|Home | About | Journals | Submit | Contact Us | Français|
Dendritic ion channels are essential for the regulation of intrinsic excitability as well as modulating the shape and integration of synaptic signals. Changes in dendritic channel function have been associated with many forms of synaptic plasticity. Recent evidence suggests that dendritic ion channel modulation and trafficking may contribute to plasticity-induced alterations in neuronal function. In this review, we discuss our current knowledge of dendritic ion channel modulation and trafficking and their relationship to cellular and synaptic plasticity. We also consider the implications for neuronal function. We argue that to gain an insight into neuronal information processing, it is essential to understand the regulation of dendritic ion channel expression and properties.
Dendrites are extensive and elaborate processes emerging from the cell body of neurons. They occupy a large surface area and receive most synaptic inputs1. Their predominant function is in processing and transmitting synaptic signals to the cell body and axon initial segment, where, if threshold is reached, action potentials are initiated. This is an active process, as it is known that dendrites possess an abundance of ion channels that are involved in receiving, transforming and relaying information to other parts of the neuron1. These dendritic ion channels often differ in their biophysical properties and densities from those present in other neuronal compartments. Moreover, ion channel expression and properties may also differ within the dendritic arbor of neurons – e.g. hyperpolarization-activated cation non-selective (HCN) channels are expressed highly in the apical, but not the basal, dendritic tree of layer V cortical pyramidal neurons2-4. This adds an additional layer of complexity to neuronal information processing.
It is now evident that dendritic ion channel expression and properties are modulated by induction of Hebbian (including long-term potentiation (LTP) and long-term depression (LTD)) as well as homeostatic (non-Hebbian) forms of plasticity (for reviews see Refs. 5-7). Hebbian forms of plasticity are input specific changes in synaptic strength that largely involve postsynaptic Ca2+ entry through voltage-sensitive N-methyl-D-aspartate receptors (NMDAR; known as NMDAR-dependent plasticity). This Ca2+ influx also activates intracellular signaling pathways that modify dendritic ion channel activity, local excitability and, perhaps, cell-wide excitability, or “intrinsic plasticity”8, 9 (Figure 1). Often these activity-dependent changes in dendritic ion channel function are stabilizing and limit the extreme neuronal activity (spiking) that might otherwise result from sustained synaptic efficacy. This “homeostatic plasticity”10 provides negative feedback control of Hebbian synaptic plasticity. Moreover, during synaptic plasticity altered expression and function of dendritic ion channels, through their effects on membrane polarization, may also influence the threshold for further induction of plasticity, or metaplasticity7-9, providing a local mechanism of control over cell excitability.
Some of the activity-dependent changes in dendritic ion channel function described above are likely to be a consequence of altered post-translational modifications as well dendritic channel trafficking (Figure 1). Recent evidence suggests that selective targeting mechanisms determine the distribution and properties of dendritic ion channels11. Specific molecules are involved in the transport of ion channel subunits from the soma to dendrites. Additionally, mRNAs encoding ion channels can be trafficked into dendrites and locally translated, a process which may be activity-dependent12. Indeed, dendrites contain the necessary machinery that enables local protein synthesis13. Hence, expression of ion channels can be dynamically modified in dendrites in response to synaptic activity. This active modulation of ion channel function may present a sophisticated mechanism by which neurons regulate information flow and thereby, neuronal output.
Here we review recent reports on dendritic voltage-gated ion channel targeting mechanisms and plasticity, omitting ligand gated ion channel trafficking and plasticity as there have been many recent reviews on this (for example see14, 15). We begin by presenting an overview on how ion channels affect dendritic intrinsic excitability and synaptic integration.
Dendrites contain a plethora of ion channels including K+ channels. In many central neurons, the densities of most voltage-gated potassium (Kv) channels appear to be uniform or lower in distal dendrites compared with those present at the soma1. One exception appears to be the Kv4 subunit. Immunohistochemical analysis first showed a predominantly dendritic localization of Kv4 channels16 (Table 1). The Kv4 subunits form a fast activating and inactivating current in heterologous systems, reminiscent of the A-type K+ current (IA) in neurons17. Consistent with the immunohistochemical observations, electrophysiological data together with pharmacology and calcium imaging have shown that A-type K+ channels are more efficacious in the apical18-21, radial oblique22, 23 and basal4, 23, 24 dendrites than the soma of several types of central neurons. Here, A-type K+ channels play an important role in determining the amplitude and width of back-propagating action potentials18, 19, 25. They also limit the propagation of local dendritic spikes generated by spatially clustered and temporal synaptic input23 and curtail dendritic Ca2+ signals generated by synaptic input or by back-propagating action potentials22-24. Thus, these channels affect forms of plasticity that depend on back-propagating action potentials or the propagation of local dendritic spikes (i.e. spike timing-dependent plasticity)23, 26. In addition, in hippocampal neurons, altering Kv4.2 channel expression leads to an activity-dependent remodeling of synaptic NMDAR subunit composition and consequentially the ability to induce synaptic plasticity27, suggesting that regulation of these channels may act as a metaplasticity mechanism.
In contrast to Kv4 channels, neuronal Kv2.1 channels conduct delayed rectifier (IK) currents that have a high threshold of voltage activation and slow kinetics28. Kv2.1 channels are found in many mammalian central neurons including hippocampal and cortical pyramidal cells (Table 1), where they appear to be localized to the somatodendritic compartment28 (but see Ref. 29). Typically delayed rectifier currents have the primary role of repolarizing the membrane after action potentials. However, the activation and inactivation properties of Kv2.1 suggest these channels are too slow for the regulation of single action potentials and instead influence repetitive spiking30. In support of this, knockdown of Kv2.1 did not alter the shape of single action potentials but did cause hyperexcitability after repetitive (1 Hz) stimulation in hippocampal pyramidal cells31. Kv2.1 channels may therefore play an important role in dendritic integration, by suppressing hyperexcitability as repetitive signals approach the soma, and potentially contributing to homeostatic plasticity.
Dendrites and spines of several central neurons also contain calcium-activated potassium (KCa) channels32-35. Intriguingly KCa2 (small conductance calcium-activated potassium or SK) channels are located in close proximity to synaptic and extra-synaptic glutamate receptors, suggesting a synaptic function (Table 1). Indeed, these channels reduce dendritic integration by restricting compartmentalized Ca2+ spikes (plateau potentials) triggered by strong synaptic input33. In the hippocampus32 and the amygdala34, Ca2+ influx through NMDA receptors activates KCa2 channels, hyperpolarizing the membrane and promoting the NMDA receptor Mg2+ block, limiting further activation. This KCa2-mediated negative feedback on NMDA receptors therefore impacts the induction of Hebbian plasticity. Consistent with this model, pharmacological downregulation of KCa2 enhances36, while the genetic upregulation of KCa2 impairs37, hippocampal LTP induction and memory encoding.
Inwardly-rectifying K+ (Kir) channels are another group of K+ channels that are expressed throughout the CNS including the apical dendrites of neocortical and hippocampal CA1 neurons38-41 (Table 1). Kir channels are characterized by their unidirectional inward rectification that is gated by an intracellular cation block41. Therefore, at membrane potentials more negative than rest, Kir channels pass an inward current, returning the membrane to resting potential. However, at potentials more positive than rest, cations prevent an outward K+ current from hyperpolarizing the cell membrane. These fundamental rectification properties of Kir channels are essential in maintaining neuronal membrane potential. Of the seven Kir subfamilies, Kir3.x channels are unique in their activation by G-protein coupled receptors (GPCRs). Specifically, Gi- or Go-type GPCRs, such as GABAB receptors, activate Kir3 channels41-43. Which particular GPCRs interact with Kir3.x is potentially mediated by their spatial compartmentalization. For example, γ-amino butyric acid B (GABAB) receptors have been observed in close proximity to synaptic Kir3.x channels in spines, but less so in the dendritic shaft44, 45. Consistent with their synaptic localization, Kir3.2 channels mediate slow inhibitory post synaptic currents (IPSCs)46, which are potentiated following low frequency (3 Hz) stimulation in hippocampal slices. This phenomenon is mediated by GABAB receptor activation of Kir3.2 channels, and is NMDAR- and calcium-calmodulin dependent protein kinase II (CaMKII)-dependent44.
Clearly, K+ channels play a significant role in shaping dendritic excitability. Dendrites, however, contain a number of other ion channels too. Interestingly, recent evidence shows that the dendrites and spines of hippocampal and cortical neurons contain an exceptionally high density of the hyperpolarization-activated cation non-selective (HCN) channels47, 48 (Table 1). Four subtypes of HCN (HCN1-4) genes exist49, with HCN1 and HCN2 channels present predominantly in dendrites47, 48. These channels have very unusual biophysical properties, in that they are permeable to both Na+ and K+ and are activated at potentials hyperpolarized to −50 mV. Hence, they are active at rest and are involved in maintaining the neuronal resting membrane potential (RMP). Their effects on dendritic excitability, though, are complex. Block or knockdown of HCN channels causes RMP hyperpolarization but results in significantly greater numbers of dendritic action potentials, slower excitatory post-synaptic potential (EPSP) decay and enhanced EPSP summation50-55. These effects are due to increased membrane resistance51, 54 as well as alterations in the biophysical properties of other ion channels such as low voltage-activated Ca2+ channels56 and delayed rectifier K+ channels57. Hence, in spite of the RMP being hyperpolarized, a loss of Ih in distal hippocampal dendrites gives rise to enhanced LTP58 and elevated neural network excitability51, 59.
In addition to HCN and K+ channels, immunohistochemical studies have demonstrated the presence of the Na+ channel subunits, NaV1.1, NaV1.2 and NaV1.6 in dendrites and spines of hippocampal CA1 and cortical pyramidal neurons60. In agreement with this, electrophysiological studies have revealed Na+ channels in the dendrites of these neurons60, where they play a role in potentiating action potential back-propagation19, 61 and generation of dendritic spikes62 (Table 1). Action potential back-propagation and initiation of dendritic spikes are critical for the induction of some forms of Hebbian plasticity9, 62.
The initiation and expression of many forms of plasticity also often involves Ca2+ entry through voltage-gated Ca2+ channels (VGCC). To date, ten VGCC primary subunits have been cloned63. Immunohistochemical as well as electrophysiological studies have revealed the presence of all subtypes of VGCC in dendrite shafts60. Further, multiple subtypes of VGCC have been found in dendritic spines in numerous cell types64 (Table 1). VGCC opening is enhanced by action potential backpropagation, synaptic potentials and action potential backpropagation, sometimes leading to the initiation of Ca2+ spikes and plateau potentials60, 65, 66. These properties allow VGCCs to regulate the induction of synaptic plasticity60, 65, 66. Indeed, Ca2+ entry through dendritic VGCC is necessary for LTD in entorhinal cortical cells67 and hippocampal CA1 neurons68, as well as for LTP at hippocampal CA1- perforant path synapses66 and hippocampal CA1-schaffer collatoral synapses69. Moreover, Ca2+ influx via CaV1.x (L-type) Ca2+ channels in dendritic spines contributes to induction of synapse specific NMDA receptor- dependent LTP in hippocampal neurons70. Hence, the presence of voltage-gated ion channels in dendrites plays a vital role in determining their intrinsic excitability as well as shaping synaptic inputs and integration and thereby induction and maintenance of plasticity.
Cellular neuroplasticity has been hypothesized to underlie experience-dependent behaviours such as learning and memory and drug addiction (Figure 1). Uncovering the cellular and molecular mechanisms of the acquisition, storage and recollection of memories is a major topic of basic and translational neuroscience research, as alterations in these mechanisms may contribute to multiple disease pathologies, including autism, epilepsy, Alzheimer's and Parkinson's disease. For the most part, regulation of individual synaptic input strength (synaptic plasticity) has received the most attention with a focus on the trafficking and properties of the neurotransmitter receptors themselves (α–amino-3-hydroxyl-5-methylisoxazole-4 propionate receptors (AMPARs) and NMDARs). However, a confluence of recent evidence indicates that subsequent to receptor activation, synaptic responses are regulated by dendritic voltage-gated channels and that these channels themselves are targeted for modulation. To fully understand how these channels contribute to various forms of plasticity is it critical to determine how their biophysical properties and subcellular localization are modulated.
Since many forms of cellular and synaptic plasticity result in altered activity of kinases and phosphatases, it is perhaps not surprising that activity-dependent changes also affect dendritic channel expression and properties (Figure 1). In distal CA1 dendrites, protein kinase A (PKA), protein kinase C (PKC) and extracellular-signal regulated kinase/mitogen-activated protein kinase (ERK/MAPK) all downregulate A-type K+ channel activity, resulting in enhanced action potential propagation71, 72 (Table 1). In addition, LTP induction in hippocampal slices shifts the voltage-dependence of steady-state IA inactivation leftward 73. These modulations both have the effect of increasing local dendritic excitability and enhancing action potential back propagation, which would lead to a change in the ability to induce subsequent potentiation (metaplasticity).
Moreover, Kv2.1 channels have a fascinating profile of phosphorylation-regulated activation. Not only does dephosphorylation of the channel by PP2B (protein phosphatase 2B- also known as calcineurin; Table 1) cause a hyperpolarized shift in its voltage-dependent activation74, but it does so in a graded manner. Using a proteomics approach, 16 phosphorylation sites were identified on the Kv2.1 channel, 7 of which are dephosphorylated by PP2B. The more sites dephosphorylated, the greater the shift in activation, with complete dephosphorylation yielding a large (~35 mV) hyperpolarized shift75. With their slow kinetics, such changes in activation would likely produce more Kv2.1 channels being activated during repetitive stimulation, and would subsequently suppress spiking during instances of neuronal excitability, providing a mechanism of homeostatic plasticity29.
Despite its lack of voltage-dependence, KCa2.2 (SK2) channel activation is also regulated through the phosphorylation state of its multiprotein complex. KCa2.2 channel activation occurs when Ca2+ binds calmodulin (CaM), which is itself constitutively bound at the channel C-terminus76. Also associated constitutively with KCa2.2 is casein kinase II (CK2) and protein phosphatase 2A (PP2A), which regulate the phosphorylation of KCa2.2-bound CaM77. While CaM phosphorylation by CK2 leads to faster channel deactivation and a reduced Ca2+-sensitivity, dephosphorylation by PP2A increases Ca2+ sensitivity. Interestingly, phosphorylation of CaM by CK2 is Ca2+- and state-dependent, only occurring when the channels are closed. The net result is a system of bidirectional modification of KCa2.2 channel activation, where during low activity (with infrequent Ca2+ signals) activation is reduced by CK2 and during repetitive stimulation or synaptic activity (with sustained Ca2+ signals) channel activation is enhanced by PP2A.
Similar to Kv and KCa channels, the resting state of HCN channels may also be regulated by phosphorylation. Several modulators, including 3′,5′-cyclic adenosine monophosphate (cAMP) and phosphoinositides, have been found, all of which shift the activation curves of HCN channels49. Hence, variation in the activity of these molecules by GPCR activity or synaptic strength would result in altered gating of these channels, leading to changes in synaptic potential shapes and integration, thereby augmenting the intrinsic excitability of neurons. Indeed, an elegant study by Wang et al. (2007) demonstrates that in spines of prefrontal cortical neurons, activation of α2 adrenoreceptors leads to a reduction in cAMP activity and HCN function (Table 1), thereby potentiating EPSP integration and elevating neuronal firing, eventually causing an increase in working memory59.
Many forms of plasticity involve depolarization of dendrites leading to opening of Ca2+ and Na+ channels. Na+ channels located in dendritic trunks are present in a phosphorylated state in some neurons78. An altered balance of kinases and phosphatases caused by changes in GPCR activity might lead to a change in the activation and inactivation curves of these channels. This would affect the initiation and back propagation of dendritic spikes, and could thereby alter the threshold for certain types of plasticity, such as spike-timing dependent plasticity. GPCR activity may also regulate the resting state of Ca2+ channels. Indeed, in hippocampal spines, activation of PKA by stimulation of β2-adrenoreceptors has been demonstrated to facilitate CaV1.x (L-type) Ca2+ channel activity79, thereby priming the induction of synaptic plasticity. Interestingly, these same pathways may also cause depression of other Ca2+ channel subtypes and block LTP80. Hence, the phosphorylated states of voltage-dependent ion channels in dendrites are critical for the generation of plasticity. This may also affect the maintenance of plasticity and thus, metaplasticity.
Throughout this section we have focused on regulation of ion channels by protein phosphorylation. However, it is likely that future research will uncover other forms of post-translational modifications (e.g. ubiquitination and palmitoylation) that contribute to dendritic ion channel sorting and localization and are therefore also potential sources of activity-dependent regulation of dendritic excitability.
In addition to post-translational modification of channel properties, active trafficking of dendritic ion channels also influences cellular and synaptic plasticity (Figure 1). For example, Kv4.2 channels are internalized from the dendritic membrane during synaptic plasticity (Figure 2). In hippocampal slices, Kv4.2 channels are internalized after LTP induction with a pairing protocol, and in cultured neurons with activation by AMPA, potassium chloride (KCl,) or glycine81. In this study, internalization was measured by a decrease in membrane-bound Kv4.2, a reduction in IA, and by real-time observation of green fluorescent protein (GFP)-tagged Kv4.2 redistributing from the dendritic spine to the shaft (Figure 2B). These effects were also NMDAR-dependent, supporting the model that Kv4.2 internalization occurs during Hebbian synaptic plasticity. The mechanism of Kv4.2 internalization likely involves clathrin-mediated endocytosis, since blocking dynamin recruitment to clathrin-coated pits prevented GFP-Kv4.2 redistribution and IA reduction.
Multiple proteins mediate Kv4.2 targeting and membrane expression, and these molecules may also play a role in its activity-dependent trafficking. The dendritic targeting of Kv4.2 subunits is dictated by a C-terminal dileucine motif 82, and Kv4.2 is transported by the motor protein Kif17, a kinesin isoform that binds to the extreme C-terminal end of the channel83. Kv4.2 cell surface expression is further regulated by a number of auxiliary subunits, including Kv4 channel interacting proteins (KChIPs) and dipeptidyl peptidase-like type II proteins, DPP6 and DPP1028, which bind to the N-terminus84 and S1/S2 domains85 of Kv4.2, respectively. An intriguing avenue for future research will be to uncover how post-translational modifications and membrane expression are related. For example, PKA phosphorylation of Kv4.2 is required for activity-dependent internalization86. But does phosphorylation trigger internalization or is it simply required for membrane localization of the mobile pool of channels? Additionally, how do post-translational modifications interact with auxiliary subunits to affect channel complex expression and properties? Auxiliary subunits themselves could be targets for modulation. For example, it has been recently shown that Kv4.2 primary subunit phosphorylation may be required for the auxiliary protein KChIP4a to regulate channel properties87.
Similar to Kv4.2, KCa2.2 channels are also internalized during LTP (Figure 2A, Table 1)88. In hippocampal slices, KCa2.2 channels are internalized after chemically-induced LTP or after physiologically-relevant LTP induction by theta burst stimulation88. This process also requires NMDAR activation and involves channel phosphorylation by PKA88. Clathrin-mediated endocytoysis of KCa2.2 subunits has also been demonstrated in lateral amygdala spines in an NMDAR- and PKA-dependent manner following LTP89. In this study, the authors suggest there is constitutive dynamin-dependent endocytosis of KCa2.2 channels, and PKA phosphorylation of the channel during stimulation sequesters it to the cytosolic compartment. The resulting effect is a reduction in functional synaptic KCa2.2 and enhanced LTP.
The expected consequence of reducing K+ channel density during synaptic activity is to enhance dendritic excitability and reduce the probability of further LTP induction. But what is the fate of Kv4.2 and KCa2 channels after activity-dependent internalization? Are they recycled back into the membrane or degraded? If the former, is reinsertion also subject to activity-dependent regulation? The spatial restriction of signaling events that trigger dendritic ion channel trafficking during plasticity is also unclear. That is, is internalization compartmentalized to the spine or could extensive spread contribute to intrinsic plasticity?90 Recent reports show that some NMDAR-activated signalling molecules like the guanosine triphosphatase Ras spread over 10 micrometers of dendrite and invade neighbouring spines91 whereas others, such as CaMKII, remain restricted to the activated spine 70. Recent advancements in fluorescent protein labeling and live cell imaging techniques may soon provide answers to such questions.
Interestingly, the same neuronal activation that reduces surface expression of Kv4.2 and KCa2.2 channels increases the surface expression of Kir channels (Figure 2A, Table 1). In hippocampal neurons, activation with KCl, glutamate, NMDA, or glycine reduces the surface expression of endogenous Kir3.1 and Kir3.2 channels92. This occurs through NMDAR-dependent activation of protein phosphatase 1 (PP1), which dephosphorylates Kir3.1-2 channels, causing their insertion into the membrane from recycling endosomes92. This NMDAR-dependent insertion of Kir3.1-2 channels may also regulate depotentiation of synapses, an input-specific and NMDAR-dependent form of synaptic plasticity important in maintaining bidirectional modification of synapses. In a recent study, Chung et al93 demonstrate that depotentiation of hippocampal synapses requires the activation of adenosine A1 receptors, PP1 and Kir3.1-2 channels- suggesting that the activity-dependent insertion of Kir3.1-2 channels into the membrane may contribute to the mechanism of depotentiation.
Together, these exciting findings raise the possibility that the input specificity of synaptic plasticity may in part be mediated by alterations in local dendritic K+ channel expression. Regulation of local protein synthesis and lateral translocation of Kv channels are also mechanisms by which their differential expression occurs. In hippocampal neurons, local Kv1.1 channel translation in dendrites is upregulated upon NMDA receptor inhibition, suggesting that activity can regulate K+ channel expression (Table 1)94. Moreover, in hippocampal pyramidal neurons, clustered somatodendritic Kv2.1 channels disperse laterally along the membrane after neuronal stimulation and dephosphorylation by PP2B (calcineurin) (Figure 2, Table 1)74, 95, 96. This dephosphorylation and translocation is accompanied by a hyperpolarizing shift in the activation and inactivation of Kv2.196, 97, enhancing the influence of Kv2.1 during repetitive firing. Interestingly, this effect is mediated by the activation of extrasynaptic NMDARs, and may be important for the regulation of intrinsic excitability of neurons during excitotoxic events95-97.
Like K+ channels, Hebbian plasticity at selective synapses results in activity-dependent alterations in HCN channels. Induction of NMDA-receptor dependent LTP via a theta burst protocol enhances HCN expression in hippocampal CA1 neurons98, 99. This effect is dependent on Ca2+ entry via NMDAR activation of CaMKII (Figure 3, Table 1)99. Conversely, metabotropic glutamate receptor-dependent LTD results in reduced HCN expression due to Ca2+ release from internal stores and activation of PKC (Figure 3, Table 1)100. Hence, depending on the source and possibly concentration, Ca2+ can bi-directionally regulate the membrane insertion of HCN channels.
One outstanding question is whether the plasticity-induced alterations in HCN function and expression involve post-translational modifications, as has been shown for LTP-induced changes in K+ channels81, 88, 89, 92, modulation of auxiliary subunits or variations in local protein synthesis. All three mechanisms may occur. HCN mRNA is abundant in dendrites101, 102 and the possibility that synaptic activity may influence local protein synthesis (as with KV1.194) or endocytic membrane re-cycling of HCN subunits cannot be ruled out. Excitingly though, HCN channels are actively trafficked to dendrites by binding to chaperone proteins known as TPR-containing Rab8b interacting protein (TRIP8b)103-106. Moreover, TRIP8b appears to be essential for membrane expression of HCN channels in hippocampal and cortical dendrites103, 105, 106. Multiple isoforms of TRIP8b have been identified, most of which enhance expression of dendritic HCN subunits105, 106. All isoforms of TRIP8b also alter the gating of HCN channels104-106. TRIP8b, like HCN channels, has phosphorylation consensus sites for a number of kinases106, 107 including CaMKII and PKC, raising the prospect that alterations in the activity of these kinases could dynamically regulate expression of TRIP8b activity and thereby influence HCN channel expression and characteristics at selective synapses and dendritic locations. In keeping with this, activity-dependent loss of TRIP8b and thus HCN channel expression, has been demonstrated to occur following excessive neuronal activity108.
Moreover, at some synapses HCN channels, and presumably TRIP8b subunits, may be located in close proximity to GPCRs. Activation of these GPCRs may additionally modulate HCN channel activity and expression and so influence the threshold of plasticity. This is certainly the case in pre-frontal cortical neurons, whereby HCN1 channels are co-localized with α2 adrenoreceptors59. In these neurons, activation of the α2 adrenoreceptors leads to a decrease in spine cAMP and HCN1 channel activity, resulting in enhanced LTP and working memory59 (Table 1). This is very intriguing as neither the gating properties nor the expression profile of heterologously expressed HCN1 channels are significantly affected by acute changes in cAMP109. Hence, it is possible that this may be due to modulation of accessory subunits such as TRIP8b, again raising the question of whether plasticity-dependent changes of HCN channel function are due to alterations in trafficking and membrane expression of the subunits.
Additionally, there are multiple isoforms of TRIP8b, which are expressed in hippocampal and cortical neurons105, 106. Interestingly, one of these isoforms inhibits rather than enhances HCN expression105, 106, raising the possibility that plasticity induced changes in HCN channel function may involve an altered balance in the activity of these TRIP8b isoforms. Hence, plasticity may not induce changes in TRIP8b expression per se but may simply result in increased activity of one isoform over the others, causing altered HCN subunit membrane expression. These are all open questions that still need to be investigated, perhaps using new tools such as isoform-specific antibodies or transgenic mice lacking selective isoforms.
In summary, we have discussed above how the activity and expression of dendritic ion channels can be dynamically regulated by alterations in intrinsic neuronal firing and changes in synaptic activity. Whilst enormous strides have been made in understanding how several subtypes of voltage-gated ion channels are selectively targeted to dendrites and how plasticity affects the dendritic trafficking of these channels, much less is known about others. For example, dendritic Na+ and Ca2+ channel function is altered during synaptic plasticity (Table 1)80, 110 but whether these changes in function are due to variations in expression and trafficking of the subunits remains to be explored. Future studies are also required to determine how multiple trafficking events synchronize during plasticity. For instance dendritic ion channels such as Kv4.2 and KCa2.2 channels are internalized while AMPA type glutamate receptors are inserted into the membrane during LTP creating a potential traffic jam. Are these events coordinated sequentially or are they independently regulated? Related to this, do the same trafficking events that lead to plasticity-induced changes in one dendritic ion channel trigger alterations in other ion channel properties to maintain homeostasis? Is mRNA translation co-regulated for different types of dendritic ion channels? Clearly, much remains unknown, and the answers will be especially rewarding, increasing our understanding of dendritic integration, basic biological signalling mechanisms, and cellular and synaptic plasticity.
This work was supported by an MRC New Investigator Award (G0700369, MMS), a Wellcome Trust project grant (WT087363MA, MMS) and the Intramural Research Program of the National Institutes of Health and the National Institute of Child Health and Human Development (DH).