20 0 20 40

Time

FIGURE 6 APV block of LTP. These data are from recordings in vitro from mouse hippocampal slices, demonstrating the NMDA receptor-dependence of tetanus-induced LTP. Identical high-frequency synaptic stimulation was delivered in control (filled circles) and NMDA receptor antagonist (APV, open triangles) treated slices. Data courtesy of Joel Selcher.

work has shown that an NMDA receptor-independent type of LTP can be induced in area CA1, and elsewhere in the hippocampus (mossy fibers to be precise), as well as other parts of the CNS. We will return to a brief description of these types of LTP at the end of this chapter, but for now we will continue to focus on NMDA receptor-dependent types of LTP.

Early studies of LTP used mostly high-frequency (100-Hz) stimulation, in repeated 1-second-long trains, as the LTP-inducing stimulation protocol. Even though these protocols are still widely used to good effect, it is clear that such prolonged periods of high-frequency firing do not occur physiologically in the behaving animal. However, LTP can also be induced by stimulation protocols that are much more like naturally occurring neuronal firing patterns in the hippocampus. To date, the forms of LTP induced by these types of stimulation have all been found to be NMDA receptor-dependent in area CA1. Two popular variations of these protocols are based on the natural occurrence of an increased rate of hippocampal pyramidal neuron firing while a rat or mouse is

100-Hz 100-Hz 100-Hz 100-Hz

100-Hz 100-Hz 100-Hz 100-Hz

10 msec between pulses

• 5-Hz burst frequency

■ 3 trains, 20-sec intertrain interval

10 msec between pulses

• 5-Hz burst frequency

■ 3 trains, 20-sec intertrain interval

(mini

FIGURE 7 LTP Triggered by theta-burst stimulation in the mouse hippocampus. (A) This schematic depicts theta-burst stimulation. The LTP induction paradigm consists of three trains of 10 high-frequency bursts delivered at 5 Hz. (B), LTP induced with theta-burst stimulation (TBS-LTP) in hippocampal area CA1. The three red arrows represent the three TBS trains. Data courtesy of Joel Selcher.

exploring and learning about a new environment. Under these circumstances hippocampal pyramidal neurons fire bursts of action potentials at about 5 bursts/sec (i.e., 5 Hz). This is the "theta" rhythm that was discussed in the last chapter. One variation of LTP-inducing stimulation that mimics this pattern of firing is referred to as theta-frequency stimulation (TFS), which consists of 30 seconds of single stimuli delivered at 5 Hz. Another variation, theta-burst stimulation (TBS) consists of three trains of stimuli delivered at 20-second intervals, each train is composed of ten stimulus bursts delivered at 5 Hz, with each burst consisting of four pulses at 100 Hz (see Figure 7). We will return to these types of LTP induction protocols in Chapters 5 and 9, where we will discuss modulation of LTP induction and the role of LTP in learning in the animal. For now, it is worth noting that these patterns of stimulation, which are based on naturally occurring firing patterns in vivo, lead to LTP in hip-pocampal slice preparations as well.

A. Pairing LTP

Of course, one can use much more sophisticated electrophysiologic techniques than extracellular recording to monitor synaptic function. Intracellular recording and patch clamp techniques that measure electrophysiologic responses in single neurons have also been used widely in studies of LTP, and as with field recordings, you can use these techniques and observe LTP (see Box 2). Of course, these types of recording techniques perturb the cell that is being recorded from and lead to "rundown" of the postsynaptic response in the cell impaled by the electrode. This limits the duration of the LTP experiment to however long the cell stays alive— somewhere in the range of 30 minutes to an hour for an accomplished physiologist. Regardless, in these recording configurations you can induce synaptic potentiation using tetanic stimulation or theta-pattern stimulation and measure LTP as an increase in post-synaptic currents through glutamate-gated ion channels, or as an increase in postsynaptic depolarization when monitoring the membrane potential.

Control of the postsynaptic neuron's membrane potential with cellular recording techniques also allows for some sophisticated variations of the LTP induction paradigm. In one particularly important series of experiments, it was discovered that LTP can be induced by pairing repeated single presynaptic stimuli with postsynaptic membrane depolarization, so-called "pairing" LTP (6). (See Figure 8.)

The basis for pairing LTP comes from one of the fundamental properties of the NMDA receptor (see Figure 9). The NMDA receptor is both a glutamate-gated channel and a voltage-dependent one. The simultaneous presence of glutamate and a depolarized membrane is necessary and sufficient (when the co-agonist glycine is present) to gate the channel. Pairing synaptic stimulation with membrane depolarization provided via the recording electrode (plus the low levels of glycine always normally present) opens the NMDA receptor channel and leads to the induction of LTP.

How is it that the NMDA receptor triggers LTP? The NMDA receptor is a calcium channel, and its gating leads to elevated intra-cellular calcium in the postsynaptic neuron. We will return to this calcium influx that triggers LTP in the next chapter, and indeed most of the rest of this book deals with the various processes this calcium influx triggers.1

These properties, glutamate dependence and voltage-dependence, of the NMDA receptor allow it to function as a coincidence detector. This is a critical aspect of NMDA receptor regulation and allows for a unique contribution of the NMDA receptor to information processing at the molecular level. Using the NMDA receptor, the neuron can trigger a unique event, calcium influx, specifically when a particular synapse is both active presynaptically (glutamate is present in the synapse) and postsynap-tically (when the membrane is depolarized).

!It is important to remember that it is not necessarily the case that every calcium molecule involved in LTP induction actually comes through the NMDA receptor. Calcium influx through membrane calcium channels and calcium released from intracellular stores may also be involved. In fact, it is an interesting "thought experiment" to try to design a way to test this idea—as a practical matter it is much more difficult to determine than one might first think.) Mechanistically the gating of the NMDA receptor/channel involves a voltage-dependent Mg2+ block of the channel pore. Depolarization of the membrane in which the NMDA receptor resides is necessary to drive the divalent Mg cation out of the pore, which then allows calcium ions to flow through. Thus, the simultaneous occurrence of both glutamate in the synapse and a depolarized postsynaptic membrane is necessary to open the channel and allow LTP-triggering calcium into the postsynaptic cell.

■20 -10 0 10 20 30 Time (min)

FIGURE 8 Pairing LTP. (A) LTP of synaptic transmission induced by pairing postsynaptic depolarization with synaptic activity. The upper panels illustrate postsynaptic currents recorded directly from the postsynaptic neuron using voltage clamp techniques (see Box 2). The lower panels are a pairing LTP experiment (upper), and control, nonpaired pathway (lower). In the pairing LTP experiment hippocampal CA1 pyramidal neurons were depolarized from -70 to 0 mV, while the paired pathway was stimulated at 2 Hz 40 times. Control received no stimulation during depolarization. From Malinow and Tsien (26). Reproduced with permission from Nature Publishing Group. (B) Pairing small EPSPs with back-propagating dendritic action potentials induces LTP. (a) Subthreshold EPSPs paired with back-propagating action potentials increase dendritic action potential amplitude. Voltage clamp recording at approximately 240 pm from soma, that is, in the dendritic tree of the neuron (see Figure 10). Action potentials were evoked by 2-ms current injections through a somatic whole-cell electrode at 20-ms intervals. Alone, action potential amplitude was small (unpaired). Paired with EPSPs (5 stimuli at 100 Hz), the action potential amplitude increased greatly (paired). (b) The grouped data show normalized EPSP amplitude after unpaired and paired stimulation. The pairing protocol shown in A was repeated five times at 5 Hz at 15-second intervals for a total of two times. (c) A similar pairing protocol was given with and without applying the sodium channel blocker tetrodotoxin (TTX, to block action potential propagation) to the proximal apical dendrites to prevent back-propagating action potentials from reaching the synaptic input sites. LTP was induced only when action potentials fully back-propagated into the dendrites. Reproduced with permission from Magee & Johnston (10). Copyright 1997 American Association for the Advancement of Science.

This confers a computational capacity at the molecular level. Using the NMDA receptor the neuron can confer a property of associativity on the synapse. This attribute is nicely illustrated by "pairing" LTP, as described previously, where low-frequency synaptic activity paired with postsynaptic depolarization can lead to LTP. The associative property of the NMDA receptor allows for many other types of sophisticated information processing as well, however. For example, activation of a weak

FIGURE 9 Coincidence detection by the NMDA receptor. The simultaneous presence of glutamate and membrane depolarization is necessary for relieving Mg2+ blockade and allowing calcium influx.

input to a neuron can induce potentiation, provided a strong input to the same neuron is activated at the same time (7). These particular features of LTP induction have stimulated a great deal of interest because they are reminiscent of classical conditioning, with depolarization and synaptic input roughly corresponding to unconditioned and conditioned stimuli, respectively.

The associative nature of NMDA receptor activation also allows for synapse specificity of LTP induction, which has been shown to occur experimentally. If you pair postsynaptic depolarization with activity at one set of synaptic inputs to a cell, while leaving a second input silent or active only during periods at which the postsynaptic membrane is near the resting potential, you get selective potentiation of the paired input pathway.

Similarly, in field stimulation experiments, LTP is restricted to tetanized pathways— even inputs convergent on the same dendritic region of the postsynaptic neuron are not potentiated if they see only baseline synaptic transmission in the absence of synaptic activity sufficient to depolarize the postsynaptic neuron adequately (8). This last point illustrates the basis for LTP "cooperativity." LTP induction in extracellular stimulation experiments requires cooperative interaction of afferent fibers, which in essence means that there is an intensity threshold for triggering LTP induction. Sufficient total synaptic activation by the input fibers must be achieved such that the postsynaptic membrane is adequately depolarized to allow opening of the NMDA receptor (9).

B. Dendritic Action Potentials

In the context of the functioning hip-pocampal neuron in vivo, the associative nature of NMDA receptor activation means that a given neuron must reach a critical level of depolarization in order for LTP to occur at any of its synapses. Specifically, in the physiologic context, the hippocampal pyramidal neuron generally must reach the threshold for firing an action potential, although there are some interesting alternatives to this that we will discuss later in this chapter and in Chapter 5. Even though action potentials are, of course, triggered in the active zone of the cell body, hippocam-pal pyramidal neurons along with many other types of CNS neurons can actively propagate action potentials into the dendritic regions: the so-called back-propagating action potential (10). (See Figure 10.) These dendritic action potentials are just like action potentials propagated down axons in that they are carried predominantly by voltage-dependent ion channels such as sodium channels. The penetration of the back-propagating action potential into the dendritic region provides a wave of membrane depolarization that allows for

FIGURE 10 Back-propagating action potentials in dendrites of CA1 pyramidal neurons. (A) This panel shows the recording set-up, with a bipolar stimulating electrode used to trigger action potentials at the cell body region (lower left), a recording electrode in the cell soma to monitor firing of an action potential, and a recording electrode in the dendrites (upper right) to monitor propagation of the action potential into the distal dendritic region. (B) Traces here indicate the data recorded from the soma (lower) and dendritic (upper) electrodes. The left-hand traces (labeled AP) indicate the membrane depolarization achieved at the soma and dendrite when an action potential is triggered and propagates into the dendritic region. Note that the dendritic action potential is of lower magnitude and broader owing to the effects of dendritic membrane biophysical properties as the action potential propagates down the dendrite. The right-hand side shows current flow through "A-type" voltage-dependent potassium currents observed in the soma and dendrites. The density of A-type potassium currents increases dramatically as one progresses outward from the soma into the dendritic regions, as illustrated by the much larger potassium current observed in the distal dendritic electrode. These voltage-dependent potassium channels are key regulators of the likelihood of back-propagating action potentials reaching various parts of the dendritic tree. Data and figures reproduced with permission from Yuan et al. (27).

FIGURE 10 Back-propagating action potentials in dendrites of CA1 pyramidal neurons. (A) This panel shows the recording set-up, with a bipolar stimulating electrode used to trigger action potentials at the cell body region (lower left), a recording electrode in the cell soma to monitor firing of an action potential, and a recording electrode in the dendrites (upper right) to monitor propagation of the action potential into the distal dendritic region. (B) Traces here indicate the data recorded from the soma (lower) and dendritic (upper) electrodes. The left-hand traces (labeled AP) indicate the membrane depolarization achieved at the soma and dendrite when an action potential is triggered and propagates into the dendritic region. Note that the dendritic action potential is of lower magnitude and broader owing to the effects of dendritic membrane biophysical properties as the action potential propagates down the dendrite. The right-hand side shows current flow through "A-type" voltage-dependent potassium currents observed in the soma and dendrites. The density of A-type potassium currents increases dramatically as one progresses outward from the soma into the dendritic regions, as illustrated by the much larger potassium current observed in the distal dendritic electrode. These voltage-dependent potassium channels are key regulators of the likelihood of back-propagating action potentials reaching various parts of the dendritic tree. Data and figures reproduced with permission from Yuan et al. (27).

the opening of the voltage-dependent NMDA receptor/ion channels. Active propagation of the action potential is necessary because the biophysical properties of the dendritic membrane dampen the passive propagation of membrane depolarization; thus, an active process such as action potential propagation is required. As a generalization in many instances in the intact cell, back-propagating action potentials are what allow sufficient depolarization to reach hippocampal pyramidal neuron synapses in order to open NMDA receptors. In an ironic twist, this has brought us back to a more literal reading of Hebb's Postulate, where as we discussed in the introduction to this Chapter, Hebb actually specified firing of the postsynaptic neuron as being necessary for the strengthening of its connections.

In fact, the timing of the arrival of a dendritic action potential with synaptic glutamate input appears to play an important part in precise, timing-dependent

FIGURE 11 The timing of back-propagating action potentials with synaptic activity determines whether synaptic strength is altered, and in which direction. Precise timing of the arrival of a back-propagating action potential (a "spike") with synaptic glutamate determines the effect of paired depolarization and synaptic activity. A narrow window when the arrival of the synaptic EPSP immediately precedes or follows the arrival of the back-propagating action potential determines whether synaptic strength is increased, decreased, or remains the same. See text for additional discussion. Figure adapted with permission from Bi and Poo (12).

FIGURE 11 The timing of back-propagating action potentials with synaptic activity determines whether synaptic strength is altered, and in which direction. Precise timing of the arrival of a back-propagating action potential (a "spike") with synaptic glutamate determines the effect of paired depolarization and synaptic activity. A narrow window when the arrival of the synaptic EPSP immediately precedes or follows the arrival of the back-propagating action potential determines whether synaptic strength is increased, decreased, or remains the same. See text for additional discussion. Figure adapted with permission from Bi and Poo (12).

triggering of synaptic plasticity in the hippocampus (see reference 10 and Figures 8 and 11). It has been observed that a critical timing window is involved vis-à-vis back-propagating action potentials: glutamate arrival in the synaptic cleft must slightly precede the back-propagating action potential in order for the NMDA receptor to be effectively opened. This timing dependence arises in part due to the time required for glutamate to bind to and open the NMDA receptor. The duration of an action potential is, of course, quite short so in essence the glutamate must be there first and already bound to the receptor in order for full activation to occur. (Additional factors are also involved; see references 11, 12, and 13 for a discussion).

This order-of-paring specificity allows for a precision of information processing— not only must the membrane be depolarized, but also, as a practical matter, the cell must fire an action potential. Moreover, the timing of the back-propagating action potential arriving at a synapse must be appropriate. It is easy to imagine how the nervous system could capitalize on these properties to allow for forming precise timing-dependent associations between two events.

One twist to the order-of-paring specificity is that if the order is reversed and the action potential arrives before the EPSP, then synaptic depression is produced. The mechanisms for this attribute are under investigation at present—one hypothesis is that the backward pairing by various potential mechanisms leads to a lower level of calcium influx, which produces synaptic depression (see chapter 5).

The role of back-propagating action potentials in regulating NMDA receptor activation raises the interesting possibility of local effects on dendritic membrane excitability regulating LTP induction in a dendritic branch-specific (or even branch subregion-specific) manner. Because den-drites are highly branched, nonuniform entities, regulating the ability of a back-propagating action potential to penetrate into a specific dendritic region could control the capacity of LTP induction for all the synapses in that branch. Although highly hypothetical at this point, such a mechanism could confer additional information processing capacity on hippocam-pal pyramidal neurons. Similarly, but even more speculatively, locally constrained action potential generation only within a dendritic subregion would allow for highly localized NMDA receptor activation as well.

Branch-specific regulation of LTP induction could occur through two easily conceived mechanisms. The first is through neuromodulation of local voltage-dependent Na+ and K+ channels, for example in the proximal region of a branch (11). Shutting down the capacity of an action potential to pass a branch point by down-regulating Na+ channels or up-regulating K+ channels is a possibility, based on the known capacity of neurotransmitter receptors and their attendant signal transduction mechanisms to regulate these channels (see Figures 10 and 12). The idea here is that local activation of NE, ACh, or serotonin receptors could increase (or decrease) the likelihood of LTP induction for an entire branch by controlling the likelihood of action potential propagation into that branch.

A second possible mechanism for regulating branch-specific action potential back-propagation is activity-dependent (10). The occurrence of EPSPs boosts action potential back-propagation in a given dendritic region through modulation of local voltage-dependent K+ channels. In brief, membrane depolarization by an EPSP leads to a transient decrease in voltage-dependent K+ channel function, by causing voltage-dependent inactivation of these channels. Thus, synaptic input into a particular branch point can increase the likelihood of action potential propagation into that region owing to increased membrane excitability because of diminution of active K+ channels (see Figures 8 and 12). The duration of this effect is, of course, limited by the time-course of recovery of the K+ channels, which is typically fairly quick. It is important to note that this mechanism and the one described in the previous paragraph are not mutually exclusive, but rather they might act in concert to confer additional information processing sophistication.

In later chapters, we will discuss additional implications of this type of information processing in more detail. Thus, we will discuss in more detail the molecular mechanisms by which these local effects regulating membrane depolarization within specific dendritic branches or dendritic subregions may be achieved. Moreover, we will discuss the signal transduction mechanisms by which modulatory neurotrans-mitter systems can regulate the likelihood of action potential back-propagation by controlling dendritic potassium channels, and we will discuss how this might allow

Was this article helpful?

0 0
Advanced Memory Techniques

Advanced Memory Techniques

A course in techniques and skills for mentalists, magicians and students. For students, improve your grades with less effort! But this book is also.... The ideal for any stage mentalist or magician by establishing credibility of amazing skills with an easy to follow instructional book on using the amazing power of your memory.

Get My Free Ebook


Post a comment