Protein kinase Cε: function in neurons

Authors


Y. Shirai, Laboratory of Molecular Pharmacology, Biosignal Research Center, Kobe 657-8501, Japan
Fax: +81 78803 5971
Tel: +81 78803 5962
E-mail: shirai@kobe-u.ac.jp

Abstract

Protein kinase Cε is expressed at higher levels in the brain compared to other tissues such as the heart and kidney, suggesting that it plays an important role in the nervous system. Several neural functions of PKCε, including neurotransmitter release and ion channel regulation, have been identified using PKCε knockout mice. In this review, we focus on the involvement of protein kinase Cε in neurite outgrowth, presynaptic regulation, alcohol actions, ischemic preconditioning and pain.

Abbreviations
ERK

extracellular signal-regulated kinase

GABAA

γ-aminobutyrate type A

KO

knockout

LTP

long-term potentiation

NMDA

N-methyl-d-aspartate

PIP2

phosphatidylinositol 4,5-bis phosphate

PKCε

protein kinase Cε

RhoA

Ras homolog gene family, member A

ROCK

Rho-associated coiled-coil-containing protein kinase

Expression of PKC- in the brain

Northern- and immunoblot analyses of the brain have revealed the predominant presence of protein kinase Cε (PKCε) mRNA and protein, respectively [1,2]; the subtype has been cloned from rat brain [3]. Immunohistochemistry of PKCε reveals the most abundant expression of the enzyme in the hippocampus, olfactory tubercle and Calleja’s islands, with moderate expression in the cerebral cortex, anterior olfactory nuclei, accumbens nucleus, lateral septal nuclei and caudate putamen (Table 1). The distribution of PKCε in the brain is consistent with results of in situ hybridization [4]. Interestingly, the immunoreactivity of the protein is evident in the nerve fibers, and precise observation using electron microscopy reveals its presynaptic localization [4,5]. These findings suggest the involvement of PKCε in neurite outgrowth and presynaptic functions such as neurotransmitter release.

Table 1.   Relative densities of PKCε immunoreactivity in the rat brain. 0, no immunoreactivity; 1, faint immunoreactivity; 2, lower immunoreactivity; 3 moderate immunoreactivity; 4, moderately dense immunoreactivity; 5, most dense immunoreactivity. Modified from Table 1 of [4].
RegionRegion
Olfactory bulbThalamus
 Olfactory nerve layer0 Reticular nucleus1
 External plexiform layer3 Dorsal nuclei1
 Internal granular layer2 Ventroposterior nucleus1
 Glomerular layer2 Lateral geniculate nuclei1
 Mitral cell layer1 Medial geniculate nuclei2
Anterior olfactory nuclei3Cerebellar cortex
Amygdara2 Molecular layer2
Caudate-putamen3 Purkinje cell layer0
Globus pallidus2 Granular layer1
Accumbens nucleus3Substantia nigra 
Olfactory tubercle4 Pars compacta1
Callrja’s island4 Pars reticulate2
Septal areaMammilary nuclei2
 Laternal septal nucleus3Suprior colicullus1
 Medial septal nucleus1Inferior colicullus1
 Diagonal band1Pontine nuclei1
 Septo-hippocampal nucleus4Locus coeruleus1
HabnullaMesencephalic trigeminal nucleus2
 Medial0Inferior olive2
 Lateral0Vestbular nuclei0
NeocortexCochlear nucleus0
 Layer I3Solitary nucleus0
 Layer II3Gracile nucleus0
 Layer III1Cuneate nucleus0
 Layer IV2Spinal cord 
 Layer V2 Substantia gelatinosa2
 Layer VI2 Vental horn1
HippocampusWhite matter
 CA14 Optic nerve0
 CA24 Anterior commissure0
 CA35 Corpus callosum0
Dentate gyrus Pyramidal tract2
 Hilus4 Facial nerve1
 Granular layer2 Cochlear nerve1
 Molecular layer3 Cerebral peduncle1
   Spinal trigeminal tract1
   Medial longitudinal fasciculus1
   Inferior cerebellar peduncle1

As little is known about the role of PKCε in brain development, we measured changes in the amount of PKCε protein in the rat brain from birth to day 28. Substantial amounts of PKCε were already detected at birth but substantial increases were detected in the forebrain between days 5 and 7, and in the hindbrain between days 7 and 14 (Fig. 1). This remarkable increase suggests the importance of PKCε for neural network construction because its timing is coincident with synapse formation in the brain.

Figure 1.

 The ontogeny of PKCε in the rat brain as assessed by immunoblot analysis. The brain was dissected from rats of various ages and cut at the level of the caudal end of the inferior colliculus. The rostal half (F) and the caudal half (H) were homogenized and the homogenates (50 μg of protein) were subjected to SDS/PAGE, followed by a western blot for PKCε.

Neurite outgrowth

Overexpression of PKCε promotes nerve growth factor-induced neurite outgrowth, whereas its down-regulation inhibits this [6]. Interestingly, the phenomenon is independent of its catalytic activity because expression of the regulatory domain alone (εRD) induces neurite outgrowth [7]. Larsson et al. [8] demonstrated that the actin binding site between the C1A and C1B is important for morphological change of neurons. They also reported that the PKCε-induced neurite outgrowth is blocked by active Ras homolog gene family, member A (RhoA) and led by inhibition of the RhoA effector Rho-associated coiled-coil-containing protein kinase (ROCK), indicating that attenuation of the RhoA-ROCK pathway is involved in the process [9]. Additionally, activation of Cdc42 is implicated in PKCε-induced neurite outgrowth [10].

How PKCε regulates RhoA and Cdc42 is still unknown. We have recently shown that PKCε binds to phosphatidylinositol 4,5-bis phosphate (PIP2) and that the PIP2 binding is necessary for PKCε-induced neurite induction and its membrane localization [11]. The PIP2 binding of PKCε may influence the function of actin binding proteins, leading to actin rearrangement and neurite induction. In addition, the binding of PKCε to PIP2 may contribute towards inhibition of the RhoA-ROCK pathway. For example, the binding of PKCε to PIP2 may reduce the level of free PIP2 that can activate RhoA by regulating the open state of the RhoA/RhoGDI complex [12]. Alternatively, PKCε may attenuate RhoA via RhoGAP, which is reported to bind to PKCε and induce neurite outgrowth [13]. Interestingly, the binding of full length PKCε to RhoGAP and PIP2 is dependent on 12-O-tetradecanoylphorbol 13-acetate, but εRD can bind to PIP2 without 12-O-tetradecanoylphorbol-13-acetate [11,13]. These results suggest that the open conformation is important for the binding of PKCε to RhoGAP and PIP2. Actin binding to PKCε may regulate the neurite cytosckelecton as well as induce the open conformation of PKCε, permitting its interaction with RhoGAP and/or PIP2. Taken together, these studies suggest a model for the involvement of PKCε in neurite function. Hormones and neurotransmitters, including nerve growth factor, activate PKC, resulting in a conformational change that accompanies translocation to the plasma membrane. Activated PKCε on the plasma membrane interacts with actin, RhoGAP and PIP2, resulting in neurite outgrowth by inhibiting the RhoA-ROCK pathway, activation of Cdc42 and cytoskeletal rearrangement (Fig. 2). Although the C1B and V3 regions of PKCε play a function in the induction of neurite outgrowth, the mechanism by which this occurs has yet to be determined [11,14].

Figure 2.

 Proposed model for neurite induction by PKCε. Activation of PKCε results in an open conformation and translocation to the plasma membrane. The PKCε on the plasma membrane interacts with actin, RhoGAP and PIP2 via the actin binding site C1 and V3 regions, resulting in neurite outgrowth by inhibiting the RhoA-ROCK pathway, activation of Cdc42, and cytoskeletal rearrangement.

Presynaptic functions

Long-term potentiation (LTP) is at least one component in the complex mechanism of learning and memory. There is general agreement that calmodulin-dependent kinase II plays an essential role in this phenomenon. PKC is also thought to be necessary, but not sufficient, to induce LTP based on the findings that phorbol esters mimic LTP and PKC inhibitors prevent it [15].

The mechanisms underlying LTP have been most extensively studied in the hippocampus, although LTP occurs in a number of brain regions. There are two types of LTP in the hippocampus [15]; one is LTP in the pathway from the Schaffer collaterals to CA1 (SC-CA1) and the other is the pathway from the mossy fibers to CA3 (MF-CA3). The former is calcium-dependent and involves N-methyl-d-aspartate (NMDA) receptor phosphorylation by PKCγ in the postsynaptic neurons [16]. The latter appears to be mediated by presynaptic events. PKCε is present at the terminals of neurons and is localized at the presynapses of the mossy fibers, consistent with a role for PKCε in LTP at MF-CA3 [3,4]. Indeed, PKCε at the nerve terminal is involved in phorbol ester-induced enhancement of glutamate exocytosis [17] and in phorbol ester-induced synaptic potentiation [19]. Thus, PKCε at the nerve terminal would be expected to contribute to the MF-CA3 LTP by increasing presynaptic neurotransmitter release. Generally, sustained activation of PKC is needed for the presynaptic regulation of neural plasticity [4]. The importance of the actin binding site in both the sustained activation of PKCε and enhanced exocytosis has been reported [18]. However, how PKCε is activated presynaptically during LTP is not fully understood. One attractive mechanism is that of the ‘retrograde messenger’. Arachidonic acid produced at a postsynaptic site could diffuse to the presynaptic terminal to activate PKCε. Indeed, arachidonic acid is released from cultured neurons by activation of NMDA receptors [19] and subtype-specifically activates PKCε [20].

Presynaptic PKCε is also important for synaptic maturation. It is well known that co-culture of purified neurons with astrocytes facilitates synaptogenesis and synapse maturation. Hama et al. [21] reported that contact of neurons with astrocytes enhances the excitatory postsynaptic potential and induces excitatory synapses, and that facilitated excitory synaptogenesis is blocked by inhibitors of PKC [19]. Of the several PKC subtypes present in the presynaptic neurons, Hama et al. [19] propose that PKCε plays a key role in the phenomenon because arachidonic acid production by phospholipase A2 is necessary for PKC activation and synaptogenesis [19]. These results strongly suggest that PKCε is involved in presynaptic modulation and regulation, although there is no consensus on the involvement of PKCε in LTP.

Actions of alcohol

PKCε is thought to mediate several actions of ethanol. For example, ethanol stimulates the translocation of PKCε in NG108-15 cells [22] and chronic ethanol exposure increases the amount of PKCε in NG108-15 and PC cells [23,24]. More directly, PKCε null mice show a higher sensitivity than their wild-type littermates to the acute behavioral effects of ethonal, and demonstrate a marked reduction in ethanol self-administration [25]. The involvement of PKCε in the actions of alcohol was confirmed in PKCε transgenic mice [26]. Conditional expression of PKCε in the basal forebrain, amygdala and cerebellum of PKCε null mice rescues the hypersensitivity and restores ethanol consumption. Also, doxycycline-induced reduction of PKCε expression results in a knockout (KO) mice-like phenotype [26]. As the conditional transgenic mice do not express PKCε in the hippocampus, these effects of PKCε on the response to ethanol are unlikely to be the result of responses in the hippocampus.

The hypersensitivity to and avoidance of ethanol in PKCε KO mice appears to be mediated by the γ-aminobutyrate type A (GABAA) receptor. This is based on studies demonstrating that KO mice showed a greater increase in locomotor activity than wild-type mice in response to pentobarbital and diazepam, which are allosteric activators of the GABAA receptor [25,26]. GABAA receptors are ligand-gated Cl channels that are considered to be an important target of ethanol. GABAA receptors are pentameric proteins complexes comprising eight different classes. Most GABAA receptors are compossed of two α, two β and one γ2 subunit. Recently, it has shown that PKCε directly phosphorylates S327 in the large intracellular loop of the γ2 subunit and that mutation of this site enhanced the ethanol-induced GABA-stimulated current [27]. These results confirm that PKCε regulates the sensitivity of GABAA receptors to ethanol via direct phosphorylation.

Finally, chronic ethanol exposure up-regulates N-type calcium channels. Because this up-regulation can be inhibited by a selective inhibitor of PKCε [28], ethanol induces N-type calcium channel expression in a PKCε-dependent manner. These results suggest that inhibition of PKCε might comprise a viable treatment for alcoholism.

Ischemic preconditioning

Subleathal and mild ischemic insult, or ‘preconditioning’, promotes tolerance against more severe subsequent ischemic insults in organs such as the heart and brain. This phenomenon involves many factors, including PKC [29–31]. Involvement of PKCε in the preconditioning has been well established in cardiac cells using PKCε-specific peptide activators and inhibitors [29,31] and has been confirmed using PKCε KO mice [32]. Unlike wild-type and heterozygous mice, preconditioning in PKCε KO mice does not reduce infarct size caused by ischemia reperfusion, implicating the involvement of PKCε in preconditioning. These results were obtained from a non-neural system but PKCε acts in a similar way in neurons.

Indeed, the role of PKCε in neural preconditioning has been investigated using hippocampal slices [33,34] and primary cultured neurons [35,36] as well as PKCε-specific peptide activators and inhibitors. According to these studies, NMDA and adenosine receptor-mediated neural preconditioning require PKCε activation. Although the mechanism of PKCε-mediated neural preconditioning is not fully understood, inhibition of extracellular signal-regulated kinase (ERK) attenuated the adenosine receptor-mediated neural preconditioning, implicating the involvement of ERK in preconditioning [34]. These findings suggest that PKCε may have a protective role in stroke. Recently, Shimomura et al. [37] demonstrated that the levels of PKCε are markedly lower in the core of focal cerebral ischemia and that this loss was prevented by hypothermia, which is known to be neuroprotective [37]. How hypotheramia alters levels of PKCε is currently unknown.

Pain

PKCε also localizes and functions in peripheral neurons such as nociceptive neurons. The nociceptive sensory neurons express the capsaicin receptor TRPV1, which is a nonselective cation channel activated not only by capsaicin, but also by heat (> 43 °C). TRPV1 is essential for the sensation of thermal and inflammatory pain [38] and pro-inflammatory signals, including ATP and bradykinin, enhance TRPV1 activity in a PKC-dependent manner [39,40]. Among the PKC subtypes, PKCε has been reported to be predominantly and specifically involved in nociceptor sensitization [39,40]. Indeed, PKCε directly phosphorylates Ser502 and Ser800 of TRPV1 [41]. It has also been shown that desensitization of TRPV1 is regulated by PKCε-mediated phosphorylation at Ser800 [42]. Furthermore, phosphorylation by PKCε appears to contribute to the proteinase-activated receptor 2-mediated potentiation of TRPV1 [43]. These findings suggest that PKCε may be a therapeutic target for regulating TRPV1 and pain.

Other functions

PKCε also modulates the Na+ channel in hippocampal neurons [44]. Acetylcholine binding to muscarinic receptors activates G-proteins, phospholipase C and PKCε, resulting in a reduction of the peak of the Na+ current in hippocampal neurons. This reduction is not observed in PKCε KO mice, implicating PKCε in the regulation of Na+ channels. Such modulation of Na+ channels by PKCε is likely to affect integration of depolarizing inputs in dendrites and the threshold and frequency of firing of action. In addition, PKCε is involved in phorbol ester-induced secretion of β-amyloid precursor protein and the reduction of β-amyloid (Aβ) peptides [45]. Neural overexpression of PKCε decreased Aβ levels via endothelin-coverting enzyme [46], suggesting that PKCε is one of the regulators of α-secretase and Aβ production.

PKCε in the non-neural system is sometimes related to the neuronal system. For example, astrocytes have traditionally been considered as passive bystanders in the formation and operation of the neural network, but accumulating evidence argues against such a model, and instead supports a model in which astrocytes play a critical role in the creation and control of synapses. Interestingly, differentiation of astrocytes, in part, is regulated by PKCε. Sterinhart et al. [47] demonstrated that 4β-phorbol 12-myristate 13-acetate and PKCε overexpression induces the differentiation of multipotential neural precursor cells to astrocytes, and that this induction is inhibited by a kinase negative mutant of PKCε [47]. They also demonstrated the involvement of Notch in this processes.

Perspective

As described above, PKCε plays several important roles in neurons. However, its precise function and mechanism of action in neurons is yet to be fully understood compared to the other functions determined for this enzyme [48]. One of the reasons is that several PKC isoforms exist in the same neuron [3,4] and different PKC subtypes can have opposing functions. For example, PKCγ is also involved in GABAA receptor regulation in response to ethanol but, in contrast to PKCε, PKCγ plays an inhibitory role in ethanol-induced enhancement of GABAA receptor-mediated inhibitory postsynaptic currents in CA1 [49]. Ethanol enhanced inhibitory postsynaptic currents in wild-type mice, but not in PKCγ null mice. By contrast, these currents are potentiated in PKCε KO mice. In addition, unlike the protective effect of PKCε on reperfusion injury, PKCδ exacerbates injury [30].

The PKCε KO mice have been an invaluable model for discriminating between the effects of εPKC and other PKCs. Notably, the importance of PKCε in the ethanol sensitivity and in Na+ channel regulation have been clearly demonstrated using KO mice. By contrast, the involvement of PKCε in the LTP or its protective effect on stoke have not been validated using PKCε KO mice. It is possible that such experiments are ongoing or that the KO mice might be somehow different from wild-type mice due to compensation by other PKC subtype(s). Because PKCε obviously has important functions in neurons, more specific inhibitors and activators would be useful to define the precise and subtype-specific functions of PKCε. With respect to delivery to the brain, development of a chemical inhibitor of PKCε that crosses the blood–brain barrier is necessary because the PKCε inhibitors used so far are peptides that cannot be employed effectively for in vivo neural studies. These PKCε-specific inhibitors or activators are expected to be utilized as a drug for stroke, Alzheimer’s disease and pain because the functions of the enzyme have been reported as descibed above.

Ancillary