Address correspondence and reprint requests to Dr Vilma R. Martins, International Research Center, A.C. Camargo Hospital, Rua Tagua 440, São Paulo 01508-010 Brazil. E-mail: firstname.lastname@example.org
Prion protein (PrPC) is a cell surface glycoprotein that is abundantly expressed in nervous system. The elucidation of the PrPC interactome network and its significance on neural physiology is crucial to understanding neurodegenerative events associated with prion and Alzheimer's diseases. PrPC co-opts stress inducible protein 1/alpha7 nicotinic acetylcholine receptor (STI1/α7nAChR) or laminin/Type I metabotropic glutamate receptors (mGluR1/5) to modulate hippocampal neuronal survival and differentiation. However, potential cross-talk between these protein complexes and their role in peripheral neurons has never been addressed. To explore this issue, we investigated PrPC-mediated axonogenesis in peripheral neurons in response to STI1 and laminin-γ1 chain-derived peptide (Ln-γ1). STI1 and Ln-γ1 promoted robust axonogenesis in wild-type neurons, whereas no effect was observed in neurons from PrPC-null mice. PrPC binding to Ln-γ1 or STI1 led to an increase in intracellular Ca2+ levels via distinct mechanisms: STI1 promoted extracellular Ca2+ influx, and Ln-γ1 released calcium from intracellular stores. Both effects depend on phospholipase C activation, which is modulated by mGluR1/5 for Ln-γ1, but depends on, C-type transient receptor potential (TRPC) channels rather than α7nAChR for STI1. Treatment of neurons with suboptimal concentrations of both ligands led to synergistic actions on PrPC-mediated calcium response and axonogenesis. This effect was likely mediated by simultaneous binding of the two ligands to PrPC. These results suggest a role for PrPC as an organizer of diverse multiprotein complexes, triggering specific signaling pathways and promoting axonogenesis in the peripheral nervous system.
Prion protein (PrPC) is a ubiquitous cell surface protein anchored through a glycosylphosphatidylinositol moiety that is associated with transmissible spongiform encephalopathies (TSEs). The conversion of PrPC to the infectious isoform PrPSc in TSEs, results in neuronal loss and progressive neurodegeneration (Moore et al. 2009; Hajj et al. 2012). Depletion of PrPC in mice abrogates prion infection and, although no gross phenotype was detected in these animals under normal conditions [reviewed by (Weissmann and Flechsig 2003)], increased neuronal sensitivity to brain injuries was observed [for review, see (Linden et al. 2008)]. Notably, studies have demonstrated that PrPC ablation in neurons, but not in Schwann cells, promoted an adult-onset chronic demyelinating polyneuropathy that affected sensory and motor pathways of the peripheral nervous system (PNS) (Bremer et al. 2010). Furthermore, severe neurodegeneration occurred when PrPC deletion mutants were expressed in the absence of the wild-type protein (Shmerling et al. 1998; Li et al. 2007).
Several physiological functions of PrPC have been described during the past decade, but the role of PrPC loss-of-function in neurodegeneration remains unknown (Caughey and Baron 2006; Aguzzi et al. 2008). PrPC accumulates in focal adhesions and regulates focal adhesion formation and filopodia extension (Schrock et al. 2009). Furthermore, PrPC depletion in zebrafish results in impaired cell adhesion, which leads to gastrulation arrest (Malaga-Trillo et al. 2009). Many roles for PrPC have been characterized in the mammalian nervous system, several of which are mediated via interactions with cell surface molecules (Linden et al. 2008; Martins et al. 2010; Biasini et al. 2012). PrPC interacts directly with neural cell adhesion molecule (NCAM) (Schmitt-Ulms et al. 2001), leading to stabilization of NCAM in lipid rafts and activation of p59fyn to induce NCAM-dependent neuritogenesis (Santuccione et al. 2005). Components of the extracellular matrix (ECM) and their cell surface receptors have been demonstrated to interact with PrPC, showing that it may regulate ECM organization and function (Linden et al. 2008). Notably, the ECM is a well-known microenvironment with important roles in cell survival, migration, and differentiation (Venstrom and Reichardt 1993). The 37-kDa/67-kDa laminin receptor precursor (37LRP/67LR), which is involved in dynamic cellular processes (e.g., increased filopodia, directional motility, modulation of gene expression, and facilitation of Ln-integrins interaction) (Nelson et al. 2008), is one of the ECM receptors known to interact with PrPC (Rieger et al. 1997; Gauczynski et al. 2001). The interaction between 37LRP/67LR and PrPC could contribute to cell-to-cell communication essential for cell survival in a physiological context (Gauczynski et al. 2001). PrPC also participates in axonogenesis by modulating αvβ3 and β1 integrin activity through its interactions with vitronectin and fibronectin, respectively (Hajj et al. 2007; Loubet et al. 2012). In contrast, PrPC binds directly to contactin-associated protein (Caspr), an inhibitor of neuritogenesis, preventing its shedding by reelin and potentiating its inhibitory effect during neurite outgrowth (Devanathan et al. 2010).
PrPC directly interacts with the laminin γ1 chain (Ln-γ1) (between amino acids 1575-1584 of Ln and 173-182 of PrPC), inducing neuritogenesis in hippocampal neurons and PC12 cells (Graner et al. 2000a,b; Coitinho et al. 2006; Beraldo et al. 2011). The signal transduction pathway involved in Ln-γ1/PrPC-dependent effects requires the activation of group I metabotropic glutamate receptors (mGluR1/5), which leads to phospholipase C (PLC) activation and an increase in intracellular calcium levels (Beraldo et al. 2011). PrPC also interacts with stress inducible protein 1 (STI1) (Zanata et al. 2002), a co-chaperone that associates with Hsp70 and Hsp90 to facilitate protein folding and maturation (Nicolet and Craig 1989). STI1 is secreted into the extracellular milieu by astrocytes (Lima et al. 2007), and its interaction with PrPC (at amino acids 113-128) recruits α7 nicotinic acetylcholine receptors (α7nAChR), which are responsible for transducing trophic signals in hippocampal neurons (Beraldo et al. 2010). Taken together, the existence of a large number of PrPC-interacting molecules suggests that PrPC is a key member of a multiprotein complex in the plasma membrane that organizes signaling modules (Linden et al. 2008; Martins et al. 2010) and may fine tune neurite outgrowth.
Although a key role for PrPC interaction with Ln or STI1 has been demonstrated, the ability of PrPC to form the multi-protein complexes with more than one ligand was never addressed. In this study, we used simultaneously two PrPC ligands to observe how different combinations affect cellular signaling and axonogenesis in dorsal root ganglia (DRG) neurons.
Materials and methods
DRG primary cultures and morphometric evaluation
Primary DRG cultures were obtained from E12.5 wild-type (Prnp+/+) and PrPC knockout (Prnp0/0) mouse embryos (Bueler et al. 1992). DRGs were dissected in Hank's Balanced Salt Solution (HBSS; Invitrogen, Carlsbad, CA, USA) and trypsinized (0.25%) for 20 min at 37°C. Trypsin was inactivated with 10% Fetal Calf Serum (Vitrocell, Campinas, SP, Brazil) in Neurobasal medium (Invitrogen), and after washing with HBSS, the cells were mechanically dissociated in Neurobasal medium. Cells (5 × 104) were transferred to coverslips pre-coated with poly-l-lysine (PL, 5 μg/mL) and cultured in Neurobasal medium supplemented with B-27 (Invitrogen), glutamine (2 mM; Sigma, St. Louis, MO, USA), streptomycin (100 μg/mL), penicillin (100 U/mL), and 50 ng/mL nerve growth factor (NGF; Sigma). Cells were treated with recombinant STI1 (purified as described by (Zanata et al. 2002), full-length Ln [purified from EHS tumors as described by (Paulsson et al. 1987)], synthetic Ln-γ1 peptide (Ln-γ1, RNIAEIIKDI; Neosystem, Strasbourg, France), or scrambled Ln-γ1 peptide (SCR, IRADIEIKID; Neosystem). In axonogenesis assays, Ln, Ln-γ1, or SCR peptide were pre-adsorbed to the PL-coated coverslips and cells were cultured on the substrates. To allow proper adsorption, the peptides were conjugated to bovine serum albumin (BSA). Cells were incubated at 37°C in a humidified atmosphere containing 5% CO2 for 2 or 6 h, fixed with 4% paraformaldehyde for 20 min, washed with phosphate-buffered saline (PBS), and stained with hematoxylin. Images were acquired using an Olympus IX70 inverted microscope equipped with a DP30BW camera (Shinjuku, Tokyo, Japan). Morphometric analyses were performed using ImageJ software (National Institutes of Health, Bethesda, MD, USA) with the NeuronJ plugin. The morphometric parameters evaluated included the percentage of cells with axons and the length of axons. Three to five fields (approximately 300 hundred cells) per treatment were analyzed.
Prnp+/+ and Prnp0/0 DRGs were plated on glass coverslips and fixed with 4% paraformaldehyde containing 0.12 M sucrose in PBS. For permeabilization, cells were incubated with 0.2% Triton X-100 in PBS for 5 min at 25°C. After rinsing with PBS, the cells were blocked with PBS containing 20% horse serum. The cells were incubated overnight with anti-βIII-tubulin antibody (1 : 100; Millipore, Billerica, MA, USA), anti-mGluR1 (1 : 100; Cell Signaling, Danvers, MA, USA), or anti-α7nAChR (1 : 50; Abcam, Cambridge, MA, USA) in PBS containing 1% horse serum at 25°C. The cells were then incubated with secondary antibody, anti-mouse Alexa Fluor® 546 (1 : 1000; Invitrogen), and the nuclei were stained with 4′,6-diamino-2-phenylindole (DAPI, 1 : 1000) or TO-PRO-3 iodide (1 : 500) for 1 h at 25°C. After washing, the coverslips were mounted on slides using Fluorsave Reagent (Calbiochem, San Diego, CA, USA). Immunolabeled cells were imaged using a BX61 Olympus Fluorescence microscope (Shinjuku, Tokyo, Japan) and Leica TCS SP5 II confocal microscope (Wetzlar, Germany).
Surface plasmon resonance
Surface plasmon resonance (SPR) experiments were performed using Biacore X system (GE Healthcare Life Sciences, Pittsburgh, PA, USA) equipped with a CM5 sensor chip. All proteins and peptides used were of > 95% purity (checked by sodium dodecyl sulfate–polyacrylamide gel electrophoresis, SDS-PAGE). Recombinant proteins were produced as previously described (Zanata et al. 2002; Ostapchenko et al. 2008). Recombinant PrP was immobilized on a CM5 chip using a standard amine-coupling procedure to the level of approximately 5000 response units (RU), (Fischer 2010). The chip was equilibrated in 25 mM HEPES, 150 mM NaCl, pH 7.0, and different concentrations of STI1, Ln-γ1 peptide, BSA-Ln-γ1, BSA or their mix were injected at 5 μL/min for 360 s. After each injection off-kinetics were recorded for 100 s, followed by wash and regeneration procedures. Data were analyzed with the supplied Biacore software and Microsoft Excel (Microsoft, Redmond, WA, USA).
Evaluation of signaling pathways in axonogenesis
Signaling pathways involved in DRG axonogenesis were examined using specific pharmacological inhibitors of extracellular signal-regulated protein kinases 1 and 2 (Erk1/2) (U0126, 50 nM), cyclic AMP-dependent Protein Kinase A (PKA) (KT5720, 60 nM), calcium/phosphatidyl-inositol triphosphate-dependent protein kinase C (PKC) (chelerythrine chloride, 100 nM), phosphoinositide 3-kinase (PI3-K) (LY294002, 5 μM), phospholipase C (PLC) (U73122, 500 nM), inositol 1,4,5-triphosphate receptor (InsP3-R) (2-APB, 50 μM), metabotropic glutamate receptor 1 (LY367389, 100 μM), and α7 nicotinic acetylcholine receptor (α-bungarotoxin, 10 nM). After 30 min of pre-incubation with inhibitors, the cells were plated in the presence recombinant STI1, Ln, Ln-γ1 peptide, or SCR peptide, as described above, and cultured for 2 or 6 h at 37°C in a humidified atmosphere containing 5% CO2. All inhibitors were purchased from Calbiochem, except for the LY367389 and α-bungarotoxin, which were purchased from Tocris Biosciences (Bristol, UK).
Cells were washed with ice-cold PBS and lysed with Laemmli sample buffer. Extracts were gently sonicated to disrupt DNA and used for SDS-PAGE (10% gels). Proteins were transferred to nitrocellulose membranes (Amersham, UK) and used for western blotting to detect P-(T202/Y204)-ERK1/2 and total ERK1/2. Bands intensity was quantified using ImageJ software (NIH). Values were represented as relative levels of the phosphorylated-ERK1/2 to total ERK1/2 ratio.
Fluorescence calcium measurements with Fluo-3 indicator
DRG neurons were plated on 50-mm glass bottom microwell dishes (Mat-Tek Corporation, Ashland, MA, USA) for 1 h with neurobasal medium containing NGF. The cells were incubated with the fluorescent calcium indicator Fluo-3-AM (10 μM; Invitrogen) for 30 min in the presence of 2 mM CaCl2 at 37°C in a humidified atmosphere containing 5% CO2. The cells were then washed 3x with HBSS and maintained in Krebs buffer (118 mM NaCl, 4.7 mM KCl, 2.5 mM CaCl2, 1.2 mM MgSO4, 1.2 mM KH2PO4, 25 mM NaHCO3, 5.55 mM glucose, pH 7.4) containing 2 mM CaCl2. In some experiments, extracellular calcium was removed by washing the cells with HBSS and incubating them in a calcium-free Krebs buffer containing 2 mM EGTA. The variation in cytoplasmic fluorescence intensity after treatments was analyzed by confocal microscopy using a Leica TCS SP5 II laser scanning confocal system. Drugs were added to the medium after 50 s of imaging, and fluorescence intensity was quantified using LAS AF software by determining the initial and final fluorescence ratio (F1/F0). Time-lapse curves were smoothened using GraphPad Prism software (La Jolla, CA, USA).
DRGs from 12-day-old embryos were used for mRNA analysis, and cortical tissue from adult mice was used as a positive control. The embryonic DRGs and cortical tissue were homogenized in Trizol, and total RNA was extracted using the Aurum Total RNA for fatty and fibrous tissue kit from Biorad (Hercules, CA, USA). The quantity and quality of the RNA in the extracted samples was determined by microfluidic analysis (Agilent Technologies' Bioanalyzer, Santa Clara, CA, USA). Corresponding cDNAs were synthesized using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystem, Foster City, CA, USA) according to the manufacturer's instructions. The cDNA was subjected to RT-PCR using the REDExtract-N-Ampä Tissue PCR Kit (Sigma). Following an initial denaturation step (95°C, 3 min), the samples were subjected to 40 cycles of 95°C for 10 s and 60°C for 30 s for annealing and extension. In each experiment, a non-template reaction was used as the negative control. The primer sequences were as follows:
The results represent the mean ± standard error of at least three independent experiments. One-way anova with Dunnet ‘s post-hoc test was used to analyze the data from the axonogenesis and calcium experiments.
Ln-γ1 and STI1 promote axonogenesis by interacting with PrPC
To determine the effect of Ln-γ1 and STI1 on axonogenesis (sprouting processes) and axonal extension (axon growth and maintenance) in peripheral neurons, DRG neurons were treated with STI1, Ln, or Ln-γ1 (an Ln peptide corresponding to the PrPC binding domain of the γ1 chain). Axonogenesis was quantified using hematoxylin-stained Prnp+/+ and Prnp0/0 neurons (Fig. 1b and c). An increased number of neurons with short axons was observed following STI1 or Ln-γ1 treatment, and a more robust effect was detected when neurons were plated on Ln (Fig. 1a). Treatment with either STI1 or Ln-γ1 increased the number of neurons with axons in a PrPC-dependent manner after 6 h (Fig. 1b), whereas full-length Ln promoted an extensive increase in the number of cells with axons independently on their PrPC expression. Furthermore, treatment with either STI1 or Ln-γ1 had no effect on axon length (Fig. 1c); in contrast, axon length was altered in both wild-type and PrPC-null neurons treated with full-length Ln. This effect was expected, as Ln is known to present other surface receptors besides PrPC, including a number of integrin subunits and proteins directly associated with the formation of cellular extensions (Belkin and Stepp 2000). A scrambled Ln-γ1 peptide (SCR) had no effect on axonogenesis in either wild-type or PrPC-null neurons. The present results provide evidence that PrPC-STI1 and PrPC-Ln-γ1 complexes promote similar effects in hippocampal (Beraldo et al. 2010, 2011) and DRG neurons.
Ln-γ1 and STI1 cooperate to stimulate PrPC-dependent axonogenesis in DRG neurons
To investigate whether PrPC ligands can mutually cooperate during axonogenesis, suboptimal concentrations of Ln and STI1, as determined by a dose-response curve for each ligand alone (Figure S1), were combined (Fig. 2). Based on the robust axonogenesis response to Ln after 6 h (Fig. 1b and c), the treatment with Ln plus STI1 was performed for 2 h (Fig. 2a). When combined, Ln and STI1 at concentrations 10- and 5-fold lower (0.2 μg/mL and 0.1 μM) than those necessary to induce axonogenesis, respectively, resulted in a positive synergistic effect on the percentage of neurons with axons (Fig. 2b) and on axonal length (Fig. 2c). Remarkably, these effects depend on the neuronal expression of PrPC, as they are observed in Prnp+/+ neurons, but not in Prnp0/0 neurons. We also investigated the synergism between STI1 and Ln in hippocampal neurons. All the combinations between STI1 and laminin that we have so far tried were unsuccessful to demonstrate a synergistic effect between those molecules in hippocampal neurons (data not shown).
Similar to the results observed following treatment with Ln plus STI1, combinations of suboptimal concentrations of Ln-γ1 (3.7 μΜ) with STI1 (0.1 μΜ) also stimulated axonogenesis (Fig. 2a, lower panels and Fig. 2d) and axonal growth (Fig. 2e) only in Prnp+/+ neurons. Notably, a synergistic effect on axonogenesis is observed when PrPC binds simultaneously to STI1 and Ln-γ1 peptide. These findings suggest that perhaps STI1 and Ln-γ1 are able to bind simultaneously to PrPC. To test this hypothesis, we performed SPR studies using STI1 and BSA-Ln-γ1 construct (Ln-γ1 has a small size and does not provide a sufficient SPR signal) as ligands for recombinant PrP immobilized on a CM5 chip (Fig. 2f). BSA-Ln-γ1 peptide was added at 3.7 μM concentration, presenting significant binding to PrP and, importantly, when STI1 and BSA-Ln-γ1 were mixed before injections, the SPR response corresponded to the sum of the values of each peptide individually [Fig. 2f, compare binding curves for STI1 + BSA-Ln-γ1 (dark gray) and STI1/BSA-Ln-γ1 (black dashed line)]. Notably, BSA-Ln-γ1 response was not caused by the presence of BSA, which by itself did not reveal significant binding to PrP in the same SPR setup (dashed gray line). Thus, the SPR analyses showed that Ln-γ1 and STI1 bind simultaneously to PrPC.
Ln-γ1 and STI1 engagement to PrPC triggers a wide range of signaling pathways
We then carried out experiments to determine which signaling pathways are involved in PrPC-dependent axonogenesis mediated by STI1 (0.5 μM), Ln-γ1 peptide (37 μM), and their combination (0.1 μM and 3.7 μM, respectively). Erk1/2 and PKC inhibitors (U0126 [U0] and chelerythrine chloride [CH], respectively) blocked the STI1, Ln-γ1, and Ln-γ1 + STI1-induced axonogenesis (Fig. 3a–c), but they had no effect on Ln-mediated axonogenesis (Fig. 3d). LY294002 (LY2), which blocks PI3-K activity, appeared to affect only Ln-induced axonal growth (Fig. 3d), as no significant inhibition was observed with STI1, Ln-γ1, or Ln-γ1 + STI1 treatments (Fig. 3a–c). Inhibition of PKA activity with KT5720 did not alter the axonogenesis induced by any of the PrPC ligands studied herein (Fig. 3a–d). Control experiments demonstrated that inhibitors alone did not modify axonogenesis (Figure S2). Moreover, to demonstrate that STI1 and Ln-γ1 were able to synergistically activate cell signaling, DRG neurons were treated with STI1 (0.1 μM), Ln-γ1 (3.7 μM), or STI1 (0.1 μM) plus Ln-γ1 (3.7 μM) and cell extracts were submitted to immunoblotting to evaluate Erk1/2 phosphorylation (Fig. 3e–g). Neurons treated with Ln-γ1 (3.7 μM) or STI1 (0.1 μM) did not present any induction of ERK1/2 activity (Fig. 3e and f, respectively). However, when these molecules were combined and used to treat cells an increased Erk1/2 phosphorylation was observed after 5 and 10 min of incubation (Fig. 3g).
Ln-γ1 and STI1 promote a potentiation in [Ca2+]i mobilization in Prnp+/+ DRG neurons
On the basis of previous evidence showing that STI1 and Ln-γ1 promote their effects on hippocampal neuritogenesis using Ca2+ as second messenger (Beraldo et al. 2010, 2011), we carried out experiments to determine whether calcium dynamics play a role in PrPC-mediated signaling in DRGs. DRG neurons were labeled with the fluorescent calcium indicator Fluo3-AM, and fluorescence intensity was evaluated following treatment with STI1, Ln-γ1, or Ln-γ1 + STI1. Treatment with STI1 (0.5 μM, Fig. 4a) or Ln-γ1 (37 μM, Fig. 4b) induced a two to threefold increase in intracellular calcium levels in Prnp+/+ neurons (solid lines), whereas no effect was observed in Prnp0/0 neurons (dotted lines). The addition of thapsigargin (THG), which is an endoplasmic reticulum (ER) Ca2+-ATPase inhibitor that promotes Ca2+ leakage to the cytoplasm, to Prnp0/0 neurons led to an increase in intracellular Ca2+ levels, showing that the lack of an effect on Prnp0/0 neurons by STI1 and Ln-γ1 was not caused by decreased cell viability or impairment of intracellular Ca2+ stores (Fig. 4a–b). Suboptimal concentrations of STI1 (0.1 μM, Fig. 4c, gray line) or Ln-γ1 (3.7 μM, Fig. 4c, dotted line) did not alter intracellular Ca2+ levels; however, when combined, they promoted a twofold increase in Ca2+ signal (Fig. 4c, solid line). The relative Ca2+ levels for each treatment performed in five independent experiments are shown in Fig. 4d. Taken together, these results indicate that the simultaneous binding of STI1 and Ln-γ1 peptide to PrPC potentiates an increase in intracellular calcium levels.
To elucidate the mechanism(s) by which PrPC ligands modulate intracellular calcium levels, we determined which Ca2+ source (extracellular influx or mobilization of intracellular stores) occurred in response to the PrPC/ligand associations. The increase in intracellular Ca2+ levels in STI1-treated DRG neurons (Fig. 5a, solid line) was completely abolished when extracellular Ca2+ was removed (Fig. 5a, dotted line). This finding suggests that PrPC binding to STI1 promotes the activation of Ca2+ channels in the plasma membrane. In contrast, Ca2+ withdrawal (Fig. 5b, dotted line) did not affect the increase in intracellular Ca2+ levels in Ln-γ1-treated DRG neurons (Fig. 5b, solid line); however, inhibition of PLC with U73122 (U73) completely abrogated Ca2+ signaling in Ln-γ1-treated DRG neurons (Fig. 5b, gray line). When extracellular Ca2+ was depleted and the cells were stimulated using a combination of suboptimal concentrations of Ln-γ1 + STI1, the Ca2+ levels increased, although they were lower (Fig. 5c, dotted line) than those observed in the presence of extracellular Ca2+ (Fig. 5c, solid line). Therefore, both intracellular Ca2+ stores and Ca2+ influx contribute to global Ca2+ signaling mediated by the combined treatment with Ln-γ1 and STI1. The relative levels of Ca2+ increase mediated by PrPC ligands in the absence of extracellular Ca2+ are shown in Fig. 5d.
Axonogenesis was also investigated using PLC (U73) and InsP3R (2-APB) inhibitors (Fig. 5e). Inhibition of PLC activity abolished axon elongation mediated by STI1, Ln-γ1, or the suboptimal combination of Ln-γ1 and STI1. On the other hand, the blockade of InsP3R inhibited the effects mediated by Ln-γ1 and Ln-γ1 + STI1 without affecting STI1-mediated responses. Taken together, these results suggest that Ln-γ1 binding to PrPC promotes the mobilization of intracellular Ca2+ by modulating PLC and InsP3 receptors (Fig. 4), which leads to PKC (Fig. 3) activation and promotes axonogenesis. On the other hand, STI1 bound to PrPC induces an extracellular Ca2+ influx, which activates PLC and promotes axonogenesis.
Participation of neurotransmitter receptors in Ln-γ1 and STI1-mediated effects
On the basis of previous evidence showing that PrPC interacts with mGluRs and α7nAChR to mediate Ln-γ1 and STI1 signals, respectively, in hippocampal neurons (Beraldo et al. 2010, 2011), we next investigated whether these receptors could be also involved in axonogenesis in peripheral neurons. RT-PCR analyses showed that DRG neurons expressed both mGluR1 and α7 mRNAs (Fig. 6a). Cortical brain tissue from adult mice was used as positive control. Neurons were also labeled with specific antibodies to detect mGluR1 and α7nAChR (Fig. 6b). In agreement with our previous results from experiments with hippocampal neurons (Beraldo et al. 2011), PrPC engagement with Ln-γ1 in DRG neurons stimulated a Ca2+ pathway (Fig. 6c) and axonogenesis (Fig. 6d), both of which were sensitive to the mGluR1 antagonist LY367389 (LY367). In contrast to our previous results using hippocampal neurons (Beraldo et al. 2010), the α7nAChR antagonist α-bungarotoxin did not inhibit the effect of STI1-induced calcium influx (Fig. 6g) or axonogenesis (data not shown). The mGluR1 inhibitor (LY367389) also did not disturb STI1-induced Ca2+ signaling (Fig. 6g, gray line), suggesting a lack of cross-talk between STI1-induced pathways and mGluR1. Remarkably, the combination of suboptimal concentrations of Ln-γ1 and STI1 favored a pathway independent of mGluR1, as LY367389 was unable to block both calcium mobilization (Fig. 6e) and axonogenesis (Fig. 6f). Considering the lack of effect of α-bungarotoxin upon STI1 signaling in DRG (Fig. 6g) and the participation of PLC in axonogenesis mediated by STI1 and STI1 plus Ln-γ1 (Fig. 5e), we speculated that another transmembrane partner could be involved in transducing these signals. One possible candidate could be the C-type transient receptor potential proteins (TRPC) that are modulated by PLC via DAG production (Albert 2011). The experiments demonstrated that SKF96365, a relatively non-specific inhibitor of TRPC channels (Li et al. 2005; Jia et al. 2007) completely abolished STI1- and STI1 + Ln-γ1-induced axonogenesis (Fig. 6h).
Collectively, these results suggest that in DRGs, Ln-γ1 and STI1 exhibit trophic properties similar to those observed in hippocampal neurons, but the mechanisms underlying their signaling are only partially similar to those found in hippocampal neurons (Beraldo et al. 2010, 2011).
The current study tests the hypothesis that multiple signaling activators of PrPC cooperate to trigger physiological responses by this GPI-anchored protein. Herein, we demonstrate that Ln-γ1 and STI1 induce axonogenesis in a PrPC-dependent manner in DRG neurons, similar to previous observations in hippocampal neurons (Graner et al. 2000a; Lopes et al. 2005; Beraldo et al. 2010). Moreover, there was a synergistic effect on axonogenesis when Ln-γ1 and STI1 were combined in concentrations that were unable to promote effects when either molecule was used alone. The interactions between Ln-γ1 and/or STI1 with PrPC trigger Erk1/2 pathways and induce rapid intracellular calcium mobilization by distinct mechanisms. STI1 induces Ca2+ influx through the plasma membrane, whereas the Ln-γ1 peptide increases intracellular Ca2+ by activating InsP3-R and mobilizing Ca2+ from intracellular stores (ER). Similar to hippocampal neurons (Beraldo et al. 2011), in DRG neurons, the signaling mediated by PrPC-Ln-γ1 peptide depends on group I mGluRs. On the other hand, differently from hippocampal neurons (Beraldo et al. 2010), signaling triggered by PrPC-STI1 in DRG is independent of α7nAChR. Remarkably, when combined, STI1 and Ln-γ1 peptide induced PrPC-dependent axonogenesis without participation of mGluRs and α7nAChR.
Several reports have shown that Ln-γ1 has important neurotrophic properties and plays a significant role in development and disease (Chen and Strickland 2003; Liebkind et al. 2003; Wiksten et al. 2003, 2007; Chen et al. 2009), and that the terminal KDI tripeptide is responsible for the biological functions of the Ln-γ1 peptide (Liesi et al. 2001; Wiksten et al. 2004a,b; Vaananen et al. 2006). Recently, we demonstrated that PrPC interacts directly with group I metabotropic glutamate receptors (mGluRs) in these neurons to transduce Ln-γ1 signals, which in turn activates PLC and produces InsP3, promoting intracellular calcium mobilization from the ER (Beraldo et al. 2011). Our present data demonstrated that similar underlying mechanisms are present in peripheral neurons. As demonstrated herein and by other groups (Alvarez et al. 2000; Hofmann et al. 2001; Carlton and Hargett 2007), DRG neurons express mGluRs, which participate in several events, such as mechanoception and nociception (Crawford et al. 2000; Zhou et al. 2001; Neugebauer and Carlton 2002; Li and Pan 2010). These results suggest a widespread and conserved mechanism for the transduction pathway involved in PrPC-Ln-γ1 signaling in different cell types.
As the initial characterization of STI1 as a PrPC ligand (Martins et al. 1997; Zanata et al. 2002), several reports have provided support for a role of STI1 as a soluble neurotrophic factor (Chiarini et al. 2002; Lopes et al. 2005; Lima et al. 2007; Caetano et al. 2008; Arantes et al. 2009; Hajj et al. 2009; Santos et al. 2011). STI1 is secreted by many cell types (Eustace and Jay 2004; Erlich et al. 2007; Lima et al. 2007; Wang et al. 2010; Walsh et al. 2011), which allows it to interact with PrPC at the cell surface (Zanata et al. 2002; Lopes et al. 2005; Caetano et al. 2008). The signaling pathways required for STI1 to transduce neuritogenic signals were also described in hippocampal neurons and involve the activation of PI3K-Akt-mTOR and ERK1/2 and require calcium influx through α7nAChR (Lopes et al. 2005; Beraldo et al. 2010; Roffe et al. 2010). Similar to results from CNS neurons (Beraldo et al. 2010), the present data show that in DRG neurons, PrPC/STI-induced axonogenesis depends on ERK1/2 activation and calcium influx. On the other hand, the PrPC/STI1 mechanism in DRG neurons also differs in some ways from that in CNS neurons, including the involvement of PLC and PKC pathways and the independence of α7nAChR. Importantly, PLC and PKC pathways participate in the inhibition of the proliferation of astrocytes by PrPC-STI1 (Arantes et al. 2009). These data suggest that STI1-PrPC engagement could activate calcium channels modulated by PLC via DAG production. One class of calcium channels that satisfies these criteria is the family of TRPC (Tu et al. 2009; Albert 2011). The TRPC channels are present in peripheral nervous tissue and have a variety of gating mechanisms, which can be activated by ligand binding, voltage gating, and changes in temperature (Patapoutian 2005; Alessandri-Haber et al. 2009; Ciobanu et al. 2009). These channels have also been implicated in diverse pathological states, including neurodegenerative disorders and cancer (Yamamoto et al. 2007; Nilius and Owsianik 2011; Santoni and Farfariello 2011). Our work presents evidence for the involvement of TRPC channels in STI1-mediated responses. However, more experiments will be necessary to characterize the connection between PrPC-STI1 signaling axis with the activation of this large family of channels.
It is also remarkable that Ca+2 plays a key role in PrPC-Ln-γ1 and PrPC-STI1 signaling, emerging as one of the main signals responsible for mediating PrPC function (Sorgato and Bertoli 2009). The first evidence for a link between prion protein and calcium dynamics came from experiments showing a reduction in bradykinin-induced intracellular calcium mobilization in neuronal cell lines infected with prions (Kristensson et al. 1993; Wong et al. 1996). Thereafter, it was shown that PrPC-null mice present altered intracellular calcium homeostasis in cerebellar granule neurons (Herms et al. 2000) and in hippocampal CA1 neurons (Fuhrmann et al. 2006). Moreover, the hyperexcitability attributed to PrPC-null mice (Collinge et al. 1994; Colling et al. 1996; Walz et al. 1999; Mallucci et al. 2002) may be because of the absence of PrPC inhibitory modulation of the ionotropic glutamate receptor (NMDAR) currents, which could lead to higher excitability (Khosravani et al. 2008). Heterologous expression of PrPC in CHO cells also alters intracellular calcium fluctuation in different cell compartments following the activation of purinergic receptors (Brini et al. 2005).
Our previous study mapped different STI1 and Ln-γ1 binding domains on the PrPC molecule. Although STI1 binds to amino acids 113–128 on PrPC (Zanata et al. 2002), Ln-γ1 interacts within the region of amino acids 173–182 of PrPC (Coitinho et al. 2006). In accord to that, present data shown that both molecules can bind to the same PrPC molecule and cooperate to achieve optimal effects on cellular signaling. In addition, the experiments conducted herein showed that combined treatment with Ln and STI1 causes a robust and synergistic effect on axonogenesis, with growth after only 2 h. Strikingly, even though the response to Ln alone is not dependent on PrPC, the combinatory stimuli with STI1 and Ln is dependent on PrPC. These data suggest at least two possibilities: (i) suboptimal activity levels from integrins triggered by Ln could have synergistic effects with PrPC-STI1 pathways; (ii) at suboptimal concentration Ln could binds preferentially to PrPC through its γ1 chain cooperating with STI1 to stimulate axonogenesis. The mechanisms associated with this phenomenon deserve further exploration.
There are numerous examples of synergism occurring between molecules involved in neurotrophic activities in the literature (Krieglstein 2004; Sharma 2007). Neurotrophic factors from the neurotrophin family (e.g., NGF), glial cell-line derived neurotrophic factor family ligands (e.g., GDNF), and neuropoietic cytokines (e.g., ciliary neurotrophic factor, CNTF) are known to have synergistic effects on neuronal differentiation, survival, synaptic function, and axonal outgrowth/regeneration (Chao 2003; Sharma 2007). Furthermore, there are reports showing cooperation between Ln and NGF in increasing Ln/integrin-induced neurite outgrowth in sensory neurons (Tucker et al. 2005) and PC12 cells (Achyuta et al. 2009).
In the current study, STI1 and Ln-γ1 acted synergistically to transduce neurotrophic signals through PrPC. Remarkably, Ca2+ signaling as well as axonogenesis mediated by PrPC binding to Ln-γ1, and STI1 in DRG are independent of mGluRs. Remarkably, our data point that TRPC channels participate in this event.
The results of the current study indicate that complexes formed between PrPC and its ligands vary depending on the cell type and ligand involved. These findings provide support for the idea that PrPC organizes signaling platforms in plasma membrane microdomains (Linden et al. 2008) and also strengthens the scaffold concept, which suggests that PrPC function could be allosterically modulated by distinct binding molecules leading to different signals and multiple biological responses (Linden et al. 2012). According to this concept, the individual or combinatory binding of Ln-γ1 and STI1 could result in different conformational rearrangements of both PrPC and its ligands leading to the recruitment of diverse transmembrane proteins that participate in a number of downstream signaling pathways. Furthermore, these data provide new cellular substrates to study the molecular basis of peripheral nerve development and to understand the process associated with axonal regeneration.
This study was supported by Fundação de Amparo à Pesquisa do Estado de São Paulo, National Institute for Translational Neuroscience (CNPq/MCT), Ludwig Institute for Cancer Research, Canadian Institutes of Health Research (CIHR), The Alzheimer's Association (USA) and PrioNet-Canada. Fellowships from FAPESP to T.G.S, F.H.B, G.N.M.H, M.H.L, F.C.S.L and M.R. We are also thankful to Professor Regina Markus (University of São Paulo) for fruitful discussions. The authors declare no conflict of interests.