NGF Upregulates the Plasminogen Activation Inhibitor-1 in Neurons via the Calcineurin/NFAT Pathway and the Down Syndrome-Related Proteins DYRK1A and RCAN1 Attenuate This Effect

Background Plasminogen activator inhibitor 1 (PAI-1) is a key regulator of the plasminogen activation system. Although several lines of evidence support a significant role of PAI-1 in the brain, the regulation of its expression in neurons is poorly understood. In the present study we tested the hypothesis that NGF induces the upregulation of PAI-1 via the calcineurin/nuclear factor of activated T cells (NFAT) pathway and analysed whether the overexpression of the Down syndrome-related proteins DYRK1A and RCAN1 modulated the effect of NGF on PAI-1 expression. Results NGF upregulated PAI-1 mRNA levels in primary mouse hippocampal neurons cultured for 3 days in vitro and in the rat pheochromocytoma cell line PC12. Reporter gene assays revealed that NGF activated the calcineurin/NFAT pathway in PC12 cells. Induction of PAI-1 by NGF was sensitive to the calcineurin inhibitor FK506 and the specific inhibition of NFAT activation by the cell permeable VIVIT peptide. Activation of calcineurin/NFAT signalling through other stimuli resulted in a much weaker induction of PAI-1 expression, suggesting that other NGF-induced pathways are involved in PAI-1 upregulation. Overexpression of either DYRK1A or RCAN1 negatively regulated NFAT-dependent transcriptional activity and reduced the upregulation of PAI-1 levels by NGF. Conclusion The present results show that the calcineurin/NFAT pathway mediates the upregulation of PAI-1 by NGF. The negative effect of DYRK1A and RCAN1 overexpression on NGF signal transduction in neural cells may contribute to the altered neurodevelopment and brain function in Down syndrome.


Introduction
The serine protease inhibitor plasminogen activator inhibitor-1 (PAI-1) is a component of the plasminogen activation system. PAI-1 forms stable complexes with both tissue plasminogen activator (tPA) and urokinase plasminogen activator (uPA) and thus prevents the activation of plasminogen to plasmin [1].
PAI-1 is expressed in a variety of tissues including the brain [2]. As a regulator of plasminogen activation, PAI-1 is involved in the degradation of the extracellular matrix, which is important for synaptic remodelling, as well as in the regulation of the plasminmediated cleavage of proBDNF to BDNF [3]. As a tPA inhibitor, PAI-1 protects neurons from the tPA potentiation of the NMDAinduced calcium influx and necrosis [4]. PAI-1 has also a neurotrophic/neuroprotective role in PC12 cells, which has been suggested to be independent of the plasminogen activation system [5]. During the last decade PAI-1 has also attracted increasing attention for its potential function in Alzheimer disease both as a neurotrophic factor and as an inhibitor of plasmin's a-secretase cleavage activity [5,6].
Dual-specificity tyrosine phosphorylation-regulated kinase 1A (DYRK1A) is a protein kinase member of the DYRK family that phosphorylates serine and threonine residues in its substrates and autophosphorylates a tyrosine residue in its activation loop [7]. The human DYRK1A gene is located on chromosome 21 [8] and its overexpression is considered to be related to Down syndrome phenotypes. Both murine Dyrk1A and human DYRK1A genes are expressed in several regions of the brain [9,10,11], where they play a role in neuronal development [12] and adult brain function [13].
Transgenic mice that carry an extra copy of the Dyrk1A gene exhibit neurodevelopmental delay, motor abnormalities and cognitive deficits that resemble those described in Down syndrome individuals [14,15,16].
Regulator of calcineurin 1 (RCAN1), which was formerly known as DSCR1 or calcipressin 1, binds to and inhibits the protein phosphatase calcineurin and its signalling pathways [17]. The RCAN1 gene is also located on the human chromosome 21, is highly expressed in the human brain [18] and is considered to be associated with Down syndrome traits. RCAN1 overexpression in transgenic mice affects the visuo-spatial learning and memory tasks and causes structural brain abnormalities in those areas affected in DS [19,20].
NFAT is a family of transcription factors consisting of the five members NFAT1-5. Except for NFAT5 all of them are calciumregulated [21]. In the basal state the regulatory domain of NFAT is heavily phosphorylated and thus the transcription factor is inactive and localized in the cytoplasm. Rise in intracellular Ca2+ results in the activation of the calmodulin-dependent phosphatase calcineurin, which in turn dephosphorylates NFAT. After dephosphorylation, a nuclear localization sequence is exposed and NFAT translocates to the nucleus where it binds to target DNA sequences either as homodimers, heterodimers or through interaction with other transcription factors [22,23,24]. Benedito et al. [25] showed that NFAT3 plays a critical role in mediating survival of granule neurons of the developing cerebellum and Graef et al. [26] showed that NFATc2/c3/c4 null mice have defects in axon outgrowth. DYRK1A and RCAN1 are negative regulators of NFAT activity. RCAN1 affects NFAT localization by inhibiting the enzymatic activity of calcineurin [27] and DYRK1A phosphorylates NFAT and promotes its nuclear export by priming the subsequent phosphorylation by GSK3 [28,29]. Furthermore, DYRK1A was shown to interact with RCAN1, extending its half life and enhancing its binding on calcineurin [30]. Taken together, DYRK1A and RCAN1 can act synergistically to block NFATdependent transcription, suggesting that the reduced NFAT activity resulting from the 1.5-fold overexpression of DYRK1A and RCAN1 could be the basis of many features of Down syndrome [28,31].
In the PC12 cell line, PAI-1 is induced by NGF [32,33]. Furthermore, NGF causes Ca2+ mobilization and consequently NFAT activation in several cell lines [34,35]. Given that Down syndrome individuals have reduced PAI-1 in blood compared with controls [36], the aim of the present study is: a) to test the hypothesis that the calcineurin/NFAT pathway is involved in the NGF-induced upregulation of PAI-1 in neuronal cells and b) to analyse whether the overexpression of the Down syndrome-related proteins DYRK1A and RCAN1 can regulate the NGF effect on PAI-1 by modulating NFAT deactivation. A, PC12 cells were treated for 2.5 h with 50 ng/ml NGF or were not treated (C). PAI-1 mRNA was quantified by qRT-PCR. B, PC12 cells were treated with 50 ng/ml NGF for the times indicated. PAI-1 levels in total cell lysates were determined by quantitative evaluation of immunoblots as shown in the lower panel. The upper one of the double band represents the active form of PAI-1, while the lower band represents the cleaved form. Intensities of both bands were combined for quantification and normalized to the TUJ1 signal. C, PC12 cells were treated with 50 ng/ml NGF for 3 h in the presence or absence of 300 nM K252A. The inhibitor was added 30 min before NGF. PAI-1 protein levels in total cell lysates were determined by quantitative evaluation of immunoblots. D, Primary hippocampal cultures from mouse embryos at DIV3 were treated for 2 h with 100 ng/ml NGF before PAI-1 mRNA levels were quantified by qRT-PCR. The diagrams present means 6 SEM from 3 independent experiments. C, Control; *, p,0.05; **, p,0.01. doi:10.1371/journal.pone.0067470.g001

NGF Induces PAI-1 Expression
PC12-NFAT-Luc cells expressing luciferase under the control of an NFAT-driven promoter [37] were used for the current study and will be referred to as PC12 throughout. NGF was tested for its ability to induce PAI-1 in this PC12 subclone and resulted in a 12fold induction of PAI-1 mRNA levels (Fig. 1A). In a time-course experiment (Fig. 1B) the cellular (non-secreted) PAI-1 protein levels increased after 2 h of NGF treatment and reached the maximum level after 3-4 h. The Trk inhibitor K252a abolished the effect of NGF on PAI-1 protein levels (Fig 1C), suggesting that NGF exerts its effect via the TrkA receptor. Next we asked whether NGF induces PAI-1 expression also in primary neurons. Cortical and hippocampal neuronal cultures from mouse embryos at embryonic day 17 (E17) were prepared. On day in vitro 3 (DIV3) treatment with NGF for 2 h increased the PAI-1 mRNA levels about 2.5-fold in hippocampal cultures ( Fig. 1D) but had no effect on cortical cultures (data not shown). At later stages of the culture (DIV14) NGF had no effect on PAI-1 expression in either hippocampal or cortical cultures (data not shown).

NGF Upregulates PAI-1 via the Calcineurin/NFAT Pathway
To examine whether NGF can activate NFAT in our cell model, PC12-NFAT-Luc cells were treated with NGF in the presence or absence of FK506, a specific inhibitor of calcineurin [38]. NGF increased the NFAT-dependent promoter activity about 11-fold, while FK506 reduced this effect by 75% ( Fig. 2A). Next we tested whether calcineurin mediates the NGF-derived upregulation of PAI-1. As shown in figure 2B, NGF upregulated the cellular PAI-1 protein about 3-fold and FK506 reduced by 50% the PAI-1 levels in NGF-treated cells. To further prove the involvement of the NFAT, we used the cell permeable VIVIT peptide, which selectively interferes with the calcineurin/NFAT interaction and thereby inhibits NFAT activation without affecting calcineurin catalytic activity [39]. Figure 2C shows that VIVIT inhibited the NGF effect on PAI-1 protein levels. To confirm that NGF upregulates PAI-1 by transcriptional regulation, the PAI-1 mRNA levels were quantified in a time-course experiment. FK506 partially inhibited the NGF-induced increase of PAI-1 transcripts and this inhibition was maximal around the time point that PAI-1 mRNA levels peaked (Fig. 2D). Throughout the experiment NGF constantly induced NFAT promoter activity (Fig. 2E) as reflected by the linear accumulation of luciferase over time.

NFAT Activation is not Sufficient for NGF-dependent PAI-1 Induction in PC12 Cells
To further elucidate the contribution of activated NFAT on PAI-1 expression, PC12 cells were treated with ATP, which causes Ca2+ influx via P2X receptors, or the ionophore calcimycin. These stimuli increase the intracellular Ca2+ level and thereby activate the NFAT-mediated gene transcription in PC12 cells [37,40]. Both treatments activated NFAT-dependent promoter activity in our cell system and slightly increased PAI-1 transcription but not PAI-1 protein levels (Fig. 3). However, NGF produced a much greater induction of PAI-1 expression than ATP, although ATP was the strongest inducer of NFAT transcriptional activity in the promoter assay. This result indicates that pathways other than calcineurin/NFAT must be involved in the induction of PAI-1 expression by NGF. The finding of a weaker increase of PAI-1 protein levels as compared to mRNA levels may be due to the different methodology or may reflect effects of translational and post translational regulation.

Potential Role of PKC in the NGF-driven Induction of PAI-1 in PC12 Cells
Since NGF is known to activate PKC [41] and because PAI-1 expression has been shown to be regulated by PKC in other cell types [42], we asked whether PKC was involved in the NGFinduced upregulation of PAI-1 in PC12 cells. Treatment of PC12 cells with the PKC inhibitor Gö6976 reduced the effect of NGF on both PAI-1 protein and mRNA levels ( Fig. 4A upper panel, 4B) without affecting NFAT activation (Fig. 4C), confirming the notion that PAI-1 induction by NGF depends on a pathway separate from NFAT activation. We used a phospho-specific PKC antibody, which detects only the autophosphorylated/activated forms of most PKC isoforms, to monitor the activation of PKC by NGF. NGF altered the phosphorylation pattern of at least one of the PKC isoforms, as shown by the strong increase of the most slowly migrating band ( Fig. 4A lower panel, arrowhead). This increase was reduced by treatment with Gö6976.

DYRK1A Overexpression Counteracts the Effect of NGF on PAI-1
Given that NGF upregulated PAI-1 in an NFAT-dependent manner and that DYRK1A phosphorylates NFAT and promotes its nuclear export [29], we examined the potential effect of DYRK1A on the NGF-induced PAI-1 expression. PC12-NFAT-Luc cells were cotransduced with the FUW-tetO-DYRK1A and FUW-M2rtTA lentiviruses in order to conditionally overexpress DYRK1A under the control of a doxycycline dependent promoter. Treatment with doxycycline resulted in about 3-fold higher levels of DYRK1A in these cells (Fig. 5A). Figure 5B shows that the overexpression of DYRK1A inhibited the NGF induced activation of the NFAT by more than 60%. FK506 totally blocked the NGF effect, while DYRK1A overexpression had no additional effect. Similarly, DYRK1A overexpression reduced by 75% the NGF-induced increase of PAI-1 mRNA levels (Fig. 5C). Overexpression of DYRK1A resulted in a 3-fold increase of DYRK1A kinase activity, whereas NGF treatment had no effect on DYRK1A activity in PC12 cells ( Figure 5D). Throughout the 6-hours time-course experiment, DYRK1A overexpression partially blocked the NFAT activation ( Fig. 6A) and reduced the cellular PAI-1 protein levels ( Fig. 6B). Next we investigated whether DYRK1A overexpression also reduced the amount of cellular tPA/PAI-1 complexes and the free PAI-1 secreted by the cell. Serpin class protease inhibitors like PAI-1 form covalent complexes with their cognate proteases that do not dissociate during SDS-PAGE [43] (Fig. 6C, right panel). Quantitative evaluation of the bands corresponding to the tPA/ PAI-1 complex and of total secreted PAI-1 in the supernatant (free, latent and complexed) suggested that all forms of PAI-1 were reduced by DYRK1A overexpression (Fig. 6C and D). Cellular PAI-1 levels showed the highest NGF effects, probably because the major part of secreted PAI-1 is trapped in cellular tPA/PAI-1 complexes (see panel C).

RCAN1 Overexpression Decreases the Effect of NGF on PAI-1 Expression
Since RCAN1 is also a regulator of NFAT activity, we next studied the effect of RCAN1 overexpression on the NGF-induced PAI-1 expression. PC12 cells were transduced with the FUW-tetO-HA.RCAN1 and FUW-M2rtTA lentiviruses. Figure 7A shows that doxycycline treatment resulted in about 3-fold  overexpression of RCAN1. Throughout a time-course experiment RCAN1 overexpression reduced the NFAT promoter activity (Fig. 7B) and the cellular PAI-1 protein levels (Fig. 7C). Since the transcription of RCAN1 is known to be affected by NFAT [37], we next asked whether in addition to PAI-1, RCAN1 was also upregulated by NGF in PC12 cells. Figure 7D shows that NGF induces the transcription of RCAN1.

DYRK1A Affects the Differentiation of PC12 Cells
To test whether DYRK1A overexpression influences the NGFinduced neuronal differentiation of PC12 cells, cells were treated with NGF in the presence/absence of doxycycline for DYRK1A overexpression (Fig. 8A, 8B). After 2 days the fraction of the cells with one or more neurites at least 1.5-time longer than the cell soma was determined. Figure 8C shows that DYRK1A overexpression promoted PC12 differentiation by about 40%.

Discussion
The present study shows that NGF regulates gene expression in neuronal cells through the calcineurin/NFAT pathway and that this effect is modulated by proteins that are overexpressed in Down syndrome. Specifically, we show that NGF induces PAI-1 expression in the neural cell line PC12 via the activation of NFAT and that overexpression of DYRK1A or RCAN1 counteracts the NGF-induced activation of NFAT and upregulation of PAI-1 expression.

NGF Induces PAI-1
The NGF induction of PAI-1 expression in the neural cell line PC12 cells has been previously described [32,33]. Extending the previous observation that this effect was inhibited by the unspecific tyrosine kinase inhibitor genistein [33], we found that it was sensitive to the TrkA inhibitor K252a. This pathway was also active in cultured murine neurons. NGF upregulates PAI-1 in hippocampal neurons during early neuritogenesis (DIV3) but not during the beginning of synaptogenesis (DIV14). A putative role of PAI-1 during neuronal differentiation is supported by the observation that treatment with PAI-1 promoted the NGFinduced differentiation of PC12 cells [44]. Furthermore, given the documented role of the plasminogen activation system in cell migration, PAI-1 may also be involved in the migration of neuronal progenitors in the hippocampus [45].

NGF Induces PAI-1 via the Calcineurin/NFAT Pathway
The NGF-induced upregulation of PAI-1 expression was sensitive to the calcineurin inhibitor FK506 and the VIVIT peptide, a specific inhibitor of NFAT activation. These results provide clear evidence that the calcineurin/NFAT pathway is critical for this gene regulatory effect of NGF. Furthermore, NGF robustly activated NFAT-dependent promoter activity in the PC12 cells and induced the expression of the NFAT-responsive RCAN1-4 mRNA. A previous study by Minneman et al. [46] had observed no effect of NGF on NFAT-driven promoter activity in PC12 cells, possibly due to differences between the PC12 subclones used in the studies. Although NFAT is a key transcriptional regulator of neuronal development and function, it is poorly characterized as a downstream target of NGF in neurons. Groth et al. [35] showed that NGF increased the NFAT-dependent transcription of BDNF and COX-2 in dorsal root ganglion cells, whereas Nguyen et al. [47] reported that NGF repressed the expression of GAP-43 (growth associated protein 43) in PC12 cells and in cultured cortical neurons through NFAT3. Our results support the hypothesis that the calcineurin/NFAT pathway mediates gene regulatory effects of NGF in neurons.

Calcineurin/NFAT Pathway is not Sufficient for Maximal PAI-1 Induction
Treatment with calcimycin alone was previously shown to increase PAI-1 secretion in the human U937 lymphoma and HepG2 hepatoma cell lines [48,49]. Interestingly, ATP and calcimycin-induced NFAT activation did not increase PAI-1 expression to the same degree as NGF. This result suggests that additionally to the activation of NFAT1-4 (the calcium-responsive NFAT isoforms) non calcium-responsive pathways must be activated for the maximal upregulation of PAI-1 in PC12 cells.

A Putative Role of PKC
NGF induced the activation of at least one PKC isoform in PC12 cells (Fig. 4A). Both PKC activation and PAI-1 induction but not NFAT activation were blocked by Gö6976, which inhibits PKCa and PKCb1 but does not inhibit PKCd, PKCe and PKCdef (IC50..3 mM [50]). These results suggest that PKC is also involved in the regulatory effect of NGF on PAI-1 expression. However, it must be noted that other kinases such as MSK1 and CHK1 are also sensitive to this inhibitor [51]. Studies in several cell types support the idea that PAI-1 regulation is often mediated by activation of multiple signalling pathways and that PKC is often one of them [52,53,54]. In accordance with the effect of Gö6976 in PC12 cells, PKCa [55] and PKCb [56,57] have been shown to regulate PAI-1 expression in other cell types.

DYRK1A and RCAN1 Modulate the NGF-induced Upregulation of PAI-1
We show here that the overexpression of the Down syndromerelated genes Dyrk1A and Rcan1 reduce the effect of NGF on NFAT promoter activity and the upregulation of PAI-1 in PC12 cells. DYRK1A negatively regulates the calcineurin/NFATmediated transcription by phosphorylating NFAT and thereby promoting its nuclear export [28,29]. RCAN1 has a dual effect on calcineurin/NFAT signalling by functioning as an inhibitor or a facilitator depending on its cellular levels [58,59,60]. RCAN1 executes its inhibitory role by directly binding calcineurin and inhibiting the dephosphorylation of NFAT [17]. In PC12 cells NGF had no effect on DYRK1A kinase activity but increased the mRNA levels of RCAN1-4, suggesting the existence of a mechanism of negative feedback for the NFAT activation. Both DYRK1A and RCAN1 are overexpressed in neurons trisomic for chromosome 21 [61,62]. Furthermore, DYRK1A has recently been reported to phosphorylate RCAN1 and to enhance its ability to inhibit calcineurin [30]. This is in line with the suggestion that the simultaneous overexpression of both proteins in Down syndrome inhibits NFAT signalling in a cooperative manner [31].

Effect of DYRK1A on PC12 Differentiation
Overexpression of DYRK1A reduced the NGF-induced activation of NFAT but promoted the NGF-induced differentiation of PC12 cells. This result suggests that maximal activation of the NFAT is not essential for the differentiation effect of NGF. Indeed, a previous study showed that DYRK1A overexpression potentiates NGF-mediated PC12 neuronal differentiation by up-regulating the Ras/MAP kinase signaling pathway [63].

Implications for the Role of the Pathway in Down Syndrome
The present results show that once NGF is bound on the TrkA receptor in PC12 cells, it upregulates PAI-1 via the calcineurin/ NFAT pathway and that the overexpression of DYRK1A and RCAN1 partially blocks this effect of NGF (Fig. 9). Studies in Down syndrome mouse models overexpressing DYRK1A will be needed to assess whether these mechanism are also operative in vivo in the intact brain. Nevertheless, given that the blood PAI-1 levels in Down syndrome individuals are lower than in controls [36] we hypothesize that the attenuation of the NGF-induced PAI-1 by overexpression of DYRK1A and RCAN1 due to trisomy 21 may be relevant to several Down syndrome features. Below we want to present some considerations in this respect. Firstly, DYRK1A is downregulated during brain development as the neurons begin to migrate away from the ventricular zone and is again expressed once they reach their target positions [12].
Considering the known role of PAI-1 in cell migration [45], inhibition of the NFAT-mediated regulation of PAI-1 through DYRK1A overexpression could be involved in the abnormal neuronal migration in Down syndrome. Secondly, individuals with Down syndrome have a low risk for solid tumors, which has been suggested to be related to the negative regulation of angiogenesis by RCAN1 [64,65]. Conversely, both PAI-1, which is used as a prognostic marker for cancer, and NGF can induce tumor angiogenesis [66,67]. The negative effect of DYRK1A and RCAN1 on PAI-1 expression could contribute to the reduced cancer risk in Down syndrome.

Conclusions
Here we show that NGF upregulates PAI-1 expression in young hippocampal murine cultures and that in neural PC12 cells this effect is mediated by at least two pathways, the calcineurin/NFAT and the PKC pathways. Overexpression of the Down syndromerelated proteins DYRK1A or RCAN1 attenuates the activation of NFAT by NGF and blocks the upregulation of PAI-1 expression.
In vivo experiments need to be conducted to test the hypothesis that these newly characterized pathways may be related to the pathophysiological alterations associated with Down syndrome such as intellectual disability and altered angiogenesis in tumor development.    Reagents Murine 7S NGF was obtained from Invitrogen. Rat recombinant PAI-1, calcimycin, FK506 (also known as Tacrolimus), 11R-VIVIT, calcimycin (also known as A23187) and Gö6976 were purchased from Calbiochem. ATP and doxycycline were purchased from Sigma. K252A was purchased from ENZO. Plasmids A lentiviral vector system was used to generate PC12 cell lines for the inducible overexpression of DYRK1A or RCAN1. In brief, rat DYRK1A cDNA was amplified by PCR and subcloned into the lentiviral FUW-tetO backbone after removing the hMYC insert of the FUW-tetO-hMYC plasmid (Addgene 20723) by EcoRI digestion, resulting in the final vector FUW-tetO-DYRK1A. The plasmid pHA-CALP1L [68], which encodes the human full-length RCAN1 isoform 1.1 was kindly gifted by Dr Susana de la Luna (Center for Genomic Regulation, Barcelona). The EcoRI restriction site, which was located downstream of the ATG, was repositioned upstream the start codon (without disturbing the amino acid sequence) and the HA-RCAN1 fragment was obtained after digestion with EcoRI. The HA-RCAN1 fragment was then inserted in the FUW-tetO backbone and the final plasmid was named FUW-tetO-HA.RCAN1.

Cell Lines and Cell Culture
The construction of PC12-NFAT-Luc cells has previously been described [37]. All PC12 subclones were grown in high glucose (4.5 g/L) Dulbecco's modified Eagles medium (DMEM) with Lglutamine and sodium pyruvate, supplemented with 5% foetal bovine serum (PAA Laboratories), 10% horse serum (PAA) and 25 mM Hepes buffer (Sigma). Cells were grown to 60-80% confluency and cultures were split every 48-72 h without trypsinization. Lenti-X 293T cells (Clontech) were grown in high glucose (4.5 g/L) DMEM with L-glutamine and sodium pyruvate, supplemented with 10% foetal bovine serum. All cells were kept in 5% CO 2 at 37uC in T75 filter flasks (Sarstedt). Western Blotting PC12 cells were plated at a density of 100,000 cells/cm2 in 12well dishes one day before the treatments. Cells were lysed with 70 ml/well boiling SDS-buffer (20 mM Tris pH 7.4; 1% SDS). Proteins were separated by 10% SDS-PAGE, transferred to nitrocellulose membranes and blocked overnight at 4uC in Trisbuffered saline, 0.01% Tween 20 (TBS-T) containing 3% (w/v) BSA before they were incubated with the primary antibody overnight at 4uC. The following antibodies were used: rabbit polyclonal PAI-1 antibody (Abcam, #Cab7205), rabbit polyclonal phospho-PKC (pan betaII Ser660) antibody (Cell Signaling, #9371), mouse monoclonal DYRK1A antibody (Abnova, #1859), rabbit polyclonal DSCR1 antibody (Abgent, #AP6315c) the chicken polyclonal antibody for class III beta-tubulin (Neuromics, #CH23005). The last was used as loading control since 6 h of NGF treatment did not affect its expression. The Western blots were developed using horseradish peroxidasecoupled secondary antibodies and chemiluminescence detection. Signal intensities were quantified with a LAS-3000 CCD imaging system and the AIDA Image Analyzer 5.0 program (Raytest, Straubenhardt, Germany).

ELISA
Quantification of free, latent and complexed PAI-1 secreted into the medium was carried out by using the RPAIKT-TOT ELISA kit (Molecular Innovations), following the manufacturer's instructions.

Reporter Gene Assays
PC12-NFAT-Luc cells contain a stably integrated reporter gene plasmid for luciferase expression under the control of four direct repeats of the NFAT binding sequence from the IL-2 gene promoter [37]. Luciferase assays were performed for measuring of NFAT activity using the Luciferase Assay System (Promega). 200,000 cells/cm2 were plated one day before the treatments in 96-well plates. After the treatments the medium was removed and the cells were lysed in 30 ml passive lysis buffer by vigorous shaking for 1 min at room temperature. Luciferase activity was determined by mixing 20 ml lysate with 80 ml luciferase assay mix and measuring light emission in an Orion Microplate Luminometer (Berthold Detection Systems). All data were obtained from triplicate wells.

Primary Hippocampal Cultures
Primary hippocampal cultures were prepared from embryonic day 17 mice as described by Banker and Goslin [72] with slight modifications. Briefly, pregnant C57BL/6J mice were anesthetized in a CO 2 chamber and sacrificed by cervical dislocation, embryos were removed and hippocampi and cerebral cortices were dissected. Cells were dissociated by trituration and plated in polyethylenimine-coated 12-well plates (130,000 cells/cm2) in Neurobasal medium containing Glutamax (Invitrogen) and B27 supplement (Invitrogen).

Immunoprecipitation and DYRK1A Kinase Assays
Cells were plated one day before the treatment in 10 cm dishes (130,000 cells/cm2). After treatment the cells were lysed in immunoprecipitation lysis buffer (50 mM Hepes pH 7.5, 150 mM NaCl, 15% (v/v) glycerol, 2 mM EDTA, 10mM NaF, 5mM NaPP, 1% (v/v) Igepal CA-630 and protease inhibitors. Lysates were precleared by incubation with Protein A Sepharose beads (Amersham) for 1 hour in a rotator at 4uC. The supernatants were rotated overnight at 4C with 1 mg of the indicated antibodies and the next day EZview Red Protein G Affinity Gel (Sigma) was added for 2 additional hours. The beads were washed twice with immunoprecipitation buffer, once with immunoprecipitation buffer without protease inhibitors and detergent and once with kinase buffer (25 mM Hepes pH 7.5, 0.5 mM MgCl 2 , 0.5 mM dithiothreitol). The kinase assays were performed with the biotinylated peptide SAPtide (Bio-RRARKLTATPTPLGG) as described previously [73]. In brief, immunocomplexes were incubated for 30 min at 30uC in 20 ml of kinase buffer, at a final concentration of 10 mm ATP and (c-32P)ATP (100-150 mCi/ pmol) and with 100 mm of the peptide. Incorporation of 32P into SAPtide was determined with the help of the SAM2 biotin capture membrane according to the manufacturer's instructions (Promega).

Determination of Neurite Outgrowth
PC12 cells were plated in collagen treated 6-well plates (5,200 cells/cm2) in growth medium. After 2 h the medium was replaced by DMEM without supplements and the next day by growth medium containing NGF (50 ng/ml) in the presence/absence of doxycycline. PC12 cell differentiation was determined 2 days later when differentiation was not yet completed to facilitate detection of treatment differences. Neurite length was scored by bright field microscopy. Cells having one or more neurites of a length greater than 1.5-fold the diameter of the cell body were scored as positive. At least 200 cells were counted for each treatment. Data were obtained from three independent experiments.

Statistical Analysis
All data derived from three independent experiments, except data shown in figure 2C. The GraphPad Prism 5.0 program (GraphPad Software, La Jolla, CA) was used for statistical analysis. Results were tested for statistical significance by ANOVA, Student's t test or One sample t-test as appropriate.