In Vivo Cytotoxicity of the Prion Protein Fragment 106–126*

Transmissible spongiform encephalopathies are fatal neurological diseases characterized by astroglyosis, neuronal loss, and by the accumulation of the abnormal isoform of the prion protein. The amyloid prion protein fragment 106–126 (P106–126) has been shown to be toxic in cultured hippocampal neurons (1). Here, we show that P106–126 is also cytotoxic in vivo. Taking advantage of the fact that retina is an integral part of the central nervous system, the toxic effect of the peptide was investigated by direct intravitreous injection. Aged solutions of P106–126 induced apoptotic-mediated retinal cell death and irreversibly altered the electrical activity of the retina. Neither apoptosis nor electroretinogram damages were observed with freshly diluted P106–126, suggesting that the toxicity is linked to the aggregation state of the peptide. The retina provides a convenient in vivo system to look for potential inhibitors of cytotoxicity associated with spongiform encephalopathies.

Transmissible spongiform encephalopathies are fatal neurological diseases characterized by astroglyosis, neuronal loss, and by the accumulation of the abnormal isoform of the prion protein. The amyloid prion protein fragment 106 -126 (P106 -126) has been shown to be toxic in cultured hippocampal neurons (1). Here, we show that P106 -126 is also cytotoxic in vivo. Taking advantage of the fact that retina is an integral part of the central nervous system, the toxic effect of the peptide was investigated by direct intravitreous injection. Aged solutions of P106 -126 induced apoptotic-mediated retinal cell death and irreversibly altered the electrical activity of the retina. Neither apoptosis nor electroretinogram damages were observed with freshly diluted P106 -126, suggesting that the toxicity is linked to the aggregation state of the peptide. The retina provides a convenient in vivo system to look for potential inhibitors of cytotoxicity associated with spongiform encephalopathies.
Transmissible spongiform encephalopathies (TSE) 1 are neurological disorders having common pathological signs like vacuolization of the neurophils, astrocytosis, and loss of neurons (2). TSE are characterized by the accumulation of large aggregates of the prion protein, PrP-res, in both human and animal brains. The abnormal PrP-res isoform has a high content of ␤-sheet secondary structure, forms amyloid fibrils, and is partially resistant to proteolysis (3)(4)(5). The precursor of the pathological isoform PrP-res is a normal host glycoprotein (termed PrP-sen) widely expressed in the central nervous system and in peripheral tissues. Unlike PrP-res, PrP-sen is a protease-sensitive protein, soluble in mild detergents, and has mainly ␣-helix structures (6 -8). Interestingly, the development of spongiform pathologies requires the presence of both isoforms PrP-res and PrP-sen (9), the level of PrP-sen expression being an essential factor in the pathology (10). In experimental transmis-sions with transgenic mice, it was found that increasing the number of PrP gene copies led to a decreased incubation time of the disease (11,12).
The in vivo mechanisms leading to the synaptic loss and neuronal death are not elucidated, but the possible involvement of apoptotic processes has been postulated and is consistent with the lack of inflammation stigmata in TSE-affected brains. The neural tissue damages could be due exclusively to PrP-res deposits, and/or to disruption of PrP-sen physiological functions, and/or to cytotoxic effects induced by PrP fragments or PrP-res aggregates. PrP-res and a synthetic peptide fragment (P106 -126) have been shown to be cytotoxic in vitro likely via the programmed cell death pathway (1,13). Like PrP-res, the P106 -126 peptide is partially resistant to proteinase K, presents a high ␤-sheet enriched structure, and forms amyloid fibrils in vitro (14). Hope et al. (15) have shown that the expression of PrP-sen is necessary for P106 -126 to exert its cytotoxic effects.
Whether the apoptotic process could account for in vivo neurodegeneration taking place during the development of TSE remains unclear. The induction of apoptosis has been suggested in scrapie-infected mouse brain (16), as well as in Creutzfeldt-Jakob disease (17) and bovine spongiform encephalopathy (18).
In the present report, we examined whether the PrP fragment P106 -126 can induce apoptotic-mediated cell death in vivo. To address this question, we directly administrated P106 -126 into rat eyes taking advantage of the fact that the retina is an integral part of the central nervous system. Apoptotic process induced by P106 -126 was demonstrated by means of the terminal deoxynucleotidyl transferase dUTP-end labeling (TUNEL) staining and of the DNA fragmentation technique. Neuronal death was confirmed by measurement of the electrical activity of the retina.

EXPERIMENTAL PROCEDURES
Peptides-The human sequence of the prion protein fragment (P106 -126; KTNMKHMAGAAAAGAVVGGLG) was purchased from Bachem. The prion fragment used as a control peptide (P89 -103; GQGGGTH-NQWNKPSK) was synthesized by Eurogentec. Lyophilized peptides were dissolved in deionized water at a concentration of 5 mM, distributed into 20-l aliquots, and stored at Ϫ20°C. Peptides were diluted twice with 2ϫ PBS minus Ca 2ϩ and immediately used or aged by incubation at room temperature for 3 days.
Intravitreal Injection-Adult male albino Wistar rats (35-55 days old) were anesthetized with an intraperitoneal injection of 60 mg/kg sodium pentobarbital followed by a topical application of 0.4% oxybuprocaine hydrochloride. The injections of 2 l of peptides or vehicle (phosphate-buffered solution) were done unilaterally with a 30-gauge needle introduced into the posterior chamber on the upper pole of the eye directed toward the center of the vitreous. The injections were performed slowly to allow a better diffusion of the peptide and avoid any ocular hypertension. For negative controls, we both considered the uninjected contralateral eyes and eyes injected in the same conditions with 2 l of vehicle (PBS). At least four animals were used in each group of experimental conditions.
In Situ Labeling by the TUNEL Method-Rats were euthanatized with an overdose of sodium pentobarbital 3 days post-intravitreal injections. The eyes were enucleated, and a puncture was made at the limbus to permit the infusion of the fixative solution. The eyes were immediately fixed in ice-cold 4% paraformaldehyde in PBS for 4 h then cryoprotected overnight in PBS containing 20% sucrose and embedded in Tissue-Tek ® (Sakura). Frozen sections (10 m) were cut on a cryostat (Leica), and slides were heated at 50°C for 60 min then stored at Ϫ20°C until use. The in situ cell death detection was performed fol-lowing the manufacturer's recommendations (Roche Molecular Biochemicals) then revealed using a 3,3Ј-diaminobenzidine (DAB) substrate kit (Vector Laboratories) and processed for detailed examination by light microscopy. Morphological and histological observations of retinas were done by staining the sections with 1% cresyl violet.
Dose-Response of the Effect of PrP P106 -126 -2 l of PBS alone or of 3-day-aged peptide P106 -126 diluted extemporarily in PBS (corresponding to 0.05, 0.5, and 5 nmol) were injected as described above. For each condition, four rats were treated, and retina sections were processed simultaneously for the TUNEL-labeling assay. To evaluate the number of TUNEL-positive cells, photographs of retinal sections were collected at an original magnification of ϫ 10. The intensity of gray of each layer (ganglion cell, G; inner nuclear, INL; outer nuclear, ONL) was quantified using the NIH image software system. The gray intensity was expressed per surface unit in arbitrary units.
DNA Fragmentation Analysis-After enucleation, the retinas were dissected from the retinal epithelium and choroid. Proteins were then digested with 200 g/ml of proteinase K in lysis buffer (10 mM Tris, pH 7.4, 5 mM EDTA, 1% SDS) for 2 h at 55°C. Retinal DNA was extracted by phenol-chloroform solutions, and the aqueous phase was incubated with Dnase-free RNase A (100 g/ml) and then precipitated in a solution of 0.3 M sodium acetate in ethanol overnight at Ϫ20°C. Precipitated DNA samples were resuspended in distilled water, and the DNA concentration was determined by measuring the absorbance at 260 nm. Samples of 10 g of DNA were electrophoresed through 1.2% agarose gel containing 1 g/ml of ethidium bromide. DNA bands were visualized by a UV light transilluminator and photographed.
Electroretinograms (ERGs)-Dark-adapted ERGs were performed as described previously (19). Full-field ERG responses were obtained overnight with dark-adapted rats prepared under dim red light before recording. The pupils of anesthetized rats were dilated with a drop of 0.5% tropicamide. A silver chloride ring-recording electrode was placed on the cornea, and the reference electrode, with a silver-silver chloride tip, was connected to the ear. Light stimulus (10 s) was provided by a stroboscopic flash (Grass Instruments Inc.) placed 0.25 m in front of the rat. The ERGs were recorded using the visual electroretinogram test system (LKC Technologies, Inc.) and then stored and analyzed. Amplitude of the a-wave was measured from the base line to the trough of the a-wave; b-wave amplitude was measured from the trough of the a-wave to the peak of the b-wave. The averaged responses represent the mean of five white flashes delivered four min apart. Electroretinograms were recorded before treatment and then at different times of recovery (1, 2, 3, and 7 days).
Statistical Analysis-The amplitude values of a-and b-waves obtained with each animal were normalized to base-line amplitudes at each time point of the study and expressed as percentage of base-line value. Data were expressed as mean Ϯ S.E. of values obtained with 4 rats per group. Results from the ERG recordings were not distributed normally and had unequal variance. Therefore, the Mann Witney U test was used to compare the peptide and the vehicle-treated group of rats. Differences were considered statistically significant when P values were less than 0.02.

In Vivo Induction of Retina Cells Apoptosis by Peptides-
The effect of freshly prepared or 3-day-aged P106 -126 and soluble P89 -103 PrP peptides on the retina cell death was examined after intravitreal injection. Fig. 1 shows representative photomicrographs of treated retina sections processed through the TUNEL labeling assay (Fig. 1, B-E) and the histology of PBStreated retina stained with cresyl violet (Fig. 1A). After intravitreal injections of aged P106 -126 (5 nmol), TUNEL-positive cells were observed in the inner and outer nuclear layers, as well as in the ganglion cells layer (Fig. 1D). When freshly prepared P106 -126 was intravitreally injected, only occasional TUNEL-positive cells were observed (Fig. 1C) suggesting that the cytotoxic effect of the peptide was linked to its physicochemical state. Intact retinas (data not shown) and retinas treated with PBS alone (Fig. 1B) or soluble PrP peptide P89 -103 (Fig. 1E) showed only few TUNEL-positive cells. No differences in labeling patterns were observed between the lower and the upper poles of the retina where injections were done.
Dose-Response Effect of PrP Peptide P106 -126 on TUNEL Labeling-To assess whether the cytotoxic effect was dose-de-pendent, rat eyes were treated for 72 h with PBS alone or containing 0.05, 0.5, or 5 nmol of aged P106 -126. Corresponding retina sections were processed through the TUNEL-labeling assay and revealed with the DAB substrate. We previously checked that the intensity of gray coloration developed by the DAB substrate was directly proportional to the number of TUNEL-positive cells. This estimation method allowed us to screen a large surface layer and an important number of retinal sections. Fig. 2 is a histogram representation of the relative gray intensity obtained for the inner and outer nuclear layers and the ganglion cells layer. The gray intensity of the three layers increases with the quantity of injected P106 -126. Although 0.05 nmol of injected P106 -126 was not significantly toxic for the retina cells, higher quantities induced TUNELlabeled positive cells in ganglion cells and in inner and outer nuclear layers. When 5 nmol of P106 -126 were injected, the TUNEL-labeling method showed that almost all the nuclei from the three layers were positive.
DNA Fragmentation Kinetics-To better discriminate between necrosis and apoptotic cell death, the fragmentation of the retinal DNA was studied. Four eyes were injected with vehicle alone or with 5 nmol of aged peptide P106 -126 and analyzed at different times after inoculation. The peptidetreated retinas showed a ladder pattern of DNA degradation with bands corresponding to multiples of 180 to 200 base pairs (Fig. 3). The DNA laddering was clearly but slightly observable 24 h after injection and was more intense after 48 and 72 h. No DNA degradation was observed either in the PBS-treated retinas (Fig. 3, control lane) or in retinas treated with freshly prepared peptide P106 -126 (data not shown).
Electroretinographic Studies-The electrical activity of peptide-treated or untreated retinas was monitored by recording the ERGs. Fig. 4 shows the effect of intravitreal injections on the electroretinograms and describes the variation of the a-and b-wave amplitudes several days after inoculation. When compared with vehicle-treated group, the aged P106 -126-treated group exhibited significant deficits in both a-and b-wave amplitudes, direct evidence of persistent and long-term damages to retinal function. In the peptide-injected group, the a-wave amplitude decreased significantly at day 1 (71 Ϯ 8%, p Ͻ 0.02) and continued to decline slowly from day 2 to day 7. Unlike the a-wave, the b-wave amplitude markedly decreased at day 1 (57 Ϯ 8%) and then partly recovered at day 2 (78 Ϯ 9%) and further decreased from day 3 (63 Ϯ 6%) to day 7 (56 Ϯ 7%). When electroretinogram recordings were done after injection of control peptide P89 -103 or freshly prepared P106 -126, no significant decrease on the a-and b-wave amplitudes could be observed (data not shown). In Vivo Cytotoxicity of Peptide P106 -126 36488 DISCUSSION For the last decade, many reports have shown that PrP peptide P106 -126 could provide insights into the understanding of some aspects of TSE pathology (1,15,20,21). Although P106 -126 does not cause infection or induce in vitro conversion of PrP-sen to PrP-res isoform, it shares a number of physicochemical properties with the PrP-res molecule. For instance, P106 -126 is partially proteinase K-resistant and displays high ␤-sheet structure (14). Amyloid fibrils formed in vitro by P106 -126 are birefringent under polymerized light when stained by Congo red (22). Like PrP-res, which causes cell death in vitro (13), P106 -126 can induce apoptotic cell death (1,23). In addition, P106 -126 exerts its toxic effects through direct and specific interactions with PrP-sen-expressing cells (20,24).
In contrast with in vitro observations, neurotoxicity of P106 -126 has never been reported after direct administration into the brain. Several explanations can be proposed for this discrepancy, including the rapid clearance of the peptide by the brain. Supporting this idea, PrP-res is detectable in the brain of intracerebrally infected mice a few days after infection, then it decreases to undetectable levels, and last the accumulation of the endogenous PrP-res starts (25). For several reasons, the retina seems to be an interesting in vivo model. First, retina is an integral part of the central nervous system; second, it is a closed system with a low content of proteolytic enzymes. Furthermore, the PrP-sen isoform is expressed in photoreceptor cells of adult retina (26). Finally, widespread retinopathy has been observed in scrapie-affected hamsters (27)(28)(29). Therefore, the retina is a well suited model to study the molecular events leading to neuronal death in the course of the disease. For the first time, we show that aged P106 -126 induces the death of retinal cells according to an apoptotic process and provides a significant reduction of the electrical activity of the retina. The observed effects of the peptide are monitored by DNA fragmentation and electroretinogram recording analysis and can be evidenced as soon as 24 h after intravitreal injection. Moreover, P106 -126 causes long-term damages on the physiological activity of retina, because no recovery of the deleterious effects can be observed 7 days after peptide inoculation.
The molecular mechanisms involved in the P106 -126 toxicity are not completely understood. The relationship between toxicity and physicochemical properties of amyloid peptides has been extensively studied in vitro (30,31). The ␤-sheet structure adopted by P106 -126 and its aggregation state are crucial (1,15). Our results are consistent with these previous reports, because intravitreous injections of freshly diluted P106 -126 failed to cause cell death and modification of electrical activity of the retina. Moreover, Jen et al. (32) have shown recently that the amyloid peptides A␤1-40 and A␤1-42 involved in Alzheimer's disease also induced photoreceptor cells apoptosis after intravitreous injections. Here again, cell death was observed only with aged (aggregated) peptides, whereas freshly prepared solutions were devoid of cytotoxicity. As a positive control to our own experiments, the apoptotic response induced by aged solutions of A␤1-40 have been reproduced in the present work (data not shown). There exists no sequence homology between A␤1-40 and P106 -126 peptides. However, both peptides share the same propency to form ␤-sheet structures and sedimentable aggregates. Taken together these data strongly suggest that the specific mode of the aggregation of P106 -126, rather than its amino acid sequence, is responsible for the observed in vivo cell death.
Apart from programmed-cell death, other molecular mechanisms of P106 -126 toxicity have been proposed but are still controversial (33). The possible insertion of the peptide into the cell membrane could lead to channel formation, the subsequent disruption of ion homeostasis being responsible for the cell death (34,35).
Our results strongly suggest that at least part of the neuronal loss observed in TSE diseases is because of the toxic effects of PrP-res and/or its degradation products. In contrast with Alzheimer's disease in which the overproduction of the amyloid peptides A␤1-40 and A␤1-42 is well documented, the catabolism of the PrP molecules remains unclear. An amino-terminally truncated fragment of PrP insoluble in detergents and resistant to protease has been purified from brains of FIG. 2. Dose-response histograms of the cytotoxic effect induced by the PrP peptide P106 -126 on retina cells. Rat eyes were intravitreally injected with 2 l of PBS alone or containing 0.05, 0.5, and 5 nmol of 3-day-aged peptide P106 -126. 72 h after injection, rats were sacrificed, and the retinas were fixed and processed by the TUNEL-labeling method as described under "Experimental Procedures." Each group represents four animals injected unilaterally; ten sections were cut per eye. The relative intensity of gray per surface unit was determined in a 10-fold magnification photograph of retina sections and quantified with the NIH image software. The data represent the means Ϯ S.E. of two independent experiments expressed in arbitrary units. G, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer.

FIG. 4. Electroretinogram profiles (A) and a-and b-wave amplitudes (B) in PrP peptide P106 -126 and PBS-injected groups.
ERGs were recorded before (0) and after 1, 2, 3, and 7 days of recovery. The black and gray marks correspond to the intravitreal injections of vehicle alone and of aged P106 -126 (5 nmol), respectively. The a-and b-wave amplitudes were normalized and expressed as the percentage of the control value. Each bar and error represents the mean Ϯ S.E. (n ϭ 6). *, statistically significant differences between control and peptidetreated groups (p Ͻ 0.02 versus control).
Creutzfeldt-Jakob disease affected patients (36). Caughey et al. (37) have also shown that a PrP-res isoform truncated at residue 90 is formed in scrapie-infected mouse neuroblastoma cells. Furthermore, fragments of PrP-sen have been identified in extracts of normal brain, and the precise cleavage site has been mapped to His-111 or Met-112 (36). Interestingly, the cleavage occurring in physiological conditions disrupts the neurotoxic region of PrP-sen comprising amino acid residues 106 -126. Thus, pathological conditions could promote the appearance of PrP toxic fragments that are never formed in physiological situation.
In conclusion, our study provides the first in vivo evidence of the involvement of an apoptotic process in the cytotoxic mechanisms of the prion protein fragment P106 -126 suggesting that similar mechanisms also occur in TSE. The retina model used in the present work is a powerful tool to further identify the molecular mechanisms leading to the neuronal loss that could be a general characteristic of several neurodegenerative diseases. Finally, the retina model should serve for screening drugs able to limit cell damage and prevent neuronal loss.