c-Jun Expression, activation and function in neural cell death, inflammation and repair

Abstract

Up-regulation of c-Jun is a common event in the developing, adult as well as in injured nervous system that serves as a model of transcriptional control of brain function. Functional studies employing in vivo strategies using gene deletion, targeted expression of dominant negative isoforms and pharmacological inhibitors all suggest a three pronged role of c-Jun action, exercising control over neural cell death and degeneration, in gliosis and inflammation as well as in plasticity and repair. In vitro, structural and molecular studies reveal several non-overlapping activation cascades via N-terminal c-Jun phosphorylation at serine 63 and 73 (Ser63, Ser73), and threonine 91 and 93 (Thr91, Thr93) residues, the dephosphorylation at Thr239, the p300-mediated lysine acetylation of the near C-terminal region (Lys268, Lys271, Lys 273), as well as the Jun-independent activities of the Jun N-terminal family of serine/threonine kinases, that regulate the different and disparate cellular responses. A better understanding of these non-overlapping roles in vivo could considerably increase the potential of pharmacological agents to improve neurological outcome following trauma, neonatal encephalopathy and stroke, as well as in neurodegenerative disease.

Abbreviations used

AA

amino acid

AP-1

activator protein 1 transcription factor

ATF

activating transcription factor

ERK

extracellular signal regulated kinase

FBW7

F-box and WD repeat domain containing 7 ubiquitin ligase

GSK3

glycogen synthase kinase 3

JIP

JNK interacting protein

JNK

N-terminal Jun Kinase

jun^ΔN

nes::cre mediated removal of c-Jun in brain neuroepithelium-derived cells

MCAO

medial cerebral artery occlusion

Ser

serine

Thr

threonine

v-Jun

viral Jun protein

In addition to the intricate structure established during embryogenesis and postnatal development, the nervous system also exhibits an elaborate response aimed at preventing excessive damage and enabling neural repair. This involves new synthesis of many different proteins and mRNAs, and by extension the activation of transcription factors to orchestrate the required cellular and molecular changes. The expression and activation of c-Jun, a prominent member of the activator protein 1 (AP-1) family of transcription factors, is a particularly informative case in point. Neural removal of both copies of the jun gene (jun^ΔN) will strongly affect the post-traumatic neuronal cell death, inflammation and neural repair (Raivich et al. 2004), underscoring the role of c-Jun in orchestrating the cellular response to injury. Our aim is to provide an overview on its molecular structure, chemical modification and the signaling pathways subserving the c-Jun function in the damaged and recovering nervous system.

Structure and molecular biology

The AP-1 family of transcription factors consists of several groups of basic region leucine zipper proteins: the Jun, the Fos and the activator transcription factor (ATF) subfamilies. AP-1 proteins have to dimerise to form functional sequence-specific transcription factors that bind to their cognate DNA target sites. Overall, there is a great number of different homo- and heterodimeric combinations with different regulatory properties determined by subunit composition that can form and differentially modulate gene transcription (Angel and Karin 1991; Whitmarsh and Davis 2000). While Fos proteins (c-fos, FosB, Fra-1, and Fra-2) do not form homodimers but heterodimerize with members of the Jun family, the Jun family proteins (c-Jun, JunB, and JunD) can also form homodimers (Angel and Karin 1991; Jochum et al. 2001). In addition, Jun proteins can also heterodimerize with other transcription factors, including members of the ATF family (Hai and Hartman 2001) and other basic zipper containing transcription factors such as CBP, MyoD, NFat or c-rel (Herdegen and Leah 1998).

On the structural level, the c-Jun molecule is composed of the N-terminal region containing the c-Jun N-terminal kinases (JNK) binding and phosphorylation sites and the transactivation domain (aa1–190), a hinge region (aa191–256), a basic, positively charged domain of 18 amino acids (aa257–276), the leucine zipper (aa280–308) and the final, C-terminal aminoacid sequence shown in Fig. 1. Changes in the activation of c-Jun-mediated transcription apparatus depend on the chemical modification at three different sites: the now classical phosphorylation of the N-terminal region serine 63 and 73 (Ser63&73) and threonine 91 and 93 (Thr91&93) residues (Smeal et al. 1991; Morton et al. 2003), the dephosphorylation of Thr239 (Morton et al. 2003) and the acetylation of lysine residues in the basic, near C-terminal region aa257–276 (Vries et al. 2001). To a large extent, this chemical modification is under the direct or indirect control of different components of the MAPK family of serine and threonine phosphorylation enzymes that include the JNK 1, 2 and 3, extracellular signal regulated kinases (ERK) 1&2 and the p38 MAPK isoforms (Whitmarsh and Davis 2000).

Figure 1.  Overview of diverging c-Jun activation pathways, via JNK, GSK3 and p300. JNKs activate c-Jun through N-terminal phosphorylation, but have a multitude of additional targets, and can also exert indirect effects by activating the transcription factor complex (TFC) controlling jun mRNA synthesis. The Phosphorylation of Thr91&93 is facilitated by the presence of phosphorylated Thr95. GSK3 mediated phosphorylation of Thr239, assisted by presence of phosphorylated Ser243 attracts the FBW7 ubiquitin ligase (FBW7L) and targets the phosphorylated protein for ubiquitination and degradation; N-terminal phosphorylation also appears to enhance this Jun degrading activity. Removal of negative charges at the 239 and 243 positions stabilizes c-jun; the identity (?) of the Thr239-P dephosphorylase is currently unresolved. Finally, p300-mediated acetylation of the C-terminal lysine residues will also result in JNK-independent, c-Jun activation. On the regulatory level, JNK activity is switched on by its phosphorylation via activated MKK4 or MKK7; GSK3 is inactivated by phosphorylation through PI3-Kinase and ERK1&2, and p300 is activated by phosporylation via ERK1&2. Sharp arrows show activating, and blocked arrows the inhibitory effects, stippled arrows – facilitating effects. The N-terminal and C-terminal ends of c-Jun are marked by N and C, respectively. Ac, acetylating; dP, dephosphorylating; and P, phosphorylating effect; DEP, dependent; IND, independent; JBD, JNK binding domain; TAD, transactivation domain; BR, basic region; LZ, leucine zipper.

In the case of the JNK1–3, both Ser73 and Ser63 phosphorylation are critical phosphoacceptor sites of JNK mediated effects. Replacement of JNK phosphorylation motifs with those for protein kinase A will abolish the c-Jun-mediated JNK effects (Smeal et al. 1991); replacement of Ser63&73 with alanine residues (junAA) will prevent JNK-mediated (Yang et al. 1997) excitotoxic cell death following exposure to kainate (Behrens et al. 1999). On the whole, phosphorylation of Ser73 seems to be more critical. Phosphorylated Ser63 usually occurs to a somewhat lower extent in neuronal as well as in non-neuronal cells (Smeal et al. 1991;Dragunow et al. 2000; Ward and Hagg, 2000, Thakur et al. 2007; see however Pearson et al. 2006) and has been suggested to be insufficient for stimulating the activating function of c-Jun (Smeal et al. 1991; Pulverer et al. 1991). Interestingly, selective replacement of just Ser73 with alanine will reduce, while that of Ser63 enhance the c-Jun mediated neurite outgrowth in the PC12 cell model of neuronal differentiation, pointing to the importance of Ser73 (Dragunow et al. 2000).

In vitro, activated JNKs will also lead to a more modest phosphorylation of Thr91 and 93. The latter two sites depend on the facilitating activity from adjacent, negatively-charged Thr95 for their maximal response in enhancing cell death in response to DNA damage (Vinciguerra et al. 2007). Here, the phosphorylation of the Thr95 residue, mediated by activation of ataxia teleangiectasia mutated protein kinase and her downstream checkpoint kinases in response to DNA damage and other forms of cell injury, and followed by strong phosphorylation of Thr91&93 may present a mechanism that translates persistent ongoing stress into pronounced apoptotic response. In the absence of JNK1&2, activation of ERK will also lead to a moderate phosphorylation of all four aminoacids (Morton et al. 2003), but the effects in vivo may be sub-threshold.

According to Wei et al. (2005), c-Jun molecules carrying phosphorylated Thr239 and Ser243 function as docking targets for the F-box&WD domain repeated 7 (FBW7) ubiquitin ligase to allow the subsequent degradation of c-Jun. Ser243 phosphorylation is a priming event for glycogen synthase kinase-3 (GSK3)-mediated phosphorylation of the Thr239 residue (Morton et al. 2003). This phosphorylation is blocked by the exchange of Ser243 for phenylalanine in the transforming, viral form of the protein, causing the stabilization of viral Jun protein (v-Jun) (Boyle et al. 1991; Wei et al. 2005). Similar stabilization of c-Jun is observed following phosphorylation-induced inhibition of GSK3 by ERK and phospho-inositide-3-kinase signaling cascades or by the low molecular weight GSK3 inhibitors kenpaullone and lithium (Morton et al. 2003). Both phosphates, at Thr239 and Ser243, appear required for the FBW7 ubiquitin ligase engagement, so loss of both negative charges in v-Jun, or just one at Thr239 following GSK3 inhibition, may be sufficient to produce the stabilizing effect. Although the identity of the Thr239 phosphatase causing the terminal c-Jun stabilization is currently still unknown, it is possible that it is also activated by the ERK and phospho-inositide-3-kinase signaling.

Significantly, complete abrogation of N-terminal phosphorylation of c-Jun at all four major phosphorylation sites, i.e. Ser63&73 and Th391&93, will also result in a block of FBW7-mediated degradation independently of the Thr239 and Ser243 pair, that forms an alternative degradation pathway for prefrential removal of the N-terminal activated c-Jun (Nateri et al. 2004). It is possible that differently spliced FBW7 isoforms diverge in their preference for phosphorylated Ser63-Th93 or Thr239&Ser243, as a way to explain the different substrate targets observed by Nateri et al. (2004) and Wei et al. (2005). However, v-Jun also carries a large aa32–58 deletion in the JNK docking domain. This precludes JNK binding and N-terminal phosphorylation of v-Jun (Bos et al. 1990; May et al. 1998), and could be the reason why v-Jun only depends on the mutation of the Thr239&Ser243 pair in order to escape the FBW7-mediated degradation. It should be mentioned that although v-Jun is normally a transforming oncoprotein, it can sometimes enhance cell death, for example in Schwann cells exposed to transforming growth factor beta 1 (Parkinson et al. 2001).

Finally, charge modification in the C-terminal near basic region can strongly affect the efficiency of gene transcription, at least in vitro. Epidermal growth factor-induced expression of keratin 16 in immortalized epidermal keratinocytes depends on the acetylation of lysine residues in the basic domain of the DNA binding, C-terminal region by the p300 acetylase, which forms a DNA binding complex together with acetylated c-Jun (Vries et al. 2001; Wang et al. 2006). Conservation of the positive charge but removing the acetylation acceptor sites by replacement of the three lysines Lys268, 271 and 273 with arginine residues, completely inhibits the epidermal growth factor/Jun-mediated effects (Wang et al. 2006). Individual amino-acid mutation analysis reveals Lys271 as the primary acetyl acceptor residue, as well as the key component for mediating repressive effects of E1A (Vries et al. 2001). Interestingly, MAPK (ERK) kinase/ERK signaling also plays a pivotal role in nuclear recruitment of p300 and its association with c-Jun (Wang et al. 2006), providing two different main action points for ERK mediated effects – the dephosphorylation of Thr239, and the acetylation of the lysine residues in the Jun DNA binding region.

c-Jun and neural cell death

Strong expression of the c-Jun gene and protein is known to precede or coincide with periods of intense cell death during embryonic development (Sun et al. 2005), following trauma (Herdegen et al. 1991; Jenkins and Hunt 1991; Raivich et al. 2004), brain ischemia (Kindy et al. 1991; Wessel et al. 1991), and seizures (Morgan and Curran 1988; Gall et al. 1990; Gass et al. 1993). Similar induction is also present in human neurodegenerative diseases such as Alzheimer’s dementia (Pearson et al. 2006; Thakur et al. 2007), amyotrophic lateral sclerosis (Migheli et al. 1997) and following exposure to neurotoxic chemicals such as MPTP, causing degeneration of dopaminergic neurons in the substantia nigra, a model of Parkinson’s disease (Oo et al. 1999; Saporito et al. 2000). In the latter, inhibition of JNK or upstream signaling will reduce MPTP-mediated dopaminergic cell loss (Saporito et al. 1999; Xia et al. 2001; Wang et al. 2004), suggesting possible use in Parkinson’s disease (Silva et al. 2005; Borsello and Forloni 2007).

In vitro, withdrawal of trophic support through removal of neurotrophins or of high, depolarizing concentrations of potassium, will cause rapid neuronal cell death which could be used to explore cell death pathways, as a model for in vivo conditions. Thus, withdrawal of NGF leads to a rapid accumulation of N-terminal phosphorylated c-Jun protein, and Jun-dependent up-regulation of the bcl2-proapoptotic family member bim (Whitfield et al. 2001) and ATF3 (Mei et al. 2008). Over-expression of c-Jun induces cell death in PC12 cells; dominant-negative Jun (DN-Jun) lacking aa25–181 (Ham et al. 1995) as well as complete jun gene excision (Palmada et al. 2002) prevents this form of neuronal cell death. Interestingly, coexpression of the dominant negative-Jun with the JNK-blocking JNK binding domain of JNK interacting protein 1 (JIP1) shows involvement of the same pathway in the JNK and Jun-mediated death (Eilers et al. 2001; Harding et al. 2001). Moreover, JNK activation can also lead to increased phosphorylation of the c-Jun containing transcription factor complex providing a positive feed-back loop on jun mRNA synthesis (Nateri et al. 2005; Parkinson et al. 2008).

In the same vein, CEP-1347 inhibition of mixed lineage kinases, the upstream inducers of p38 and JNK, also promotes the survival of embryonic sensory, sympathetic and motor neurons, as well as the PC12 cells following withdrawal of trophic support (Borasio et al. 1998; Maroney et al. 1999). However, inhibition of all three JNK isoforms frequently has a greater protective effect, compared with Ser63&73 phosphoacceptor site replacement or small interfering RNA (siRNA)jun inhibition. This enhanced protection is particularly pronounced with nuclear translocating, but not cytoplasmic translocating forms of dominant negative JNK and upstream stress activated protein kinase 1 phosphorylation enzymes (Björkblom et al. 2008). Moreover, this points to a critical role for nuclear mixed lineage kinase/JNK substrates, for example the Nup214 subunit of the nuclear pore complex, in mediating cell death following the withdrawal of trophic support (Besirli et al. 2005).

Similar protection through inhibition of JNK is also observed in excitotoxic, hypoxic/ischemic and inflammatory, neuronal and glial cell models. Exposure to p75 neurotrophin receptor ligands, inflammatory stimuli or reactive oxygen radicals, causes a strong up-regulation in JNKs, with downstream increase in p53 and cell death in astrocytes, adult oligodendrocytes and particularly, in oligodendroglial precursors (Casaccia-Bonnefil et al. 1996; Ladiwala et al. 1998; 1999; Vollgraf et al. 1999). These effects are inhibited by specific, small molecular weight JNK antagonists (Jurewicz et al. 2003, 2006; Pirianov et al. 2006; Fernandes et al. 2007). Application of DJNKI1, a cell permeable, JNK binding and inhibiting fragment of the JIP1, will also block phosphorylation of Jun, up-regulation of c-fos as well as cell death in cultured cortical neurons in response to the stimulation with NMDA (Borsello et al. 2003). Importantly, similar protection following excitotoxic kainate stimuli is also observed in junAA neurons lacking the Ser63&73 phosphoacceptor sites (Behrens et al. 1999) normally phosphorylated by the JNKs, but in this model particularly by JNK3 (Yang et al. 1997).

How far are these effects representative of the situation in vivo? Genetic deletion of JNK3 isoform protects against hypoxic ischemic neonatal as well as adult forms of hypoxic ischemic brain damage, blocking the post-ischemic synthesis of Fas, induction of the proapoptotic bcl2-members bim and PUMA, and mitochondrial release of cytochrome c (Kuan et al. 2003; Pirianov et al. 2007). Preliminary data also show an almost complete block of neonatal hypoxic ischemic insult in animals with neural deletion of the whole jun gene (jun^ΔN) using the loxP (floxed) tagging of the jun gene (jun^f/f) and the nestin-promoter driven cre (nes::cre) recombinase (Raivich and Behrens, unpublished). A similar match is also observed with excitotoxic brain damage, with demonstrated protection following application of kainate in JNK3 null (Yang et al. 1997) and in junAA mice (Behrens et al. 1999).

However, these effects diverge in adult axotomy and medial cerebral artery occlusion (MCAO), the trauma and stroke models, respectively. In MCAO, there is a highly effective inhibition of brain tissue loss following application of small JNK antagonists or DJNKI1 but no obvious effect in the junAA mice with phosphoacceptor site substitution (Brecht et al. 2005) or with neural deletion of the whole c-Jun gene (Vogel et al. 2007). Likewise, deletion of JNK1&2 or their upstream activating enzymes will have profound effects on the formation of neural tube, developing brain morphology as well as embryonic neuronal cell death (Kuan et al. 1999; Sabapathy et al. 1999; Wang et al. 2007). In contrast, the effects of junAA substitution or the neural jun deletion are minor. There is no obvious morphological neural tube abnormality in the junAA or jun^ΔN mutants, a 5% numerical increase in adult, jun^ΔN facial motoneurons, and a possibly transient, 40% increase in that of junAA neonatal sympathetic neurons (Roffler-Tarlov et al. 1996; Raivich et al. 2004; Besirli et al. 2005). Interestingly, deletion of JNK3 also protects motoneurons and dorsal root ganglia sensory neurons against neonatal axotomy-induced death, but through mechanisms that are independent of c-Jun phosphorylation (Keramaris et al. 2005).

Finally, in adult facial axotomy, neural deletion of the jun gene in the jun^ΔN mutants abrogates the post-traumatic neuronal cell death (Raivich et al. 2004), in the absence of obvious effects in the junAA mice (Brecht et al. 2005). Moreover, these effects appear part of a larger, neuronal regeneration gene program, which is strongly altered in the jun^ΔN, but unaffected in the junAA mice. Thus, there are apparently three different sets of signaling effects: (A) the JNK independent effects of c-Jun in adult peripheral axotomy, (B) the non-jun effects of JNK in brain development & MCAO, and (C) the classical JNK→c-Jun signaling, that seems to be restricted to hypoxic-ischemic and excitotoxic sets of effects, and appears to underscore their functional relatedness.

Glial and inflammatory changes

Appearance of glial c-Jun is a distinct feature in injury, neuroinflammatory and neurodegenerative disease. In the affected central nervous system, glial c-Jun immunoreactivity is the primarily astrocytic: it is present in cerebral ischemia (Kato et al. 1995), in several forms of neurodegenerative disease including Alzheimer’ dementia (Anderson et al. 1994) and amyotrophic lateral sclerosis (Migheli et al. 1997), in mechanic allodynia models of neuropathic pain (Zhuang et al. 2006), and in toxin-induced degeneration of cholinergic (Rossner et al. 1997) and dopaminergic (Nakagawa and Schwartz 2004) neurite terminals. In the peripheral nervous system, c-Jun is also strongly expressed in embryonic and adult denervated Schwann cells (Vaudano et al. 1996), regulating cell survival, myelination and de the denervation and demyelination (Parkinson et al. 2001, 2008).

In vitro, JNK signaling plays a prominent role in regulating astrocytic, microglial and oligodendroglial response to a variety of hypoxic/ischemic, excitatory, and inflammation-associated stimuli, regulating cellular swelling, synthesis and release of neurotrophins and cytokines, reactive oxygen radicals as well as matrix degrading enzymes (e.g. Hidding et al. 2002; Waetzig et al. 2005; Jayakumar et al. 2006; Kim et al. 2008; Tanaka et al. 2008). These effects were summarized in several recent reviews on the subject of MAPK, their inhibitors and their inherent, in vivo therapeutic potential (Ferrer et al. 2005; Waetzig and Herdegen 2005; Ji et al. 2007). Regrettably, most reports have only focused on the JNKs and other signaling enzymes. However, a smaller selection of studies do show that at least some (but not all) of the reported effects are because of the expression and activation of the glial c-Jun AP-1 transcription factor (Hidding et al. 2002; Zhuang et al. 2006; Kim et al. 2008).

Developing, myelinating and demyelinating Schwann cells are particularly sensitive to the effects of c-Jun, but, unlike most of the cases listed above, via a JNK-independent mechanism. Enforced expression of c-Jun will inhibit myelination in cocultures with sensory neurons by antagonizing Krox20. The fact that similar inhibition is observed in cells transfected with the JNK-insensitive Ala63&73 (junAA) as well with the JNK-mimicking asparagine (Asp) 63&73 (junAsp2) mutant, with negative charges in position 63 and 73 position argues against direct involvement of JNK (Parkinson et al. 2008). In the reverse case, deletion of the whole c-Jun gene using cre recombinase driven by P0 promoter (P0::cre) caused a very pronounced delay in demyelination following explantation in vitro. Similar delay in myelin degradation was also observed in denervated Schwann cells left in situ in the first week following sciatic nerve cut (Parkinson et al. 2008), underscoring the critical in vivo role of c-Jun in switching off the myelinating cell phenotype in vivo (Salzer 2008).

Neural deletion of jun also interferes with the glial and inflammatory response, with a much weaker microglial activation and T-cell recruitment following peripheral, facial motor nerve cut in jun^ΔN mutant mice. This decrease in inflammation-associated effects in the affected brainstem cranial motor nucleus was already observed at day 4, before the normal onset of post-traumatic neuronal cell death at day 14 (Raivich et al. 2004). As only neurons, but not astrocytes, microglia or T-cells express c-Jun in this model, the activation and recruitment effects appear neuronal in origin. Similar decrease in microglial, but also in astrocyte activation in the absence of neural c-Jun, was also observed in the super-oxide dismutase G93A model of familial amyotrophic lateral sclerosis (Acosta-Saltos et al. 2007). Deletion of jun in myelinating Schwann cells using P0-promoter driven cre recombinase (P0::cre) causes a defect in leukocyte recruitment, in this case of macrophages, into the cut peripheral nerve (Arthur-Farraj et al. 2007), underscoring c-Jun function in injured neurons and denervated Schwann cells in switching on local inflammation.

Axonal outgrowth, regeneration and degeneration

As with cell death, inflammation and glial response, there is disparity in the effects of c-Jun and upstream JNK signaling on the outgrowth, maintenance and regeneration of axons. Unlike c-Jun itself, JNK, its upstream kinases, the JNK scaffolding protein JIP1 and the transcription factors ATF2 and ATF3 are axonally transported, providing a JNK-mediated retrograde communication pathway between the nerve terminals and the c-Jun containing neuronal cell body (Lindwall and Kanje 2005). Deletion of JNK1 leads to a disruption in the formation of anterior commissure (Chang et al. 2003). There are only minor effects in mutants lacking JNK2 or JNK3 alone, or minor additional effects with combinations of JNK1&JNK3 or JNK2&JNK3 (Gelderblom et al. 2004). Deletion of JNK scaffolding protein JSAP-1 will also cause an axon guidance defect of the telencephalic commissures (Kelkar et al. 2003), with a partial rescue following transgenic expression of JIP1, a functionally related scaffolding protein, in embryonic glial cells (Ha et al. 2005).

Effects on axonal outgrowth are also observed in the adult, at least in Drosophila. Neuronal over-expression of constitutively active Hemipterous, the drosophila JNK phoshorylating & activating kinase will strongly enhance central axonal regeneration. The reverse over-expression of dominant negative form of Basket, the single Drosophila JNK isoform, has a moderate, outgrowth-inhibiting effect (Ayaz et al. 2008).

N-terminal Jun Kinases are also active in inflammatory axonopathy and degenerative processes. Near exposure of cultured neurons to highly activated microglia will also cause a disruption of anterograde transport, because of nitric oxide and tumour necrosis factor-mediated activation of axonal JNK and the unloading of transported cargoes from their tubulin tracks (Stagi et al. 2005, 2006). These effects are due to the dissociation of the kinesin 5FB, JNK and beta tubulin-III complex. They can be blocked by specific JNK inhibitors (Stagi et al. 2006) and underscore the participation of JNK mediated signals in inflammation-associated axonal degeneration.

Removal of the Ser63&73 c-Jun phosphoacceptor sites in junAA or neural deletion of c-Jun does not affect the regular development of central white matter tracts (Behrens et al. 1999; Raivich et al. 2004), emphasizing the non-Jun targets of embryonic JNK action. However, absence of neural c-Jun because of nes::cre driven deletion of floxed jun gene interferes with the cell body response to axonal injury in the adult nervous system. Axotomized motoneurons lacking c-Jun gradually become atrophic but fail to die and show reduced perineuronal sprouting, processes readily observed in the wild type cells. Compared with controls that do not express cre, the neural Jun-deficient mice display a 4-fold decrease in the speed of regeneration, the reinnervation of target muscle and functional recovery. Expression of CD44, galanin, and alpha7beta1 integrin, molecules known to be involved in regeneration and carry AP-1 responsive elements in their promoter regions are greatly impaired, suggesting a mechanism for c-Jun-mediated axonal growth (Raivich et al. 2004). Interestingly, preliminary studies using junAA mice show only moderate effects of JNK deletion or Ser63&73 phosphoacceptor removal (Raivich, Herdegen, and Behrens, in preparation), again pointing to a non-JNK transduced action of c-Jun.

In conclusion, the current data reinforce the general notion of just a partial overlap between c-Jun and JNK-mediated mechanisms in vivo, underscoring the considerable, Jun-independent effects of JNKs injury-activated signaling kinases and JNK-independent effects of c-Jun. Studies in vitro suggest additional turn-on and -off switches located in the central and C-terminal Jun domains, for example through dephosphorylation of Thr239, acetylation of lysines 268, 271 and 273, as well as the FBW7 ligase-mediated degradation. Even this may represent just a tip of the iceberg in terms of signaling involving this ubiquitous transcription factor up-regulated in the injured and recovering nervous system. Identification of their specific functions in vivo, and pharmacological modulation of their activity could help improve neurological outcome following trauma, neonatal encephalopathy and stroke, or in neurodegenerative and neuroinflammatory disease.

Acknowledgements

I thank Dr Axel Behrens from Cancer Research, UK, for his many helpful comments and critical suggestions during the preparation of the manuscript, Alejandro Acosta for the help with figure 1 and Crystal Ruff for reading the manuscript.