Oral Prion Disease Pathogenesis Is Impeded in the Specific Absence of CXCR5-Expressing Dendritic Cells

ABSTRACT After oral exposure, the early replication of certain prion strains upon stromal cell-derived follicular dendritic cells (FDC) in the Peyer's patches in the small intestine is essential for the efficient spread of disease to the brain. However, little is known of how prions are initially conveyed from the gut lumen to establish infection on FDC. Our previous data suggest that mononuclear phagocytes such as CD11c+ conventional dendritic cells play an important role in the initial propagation of prions from the gut lumen into Peyer's patches. However, whether these cells conveyed orally acquired prions toward FDC within Peyer's patches was not known. The chemokine CXCL13 is expressed by FDC and follicular stromal cells and modulates the homing of CXCR5-expressing cells toward the FDC-containing B cell follicles. Here, novel compound transgenic mice were created in which a CXCR5 deficiency was specifically restricted to CD11c+ cells. These mice were used to determine whether CXCR5-expressing conventional dendritic cells propagate prions toward FDC after oral exposure. Our data show that in the specific absence of CXCR5-expressing conventional dendritic cells the early accumulation of prions upon FDC in Peyer's patches and the spleen was impaired, and disease susceptibility significantly reduced. These data suggest that CXCR5-expressing conventional dendritic cells play an important role in the efficient propagation of orally administered prions toward FDC within Peyer's patches in order to establish host infection. IMPORTANCE Many natural prion diseases are acquired by oral consumption of contaminated food or pasture. Once the prions reach the brain they cause extensive neurodegeneration, which ultimately leads to death. In order for the prions to efficiently spread from the gut to the brain, they first replicate upon follicular dendritic cells within intestinal Peyer's patches. How the prions are first delivered to follicular dendritic cells to establish infection was unknown. Understanding this process is important since treatments which prevent prions from infecting follicular dendritic cells can block their spread to the brain. We created mice in which mobile conventional dendritic cells were unable to migrate toward follicular dendritic cells. In these mice the early accumulation of prions on follicular dendritic cells was impaired and oral prion disease susceptibility was reduced. This suggests that prions exploit conventional dendritic cells to facilitate their initial delivery toward follicular dendritic cells to establish host infection.

protein (PrP C ), in affected tissues. Prion infectivity copurifies with PrP Sc and constitutes the major component of the infectious agent (1). Many prion diseases, including natural sheep scrapie, bovine spongiform encephalopathy, chronic wasting disease in cervid species, and kuru and variant Creutzfeldt-Jakob disease in humans, are acquired by peripheral exposure, such as by oral consumption of prion-contaminated food or pasture.
After ingestion many prion isolates initially accumulate and replicate first upon follicular dendritic cells (FDC) within the gut-associated lymphoid tissues (GALT) in the small intestine, such as the Peyer's patches, before they spread to the nervous system (termed neuroinvasion) (2)(3)(4)(5)(6)(7)(8)(9). After their replication upon FDC studies in experimental rodents show that the prions subsequently infect neurons of both the sympathetic and parasympathetic nervous systems and spread along these to the central nervous system (CNS), where they ultimately cause neurodegeneration (10)(11)(12).
Mononuclear phagocytes arise from precursors in the bone marrow and comprise a heterogeneous population of monocytes, conventional dendritic cells (DC), and tissue macrophages. The different mononuclear phagocyte populations display a diverse range of roles during prion disease (13). For example, tissue macrophages appear to aid the sequestration and clearance of prions (14,15), and microglia help protect the brain against prion-induced neuropathology (16). Conventional DC are a distinct lineage from the stromal derived FDC (17,18) and are strategically located throughout the body to sample the local environment for pathogens and their antigens. After antigen uptake, these cells typically undergo a degree of maturation and migrate toward the draining lymphoid tissue to initiate a specific immune response. Some conventional DC populations can retain antigens in their native states and transfer them intact to naive B cells in order to initiate a specific antibody response (19). The migratory characteristics of conventional DC are exploited by some pathogens to facilitate their delivery to lymphoid tissues (20)(21)(22)(23). Independent studies have shown that the early replication of prions upon FDC in the draining lymphoid tissue is impeded in the transient absence of conventional DC at the time of exposure (24)(25)(26)(27). This suggests that prions may also exploit these cells to establish host infection after peripheral exposure, perhaps using them as "Trojan horses." Whether conventional DC propagate prions toward FDC within the B cell follicles is not known. Treatments that prevent the uptake and propagation of prions to FDC can significantly delay disease pathogenesis and reduce susceptibility (24)(25)(26)(27)(28)(29). A thorough understanding of the cellular mechanisms used by prions to establish infection upon FDC in the GALT may help to identify novel targets for prophylactic intervention.
Chemokines attract lymphocytes and leukocytes to lymphoid tissues and control their positioning within them. The chemokine CXCL13 is expressed by FDC and follicular stromal cells in the B-cell follicles of lymphoid tissues and mediates the homing of CXCR5-expressing cells toward them (30,31). The migration of certain populations of conventional DC toward the FDC-containing B-cell follicles is also mediated by CXCL13-CXCR5 signaling (32)(33)(34). We therefore reasoned that if conventional DC were important for the efficient propagation of prions toward FDC in order to establish host infection, this activity would be impeded and disease susceptibility reduced in mice which specifically lacked CXCR5-expressing conventional DC. Since mice that lack Peyer's patches are refractory to oral prion infection (3,5,6,35), CXCR5 Ϫ/Ϫ mice and CXCL13 Ϫ/Ϫ mice were unsuitable for use here because they also lack most Peyer's patches and certain lymph nodes (31,36). The few Peyer's patches that do develop in these mice are much smaller, and their microarchitecture is grossly disturbed since the B cells are unable to form organized follicles (31,36). Furthermore, as a consequence of the disturbed microarchitecture in the spleens of CXCR5 Ϫ/Ϫ mice, prion neuroinvasion after intraperitoneal exposure occurs at a higher rate because their FDC are abnormally superimposed over sympathetic nerves (37). Therefore, in the present study a unique compound transgenic mouse model was created in which Cxcr5 was specifically ablated in CD11c ϩ conventional DC. These CXCR5 ΔDC mice were then used to test the hypothesis that conventional DC play an important role in the efficient B-cell follicles of lymphoid tissues and mediates the homing of CXCR5-expressing cells toward them (30,31). Ex vivo chemotaxis assays confirmed that the migration of CD11c ϩ conventional DC from CXCR5 ΔDC mice toward CXCL13 was significantly impeded compared to conventional DC from CXCR5 F/F control mice ( Fig. 2; P Ͻ 0.024). In contrast, the chemotaxis of B cells (B220 ϩ cells) from CXCR5 ΔDC mice toward CXCL13  was equivalent to that observed from cells from CXCR5 F/F mice. The ability of conventional DC to migrate toward the chemokine CCL21 (which signals via CCR7) was also similar in cells from each mouse group. Lymphoid tissue microarchitecture in CXCR5 ⌬DC mice. The sampling of prions across the intestinal epithelium by M cells, and their subsequent early replication upon PrP C -expressing FDC is obligatory for efficient neuroinvasion after oral exposure (4-6, 29, 42). We therefore determined whether the development of these cells was affected in the lymphoid tissues of CXCR5 ΔDC mice. Whole-mount immunostaining using the mature M-cell marker glycoprotein 2 (GP2 [43,44]) revealed similar densities of GP2 ϩ M cells in the follicle-associated epithelia (FAE) overlying the Peyer's patches and CXCR5 ΔDC mice and CXCR5 F/F control mice ( Fig. 3A and B). The size of the FAE was also equivalent (Fig. 3C). Unlike the disturbed distribution of FDC in the spleens of CXCL13 Ϫ/Ϫ mice (37), those in the lymphoid tissues of CXCR5 ΔDC mice formed distinct networks. Furthermore, immunohistochemistry (IHC) and morphometric analysis suggested that FDC size (the area occupied by CD35 ϩ immunostaining) was similar in the mesenteric lymph nodes (MLN) and spleens from CXCR5 ΔDC and CXCR5 F/F mice ( Fig. 3H and K, respectively), although those in the Peyer's patches of CXCR5 ΔDC mice were slightly smaller (Fig. 3E). However, the abundance of PrP Cϩ immunostaining upon FDC in Peyer's patches and MLN from each mouse group was similar ( Fig. 3F and I, respectively), although a marginal reduction was in observed in the spleens of CXCR5 ΔDC mice (Fig. 3L).
Prion accumulation in lymphoid tissues is reduced in orally exposed CXCR5 ⌬DC mice. In order to determine the effects of conventional DC-specific CXCR5-deficiency on oral prion pathogenesis, groups of CXCR5 ΔDC mice and CXCR5 F/F mice were orally exposed to ME7 scrapie prions and Peyer's patches, MLN, and spleens collected at intervals afterwards (n ϭ 4/group). Prion disease-specific, abnormal accumulations of PrP (referred to as PrP d ) characteristically present only in prion-affected tissues were detected by IHC (4-6, 29, 45-48). Paraffin-embedded tissue (PET) blot analysis of adjacent membrane-bound sections was used to confirm that these PrP d aggregates contained prion disease-specific, relatively proteinase K (PK)-resistant PrP Sc (49).
As anticipated, by 70 days postinfection (dpi), abundant PrP Sc accumulations were detected in association with FDC (CD21/35 ϩ cells) in the majority of the Peyer's patches from CXCR5 F/F control mice (Fig. 4A). By 105 dpi, heavy FDC-associated PrP Sc accumulations were detected in all the Peyer's patches of CXCR5 F/F mice ( Fig. 4B and C). However, the incidence of the PrP Sc accumulation in Peyer's patches of CXCR5 ΔDC mice was reduced at each of these times after prion exposure ( Fig. 4A to C). and B220 ϩ lymphocytes (B cells) which had migrated into the lower chamber after 24 h was determined by flow cytometry. Med., medium alone was used as a control. The ability of CD11c ϩ conventional DC from CXCR5 ΔDC mice to migrate toward CXCL13 was significantly impeded. *, P Ͻ 0.024 (n ϭ 5 wells/treatment).
Within weeks of oral exposure, high levels of ME7 scrapie prions first accumulate upon FDC in the Peyer's patches and are subsequently disseminated via the blood and lymph to most other lymphoid tissues (4-6, 24, 29, 50). In the spleens of orally exposed CXCR5 F/F control mice, some FDC-associated PrP Sc accumulations were first evident at 70 dpi ( Fig. 5A to C). However, the abundance of these FDC-associated PrP Sc accumulations was reduced in the spleens of CXCR5 ΔDC mice at 70 and 105 dpi ( Fig. 5A to C). To compare the prion infectivity levels within these tissues, spleen homogenates were prepared and injected intracerebrally (i.c.) into groups of tga20 indicator mice (n ϭ 4 recipient tga20 mice/spleen homogenate). High levels of prion infectivity were detected in the majority of the spleens collected from the CXCR5 F/F control mice at 70 dpi and had increased in magnitude by 105 dpi (Fig. 5D). In contrast, significantly lower levels of prion infectivity were detected in the spleens of orally exposed CXCR5 ΔDC mice ( Fig. 5D; 70 dpi, P Ͻ 0.0004; 105 dpi, P Ͻ 0.009; log-rank [Mantel-Cox] test). Only trace levels of prion infectivity were detected in three of four spleens from the CXCR5 ΔDC mice analyzed at 70 dpi and in two of the four spleens analyzed at 105 dpi.
Together, these data show that in the specific absence of CXCR5-expressing conventional DC at the time of oral exposure, the early accumulation of prions upon FDC

FIG 4
The early accumulation of prions in the Peyer's patches is delayed in CXCR5 ΔDC mice. Mice were orally exposed to ME7 scrapie prions; Peyer's patches were collected at 70 and 105 dpi and analyzed by IHC and PET immunoblot analysis. At 70 (A) and 105 (B) dpi, the frequencies of FDC networks containing PrP Sc in the Peyer's patches from each mouse group were compared (n ϭ 4 mice/group). Horizontal bar, median. (C) IHC analysis showed that at 105 dpi, high levels of disease-specific PrP (PrP d ; brown, middle row, arrows) were detected in association with FDC (CD21/35 ϩ cells; brown, upper row) in Peyer's patches from CXCR5 F/F control mice. Sections were counterstained with hematoxylin to detect cell nuclei (blue). Analysis of adjacent sections by PET immunoblot analysis confirmed the presence of prion-specific PK-resistant PrP Sc (blue/black). In contrast, although PrP Sc was detectable in association with FDC in some of the tissues from orally exposed CXCR5 ΔDC mice (right-hand column), many FDC in the Peyer's patches of CXCR5 ΔDC mice lacked PrP Sc accumulation (middle column). Scale bar, 100 m.

FIG 5
Comparison of prion accumulation in the spleens and MLN of orally exposed CXCR5 F/F and CXCR5 ΔDC mice. Mice were orally exposed to ME7 scrapie prions, and spleens and MLN were collected at 70 and 105 dpi. At 70 (A) and 105 (B) dpi, the frequencies of FDC networks containing PrP Sc in the spleens from CXCR5 ΔDC mice were reduced compared to CXCR5 F/F control mice (n ϭ 4 mice/group). Horizontal bar, median. (C) IHC analysis showed that at 105 dpi high levels of disease-specific PrP (PrP d ; brown, middle row, arrow) were detected in association with FDC (CD21/35 ϩ cells; brown, upper row) in spleens from CXCR5 F/F control mice. Sections were counterstained with hematoxylin to detect cell nuclei (blue). Analysis of adjacent sections by PET immunoblot analysis confirmed the presence of prion-specific PK-resistant PrP Sc (blue/black). In contrast, PrP Sc was not detectable in association with the majority of the FDC in spleens (Continued on next page) in the Peyer's patches and spleen is significantly delayed. However, abundant levels of PrP Sc were evident in association with FDC in the MLN of CXCR5 ΔDC and CXCR5 F/F mice at each time point after oral exposure to ME7 scrapie prions ( Fig. 5E to G). This implied that the accumulation of prions upon FDC in the MLN was unaffected by an absence of CXCR5-expressing conventional DC.
Oral prion disease susceptibility is reduced in the specific absence of CXCR5expressing conventional DC. Since oral exposure to a limiting dose of ME7 scrapie prions typically yields a disease incidence of Ͻ100% in wild-type (control) mice, it was used here to enable the effects of conventional DC-restricted CXCR5-deficiency on survival time and prion disease susceptibility to be determined. As anticipated, the majority of the orally exposed CXCR5 F/F (control) mice succumbed to clinical prion disease ( Fig. 6A; mean, 376 Ϯ 12 days; median, 382 days; n ϭ 6/8). In contrast, disease susceptibility was significantly reduced in the orally exposed CXCR5 ΔDC mice, since only two of seven mice succumbed to clinical prion disease with individual survival times of 343 and 371 days: the remaining five mice did not develop clinical signs of prion disease up to at least 501 dpi ( Fig. 6A; P Ͻ 0.023 [Fisher exact test]). All of the brains from the mice in each group that developed clinical signs of prion disease had the characteristic spongiform pathology (vacuolation), disease-specific PrP accumulation, astrogliosis, and microgliosis and typically associated with terminal infection with ME7 scrapie prions ( Fig. 6B and C). The distribution and severity of the spongiform pathology was also similar in the brains of all the clinically affected mice (Fig. 6D), indicating that conventional DC-specific CXCR5 deficiency did not alter the course of CNS prion disease once neuroinvasion had occurred. No histopathological signs of prion disease were detected in the brains of any of the clinically negative mice.
After oral exposure, high levels of ME7 scrapie prions are maintained upon FDC in lymphoid tissues until the terminal stages of disease (5,6,24,29). Here, FDC-associated PrP Sc accumulations were detected in the Peyer's patches, MLN, and spleens of all clinically affected CXCR5 ΔDC and CXCR5 F/F mice (Fig. 7). In comparison, no evidence of PrP Sc accumulation was observed in any of the tissues from the clinically negative mice (Fig. 7). These data indicate that all of the clinically negative mice were free of prions in their lymphoid tissues and brains and were highly unlikely to succumb clinical prion disease after substantially extended survival times. Together, these data show that CXCR5-expressing conventional DC are essential for efficient prion neuroinvasion after oral exposure.
Conventional DC-specific CXCR5 deficiency does not influence prion disease susceptibility when infection is established directly within the CNS. When groups of CXCR5 ΔDC and CXCR5 F/F mice were injected i.c. with ME7 scrapie prions directly into the CNS, all mice succumbed to clinical disease with similar survival times (CXCR5 ΔDC mice, 157 Ϯ 2 days, n ϭ 4/4; CXCR5 fl mice, 162 Ϯ 3 days, n ϭ 4/4; P ϭ 0.302). Histopathological analysis of the brains from each group of clinically affected mice revealed the characteristic neuropathology and PrP Sc accumulation associated with terminal infection with ME7 scrapie prions ( Fig. 8A and B). The distribution and severity of the spongiform pathology was also similar in the brains of the clinically affected CXCR5 ΔDC and CXCR5 F/F mice (Fig. 8C). Therefore, a CXCR5 deficiency specifically in FIG 5 Legend (Continued) from orally exposed CXCR5 ΔDC mice. (D) Prion infectivity levels were assayed in spleens from CXCR5 F/F and CXCR5 ΔDC mice (n ϭ 4 spleens/group) collected at 70 and 105 dpi. Prion infectivity titers (log 10 ID 50 /g tissue) were determined by the injection of tissue homogenates into groups of tga20 indicator mice (n ϭ 4 recipient mice/spleen). Each symbol represents data derived from an individual spleen. Data below the horizontal line indicate disease incidence in the recipient mice Ͻ100% and considered to contain trace levels of prion infectivity. IHC and PET blot analysis suggested that the frequencies of FDC networks containing PrP Sc in the MLN of CXCR5 ΔDC and CXCR5 F/F control mice were similar at 70 (E) and 105 (F) dpi (n ϭ 4 mice/group). (G) Analysis of adjacent sections showed that at 105 dpi high levels of PrP Sc (blue/black, upper row, arrows) were detected in association with FDC (CD21/35 ϩ cells; brown, low row) in MLN from each mouse group. Sections were counterstained with hematoxylin to detect cell nuclei (blue). , reactive astrocytes expressing GFAP (brown), and active microglia expressing IBA1 (brown) were detected in the brains of all orally exposed mice with clinical prion disease. None of these histopathological signs of prion disease were detected in the brains of any of the clinically negative mice up to at least 501 days after oral exposure. Clin., clinical prion disease status; pos., clinically positive; neg., clinically negative; individual survival times are shown. Sections were counterstained with hematoxylin to detect cell nuclei (blue). (C) Immunoblot analysis of brain tissue homogenates confirmed the presence of high levels of prion-specific, relatively PK-resistant PrP Sc within the brains of the clinically affected mice from each group. However, no PrP Sc was detected in the brains of any of the clinically negative CXCR5 ΔDC mice. Samples were treated in the presence (lower (Continued on next page) conventional DC did not affect prion disease pathogenesis or susceptibility when the infection was established directly within the CNS. These data are consistent with the independent observation that prion propagation within the CNS is not affected in CXCR5 Ϫ/Ϫ mice (37). Mice were orally exposed to ME7 scrapie prions, and Peyer's patches, MLN, and spleens were collected from all clinically affected mice and from those free of clinical signs of prion disease at the end of the experiment at 501 dpi. Clin., clinical prion disease status; pos., clinically positive; neg. clinically negative. Individual survival times are shown. High levels of PrP Sc (PET immunoblot; black, arrows) were detected in association with follicular dendritic cells (CD21/35 ϩ cells; brown, arrows) in the Peyer's patches, MLN, and spleens from all clinically affected mice. In contrast, no PrP Sc was detected in tissues from any of the clinically negative survivors at 501 dpi. Sections were counterstained with hematoxylin to detect cell nuclei (blue). CXCR5 F/F Clin. pos, n ϭ 6 mice; CXCR5 F/F Clin. neg, n ϭ 2 mice; CXCR5 ΔDC Clin. pos, n ϭ 2 mice; CXCR5 ΔDC Clin. neg, n ϭ 5 mice.

FIG 6 Legend (Continued)
panel) or absence (upper panel) of PK before electrophoresis. After PK treatment, a typical three-band pattern was observed between molecular mass value of 20 to 30 kDa, representing unglycosylated, monoglycosylated, and diglycosylated isomers of PrP (in order of increasing molecular mass). (D) The severity and distribution of the spongiform pathology (vacuolation) within each brain was scored on a scale of 1 to 5 in nine gray-matter areas: G1, dorsal medulla; G2, cerebellar cortex; G3, superior colliculus; G4, hypothalamus; G5, thalamus; G6, hippocampus; G7, septum; G8, retrosplenial and adjacent motor cortex; and G9, cingulate and adjacent motor cortex. Each point represents the mean vacuolation score Ϯ SE.
Splenectomy before oral prion exposure does not influence disease susceptibility. The data presented above show that the early accumulation of prions in the Peyer's patches (Fig. 4) and spleens (Fig. 5) of orally exposed CXCR5 ΔDC mice was impaired. Although early prion replication upon FDC in the Peyer's patches in the small intestine is essential for efficient neuroinvasion after oral exposure (5, 6, 45), we considered it plausible that the effects observed here on disease susceptibility were due to impaired neuroinvasion from the spleen. If the spleen did play an important role in prion neuroinvasion after oral exposure, we reasoned that splenectomy before prion exposure would similarly impede neuroinvasion and reduce disease susceptibility. To test this hypothesis, wild-type mice were surgically splenectomized and 8 days later orally exposed to a 0.1% dose of ME7 scrapie prions as described above. A parallel group of mice were sham operated before oral prion exposure as a control. Our data show that splenectomy did not influence prion disease pathogenesis or susceptibility as mice from each treatment group succumbed to clinical disease with similar survival times and disease incidences ( Fig. 9A; splenectomized mice, mean ϭ 371 Ϯ 13 days, median ϭ 364 days, n ϭ 6/11; sham-operated mice, mean ϭ 367 Ϯ 11 days, median ϭ 364 days, n ϭ 7/11 [P ϭ 1.0, Fisher exact test]). The distribution and severity of the spongiform pathology were also similar in the brains of all the clinically affected in the brains of all clinically affected mice from each group. Sections were counterstained with hematoxylin to detect cell nuclei (blue). (B) Immunoblot analysis of brain tissue homogenates confirmed the presence of high levels of prion-specific, relatively PK-resistant PrP Sc within the brains of mice from each group. Samples were treated in the presence (ϩ) or absence (Ϫ) of PK before electrophoresis. After PK treatment, a typical three-band pattern was observed between molecular mass values of 20 to 30 kDa, representing unglycosylated, monoglycosylated, and diglycosylated isomers of PrP (in order of increasing molecular mass). (C) The severity and distribution of the spongiform pathology (vacuolation) within the brains of all clinically affected CXCR5 ΔDC and CXCR5 F/F mice was similar. The severity of the vacuolation in each brain was scored on a scale of 1 to 5 in the following gray matter regions: G1, dorsal medulla; G2, cerebellar cortex; G3, superior colliculus; G4, hypothalamus; G5, thalamus; G6, hippocampus; G7, septum; G8, retrosplenial and adjacent motor cortex; and G9, cingulate and adjacent motor cortex. Each point represents the mean vacuolation score Ϯ the SE. splenectomized and sham-operated mice (Fig. 9B). These data clearly show that the spleen does not play an important role in neuroinvasion after oral prion exposure. This suggests that the major effects observed on oral prion disease pathogenesis and susceptibility in CXCR5 ΔDC mice were not due to impaired neuroinvasion from their spleens.

DISCUSSION
The early replication of prions upon FDC in Peyer's patches is essential for efficient neuroinvasion from the small intestine after oral exposure (3)(4)(5)(6)35). The cellular mechanism by which orally acquired prions are initially conveyed to FDC in order to establish host infection was uncertain. Production of the chemokine CXCL13 by FDC and follicular stromal cells plays an important role in attracting CXCR5-expressing cells toward the FDC-containing B-cell follicles of secondary lymphoid organs (31,36,51). Here, a unique compound transgenic mouse model was created in which CXCR5 deficiency was restricted to CD11c ϩ conventional DC. In the specific absence of CXCR5-expressing conventional DC in CXCR5 ΔDC mice, the early accumulation of prions upon FDC in the Peyer's patches and spleen was reduced, significantly reducing disease susceptibility. Our data suggest that CXCR5-expressing conventional DC are required for the efficient delivery of orally acquired prions to FDC in the Peyer's patches of the small intestine in order to achieve host infection.
The data presented here help advance our understanding of how orally acquired prions utilize an elegant cellular relay to establish host infection from the lumen of the small intestine. Prions are first transported across the gut epithelium by M cells in the FAE overlying the Peyer's patches (29,45,52,53). This initial transport of prions across the gut epithelium by M cells is essential to establish infection in Peyer's patches (29,45). Antigens and particles that have been transported across the gut epithelium by M cells are released into their underlying basolateral pockets where they are sampled by The severity of the vacuolation in each brain was scored on a scale of 1 to 5 in the following gray-matter regions: G1, dorsal medulla; G2, cerebellar cortex; G3, superior colliculus; G4, hypothalamus; G5, thalamus; G6, hippocampus; G7, septum; G8, retrosplenial and adjacent motor cortex; and G9, cingulate and adjacent motor cortex. Each point represents the mean vacuolation score Ϯ the SE. mononuclear phagocytes (54). Our data suggest that prions are subsequently acquired by conventional DC (24,55) and propagated by them in a CXCL13-CXCR5-dependent manner toward FDC within the B-cell follicles of Peyer's patches. The prions are then acquired by FDC and amplified upon their surfaces above the threshold required to achieve neuroinvasion (4-6, 42, 55). Studies from experimental mice indicate that prions can establish infection in enteric nerves within 21 days after oral exposure (4,55). How the prions spread between the FDC and enteric nerves is not known, but a role for conventional DC in this process has been proposed (56)(57)(58).
Whether these apparent prion propagation activities are restricted to certain conventional DC populations is uncertain, but the specific depletion of CD8 ϩ CD11c ϩ cells does not influence oral prion disease pathogenesis, implying that CD8 ϩ conventional DC do not play a role (26). Indeed, within Peyer's patches CD8 ϩ conventional DC are rarely encountered within the subepithelial dome region immediately beneath the M-cell-containing FAE. Instead, these cells are mostly localized within the T-cell-rich interfollicular regions (59). Splenic plasmacytoid DC also accumulate high levels of infectious prions during infection (60) but are unlikely to play a significant role in prion propagation since they do not migrate in the lymphatics during the steady state or after activation (61).
Whether the initial uptake of prions by conventional DC involves a specific receptor is uncertain. Most mononuclear phagocyte populations express cellular PrP C (62-64), but the propagation of prions from peripheral exposure sites to FDC is not affected by the absence of PrP C expression in hematopoietic cells (42,48,(65)(66)(67). Conventional DC, like the FDC in the B-cell follicles (28,(68)(69)(70)(71), can acquire prions after their opsonization by complement components such as C1q and C3 (72,73). Depending on their location, subset, and activation status, certain populations of conventional DC express a variety of complement receptors (CR), including CR1 (CD35), CR2 (CD21), CR4 (CD11c/CD18), calreticulin, CD93, and SIGN-R1 (CD209b), but it is uncertain whether they also facilitate the complement-mediated uptake of prions by these cells (72,73). However, SIGN-R1 is unlikely to play a role since its transient downregulation before prion exposure does not influence disease pathogenesis (74). The possibility also cannot be excluded that conventional DC simply acquire prions nonspecifically as they constitutively sample their microenvironments via fluid-phase micropinocytosis.
Although the data in the present study suggest that CXCR5-expressing conventional DC are required for the efficient propagation of prions to FDC in Peyer's patches, it is not known how they are transferred between these cell populations. Follicular B cells within the subcapsular sinus of lymph nodes acquire lymph-borne immune complexes via their CR and deliver them to FDC. The higher immune-complex-binding affinity of FDC appears to enable them to strip the B cells of their immune-complex cargo as they migrate into the follicles (75)(76)(77). Conventional DC can retain PrP Sc on both the cell surface and in intracellular compartments (73). Since the expression level of CR1 and CR2 on FDC is greater than the surrounding lymphocytes and leukocytes, it is plausible that complement-opsonized prions are stripped from the surfaces of conventional DC by FDC in a similar manner. Alternatively, it is plausible that conventional DC might indirectly deliver infectious prions toward FDC. Since CR-expressing marginal zone B cells in the spleen can shuttle complement-opsonized antigens to the FDC (78), B cells may similarly strip opsonized prions from conventional DC in the vicinity of the B cell follicle and deliver them to the FDC.
Tunneling nanotubes (TNT) are thin, membrane-bound cylinders of cytoplasm that can connect neighboring cells to enable cell-to-cell communication and the intercellular transfer of plasma membrane or cytoplasmic components. These TNT structures are exploited by HIV-1 as a means of intercellular transfer between T cells (79) and to shuttle virus-encoded immunosuppressive factors from infected macrophages to B cells to suppress the humoral response (80). In vitro coculture studies suggest prions may also transfer between conventional DC to neurons via endolysosomal vesicles within TNT (57, 58, 81, 82). Infectious prions may also be released from infected cells in the form of small endosome-derived vesicles termed exosomes (83) and, in doing so, enhance their ability to infect neighboring cells (60). Whether the transfer of prions between conventional DC and FDC occurs in vivo via one or a combination of the above processes remains to be determined.
We have previously shown that although the transient depletion of CD11c ϩ cells impedes the accumulation of prions in Peyer's patches and reduces disease susceptibility, a small number of mice did develop clinical disease (24). In the present study, we observed a similar effect on prion disease pathogenesis in the absence of CXCR5expressing conventional DC: prion accumulation was impeded in the Peyer's patches and spleens of orally exposed CXCR5 ΔDC mice, and disease susceptibility was reduced, but a small number of mice eventually succumbed to clinical prion disease. These studies suggest that conventional DC provide an efficient route by which prions are initially conveyed to FDC. However, in their absence a limited amount of prions are able to avoid clearance by cells such as tissues macrophages (14,15) and reach the FDC via alternative and less efficient routes (27,73,84).
Studies in experimental mice show that after oral exposure, prions accumulate first in the Peyer's patches in the small intestine and then spread via the blood and lymph to most other secondary lymphoid tissues, including the spleen (4-6, 24, 50). Analysis of the pathogenesis of natural sheep scrapie suggests a similar temporal distribution (9). In the present study, the early accumulation of prions within the MLN of CXCR5 ΔDC mice was not impaired, suggesting that the prions had established infection within the MLN independently of CXCR5-expressing conventional DC. Although prions also replicate upon FDC in the MLN soon after oral exposure, our studies in mice have shown that the MLN do not influence prion neuroinvasion from the intestine (5,6,45).
We also showed that early prion replication in the spleens of CXCR5 ΔDC mice was significantly reduced. This raised the possibility that the significantly reduced prion disease susceptibility observed in these mice was, at least in part, due to impaired neuroinvasion from their spleens. However, splenectomy of conventional mice prior to oral exposure with ME7 scrapie prions did not influence survival times or disease susceptibility. These data are consistent with an earlier study using mouse-passaged 139A scrapie prions which also suggested that the spleen does not play a role in prion disease pathogenesis after intragastric exposure (85), and our subsequent demonstrations that the GALT in the small intestine are the important sites of early prion accumulation and neuroinvasion after oral exposure (5,6).
Many studies have defined conventional DC based on their expression of distinct cell surface markers such as the integrin CD11c. Murine conventional DC express CD11c highly, and this has been used as a reliable marker for these cells in a wide range of studies. However, the expression of this integrin is not restricted to conventional DC since certain macrophage populations and activated monocytes can also express CD11c to some degree (86,87). The recent discovery that expression of the zinc finger transcription factor ZBTB46 is restricted to conventional DC (88,89) provides an excellent opportunity to create conditional knockout mouse models to further define the separate roles of conventional DC and tissue macrophages in prion disease pathogenesis (90).
In conclusion, our data demonstrate that prions exploit conventional DC in order to facilitate their efficient propagation to FDC in Peyer's patches in order to establish host infection. The status of conventional DC is dramatically influenced by microbial stimuli and inflammatory conditions in the intestine and can enhance the uptake of certain pathogens from the gut lumen (91)(92)(93). Treatments that prevent the initial uptake and accumulation of prions in Peyer's patches can significantly reduce disease susceptibility (4,24,29). Therefore, a thorough understanding of the cellular and molecular mechanisms which prions exploit to establish infection upon FDC in the GALT may help to identify important factors that influence disease susceptibility or novel targets for prophylactic intervention.

MATERIALS AND METHODS
Mice. The following mouse strains were used in this study where indicated: CD11c-Cre (38); and tga20, which overexpress PrP C (94). Cxcr5 F/F mice were produced by Ozgene (Bentley DC, Australia) and created by introducing loxP sites flanking exon 2 of the Cxcr5 gene via homologous recombination. All mice were bred and maintained on a C57BL/6J mice background and maintained under specificpathogen-free conditions. In some studies C57BL/Dk mice were also used, where indicated. All studies and regulatory licenses were approved by University of Edinburgh's ethics committee and carried out under the authority of a UK Home Office Project License.
Macroscopic analysis of secondary lymphoid tissues. Mice were injected intraperitoneally with 300 l of 1% Chicago Sky Blue 6B in sterile phosphate-buffered saline. Mice were culled 7 days later, and the presence or absence of secondary lymphoid organs was determined macroscopically (41).
Chemotaxis assays. Ex vivo chemotaxis assays were performed as described previously (33,95). Briefly, MLN cells in RPMI medium with 5% fetal calf serum (Invitrogen) were placed in the upper chamber of a 5-m Transwell insert in a 24-well plate (Corning, Corning, NY). The lower chamber contained 200 ng of either CXCL13 or CCL21 (Peprotech, London, UK)/ml. Five technical replicates were performed. After 24 h, the upper chamber was discarded, and the cells in the lower chamber were collected and immunostained with anti-CD3 (clone 17A2) to detect T cells, anti-B220 (clone RA3-2GB) to detect B cells, and anti-CD11c (clone N418; BioLegend, London, UK) and resuspended in 500 l of fluorescence-activated cell sorting buffer. The number of migrated cells was counted for 60 s using a LSRFortessa flow cytometer (BD Biosciences).
Prion exposure and disease monitoring. For oral exposure, mice were fed individual food pellets dosed with 50 l of a 0.1% (wt/vol) dilution of scrapie brain homogenate (containing approximately 2.5 ϫ 10 3 i.c. 50% infectious dose [ID 50 ] units) prepared from mice terminally affected with ME7 scrapie prions according to our standard protocol (4-6, 24, 29, 96). During the dosing period mice were individually housed in bedding-and food-free cages with water provided ad libitum. A single priondosed food pellet was then placed in the cage. The mice were returned to their original cages (with bedding and food ad libitum) as soon as the food pellet was observed to have been completely ingested. The use of bedding-and additional food-free cages ensured easy monitoring of consumption of the prion-contaminated food pellet. For i.c. exposure, mice were injected with 20 l of a 1% dilution of scrapie brain homogenate. After prion exposure, mice were coded and assessed weekly for signs of clinical disease and culled at a standard clinical endpoint. The clinical endpoint of disease was determined by rating the severity of clinical signs of prion disease exhibited by the mice. The mice were clinically scored as "unaffected," "possibly affected," or "definitely affected" according to standard criteria that typically present in mice with terminal ME7 scrapie prion disease. Clinical signs following infection with the ME7 scrapie agent may include weight-loss, starry coat, hunched and jumpy behavior (at early onset) progressing to limited movement, upright tail, wet genitals, decreased awareness, discharge from eyes and/or blinking eyes, and ataxia of the hind legs. The clinical endpoint of disease was defined in one of the following ways: (i) the day on which a mouse received a second consecutive "definite" rating, (ii) the day on which a mouse received a third "definite" rating within four consecutive weeks, or (iii) the day on which a mouse was culled in extremis. Survival times were recorded for mice that did not develop clinical signs of disease during the observation period. Prion diagnosis was confirmed by histopathological assessment of vacuolation in the brain. For the construction of lesion profiles, vacuolar changes were scored in nine distinct gray-matter regions of the brain, as described previously (97).
For bioassay of prion infectivity, individual spleens were prepared as 1% (wt/vol) homogenates. For each tissue homogenate, groups of tga20 indicator mice (n ϭ 4/homogenate) were injected i.c. with 20 l of homogenate. The prion infectivity titer in each sample was determined from the mean incubation period in the indicator mice by reference to a dose/incubation period response curve for ME7 scrapie-infected spleen tissue serially titrated in tga20 mice using the following relationship: y ϭ 9.4533 to 0.0595x (where y is the log ID 50 U/20 l of homogenate, and x is the incubation period [R 2 ϭ 0.9562]). Since the expression level of cellular PrP C controls the prion disease incubation period, tga20 mice overexpressing PrP C are extremely useful as indicator mice in scrapie agent infectivity bioassays because they succumb to disease with much shorter incubation times than conventional mouse strains (94).
Splenectomy. Groups of C57BL/Dk mice were surgically splenectomized and orally exposed to prions 8 days later. Briefly, an ϳ1-cm incision was made in the left upper quadrant of the abdominal wall, and the spleen was identified. Blood supply to the organ was tied off with sutures before removal. The incision site was then also closed with sutures. Isoflurane anesthesia was used throughout the surgical procedure. After surgery, the mice were closely monitored throughout the recovery period and given buprenorphine hydrochloride as an analgesic. Sham-operated mice (laparotomy incision and suture repair with no spleen removal) were used as a control.
Intestines, MLN, and spleens were also removed and snap-frozen at the temperature of liquid nitrogen. Serial frozen sections (6 m in thickness) were cut on a cryostat and immunostained with the following antibodies: FDC were visualized by staining with MAb 8C12 to detect CR1 (CD35; BD Biosciences) and cellular PrP C was detected using PrP-specific polyclonal antibody (pAb) 1B3 (98). Where appropriate, sections were counterstained with DAPI (4=,6=-diamidino-2-phenylindole) to detect cell nuclei (Life Technologies).
For the detection of disease-specific PrP (PrP d ) in intestines, MLN, spleens, and brains, tissues were fixed in periodate-lysine-paraformaldehyde fixative and embedded in paraffin wax. Sections (thickness, 6 m) were deparaffinized and pretreated to enhance the detection of PrP d by hydrated autoclaving (15 min, 121°C, hydration) and subsequent immersion in formic acid (98%) for 5 min. The sections were then immunostained with 1B3 PrP-specific pAb. For the detection of astrocytes, brain sections were immunostained with anti-glial fibrillary acidic protein (GFAP; Dako, Ely, UK). For the detection of microglia, deparaffinized brain sections were first pretreated with citrate buffer and subsequently immunostained with anti-ionized calcium-binding adaptor molecule 1 (Iba-1; Wako Chemicals GmbH, Neuss, Germany). For the detection of FDC in intestines, MLN, and spleens, deparaffinized sections were first pretreated with Target Retrieval Solution (Dako) and then immunostained with anti-CD21/35 (clone 7G6; BD Biosciences). PET immunoblot analysis was used to confirm the PrP d detected by IHC was proteinase K-resistant PrP Sc (49). Membranes were subsequently immunostained with 1B3 PrP-specific pAb.
For light microscopy, following the addition of primary antibodies, biotin-conjugated species-specific secondary antibodies (Stratech, Soham, UK) were applied, and immunolabeling was revealed using horseradish peroxidase-conjugated avidin-biotin complex (ABC kit; Vector Laboratories, Peterborough, United Kingdom) and visualized with DAB (diaminobenzidine; Sigma). Sections were counterstained with hematoxylin to distinguish cell nuclei. For fluorescence microscopy, following the addition of primary antibody, streptavidin-conjugated or species-specific secondary antibodies coupled to Alexa Fluor 488 (green), Alexa Fluor 594 (red), or Alexa Fluor 647 (blue) dyes (Life Technologies) were used. The sections were subsequently mounted in fluorescent mounting medium (Dako). Images of whole-mount immunostained intestinal pieces and cryosections were obtained using a Zeiss LSM710 confocal microscope (Zeiss, Welwyn Garden City, UK).
Image analysis. For morphometric analysis, images were analyzed using ImageJ software (http:// rsb.info.nih.gov/ij/) as described on coded sections (42,45,99). Background intensity thresholds were first applied using an ImageJ macro which measures pixel intensity across all immunostained and nonstained areas of the images. The obtained pixel intensity threshold value was then applied in all subsequent analyses. Next, the number of pixels of each color (black, red, green, yellow, etc.) were automatically counted. For these analyses, data are presented as the proportion of positively stained pixels for a given IHC marker per total number of pixels (all colors) in the specific area of interest. In each instance, typically six images were analyzed per mouse, from tissues from multiple mice per group (n ϭ 6 to 8 mice/group). Details for all of the sample sizes for each parameter analyzed are provided in the figure legends.
Statistical analyses. Unless indicated otherwise, data are presented as means Ϯ the standard errors (SE), and significant differences between samples in different groups were evaluated by Student t test. In instances where there was evidence of nonnormality, data were analyzed by nonparametric analysis of variance (Kruskal-Wallis test) with Dunn's multiple comparison post hoc test. P values of Ͻ0.05 were accepted as significant.