Glycosylation of PrPC Determines Timing of Neuroinvasion and Targeting in the Brain following Transmissible Spongiform Encephalopathy Infection by a Peripheral Route

ABSTRACT Transmissible spongiform encephalopathy (TSE) infectivity naturally spreads from site of entry in the periphery to the central nervous system where pathological lesions are formed. Several routes and cells within the host have been identified as important for facilitating the infectious process. Expression of the glycoprotein cellular PrP (PrPC) is considered a key factor for replication of infectivity in the central nervous system (CNS) and its transport to the brain, and it has been suggested that the infectious agent propagates from cell to cell via a domino-like effect. However, precisely how this is achieved and what involvement the different glycoforms of PrP have in these processes remain to be determined. To address this issue, we have used our unique models of gene-targeted transgenic mice expressing different glycosylated forms of PrP. Two TSE strains were inoculated intraperitoneally into these mice to assess the contribution of diglycosylated, monoglycosylated, and unglycosylated PrP in spreading of infectivity to the brain. This study demonstrates that glycosylation of host PrP has a profound effect in determining the outcome of disease. Lack of diglycosylated PrP slowed or prevented disease onset after peripheral challenge, suggesting an important role for fully glycosylated PrP in either the replication of the infectious agent in the periphery or its transport to the CNS. Moreover, mice expressing unglycosylated PrP did not develop clinical disease, and mice expressing monoglycosylated PrP showed strikingly different neuropathologic features compared to those expressing diglycosylated PrP. This demonstrates that targeting in the brain following peripheral inoculation is profoundly influenced by the glycosylation status of host PrP.

Transmissible spongiform encephalopathies (TSE) or prion diseases are a group of fatal neurodegenerative diseases which include Creutzfeldt-Jakob disease (CJD) in humans, scrapie in sheep and goats, bovine spongiform encephalopathies (BSE) in cattle, and chronic wasting disease (CWD) in deer and elk (30). These diseases can be sporadic, familial, or acquired by infection, and the common hallmark is a distinct pathology in the central nervous system (CNS) characterized by neuronal loss, spongiform degeneration, and gliosis (38,46).
Expression of the host-encoded cellular PrP (PrP C ) is fundamental for the onset of disease since PrP-deficient mice are refractory to TSE infection (11,31). PrP C is a glycoprotein with two consensus sites for attachment of N-linked glycans (at codons 180 and 196 in the mouse) which are variably occupied, producing di-, mono-, and unglycosylated PrP (43). The diversity in glycosylation, combined with the complexity of added sugars, results in a large number of glycosylated forms of PrP (41). A central event associated with TSE infection is the conformational conversion of PrP C into an abnormal protease-resistant form, PrP Sc (39). PrP Sc is deposited in brain and, in some but not all cases, in peripheral organs of individuals affected by TSE (21).
Although the pathology associated with TSE is found in the brain, the periphery is the most natural route of acquiring infection. Evidence suggests that oral transmission via contaminated food is linked with transmission of BSE to humans, resulting in variant CJD (vCJD) (10,47), and blood transfusion has been identified as a probable route of human-to-human transmission of vCJD (23,27,36). Moreover, parenteral administration of contaminated human tissue-derived therapeutics has been shown to facilitate iatrogenic spread of these diseases (8,46). It is therefore important to understand the mechanisms that allow the infectious agent to propagate in the periphery and be transported to the CNS prior to the onset of neurodegeneration in the brain.
Many studies have been conducted to understand routes of transmission (for a review see references 1 and 29). Lymphoid tissues such as the spleen have been shown to play a fundamental role in agent replication and propagation in the very early stages of disease. Indeed, studies of splenectomized and asplenic mice have shown the lymphoreticular system (LRS) to be an important site for TSE agent replication (14,26). The periphery also appears to have a role in processing the infectious agent following intracerebral (i.c.) inoculation as PrP Sc accumulates in the spleen shortly after inoculation and before accumulation of the abnormal protein in the brain (15,17). Within the LRS, follicular dendritic cells (FDC) have been shown to be important for the uptake of infectivity and subsequent spreading toward the CNS (7,28,33,35). Several studies have also suggested the peripheral nervous systems (PNS) as a potential route of infectivity to the brain, implicating the vagus and sciatic nerves in this process (5,20,25,34).
Expression of PrP C in the peripheral tissues appears to be an important prerequisite for the transport of infectivity to the CNS following peripheral routes of inoculation. Indeed, it has been proposed that a continuous chain of cells expressing PrP C is fundamental for TSE neuroinvasion (6,40), with overexpression of endogenous PrP in the PNS greatly facilitating the spread of infectivity (19). Thus, host PrP appears to have a fundamental role in the uptake, transport, and replication of the infectious agent (6). Moreover, it has been suggested that the different PrP C glycoforms may influence the timing of neuroinvasion by directly influencing the interaction with the infectious agent (19). However, the mechanism by which the different glycoforms are involved in these processes remains to be determined.
In order to investigate the role of PrP C glycosylation in TSE disease after peripheral infection with different TSE strains, we have used our inbred gene-targeted transgenic mice expressing different glycosylated forms of PrP. These mice expressed PrP with no sugars at the first (designated G1/G1 in homozygous mice) or the second glycosylation site (G2/G2) or both (G3/ G3) under the control of the endogenous PrP promoter (13). We have previously shown that following intracerebral inoculation, all glycotypes are susceptible to infection with at least one TSE strain and that the type of PrP glycosylation in the host influenced the incubation period but not the distribution of pathological lesions in the brain (45). Here, we examine the influence of host PrP glycosylation on the peripheral acquisition of infection and demonstrate that, unlike the intracerebral route, mice without PrP glycosylation were resistant to disease and that the different glycoforms had a profound influence on not only the timing of disease but also the type and distribution of the PrP Sc deposits in the brain.

MATERIALS AND METHODS
Transgenic mouse lines. Inbred gene-targeted transgenic mouse lines G1, G2, and G3, and the corresponding inbred 129/Ola wild-type control line have been described previously (13). Edinburgh PrP null mice (31) were used as negative controls.
Genotyping of mouse tail DNA. All the transgenic mice used in this study were genotyped twice: before inoculation and at the end of the experiments. A portion of tail was removed from each mouse. DNA was prepared from a 1-cm piece of tail by digestion overnight at 37°C in tail lysis buffer (300 mM sodium acetate, 1% SDS, 10 mM Tris, pH 8, 1 mM EDTA, 200 g/ml proteinase K) and subsequent extraction with an equal volume of phenol-chloroform. DNA was precipitated with isopropanol, washed with 70% ethanol, and resuspended in 100 l of TE buffer (10 mM Tris, 1 mM EDTA, pH 7.4). The mismatch PCR method to identify the different transgenics has been described elsewhere (13).
Preparation of inoculum and injection. Inocula were prepared from the brains of C57BL mice with terminal ME7 or 79A (mouse-adapted scrapie strains) TSE disease. A 1% homogenate of each sample was prepared in sterile saline prior to use as an inoculum. All experimental protocols were submitted to the Local Ethical Review Committee for approval before mice were inoculated. All experiments were performed under license and in accordance with the United Kingdom Home Office Regulations under the Animals (Scientific Procedures) Act, 1986.
Scoring of clinical TSE disease. The presence of clinical TSE disease was assessed as described previously (16). Animals were scored for clinical disease without reference to the genotype of the mouse. Genotypes were confirmed for each animal by PCR analysis of tail DNA at the end of the experiment. Incu-bation times were calculated as the interval between inoculation and culling due to terminal TSE disease. Mice were killed by cervical dislocation at the terminal stage of disease, at termination of the experiment (between 500 and 700 days), or for welfare reasons due to intercurrent illness.
Lesion profiles. Sections were stained with hematoxylin and eosin and scored for vacuolar degeneration on a scale of 0 to 5 in nine standard gray matter areas and on a scale of 0 to 3 in three standard white matter areas as described previously (16).
Immunohistochemical analysis in brain. Sections were stained for disease associated PrP using monoclonal antibody 6H4 (1/1,000; Prionics). Antigen retrieval by autoclaving at 121°C for 15 min and 10 min in formic acid (98%) was used to facilitate detection of the antigens. Sections were then blocked with normal serum prior to incubation with the primary antibody. Antibody binding was detected with a catalyzed signal amplification system (Dakocytomation) and visualized with diaminobenzidine (DAB). In all the experiments normal brain homogenate inoculum and PrP null negative controls were used. Images were taken using a Nikon Eclipse E800 microscope. Sections were analyzed by an observer blinded to the mouse genotype and type of inoculum.
PrP Sc extraction from spleens. Mice were killed by cervical dislocation; spleens were removed, flash frozen in liquid nitrogen, and stored at Ϫ70°C until required. Whole spleens were weighed, and PrP Sc was extracted using the method of centrifugal concentration from detergent solution (15). The spleen was homogenized in 3 ml of 0.2 M potassium chloride and 20 l of 100 mM phenylmethylsulfonyl fluoride (PMSF) and centrifuged at 1,500 rpm for 10 min, and the subsequent supernatant was centrifuged at 50,000 rpm for 30 min, all at 4°C. This pellet was resuspended in 2 ml of 0.1 M Tris-hydrochloric acid (Tris-HCl, pH 7.4), and the suspension was divided into two equal parts: one part was left at 4°C with 40 l of 100 mM PMSF for samples that were not digested with proteinase K (ϪPK; Sigma), and the other part was incubated with 3 l of 18 mg/ml PK at 37°C for the samples digested with PK (ϩPK), both for 1 h. Twenty microliters of PMSF, 1 ml of 2% sarcosyl, and 2 l of ␤-mercaptoethanol were added, and the samples were incubated at 37°C for an additional 1 h. The samples were then overlaid on 200 l of 20% sucrose and centrifuged at 50,000 rpm for 2 h, and the pellets were stored at Ϫ20°C.
Western blotting. Mice were killed by cervical dislocation, and brains and spleens were removed, flash frozen in liquid nitrogen, and then stored at Ϫ70°C until required. Spleen preparation has already been described above. Brain homogenates (10%, wt/vol) were prepared in NP-40 lysis buffer (1% Nonidet P-40, 0.5% sodium deoxycholate, 150 mM NaCl, 50 mM Tris-HCl, pH 7.5, and 1 mM PMSF). The homogenate was centrifuged at 16,000 rpm for 10 min at 4°C, and supernatant was isolated. Total protein was denatured in 1ϫ Novex Trisglycine SDS sample buffer (Invitrogen Life Technologies) and 1ϫ NuPage sample reducing agent (Invitrogen Life Technologies) for 30 min at 95°C. Proteins were separated by gel electrophoresis at 125 V using 12% Novex Tris-glycine gels (Invitrogen Life Technologies). Proteins in the acrylamide gel were transferred to polyvinylidene difluoride (PVDF) membrane at 2 mA/cm 2 of gel using a semidry transfer blotter (Bio-Rad) in 1ϫ transfer solution (48 mM Tris, 39 mM glycine, 0.375% SDS, 20% methanol). Presence of PrP was assessed using the anti-PrP monoclonal antibodies 8H4 or 7A12 (1/10,000; kind gift of M. S. Sy) (48).
Immunohistochemical analysis of spleens. Tissues were flash frozen in liquid nitrogen and embedded in optimal cryotomy temperature (OCT) compound prior to sectioning on a Leica cryostat. Serial sections (10 m) were cut and air dried overnight on SuperFrost Plus slides. Sections were fixed in acetone for 10 min. All labeling was done at room temperature in a humid chamber. For light microscopy, sections were blocked in normal mouse serum (1/20) for 15 min prior to incubation for 60 min with the primary antibody: 8H4 for PrP detection or the rat anti-mouse FDC-M2 monoclonal antibody for FDC (1/800; AMS Biotechnology). After being washed in Tris (pH 7.6)-bovine serum albumin (BSA) buffer, sections were further incubated for 60 min in mouse anti-rat biotinylated serum (1/500; Jacksons). After further washing, sections were stained with streptavidin-alkaline phosphatase reagent and then Vector red alkaline phosphatase substrate (Vector) according to the manufacturer's protocols. Finally, sections were counterstained with hematoxylin and rinsed in Scott's tap water prior to dehydration and mounting. Confocal microscopic studies were performed in sections probed with antibodies CD16/CD32 at a 1/100 dilution (BD Biosciences). Following washing in Tris (pH 7.6)-BSA buffer, sections were further incubated for 60 min in mouse anti-rat biotinylated serum (1/500; Jacksons). After further washing, sections were incubated with streptavidin-Alexa Thioflavin-S treatment. Brain sections were stained with hematoxylin for 30 s and then washed in water. Sections were treated with a thioflavin-S solution (10 mg/ml) for 5 min and then immersed in 70% ethanol solution for 5 min. After sections were rinsed in water, they were mounted using fluorescent mounting medium (Dakocytomation) and analyzed using a Hamamatsu camera and Image Pro Plus.

Host PrP glycosylation regulates the timing of TSE neuroinvasion.
To assess the effect of glycosylation in the TSE trafficking following peripheral exposure, wild-type and glycosylation-deficient mice were inoculated intraperitoneally (i.p.) with the ME7 or 79A mouse-adapted scrapie strain and monitored for the appearance of clinical signs of disease (16). G2 homozygous mice (G2/G2) showed a very long incubation period of disease compared to wild-type mice following i.p. inoculation with 79A. All G2/G2 animals succumbed to TSE disease, with an average incubation period of 307 (Ϯ 4.5) days, whereas the average incubation period in wild-type animals was 201 (Ϯ 2.2) days (Table 1). However, inoculation of G2/G2 mice with ME7 produced only a modest increase in the incubation period, with an average of 280 (Ϯ4) days compared to 257 Ϯ 3 days for wild-type mice inoculated with the same strain (Table 1). Thus, lack of sugars at the second site delays the onset of disease, but the extent of this delay is TSE strain specific.
A very long incubation period and low susceptibility were also observed in G1 homozygous mice (G1/G1) inoculated i.p. with 79A compared with the wild-type mice. In this case only 11/18 G1 animals succumbed to disease, with an average incubation period of 310 (Ϯ 35) days (Table 1), whereas no G1/G1 mice showed any clinical disease signs after more than 700 days following peripheral administration of ME7. No clinical signs of disease were observed in G3 homozygous mice (G3/G3) after i.p. inoculation with either ME7 or 79A (Table 1). Therefore, host PrP glycosylation clearly influences the incubation time and susceptibility to disease following an intraperitoneal inoculation.
Glycosylation state of PrP C in the host determines the targeting of pathological lesions in the brain following peripheral infection. Western blot analysis using the anti-PrP antibodies 8H4 or 7A12 revealed the presence of PK-resistant unglyco-   sylated and monoglycosylated PrP Sc in brains of all clinically positive G2/G2 animals inoculated with either ME7 or 79A (Fig. 1). A PK-resistant PrP Sc characterized by lack of the diglycosylated band was also detected in 11 out of 18 clinically positive G1/G1 mice inoculated with 79A but not in the seven clinically negative G1/G1 animals after inoculation with the same strain (Fig. 1A). No PrP Sc was detected in brain from G1/G1 mice inoculated with ME7 (Fig. 1B), and this was also the case for G3/G3 mouse brains after challenge with 79A or ME7 (Fig. 1). PrP Sc deposition was assessed in the brains of mice inoculated with 79A and ME7 strains by immunohistochemistry (IHC) using 6H4 antibody. This analysis revealed a remarkable difference in both the amount and distribution of PrP Sc deposition between wild-type mice and G2/G2 mice injected with 79A. In wild-type mice a mild to moderate, fine punctate PrP Sc immuno-positivity was observed in the thalamus and midbrain. Sparse PrP Sc deposits were also occasionally seen in other brain areas ( Fig. 2A to C). In G2/G2 animals fine punctate PrP Sc immuno-positive deposits were seen in several brain regions: septum, thalamus, hippocampus, hypothalamus, midbrain, and brain stem. In several G2/G2 mice high levels of PrP Sc accumulation were also observed in the cerebral cortex, whereas in wild-type animals this area of the brain had only sparse or no PrP Sc deposition (Fig. 2D to F).
FIG. 2. PrP Sc deposition in brain of wild-type and transgenic homozygous mice after peripheral administration of 79A. Brains from clinically positive and clinically negative mice after 79A challenge were immunostained for PrP Sc with the monoclonal antibody 6H4 and analyzed by light microscopy using a Nikon Eclipse E800 microscope. (A to C) Wild-type animal brain showing a characteristic fine punctate PrP deposition. In these mice little involvement of the hippocampus and cortex areas (A) and of the midbrain (B) was observed, with most of the PrP Sc deposition present in the thalamus (C). (D to F) G2/G2 brains showing a very strong mantle-like PrP Sc deposition in several brain regions such as cortex and hippocampus (D), midbrain (E), and thalamus (F). (G to K) G1/G1 mouse brain showing a very peculiar PrP Sc deposition characterized by aggregates of PrP Sc in very specific brain regions. In contrast to G2/G2 mice, G1 animals did not show any PrP Sc deposition in many brain areas, including the hippocampus and cortex (G). However, some PrP Sc deposition in the form of fine punctuate deposits was observed in the habenula (H). A closer examination of these brains revealed the presence also of unusual PrP Sc small aggregates in the thalamus (I) that were thioflavin-S negative (J). These aggregates were PrP Sc specific since they were not present in control brain sections treated just with normal mouse serum without any anti-PrP antibody (K). (L) G3/G3 clinically negative brain showing no PrP deposition in any area of the brain examined. Magnifications, ϫ4 (A to G, J, and L), ϫ10 (H), and ϫ20 (I). In ME7-inoculated wild-type animals, fine punctate and coarse (confluent) PrP Sc immuno-positive deposits were seen in multiple brain areas including the septum, cerebral cortex, thalamus, hippocampus, hypothalamus, midbrain, brain stem, and cerebellum ( Fig. 3A to C). Similar topographical distributions of fine punctate and coarse PrP Sc deposits were observed in wild-type and G2/G2 mice. However, PrP Sc deposition in the ME7-infected G2/G2 mice was also characterized by the presence of large, thioflavin S-positive amyloid plaques in all of the affected regions. (Fig. 3D to G).
The degree of spongiform degeneration in the CNS as determined by lesion profile analysis is an important parameter to define TSE disease (9). G2/G2 animals inoculated with 79A showed a higher degree of spongiform degeneration than wildtype mice in a number of areas of the gray matter: hypothalamus, thalamus, hippocampus, septum, and forebrain cortex. However, in other regions, such as cerebral cortex or medulla, no differences were observed. In the white matter areas analyzed (cerebellar white matter, mesencephalic tegmentum, and pyramidal tract), G2/G2 mice demonstrated a milder spongiosis (Fig. 4A). Differences in spongiform degeneration were also observed in G2/G2 mice inoculated with ME7, with more extensive spongiosis in the hypothalamus region in the transgenic mice than in wild-type mice and less in other areas such as the hippocampus, septum, cerebral cortex, and forebrain cortex in the gray matter and the pyramidal tract of the white matter (Fig. 4B).
An even more striking difference in pathology was observed in G1/G1 mice inoculated with 79A compared to wild-type animals in both the distribution and intensity of the PrP Sc deposition and spongiosis. In the G1/G1 mice sparse, fine punctate PrP Sc deposits were seen in habenula, thalamus, midbrain, and brain stem. In addition, coarse punctate PrP Sc deposits were seen in the lateral thalamic region (Fig. 2G to I).
To further investigate if this pattern was composed of amyloid deposits, infected G1/G1 mouse brain sections were treated with thioflavin-S (37). This analysis, however, did not reveal any fluorescent deposits in the parenchyma (Fig. 2J). A further unusual pathology in brain was also observed in these mice once a spongiform degeneration profile was performed. Indeed, in G1/G1 mouse brains spongiosis was restricted to the hypothalamus and dorsal medulla areas, whereas in wild-type mice a higher degree of spongiform degeneration was observed in all the gray and white matter regions analyzed (Fig. 4C). No spongiform degeneration or PrP Sc deposition was found in any clinically negative G1/G1 and G3/G3 mice after inoculation with 79A or ME7, supporting the observation that these mice are resistant to peripheral inoculation with these strains (Fig.  2L and 3H and I). A single copy of wild-type PrP restores the incubation time in PrP glycosylation-deficient mice. Heterozygous mice expressing both wild-type and G2 PrP (Wt/G2) were also inoculated i.p. with 79A or ME7. Wt/G2 mice showed an incubation time strikingly similar to that of homozygous wild-type mice. Indeed, Wt/G2 mice inoculated with 79A died in an average incubation period of 205 (Ϯ 3) days, whereas wild-type animals succumbed with an average incubation period of 201 (Ϯ 2.2) days ( Table 2). Similar incubation times were also observed between heterozygous and wild-type mice inoculated with ME7 (Wt/G2, 253 Ϯ 5.5 days; Wt/Wt, 257 Ϯ 2.8 days) ( Table  2). Thus, a single copy of the wild-type Prnp gene was able to restore the incubation times of disease observed in the wildtype mice.
Western blot analysis performed in brains of the clinically positive Wt/G2 heterozygous animals, after exposure to 79A or ME7, revealed the presence of PK-resistant diglycosylated PrP Sc . The glycoprofile of PrP Sc in these mice was characterized by having a monoglycosylated band more intense than the diglycosylated band (Fig. 1). This differs from the glycoprofile observed in wild-type homozygous animals characterized by a predominant diglycosylated band and reflects the glycosylation status of PrP C in Wt/G2 mice.
Despite identical incubation times, immunohistochemical analysis of PrP Sc deposition in brain sections revealed differences between wild-type and heterozygous animals. Analysis of brain sections from Wt/G2 mice inoculated with the 79A strain showed accumulation of PrP Sc in several areas of the brain. However, the severe deposition of PrP Sc in the cerebral cortex detected in G2/G2 mice was not apparent in heterozygous animals ( Fig. 5A and B). Analysis of brain sections from Wt/G2 mice inoculated with the ME7 strain shows fine punctate PrP Sc deposits in several areas. Moreover, coarse and plaque-like PrP Sc deposits previously observed in G2/G2 mice were also present in the heterozygous animals; however, the size and numbers of these deposits were reduced compared to those seen in G2 homozygous animals ( Fig. 5D and E). The FIG. 4. Lesion profile analysis of wild-type and PrP glycosylation mutant homozygous mice after peripheral inoculation with 79A or ME7. Spongiform degeneration in brains of the clinically positive wild-type, G2, and G1 mice was analyzed after peripheral infection with 79A or ME7. For this purpose nine gray matter areas were scored on a scale of 0 to 5 (y axis). Gray matter areas are represented on the x axis by the following numbers: 1, dorsal medulla; 2, cerebellar cortex; 3, superior colliculus; 4, hypothalamus; 5, medial thalamus; 6, hippocampus; 7, septum; 8, cerebral cortex; and 9, forebrain cortex. Three white matter areas were scored on a scale of 0 to 3 (y axis). White matter areas are represented on the x axis by the following numbers: 11, cerebellar white matter; 12, mesencephalic tegmentum; and 13, pyramidal tract. The magnitude of spongiform degeneration in the transgenic mice was then compared with the one in wild-type mice.  immunohistochemical characteristics of PrP Sc accumulation in heterozygous mice can be best described as intermediate between those seen in G2 and wild-type homozygous animals. The spongiform degeneration analysis revealed no differences in the brain areas affected between Wt/G2 and Wt/Wt mice inoculated with 79A (Fig. 5C). However, Wt/G2 mice inoculated with ME7 have reduced vacuolation in specific brain areas compared to wild-type mice, e.g., in the hippocampus, septum, cerebral cortex, forebrain cortex, and the pyramidal tract (Fig. 5F). Glycosylation-deficient mice possess mature and functional FDC that express PrP C . Since FDC have been suggested to be a pivotal component in the uptake and replication of infectivity in the periphery, we assessed if the delay or resistance observed in the glycosylation-deficient mice could be linked to any possible FDC abnormality.
FDC networks were detected using the FDC-M2 antibody. The epitope recognized by this antibody has been shown to be the activated form of the complement receptor C4 (44). FDC-M2 labels mature FDC networks within the germinal center as well as parts of the primary follicle. Initial light microscopy analysis revealed no overt differences between the FDC networks of transgenic and wildtype mice (Fig. 6A). To further explore the functionality of these networks, immunolabeling was carried out using an antibody which recognizes an epitope common to both CD16 and CD32, which are both low-affinity receptors for the immunoglobulin Fc portion expressed by functional FDC, as well as other cell types (18). No differences were evident in the labeling between the different genotypes (Fig. 6B).
It was then determined whether PrP C localization within these FDC networks is altered in the transgenic mice compared with wild-type mice. For labeling of PrP C , the rabbit polyclonal antibody used was 1B3, which targets the amino acid residues 14 to 36, 83 to 102, 119 to 139, and 188 to 212 on the PrP molecule (15). Confocal analysis of double immunofluorescent labeling of PrP and FDC networks showed that PrP C colocalizes within the FDC networks of each of the transgenic genotypes and wild-type samples (Fig. 6C). Thus, the glycosylation-deficient mice, like wild-type mice, appear to possess an intact FDC network which expresses PrP C .
Host PrP glycosylation influences the peripheral stages of TSE infectivity. Following peripheral inoculation of TSE-infected material, accumulation and amplification of infectivity in the LRS occur rapidly, with infectivity levels in these tissues reaching a plateau after several weeks (14)(15). We therefore analyzed the presence of PrP Sc deposition in spleens of clinically positive or clinically negative glycosylation-deficient transgenics and wild-type mice after i.p. inoculation with the 79A or ME7 strain by IHC. This analysis showed a correlation between the amount of PrP deposition in spleen and subsequent clinical disease. Clinically positive Wt/Wt and G2/G2 mice after 79A or ME7 infection or G1/G1 mice after 79A FIG. 5. Brain pathology analysis in G2 heterozygous animals challenged with 79A or ME7. Brains from clinically positive Wt/G2 heterozygous animals inoculated with 79A or ME7 were immunostained for PrP with monoclonal antibody 6H4 and analyzed by light microscopy using a Nikon Eclipse E800 microscope. Analysis of the spongiform degeneration in the nine gray matter areas and the three white matter areas described in the legend of Fig. 4 was also carried out in Wt/G2 brains and compared with spongiform degeneration in Wt/Wt mice. (A and B) Wt/G2 brain after inoculation with 79A i.p. showed a PrP deposition intermediate between that of wild-type and G2 homozygous animals. PrP Sc accumulation was indeed observed in regions such as thalamus, hippocampus, and cerebral cortex (A) or midbrain (B). However, this deposition was less intense than the one observed in G2 homozygous animals (compare Fig. 2). (C) Although there were some differences in terms of PrP Sc deposition between Wt/Wt and Wt/G2 mice, no differences were observed between these two lines in terms of spongiosis in all the brain areas analyzed. (D and E) Wt/G2 mice brains after peripheral infection with ME7 also resembled an intermediate pathology between wild-type and G2 homozygous animals. PrP Sc deposits and small plaques were indeed observed in regions such as the hippocampus (D) and the septum (E) (compare Fig. 3). (F) Heterozygous animals presented a vacuolation profile that differed from that of wild-type mice infected with ME7, with lower degrees of vacuolation in several brain areas: 6, hippocampus; 7, septum; 8, cerebral cortex; and 9, forebrain cortex. Magnification, ϫ4 (A, B, D, and E).
challenge showed large amounts of PrP Sc throughout the spleen (Fig. 7A to D and F). However, complete absence of PrP deposition was observed in spleens of clinically negative G1/G1 mice inoculated with ME7 and of clinically negative G3/G3 mice challenged with 79A or ME7 (Fig. 7E, G, and H). Very small PrP Sc deposits in the spleen were also observed in some of the seven G1/G1 mice that were not affected by a clinical TSE disease after 79A challenge, suggesting that in these mice PrP Sc amplification levels were not sufficient to further elicit TSE clinical disease (Fig. 7I). Total lack of glycosylation in PrP C prevents trafficking of PrP Sc from brain to spleen. After intracerebral inoculation, PrP Sc is often detected first in the spleen, suggesting that infectivity travels from the site of injection in the brain to the LRS and replicates in the periphery prior to causing disease in the CNS (15,17). To test the influence of host PrP glycosylation in transport of PrP Sc from brain to spleen, we analyzed by Western blotting the presence of PrP Sc in spleens of the three glycosylation mutant mice infected intracerebrally in previous experiments (45) with ME7 and 79A. Wild-type and G2 mice are susceptible to infection with ME7, and PrP Sc was detected in the spleens. However, G1/G1 and G3/G3 mice did not develop any clinical disease after challenge with ME7, and no PrP Sc was detected in spleens (Fig. 8A), suggesting that lack of replication in the LRS may prevent clinical disease in these mice, as previously suggested (7). Presence of PrP Sc was observed in the spleens of clinically positive G1/G1 and G2/G2 mice infected with 79A, but surprisingly no PrP Sc was found in the spleens of G3/G3 mice that were clinically positive (Fig.  8B). These results suggest that monoglycosylated but not unglycosylated PrP is able to sustain transport of infectivity between the brain and the periphery and that while replication in the spleen is important for establishing disease with ME7 via intracerebral or intraperitoneal routes, this step may not be mandatory for establishing infection with 79A.

DISCUSSION
The mechanisms regulating peripheral propagation of the different TSE strains are not yet fully understood. Different organs, cells, and routes have been shown to be important according to which strain of agent is infecting a particular host. Some strains require amplification in the LRS before causing a productive infection in the CNS, whereas with other strains, such as naturally occurring BSE in cattle, the involvement of the LRS is limited (3-4, 12, 22, 42). A number of factors are likely to contribute to this diversity in neuroinvasion, including strain, host PrP genotype, and the route of entry. Our results FIG. 6. Characterization of spleen morphology in glycosylation-deficient mice. In order to assess if a lack of sugars can induce an overt phenotype in spleen, a number of morphological assays in spleen derived from transgenic, wild-type, and PrP knockout animals (null) were performed. (A) FDC networks were detected using the FDC-M2 antibody, which labels mature FDC networks within the germinal center. With light microscopy analysis using a Nikon Eclipse E800 microscope, no differences were highlighted between wild-type and glycosylation-deficient mice. (B) CD16/CD32 antibody was used to investigate the functionality of FDC by confocal microscopy analysis. Again, there were no differences evident in the labeling between the different genotypes. Scale bar, 50 m. (C) PrP localization within the FDC network was analyzed by confocal microscopy analysis. Double immunofluorescent labeling of PrP with 1B3 antibody (green) and of FDC networks with FDC-M2 antibody (red) showed that PrP C colocalizes within the FDC networks (yellow) of each of the transgenic genotypes and wild-type samples. Scale bar, 50 m. now suggest that glycosylation status of the host PrP may be an important factor in this process and, in particular, in determining the timing of neuroinvasion and the final targeting in the CNS.
Glycosylated PrP C appears to be required at the early stages of the infectious process, which may involve uptake and transport from the site of infection to the spleen or amplification of infectivity in the spleen. Clinical disease was observed in each of the PrP glycosylation-deficient mice following intracerebral inoculation with at least one strain (45), albeit with very long incubation periods and limited susceptibility in the G3 mice. In this study this was not the case as following intraperitoneal inoculation the G3 mice appeared to show complete resistance to infection with two different TSE strains. This may be due to an inability of unglycosylated PrP C to sustain the transport of infectivity from the periphery to the brain and/or vice versa. Therefore, PrP glycosylation plays a central role in the spread of infectivity from the periphery to the CNS. Moreover, our results suggest that replication or transport in the LRS is a limiting factor during spreading of TSE infectivity.
The role played by PrP C glycotypes in the pathogenesis of disease appears to be strain specific. G2 mice developed disease after ME7 i.p. inoculation with only a short delay of 21 days compared to wild-type mice. Similarly, when ME7 was injected directly into the CNS, G2 mice developed TSE disease with incubation periods similar to those of wild-type mice (45). Thus, the monoglycosylated G2 PrP protein appears to be able to replicate and transport the ME7 agent almost as efficiently as the diglycosylated PrP. However, ME7 fails to establish disease following peripheral challenge of G1/G1 and G3/G3 mice, and PrP Sc is absent from spleens and brains of these mice. This is in accordance with what was observed following direct inoculation of ME7 into the brain; although these transgenic mice did not develop any clinical TSE disease, G3 mice presented some degree of PrP deposition in the form of amyloid plaques (45). Thus, the first glycosylation site appears critical for the replication of ME7 in both the CNS and periphery. Failure to establish disease is therefore likely to be due to a failure of the host to replicate the agent.
G1 and G2 homozygous mice were susceptible to infection following peripheral challenge with 79A; however, the incubation time was more than 100 days longer than for wild-type animals. This delay after peripheral inoculation is much longer than that observed following inoculation of the same strain intracerebrally; in G1/G1 and G2/G2 mice the incubation of clinical prion disease was 40 and 20 days longer, respectively, than in wild-type mice (45). This suggests either that there is a difference in efficiency of replication of the agents in the pe-FIG. 7. PrP Sc deposition in spleens of mice after peripheral inoculation with strain 79A or ME7. Spleens from clinically positive and clinically negative mice were immunostained for PrP with the monoclonal antibody 6H4 and analyzed by light microscopy using a Nikon Eclipse E800 microscope and a 10ϫ objective. PrP Sc deposition was observed in spleens of wild-type (A), G1/G1 (B), and G2/G2 (C) clinically positive animals after inoculation with 79A. PrP Sc deposition was also observed in spleens of wild-type (D) and G2/G2 (F) clinically positive animals but not in G1/G1 clinically negative mice (E) after ME7 infection. PrP Sc accumulation was also not observed in G3/G3 clinically negative mice after i.p. inoculation with either 79A or ME7 (G and H). Slight PrP Sc deposition was observed in spleen of some clinically negative G1 mice after challenge with 79A (I). riphery and the CNS or that lack of diglycosylated protein in this case is delaying the transport to the CNS. However, the incubation time in G2 mice was rescued when a wild-type PrP was expressed alongside the monoglycosylated PrP. In heterozygous PrP mice a gene dosage effect on the length of TSE clinical disease onset was observed due to the contribution of both alleles. This effect was previously described in wild-type/ knockout PrP heterozygous transgenic mice (PrP ϩ/Ϫ ) in which the incubation time was almost doubled compared to incubation in PrP wild-type homozygous animals (32). In Wt/G2 mice it was therefore surprising to observe the appearance of clinical disease at the same time as in Wt/Wt mice when it might have been expected that incubation periods in these mice would fall somewhere between those of wild-type and G2 homozygous mice. This suggests that the mono-and diglycosylated forms of PrP can act together to facilitate replication and transport of the agent. This suggestion is also supported by the observation that heterozygous animals presented a brain pathology that was an intermediate between that observed in wild-type and G2 homozygous mice, indicating that both alleles are also contributing to the final targeting in brain. Thus, the provision of diglycosylated PrP clearly provides an important function in the disease process and can overcome the incubation time delays observed with only monoglycosylated and unglycosylated PrP. The discrepancy between the hybrid pathology and the wild-type incubation times observed in the heterozygote mice demonstrates that these are an important model for defining the mechanisms regulating the neuroinvasion of TSE infectivity, and further studies are under way to address this issue.
Host PrP glycosylation appears to determine the targeting in the CNS following peripheral challenge. Absence of fully glycosylated PrP had a profound effect in determining the brain area targeted by PrP Sc deposition after infection with 79A. G1 mouse brains were characterized by PrP Sc accumulation in very restricted areas like the habenula and the thalamus. This deposition was remarkably different from the characteristic widespread distribution of PrP Sc in many brain areas observed in wild-type mice. G2 homozygous animals also presented changes in 79A targeting compared to wild-type animals, with cortex and midbrain heavily affected. This different targeting may be due to an effect of host PrP glycosylation in determining routes of propagation of infectivity from the periphery to the CNS or to the fact that the transgenic mice developed clinical disease with a much longer incubation time than the wild types. However, even in cases where there were modest alterations in the incubation times, in the glycosylation-deficient transgenic mice there were dramatic differences observed in PrP Sc deposition. Absence of a fully glycosylated PrP appears to facilitate the deposition of amyloid plaque formation in the brain, compared with fine punctuate staining observed in wild-type mice after ME7 inoculation. This contrasts with the targeting in the brain following intracerebral challenge, where no difference in targeting was observed in the different lines of mice (45). After inoculation of 79A or ME7 directly into the brain, clinically positive G1/G1 and G2/G2 mice, indeed, did not present any differences in terms of spongiform degeneration or PrP Sc deposition compared to wild-type animals. Wildtype and glycosylation-deficient mice had the same vacuolation profiles and the same widespread, fine punctuate PrP Sc depositions throughout the brain (45). These results may also suggest that the delay observed in the transgenic animals may be due to an initial targeting of infectivity in different brain areas that may delay the neurotoxic effect, as previously proposed (24). Whether differences in routing to the CNS are responsible for the differences in targeting observed in the glycosylation-deficient mice remains to be established.
A difference in brain pathology between wild-type and gly-FIG. 8. Travel of PrP Sc from brain to spleen. In order to assess influence of host PrP glycosylation in trafficking of PrP Sc from the CNS to the periphery, spleens from glycosylation-deficient mice and wild-type animals inoculated intracranially with ME7 or 79A were tested for presence of PK-resistant PrP by Western blotting using the anti-PrP antibody 8H4. (A) Western blot analysis of spleen from mice infected i.c. with ME7 after pretreatment with PK. PK-resistant PrP was observed only in clinically positive wild-type and G2/G2 mice; no PrP Sc was found in spleen of clinically negative G1/G1 and G3/G3 mice. (B) Western blot analysis of spleen from mice infected i.c. with 79A after pretreatment with PK. PrP Sc was recovered in all clinically positive G1/G1, G2/G2, and wild-type mice; however, no PK-resistant PrP was found in clinically positive G3/G3 animals.
VOL. 84, 2010 ROLE OF PrP GLYCOSYLATION IN PERIPHERALLY ACQUIRED TSE cosylation-deficient mice was also confirmed by the analysis of vacuolation damage in the CNS. Differences were observed in G1 and G2 mice after infection with both 79A and ME7 in several brain areas. This analysis also revealed a discrepancy between PrP Sc accumulation and spongiform degeneration. For example, the strong accumulation of PrP Sc observed in the cortex of G2 but not of wild-type animals was not linked with a difference in terms of spongiosis in the same areas. This observation may support previous observations suggesting that PrP Sc presence is not always associated with vacuolation (2,37). In summary, we have shown that PrP is a requirement for the peripheral events in TSE disease. In this process PrP glycosylation is a key component to determine the timing of neuroinvasion and the targeting in the CNS, and this effect varies according to TSE strain. Indeed, a fully glycosylated PrP molecule is required for the most efficient neuroinvasion by 79A. However, ME7, while less dependent on a fully glycosylated PrP, appears to be totally dependent on glycosylation at the first site. Finally, the results presented here indicating a role of N-linked glycans in the spread of TSE infectivity may have some important implications for the development of new therapeutic approaches. For example, destabilizing sugar-mediated interactions between host and the infectious agent would appear to provide an important approach to block propagation of infectivity.