Recombinant prion protein induces rapid polarization and development of synapses in embryonic rat hippocampal neurons in vitro


Address correspondence and reprint requests to Steinunn Baekkeskov, Diabetes Center, University of California San Francisco, 513 Parnassus Ave., HSW-1090, San Francisco, CA 94143–0534, USA.


While a β-sheet-rich form of the prion protein (PrPSc) causes neurodegeneration, the biological activity of its precursor, the cellular prion protein (PrPC), has been elusive. We have studied the effect of purified recombinant prion protein (recPrP) on rat fetal hippocampal neurons in culture. Overnight exposure to Syrian hamster or mouse recPrP, folded into an α-helical-rich conformation similar to that of PrPC, resulted in a 1.9-fold increase in neurons with a differentiated axon, a 13.5-fold increase in neurons with differentiated dendrites, a fivefold increase in axon length, and the formation of extensive neuronal circuitry. Formation of synaptic-like contacts was increased by a factor of 4.6 after exposure to recPrP for 7 days. Neither the N-terminal nor C-terminal domains of recPrP nor the PrP paralogue doppel (Dpl) enhanced the polarization of neurons. Inhibitors of protein kinase C (PKC) and of Src kinases, including p59Fyn, blocked the effect of recPrP on axon elongation, while inhibitors of phosphatidylinositol 3-kinase showed a partial inhibition, suggesting that signaling cascades involving these kinases are candidates for transduction of recPrP-mediated signals. The results predict that full-length PrPC functions as a growth factor involved in development of neuronal polarity.

Abbreviations used

circular dichroism






extracellular signal-regulated kinase


microtubule-associated protein 2


full-length recombinant mouse doppel protein


full-length recombinant mouse prion protein


phosphate-buffered saline


phosphatidylinositol 3-kinase


protein kinase A


protein kinase C


cellular prion protein


scrapie form of the prion protein


recombinant prion protein


sodium dodecyl sulfate – polyacrylamide gel electrophoresis


full-length recombinant Syrian hamster prion protein

Conversion of PrPC into an alternatively folded isoform, PrPSc, is a fundamental event in transmissible spongiform encephalopathies or prion diseases (Prusiner 1998). PrPC is a ubiquitous cell-surface membrane protein, which is anchored to the membrane by a C-terminal glycosyl phosphatidyl inositol moiety, but can also acquire a transmembrane topology or be secreted (Hegde et al. 1998; Ermonval et al. 2003; for review). The chromosomal PrP gene, Prnp, is a member of the Prn gene family that also includes the gene Prnd encoding Dpl. Dpl shares ∼25% sequence similarity with PrP but lacks the N-terminally located octameric repeats and hydrophobic regions present in PrP (Moore et al. 1999). In contrast to PrP, Dpl has a very low expression in the adult brain, is expressed primarily in testis, and less so in other peripheral tissues (Li et al. 2000). The functions of PrP and Dpl have remained elusive. Mice with an ablated PrP gene (Prnp0/0) are resistant to infection with prions (Büeler et al. 1993; Prusiner et al. 1993), while showing normal development and behavior (Büeler et al. 1992). However, recent studies have revealed deficits in hippocampal-dependent spatial learning and hippocampal synaptic plasticity in Prnp0/0 mice (Criado et al. 2005). Deletion of the Prnd gene in mice results in male sterility but otherwise normal behavior (Behrens et al. 2002) and development of prion disease appears unabated (Behrens et al. 2001). Over-expression of Dpl in the brain of Prnp0/0 mice results in ataxia and cerebellar degeneration (Moore et al. 1999), a phenotype that is not observed in mice concomitantly expressing PrP (Nishida et al. 1999). Thus, the knock-out and transgenic mouse models have suggested intricate and perhaps antagonistic functional networks involving PrPC and Dpl.

One approach to elucidating the function of PrPC has been to identify interacting molecules. PrPC binds copper (Brown et al. 1997) and five octapeptide repeats are responsible for the coordination of bivalent copper ions (Burns et al. 2003). PrPC interacts or associates with a number of proteins including bacterial HSP-60 (Edenhofer et al. 1996), the 37-kDa/67-kDa laminin receptor (Rieger et al. 1997), laminin (Graner et al. 2000), the Grb2 protein, which is central in many signal transduction pathways (Spielhaupter and Schaltz 2001), the lipid raft protein caveolin 1 (Mouillet-Richard et al. 2000), and the neuronal cell adhesion molecule NCAM (Schmitt-Ulms et al. 2001). However, while these studies identified molecules interacting with PrPC, they did not clarify its function (Aguzzi and Hardt 2003; for review). Antibody-mediated cross-linking of PrP on the surface of differentiated 1C11 neuronal cells, mimicking interaction with a ligand, results in activation of p59Fyn kinase (Mouillet-Richard et al. 2000), production of NADPH oxidase-dependent reactive oxygen species and the phosphorylation of extracellular signal-regulated kinases 1/2 (ERK1/2), suggesting a coupling of membrane-anchored PrPC to a signal transduction pathway (Schneider et al. 2003). Recently, cis- and trans-association of PrP with NCAM on the surface of mouse neurons was shown to result in recruitment of NCAM to lipid rafts and activation of p59Fyn (Santuccione et al. 2005). A role of PrPC in axon growth was proposed by Sales et al. (2002), who detected increased expression of the protein on the surface of elongating retinal axons in hamsters. A PrP-Fc fusion protein was shown to bind strongly to the granule cell layer of mouse cerebellum, suggesting the presence of a PrPC-interacting ligand (or ligands) in this region (Legname et al. 2002). In studies that did not distinguish axons and dendrites, it was shown that exposure of neonatal mouse cerebellar or hippocampal neurons to a PrP-Fc fusion protein enhanced the total length of neurites by a factor of ≤ 2 (Chen et al. 2003; Santuccione et al. 2005). Recent evidence suggests that the effect of PrP-Fc on neurite outgrowth in cultures of mouse neurons is mediated through its binding to NCAM (Santuccione et al. 2005).

In this study, we have developed an assay to study the effect of recPrP on fetal rat hippocampal neurons in culture. We report that in vitro, purified full-length recPrP with a conformation similar to native PrPC (Pan et al. 1993), dramatically enhances the development of neuronal polarity, including neurite definition and growth, axonal length, formation of neuronal networks and development of synaptic-like contacts. In the same assay, purified N-terminal or C-terminal domains of PrP fail to enhance polarization. Furthermore, purified recombinant Dpl (recDpl) does not enhance polarization and instead appears to be toxic to neurons.

Materials and methods

Antibodies and materials

The mouse monoclonal antibody 3F4, which specifically recognizes SHaPrP, was described earlier (Kascsak et al. 1987). Mouse phospho-specific monoclonal antibody against phosphorylated NF-200 was from Sigma (St Louis, MO, USA). Rabbit polyclonal antibody to microtubule-associated protein 2 (MAP2) and mouse monoclonal antibodies to Tau1 and to the postsynaptic density protein 95 (PSD-95) were from Chemicon (Temecula, CA, USA). Rabbit anti-synaptophysin antibody was from Zymed Laboratories Inc. (South San Franscisco, CA, USA). Cy3-conjugated secondary antibodies against mouse IgG were from Jackson Immuno-Research Laboratories (West Grove, PA, USA) and Alexa Fluor 488-conjugated secondary antibodies against rabbit IgG were from Molecular Probes (Eugene, OR, USA). NGF 2.5S was from Harlan Inc. (Indianapolis, IN, USA); fibronectin-like engineered protein polymer-plus was from Sigma. Cyclized human amylin was kindly provided by Dr Per Westermark (Uppsala University, Sweden). Inhibitors of signal transduction molecules were from Calbiochem (San Diego, CA, USA).

Protein production and purification

The recombinant PrP constructs used in this study are shown in Fig. 1(a). Recombinant plasmids were expressed in Escherichia coli BL21(DE3) (Novagen, Madison, WI, USA). Purified proteins were lyophilized and solubilized in distilled water before circular dichroism (CD) analyses and addition to neuronal cultures. Purity of proteins was estimated by sodium dodecyl sulfate – polyacrylamide gel electrophoresis (SDS–PAGE) analyses (Fig. 1b). Conformation of purified proteins was analyzed using CD spectroscopy. Analyses of secondary structure was carried out using the program SELCON3 ( (Sreerama and Woody 1993) (Fig. 1c).

Figure 1.

PrP and Dpl constructs and proteins used in this study. (a) Schematic drawing of the PrP and Dpl constructs. (b) SDS–PAGE analysis of 2 μg of purified recombinant proteins. Lanes 1, 7: molecular weight markers; lane 2: MoDpl(27–155); lane 3: MoPrP(23–230); lane 4: SHaPrP(90–231); lane 5: SHaPrP(29–231); lane 6: SHaPrP(23–98). (c) CD spectra for the purified recombinant proteins. Percentages refer to content of α–helices and β-sheets. The remaining structures (not listed) are turns and random coil.

Full-length recombinant mouse PrP [MoPrP(23–230)] was expressed from the pET11a plasmid. The bacterial pellet was re-suspended in 25 mm Tris-HCl and 5 mm EDTA, pH 8.0, and processed twice in a Microfluidizer M-110 EH (Microfluidics, Newton, MA, USA). Inclusion bodies were collected by centrifugation and solubilized in five volumes of 8 m urea, 10 mm MOPS (pH 7.0) by agitation overnight at 20°C. MoPrP(23–230) was purified by column chromatography using carboxymethyl Sepharose (Amersham Bioscience, Picataway, NJ, USA) followed by C4 reverse-phase media (Phenomenex, Torrance, CA, USA). The purity of MoPrP after the C4 column was ≥ 95% (Fig. 1b). Forty-four percent of purified MoPrP was folded in an α-helical conformation and 14% in a β-sheet conformation (Fig. 1c).

Recombinant full-length SHaPrP [SHaPrP(29–231)] and the C-terminal region [SHaPrP(90–231)] were produced from plasmid pNT3A and purified as previously described (Mehlhorn et al. 1996). SHaPrP(29–231) and SHaPrP(90–231) used in this study were ≥ 95% pure as estimated by SDS–PAGE (Fig. 1b). The percentages of α-helix and β-sheet were 40 and 14%, respectively, for SHaPrP(29–231) and 77 and 7%, respectively, for SHaPrP(90–231) (Fig. 1c).

A bacterial construct for the expression of recombinant full-length Dpl [MoDpl(27–155)] in the expression vector pET11d (Novagen) was generated by standard procedures. E. coli cultures were processed in a microfluidizer as described above. Recombinant MoDpl was purified on a Mono S FPLC column. The purity of recombinant MoDpl used in this study was ≥ 95% (Fig. 1b). The percentages of α-helix and β-sheet were 31 and 15%, respectively (Fig. 1c).

Synthetic SHaPrP(23–98) peptide was obtained as previously described (Ball et al. 2001) (Figs 1a and b). CD analysis did not reveal detectable secondary structure of the peptide in solution (results not shown).

Neuronal culture

Primary hippocampal neurons were prepared from embryonic rat brains (E18/E19) as described previously (Kanaani et al. 2002). Neurons were plated onto glass coverslips coated with poly d-lysine at a density of approximately 5 × 104 cells/well of a 24-well tissue culture plate in 0.5 mL of neurobasal medium with 1 × B27 supplement, 0.5 mm glutamine, and 1 × penicillin/streptomycin. Approximately 20 h after plating, recPrP, recDpl, or control proteins, prepared as stock solutions in water, were added to the culture medium at indicated concentrations. Kinase inhibitors, prepared as 500-fold stocks in dimethylsulfoxide (DMSO), were added to the neuronal culture medium 30 min prior to adding recPrP. Neurons were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) (pH 7.4) or methanol (−20°C) (immunostaining for synaptophysin and PSD-95). All procedures involving rats had received prior approval by the Committee on Animal Research at University of California San Francisco.

Immunofluorescence analysis and quantitative measurement of neurite outgrowth and formation of synaptic like contacts

Immunoflourescence analyses were carried out as described (Kanaani et al. 2004) with some modifications. After fixation, neurons were washed three times with PBS, followed by incubation for 20 min at room temperature with a blocking solution (2% normal goat serum in PBS) with or without 0.3% Triton X-100. Neurons were incubated with a primary antibody in blocking solution for 1 h at room temperature, washed three times with PBS, and incubated for additional 1 h at room temperature with the appropriate fluorochrome-conjugated secondary antibody in a blocking solution. After three washes with PBS, the coverslips were mounted on slides with mounting solution. Nuclear staining was performed by mounting the coverslips on slides with a mixture of Fluoromount-G mounting medium (Southern Biotechnology Associates, Birmingham, AL, USA) and Vectashield mounting medium with 4′6-diamidino-2-phenylindole (DAPI) (Vector Laboratories, Burlingame, CA, USA). Fluorescent images were acquired with a Leica TCS NT laser scanning confocal microscope with a Krypton-Argon and UV lasers. Images of neurons immunostained with synaptophysin and PSD-95 were collected on a Leica TCS SP2 AOBS confocal microscope. Usually, eight consecutive horizontal optical sections were line averaged and collected at 0.2–0.5-μm intervals. For the quantitative analyses of axon length and the number of synaptic clusters, images of neurons at similar density in different conditions were selected randomly from the middle of the coverslips and away from the edges, where cell density tends to be too high for analyses of individual neurons. For axon length measurements, projections of the eight optical sections were analyzed with ImageJ software ( In these images, axons were identified as the processes that are predominantly labeled with the axon-specific microtubule-associated protein Tau1, whereas dendrites are predominantly labeled with the microtubule-associated protein MAP2. For each condition, a total of 25–40 neurons from two independent experiments involving multiple coverslips each were analyzed. For the quantification of synaptic-like contacts, projections of eight optical sections collected at 63 × 2 magnification, were analyzed. Synaptophysin and PSD-95 co-localized puncta were manually counted for a total of 25 neurons from each culture. A two-sample unequal variance two-tailed t-test was used to identify statistically significant differences.


RecPrP enhances establishment of polarity and maturation of synapses in cultured fetal rat hippocampal neurons

We established an in vitro assay to study the effect of purified PrP on the polarization and neurite outgrowth of fetal hippocampal rat neurons. One-day-old fetal rat hippocampal neurons were cultured with and without PrP, fixed, immunostained, and analyzed by confocal microscopy. Tau1 or phosphorylated neurofilament-200 were used as markers for axons, and MAP2 served to identify dendrites. Fetal rat hippocampal neurons proceed through five defined stages of growth and polarization following plating in vitro (Craig and Banker 1994; for review). At stage 1, cultured neurons display one or more lamellopodia on their surface. At stage 2, neurons extend short undifferentiated neurites. At stage 3, one of the initially equivalent neurites grows rapidly and becomes the axon. The other neurites extend and develop into dendrites during stage 4. Finally, stage 5 is marked by the formation of a network of interconnecting neurons, maturation of axonal and dendritic arbors, branching of dendrites, and synaptogenesis. Incubation of 1-day-old fetal rat hippocampal neurons for 22 h with full-length Syrian hamster recPrP [SHaPrP(29–231)] (Fig. 1) had a dramatic effect on development of neuronal polarity and neurite extension. While untreated hippocampal neurons plated at low density were still at stage 1 or 2 after 2 days in culture, addition of SHaPrP(29–231) during the second day stimulated polarization in a dose-dependent manner and resulted in progression to stages 4–5 of development (Figs 2a–d). Thus, cultures with 0.45 μm SHaPrP(29–231) showed an increase in differentiation of axons (1.9-fold) and dendrites (13.5-fold) (compare Figs 2a and c; Fig. 3), a fivefold increase in axon length (Fig. 2e), and the formation of interneuronal connections in ≥ 70% of cells (Fig. 2c). While 0.45 μm SHaPrP(29–231) was observed to be the maximal concentration to enhance axonal elongation, higher concentrations of SHaPrP(29–231) (0.9 and 1.8 μm) continued to enhance formation of neuronal circuitry to involve ≥ 90% of cells (Fig. 2d). In control experiments, cultures were incubated with different concentrations of proteins that have been shown to enhance neuritogenesis in other more long-term assays (Koo et al. 1993; Gomez and Letourneau 1994; Kimpinski et al. 1997), including NGF 2.5S (50–400 ng/mL; 3.85–30.80 nm), fibronectin (0.07–0.28 μm), and cyclized human amylin (0.64–5.12 μm). Overnight incubation of primary hippocampal neurons with these proteins did not enhance neuritogenesis and there was no effect on axonal elongation (supplementary Fig. 1). Thus, the stimulatory effect on polarization and growth of axons and dendrites observed in our overnight assay appears to be specific to recPrP.

Figure 2.

Full-length recPrP induces rapid polarization of rat hippocampal neurons in a dose-dependent manner. (a–d) Confocal analyses of 2-day-old, low-density cultures of fetal hippocampal neurons after 22 h of incubation with increasing concentrations of SHaPrP(29–231). After fixation, cells were immunostained for the axon-specific microtubule-associated protein Tau1 (red), dendrite-specific microtubule-associated protein (MAP2; green), and the nuclear stain DAPI (blue). Untreated hippocampal neurons (a) form lamellopodia and a few have an emerging short axon (arrowheads, stages 2–3). Incubation with 0.11 μm SHaPrP(29–231) (b) results in elongation of all axons (arrowheads) and the emergence of differentiated dendrites, expressing MAP2, in some neurons (stages 3–4). With 0.45 and 1.8 μm SHaPrP(29–231) (c and d, respectively), all neurons have well-defined dendrites expressing MAP2 and axons reach maximum length (end of stage 4). Neurons form contacts, a prerequisite for synaptogeneses (stages 4–5). Scale bars, 10 μm. (e) Quantification of the length of axons after incubation with different concentrations of SHaPrP(29–231). The length of the axon, defined as the process predominantly labeled with Tau1, was determined using ImageJ software. For each concentration of SHaPrP(29–231), a total of 20–25 cells from two independent experiments, each involving multiple coverslips, were analyzed. Data represent mean ± SD. *p < 0.001 versus control.

Figure 3.

RecPrP enhances differentiation of axons and dendrites. Axons were defined as processes predominantly labeled with Tau1 and the number of neurons containing an axon was analyzed in cultures with and without 0.45 μm SHaPrP(29–231). Dendrites were defined as processes labeled with MAP2 and the number of neurons containing dendrites was analyzed in cultures with and without 0.45 μm SHaPrP(29–231). Data represent mean ± SD for five independent experiments each involving 40–50 neurons. *p < 0.001 versus control without recPrP.

We next analyzed the formation of synaptic-like contacts in cultures with and without PrP (Fig. 4). Synaptic-like contacts were detected as yellow puncta, representing co-clustering of immunoreactivity to the pre- and postsynaptic markers synaptophysin and PSD-95, respectively. Incubation of 1-day-old fetal rat hippocampal neurons for 2 days with 0.45 μm SHaPrP(29–231) resulted in an increase of synaptophysin clustering in the axon (results not shown). At this time point, the majority of the synaptophysin clusters was localized in axons without contacts to dendrites, indicating that they do not represent synaptic-like contacts. After 7 days in culture with 0.45 μm SHaPrP(29–231), however, the majority of synaptophysin clusters co-localized with PSD-95 (Fig. 4) (mean ± SD: 55 ± 10 co-clusters/neuron; n = 25), while such co-clustering in cultures without SHaPrP(29–231) was rare (mean ± SD: 12 ± 5 co-clusters/neuron; n = 25; p < 0.001) (Fig. 4c). Thus, the enhancing effect of SHaPrP(29–231) on neuronal development in our assay is not limited to the early stages of neuritogenesis, but extends to the formation of mature synaptic-like contacts.

Figure 4.

RecPrP enhances the formation of synaptic-like contacts. (a, b) Confocal analysis of 8-day-old cultures of fetal hippocampal neurons after 7 days of incubation without (a) or with (b) 0.45 μm SHaPrP(29–231). Neurons were fixed with methanol, immunostained for synaptophysin (green) and PSD-95 (red), and the nucleus stained with DAPI (blue). (c) Quantitative analyses of presynaptic synaptophysin and postsynaptic PSD-95 co-clustering in cultures with and without 0.45 μm SHaPrP(29–231). Neurons treated with full-length recPrP show a ∼5-fold increase in co-localization of presynaptic synaptophysin and postsynaptic PSD-95, indicating mature synaptic-like contacts. Data represent mean ± SD of co-clusters per neuron for a total of 25 neurons from each culture. *p < 0.001 versus control. Scale bars, 10 μm.

The effect of recPrP on neuritogenesis is dependent on the full-length molecule, is not species-specific and is not observed for recDpl

We tested whether the stimulatory effect of recPrP was dependent upon the full-length molecule. Primary hippocampal neurons were incubated with two fragments of recPrP: (i) SHaPrP(90–231), which represents the C-terminal globular domain, and (ii) SHaPrP(23–98), the N-terminal domain (Fig. 1). Over the concentration range tested (0.16–1.32 μm), neither SHaPrP(90–231) nor SHaPrP(23–98) had a detectable effect on the morphology of neuronal cultures, indicating that the growth-promoting activity of recPrP requires the full-length protein (Fig. 5).

Figure 5.

The growth-promoting activity of recPrP is not species specific but requires the full-length molecule. Comparative analyses of the effect of full-length recMoPrP, full-length recSHaPrP, and fragments of SHaPrP on axonal outgrowth of hippocampal neurons. (a–c) Confocal analyses of 2-day-old, low-density cultures of fetal hippocampal neurons after overnight incubation with 0.45 μm of the indicated recombinant proteins. Fixed cells were immunostained with MAP2 (green) and DAPI (blue). The green color for MAP2 has been enhanced to show neuronal networks. Scale bars, 10 μm. (d) Quantification of the length of axons after overnight incubation with 0.45 μm of the indicated recombinant proteins. The length of the axon (arrowhead), identified as the process that extends farthest from the cell body and stains only weakly for MAP2, was determined using ImageJ software. For each recombinant protein, a total of 20–25 cells from two independent experiments, each involving three coverslips, were analyzed. Data represent mean ± SD. Neurons incubated with the N-terminal [PrP(23–98)] or C-terminal [PrP(90–231)] domains of SHaPrP are indistinguishable from the control without recPrP. However, the effect of full-length recMoPrP is similar to that of full-length recSHaPrP and both are significantly different from the control without recPrP. *p < 0.001 versus control.

While the effect of PrP seems to depend upon the full-length protein, it does not rely on a species-specific sequence. Purified full-length mouse recPrP [MoPrP(23–230)] had a similar dose-dependent effect as SHaPrP(29–231) over the concentration range tested (0.11–1.74 μm) (Fig. 5d and data not shown).

We also addressed the question whether full-length mouse recDpl [MoDpl(27–155)] has an effect on hippocampal neurons in vitro. In the concentration range of 0.33–1.32 μm, MoDpl(27–155) did not enhance neuritogenesis [Fig. 6, compare (a and b) with (d)] and rather appeared to inhibit the development of axons and dendrites [Fig. 6, compare (a) and (b) with (c)] and induce neuronal death (data not shown).

Figure 6.

RecDpl does not induce neurite growth. Confocal immunofluorescence (a, c, d) and reflective contrast image (b) analyses of 2-day-old, neuronal cultures incubated overnight with either 0.66 μm MoDpl(27–155) (a, b), no addition (c), or 0.45 μm SHaPrP(29–231) (d). While neurons incubated with SHaPrP(29–231) have formed interconnected neurons (stages 4–5 in development) and untreated neurons grow short axons, neurons incubated with MoDpl(27–155) remain at stages 1–2 displaying lamellopodia and immature neurites. (a, c, d) Cells were immunostained for MAP2 (green). Axons are indicated with arrowheads. Scale bars, 10 μm.

Inhibitors of protein kinase C (PKC), p59Fyn, and PI3-kinase block the transduction of recPrP-mediated signals

The 3F4 monoclonal antibody is directed to an epitope that is specific for hamster PrPC and is absent in rat PrPC (Kascsak et al. 1987). Immunostaining of neurons incubated with SHaPrP(29–231) using the 3F4 antibody revealed binding of the hamster recPrP to the surface of the soma, dendrites, and axons of rat hippocampal neurons both in the presence (Figs 7b and c) and absence (results not shown) of TX-100. The pattern of surface immunoreactivity observed with the 3F4 antibody in the presence of TX-100 confirms the lack of reactivity of this antibody with endogenous rat PrP and suggests that bound SHaPrP(29–231) is not internalized. Thus, SHaPrP(29–231) appears to bind to a ligand that is distributed throughout the neuronal surface.

Figure 7.

RecPrP binds to the surface of axon, dendrites, and soma. Confocal analyses of 2-day-old, low-density cultures of fetal hippocampal neurons after overnight incubation without (a) or with (b, c) 0.45 μm SHaPrP(29–231). (a, b) Immunostaining with the 3F4 antibody is shown in red on reflective contrast images; and in green in high-resolution confocal layers (c). The neurons cultured without recPrP extend filopodia [arrows, (a)]. The neurons treated with recPrP form axons (arrowheads), and dendrites (arrows). The 3F4 antibody specifically recognizes recSHaPrP on the surface of axons (arrowheads), dendrites (arrows), and soma (b, c) but does not react with endogenous rat PrPC (a). Scale bars, 10 μm.

Polarization of neurons and the selection of a single future axon among multiple budding neurites in cultured rat hippocampal neurons requires the activity of the growth factor receptor tyrosine kinase, phosphatidylinositol 3-kinase (PI3-kinase) as well as atypical PKC (Shi et al. 2003; Menager et al. 2004). Fas engagement on neurons has been shown to induce neurite growth through ERK activation (Desbarats et al. 2003). In a study of neurite outgrowth in cultures of mouse cerebellar granule neurons, the enhancing effect of PrP-Fc on total neurite length was partially blocked by inhibitors of c-AMP-dependent protein kinase A (PKA), ERK, and the Src kinase family, including p59Fyn, but not by an inhibitor of PI3-Kinase (Chen et al. 2003). ERK1/2 activation was also shown to be a feature of the signaling pathway induced by antibody cross-linking of surface PrPC on the murine 1C11 and GT1-7 neuronal cell lines (Schneider et al. 2003; Monnet et al. 2004).

To assess possible roles for some of these and other signaling pathways in the polarization and axon growth-enhancing activity of recPrP on rat hippocampal neurons, a panel of kinase inhibitors was tested for their ability to inhibit the effect of recPrP on axon elongation (Fig. 8). In these experiments, herbimycin A (0.5 μm), an inhibitor of the Src kinase family, and PP2 (1 μm), an inhibitor of the Src kinase p59Fyn, blocked the enhancing effect of SHaPrP(29–231) on axon growth by 89 and 74%, respectively. Furthermore, bisindolmaleimide I (Bis, 5 μm), an inhibitor of most known isoforms of PKC, blocked the effect of SHaPrP(29–231) by 86%. Finally, Wortmannin (0.2 μm) and Ly 294002 (20 μm), inhibitors of PI3-kinase, showed a partial inhibition (∼40%) of the effect of recPrP on axon elongation. In contrast, PD 98059 (50 μm, an inhibitor of ERK) AG1478 (0.25–25 μm; a selective inhibitor of growth factor receptor kinase), Rp-8-Br-cAMP (10 μm; an inhibitor of cAMP-dependent PKA), pertussis toxin (200 ng/mL; an inhibitor of Gi-protein signaling), and PP3 (1 μm; a negative control for PP2) did not inhibit the effect of recPrP on axon elongation (Fig. 8).

Figure 8.

Effect of kinase inhibitors on recPrP-induced axonal elongation. Analyses of the ability of inhibitors of different kinases to block axonal growth induced by 0.45 μm SHaPrP(29–231). (a–d) Confocal analyses of 2-day-old, low-density cultures of fetal hippocampal neurons after overnight incubation with SHaPrP(29–231) and the indicated inhibitors (in DMSO) or DMSO alone. PrP was added to the culture medium 30 min after DMSO or the inhibitors. Fixed neurons were immunostained for Tau1 (red) and MAP2 (green). The nucleus was stained using DAPI (blue). Scale bars, 10 μm. (e) Quantification of the length of axons after overnight incubation without or with 0.45 μm SHaPrP(29–231) and the indicated inhibitors. Quantification was performed as described in the legend to Fig. 2(e). *p < 0.001 versus control treated with SHaPrP(29–231) and DMSO but no inhibitor.


A possible role of PrPC in neurite differentiation and growth

Polarization of neurons and compartmentalization into axons and dendrites are fundamental for the function of the nervous system. In this study, we show that incubation of newly plated fetal hippocampal neurons with recPrP rapidly enhances the development of polarized and interconnected neurons. Specifically, overnight incubation of 1-day-old fetal neurons with purified full-length recPrP results in a several-fold increase in neurite differentiation, and a 5-fold increase in axon length. The effect of recPrP is not species-specific as both Syrian hamster and mouse recPrP have a similar effect. The specificity of recPrP is suggested by the lack of effect of purifed mouse doppel protein, NGF, amylin, and fibronectin on neurite outgrowth and/or differentiation in our overnight assay.

While neuronal networks are already visible after overnight culture with recPrP, formation of synaptic-like contacts defined by co-clustering of the presynaptic marker synaptophysin and the postsynaptic marker PSD-95 requires a longer incubation with recPrP. After 7 days in culture with recPrP, synaptic-like contacts are increased 5-fold compared with cultures without recPrP. At this stage in culture, the enhancement of synaptic-like contacts in cultures with recPrP appears to be the result of a combination of an increase in axonal length and branching and a 2–3-fold increase in the number of synaptophysin clusters (results not shown), while length and number of dendrites and number of PSD-95 clusters are similar in cultures with and without recPrP.

In the experimental conditions described here, fetal hippocampal rat neurons, seeded at low density (50 × 104/well) to facilitate accurate analysis of axonal length and synapse formation of individual neurons, are highly viable, and cell death is rare. Moreover, the cultures are highly homogenous for hippocampal neurons. Importantly, the effect of recPrP both on axonal elongation during a 1-day culture and on synaptogenesis during a 7-day culture is substantial (factor of five increase), and easy to observe and quantify. In comparison, previous studies on the effect of MoPrP-Fc on neonatal cerebellar granule and hippocampal neurons did not distinguish axons and dendrites, but reported an increase in total neurite length by less than a factor of two (Chen et al. 2003; Santuccione et al. 2005). An enhanced survival of neonatal mouse neurons in the presence of MoPrP-Fc (Chen et al. 2003) was also reported, an effect that was not observed in our analyses of rat hippocampal neurons, which evidence 96–99% viability both in the presence and absence of recPrP. We have conducted a series of experiments to study the effect of purified SHaPrP(29–231) on embryonic (E18/19) and neonatal (P1) mouse hippocampal neurons in culture. Both embryonic and neonatal mouse cultures were more heterogeneous than rat cultures with a high percentage of glial cells in addition to neuronal cells (supplementary Fig. 2). Embryonic mouse hippocampal neurons required plating at a twofold higher density than rat neurons in order for neurons to survive, a constraint that hampered analyses of individual neurons. Cultures of neonatal mouse hippocampal neurons, in contrast, could be plated at a similar density as rat hippocampal neurons, albeit at lower viability, and were amenable to analyses of individual neurons. In these similar conditions, the effect of SHaPrP(29–231) on cultures of neonatal mouse hippocampal neurons was strikingly different from the effect on embryonic rat neurons (supplementary Fig. 2): recPrP induced the differentiation and growth of multiple neurites with axon-like characteristics (i.e. predominant expression of Tau1 and low expression of MAP2), in a starburst effect, rather than inducing a single axon. These differences in the effect of recPrP on mouse and rat neurons suggest that endogenous cues required for selection of a single axon and the interplay with PrP may vary either between rat and mouse, or between embryonic and neonatal neurons. As reported for mouse cerebellar granule neurons and MoPrP-Fc, incubation with SHaPrP(29–231) significantly enhanced the viability of hippocampal mouse neurons (∼twofold), suggesting that, in conditions where cell viability is compromised, recPrP may indeed rescue neurons from cell death. Thus, cultures of neonatal mouse hippocampal neurons may offer an assay for the effect of PrP on cell survival. However, the assay we present using embryonic rat hippocampal neurons appears far more tractable for studies of neuronal polarization and allows detailed analyses of the effect of recPrP on axonal differentiation and growth and synapse formation.

The nature of the implicit cell-surface receptor mediating the effect of purified recPrP on the development of neuronal polarity in our assay remains to be elucidated. Several cell-surface proteins known to interact with PrPC, including NCAM (Schmitt-Ulms et al. 2001; Santuccione et al. 2005) and the laminin receptor precursor (Rieger et al. 1997; Gauczynski et al. 2001), are implicated in neurite outgrowth. Most importantly, there is evidence to suggest that the interaction of PrP-Fc with NCAM in mouse neurons in culture promotes neurite outgrowth (Santuccione et al. 2005). Whether the association of PrP with the laminin receptor results in a cumulative effect on neurite outgrowth or controls a distinct aspect of neuronal differentiation is currently unknown. A blocking of those candidate receptors for PrP in our assay using antibodies has been hampered by toxicity of such blockade for embryonic rat neurons, independent of the presence of recPrP. The role of those receptors in the effect of recPrP on differentiation of fetal rat neurons remains the subject of further investigations. Another candidate for mediating the effect of recPrP is surface-anchored PrPC itself. Gauczynski et al. (2001) did not observe differences in the binding of human recPrP to the surface of mouse neurons from either wild-type or Prnp0/0 mice, suggesting that membrane-anchored PrPC is not the main receptor for human recPrP. Furthermore, no difference was found in the enhancing effect of exogenous PrP-Fc on total neurite length in cultures of mouse cerebellar granule or hippocampal neurons derived from either wild-type or Prnp0/0 mice (Chen et al. 2003; Santuccione et al. 2005). Our experiment also did not reveal a difference in the response of hippocampal neurons from either wild-type or Prnp0/0 mice to recPrP (results not shown). Taken together, these results suggest that membrane-anchored PrPC is not a receptor for trans interactions with PrP.

Our results with fetal rat neurons confirm previous data suggesting a role of p59Fyn in the effect of PrP-Fc on polarization of neonatal mouse neurons (Chen et al. 2003). Activation of p59Fyn appears to be mediated through association of PrP-Fc with NCAM (Santuccione et al. 2005). However, in contrast to the studies of mouse neurons incubated with PrP-Fc, a role for ERK and/or PKA (Chen et al. 2003) was not detected in our assay using rat hippocampal neurons. The blocking effect of inhibitors to PKC and PI3-kinase in our assay is consistent with their role in polarization of neurons (Shi et al. 2003; Menager et al. 2004), but further studies are needed to assess whether these kinases are directly involved in transduction of recPrP-mediated signals.

PrPC appears to fold into multiple forms with different topology in the endoplasmic reticulum (Hegde et al. 1998). The detection of two transmembrane forms of opposite membrane orientation, a GPI-linked form, a secreted form, as well as differentially glycosylated forms (Ermonval et al. 2003; for review), suggests the possibility of more than one function in vivo, and perhaps interactions with multiple ligands and/or receptors. Furthermore, the ratio of α-helix versus β-sheet in purified recPrP may well be a critical parameter for function. The in vitro assay described here, measuring the effect of recPrP on axonal elongation in 1–2 day-old cultures of primary hippocampal neurons, provides an approach to dissect the functions of different PrPC isoforms/conformers and their interaction with other cellular factors.


This work was supported by NIH grant AG021601, cores from NIH grant P30 DK063720, and by a gift from the G. Harold and Leila Y. Mathers Charitable Foundation. We thank Ms Radha Verman and Mr Patrick Culhane for expert technical help, Drs Marc Tessier-Lavigne, for help and discussions in the early phase of this study, and Peter Walter for use of a Leica TCS NT confocal microscope. SBP and GL have a financial interest in InPro Biotechnology Inc.