical conditioning. However, whilst the LTP generated in CA1 by the induction protocols outlined above has the appropriate properties for a memory mechanisms, this LTP only reflects a ‘memory’ of the brain having been (artificially) electrically stimulated and does not provide evidence that learning itself induces LTP in vivo. Learning induced LTP is difficult to demonstrate for 2 reasons. Firstly, many Hc-dependent learning tasks are iterative and thus require multiple training trials for the memory to be formed. Differences in learning rates across animals would obscure time-sensitive markers of LTP. Secondly, the synaptic changes may be sparse and widely distributed, meaning that potentiated synapses would be difficult to locate within a vast sea of synapses. Whitlock et al (2006) were able to overcome this by training adult rats using the inhibitory avoidance (IA) paradigm, which has been shown to create a stable memory trace in a single trial, and to cause substantial changes in gene expression in CA1. They showed that IA results in an immediate NMDAR-dependent increase in phosphorylation of GluR1 at Ser⁸³¹, the delivery of GluR1 and GluR2 (but not NR1) to the synaptoneurosome (SNS) biochemical fraction, and an increase in the slope of the evoked fEPSP, thus mimicking the effects of HFS. In addition, they showed that these IA-induced increases in evoked fEPSP partially occluded subsequent LTP by FHS, providing evidence that LTP and learning induce change by a common mechanism

 

There is much evidence supporting the importance of the hippocampus in certain kinds of memory. The famous case of patient HM shows that damage to/lesioning of Hc results in profound amnesia, whilst electrophysiolocial recordings and molecular imaging in animals, and MRI in humans provides correlative evidence that certain types of learning involve Hc activity. However, the exact role of Hc in memory remains contested. The theory that is perhaps most influential, put forward by O’Keefe and Nadel in 1978, posits that the primary role of Hc is to encode spatial information and form spatial memories. This is based on the discovery of place cells in CA1 (O’Keefe & Dostrovsky, 1971), and the observation that lesions in Hc impair spatial memory, particularly the acquisition of associative spatial reference memories such as during the Morris Watermaze task. Thus, the hypothesis that LTP-like mechanisms in the CA1-subregion of Hc underlie Hc-dependent forms of associative spatial learning has emerged. Based on this, manipulation that prevent Hc LTP should prevent Hc-dependent forms of learning, and this has been studied using both pharmacological and genetic tools, as is discussed below.

 

In 2006, Pastalkova et al injected a cell-permeable PKM inhibitor into the rat Hc. Though found that this both reserved LTP maintenance in vivo and also produced persistent loss of 1-day-old spatial information. Furthermore, they found that the inhibitor could block LTP and impair Hc-dependent memory even if administered days after acquisition. Thus, they argued that the mechanism maintaining LTP also sustains spatial memory. Morris et al (1986) showed the importance of NMDAR-dependent LTP in spatial memory by applying AP5 to rats by intracerebroventricular infusion using osmotic minipumps. They trained control and AP5-infused rats on the morris watermaze task, and found that the AP5 rats stabilized escape latencies at a higher level (were impaired) and were also impaired in the transfer test. Furthermore, as the AP5 mice performed as well as controls in a visual discrimination task, they argued that chronic AP5 infusion led to a spatial learning impairment that was not caused by secondary sensorimotor or motivational impairment. However, other studies have reported that chronic AP5 application results in sensorimotor deficits, and even in the above studies, the AP5 rats were more prone to falling off the platform. Also, in the above study, AP5 application is not Hc-specific as it diffuses all around the brain, and thus it is not possible to conclude that the effect seen is due to blocking Hc NMDAR. Furthermore, Bannerman et al (1995) found that NMDAR-dependent LTP was not required for spatial navigation or to form associations between a spatial location and a platform. They pre-trained rats in a watermaze task before implanting the minipumps and training them in a second task. This pretraining ameliorated the effects of the AP5, with the rats showing no impairment in a transfer test. They showed that the Hc was still necessary for spatial learning even after pretraining, as rats lesioned after pretraining were significantly impaired. They argued that though NMDAR-dependent LTP was required for some component of spatial learning, it may not be required for encoding spatial representations of a specific environment. Thus, we see that no clear conclusions can be drawn from these pharmacological studies.

 

Other labs have used a genetic approach to study the involvement of hippocampal LTP in learning and memory. Silva et al (1992) showed that -CaMKII knockout (KO) mice exhibit mostly normal behaviours and intact postsynaptic mechanisms (including NMDAR function). However, these mice exhibit specific learning impairments, indicating the importance of -CaMKII in spatial learning but not in non-spatial learning. Tsien et al (1996) used the Cre/LoxP method to create a mouse strain with CA1 pyramidal cell-specific GluN1 KO. LTP could not be induced at these synapses, and the mice showed significant impairment in spatial tasks, though not in non-spatial tasks. However, it was later found by other labs that the KO was not truly Hc-specific, and indeed, other labs obtained different results when using confirmed Hc-specific GluN1 KO mice. Bannerman et al (2012) showed that whilst LTP could not be induced in CA1 of these mice, they showed no impairment in the watermaze task of transfer task. However, in a spatial discrimination watermaze task (using two visually identical beacons), naïve KO mice were significantly impaired, making more choice errors than controls, though performing as well as controls in the transfer task. They observed that the number of choice errors varied systematically a function of where the trial was started from. Thus, it was postulated that the KO did not result in a spatial learning deficit, but rather an inability to use spatial cues to behaviourally inhibit strong conditioned responses. The group argued that hippocampal LTP was not required to encode associative spatial memories, but rather were important for when one must disambiguate between overlapping or competing memories. They did not deny that synaptic plasticity is important for learning and memory, as has been shown by other studies, but their evidence suggest that perhaps the role of the Hc must be reconsidered. Hc could still be important for learning and memory, but may not be the site at which memories are encoded. Thus, we see that even genetic studies are not able to offer clear conclusion on the topic.

 

Furthermore, it is important to note that in the above experiments, it is difficult to conclude that LTP has been completely abolished. Though we see that experimenter-induced LTP has been abolished, the Hc is likely to use different protocols to establish LTP (indeed, some LTP induction protocols are not physiologically plausible). It is possible that behaving animals may still be able to generate sharp-wave ripples: naturally occurring high-frequency waveforms generated by synchronous firing of CA3 pyramidal cells that can facilitate LTP induction. Indeed, Hoffman et al (2002) showed that, whilst LTP induced by a brief burst of 100Hz stimulation was absent in GluN1 KO mice CA1, LTP could still be induced by theta-burst pairing protocol, which mimics the theta waves generated in Hc of rodents when they explore an environment.

 

When designing experiments testing the causal link between memory and synaptic plasticity, it is important to not only consider synaptic plasticity at single neurons, but also how networks of neurons act together to encode neurons. The necessity for synaptic plasticity in memory encoding could be shown in experiments using circuit-specific memory erasure (that is to say, the silencing of neurons containing synapses that were modified during the acquisition of memories), though this will require the advent of more technologies. One approach could be the generation of a transgenic animal model in which promoters from activity-dependent genes are used to drive the expression of transgenes specifically in recently potentiated cells. These transgenes could be used to reversibly inactivate Hc neuronal spiking by driving the expression of membrane proteins that generate appropriate changes in excitability, thus enable us to study the effect on the memory when the neurons associated with it are silenced. Evidence has shown that the expression of several immediate-early genes (IEGs) such as Arc/Arg3.1 are upregulated by LTP-inducing protocols in vitro, in vivo, and in behavioural training. The drosophila melanogaster GPCR for allatostatin (AlstR) has been shown to be able to couple to mammalian G-protein-activated inwardly rectifying K⁺ channels, thus hyperpolarising and silencing the neuron upon allatostatin application. In a transgenic mouse, the activity-dependent Arc/Arg3.1 promoter could be used to drive AlstR expression, meaning that cells with potentiated synapses will express the receptor. As the receptors are internalised and degraded after some time, the application of allatostatin would silence neurons with recently potentiated synapses, meaning that this method could be used to erase recent memories whilst sparing more remote memories. This would enable the study of the causal link between synaptic plasticity and memory storage. In addition, future studies must aim to also show the sufficiency of plasticity for memory storage. Such studies may appear impossible, as it would be unclear which synapses within a sea of synapses in a network should by modified. Perhaps one way of getting around this would be to use a standard training procedure to form a Hc-dependent memory, then erase this memory, and finally attempt to re-install the memory using the knowledge gained about synaptic changes during the original learning. Thus, we see that the full understanding of memory and the neural circuity involved may not be achieved until tools have been developed to study networks at large.

 

In conclusion, we have discussed the properties of the LTP that can be induced in CA1 region of Hc, and seen that behavioural learning is able to induce LTP-like changes that mimic and occlude LTP (indicating that they act by common mechanisms). We have discussed a number of properties of LTP that make it a promising candidate for a cellular substrate of memory, but although increasing numbers of experiments provide evidence for a causal link between LTP and memory, definitive evidence that LTP is necessary for Hc-dependent learning is lacking. Indeed, a number of experiments provide opposing evidence, claiming that NMDAR-dependent LTP in Hc is not required for Hc-dependent forms of associative spatial learning. This being said, it is also important to consider that though LTP may not be necessary for learning (perhaps due to another mechanism), LTP may be the brain’s default choice for memory encoding when available, and may be the most physiologically relevant form of memory encoding. Finally, we have discussed how, in order to fully understand the link between plasticity and memory, neural networks must be studied.

 

 

 

 

 

 

 

 

References:

Andersen, P., Sundberg, S. H., Sveen, O., Swann, J. W., & Wigström, H. (1980). Possible mechanisms for long-lasting potentiation of synaptic transmission in hippocampal slices from guinea-pigs. The Journal of Physiology, 302, 463–82. Retrieved from

Bannerman, D. M., Good, M. A., Butcher, S. P., Ramsay, M., & Morris, R. G. (1995, November 9). Distinct components of spatial learning revealed by prior training and NMDA receptor blockade. Nature. Nature Publishing Group.

Barkus, C., Dawson, L. A., Sharp, T., & Bannerman, D. M. (2012). GluN1 hypomorph mice exhibit wide-ranging behavioral alterations. Genes, Brain and Behavior, 11(3), 342–351.

Barria, A., Muller, D., Derkach, V., Griffith, L. C., & Soderling, T. R. (1997). Regulatory phosphorylation of AMPA-type glutamate receptors by CaM-KII during long-term potentiation. Science (New York, N.Y.), 276(5321), 2042–5.

Barrionuevo, G., & Brown, T. H. (1983). Associative long-term potentiation in hippocampal slices. Proceedings of the National Academy of Sciences, 80(23), 7347–7351.

Benke, T. A., Luthi, A., Isaac, J. T. R., & Collingridge, G. L. (1998). Modulation of AMPA receptor unitary conductance by synaptic activity. Nature, 393(June), 793–797.

Bliss, T. V. P., & Lomo, T. (1973). Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the preforant path. Journal of Physiology, 232(2), 331–356.

Collingridge, G. L., Kehl, S. J., & McLennan, H. (1983). Excitatory amino acids in synaptic transmission in the Schaffer collateral-commissural pathway of the rat hippocampus. The Journal of Physiology, 334, 33–46.

Derkach, V., Barria, A., & Soderling, T. R. (1999). Ca2+/calmodulin-kinase II enhances channel conductance of alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionate type glutamate receptors. Proceedings of the National Academy of Sciences of the United States of America, 96(6), 3269–74.

Frey, U., Huang, Y. Y., & Kandel, E. R. (1993). Effects of cAMP simulate a late stage of LTP in hippocampal CA1 neurons. Science (New York, N.Y.), 260(5114), 1661–4.

Hoffman, D. A., Sprengel, R., & Sakmann, B. (2002). Molecular dissection of hippocampal theta-burst pairing potentiation. Proceedings of the National Academy of Sciences of the United States of America, 99(11), 7740–5.

Ling, D. S. F., Benardo, L. S., Serrano, P. A., Blace, N., Kelly, M. T., Crary, J. F., & Sacktor, T. C. (2002). Protein kinase Mzeta is necessary and sufficient for LTP maintenance. Nature Neuroscience, 5(4), 295–296.

Lisman, J. (1989). A mechanism for the Hebb and the anti-Hebb processes underlying learning and memory. Proceedings of the National Academy of Sciences of the United States of America, 86(23), 9574–8.

Lledo, P. M., Hjelmstad, G. O., Mukherji, S., Soderling, T. R., Malenka, R. C., & Nicoll, R. A. (1995). Calcium/calmodulin-dependent kinase II and long-term potentiation enhance synaptic transmission by the same mechanism. Proceedings of the National Academy of Sciences of the United States of America, 92(24), 11175–9.

Malenka, R. C., Kauer, J. A., Perkel, D. J., Mauk, M. D., Kelly, P. T., Nicoll, R. A., & Waxham, M. N. (1989). An essential role for postsynaptic calmodulin and protein kinase activity in long-term potentiation. Nature, 340(6234), 554–557.

Malenka, R. C., Kauer, J. A., Zucker, R. S., & Nicoll, R. A. (1988). Postsynaptic calcium is sufficient for potentiation of hippocampal synaptic transmission. Science (New York, N.Y.), 242(4875), 81–4.

Malinow, R., Schulman, H., & Tsien, R. W. (1989). Inhibition of postsynaptic PKC or CaMKII blocks induction but not expression of LTP. Science (New York, N.Y.), 245(4920), 862–6.

Martin, S. J., Grimwood, P. D., & Morris, R. G. M. (2000). Synaptic Plasticity and Memory: An Evaluation of the Hypothesis. Annual Review of Neuroscience, 23(1), 649–711.

Morris, R. G. R., Anderson, E., Lynch, G. S. G., Baudry, M., Anderson, A., & Baudry, B. (1986). Selective impairment of learning and blockade of long term potentiation by an N-methyl-D-aspartate receptor antagonist, APV-5. Nature, 319(6056), 774–776.

Neves, G., Cooke, S. F., & Bliss, T. V. P. (2008). Synaptic plasticity, memory and the hippocampus - a neural network aproach to causality. Nature Reviews Neuroscience, 9(January 2008), 65–75.

Nowak, L., Bregestovski, P., Ascher, P., Herbet, A., & Prochiantz, A. (1984). Magnesium gates glutamate-activated channels in mouse central neurones. Nature, 307(5950), 462–465.

O’Keefe, J., & Dostrovsky, J. (1971). The hippocampus as a spatial map. Preliminary evidence from unit activity in the freely-moving rat. Brain Research, 34(1), 171–175.

Pastalkova, E., Serrano, P., Pinkhasova, D., Wallace, E., Fenton, A. A., & Sacktor, T. C. (2006). Storage of Spatial Information by the Maintenance Mechanism of LTP. Science,

Regehr, W. G., & Tank, D. W. (1990). Postsynaptic NMDA receptor-mediated calcium accumulation in hippocampal CA1 pyramidal cell dendrites. Nature, 345(6278), 807–10.

Silva, A. J., Stevens, C. F., & Tonegawa, S. (1992). Deficient hippocampal long-term potentiation in α-calcium- calmodulin kinase II mutant mice. Science, 257(July), 201–

Tsien, J., Huerta, P., & Tonegawa, S. (1996). The essential role of hippocampal CA1 NMDA receptor-dependent synaptic plasticity in spatial memory. Cell, 87(7), 1327–1338.

Volk, L. J., Bachman, J. L., Johnson, R., Yu, Y., & Huganir, R. L. (2013). PKM-ζ is not required for hippocampal synaptic plasticity, learning and memory. Nature, 493(7432), 420–3.

Whitlock, J. R., Heynen, A. J., Shuler, M. G., & Bear, M. F. (2006). Learning Induces Long Term Potentiation in the Hippocampus. Science, 313(5790), 1093–1097.

Filed Under: Synaptic Plasticity Memory