MAPK AS A SIGNAL INTEGRATOR CONTROLLING Kv4.2
While regulation of cell proliferation is the best-studied function of the ERK mitogen Activated Protein Kinase (MAPK) cascade, it is now known that hippocampal ERK activation is necessary for LTP and a wide variety of forms of hippocampus-dependent memory formation (see reference 92). It is interesting to consider that this cascade so critical for normal development is utilized for memory formation in the adult, suggesting a generalized mechanistic conservation between development and adult learning (see reference 52).
Regulation of the ERK cascade is complex, but this complexity allows for some interesting possibilities in terms of hippocampal information processing and neuronal coincidence detection. The ERK cascade, like MAPK cascades in general, is distinguished by a characteristic core cascade of three kinases (see figure). The first kinase in the sequence is Raf-1 (or B-Raf), which activates the second kinase,
MEK, by serine/threonine phosphorylation. MEKs are "dual-specificity" kinases, which means they phosphorylate both a threonine and tyrosine side chain in their substrates. Via this dual phosphorylation, they activate a downstream MAP kinase (p44 MAPK = ERK1, p42 MAPK = ERK2). One important feature of the cascade is that ERK (both ERK1 and ERK2) activity is exclusively regulated by MEK. Dual phosphorylation by MEK both necessary and sufficient for ERK activation. This allows the use of MEK inhibitors to selectively block activation of the ERKs.
Several second-messenger-regulated kinases have been shown to activate the ERK/MAPK cascade in the hippocampus. Stimulation of protein kinase C produces a robust activation of ERK2 in acute hip-pocampal slices, and activation of the cAMP cascade also leads to secondary activation of MAPK in hippocampal area CA1 (see figure). In addition, activation of P-adrenergic receptors (PARs) using
MAPK AS A SIGNAL INTEGRATOR CONTROLLING Kv4.2
BOX 1 MAPK Activation may underlie the effects of Isoproterenol (ISO) plus carbachol on 5-Hz stimulation-induced LTP. The digram to the left illustrates the basic components of the ERK MAP kinase cascade in the hippocampus. (A) The MEK inhibitor U0126 inhibits the induction of LTP by 5 seconds of 5-Hz stimulation delivered during the coapplication of ISO and carbachol. A 5-second train of 5-Hz stimulation delivered at the end of a 10-minute bath application of ISO plus carbachol (200 nM each; the presence of agonists in the bath indicated by the bar) induced robust LTP in vehicle (0.2% DMSO) control experiments (open symbols; fEPSPs were potentiated to 163.9 ± 6.8% of baseline; n = 5) but had little effect on synaptic strength in slices continuously bathed in 20 pM U0126 (filled symbols; fEPSPs were 112.8 ± 7.3% of baseline; n = 5). The traces show superimposed fEPSPs recorded during baseline and 45 minutes post-5-Hz stimulation in a control experiment (left) and in a slice bathed in U0126 (right). Calibration: 1 mV, 5 msec. (B) Synergistic activation of MAPK by coactivation of P-adrenergic and cholinergic receptor agonists. (B1) Representative Western immunoblots showing protein bands visualized with antibodies to dually phosphorylated p42/44 MAPK (PP) and total p42/44 MAPK (Total) in control, untreated slices (Con), and slices bathed for 10 minutes in a CSF containing 200 nM ISO, 200 nM carbachol (CCh), or ISO plus carbachol (ISO 1 CCh). (B2) Average results ± SEM from seven separate experiments like that shown in B1. Only coapplication of ISO plus carbachol induced a statistically significant (*p < .05) increase in phospho-p42 MAPK levels. Reproduced from Watabe, Zaki, and O'Dell (2).
isoproterenol application leads to MAPK activation in area CA1, an effect attenuated by PKA inhibition. Metabotropic glutamate receptors, muscarinic acetylcholine receptors, DA receptors, alpha7 nicotinic acetyl-choline receptors, and serotonin receptors all also lead to ERK activation in the hippocampus. Moreover, regulation of ERK activation in the hippocampus is not limited to neurotransmitter receptors. One of the most widely studied activators of hippocam-pal ERKs is BDNF; BDNF receptors couple to
MAPK AS A SIGNAL INTEGRATOR CONTROLLING Kv4.2
ERK activation in hippocampal neurons, and the ERK activation contributes to BDNF-induced synaptic plasticity in area CA1. Other intriguing possible regulators of ERK in the hippocampus include a novel GTPase activating protein, SynGAP, that potentially links Ca2+/calmodulin activation to ERK stimulation and reactive oxygen species, including superoxide, that can lead to ERK activation in the hippocampus.
This wide variety of upstream regulators of ERK suggests that this signal transduc-tion cascade may serve to integrate diverse cell-surface signals into a coherent intracel-lular response. Especially intriguing is the possibility that this signal integration may not simply serve to sum up signals but rather, in some cases, serve to allow synergistic effects or coincidence detection. Recent important results from Tom O'Dell's laboratory suggest that this type of information processing is indeed occurring. For example, Watabe, Zaki, and O'Dell (see figure) found synergistic activation of hippocampal ERKs by convergent sub-threshold activation of P-adrenergic receptors and mus-carinic acetylcholine receptors. These data suggest that the ERK cascade can serve as a coincidence detector in its own right.
In considering a role for the ras/raf/ MEK/ERK pathway in learning and memory, it certainly warrants emphasis that this pathway has been implicated in learning and memory in a wide variety of species. Published and unpublished studies in Aplysia, Lymnaea, Hermissenda, C. elegans, crayfish, Zebra Finch, Drosophila, mice, and rats have all directly or indirectly implicated a role for this cascade in learning and memory. Studies that we will discuss in Chapter 10 also suggest that it is appropriate to add the human to this list.
H stands for hyperpolarization—these are cation (sodium + potassium) channels that are partially active at the resting membrane potential and that open more with membrane hyperpolarization. Thus, they are voltage-dependent, hyperpolarization-activated channels. They also are directly gated by cyclic nucleotides (cAMP and cyclic Guanosine Mono Phosphate cGMP), which open the channels, but there also is good evidence that some of the enhancing effects of cAMP on Ih are mediated by PKA-dependent phosphorylation. The net effect of H channel function in dendrites is to dampen membrane excitability as other potassium channels do, at least in CA1 pyramidal neurons. However, they dampen membrane excitability without significantly attenuating the peak depolarization of back-propagating action potentials (hyperpolarization activated therefore shut down when the membrane is strongly depolarized, get it?).
What does augmentation of H channels do, then? As a first approximation, you can think of enhancing H channel function as sharpening a back-propagating action potential, narrowing the window of membrane depolarization (36). They are open and counteracting membrane depolarization when the membrane is modestly depolarized, but inactive when the membrane is at the peak of the action potential. This effect might play an important role in timing-dependent plasticity mechanisms, restricting the time-frame over which associative actions like NMDA receptor activation might occur. A second, more tonic effect of augmenting H channels is to limit the ability of modest depolarization to penetrate a dendritic region. Overall, one can think of H channels as a cyclic nucleotide/PKA-dependent filtering mechanism for limiting membrane depolariza-tion—in other words, a mechanism for enhancing the signal-to-noise ratio for depolarization-dependent associative events. cAMP/PKA, acting through H channels, might serve to ensure that associative events occur only when sufficiently robust electrical signals are seen.
B. Voltage-Dependent Sodium Channels (and Calcium Channels?)
Just like in axons, propagation of action potentials along dendrites depends on voltage-gated sodium channels. In situations where back-propagating action potentials provide the depolarization necessary for NMDA receptor activation, this effect is of course dependent on the function of these channels. Work from Costa Colbert and his colleagues has demonstrated that this is a potential site of plasticity for the regulation of LTP induction (37). Specifically, Costa has shown that PKC can regulate the rate of inactivation of sodium channels in pyramidal neuron dendrites, a mechanism allowing PKC control of the extent of action potential back-propagation. Specifically, PKC decreases the extent of sodium channel inactivation, allowing for repetitive action potentials (where Na channel inacti-vation is relevant) to penetrate the dendrites more effectively (30, 38). By this mechanism, PKC activation could promote the induction of LTP.
Back-propagating action potentials also activate voltage-dependent calcium channels in the dendritic membrane (30, 39). This is a critical source of dendritic calcium influx, but under certain conditions voltage-gated calcium channels (VGCCs) can contribute to the action-potential associated membrane depolarization as well. In various experimental circumstances, VGCCs can even propagate a dendritic "calcium spike," a regenerative action potential mediated by calcium flux across the membrane. Thus, modulation of calcium channels, aside from being a mechanism for regulating calcium influx, also can theoretically contribute to controlling local membrane depolarization.
In the next chapter we will cover in detail a variety of mechanisms by which protein kinases augment AMPA receptor function. This is one of the principal mechanisms by which synaptic strength is enhanced during the expression of E-LTP. However, it should not escape our attention that these mechanisms could play a critical role in the induction of LTP as well. In this context it is important to note that CaMKII, PKC, and PKA all enhance AMPA receptor function—any cell surface receptor or calcium influx process that leads to activation of these kinases could influence the likelihood of LTP induction by augmenting AMPA receptor membrane depolarization.
This amplification of AMPA receptor function could be important in several contexts. First, AMPA receptors provide the initial depolarization of the membrane that brings the cell to threshold for firing an action potential. An augmentation of AMPA receptor function means that, for any given level of glutamate at the synapse, there is greater depolarization, at least until the concentration of glutamate reaches a saturating level. Thus, the cell is more likely to reach threshold for firing an action potential, and thus be more likely to trigger NMDA receptor activation. In "non-silent" synapses where AMPA receptors and NMDA receptors are both present, AMPA receptor augmentation will directly lead to greater local membrane depolarization and enhanced NMDA receptor function. This is an appealing model in the context of second-messenger-coupled neurotransmitter receptors modulating LTP induction.
Second, enhanced AMPA receptor function can serve as a temporal integration mechanism. Protein kinase activation and substrate phosphorylation typically are much longer lasting (relatively) than glutamate elevation at the synapse. Depolarization-associated calcium influx or second messenger elevation can set in motion a positive feedback mechanism over time, such that subsequent stimulation of AMPA receptors is enhanced relative to an initial response. This could occur via calcium-activated kinases phosphorylating AMPA receptors, or even through AMPA receptor insertion into the membrane. Although this idea is quite speculative at this time, these mechanisms nicely fit the classical definition of a temporal integration system.
As described in the last chapter, GABA-gated chloride ion channels (GABA-A receptors) control numerous processes relevant to the triggering of LTP. Because these processes were described in the last chapter in the context of hippocampal circuit information processing mechanisms, I won't cover them further here. Moreover, at present the role of GABA-A receptors has largely been studied in the context of controlling the likelihood of action potential firing and membrane depolarization in the cell-body region, as opposed to local control of the membrane electrical properties in the distal dendritic regions. While it is clear that dendritic GABA-A receptor activation could play a powerful role in controlling
Was this article helpful?
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.