Long-Term Potentiation and the Computational Synapse Lynn E

Proc. Natl. Acad. Sci. USA Vol. 95, pp. 4086–4088, April 1998

Commentary

Long-term potentiation and the computational synapse Lynn E. Dobrunz Howard Hughes Medical Institute and Molecular Neurobiology Laboratory, The Salk Institute, 10010 North Torrey Pines Road, La Jolla, CA 92037

Not all synapses in the brain are the same. Some transmit two closely spaced stimuli, whereas synapses in the lateral signals from one neuron to the next more efficiently than pathway grow stronger with the same stimulus pair (10). others; the efficiency is referred to as synaptic strength. Furthermore, the two pathways are different in terms of the Currently the most widely accepted theory of memory is that stimulation needed to induce LTP (11). memories are stored as patterns of synaptic strength. Long- In the current paper, Min et al. (1) address the important term potentiation (LTP) is one way that the strength of question of whether the differences between the inputs from synapses is changed and perhaps memories are encoded. these two pathways represent two distinct forms of LTP. They Because of this, LTP is one of the most widely studied topics use a trick for estimating the probability that stimulation of neuroscience research. In particular, LTP in the mammalian causes transmitter release, and see if the probability is in- hippocampus has been the focus of intensive research by many creased by LTP. The trick is to add a drug that only blocks scientists, as well as the source of considerable controversy. synapses that release transmitter, and see how quickly the Although a great deal of effort has been devoted to elucidating synapses are blocked. The drug, MK-801, acts at NMDA the mechanism of LTP expression in the CA1 region of receptors and irreversibly blocks synapses where transmitter hippocampus, other exciting questions concerning LTP are has been released (12). The rate of MK-801 blocking therefore addressed by two papers in this issue of the Proceedings. Min is related to transmitter release probability: faster blocking et al. (1) investigate synapses in the dentate gyrus region of the indicates an increase in release probability, whereas a slower hippocampus, and ask the question ‘‘Do synapses from the two rate indicates a decrease (13, 14). input pathways with different properties exhibit the same type A key finding of this paper is that LTP causes an significant of LTP?’’ For synapses in CA1, Peterson et al. (2) raise the increase in the MK-801 blocking rate in synapses from the critical question ‘‘Are changes in synaptic strength graded or medial path, whereas no change was observed in the blocking all-or-none?’’ rate in the lateral path synapses (1). The authors conclude that The hippocampus is a popular site for electrophysiology LTP in the medial perforant path is accompanied by an research because most of its primary cell types are separated increase in neurotransmitter release probability. Their inter- into easily recognizable regions, and the axonal fiber tracts pretation of the results from the lateral path, however, is more connecting these regions are relatively simple and well defined, complex. Although the LTP in the lateral path appears to be as illustrated in Fig. 1. In the dentate gyrus the principal cell different, the authors point out that in the lateral path an type is the granule cell, and the majority of granule cell input increase in release probability after LTP still could be present, comes from a fiber tract called the perforant path. The but difficult to detect using the MK-801 method. To explain perforant path contains axons that originate in the entorhinal this, they invoke the currently popular hypothesis of intersyn- cortex and provide a major sensory input to the hippocampus. aptic cross-talk. The perforant path splits into two anatomically distinct path- To understand this possibility, we need to take a step back ways, called the medial perforant path and the lateral per- and review the meaning of ‘‘intersynaptic cross-talk.’’ Inter- forant path. These fibers form excitatory synapses on the synaptic cross-talk refers to the idea that neurotransmitter middle third and outer third (respectively) of the granule cell released by one synapse may spill over and be detected by a dendritic tree. This anatomical organization enables the study neighboring synapse. This would involve transmitter diffusing of two different populations of synapses onto the same cell out of the synaptic cleft and over to a neighboring cleft in a high type using extracellular stimulation. Axons in the medial and enough concentration to activate postsynaptic receptors. The lateral perforant pathways originate from different regions of idea of intersynaptic cross-talk challenges the traditional as- the entorhinal cortex and carry different kinds of sensory sumption that a synapse is a private connection between an inputs (3). The locations of the medial and lateral perforant axon and a dendrite, and that all synapses act independently of path are shown in Fig. 1; the two types of synapses studied by each other. Excitatory synapses in cortex are on average, Min et al. (1) are represented in red (medial) and yellow however, only separated from a neighboring synapse by 1 ␮m, (lateral). and because glutamate can diffuse on the order of 1 ␮m2͞msec, The two types of synapses are similar in that they both the idea of glutamate spillover is not unreasonable. NMDA release the neurotransmitter glutamate, and, like most cortical receptors have an affinity for glutamate, which is two orders synapses, contain two types of postsynaptic glutamate recep- of magnitude greater than that of AMPA receptors (15). In tors, NMDA and AMPA receptors. At these synapses, like addition, NMDA receptors are tonically activated by baseline others, neurotransmitter is released probabilistically; stimulat- levels of exogenous glutamate (16). As a result, NMDA ing the synapse may cause either transmitter release or no receptors are more likely to be affected by glutamate spillover transmitter release. But differences between the medial and than AMPA receptors. Consequently, synaptic transmission lateral synapses also have been found. Many pharmacological involving AMPA receptors could have little or no cross-talk, manipulations produce markedly different results in the two whereas synaptic transmission mediated by NMDA receptors populations (4–7). Physiological differences occur as well (8, could involve much more cross-talk. Min et al. (1) measured a 9). For example, synapses in the medial perforant path grow greater amount of transmitter sensed by the NMDA receptors weaker (are less likely to release transmitter) on the second of than the AMPA receptors in the lateral perforant path synapses, which supports the idea of glutamate spillover selec- © 1998 by The National Academy of Sciences 0027-8424͞98͞954086-3$2.00͞0 tively activating neighboring NMDA receptors in these syn- PNAS is available online at http:͞͞www.pnas.org. apses. This same result has been shown recently at the CA3 to

4086 Downloaded by guest on September 27, 2021 Commentary: Dobrunz Proc. Natl. Acad. Sci. USA 95 (1998) 4087

indicating that once potentiated, a synapse is not capable of showing further potentiation. In addition, synapses seem to have (at least) two threshold levels for the induction of potentiation: some were able to be potentiated by the weak protocol, whereas others required the stronger stimulus protocol to potentiate. Interestingly, different synapses also showed different amounts of potentiation, although on average the amount of potentiation observed was not different between the weak and strong protocols. This exciting result suggests that at these synapses, information is stored in a discrete or digital manner rather than a graded or analog manner. The authors suggest that in the context of synaptic memories there exist only two stable states, FIG. 1. Synaptic connections in the hippocampus and entorhinal ‘‘potentiated’’ and ‘‘not potentiated.’’ Because these synapses cortex. Lateral perforant path (dotted green) and medial perforant are also capable of long-term depression (LTD), which might path (solid green) provide inputs from the entorhinal cortex to the dentate gyrus of the hippocampus. Perforant path axons form synapses or might not be a reversal of LTP, it appears that at least three onto dentate granule cells (lateral shown in yellow, medial shown in states must be present: potentiated, neither potentiated nor red); these synapses are studied in Min et al. (1). Axons from the CA3 depressed, and depressed. One complication to the idea of region of hippocampus form synapses onto cells in CA1 (shown in information being digitally encoded in synaptic strength is that purple); these synapses are investigated in Peterson et al. (2). several forms of short-term plasticity exist that transiently change synaptic strength at single synapses (22–28). This would CA1 synapse in hippocampus (shown in purple in Fig. 1) (17). lead to the appearance of many more apparent states depend- However, in both of these systems this difference between ing on the history of use of the synapse. Two other exciting transmitter sensed by NMDA and AMPA receptors is greatly reduced or absent when experiments are performed at phys- questions for future study: What determines the threshold for iological temperatures rather than room temperature (1, 18). LTP at a single synapse? And what determines the magnitude In the medial perforant path there was no evidence of gluta- of potentiation achieved? mate spillover onto neighboring synapses (1). Although at It remains to be determined, however, as to whether the some synapses intersynaptic cross-talk seems theoretically all-or-none potentiation observed in Peterson et al. (2) is possible, its existence at these synapses remains controversial actually LTP. One limitation of the work is that responses were (19–21). not measured long enough after potentiation to determine If intersynaptic cross-talk does occur at lateral path syn- whether it was actually a long-term effect. In light of the apses, as the authors suggest, then the trick of using the significance of the findings, additional experiments should MK-801 blocking rate to detect changes in release probability help to clarify whether this all-or-none potentiation is in fact does not work. LTP may, in fact, cause an increase in release LTP. probability at lateral perforant path synapses, in which case the The thought-provoking results presented in both Min et al. medial and lateral path LTP mechanisms are not different. (1) and Peterson et al. (2) provide important links between the Peterson et al. (2) explore a different feature of LTP at cellular physiology of long-term potentiation and the compu- excitatory synapses between CA3 axons and CA1 pyramidal tational function of hippocampal synapses. Intersynaptic cross- cells in the hippocampus (represented in purple in Fig. 1). They talk provides a major challenge to traditional ideas about the ask the novel question of whether potentiation at single computational importance of synapse specificity. And the synapses happens in a graded fashion, or is instead an all-or- existence of all-or-none potentiation as a form of digital signal none process. Although graded levels of LTP can be observed storage may provide a critical clue for deciphering what in populations of synapses, it is not known whether single information is being stored in the strength of synapses. It also synapses are also capable of multiple LTP levels. This impor- raises the question of whether the postsynaptic cell can ‘‘read tant question has implications both for the mechanism under- out’’ the digital level encoded in the synapse and make use of lying LTP and for theories of information storage and perhaps that information. If so, neural networks and computational memory formation in the brain. models of central nervous system function will need to incor- The authors use two protocols for inducing potentiation: porate this fundamental fact about synaptic strength. one weaker, the other much stronger. When these two protocols are used successively to potentiate a population of syn- 1. Min, M.-Y., Asztely, F., Kokaia, M. & Kullmann, D. M. (1998) apses, the response is graded amounts of potentiation. The Proc. Natl. Acad. Sci. USA 95, 4702–4707. weak pairing produces a small, nonsaturating amount of 2. Peterson, C. C. H., Malenka, R. C., Nicoll, R. A. & Hopfield, J. J. potentiation, after which the strong pairing protocol is able to (1998) Proc. Natl. Acad. Sci. USA 95, 4732–4737. produce even more potentiation. What happens at the indi- 3. Witter, M. P. (1993) Hippocampus 3, 33–44. vidual synapses? Do they also potentiate in a graded fashion? 4. Dahl, D., Burgard, E. C. & Sarvey, J. M. (1990) Exp. Brain Res. Or is there all-or-none potentiation of first some of the 83, 172–177. synapses with the weak stimulation and then additional syn- 5. Kahle, J. S. & Cotman, C. W. (1989) Brain Res. 482, 159–163. apses with the strong protocol? 6. Bramham, C. R., Errington, M. L. & Bliss, T. V. (1988) Brain Res. Peterson et al. (2) repeat the experiment using minimal 449, 352–356. stimulation techniques to measure the response of a single 7. Koerner, J. F. & Cotman, C. W. (1981) Brain Res. 216, 192–198. Brain Res. 303, synapse. The weak stimulus protocol was applied first to try to 8. Abraham, W. C. & McNaughton, N. (1984) 251–260. cause potentiation; it then was followed by the stronger 9. McNaughton, B. L. & Barnes, C. A. (1977) J. Comp. Neurol. 175, protocol, to see whether potentiated synapses could exhibit 439–454. further potentiation. Although some synapses failed to show 10. McNaughton, B. L. (1980) Brain Res. 199, 1–19. potentiation in response to either protocol, the majority of 11. Colino, A. & Malenka, R. C. (1993) J. Neurophysiol. 69, 1150– cells exhibited potentiation in an all-or-none fashion. Either 1159. they were potentiated by the first pairing or by the second 12. Huettner, J. E. & Bean, B. P. (1988) Proc. Natl. Acad. Sci. USA pairing, but not by both. The authors interpret these results as 85, 1307–1311. Downloaded by guest on September 27, 2021 4088 Commentary: Dobrunz Proc. Natl. Acad. Sci. USA 95 (1998)

13. Hessler, N. A., Shirke, A. M. & Malinow, R. (1993) Nature 22. Dobrunz, L. E., Huang, E. P. & Stevens, C. F. (1997) Proc. Natl. (London) 366, 569–572. Acad. Sci. USA 94, 14843–14847. 14. Rosenmund, C., Clements, J. D. & Westbrook, G. L. (1993) 23. Dobrunz, L. E. & Stevens, C. F. (1997) Neuron 18, 995–1008. Science 262, 754–757. 24. Murthy, V. N., Sejnowski, T. J. & Stevens, C. F. (1997) Neuron 15. Patneau, D. K. & Mayer, M. L. (1990) J. Neurosci. 10, 2385–2399. 18, 559–612. 16. Sah, P., Hestrin, S. & Nicoll, R. A. (1989) Science 246, 815–818. 25. Turner, D. A., Chen, Y., Isaac, J. T., West, M. & Wheal, H. V. 17. Kullmann, D. M., Erdemli, G. & Asztely, F. (1996) Neuron 17, (1997) J. Physiol. 500, 441–461. 461–474. 26. Debanne, D., Gue´rineau,N. C., Ga¨hwiler, B. H. & Thompson, 18. Asztely, F., Erdemli, G. & Kullmann, D. (1997) Neuron 18, 281–293. S. M. (1996) J. Physiol. 491.1, 163–176. 19. Kullmann, D. & Asztely, F. (1998) Trends Neurosci. 21, 8–14. 27. Chen, Y., Chad, J. E. & Wheal, H. V. (1996) Neurosci. Lett. 218, 20. Barbour, B. & Ha¨usser,M. (1997) Trends Neurosci. 20, 377–384. 204–208. 21. Malenka, R. & Nicoll, R. (1997) Neuron 19, 473–476. 28. Stevens, C. F. & Wang, Y. (1995) Neuron 14, 795–802. Downloaded by guest on September 27, 2021