OUP user menu

Location, Location, Location: Altered Transcription Factor Trafficking in Neurodegeneration

Charleen T. Chu MD, PhD, Edward D. Plowey MD, PhD, Ying Wang PhD, Vivek Patel BA, Kelly L. Jordan-Sciutto PhD
DOI: http://dx.doi.org/10.1097/nen.0b013e318156a3d7 873-883 First published online: 1 October 2007

Abstract

Neurons may be particularly sensitive to disruptions in transcription factor trafficking. Survival and injury signals must traverse dendrites or axons, in addition to soma, to affect nuclear transcriptional responses. Transcription factors exhibit continued nucleocytoplasmic shuttling; the predominant localization is regulated by binding to anchoring proteins that mask nuclear localization/export signals and/or target the factor for degradation. Two functional groups of karyopherins, importins and exportins, mediate RanGTPase-dependent transport through the nuclear pore. A growing number of recent studies, in Alzheimer, Parkinson, and Lewy body diseases, amyotrophic lateral sclerosis, and human immunodeficiency virus encephalitis, implicate aberrant cytoplasmic localization of transcription factors and their regulatory kinases in degenerating neurons. Potential mechanisms include impaired nuclear import, enhanced export, suppression of degradation, and sequestration in protein aggregates or organelles and may reflect unmasking of alternative cytoplasmic functions, both physiologic and pathologic. Some "nuclear" factors also function in mitochondria, and importins are also involved in axonal protein trafficking. Detrimental consequences of a decreased nuclear to cytoplasmic balance include suppression of neuroprotective transcription mediated by cAMP- and electrophile/antioxidant-response elements and gain of toxic cytoplasmic effects. Studying the pathophysiologic mechanisms regulating transcription factor localization should facilitate strategies to bypass deficits and restore adaptive neuroprotective transcriptional responses.

Key Words
  • Alzheimer disease
  • Cyclic AMP response element-binding protein
  • Karyopherins
  • Neurodegeneration
  • Oxidative stress
  • Parkinson disease
  • Transcription factors

Introduction

With their highly polarized cell biology, neurons face unique challenges in intracellular trafficking of key regulatory signaling proteins. Whereas basal transcription of "housekeeping" genes is essential for cell survival, neurons also depend on efficient regulation of transcription in response to both physiologic and pathologic stimuli for maintenance of neuronal plasticity and survival (1). At the most basic level, transcription is regulated by the binding of transcription factors to specific DNA sequences in the promoter region of targeted genes. However, gene expression is also regulated by accessory factors, chromatin modifiers, and signaling molecules, which couple alterations of gene expressions to physiologic and pathologic stimuli. As DNA is confined to the nucleus (and the mitochondrion, which has its own system of transcriptional regulation), appropriate subcellular trafficking of transcription factors plays a particularly important role in determining their biologic activities. Recently, a growing body of literature indicates that some transcription factors display alterations not only in expression levels and phosphorylation, but also in their subcellular distribution within populations of neurons affected by neurodegenerative diseases. The implications of altered transcription factor localization are not fully understood, although functional repression of adaptive prosurvival transcriptional programs has been reported (2). In the following section we review the literature on transcription factors and related proteins that display altered subcellular localization in human neurodegenerative diseases. This review is followed by a discussion of nuclear import and export mechanisms with integration of human and experimental observations to illustrate potential neurodegenerative disease mechanisms and their therapeutic implications.

Transcription Factors in the Cytoplasm of Degenerating Human Neurons

Since numerous proteins are alternatively expressed in neurodegenerative diseases such as Alzheimer disease (AD), Parkinson disease (PD), Huntington disease (HD), and amyotrophic lateral sclerosis (ALS) (3), a cohesive conceptual model accounting for these observed changes would be beneficial. Many proteins and/or their phosphorylated forms that exhibit increased expression in neurodegeneration are also alternatively localized with regard to their normally ascribed functions. However, the significance of such changes has often been overlooked or incompletely described. Here we will highlight the importance of subcellular distribution in neurodegenerative diseases by discussing the localization of proteins with well-defined functions associated with a specific subcellular compartment-the nucleus (Table). In the interest of space, we have limited our discussion to a few key transcriptional regulators; however, we hope this review stimulates further interest in the study of how changes in subcellular distribution of proteins may contribute to neurodegenerative disease progression.

View this table:
TABLE

Nuclear Factor-κB

The first transcription factor shown to be regulated by altered subcellular distribution was nuclear factor-κB (NF-κB). The NF-κB family consists of p50, p52, p65, c-Rel, and Rel-B, which share an amino-terminal 300-amino acid Rel homology domain for DNA binding, homo- and heterodimerization and nuclear localization (4). The activity of NF-κB dimers is inhibited by sequestration in the cytoplasm through interaction with inhibitor of nuclear factor-κB (IκB) proteins, which masks the NF-κB nuclear localization and DNA binding domains. In neurons, the most common NF-κB complex consists of p65, p50, and IκBα (4). Phosphorylation or oxidation of IκB proteins results in ubiquitination and degradation of IκB. This allows NF-κB to translocate to the nucleus and increase expression of adaptive genes such as brain-derived neurotrophic factor (BDNF), inducible nitric oxide synthase, and Ca2+/calmodulin-dependent protein kinase II as well as proapoptotic genes such as p53, Fas ligand, and Bcl-Xs (5).

Several stimuli associated with neurodegenerative diseases cause nuclear translocation of NF-κB, including oxidative stress and inflammatory mediators (4), and the distribution of NF-κB has been investigated in AD, PD, and ALS. In patients with AD, increased levels of neuronal p65 immunoreactivity were observed in the vicinity of early plaques in the basal forebrain, hippocampus, and cortex (6, 7). However, active, nuclear p65 immunoreactivity was decreased near plaques in patients with AD in comparison to age-matched controls (8). In contrast, postmortem PD brains exhibited increased nuclear translocation of NF-κB in dopaminergic neurons of the substantia nigra and ventral tegmental area (9). Neither nuclear nor cytoplasmic NF-κB was found to be increased in patients with ALS compared with control patients in 1 study (10); however, another report demonstrated nuclear NF-κB in astrocytes of the ALS spinal cord, but not in motor neurons (11). Taken together, these observations suggest that nuclear translocation of NF-κB differs among several neurodegenerative diseases, which may provide clues to the distinct pathology and neuronal subpopulations affected in these diseases.

Activating Transcription Factor 2

Activating transcription factor 2 (ATF2) belongs to the basic region leucine zipper family and is an important member of the activator protein-1 (Ap-1) family. ATF2 is ubiquitously expressed in the mammalian brain. Knockout studies revealed that ATF2 is required for neuronal development and migration, but little is known about its specific functions in neuronal cells (12). Localization of ATF2 is regulated by a nuclear export signal (NES) in the leucine zipper region and 2 nuclear localization signals (NLS) in the basic region, resulting in continuous shuttling between the cytoplasm and the nucleus. Heterodimerization of ATF2 and c-Jun in the nucleus prevents the nuclear export of ATF2 and leads to enhanced transcriptional activation (13).

ATF2 expression was investigated in AD, PD, and HD. In AD, there is downregulation of ATF2 expression in the hippocampus (12) and enhanced cytoplasmic ATF2 expression in cortical neurons (14). Although these 2 findings are in distinct brain regions, both indicate a decrease in the nuclear/cytoplasmic ratio of ATF2. Many ATF2 target genes are upregulated in AD. Both c-fos and c-Jun exhibit intense immunolabeling within nuclei in hippocampus in AD brain (15). Interestingly, nuclear localization of the upstream kinases phospho-c-Jun N-terminal kinase (JNK) and phospho-JNK kinase 1 (JKK1) were found only in low-stage AD, suggesting that ATF2 responses are an early event in AD pathogenesis (16). ATF2 downregulation has also been observed in the substantia nigra pars compacta (SNc) of PD brains and in the caudate nucleus of HD brains (12); however, the possibility of early-stage increases followed by late-stage decreases was not addressed in these studies.

Cyclic AMP Response Element-Binding Protein

Cyclic AMP response element-binding protein (CREB) belongs to the family of dimerizing leucine zipper transcription factors, comprising both activating transcription factors such as CREB and activating transcription factor 1 (ATF 1) and repressive transcription factors such as cAMP-responsive element modulator (CREM) and inducible cAMP early repressor (ICER). As more than 100 genes important for neuron function contain a cAMP response element (CRE) in their promoter region, CREB has become the most widely studied transcription factor in neurons (17) and is implicated in memory, plasticity, and survival (1).

Many signaling pathways converge upon CREB, resulting in phosphorylation of Ser133, including protein kinase A (PKA), calcium-dependent kinases (CaMKs), phosphoinositide-3 kinase (PI3K)/Akt, extracellular signal regulated protein kinase (ERK), and phospholipase C (PLC) (18). Phosphorylation at other sites is also believed to modulate activity but has been less studied.

In cognitively normal populations, immunostaining for CREB is almost exclusively nuclear in the hippocampus, dorsal root ganglia, substantia nigra, and neocortex (19-21). In contrast, brains of patients with AD exhibit decreased levels of the active phosphorylated form of CREB (pCREB) (22). These findings suggest that loss of CREB function may contribute to changes observed in AD (23). Consistent with this, CREB dephosphorylation has been reported in cerebellar granule neurons overexpressing human tau (24). However, further analysis of pCREB in AD brain tissues is required to determine whether the decrease in pCREB occurs through altered expression or phosphorylation and whether subcellular compartmentalization is affected.

Although the expression of total CREB has not been characterized in the brains of patients with PD, there are distinct alterations in pCREB localization in degenerating SNc neurons. Nuclear pCREB is not detected in SNc neurons by immunohistochemistry in postmortem human tissues but is found under control conditions in cultured primary dopaminergic neurons (2). In PD brains, cytoplasmic aggregates of pCREB are observed in SNc neurons in the absence of nuclear staining, whereas glial cells in both control and PD patient tissues show nuclear pCREB staining (2). In primary midbrain cultures, 6-hydroxydopamine (6-OHDA) causes progressive accumulation of pCREB in the cytoplasm and decreased pCREB in the nucleus of dopaminergic neurons but not in nondopaminergic neurons. Downregulation of CREB-regulated prosurvival gene targets Bcl-2 and BDNF was also observed in 6-OHDA-treated cells (2). In neonatal 6-OHDA-lesioned rats, nuclear pCREB is also absent in SNc neurons, being restricted to the medial prefrontal cortex (25). As CREB nuclear activities play a central role in neuronal adaptation and survival, cytoplasmic sequestration of this transcription factor may represent a novel mechanism contributing to neuronal degeneration and death.

Less is known about the distribution of CREB or pCREB in patient material in HD. However, aberrant transcriptional regulation has been implicated from observations that mutant huntingtin protein may show DNA binding characteristics. In animal models of HD, it has been shown that sequestration of CREB binding protein into nuclear aggregates contributes to impaired CRE transactivation and neuronal cell death (26). Although other studies show an increase in CRE-reporter transcription in vivo (27), the response element sequences from different genes may be differentially modulated through cross-talk, and loss of BDNF transcription has been proposed to play a central role in HD (28). Thus, even when CREB is localized to the nucleus, adaptive transcription may still be impaired.

TAR DNA Binding Protein 43

TAR DNA binding protein 43 (TDP-43) is a highly conserved, widely expressed nuclear ribonucleoprotein. It was originally identified by its ability to bind pyrimidine-rich motifs in the human immunodeficiency virus (HIV)-1 TAR element, functioning as a transcriptional inhibitor (29). TDP-43 has subsequently been shown to regulate splicing by binding RNA (30); however, its physiologic role remains elusive. Interestingly, TDP-43, which is normally a nuclear protein, was recently reported to accumulate in cytoplasmic inclusions in both frontotemporal lobar dementias and ALS (31). Concomitantly, nuclear TDP-43 was reduced in neurons. Furthermore, TDP-43 was universally present in ubiquitin-positive, tau-free inclusions in all subtypes of frontotemporal lobar dementias, the first protein besides ubiquitin to be localized to tau-free inclusions (31). These observations suggest a role for altered TDP-43 function(s) in these neurodegenerative diseases.

p53

The transcription factor p53 regulates essential cellular functions including cell cycle arrest, DNA repair, and cell death (32). Mechanisms of p53 regulation are complex; however, its physiologic roles are tightly linked to subcellular localization. Nuclear localization of p53 enables transcriptional activation of genes that regulate cell cycle arrest and apoptotic cell death (33). Mounting evidence also suggests that p53 can activate apoptosis through a transcription-independent mechanism involving localization to the mitochondria (34).

In postmortem brain tissues from patients with AD (35), diffuse Lewy body disease, frontotemporal lobar dementia, PD, and ALS, increased p53 immunoreactivity was observed in neuronal nuclei coincident with nuclear DNA fragmentation (36). However, increased p53 in other neurons was sometimes diffuse or cytoplasmic (36). Finding nuclear p53 in conjunction with DNA fragmentation (37) is consistent with the well-characterized transcription-dependent apoptotic pathway induced by p53 in numerous non-neuronal cells. However, the localization pattern in neurons without DNA fragmentation needs to be considered further in comparison with controls to understand the significance of increased cytoplasmic p53 (34, 38). The role of p53-dependent cell death in experimental models of neurodegeneration is highly controversial. There is as much evidence to support a role for p53 in neuronal cell death as there is evidence that p53 is not required. The significance of p53 in neurodegenerative disease needs to be considered in light of its subcellular distribution and the mechanisms that regulate these changes.

Sma/Mothers Against Decapentaplegic

The transcription factors that function downstream of transforming growth factor β can be divided into receptor-regulated Sma/mothers against decapentaplegic (Smads), which interact with Smad 4 to transduce signals, and the inhibitory decoy Smads (39). Nuclear retention of RSmads is stabilized by phosphorylation-dependent binding to Smad 4, allowing recruitment of DNA-binding cofactors. Smad 2 enters the nucleus via direct nucleoporin interactions, whereas both direct FG interactions and importin-dependent pathways are important for Smads 3 and 4. Interestingly, a recent study shows decreased Smad 3 nuclear localization in tangle-bearing neurons in AD and increased cytoplasmic Smads associated with insoluble tau and granulovacuolar degeneration (40).

E2 Promoter Binding Factor 1

E2 promoter binding factor (E2F) transcription factors participate in cell cycle progression and apoptosis in non-neuronal cells (41) and have been linked to regulation of neuronal survival and death (42). The ability of E2F1 to promote cell cycle progression or apoptosis depends on transcriptional activation of a subset of E2F target genes. Transactivation by E2F1 is repressed by interaction with the retinoblastoma protein in most cell types (43); however, in neurons, a novel, melanoma antigen gene (MAGE)-like protein, necdin, has also been linked to repression of E2F1 activity (44). Endogenous E2F1 protein levels are nearly undetectable in postmitotic neurons, but E2F1 mRNA is abundantly expressed (45, 46), so the low level of E2F1 protein may contribute to maintenance of the postmitotic phenotype in neurons. E2F1 has a NLS and DNA binding activity. Moreover, treatment of primary cortical neurons with an inhibitor of nuclear export, leptomycin B, led to nuclear accumulation of E2F1, whereas untreated cells had predominantly cytoplasmic E2F1 (47). This suggests a role for active nuclear export, although the E2F1 NES remains undefined (48).

Whereas mature neurons permanently exit the cell cycle, S-phase reentry has been implicated in neurodegeneration (46). As E2F1 is principally responsible for S-phase progression, its expression in neurodegenerative diseases was explored in AD, ALS, and HIV encephalitis. In AD and ALS, E2F1 was observed predominately in the cytoplasm of cortical neurons and motor neurons of the spinal cord, respectively (49-51). In another neurodegenerative disease, HIV encephalitis, E2F1 protein was increased in neurons and glia in both the nucleus and cytoplasm (52-54); however, potential alterations in the nuclear/cytoplasmic ratio were not addressed in these studies. The presence of E2F1 in the cytoplasm has only been reported in 1 other cell type, mature skeletal myocytes, another terminally differentiated cell population (55). The role of cytoplasmic E2F1 is not known, although it has been reported that expression of E2F1 causes neuronal degeneration coincident with aberrant S-phase entry (56), and E2F1 can induce apoptosis independent of its transcriptional activity (57). In non-neuronal cells, E2F1 mutants lacking the NLS and DNA binding domain induce calpain activation, and a significant reduction in antiapoptotic MDMX protein was coincident with an increase in E2F1 in a simian model of HIV encephalitis (47). Additional studies are needed to elucidate the function(s) of cytoplasmic E2F1 and mechanisms by which the balance of nuclear versus cytoplasmic E2F1 may regulate neuronal survival.

Nuclear Factor E2-Related Factor 2

Nuclear factor E2-related factor 2 (Nrf2) is a member of the Cap ‘n’ Collar family of basic leucine zipper transcription factors, which includes Nrf1, Nrf3, p45, Bach1, and Bach2 (58). Nrf2 appears to be a master regulator of the cellular response to electrophilic xenobiotics and oxidative stress. Promoter regions of protective genes, such as those encoding phase II detoxification enzymes, the antioxidant copper zinc superoxide dismutase, and enzyme subunits involved in glutathione synthesis contain the DNA binding site for Nrf2, which is known as the electrophile response element (EpRE) or antioxidant response element.

Nrf2 activity is predominantly regulated by subcellular distribution (59), as Nrf2 mRNA levels remain relatively constant in response to oxidative stressors (60). In unstressed cells, Nrf2 is retained in the cytoplasm by the actin-binding, cytoplasmic Kelch-like ECH-associated protein 1 (Keap1) and ubiquitinated by the Cul3 ligase leading to proteasome-mediated Nrf2 degradation (60). Disruption of the Keap1:Nrf2 interaction by phosphorylation of either protein or by oxidation of Keap1 sulfhydryl residues results in nuclear import of Nrf2 and transcriptional activation (61). A variety of kinases have been implicated in phosphorylation of Nrf2 including PKC, type I transmembrane endoplasmic reticulum-resident protein kinase (PERK), PI3K, and ERK/mitogen-activated protein kinase (MAPK) (62).

Oxidative stress has been implicated in the pathologic mechanisms of many neurodegenerative diseases, including AD, PD, and ALS. Whereas Nrf2 normally shows a nuclear distribution in the hippocampus, in AD and the Lewy body variant of AD, Nrf2 accumulated in the cytoplasm of hippocampal neurons (63). In contrast, localization of Nrf2 was nuclear in SNc neurons of PD. Further investigation of Nrf2 targeted gene expression in neurodegenerative diseases and the potential effects of premortem drug therapy will aid in understanding the differences in the response of neuronal Nrf2 in these patient populations. For example, deprenyl treatment induces Nrf2 nuclear localization and EpRE-mediated gene expression in SH-SY5Y cells (64), and Nrf2 responses can protect against 6-OHDA toxicity in vitro and in vivo (65, 66). These findings suggest that altered localization and activity of Nrf2 may contribute to neurodegeneration and may be an important target for therapeutic intervention.

Cytoplasmic Retention of Activated Transcriptional Regulators: Potential Mechanisms

Given that signaling proteins and transcriptional factors are thought to exhibit dynamic shuttling between nuclear and cytoplasmic compartments, with shifts in the predominant distribution triggered in response to stimuli, there are several potential mechanisms that might result in the perturbed distribution of transcription factors observed during neurodegeneration (Fig. 1). These include compartmentalized cytoplasmic or organelle-associated activation, altered levels of or access to compartment-specific anchoring proteins, impaired nuclear import, enhanced nuclear export, inefficient degradation/dephosphorylation, and/or sequestration in protein aggregates.

FIGURE 1.

(A) Schematic diagram of a generalized signal transduction pathway resulting in activation and nuclear import of a transcription factor and subsequent transcriptional activation. Under resting conditions (1), the transcription factor is bound to a cytoplasmic anchoring protein. Receptor ligation and/or physiologic redox signals result in phosphorylation of the transcription factor and/or phosphorylation or oxidation of the anchor and release of the transcription factor (2). The exposed nuclear localization signal (NLS) is recognized by importins, which escort the transcription factor into the nucleus (3). Importin β is also involved in retrograde transport of activated signaling proteins along neuritic processes (4). Within the nucleus, RanGTP binds importin β, which promotes disassembly of the import complex (5), freeing the transcription factor to bind DNA and form active transcription regulatory complexes. Transcription may be terminated by nuclear phosphatases and/or proteasomal degradation (6). Nuclear export is mediated by complexes consisting of cargo, an exportin, and RanGTP. Near the cytoplasmic face of the nuclear envelope, Ran GTPase-activating protein (RanGAP) maintains the nuclear-to-cytoplasmic RanGTP gradient by catalyzing GTP hydrolysis (7), promoting disassembly of the export complex. Importins are also recycled to the cytoplasm through a RanGTP-mediated mechanism. (B) Proposed mechanisms through which oxidative stress could impair nuclear transport of transcription factors, depriving the cell of adaptive, prosurvival transcriptional support and eliciting potentially detrimental cytosolic effects. Under conditions of oxidative stress such as those that occur during acute and chronic neurodegenerative insults, adaptive signaling pathways are activated by phosphorylation and/or redox mechanisms. However, post-translational modifications may impair cargo recognition by importins (1). Oxidative cross-linking of proteins, the presence of mutant neurodegenerative disease-associated proteins, and/or oxidative inactivation of phosphatase or proteasomal function could promote protein aggregation, resulting in further sequestration of signaling proteins (2). Sustained cytoplasmic signaling and/or organelle-targeted activation could further contribute to pathogenic mechanisms, such as calpain activation, altered mitochondrial function or autophagic stress from dysregulated autophagy (3). Alternatively, the nuclear pore complex itself could be the target of oxidative damage, with impairment of import and/or leakage of proteins from the nucleus (4). Intranuclear aggregation or other post-translational or oxidative/nitrosative modifications may further interfere with proper transcriptional responses (5), resulting in loss of stabilizing transcription complex-DNA interactions. Dissipation of the RanGTP gradient and/or impairment of importin recycling could also affect nuclear trafficking (6). Finally, lack of transcriptional support and/or oxidative or aggregative damage to axons and dendrites could result in further impairment in transmission of signals needed for maintenance and/or repair of neurites and synaptic structures (7). ROS, reactive oxygen species.

Regulation of Classic Nuclear Import and Export

Nuclear transport of proteins has been extensively reviewed (67-69). Briefly, nuclear transport is mediated by macromolecular nuclear pore complexes composed of nucleoporins with FG repeats that are important in pore selectivity. Whereas small proteins <40 kDa and metabolites can passively equilibrate across the nuclear envelope, karyopherin proteins (importins and exportins) mediate active transport into and out of the nucleus (Fig. 1A). Classical import involves recognition of cargo bearing an NLS by the adapter importin α. The importin alpha-cargo complex is escorted through the NPC by importin β via interactions with nucleoporin FG repeat motifs. Whereas importin β can directly bind some cargo proteins, recruitment of an array of adaptors and heterodimerization with other importin β-like receptors greatly expand the cargo repertoire.

The small G protein Ran plays a critical role in both import and export cycles. RanGTP binds to importin β on the nuclear side of the NPC to induce dissociation of the complex, releasing the cargo protein into the nucleus. RanGTP-bound importins are then returned to the cytoplasm and released upon GTP hydrolysis. Nuclear export is mediated by exportins such as Crm1, which chaperone cargo displaying NESs in complex with RanGTP through the NPC. Cargo release to the cytoplasm occurs upon stimulation of RanGTP hydrolysis by RanGAP. Ran itself is returned to the nucleus by nuclear transport factor 2/p10, where it is reloaded with GTP with the assistance of a Ran guanine nucleotide exchange factor (RanGEF) located inside the nuclear envelope and released to maintain a steep nuclear-to-cytoplasmic RanGTP gradient.

Importins and Axonal Transport

In addition to mediating nuclear import, there are several other functions of importin β family members that may be relevant to neurodegeneration. First, importin β heterodimers can function as cytoplasmic chaperones, shielding highly basic proteins and preventing protein aggregation (67). Secondly, importins have been shown to play an important role in injury-induced axonal transport along microtubules (69). Injury induces localized axonal translation of importin β, which shepherds cargo along microtubes in a dynein-dependent manner involving importin α and intermediate filaments (70). Excess NLS peptides introduced to the axon competes for importin-mediated trafficking and significantly delays regenerative transcriptional responses. These studies emphasize the critical role of well-regulated trafficking of transcriptional regulators, often over long distances, for successful adaptation of the neuron to physiologic (plasticity) and pathologic (injury or disease-associated) stimuli. Impairment of adaptive signal transmission, as observed in models of oxidative neuronal injury, could contribute to loss of functional viability in the setting of chronic neurodegenerative diseases.

Localized Activation and Alterations in Anchor Proteins

Many signaling proteins and transcription factors are maintained in the cytoplasm through binding interactions with cytoplasmic anchors, which are normally released upon initiation of signaling cascades that promote transcription. For example, the interaction of the transcription factor Nrf2 with its cytoplasmic anchor Keap is disrupted by phosphorylation or cysteine oxidation of Keap, allowing nuclear translocation of Nrf2 to promote EpRE-mediated transcription (62). Similarly, the interaction of NF-κB with its anchor IκB is released upon IκB phosphorylation or oxidation. As variations on this theme, phosphorylation of the translocating kinase can cause its release from cytoplasmic anchors, as observed for ERK signaling, and, conversely, dephosphorylation of Forkhead box transcription factor O unmasks its NLS to promote nuclear import (71). Any process that results in increased binding to cytoplasmic anchors, such as oxidative cross-linking, protein aggregation, or increased anchor protein expression, could serve to tip the nuclear to cytoplasmic balance of these dynamic signaling proteins.

Association of activated signaling proteins and transcriptions factors with cytoplasmic organelles may represent another mechanism for increased cytoplasmic localization. The growing field of mitochondrial kinase and transcription factor regulation has been the subject of a recent review (72). For example, CREB and its activating kinase PKA are localized to the mitochondrial matrix (73) and function to regulate both nuclear and mitochondrial transcription. Likewise, mitochondrial reactive oxygen species mediate delayed redox activation of mitochondrial ERK in a culture model of PD (74). Additional studies of organelle-targeted signaling and activation of other scaffolded signaling complexes are likely to reveal novel physiologic as well as pathologic functions for signaling proteins and transcription factors in cytoplasmic compartments.

On the other hand, interactions with chromatin or the nuclear scaffold delay rates of nuclear export, and a few transcriptional regulators are maintained in the nucleus through interaction with anchors such as dTCF/Pangolin for Armadillo and phosphatases for ERK (75, 76), which function to regulate duration of ERK activity and its localization. Loss, inaccessibility, or decreased affinity for these binding sites may facilitate a decreased nuclear to cytoplasmic balance and cytoplasmic accumulation of transcription factors.

Nuclear Transport Impairment With Oxidative Stress

The effects of oxidative stress on nuclear import mechanisms in non-neural cell culture systems have been examined in a few studies. Hydrogen peroxide induced a dose-dependent impairment of nuclear import in permeabilized aortic smooth muscle cells (77). Subsequent studies in HeLa cells (78) demonstrated that hydrogen peroxide treatment resulted in dissipation of the nucleocytosolic Ran-GTP gradient and redistribution of nucleoporin Nup153 immunofluorescence from the nuclear periphery to throughout the nucleus. Degradation of Nup153 and other components of the nuclear transport machinery was demonstrated and partially reversed by caspase inhibitors (78). Dissipation of the Ran gradient by hydrogen peroxide was also associated with impaired nuclear export of importin α (79), which would result in a nuclear import deficit due to impaired recycling of limiting amounts of importins. Taken together, these findings may relate to apparent derangements in nuclear localization of signaling proteins in models of neurodegeneration. Although the results of Ran-related dysfunction on the localization of specific transcription factors are difficult to predict, as import and export cycles are so closely tied, it is possible that specific components of the nuclear transport machinery are sensitive to redox regulation. Depending on the extent of redox stress, there may be selective perturbations in certain signaling and transcriptional pathways that impair the neuron's ability to adapt to stresses. More severe levels of damage may then lead to widespread impairment of cellular trafficking and signaling pathways.

Another potential mechanism of impaired nuclear import may involve modifications of NLS proteins or of importins, either of which could inhibit cargo recognition. As phosphorylation of transcription factors near the basic NLS can decrease binding to importin α, oxidative or nitrosative additions may have similar effects. As there are at least 10 human importins, subsets of nuclear cargo might be affected by redox modulation of specific importins. Even if the transcription factor is successfully imported, modifications such as nitroalkylation of NF-κB can lead to decreased DNA binding and decreased transcriptional activity (80). The loss of the anchoring property of DNA interactions could also enhance nuclear export of activated transcription factors, contributing to increased steady-state concentrations of cytoplasmic transcription factors.

It is possible that cytoplasmic distribution reflects passive leakage secondary to nuclear envelope damage. Structural alterations of the nuclear envelope have been described in tissues of AD patients (81). Gross disruption of the nuclear envelope, as that occurring during hyperacute cell death or in the late stages of apoptosis, could result in leakage of nuclear proteins due to passive equilibration of soluble proteins. However, there does seem to be selectivity in the spectrum of nuclear proteins observed accumulating in the cytoplasm during oxidative neuronal injury and neurodegeneration. Thus, it is likely that more selective disruptions in nucleocytoplasmic shuttling underlie the specific alterations observed in both early and advanced stages of chronic neurodegenerative diseases.

Protein Aggregation and Altered Degradation/Dephosphorylation Mechanisms

The characteristic pathologic hallmarks of most neurodegenerative diseases include some form of ubiquitinated protein aggregates. Interestingly, abnormal accumulation of importin α has been observed in neuronal Hirano bodies in AD, suggesting the possibility of functional impairment (82). Tangles and granulovacuolar inclusions have been implicated in sequestering and preventing nuclear localization of pSmad 3 in AD brains (40). Moreover, phosphorylated CREB kinases are observed in the halo region of Lewy bodies and in association with abnormal α-synuclein deposits (83). The transcription factor Elk-1 has also been observed in glial cytoplasmic inclusions of multiple system atrophy (84), another neurodegenerative disease characterized by α-synuclein aggregation. In experimental systems, overexpression of α-synuclein drives robust ERK phosphorylation, aberrant expression of cell cycle proteins (85), and sequestration of the transcription factor Elk-1 in cytoplasmic aggregates (84). The PD toxin MPP+ causes increased levels of both α-synuclein and phosphorylated ERK (86), suggesting common mechanisms between genetic and toxin models of PD. Finally, sequestration of CREB binding protein into nuclear aggregates has been shown to contribute to impaired CRE transactivation and neuronal cell death in models of polyglutamine repeat diseases (26). Thus, even when proteins traffic correctly to the nucleus, adaptive transcription may still be impaired.

In addition to nuclear import, there are other prominent mechanisms that limit the lifetime of activated signaling proteins in the cytoplasm. These mechanisms include dephosphorylation and degradation. Thus, it is possible that accumulation of activated signaling proteins in the cytoplasm reflects dysfunction of phosphatases, which can be sensitive to oxidative inactivation (87). Alternatively, the phosphorylated transcriptional regulators may be inaccessible to phosphatases, due perhaps to sequestration in aggregates or in autophagic vacuoles. Cytoplasmic anchoring systems for several transcription factors, such as Nrf2 and NF-κB, often function to also target the factor for proteasomal degradation. Impairment of proteasome function is a feature of many neurodegenerative diseases (88), and macroautophagy may be induced as a backup clearance mechanism or in the context of autophagic cell death (89, 90). It is possible that some mislocalized proteins are merely passive markers of an underlying neurobiologic dysfunction that mediates neurodegeneration. Studies in experimental models of several neurodegenerative diseases, however, suggest that the functional consequences of altered transcription factor localization contribute actively to neurodegeneration and neuronal cell death.

Transcription Factor Localization in Neurodegeneration: Therapeutic Implications

The potential functional consequences of nonclassic transcription factor trafficking and distribution are numerous. One potential explanation for observations in neurodegenerative diseases is that the protein is normally alternatively localized in neurons compared with non-neuronal cells, based on distinct neuronal import and export machinery or distinct regulatory mechanisms for a given transcription factor in neuronal subpopulations. For example, E2F1 interacts with the neuron-specific protein, necdin (44) and exhibits distinct localization in primary neurons in vitro (91). If the protein is truly altered in its localization in subsets of neurons undergoing neurodegeneration compared with control conditions, these changes may affect neuronal function and activity in different ways. By not localizing to the nucleus, transcription factor-mediated regulation of neuronal gene expression may be lost. As many of these signaling pathways regulate neuronal survival, altered localization may contribute to cell death commitment. Loss of nuclear localization of other transcriptional regulators may also serve to decrease expression of gene products needed for synaptic maintenance, leading to functional disturbances and neurite degeneration. Alternatively, the protein may gain aberrant functions from prolonged activation in the cytoplasmic compartment. Finally, divergent roles for signaling proteins and transcription factors may be unmasked under conditions of stress or disease. A protein may be diverted from its usual role in neurons to alternative roles as part of a stress response. In the short term, this adaptive response may be tolerated; however, during chronic stress of neurodegenerative disease, loss of normal function may ultimately compromise survival.

Nongenomic Functions of Transcription Factors and Their Upstream Regulators

In addition to cytoplasmic retention of transcription factors, upstream regulatory kinases often show increased expression or altered localization in neurodegenerative diseases. In particular, mitogen-activated protein kinases act upstream of ATF2, Elk, CREB, and Nrf2. Increased activation of JNK, ERK, p38 MAPK, JKK 1, MAPK/ERK kinase 1 (MEK 1), and MAPK kinase kinase 6 (MKK6) have been reported in AD hippocampus (16, 92, 93) and in the midbrains of patients with PD, Lewy body dementia, and progressive supranuclear palsy (83, 94). Ultrastructural and dual fluorescence studies indicate distribution of phospho-ERK in fibrillar bundles, mitochondria, and autophagosomes (95). Furthermore, mitochondrial oxidative stress may be involved in localized redox activation of ERK (74).

Potential cellular effects of cytoplasmic ERK activation include regulation of autophagy (96), calpain activation (97), mitochondrial alterations (98, 99), and, interestingly, inhibition of nuclear import in hydrogen peroxide-treated cells (77). The potential roles of autophagy, calpains, and abortive cell cycle reentry in neurodegeneration have been reviewed elsewhere (46, 90, 100, 101). Other mitochondrial signaling pathways include the PKA/CREB system, in which localization is mediated by mitochondrial A-kinase anchoring proteins. Interestingly, pCREB is found in human neuronal mitochondria, and culture studies indicate that CREB is involved in modulating mitochondrial transcription and prosurvival functions (73). Thus, cytoplasmic localization of certain transcription factors may reflect normal, rather than pathologic, functions yet to be elucidated.

Loss of Neuroprotective Transcriptional Responses

In addition to pathologic gain-of-function effects attributed to sustained activation of signaling proteins in the cytoplasm, impairment of nuclear import would be expected to block prosurvival transcriptional responses by activated signaling proteins. CREB is a downstream transcriptional effector of ERK (102), which exemplifies this mechanism (2). CREB traffics to the nucleus in an importin β- and RanGTP-dependent mechanism (103), where it is phosphorylated. In contrast, cytoplasmic accumulations of pCREB are detected in primary cultures and cell lines exposed to 6-OHDA, as well as in human PD substantia nigra neurons (2). 6-OHDA decreases the expression of CRE-containing prosurvival genes BDNF and Bcl-2 (2), but treatment of cells with cAMP after injury restores CRE transactivation and confers protection (2). Interestingly, kinase-induced CRE activation requires active nuclear import, but cAMP-induced CRE activity does not (104), suggesting a mechanism to bypass deficits in active import. Alternatively, it is possible that cAMP reverses the nuclear import impairment, allowing primed cytoplasmic pCREB to regulate transcription. Regardless of the precise mechanism, these data provide proof of concept information that strategies to bypass or reverse mechanisms leading to altered nucleocytoplasmic shuttling of transcription factors can alleviate signaling dysfunction in oxidatively stressed neurons, conferring neuroprotection.

Unresolved Issues and Therapeutic Implications

Subcellular distribution of signaling proteins is important for normal cellular function and efficient responses to cellular stress. Furthermore, it is emerging as an important regulatory mechanism that is perturbed during the progression of neurodegenerative disease. Proteins localized to compartments that are distinct from their better-known functions are heterogeneous in their activities, roles, and regulatory mechanisms. This heterogeneity may account for the recalcitrance of neurodegenerative diseases to single modality therapies. Most importantly, the reasons for alternative localization are still not understood for most of these proteins. To devise mechanistically driven treatments for these diseases, it will be necessary to understand the significance of altered localization of a given group of proteins, the mechanisms accounting for altered localization, and the best way to modulate and restore appropriate localization and function. For example, simple trophic factor replacement therapies may be destined to fail if potential neurodegenerative impairments in downstream signaling to the nucleus are not considered. Future researchers should not only address altered expression levels of proteins in neurodegenerative disease but also delineate subcellular compartments of activation and the cell types impacted to yield a more complete picture of signaling dysfunction in neurodegenerative diseases.

Acknowledgment

We thank Dr. Donald DeFranco of the University of Pittsburgh for helpful discussion.

Footnotes

  • The authors' research studies on signaling in neurodegeneration have been supported by funding from the National Institutes of Health (Grants NS040817, NS053777, NS41202, and AG026389) and by the Pittsburgh Foundation John F. and Nancy A. Emmerling Fund.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 75.
  76. 76.
  77. 77.
  78. 78.
  79. 79.
  80. 80.
  81. 81.
  82. 82.
  83. 83.
  84. 84.
  85. 85.
  86. 86.
  87. 87.
  88. 88.
  89. 89.
  90. 90.
  91. 91.
  92. 92.
  93. 93.
  94. 94.
  95. 95.
  96. 96.
  97. 97.
  98. 98.
  99. 99.
  100. 100.
  101. 101.
  102. 102.
  103. 103.
  104. 104.
View Abstract