|Home | About | Journals | Submit | Contact Us | Français|
Differentiation of PC12 cells by nerve growth factor (NGF) is characterized by changes in signal transduction pathways leading to growth arrest and neurite extension. The transcription factor p53, involved in regulating cell cycle and apoptosis, is also activated during PC12 differentiation and contributes to each of these processes but the mechanisms are incompletely understood. NGF signaling stabilizes p53 protein expression which enables its transcriptional regulation of target genes, including the newly identified target, wnt7b, a member of the wnt family of secreted morphogens. We tested the hypothesis that wnt7b expression is a factor in NGF-dependent neurite outgrowth of differentiating PC12 cells. Wnt7b transcript and protein levels are increased following NGF treatment in a p53-dependent manner, as demonstrated by a reduction in wnt7b protein levels following stable shRNA-mediated silencing of p53. In addition, overexpressed human tp53 was capable of inducing marked wnt7b expression in neuronal PC12 cells but tp53 overexpression did not elevate wnt7b levels in several tested human tumor cell lines. Ectopic wnt7b overexpression was sufficient to rescue neurite outgrowth in NGF-treated p53-silenced PC12 cells which could be blocked by JNK inhibition with SP600125 and did not involve β-catenin nuclear translocation. Addition of sFRP1 to differentiation medium inhibited wnt7b-dependent phosphorylation of JNK, demonstrating that wnt7b is secreted and signals through a JNK-dependent mechanism in PC12 cells. We further identify an NGF-inducible subset of wnt receptors that likely supports wnt7b-mediated neurite extension in PC12 cells. In conclusion, wnt7b is a novel p53-regulated neuritogenic factor in PC12 cells that in conjunction with NGF-regulated Fzd expression is involved in p53-dependent neurite outgrowth through noncanonical JNK signaling.
The wnt family of secreted lipid-modified signaling proteins are involved in a spectrum of developmental processes and are homologous to the Drosophila wingless and mouse int-1 developmental control genes (Logan and Nusse, 2004). Within the nervous system, wnt family signaling has been implicated in neuronal development (Ille and Sommer, 2005), including the differentiation of neural progenitor cells (Hirabayashi et al., 2004), dendritic development (Ciani and Salinas, 2005) and adult cellular homeostasis including neurogenesis (Lie et al., 2005). Germline loss of wnt1 results in profound central nervous system developmental defects, particularly in the midbrain and cerebellum (McMahon and Bradley, 1990; Thomas and Capecchi, 1990). Dysregulated wnt signaling has also been associated with cellular transformation and tumorigenesis (Nusse, 2005). Individual members of the wnt gene family have been partially characterized based on their ability to induce in vitro cellular transformation (Wong et al., 1994) but recent evidence suggests a more complicated picture of wnt signaling based on cell type (Yi et al., 2005) and receptor availability (Mikels and Nusse, 2006). The simplified model of wnt signaling involves wnt ligand-mediated activation of frizzled receptor (Fzd) and LDL related protein 5/6 receptor (Lrp5/6) heterodimers at the plasma membrane surface, which propagate an internal signaling cascade either resulting in β-catenin stabilization and subsequent activation of the Tcf/Lef transcriptional complex (canonical model) or alternately involves the less-defined Ca2+- or c-Jun N-terminal kinase (JNK)-dependent noncanonical signaling paradigm (Gordon and Nusse, 2006).
PC12 cells have been extensively studied as an in vitro model of nerve growth factor (NGF)-induced neuronal differentiation (Fujita et al., 1989), a process which involves tumor suppressor p53 (tp53) activity for both cell cycle arrest and more recently, neurite extension (Hughes et al., 2000; Fabian et al., 2006; Brynczka et al., 2007; Poluha et al., 1997; Zhang et al., 2006). In order to understand the specific mechanism of p53 function within NGF-induced differentiation, our lab has recently identified a number of p53 transcriptional targets in differentiating PC12 cells through a chromatin immunoprecipitation cloning strategy, in which we identified NGF-regulated p53 binding to the wnt7b locus (Brynczka et al., 2007). Previous reports have suggested that wnt7b is transcriptionally regulated during hindbrain development by Pax6 in mice (Takahashi et al., 2002) and by the TTF-1, GATA6, and Foxa2 transcription factors in cultured lung epithelium from binding sites within the 5′ promoter region (Weidenfeld et al., 2002). In addition to these factors, we identified a region within the first wnt7b intron, approximately 2 kB downstream from the transcriptional start site which was occupied by p53 in a NGF-dependent manner during PC12 cell differentiation.
Wnt7b-mediated signaling has been described as acting through both canonical (Wong et al., 1994; Wang et al., 2005) and noncanonical signaling pathways (Rosso et al., 2005) depending upon both cell type and receptor availability, without inducing cellular transformation (Wong et al., 1994; Naylor et al., 2000). The developmental role of wnt7b has been partially described through the generation of null mouse models. Wnt7b-/- mice are nonviable and die shortly following parturition due to extensive lung malformation and hemorrhage due to defects in mesenchyme proliferation (Shu et al., 2002) and are also deficient in chorion/allantois fusion during placental development (Parr et al., 2001). In hippocampal neurons, wnt7b stimulates dendritic development through noncanonical signaling (Rosso et al., 2005). Neurite outgrowth in differentiating PC12 cells is a p53-dependent process (Di Giovanni et al., 2006; Bacsi et al., 2005) and the aberration of p53 transcriptional activity leads to a generalized lack of neurite extensions (Di Giovanni et al., 2006; Fabian et al., 2006). In the p53-silenced PC12 cell, both a generalized lack of neurite extension upon treatment with NGF and a decrease in expression of the p53 target gene wnt7b have been described (Brynczka et al., 2007). These results suggested that wnt7b may be involved in a signaling mechanism through which p53-regulated neurite outgrowth occurs.
NGF-dependent expression and activity of the wnt7b protein has not previously been described within the wild-type PC12 cell (Erdreich-Epstein and Shackleford, 1998). We hypothesized that wnt7b function may be related to p53-dependent neurite outgrowth during PC12 differentiation and aimed to characterize the expression and function of wnt7b within these cells. We now report that wnt7b is a p53-inducible target gene that is regulated in a time-dependent manner within PC12 cells upon NGF treatment. In combination with previously published data (Brynczka et al., 2007), we collectively demonstrate that p53 regulates wnt7b at the promoter, RNA and protein level. We further identify that p53-regulated wnt7b expression leads to NGF-induced neurite outgrowth through a noncanonical JNK-regulated signaling mechanism and define an NGF-inducible subset of wnt receptors through which this process may occur.
Rat PC12 cells (ATCC, Manassas, VA) were grown in RPMI 1640 medium supplemented with 10% horse serum, 5% fetal calf serum, 4 mM L-glutamine and penicillin/streptomycin antibiotics (cell culture reagents from Invitrogen, Carlsbad, CA) in a humidified 37° C incubator maintained at 5% CO2. PC12 cells which stably express anti-p53 shRNA (hereafter called p53sh#3 cells) have been described previously (Brynczka et al., 2007; Brynczka and Merrick, 2007) and were cultured as above. Briefly, p53sh#3 cells were generated by lentiviral infection and stable selection of infected cells carrying the blasticidin positive selection marker. PC12 cells were plated on rat tail type I collagen (Sigma, St. Louis, MO) prior to experimentation and were differentiated over indicated time intervals by the addition of 50 ng/mL NGF 2.5S (Chemicon, Temecula, CA) in RPMI 1640 medium supplemented with 1% horse serum and antibiotics. Kinase inhibitors targeting JNK (SP600125) and p38MAPK (SB202190) (Calbiochem, Darmstadt, Germany) were used at 1 μM and 25 μM, respectively. Recombinant sFRP1 (R&D Systems) was added directly to culture medium at a dose of 5 μg/mL for both control and wnt7b-expressing cells where indicated and incubated at 37° for 4 hours prior to NGF addition.
The Homo sapiens normal lung fibroblast (IMR-90), fibrosarcoma (HT-1080) and osteosarcoma (Saos-2) cell lines were cultured in DMEM supplemented with 10% fetal calf serum, 4 mM L-glutamine and penicillin/streptomycin antibiotics in a humidified 37° C incubator maintained at 5% CO2. All experiments were carried out using cells of less than 30 passages.
Overexpression studies were performed by transfection of either Rattus norvegicus (Rn) wnt7b cDNA (Genbank accession NM_001009695.1) cloned into the pDREAM 2.1 vector (Genscript, Piscataway, NJ) or Homo sapiens (Hs) tp53 cDNA (Genbank accession NM_00546.3) cloned into the pCMV6-XL4 vector (Origene, Rockville, MD). Transfections were performed 24 hours prior to NGF treatment using either Lipofectamine 2000 (Invitrogen) or FuGene (Roche, Indianapolis, IN) reagents according to manufacturer’s recommended protocol.
qPCR was performed following total RNA isolation and on-column DNase treatment (Qiagen, Valencia, CA) from indicated sample types according to manufacturer’s protocol (Qiagen). RNA concentration was determined using a NanoDrop spectrophotometer (BioRad, Richmond, CA). 1.0 μg RNA from each sample was used for cDNA synthesis by the SuperScript II reverse transcriptase according to manufacturer’s instructions (Invitrogen). qPCR was carried out with 1/20th reaction volume of cDNA per sample, HotStart master mix (SuperArray, Frederick, MD) and 0.25 μM each primer (IDT, Coralville, ID) in a GeneAmp 9700 PCR instrument (Applied Biosystems, Foster City, CA). PCR was carried out for an empirically identified cycle number in which each amplicon was measured within the linear phase of target amplification. Equivalent volumes of each PCR reaction were run on 2% TBE agarose gels containing ethidium bromide and photographed under UV illumination. RT-PCR was performed on cDNA samples isolated as above using SYBR green-based detection (Applied Biosystems, Foster City, CA) in a GeneAmp 7900 real-time PCR instrument (Applied Biosystems). Relative quantitation of gene expression was performed using the 2-ΔΔct method and compared to GAPDH levels. Gene expression in the untreated cell was used as the endogenous control for indicated treatment groups as appropriate for individual cell types. Primer sequences were used as follows: Rn (Rattus norvegicus) wnt7b exon 1 forward 5′-CTGGGAGCCAACATCATCTG-3′, Rn wnt7b exon 1 reverse 5′-TGCCCAAAGACGGTCTTCTC-3′, Rn wnt7b exon 2 forward 5′-GAGGCTGCCTTCACATACG-3′, Rn wnt7b exon 2 reverse 5′-GCCTTCTGCCTGGTTGTAG-3′, Rn wnt7b exon 3 forward 5′-GTCGGGCTCATGTACTACC-3′, Rn wnt7b exon 3 reverse 5′-GTCCTCCTCGCAGTAGTTG-3′, Hs (Homo sapiens) wnt7b forward 5′-TATCCCAGAGAGCAAAGTG-3′, Hs wnt7b reverse 5′-TGTGTTAGTGCCGAGAATC-3′, Rn GAPDH forward 5′-ATCCCATCACCATCTTCCAG-3′, Rn GAPDH reverse 5′-CCTGCTTCACCACCTTCTTG-3′, Hs GAPDH forward 5′-GGACCTGACCTGCCGTCTAG-3′, Hs GAPDH reverse 5′-TAGCCCAGGATGCCCTTGAG-3′, Rn p53 forward 5′-CAGCCAAGTCTGTTATGTGC-3′, Rn p53 reverse 5′-GTCTTCCAGCGTGATGATG-3′, Hs p53 forward 5′-AATAGGTGTGCGTCAGAAG-3′, Hs p53 reverse 5′-CTTACATCTCCCAAACATCC-3′, Fzd1 forward 5′-CGCTCTTCGTCTATCTGTTC-3′, Fzd1 reverse 5′-GTAGTCTCCCCTTGTTTGC-3′, Fzd2 forward 5′-AGGGCACTAAGAAAGAAGG-3′, Fzd2 reverse 5′-TGTAGAGCACGGAGAAGAC-3′, Fzd3 forward 5′-CAGGCACAGTAGTTCTCATC-3′, Fzd3 reverse 5′-AGCAGTCACCACACATAGAG-3′, Fzd4 forward 5′-GTTGGAAAGGCTAATGGTC-3′, Fzd4 reverse 5′-AGTCATCTGCAGAATACCG-3′, Fzd5 forward 5′-CACAGCCACATTCACTATG-3′, Fzd5 reverse 5′-GTAGCGAGTTCAGGTTTTG-3′, Fzd6 forward 5′-CAAAGGTTCCACATCTCTG-3′, Fzd6 reverse 5′-GGTCGTCTCCAGTGTAGTG-3′, Fzd7 forward 5′-GCTTTGTGTCTCTCTTTCG-3′, Fzd7 reverse 5′-CAGTTCTTTCCCTACCATG-3′, Fzd9 forward 5′-AGGTTTTGTGGCTCTCTTC-3′, Fzd9 reverse 5′-AGGGGTCTGTCTTAGTCATG-3′, Fzd10 forward 5′-GGGAGGAGGTAAAAGAAGG-3′, Fzd10 reverse 5′-GTCCCAAACGAGTAGAACAC-3′, Lrp5 forward 5′-GAGCACGTGATTGAGTTTG-3′, Lrp5 reverse 5′-CTCGGTCCAGTAGATGTAGC-3′, Lrp6 forward 5′-TGCTATGTCCTTCACTGTTG-3′, Lrp6 reverse 5′-GCCTCGATTCTCACTAAGC-3′, Ror2 forward 5′-ACCTCTGTCTGCTTCATCC-3′, Ror2 reverse 5′-CCACCCTTGAATTACATACG-3′, Ryk forward 5′-GAAAGGGTCACACTGAAAG-3′, Ryk reverse 5′-ATAGGAAGGAGGTTTCTGTG-3′.
Western blotting was performed on samples grown as above and treated as indicated. Protein isolation, electrophoresis and transfer to nitrocellulose (Invitrogen) was performed as described previously (McNeill-Blue et al., 2006). HRP-based detection was subsequently carried out with either ECL (GEH Amersham, Piscataway, NJ) or SuperSignal (Pierce, Rockford, IL) reagent and used the following antibodies: wnt7b (Q-13, Santa Cruz Biotechnology, Santa Cruz, CA), wnt7a/b (H-40, Santa Cruz), wnt7a (K-15, Santa Cruz) actin (MAb1501R, Chemicon, Temecula, CA), phosphorylated cJun (9164, Cell Signaling, Danvers, MA), phosphorylated JNK (9251, Cell Signaling), total JNK (9252, Cell Signaling), β-catenin (9581, Cell Signaling), donkey anti-rabbit IgG-HRP (GEH Amersham) and sheep anti-mouse IgG-HRP (GEH Amersham).
Indirect immunofluorescence of wnt7b or β-catenin protein was performed after fixation of cells with 4% paraformaldehyde (Electron Microscopy Sciences, Hatfield, PA). Cells were permeabilized with 0.4% Triton X-100 and incubated with either anti-wnt7b (Santa Cruz) or anti-β-catenin antibody (Cell Signaling, Danvers, MA). Protein localization was detected with goat anti-rabbit Alexa 594-conjugated secondary antibody (Invitrogen). Samples were mounted using ProLong Gold reagent (Invitrogen) containing the nuclear counterstain DAPI and photographed using an Olympus IX70 inverted microscope (Olympus, Center Valley, PA) and appropriate filter sets. Image exposure time for each fluorophore was maintained across all samples for each experiment. Merging of images was performed using AxioVision software (Carl Zeiss, Oberkochen, Germany).
Morphologic analysis and image capture of PC12 and p53sh#3 cells treated as indicated was performed at 40x magnification using microscope and software as described above. Total neurite outgrowth was measured in indicated samples by the scoring of neurites at least one cell diameter in length, with at least 100 cells counted per sample. Experiments for neurite number measurements were repeated at least twice and results from multiple experiments were compiled for statistical analysis. Pairwise comparisons were performed between samples as indicated using Student’s t-test at significance level of p ≤ 0.05. Neurite length measurements were performed by scoring neurite outgrowth relative to cell diameter of wnt7b transfected cells and wild-type cells after 48 hours NGF treatment using same samples as above. The smallest recorded neurite length measurement was 0.5 cell diameters. Neurite length was scored in at least 100 cells per sample and average lengths were compared using Student’s t-test with significance at p ≤ 0.05.
Wnt7b transcript levels were significantly elevated from basal (naïve) amounts within 3 days of NGF exposure and were maintained at elevated levels for at least 7 days with continued NGF stimulation (Figure 1A). Transcript levels were undetectably low in the mitotic cell as previously described (Erdreich-Epstein and Shackleford, 1998). Intracellular wnt7b protein levels were increased following NGF treatment within 8 hours (Figure 1B) and remained highly elevated over the course of 7 days (Figure 1C). Intracellular wnt7b protein levels increased rapidly with NGF treatment relative to mRNA levels, which suggested that wnt7b RNA was efficiently translated within PC12 cells. Levels of wnt7b protein were similarly increased relative to total p53 protein levels upon NGF stimulation as previously reported (Brynczka et al., 2007; Brynczka and Merrick, 2007), where significant increases in p53 protein above baseline levels were also detected within 8 hours of NGF treatment.
Recent studies in our lab have characterized significantly elevated p53 occupancy of a binding site within the first wnt7b intron upon NGF treatment in PC12 cells with accompanied p53-dependent transcription of the wnt7b locus (Brynczka et al., 2007). We aimed to determine whether wnt7b protein expression was also dependent upon p53 activity. Two constitutive anti-p53 shRNA-expressing PC12 cell lines were utilized in which p53 RNA and protein levels were stably decreased relative to wild-type PC12 cells (Brynczka et al., 2007; Brynczka and Merrick, 2007) to explore the effect of reduced p53 protein levels on wnt7b expression. As described above, wnt7b protein was highly elevated within 24 hours of NGF treatment in wild-type cells compared to the untreated naïve cell. However, wnt7b levels following NGF treatment in each p53-silenced cell line were significantly lower and approached the amount observed in the untreated wild-type cell (Figure 2A). Silencing of p53 levels was more efficient in p53sh#3 than p53sh#2 cells (Brynczka et al., 2007), which was reflected in the proportionately graded reduction of wnt7b levels within each cell line.
Since mouse and human p53 appear to operate in an analogous functional manner despite roughly 15% divergence in primary sequence (Luo et al., 2001), we aimed to determine whether the wnt7b gene could be also be effectively induced by Homo sapiens p53 within the Rattus norvegicus PC12 and Homo sapiens Saos-2, HT1080 and IMR-90 cell lines. Constitutive overexpression of Hs tp53 cDNA in PC12 cells led to observable increases in p53 RNA levels, which could be distinguished from endogenous Rn p53 transcripts through primer selectivity (Figure 2B). Elevated Hs p53 led to significantly increased wnt7b transcript levels in the absence of NGF treatment in PC12 cells, demonstrating that Hs p53 protein was capable of inducing Rn wnt7b transcription. No changes in endogenous Rn p53 levels were observed following Hs p53 overexpression in PC12 cells (Figure 2B), suggesting that increased wnt7b expression was solely dependent upon exogenous Hs p53. Overexpression of Hs p53 cDNA in the human p53-null Saos-2 osteosarcoma cell line significantly increased tp53 RNA levels without observable changes in wnt7b levels (data not shown). Similarly, no changes in wnt7b levels were observed following Hs p53 overexpression in either the HT1080 or IMR-90 cell lines (data not shown). Collectively, these data demonstrate that increased p53 protein expression was capable of initiating wnt7b transcription in PC12 neuronal cells.
The wnt7b protein was identified as a single 27 kDa band via immunoblot (Figure 1C) within rat PC12 cells. However, the calculated molecular weight of wnt7b derived from accession NP_001009695 is 39.3 kDa which was observed by wnt7b immunoblot following cDNA overexpression in COS monkey embryonal kidney cells (Burrus and McMahon, 1995). We also noted that a wnt7b variant with alternative first exon usage had been described within the developing chick eye (Fokina and Frolova, 2006), a region in which p53 is highly expressed (Pokroy et al., 2002). Thus, we aimed to determine whether the observed lower molecular weight wnt7b protein was generated as a transcriptional variant regulated by p53 or as a full-length transcript post-translationally modified to generate the observed protein. Constitutive expression of full-length wnt7b cDNA lacking intronic sequence, and therefore lacking the identified p53 binding site within the first intron (Brynczka et al., 2007), also produced a 27 kDa product that we observed by immunoblotting (Figure 3A). Wnt7b protein expression was increased in mitotic cells transfected with pDREAM-wnt7b, while NGF treatment alone led to elevated wnt7b levels that were further increased in transfected cells. No changes in migration distance were detected between endogenous and ectopically-expressed wnt7b.
Using primers to selectively amplify each exon within wnt7b cDNA, it was determined that each exon was represented in the expressed transcript (Figure 3B). Amplicon levels from each exon were increased as expected upon NGF treatment but differences in intensity among exons 1 through 3 were attributed to varying amplification efficiencies or PCR product sizes from relative differences in ethidium bromide signal output. Sequencing of a near full-length wnt7b cDNA amplicon determined no differences from the annotated full RNA sequence (accession# NM_001009695.1) for wnt7b when compared via pairwise alignment. Therefore, each of three exons was represented in the wnt7b transcript upon NGF-induced differentiation of PC12 cells despite SDS-PAGE migration at 27 kDa in reducing conditions. The observation that neither decreased migration nor multiple bands were identified by immunoblot in transfected cells suggested that wnt7b is post-translationally processed (i.e. proteolytic cleavage) in the differentiating PC12 cell.
Immunofluorescent measurement of wnt7b within mitotic and differentiating PC12 cells demonstrated nearly undetectable levels within mitotic cells, while transfection with pDREAM-wnt7b greatly increased wnt7b fluorescence (Figure 3C). In particular, overexpressed wnt7b was found to localize as dense foci either within or at the plasma membrane surface of the mitotic PC12 cell. Following NGF treatment, wnt7b protein was readily detected and evenly distributed throughout the cytoplasmic compartment with localized nodal regions apparent upon transfection with pDREAM-wnt7b. Both primary antibody alone and secondary (Alexa 594-conjugated) antibody alone did not produce background fluorescence (data not shown). Because the antibody used was capable of detecting both wnt7a and 7b, we also analyzed wnt7a expression to verify the specificity of immunofluorescence measurements. No wnt7a immunoreactive bands were detected on immunoblot of PC12 cellular lysate in either the mitotic or NGF-differentiated state (data not shown). Further, our inability to detect wnt7a confirms previous reports describing a lack of wnt7a expression in wild-type PC12 cells (Erdreich-Epstein and Shackleford, 1998) and supports the antibody selectivity for wnt7b detected in our immunofluorescence measurements.
In order to determine the physiological effect of wnt7b within PC12 cells, we studied phenotypic changes upon overexpression in both the wild type and p53-silenced p53sh#3 cell line (Figure 4A). NGF treatment for 24 hours induced neurite extension in wild-type cells while neurite formation was inhibited in p53sh#3 cells as previously reported (Brynczka et al., 2007) and as seen in both dominant negative (Di Giovanni et al., 2006; Eizenberg et al., 1996) and temperature-sensitive (Zhang et al., 2006) p53 expressing neuronal cells. Upon transfection with pDREAM-wnt7b, no observable changes in neurite growth were detected in mitotic cells. However, ectopic wnt7b overexpression was sufficient to recover neurite outgrowth in p53sh#3 cells in the presence of NGF (Figure 4A, results collated in Figure 6B). Significantly increased neurite length (Figure 4B) was also observed within 48 hours in wnt7b overexpressing cells concurrently treated with NGF as compared to control NGF-treated cells.
Because expression of wnt7b rescued neurite outgrowth in p53-silenced PC12 cells following NGF treatment, we aimed to describe the signaling mechanism through which wnt7b contributes to neuritogenesis. Canonical wnt signaling is known to induce the accumulation and nuclear localization of the Tcf/Lef transcriptional cofactor β-catenin (Akiyama, 2000). Subcellular localization and relative levels of β-catenin were visualized by indirect immunofluorescence following NGF treatment and wnt7b overexpression in wild-type and p53sh#3 cells (Figure 5A). β-catenin was localized primarily at intercellular junctions as previously described (Perez-Moreno and Fuchs, 2006) in each sample group irrespective of NGF stimulation, p53 silencing or wnt7b overexpression. Both primary antibody alone and secondary (Alexa 594-conjugated) antibody alone did not produce fluorescence. Absolute protein levels of β-catenin were not stabilized following NGF treatment or wnt7b overexpression in comparison to naïve cells as determined via both immunofluorescence (Figure 5A) and immunoblotting (Figure 5C). Lack of nuclear β-catenin localization or stabilization following either NGF treatment alone or NGF treatment with concomitant wnt7b overexpression demonstrated that canonical wnt signaling was not activated by NGF or wnt7b within the tested time frame.
Wnt signaling may also proceed along a noncanonical pathway, propagated through cJun N-terminal kinase (JNK) activity and the subsequent post-translational phosphorylation of downstream molecular targets (Veeman et al., 2003). Wnt7b activation of noncanonical signaling was therefore studied as a putative mechanism through which neurite outgrowth could be influenced by wnt7b. Phosphorylation of cJun, a direct downstream target of activated JNK, was measured over time in wild-type PC12 cells following NGF treatment (Figure 5B). PC12 cells treated with NGF demonstrated a rapid spike in cJun phosphorylation at 1 hour relative to those cultured in complete medium, with phosphorylation levels significantly decreased within 4 hours before becoming elevated again at 48 hours following NGF treatment. Notably, the time course of delayed cJun phosphorylation was similar to the observed increase of wnt7b levels following NGF treatment (Figure 1C). In order to determine whether increased JNK signaling activity was dependent upon wnt7b and whether wnt7b protein was secreted into culture medium, we expressed wnt7b cDNA in PC12 cells concomitantly treated with NGF and recombinant secreted Frizzled-related protein 1 (sFRP1), an endogenous extracellular wnt7b antagonist (Rosso et al., 2005) (Figure 5C). NGF treatment alone for 24 hours resulted in observable phosphorylation of JNK, which could be increased further by overexpression of wnt7b. JNK phosphorylation was nearly undetectable in NGF-treated p53sh#3 cells but was marked in cells concomitantly expressing wnt7b. Addition of sFRP1 directly to culture medium markedly reduced JNK phosphorylation in both wnt7b overexpressing and control wild-type and p53sh#3 cells treated with NGF, demonstrating that both endogenous and ectopically-expressed wnt7b must be secreted in order to activate JNK phosphorylation. In addition, neither expression of wnt7b nor addition of sFRP1 to medium resulted in changes in β-catenin protein levels in either wild-type or p53sh#3 cells, demonstrating that the wnt7b signaling axis does not incorporate β-catenin stabilization.
In order to determine whether JNK activity contributed directly to neurite outgrowth, we treated wnt7b-expressing wild-type and p53sh#3 cells with NGF and kinase inhibitors (Figure 6A). JNK and p38MAPK, an associated family member which is not recognized as a component of wnt signaling and used as a negative control, were respectively inhibited with SP600125 and SB202190. Effects upon neuritogenesis following 24 hours NGF treatment, selective kinase inhibition and wnt7b overexpression were scored in wild-type or p53sh#3 PC12 cells (Figure 6A, compiled in in6B).6B). Compared to high levels of neuritogenesis visible in both wild-type and p53sh#3 wnt7b-transfected cells treated with NGF (Figure 4A), neurite outgrowth and extension in both cell types were significantly attenuated following JNK inhibition (Figure 5A, 5C). By comparison, p38MAPK inhibition did not significantly abrogate neurite outgrowth in either cell type transfected with wnt7b. Attenuation of neurite outgrowth in wnt7b-expressing p53sh#3 cells following JNK, but not p38MAPK, inhibition demonstrated the functional involvement of JNK signaling in the process of neurite outgrowth involving wnt7b.
Wnt7b overexpression in p53sh#3 cells can rescue neurite outgrowth in the presence of NGF. However, we found that wnt7b expression without growth factor stimulation had no effect on neurite growth in PC12 cells. We hypothesized that NGF treatment caused induction of wnt receptor(s) through which the wnt7b ligand might potentially activate JNK signaling. In order to identify putative NGF-inducible wnt7b receptors, expression levels were measured by RT-PCR over time at 4 to 48 hours for all annotated wnt-associated receptors identified within the Rattus norvegicus genome (Figure 7). Compared to the untreated naïve wild-type cell, expression levels for two Fzd receptors from the 13 tested genes were significantly increased over time following NGF treatment. We observed significantly elevated Fzd7 expression within 4 hours of NGF treatment, with levels subsequently decreased over the course of 48 hours (Figure 7). In addition, Fzd9 was also identified as an NGF-inducible wnt receptor with transcript levels that were increased within 4 hours and were sustained over the course of 48 hours (Figure 6A, 6B). Observed increases in Lrp6 (Figure 6A) were not recapitulated in subsequent qPCR (Figure 6B) experiments. Levels of Fzd10 were also modestly increased within 24 hours following NGF treatment with levels near baseline at other tested time points. In contrast, we observed a significant reduction in transcript levels of Fzd5 and Fzd2.
The current study demonstrates an expanding role for p53 signaling in cellular differentiation, specifically within NGF treated PC12 neuronal cells. In addition to its known roles in DNA damage-induced growth arrest and apoptosis (Vousden, 2006; Vogelstein et al., 2000), p53 appears to be significantly involved in neuronal differentiation. p53 does so by negative regulation of the cell cycle (Hughes et al., 2000) and the subsequent positive induction of neurite outgrowth through factors such as wnt7b and other target genes recently identified by our lab (Brynczka et al., 2007) as well as the recently described actin binding protein, coronin 1b, and the GTPase rab13 transcriptional targets (Di Giovanni et al., 2006). Data presented here, along with previously published experimental data (Brynczka et al., 2007), demonstrated that p53 can positively regulate wnt7b at the promoter, RNA and protein level in a manner dependent upon NGF stimulation during differentiation of PC12 cells. Rescue of neurite outgrowth by ectopic wnt7b expression in p53sh#3 cells suggests that wnt7b-induced signaling is an important pathway through which p53 promotes neurite outgrowth during PC12 neuronal differentiation, and therefore represents a unique target gene of the p53 transcription factor through which neuronal morphology may be regulated. While these studies have described a role for p53 transcriptional activity in the differentiation of neuronal cells, the relationship is not yet clear between differentiation and other recognized functions of p53 signaling that include the DNA-damage response, cell cycle arrest and pro-apoptotic signaling. It is appropriate to consider how these other diverse functions of p53 might be integrated into neuronal differentiation. Interestingly, components of the intrinsic apoptotic pathway have functions within neuronal developmental processes that include caspase activity in synaptic plasticity (Chan and Mattson, 1999) and Bcl-XL/Bax expression in the determination of neuronal lineage (Chang et al., 2007). p53-dependent transcriptional regulation of genes involved in both cell cycle arrest and differentiation may provide a mechanism through which progression of both these processes can proceed from a single pathway node (Bacsi et al., 2005), perhaps suggesting a method in which cell cycle arrest can be ensured prior to undergoing morphological changes during differentiation. Furthermore, defects in PC12 differentiation following p53 silencing suggests that redundancy in these processes by a p53 functional equivalent does not exist in this cell type. However, the larger role for p53 during in vivo development remains to be determined, where the majority of p53-null mice are developmentally viable (Donehower et al., 1992) with a subset exhibiting neuronal malformations (Sah et al., 1995).
Wnt7b promoted neurite outgrowth and extension within differentiating PC12 cells through a signaling mechanism involving the noncanonical JNK pathway without activating canonical β-catenin nuclear translocation and signaling. The lack of neurite outgrowth in mitotic cells overexpressing wnt7b suggested that wnt7b alone was not sufficient for neurite development. Based on our gene expression studies, we propose that wnt7b expression cooperates with NGF-inducible Fzd receptor expression (Fzd9 and/or Fzd7) and decreased expression of other receptors such as Fzd5 to promote increased neurite number and length. While establishment of wnt7b as a definitive ligand for either the Fzd9 or Fzd7 receptors has not yet been confirmed, the NGF-dependent time course of Fzd9 and Fzd7 expression suggests an enticing mechanism through which PC12-acquired competence to the wnt7b ligand could lead to the induction of neurite outgrowth. Furthermore, data presented here argues against p53-dependent expression of the receptors Fzd7 or Fzd9. Importantly, we have demonstrated that introduction of wnt7b cDNA into p53sh#3 cells results in the recovery of neurite outgrowth but this rescue occurs only in the presence of NGF. The data suggest that the signaling mechanism for neurite outgrowth can occur with NGF-dependent gene expression and wnt7b expression, with concurrent p53 silencing. The lack of β-catenin involvement and the identified contribution of JNK activity in the process of neurite outgrowth suggests that wnt7b is a noncanonical wnt ligand in the differentiating PC12 cell. Since Fzd receptor type is known to lend specificity to downstream signaling upon stimulation by wnt ligands (Mikels and Nusse, 2006), both the cell type and expression complement of Fzd receptors may be involved in selective activation of downstream pathways by wnt ligands. In this regard, wnt7b reportedly activates canonical signaling in epithelial and smooth muscle vascular cells (Wang et al., 2005) but promotes osteogenesis and dendritic development through noncanonical mechanisms involving PKCdelta and JNK, respectively (Tu et al., 2007; Rosso et al., 2005). Similarly, the outcome of Fzd receptor activation may be functionally divergent as a function of cell type in which they are expressed. We observed increased expression of Fzd7 and Fzd9 along with decreased expression of Fzd5 and Fzd2 wnt receptors. The Fzd7 and Fzd9 receptors identified in this study are involved in developmental processes, where the Fzd7 receptor appears to be selectively expressed in glial precursor cells and Fzd9 in precursor neuronal cells of the developing mouse midbrain (Rawal et al., 2006). Fzd7 influences cellular morphology (Vincan et al., 2005; Vincan et al., 2007; Chen and Gumbiner, 2006) and neural crest induction (Abu-Elmagd et al., 2006) while Fzd9 is involved in normal hippocampal and behavioral development (Zhao et al., 2005) and is also expressed in neural precursor cells within the developing neural tube (Van Raay et al., 2001). Interestingly, we observed downregulation of the Fzd5 receptor, described as a receptor for the related wnt7b family member, wnt7a, in PC12 cells (Caricasole et al., 2003), although we did not observe wnt7a expression in either the undifferentiated or differentiated state. The discrete subset of NGF-upregulated and downregulated Fzd receptors suggests to us that differentiating PC12 cells may be primed to respond to wnt availability in a specific autocrine manner. Future studies concerning the activation of these receptors by wnt7b ligand within the PC12 cell type should provide additional information concerning the role of Fzd9 and Fzd7 in NGF-induced differentiation.
Wnt proteins are primarily regulated at the level of the endoplasmic reticulum by factors such as the evolutionarily conserved porcupine chaperone, which aids in the processing of nascent wnt proteins (Tanaka et al., 2000) through post-translational glycosylation and palmitoylation (Tanaka et al., 2002; Hofmann, 2000) and ultimately enables the proper folding and transport of wnt proteins from the ER for secretion. Immunofluorescence experiments shown here upon NGF treatment demonstrate dispersed cytosolic localization of wnt7b, while overexpression of wnt7b also resulted in continued cytosolic localization with a noticeable punctate, membrane-associated localization. These data suggest that wnt7b may be associated with the endoplasmic reticulum upon expression in PC12 cells prior to secretion. This observation is similar to previous reports describing wnt ER localization during processing (Tanaka et al., 2002), followed by transport to membrane regions such as lipid rafts prior to secretion (Zhai et al., 2004). Data presented here also demonstrate that processing of wnt7b in PC12 cells apparently generates a 27 kDa protein, unlike the predicted higher theoretical molecular weight of wnt7b or the demonstrated size upon wnt7b cDNA expression in COS monkey kidney cells (Burrus and McMahon, 1995). Our data indicate that processing is likely attributable to post-translational cleavage and not alternative splicing events or transcription from an alternative p53-regulated downstream transcriptional start site. Multiple pieces of evidence presented here support this conclusion: 1) expression of wnt7b cDNA lacking the intronic p53 binding site resulted in a protein of 27 kDa in size as determined via immunoblot, and 2) sequencing of endogenous wnt7b cDNA demonstrated expression of the full-length annotated gene. The significance of a truncated wnt7b protein within these cells is not yet clear but the molecular weight does not appear to be dependent upon p53-induced transcription.
Multiple reports have described tissue-specific regulation of gene loci by activated p53 (di Masi et al., 2006; Fei et al., 2002; Coates et al., 2003) which suggest that activation of targets such as wnt7b may be dependent upon transcriptional coactivators or signaling in a specific cell context. Evidence presented here supports this conclusion, as Homo sapiens p53 protein was capable of inducing transcription of the wnt7b gene in Rattus norvegicus neuronal PC12 cells but not in human normal lung fibroblasts (IMR-90), fibrosarcoma (HT-1080) or osteosarcoma (Saos-2) cell lines. Transcriptional cofactors for p53 such as JMY, hnRNP K (Moumen et al., 2005), YY1 (Gordon et al., 2006; Sui et al., 2004) and the SWI/SNF complex (Lee et al., 2002) each have a specificity for participating with p53 in gene expression according to a specific activating stimulus and cell type. Differences in chromatin assembly and accessibility for p53 binding may also play a large role in cell-type specific differences in transcriptional activity and gene expression. Accordingly, it is not yet possible to definitively determine whether wnt7b is a p53-inducible target gene in human cells, where a wider sampling of cell types or conditions may be necessary to identify wnt7b transcriptional regulation.
Based on data presented in this study, we propose a signaling paradigm as shown in Figure 8 through which the transcription factor p53 regulates PC12 neuronal differentiation following NGF treatment. Activation of the p53 transcription factor results in its increased nuclear localization and the transcriptional activation of wnt7b and other target loci. Wnt7b processing (~27 kDa) and concomitant p53-independent expression of cognate Fzd7 and/or Fzd9 receptors with NGF treatment ultimately generates a competent PC12 cell, in which the wnt7b ligand stimulates JNK signaling activity via a noncanonical wnt pathway involving the putative wnt7b receptors Fzd9 and/or Fzd7. We conclude that noncanonical wnt-regulated JNK signaling promotes the outgrowth of neurites and the morphological differentiation of the PC12 cell. Future studies should further refine the contribution of wnt7b to neurite outgrowth and determine its extensibility to other neuronal systems and species.
This work was supported by the Intramural Research Program of the NIH, National Institute of Environmental Health Sciences. We thank Dr. Kevin E. Gerrish and Dr. Serena M. Dudek for critical review of this manuscript.