Transgene-driven expression of the Doppel protein in Purkinje cells causes Purkinje cell degeneration and motor impairment

Abstract

The Doppel (Dpl) and Prion (PrP) proteins show 25% sequence identity and share several structural features with only minor differences. Dpl shows a PrP-like fold of its C-terminal globular domain and lacks the flexible N-terminal tail. The physiological functions of both proteins are unknown. However, ubiquitous Dpl overexpression in the brain of PrP knockout mice correlated with ataxia and Purkinje cell degeneration in the cerebellum. Interestingly, a similar phenotype was reported in transgenic mice expressing an N-terminally truncated PrP (ΔPrP) in Purkinje cells by the L7 promoter (TgL7-ΔPrP). Coexpression of full-length PrP rescued both the neurological syndromes caused by either Dpl or ΔPrP. To evaluate whether the two proteins caused cerebellar neurodegeneration by the same mechanism, we generated transgenic mice selectively expressing Dpl in Purkinje cells by the same L7 promoter. Such mice showed ataxia and Purkinje cell loss that depended on the level of Dpl expression. Interestingly, the effects of high levels of Dpl were not counterbalanced by the presence of two Prnp alleles. By contrast, PrP coexpression was sufficient to abrogate motor impairment and to delay the neurodegenerative process caused by moderate level of Dpl. A similar situation was reported for the corresponding TgL7-ΔPrP mice supporting the concept that Dpl and ΔPrP cause cell death, possibly by interfering with a common signaling cascade essential for cell survival.

Transmissible spongiform encephalopathies (TSEs) or prion diseases are neurodegenerative conditions including scrapie, bovine spongiform encephalopathy, and Creutzfeldt-Jakob disease (1-3). Mice lacking the cellular prion protein (PrP^C) are resistant to scrapie infection and fail to propagate prions, establishing a central role for PrP in TSE pathology (4-7). PrP^C is a membrane-associated glycoprotein encoded by a single-copy gene (8). It is expressed at high levels in the CNS, preferentially at neuronal synapses (9). Its physiological function remains elusive despite the generation of several PrP knockout mouse lines (reviewed in ref. 10). Two lines in which only the PrP coding sequence was disrupted (11, 12), hereafter named Zürich-I-type mice, behaved and developed normally, showing only minor electrophysiological deficits (13-15) and altered circadian rhythm (16). However, three other PrP knockout lines with deletions extending beyond the PrP ORF, Nagasaki-type mice, presented with progressive ataxia and cerebellar Purkinje cell degeneration (17-19). Because this phenotype was rescued by introduction of a Prnp transgene, it was originally concluded that PrP was crucial for Purkinje cell survival (17, 20). However, subsequent reports proposed that the ataxic phenotype was caused by ectopic expression of the PrP homologue Doppel (Dpl) in brain as a consequence of the gene deletion strategy used to generate the mice (18, 19, 21). Overproduction of Dpl in the CNS was ascribed to an intergenic splicing event between Prnp and the Prnd locus located 16 kb downstream of it, which places Dpl expression under the control of the Prnp promoter. Such intergenic splicing is barely detectable in wild-type animals but is greatly enhanced by the deletion of the splice acceptor site of Prnp exon 3 (18, 21). That the time to onset of disease is inversely correlated to the expression level of Dpl confirmed that its abnormal presence in brain was relevant to the ataxic phenotype (19). Coexpression of full-length PrP (19, 20), but not PrP lacking residues 23-88 (22), rescued the Nagasaki-type mice from Dpl-induced neurodegeneration. Prnp^o/o Zürich-I mice with transgene-driven ubiquitous expression of Dpl in brain developed severe ataxia associated with loss of both granule and Purkinje neurons (23). Coexpression of hamster PrP completely abrogated the cerebellar phenotype of the lower-expressing mouse lines, whereas limited neurodegeneration was seen in the cerebellar cortex of the higher-expressing ones (23). Dpl is an N-glycosylated glycosyl-phosphatidylinositol-anchored protein (24) normally expressed in various peripheral organs but not in brain of adult wild-type mice (18, 25). Its physiological function is unknown, although Dpl knockout mice have defective male gametogenesis (26, 27). Murine Dpl and PrP exhibit 25% sequence identity (18) and share structural features, including three α-helices and one antiparallel β-sheet, with only minor differences (24, 28-30).

We previously reported that PrP knockout mice expressing N-terminally truncated PrP from the so-called half-genomic vector (31) developed a cerebellar phenotype characterized by ataxia and granule cell degeneration (32). It was proposed that Purkinje cells were preserved, whereas granule cells died because expression was driven by a vector inactive in the former but very active in the latter (31, 33). In keeping with this proposal, when expression of truncated PrP was specifically targeted to Purkinje cells, mice showed ataxia and Purkinje cell death (34). In both cases, reintroduction of a PrP gene abrogated the cerebellar syndrome (32, 34). Because the overall structure of Dpl is remarkably similar to that of ΔPrP, and because full length PrP antagonizes both Dpl- and ΔPrP-induced neuronal cell death, it is tempting to assume that the mechanism of pathogenesis might be the same in both ataxic syndromes (10, 18, 19, 23, 34). If so, mice expressing Dpl in Purkinje cells should show the same phenotype as those expressing ΔPrP. We indeed found that such animals developed severe ataxia and Purkinje cell degeneration similarly to the analogous mice expressing ΔPrP (34). The presence of Prnp alleles suppresses the phenotype of the low-expressing transgenic lines but is insufficient to rescue the mice expressing Dpl at high levels.

Materials and Methods

Generation and Maintenance of the L7Dpl Transgenic Mouse Lines. The pL7-Dpl expression vector (Fig. 1) coding for Dpl under the control of the L7 promoter was generated by replacing the PrP ORF by the Dpl ORF in the PrP cDNA and inserting the hybrid cDNA into the unique BamHI site of the L7-pGEM3 plasmid (35). The final pL7Dpl vector was linearized with NotI and EcoRI, purified, and microinjected into fertilized oocytes from wild-type C57BL/6 × 129sv mice. Details are given in Supporting Text, which is published as supporting information on the PNAS web site.

Fig. 1.

L7-Dpl transgene. Shown is a schematic representation of the pL7Dpl expression vector with restriction sites used for cloning and linearization of the plasmid.

Northern Blot and Quantitative Real-Time RT-PCR Analysis. Total RNA was extracted from the cerebellum by using RNAwiz (Ambion, Austin, TX) according to the manufacturer's guidelines. Northern blot analysis was performed as previously described by using a ³²P-labeled Dpl ORF-specific probe (19). For quantitative RT-PCR analysis, 2 μg of cerebellum-extracted total RNA was reverse transcribed by using Omniscript reverse transcriptase according to the manufacturer's guidelines (Qiagen, Chatsworth, CA). One microliter of the resulting cDNA was analyzed by quantitative PCR by using the QuantiTect Probe PCR kit (Qiagen) on an ABI prism 7000 Sequence Detection System (Applied Biosystems). Dpl ORF-containing transcripts were detected by using the Prnp exon 2 primer 5′-PrP RT-PCR (5′-CCAATTTAGGAGAGCCAAGCA) and the Prnd ORF primer 3′-Dpl ORF RT-PCR (5′-GGATGGCCACCCACCAT), with a dual-labeled Dpl-specific probe (5′-TACCCAGCCGGTTCTTCATGATGACTG). The housekeeping gene GAPDH was chosen as a reference, and its mRNA was detected by using primers and probe from TaqMan Rodent GAPDH Control Reagents (Applied Biosystems).

Having confirmed that the amplification efficiencies of both the Dpl-containing sequence (target) and the GAPDH sequence were equivalent, the expression levels of the target were quantified according to the comparative method, which is based on the differences between the threshold cycle (CA) values.

Histopathology, Immunohistochemistry, and in Situ Hybridization.Brains were fixed in 10% formalin, embedded in paraffin, cut into 3-μm sections, and processed for haematoxylin/eosin staining. On selected sections, the following immunostains were carried out according to the manufacturer's instructions: glial fibrillary acidic protein; rabbit polyclonal antibody, 1:1,000; Dako, Carpenteria, CA), NeuN (antineuronal nuclei; mouse monoclonal antibody, 1:2,000; Chemicon), calbindin (Calbindin D-28K; rabbit polyclonal antiserum, 1:100; Chemicon), or the cyclin-dependent kinase inhibitor p27 (p27^Kip1; rabbit polyclonal antiserum, 1:100, Santa Cruz Biotechnology).

For immunohistochemical detection of the Dpl protein, microwave-treated sections were incubated with a rabbit polyclonal anti-Dpl antiserum (1:1,000; raised against recombinant Dpl, a gift from A. Behrens, Cancer Research UK, London). Biotinylated secondary antibodies were used for all staining, and visualization was with a horseradish peroxidase-conjugated streptavidin complex and diaminobenzidine as a chromogen.

Except for Dpl, all other immunostainings were carried out by using the automated Nexus staining apparatus (Ventana Medical Systems Tucson, AZ), following the manufacturer's guidelines. Details on in situ hybridization are given in Supporting Text.

Photographs were obtained on a ColorView II (Helperby, North Yorkshire, U.K.) digital camera mounted on a Zeiss Axioplan microscope and composed in photoshop (Adobe Systems, Mountain View, CA).

Behavioral Assessment. Mice were inspected weekly for behavioral abnormalities, such as disorientation and staggering gait that could rapidly progress to severe ataxia.

Results

Targeted Expression of the Dpl Protein to Purkinje Cells of Transgenic Mice. Transgenic mice expressing Dpl under the control of the Purkinje-cell-specific L7 promoter (35) were generated by pronuclear injection of fertilized oocytes from Prnp^+/+ (C57BL/6 × 129Sv) mice. Offspring positive for the pL7Dpl transgene (Fig. 1) were mated with Prnp^+/+ mice with the same C57BL/6 × 129Sv mixed genetic background to establish transgenic lines (TgL7Dpl). Offspring were characterized in regard to the levels of Dpl-ORF-containing transcripts in the cerebellum. Northern blot analysis revealed that the Tg15 and Tg16 transgenic lines expressed a single Prnd mRNA with about the same intensity, whereas Tg79 and Tg80 showed lower levels of expression (Fig. 6, which is published as supporting information on the PNAS web site). Expression levels were confirmed by quantitative RT-PCR (data not shown). The specificity of Dpl expression from the L7 promoter and the activity of the transgene were confirmed by in situ hybridization (Fig. 2 A-H) and immunohistochemistry (Fig. 2 I-J). All lines showed tightly regulated expression with Prnd-specific mRNA detectable only in cerebellar Purkinje cells (Fig. 2 A-H), whereas all other regions of the brain, in particular the granule cell and molecular layers, remained negative. The lines showing the highest level of Prnd mRNA were used to determine whether the L7 promoter was active during early development. In all lines examined, Dpl mRNA was detectable in the developing brain from postnatal day 1 (P1; Fig. 2 A and B) and continued to be expressed at P8 (Fig. 2 C and D), P15 (Fig. 2 E and F), and into adulthood. The presence of Dpl protein in Purkinje cells of transgenic lines was confirmed by immunohistochemistry (Fig. 2 I and J).

Fig. 2.

Analysis of Dpl mRNA and protein expression in the cerebellum of wild-type and L7Dpl transgenic mice. Digoxigenin in situ hybridization for Dpl in the cerebella of Tg15 mice at P1 (A and B), P8 (C and D), and P15 (E and F) shows the Purkinje cell-specific expression of Dpl mRNA. Note the marked loss of Purkinje cells already at P15 (E, arrowheads). F shows high magnification of an area with partially preserved Purkinje cells. A, C, and E show low magnification, and B, D, and F show high magnification. Controls hybridized with sense probe (P15, G and H) are negative, indicating the specificity of the probe. Shown is immunohistochemical detection of Dpl protein in the cerebellum of L7Dpl transgenic mice (I, arrows pointing to Purkinje cells) but not in wild-type littermate controls (J, arrow pointing to a negative Purkinje cell). [Bars = 260 μm (A), 670 μm (C), 1,400 μm (E and G), and 45 μm (B, D, F, and H-J).]

Neurological Phenotype of Mice Expressing Dpl Specifically in Purkinje Cells. It was previously reported that PrP knockout mice with ectopic expression of Dpl in brain exhibit spontaneous ataxia that can be rescued by introduction of a Prnp wild-type allele (18-20, 23). Surprisingly, hemizygous TgL7Dpl mice such as Tg15/Prnp^+/+ and Tg16/Prnp^+/+ expressing high levels of Dpl in Purkinje cells developed a neurological syndrome characterized by trembling gait and ataxia even on a Prnp wild-type background, showing the first symptoms of motor dysfunction between 4 and 8 weeks of age (Table 1). Tg79 mice with an expression level ≈15% that of the two lines mentioned above were rescued by a single Prnp wild-type allele and remained healthy up to at least 90 weeks of age. However, when the transgenes of Tg79 mice were bred into a Prnp^o/o Zürich-I background, they gave rise to ataxia at ≈22 weeks of age. Animals of the Tg80 line that contained very low levels of Dpl in Purkinje cells showed no signs of disease on either a PrP knockout or a wild-type background within the 97-week period of observation (Table 1). This suggests that the time of appearance of ataxia and the complementation effect by PrP depend on the level of Dpl expression. It was previously reported that expression of PrP from the half-genomic vector (Tga20 transgene) occurred in most areas of the brain with the exception of Purkinje cells (31) and was sufficient to rescue PrP knockout Zürich-II mice from the ataxic syndrome (19). We bred the high-expressing Tg15/Prnp^+/+ and Tg16/Prnp^+/+ animals with Tga20/Prnp^o/o transgenic mice and analyzed the F₁ progeny.

Table 1. DpI mRNA expression levels and onset of clinical phenotype.

	Gene copy number	Expression level*	Age at onset of ataxia†
Mouse line			Prnp+/+ ZH I	Prnp+/0 ZH I	Prnp0/0 ZH I	Prnp+/0 ZH I/Tg20
Tg15	10x	100	4-8 (n = 10)	n/a	n/a	4-8 (n = 10)
Tg16	3x	82	4-8 (n = 10)	n/a	n/a	4-8 (n = 10)
Tg79	2x	14	>97 (n = 7)	>80-90 (n = 10)	≈22 (n = 8)	n/a
Tg80	1x	10	>97 (n = 3)	>68-90 (n = 7)	>57-75 (n = 8)	n/a

n/a, not applicable.

^*DpI mRNA expression level, expressed in percent of levels of tg 15 mice, determined by Northern blot analysis of the cerebellum of all mouse lines. Transgene copy number and development of the phenotype were assessed as described in Materials and Methods.

^†Interval of age in weeks when 100% of the mice analyzed showed the phenotype.

Double transgenic Prnp heterozygous animals showed ataxia similarly to the same L7Dpl lines on a Prnp wild-type background (Table 1). Thus, the deleterious effect exerted by high amounts of Dpl on Purkinje cells was not counteracted by the expression of PrP on both the same and the neighboring cells. Interestingly, the death rate of L7Dpl transgenic mice was not different from that observed in wild-type animals for all lines analyzed at least up to 17 months.

Histopathology of Tg L7Dpl Mice. We examined brains of transgenic and wild-type littermates during early postnatal development and disease progression for histological changes.

The architecture of the cerebellum of Tg15/Prnp^+/+ brains taken at P1 (Fig. 3 A-F), P8 (Fig. 3 G-L), and P15 (data not shown) was indistinguishable from that of wild-type littermates. Moreover, fully differentiated and committed External Granular Layer (EGL) neurons beginning to migrate inward to form the deep fissures and folia associated with the adult cerebellum were detected by immunostaining for the cell cycle inhibitor of p27^kip1. This suggests that development and migration of the Purkinje, molecular, and granule cell layers were normal in both transgenic mice and wild-type littermates. At P8, Purkinje cells began to adopt the laminar arrangement typical of the mature cerebellum (Fig. 3J). At P15, all areas of the cerebellum, including the granule cell layer, were indistinguishable from wild-type littermates, but a moderate loss of Purkinje cells became evident (Fig. 2 E and F). Therefore, Purkinje cell degeneration likely occurs in this mouse line between P8 and P15, and Purkinje cell loss is due to a neurodegenerative process and not to abnormal development. Additional histological analyses at the time of onset of symptoms revealed that at 3 weeks of age, there was a subtotal loss of Purkinje cells in all areas of the cerebellum in Tg15 L7Dpl mice with the exception of foliae XI and X (data not shown). By 6 weeks, Purkinje cell loss (Fig. 4 B and D) was accompanied by reduction of granule cells (Fig. 4 F and H) and severe gliosis (Fig. 4 J and L), as revealed by glial fibrillary acidic protein staining. Interestingly, both granule cell loss and gliosis occurred in a strong gradient across the cerebellum, with anterior sections more severely affected than the posterior half (Fig. 4 F and J). This pattern corresponds to a general gradient laid down in the developing brain, whereby anterior regions develop marginally earlier than posterior ones, and is in keeping with the observation that anterior parts of the cerebellum are generally more susceptible to neurodegenerative processes, especially loss of Purkinje cells (36-38). At 28 weeks, Purkinje and granule cell loss as well as the extent of gliosis were no more severe than at 3 and 6 weeks (data not shown). Because the level of granule cell loss and gliosis at the early time points is minimal, any granule cell reduction seen at later time points is likely secondary to Purkinje cell loss.

Fig. 3.

Normal postnatal cerebellar development in L7Dpl and wild-type mice. Normal foliation and thickness of the EGL in Tg15 L7Dpl and wild-type mice (P1 in A and B and P8 in G and H). At P1, no loss of Purkinje cells as determined by Calbindin immunostaining (C and D) and regular thickness of the external (eEGL) and inner (iEGL) part of the EGL, as identified by p27 immunostaining, which identifies committed postmitotic EGL neurons, in contrast to the still-proliferating p27-negative eEGL (E and F). At P8, no loss of Purkinje cells (Calbindin, I and J) and a normal EGL, split into proliferating eEGL and postmitotic iEGL (K and L); see labels. [Bar = 550 μm (A and B), 950μm (G and H), and 80 μm (C-F and I-L).]

Fig. 4.

Severe loss of Purkinje cells, secondary granule cell degeneration, and gliosis in Tg15 L7Dpl mice at 6 weeks of age. (A-D) Calbindin immunohistochemical staining reveals a subtotal loss of Purkinje cells in L7Dpl transgenic mice, leaving only a small population in Lob. XI and XII (blue arrowheads in B). The red square indicates the area shown at high magnification in D, which is devoid of Purkinje cells. Shown is the control cerebellum with normal density of Purkinje cells at low (A) and high (C) magnification. Concomitant loss of granule cells is demonstrated by immunostaining for NeuN (E-H), which specifically stains cerebellar granule cells. Note the pronounced cell loss in the anterior part of the cerebellum (blue arrowheads). The red squares in E and F indicate the area shown magnified in G and H. In H, the depletion of granule cells (adjacent to a gyrus with more cells) is shown. The cell loss is accompanied by massive gliosis, as shown in I-L, again showing a caudal rostral gradient in its severity (J). Red boxes in I and J indicate the area shown at high magnification in K and L. [Bar = 1,000 μm(A, B, E, F, I, and J) and 100 μm(C, D, G, H, K, and L).]

Cerebellar cortices of the two mouse lines expressing Dpl at lower levels (Tg79 L7Dpl and Tg80 L7Dpl, the first of which has a slightly higher expression than the latter) were examined for histopathological changes. Removal of one or both PrP alleles by intercrossing with Prnp^o/o mice showed that both Tg79/Prnp^+/o and Tg79/Prnp^o/o had a severe loss of Purkinje cells by 4 weeks (Fig. 5 A and B) in a pattern consistent with the loss seen in the high-expressing lines on the Prnp^+/+ background. Interestingly, the level of granule cell degeneration and gliosis was much less extensive in these lines than in the high-expressing lines. Loss of granule cells and gliosis was observed in cerebella of Tg79/Prnp^o/o but not Tg79/Prnp^+/o at 20 weeks (data not shown). The presence of two Prnp alleles (Tg79/Prnp^+/+) had a remarkable effect on the loss of Purkinje cells in that mice showed a full rescue of the neurodegeneration at 4 weeks and a normal laminar arrangement of Purkinje cells (Fig. 5C). Both the molecular and granule cell layers remained indistinguishable from wild-type controls. However, at 72 weeks of age, the loss of Purkinje cells became evident in all areas of the cerebellum, indicating that Purkinje cell degeneration was occurring at low rate despite complementation with PrP alleles (Fig. 5D).

Fig. 5.

Complementation of Dpl-induced Purkinje cell loss by restoring PrP expression. In the higher-expressing line Tg79 (A-D), introducing one allele of Prnp leads to a partial rescue of Dpl-induced neurotoxicity (B), which is complete after introduction of a second copy of Prnp (C) at 4 weeks. However, long-term coexpression of the Dpl transgene and two Prnp alleles still causes neurodegeneration in Tg79 L7Dpl mice. Tg80 L7Dpl transgenic mice (E-H), expressing Dpl at lower levels than Tg79 mice show little Purkinje cell loss at 10 weeks (E), which can be fully compensated by introducing one (F) or two (G) Prnp alleles after 10 and 72 weeks (H). Each photograph shows a representative segment of Purkinje cells below the respective overview. [Bar = 1,000 μm (overview) and 80 μm (detail).]

Histological analysis of Tg80, the lowest Dpl-expressing line, showed Purkinje cell loss at ≈10 weeks of age on a Prnp^o/o background (Fig. 5E). However, this was not accompanied by any loss of granule cells or reactive glia. At the same time point, Tg80/Prnp^+/o and Tg80/Prnp^+/+ mice showed no degeneration or loss of Purkinje cells (Fig. 5 F and G). By 72 weeks, Tg80/Prnp^+/+ mice continued to show no neuronal loss, suggesting that a low level of Dpl expression in the Purkinje cells can be fully counteracted by PrP coexpression (Fig. 5H).

Discussion

We recently reported that transgenic mice overexpressing N-terminally truncated PrP (PrP Δ32-134) from the L7 promoter developed a cerebellar syndrome similar to that described for the Nagasaki-type PrP knockout mice with ectopic expression of Dpl in brain (34). This finding suggested that truncated PrP and Dpl may cause degeneration by the same mechanism, possibly by interfering with a common signaling pathway essential for cell survival. To further evaluate this hypothesis and to discount the possibility that the ataxic syndrome observed in the Nagasaki-type Prnp^-/- lines was caused by deregulation of other genes rather than by ectopic expression of Dpl, we generated transgenic mice expressing the Dpl protein in Purkinje cells. Interestingly, its presence exclusively in this neuronal cell type triggered a phenotype similar to that caused by ΔPrP, namely trembling gait rapidly progressing to ataxia accompanied by Purkinje cell loss. The extent of Purkinje cell degeneration was dose-dependent, with high-expressing lines showing rapid Purkinje cell death between P8 and P15, whereas lower-expressing lines had a delayed onset of Purkinje cells loss of up to 10 weeks. Importantly, initial neuronal loss was restricted to Purkinje cells, in concordance with the specificity of the L7 promoter, whereas granule cell death and gliosis were detected within the cerebellum only later, likely as a consequence of the absence of critical inputs from Purkinje cells rather than of a direct effect of the transgene.

Previous studies reported that both ΔPrP- and Dpl-associated ataxia could be rescued by coexpression of full-length PrP (19, 20, 23). Similarly, Dpl-mediated ataxia resulting from moderate but not a high level of Dpl expression in Purkinje cells could be abrogated by the presence of one or two Prnp alleles. Histological analysis of L7Dpl cerebellar cortex expressing medium levels of Dpl on a Prnp wild-type background revealed delayed Purkinje cell degeneration when compared to L7Dpl/Prnp^o/o brains, rather than full rescue. Interestingly, a similar situation was reported for the corresponding L7-ΔPrP mice (34), suggesting that, whereas PrP is sufficient to abrogate motor impairment associated with expression of ΔPrP or Dpl, it is unable to completely suppress neuronal loss.

Our findings support the view that Dpl and ΔPrP cause cell death by the same mechanism, perhaps by interfering with a cellular pathway normally controlled by full-length PrP. The molecular events responsible for Dpl- and ΔPrP-mediated toxicity remain unclear; however, several hypotheses can be postulated. That the N-terminal region of PrP is implicated in the internalization of the protein during constitutive endocytosis (39) suggests that its deletion may alter the intracellular trafficking and possibly the physiological function of the prion protein. Such a view could be extended to Dpl because of its structural similarities to ΔPrP. Moreover, several lines of evidence suggest that PrP has a protective role against apoptosis. It was reported that PrP^C inhibits Bax-induced cell death in primary neuronal culture, whereas deletion of four octapeptide repeats abolishes this beneficial effect (40). Serum withdrawal induced apoptosis in a hippocampal cell line derived from Nagasaki-type Prnp^-/- mice but not in cells derived from Prnp^+/+ animals (41). Signaling through PrP was reported to protect retinal neurons against anisomycin-induced apoptosis (42). Further, it has been reported that Nagasaki-type PrP knockout mice showed an increase in heme oxygenase 1 and nitric oxide synthase levels, suggesting enhanced oxidative stress concomitant with Dpl expression (43). The proposal that PrP is a superoxide dismutase and therefore able to remove free radical species (44) would explain its beneficial effect in counterbalancing Dpl toxicity; however, this role of PrP is disputed (45-47). Conversely, it has been reported that PrP facilitates apoptosis (48, 49).

Our results definitely confirm remarkably similar features between the phenotypes elicited by Dpl and PrP lacking the flexible N-terminal tail, the mechanism by which the two proteins cause selective neuronal cell death in the cerebellum of transgenic mice remains unknown. Identification of Dpl/ΔPrP-binding partners or downstream signaling events may provide information about the pathways involved and help elucidate the physiological role of the cellular PrP.