|Home | About | Journals | Submit | Contact Us | Français|
Dendritic cells (DC) are suspected to be involved in transmissible spongiform encephalopathies, including bovine spongiform encephalopathy (BSE). We detected the disease-specific, protease-resistant prion protein (PrPbse) in splenic DC purified by magnetic cell sorting 45 days after intraperitoneal inoculation of BSE prions in immunocompetent mice. We showed that bone marrow-derived DC (BMDC) from wild-type or PrP-null mice acquired both PrPbse and prion infectivity within 2 h of in vitro culture with a BSE inoculum. BMDC cleared PrPbse within 2 to 3 days of culture, while BMDC infectivity was only 10-fold diminished between days 1 and 6 of culture, suggesting that the infectious unit in BMDC is not removed at the same rate as PrPbse is removed from these cells. Bone marrow-derived plasmacytoid DC and bone marrow-derived macrophages (BMM) also acquired and degraded PrPbse when incubated with a BSE inoculum, with kinetics very similar to those of BMDC. PrPbse capture is probably specific to antigen-presenting cells since no uptake of PrPbse was observed when splenic B or T lymphocytes were incubated with a BSE inoculum in vitro. Lipopolysaccharide activation of BMDC or BMM prior to BSE infection resulted in an accelerated breakdown of PrPbse. Injected by the intraperitoneal route, BMDC were not infectious for alymphoid recombination-activated gene 20/common cytokine γ chain-deficient mice, suggesting that these cells are not capable of directly propagating BSE infectivity to nerve endings.
Prion diseases, such as Creutzfeldt-Jakob disease (CJD) in humans, scrapie in sheep, and bovine spongiform encephalopathy (BSE) in cattle, are transmissible spongiform encephalopathies that are induced by a novel class of infectious agents thought to be composed of a misfolded, host-encoded, protease-resistant protein, the disease-specific prion protein (PrPsc, PrPbse, or PrPcjd for the isoform of scrapie, BSE, or CJD, respectively). A new form of human disease, called variant CJD (vCJD), emerged as a consequence of transmission from cattle to human due to consumption of BSE-contaminated food. Blood is infectious in sheep infected with the cattle-derived BSE strain, in mice infected with the human-derived vCJD strain, and probably in human vCJD (10, 15, 25, 32). PrP-null (PrP0) mice are resistant to infection by prions, thus clearly establishing the role of normal, self PrP in prion propagation (6). Immune cells are probably involved in prion pathogenesis, as severely combined immunodeficient (SCID) mice are resistant to peripheral inoculation with a scrapie agent (20). Following a peripheral inoculation, prions have a primary replication phase in the spleen and the lymph nodes (LN) before reaching the brain. Splenic infectiosity is associated with B and T lymphocytes and follicular dendritic cells (FDC) (29, 34). Specific complement components are involved in the initial trapping of prions in lymphoreticular organs early after infection (21). Prion replication does not seem to alter the immune system or to be pathogenic in tissues other than the central nervous system (1). On the other hand, chronic inflammatory conditions may expand the tissue distribution of prions in inflammatory foci, with ectopic induction of PrP-expressing FDC (13).
How prions are transported from the site of peripheral exposure to FDC and to the nervous system is not known. Hematopoietic cells, such as macrophages and dendritic cells (DC), are suspected to be involved in prion neuroinvasion because of their implication in the uptake and transport of antigens from sites of exposure to lymphatic tissue (3). PrPsc was detected in LN resident macrophages in scrapie-infected sheep as well as in scrapie-infected tumor necrosis factor receptor 1-null mice (2, 14, 33). LN from these mice were infectious, suggesting that cells of the immune system, such as macrophages, DC, or lymphocytes, could act as a replication site or as a reservoir for prions. The disease-associated form of PrP was also detected in LN DC from scrapie-infected rats, as well as in cerebral blood vessel DC in human sporadic CJD and vCJD (16, 22). CD205+ DC were detected in several regions of the brain, such as the cerebral cortex and thalamus, in mice intraperitoneally infected with the mouse-adapted KFu strain of Gerstmann-Straussler-Scheinker syndrome (36). DC purified from spleens of scrapie-infected mice were infectious for immunodeficient mice following peripheral inoculation, but only when live DC were injected (4).
We focused in the present study on the role of DC in BSE physiopathology. As DC are a heterogeneous population of cells, including myeloid, lymphoid, and plasmacytoid subsets, our objective was to determine which DC populations express PrP and are responsible for transmitting the disease. Following combined intraperitoneal and intracerebral injections with BSE cow brain homogenate, only 4/22 SCID mice developed prion disease (5). In contrast, 100% of normal competent mice died of BSE. These findings indicate that the immune system probably plays an important role in the cow-mouse species barrier. To solve the species barrier problem, we used the BSE mouse-adapted prion strain 6PB1 in challenges with immunodeficient mice (23). We first investigated the uptake of PrPbse by spleen DC in vivo following a peripheral infection by BSE prions. As spleen DC cannot be maintained in culture for more than a few days, we chose the bone marrow-derived DC (BMDC) model to investigate the uptake of PrPbse by DC in vitro. We analyzed both prion infectivity and PrPbse processing in cultures of BMDC incubated with BSE prions. Then, we compared this processing to those of bone marrow-derived plasmacytoid DC (BMpDC), bone marrow-derived macrophages (BMM), and T and B lymphocytes isolated from spleen by magnetic antibody cell sorting.
Six- to 8-week-old C57BL/6J (Janvier, Le Genest-Saint-Isle, France), PrP0 (CDTA, France), B6129SF1/J (the F1 cross between C57BL/6J and 129S, stock number 101043; Jackson ImmunoResearch Laboratories), and recombination-activated gene 20/common cytokine γ chain-deficient (RAG γc0) immunodeficient mice were used for this study (7, 11, 12). RAG γc0 mice originated on a 129SF1 background were backcrossed on the C57BL/6 background. All animals were housed in level 3 care facilities of the Pasteur Institute, officially registered for prion experimental studies with rodents (Ministry agreement number A 75-15-27 for animal care facilities and agreement number 75-585 for animal experimentation). RAG γc0 immunodeficient mice were kept in isolators under specific-pathogen-free conditions.
The mouse-adapted BSE strain 6PB1 was kindly provided by D. Dormont (CEA, Fontenay-aux-Roses, France). It was stabilized and propagated in C57BL/6 mice (24). The inoculum was a brain homogenate at 10% (wt/vol) in 5% glucose solution from BSE-infected mice at the terminal stage of disease, routinely titrating 5 × 108 to 5 × 109 mean 50% lethal doses (LD50) per gram in intracerebral-infection challenges. The mean (±standard deviation) survival time was around 181 ± 5.4 days after intracerebral infection or 315 ± 27.3 days after intraperitoneal challenge (23). Healthy mouse brain homogenate was used as a negative control.
Scrapie-associated fibrils (SAF) were purified from infected brains according to the method developed by Vilette et al. (38), which was slightly modified (38). Briefly, 1 volume of 20% NaCl (Sigma, St. Louis, MO) was added to 1 volume of 20% brain homogenate (BSE or control). After homogenization and addition of 1 volume of a mix containing 20% Sarkosyl (Interchim, Montluçon, France), 2% SB3-14 (Calbiochem, VWR, Fontenay-sous-Bois, France), and 2 mM Tris-HCl (pH 7.4), samples were incubated for 15 min at room temperature. They were then centrifuged at 26,500 × g for 2 h at 20°C on a 20% sucrose cushion. Pelleted material was resuspended in Tris-borate-sodium buffer and heated at 80°C for 10 min before use for cell infections.
DC were purified by magnetic cell sorting as previously described (30). Briefly, spleens were removed from mice and treated for 45 min at 37°C with 400 U/ml of collagenase type IV and 50 μg/ml of DNase I (Roche, Mannheim, Germany) in RPMI 1640. After inhibition of collagenase activity with 6 mM EDTA in phosphate-buffered saline (PBS), spleens were dissociated in Ca2+- and Mg2+-free PBS in the presence of 2.5 mM EDTA, 0.5% fetal calf serum (Invitrogen, Cergy-Pontoise, France), and 1% antibiotics (penicillin, 100 U/ml; streptomycin, 100 μg/ml). Single spleen cell suspensions were prepared and incubated with colloidal superparamagnetic microbeads conjugated to anti-CD11c monoclonal antibody (MAb) (MACS-anti-CD11c, N418 clone; Miltenyi Biotec, Bergisch-Gladbach, Germany) by following the manufacturer's instructions. CD11c+ cells were positively selected by two serial passages with miniMACS magnetic cell sorting (Miltenyi Biotec). The purity of DC preparations was superior to 70% after the first column and superior to 93% after the second one.
T and B cells were purified from mouse spleen by magnetic cell sorting with pan-T and -B cell isolation kits, respectively, following the manufacturer's instructions (Miltenyi Biotec). The purity of both T and B populations was superior to 97%.
Complete medium consisted of RPMI 1640 containing l-alanyl-l-glutamine dipeptide (Invitrogen, Cergy-Pontoise, France) supplemented with 10% fetal calf serum (ICN Pharmaceuticals France S.A., Orsay, France), 50 μM 2-mercaptoethanol, and antibiotics (penicillin, 100 U/ml; streptomycin, 100 μg/ml).
BMM, BMDC, and BMpDC were generated from bone marrow precursors, as previously described (17). Briefly, bone marrow cells from C57BL/6 or PrP-null mice were harvested and plated at 2 × 105 cells/ml in 60-mm cell culture dishes (BD Falcon, Bedford, Mass.) in complete medium with 5% of a supernatant containing granulocyte-macrophage colony-stimulating factor (GM-CSF), 5% of a supernatant containing fms-like tyrosine kinase 3 ligand (Flt3-L), or 10% of a supernatant with macrophage colony-stimulating factor for the generation of BMDC, BMpDC, or BMM, respectively. J558 GM-CSF (a kind gift from T. Rolink, Basel, Switzerland), Sp210 Flt3-L (a kind gift from R. Rottapel, Toronto, Canada), and L929 (ATCC, Teddington, Middlesex, United Kingdom) cell lines were used to produce GM-CSF, Flt3-L, and macrophage colony-stimulating factor, respectively. BMpDC were purified prior to infection by magnetic cell sorting using a murine plasmacytoid dendritic cell antigen 1 isolation kit following the manufacturer's instructions (Miltenyi Biotec). The purity of BMpDC preparations was superior to 86%.
T and B splenic cells were infected just after purification by magnetic cell sorting. Infections of BMDC and BMM were performed on day 6. Infection of BMpDC was performed on day 6, after purification by magnetic cell sorting. For in vitro activation, BMM or BMDC were incubated at day 5 for 24 h with 10 μg/ml of lipopolysaccharide (LPS) (Invivogen, San Diego, Calif.) prior to infection. The murine SAF from 1 mg of brain homogenate of uninfected or BSE terminal-stage mice or 1 mg of direct brain homogenate was incubated with 4 × 106 cells in 60-mm cell culture dishes (BD Falcon, Bedford, Mass.). Inocula were previously heated at 80°C for 10 min to avoid bacterial contamination. The nonadherent and semiadherent cells were recovered by flushing the plates with PBS containing 5 mM EDTA and washed in PBS before fluorescence-activated cell sorter (FACS) and Western blot analysis or before injection into wild-type or RAG γc0 mice.
Briefly, tissue homogenates (5% [wt/vol] in 5% glucose) or cell lysates (8 × 105 cells in 0.5% sodium deoxycholate [Sigma] and 0.5% NP-40 [Coger, France]) were treated or not for 1 h at 37°C with 10 or 40 μg/ml of proteinase K (PK) (Sigma, France), respectively. Laemmli buffer (Bio-Rad, Ivry sur Seine, France) containing 3 M (final concentration) urea was added, and 20 μl of each extract, corresponding to 100 μg of brain or to 1.6 × 105 cultured cells, was used for PrPbse detection. Samples were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis in 12% polyacrylamide gels and electroblotted onto nitrocellulose membranes. Immunodetection was done by using a pool at 1 μg/ml of two or three mouse monoclonal anti-PrP antibodies, all raised against hamster PrP (SAF32, recognizing the octo-repeat region located in the N-terminal part of PrP; SAF84, recognizing an epitope located between 160 to 170 residues; and SAF83, with an unknown recognized epitope [all from Spi-Bio, Massy, France]), followed by a peroxidase-conjugated goat anti-mouse antibody (Jackson ImmunoResearch Laboratories). Immunoreactivity was visualized by enhanced chemiluminescence (ECL kit; Pierce, France).
Immunocytochemistry was run on fixed, frozen DC as previously described, with minor modifications (37). Cells were fixed for 15 min in 4% formaldehyde in PBS, incubated in cold acetone for 10 min, and cryoprotected in 15% sucrose for 15 min before freezing. Two combined monoclonal anti-PrP antibodies were used: SAF84 (1 μg/ml) and SAF83 (1 μg/ml) from SpiBio. SAF84 gave a weaker immunostaining than SAF83. Slides were treated with 5 M guanidine for 5 min before PrP staining. Staining was performed with a mouse ABC staining system (Santa Cruz Biotechnology). Images were taken with a Leica DMLBHC microscope by using a Leica DC 100 charge-coupled-device camera and Leica IM50 image manager software (Leica Microsystems, Rueil-Malmaison, France).
Cells were washed and resuspended in PBS containing 1% bovine serum albumin and 0.2% sodium azide (both purchased from Sigma), incubated with anti-CD16/CD32 Fc Block (2.4G2 clone; BD Pharmingen, San Diego, Calif.) to avoid antibody binding via Fc, and labeled at 4°C for 20 min with different combinations of the following fluorescein isothiocyanate-, phycoerythrin-, or allophycocyanin-conjugated MAbs: anti-B220 (RA3-6B2), anti-CD11c (HL-3), anti-CD11b (M1/70), anti-CD3 (17A2), polyclonal goat anti-mouse immunoglobulin (Ig), and appropriate Ig isotype controls. All MAbs were purchased from BD Pharmingen, except anti-PrP (SAF83; SpiBio, France). SAF83 was conjugated to fluorescein isothiocyanate (Sigma) or to Alexa 488 by use of a Zenon mouse IgG labeling kit (Molecular Probes, PoortGebouw, The Netherlands) prior to staining. Cells were analyzed with CellQuestPro software (BD Biosciences) on a FACSCalibur flow cytometer (BD Biosciences, San Diego, Calif.) situated in our level 3 care facilities.
Animals were examined three times per week for clinical signs of BSE, which could include bradykinesia, waddling gait, poor coat condition, terminal incoordination, and weight loss. BSE-infected mice, animals affected by injuries, and surviving mice were euthanized by cervical dislocation at the terminal stage of the disease or at 650 days postinoculation by following the Pasteur Institute's animal experimentation regulations. BSE diagnosis was confirmed by detection of PrPbse by Western blotting after PK digestion. The survival period was calculated as the interval between the first inoculation and the death, or in extremis, stage. The BSE incidence and the mean survival time were determined for each group of mice. Statistics were performed by use of Prism software (GraphPad Software, San Diego, Calif.).
End point titration was performed by intracerebral inoculations of serial decimal dilutions of the infectious inoculum, ranging from 10−2 to 10−10. Twenty microliters of inoculum was injected into groups of four to six C57BL/6 mice. The infectious titer was determined by the calculation of the LD50 according to the formula of Reed and Muench (35).
At day 1 or 6 postinfection, wild-type or PrP-null BMDC semiadherent cells were washed and suspended in PBS. Live BMDC (4× 106) in 500 μl PBS were immediately injected by the intraperitoneal route into groups of four to six RAG γc0 mice. In parallel, as a control, four to six RAG γc0 mice were injected by the same route with the same amount of BMDC that had been killed by three cycles of freezing and thawing.
To unravel the role of DC in BSE pathology, we investigated the expression of both normal and misfolded PrP isoforms in spleen DC of mice inoculated or not with BSE prions. For noninfected C57BL/6 mice, FACS analysis showed that splenic CD11c+ DC express membrane PrP (Fig. (Fig.1A).1A). A high expression of PrP was noted in 35% of splenic CD11c+ DC. As expected, no PrP expression was detected in PrP0 mice (Fig. (Fig.1A).1A). To determine which subsets of splenic DC express PrP, spleen DC were analyzed by flow cytometry for the expression of CD11c, CD11b, B220, and PrP. As shown in Fig. Fig.1B,1B, myeloid (CD11c+ CD11b+), lymphoid (CD11c+ CD11b−), and plasmacytoid (CD11c+ B220+) DC express PrP on their surfaces. The highest expression of PrP was detected in lymphoid DC and the lowest in plasmacytoid DC. Similarly, we found that cultured BMDC, BMpDC, and BMM express PrP on their surfaces (data not shown).
In BSE-infected C57BL/6 mice, at 45 days after intraperitoneal inoculation, PrPbse in total spleen cells was detectable by Western blotting as well as in CD11c+ DC purified by magnetic cell sorting (Fig. (Fig.2A).2A). DC from noninfected control mice were negative for PrPbse. These data were confirmed by PrP immunostaining of splenic DC purified by magnetic cell sorting (Fig. (Fig.2B).2B). A strong intracellular labeling of PrP was observed with DC purified from spleens of mice with preclinical BSE but not with those from PrP0 or noninfected C57BL/6 mice. Altogether, these data strongly suggest that murine splenic DC express normal PrP and PrPbse, and thus these cells could be infected in vivo in the BSE preclinical stage.
Cell cultures were exposed at day 6 to BSE brain homogenate or to PrPbse purified in SAF. The control inoculum was LPS, brain homogenate, or SAF from healthy mice.
Western blot analysis showed that PrPbse of the BSE inoculum was taken up within 2 h by BMDC (Fig. (Fig.3A).3A). PrPbse was then cleared in 2 to 3 days. Then, long-term cultures showed no PrPbse expression, strongly suggesting that BMDC did not multiply BSE prions in vitro (data not shown). BMDC from PrP0 mice also acquired and cleared PrPbse in vitro with a similar time course (Fig. (Fig.3B).3B). Similarly, SAF were taken up and cleared in vitro (Fig. 3A and B). Normal PrP of the control inoculum was processed and rapidly degraded by BMDC from normal or PrP0 mice (Fig. 3A and B). Similarly, BMM and BMpDC acquired and rapidly cleared PrP of the control or BSE inoculum (Fig. 3C and D). LPS activation of BMDC or BMM before BSE infection resulted in an accelerated breakdown of PrPbse (Fig. (Fig.3G).3G). Splenic T or B cells acquired only small amounts of PrPbse and did not multiply it (Fig. 3E and F). To verify that PrPbse was processed by BMDC, PrP immunocytochemistry was run on BMDC of PrP0 mice after their exposure to the BSE inoculum (Fig. (Fig.4).4). A strong intracellular labeling of PrP in PrP0 BMDC was detected after 24 h of exposure to the BSE inoculum (Fig. (Fig.4).4). As no staining was detected in BMDC treated or not with control inoculum, these data indicate that only PrPbse, but not normal PrP, was detected in BSE inoculum-treated BMDC, confirming the Western blot analysis results. Furthermore, these immunocytochemistry data also indicate that PrPbse was internalized by BMDC before clearing. Altogether, these findings highlight the specific high capacity of DC and macrophages to capture and clear PrPbse in vitro.
To quantify prion infectivity in DC cultures, a time course of in vitro infection was performed by in vivo titration of cell lysates (Table (Table1).1). After 1 or 6 days of BMDC culture with the BSE inoculum, adherent cells were washed twice before being scraped in PBS, frozen, and injected without filtration into the brains of C57BL/6 mice. We found that following 1 or 6 days of in vitro incubation with BSE brain homogenate, PrP-expressing or PrP0 BMDC were highly infectious in vivo. Following 1 day of BSE prion incubation, BMDC from C57BL/6 or PrP0 mice have an infectious titer of LD50 of 104.7 and 103.9 infectious units per 4 × 106 injected cells, respectively. BMDC were 10-fold less infectious after 6 days of culture than after 1 day.
To further investigate whether DC have the ability to propagate prion infectivity to the central nervous system, we tested, in RAG γc0 immunodeficient mice, the infectivity of BMDC exposed in vitro to BSE prions. These mice are resistant to murine BSE inoculum (brain homogenate) by the intraperitoneal route, although they are susceptible to the same inoculum after intracerebral inoculation (Table (Table2).2). While infected immunocompetent mice accumulated PrPbse in spleen and brain, RAG γc0 mice intracerebrally infected with BSE brain homogenate accumulated PrPbse in brain but not in spleen (data not shown). These findings confirm that RAG γc0 immunodeficient mice lack splenic target cells supporting BSE prion multiplication.
Four million live or dead BMDC from normal or PrP0 mice were incubated for 1 or 6 days with BSE prions and then intraperitoneally injected into RAG γc0 mice. At 6 months after injection, FACS analysis of blood showed that these mice still lacked B220 cells (data not shown). Seven hundred days after injection, no animal had developed BSE (Table (Table3),3), suggesting that BMDC did not directly release infection to nerve terminals.
The presented study shows that BMDC, BMpDC, and BMM efficiently capture and remove PrPbse in vitro. In contrast, no uptake of PrPbse was observed when splenic B and T lymphocytes were incubated with a BSE inoculum. Transferred BSE-infected BMDC did not spread BSE in RAG γc0 mice, suggesting that these cells do not have the ability to directly release infectivity to nerve terminals.
We showed here that myeloid, lymphoid, and plasmacytoid DC purified from spleen express normal PrP at levels conceivably high enough to support replication of PrPbse in vivo. We also found that cultured BMpDC and BMDC, which were mainly myeloid (CD11b+) DC, express large amounts of normal PrP on their surfaces. These data confirm and extend an earlier study showing PrP expression on myeloid DC in humans (8). DC are probably involved in PrPbse uptake in vivo, as we detected PrPbse in spleen DC shortly after intraperitoneal inoculation with BSE strain. Our findings are consistent with those of Huang et al., who also reported the presence of PrPsc in DC shortly after oral infection of rats with ME7 scrapie agent (16). Our data also fit with the demonstration of the infectivity of spleen DC from preclinical scrapie-infected mice (4). DC might be chemoattracted by PrPbse, as a neurotoxic PrP fragment, which can form insoluble amyloid fibrils similar to SAF, is chemotactic in vitro for immature DC (18, 19).
We demonstrated that cultured BMDC and BMpDC efficiently captured and cleared PrPbse within 2 to 3 days of incubation with a BSE inoculum. Our findings with the BSE strain confirm and extend earlier studies with scrapie strains, showing a PrPsc clearance in vitro by rodent BMDC or Langerhans cells (16, 26, 27). As macrophages and DC share several properties, including sample of antigens and motility, we compared the processing of PrPbse in BSE-infected cultured BMM and BMDC. We found that BMM rapidly captured and cleared PrPbse from the BSE inoculum in vitro, with a time course similar to that of BMDC. Our results are in agreement with two reports on in vitro and in vivo clearance of the scrapie agent by macrophages (9). The ability of BMDC and macrophages to capture and degrade PrPbse seems to be specific to these antigen-presenting cells, as we showed here that splenic T and B cells were not able to internalize and breakdown PrPbse. Moreover, we found that LPS-activated BMDC or BMM showed an accelerated breakdown of PrPbse, suggesting that the process of capture and clearance of PrPbse can be increased following DC maturation or macrophage activation. Our results differ from those of Mohan et al., who showed an inhibition of PrPsc clearance when skin-derived DC were activated by LPS (28). This difference could be due to the prion strain or the subset of DC. The kinetics of PrPbse degradation by C57BL/6 and PrP0 BMDC were very similar, suggesting that the normal PrP expressed by BMDC did not play a role in the capture of PrPbse.
We observed, by in vivo titration, that BMDC degrade BSE prions following in vitro exposure. Our data are consistent with those of Mohan et al., who reported a clearance of the scrapie agent by skin-derived DC following in vitro incubation (28). We found that although PrPbse was not detected within the cell culture at 6 days following incubation with BSE prions, infectivity still exists, confirming that the bioassay presents a very superior sensitivity level of infectivity. Interestingly, we observed only a 10-fold diminution of the infectious titer between days 1 and 6 following incubation with BSE inoculum, suggesting that the infectious unit in BMDC is not removed at the same rate as PrPbse is removed from these cells. PrPbse could probably be cleaved in small infectious peptides, which cannot be detected easily by Western blotting. The minimum PrP peptide size required to propagate infectivity is not yet known. As the infectious titer of C57BL/6 BMDC is 10-fold higher than that of PrP0 BMDC, these data do not exclude the possibility of some initial amplification of BSE prions in C57BL/6 BMDC versus PrP0 BMDC.
Immunodeficient RAG γc0 mice intraperitoneally challenged with BSE-infected BMDC did not develop BSE. These data suggest that BMDC do not release infection to nerve terminals and thus that these cells do not play a role in direct neuroinvasion. These results are in agreement with several earlier observations. First, SCID and RAG immunodeficient mice which possess DC are resistant to peripheral prion infection, indicating that DC are probably not target cells for prions. Second, a treatment depleting DC did not affect scrapie infectivity (31). Third, the neuroinvasion by scrapie following inoculation via the skin was not impaired in CD40 ligand-deficient mice with blocked Langerhans cells, suggesting that this subset of DC is not involved in scrapie transport to draining lymphoid tissue (27). Altogether, these findings do not support the hypothesis that BMDC play a role in direct prion neuroinvasion. Our results differ from an experiment of Aucouturier et al., where the infectivity of spleen DC could be transferred, by a peripheral route, to RAG-deficient mice (4). This discrepancy might be due to the presence of FDC copurified with spleen DC or to the level of infectivity isolated from the in vivo versus the in vitro model. On the other hand, our results do not definitively exclude a role of BMDC in indirect neuroinvasion by prions. Indeed, one can hypothesize that neuroinvasion could occur if the BSE-infected cultured BMDC were allowed to interact with FDC or other immune cells of which RAG γc0 mice are devoid. Our in vitro data clearly show that BMDC capture PrPbse and internalize it. One can suppose that, like for other antigens, BMDC cleave PrPbse into peptides and express it in the cell surface for interaction/activation of immune cells such as T lymphocytes. If these cells are absent, as in the case of RAG γc0 mice or in BMDC cultures, the fate of those peptides is probably further reinternalization, digestion, and thus clearance of prion infectivity. If immune cells and FDC are present, one cannot exclude the possibility that BSE-infected BMDC might transfer infectivity to these cells and then to the nervous system.
We thank Monique Dubois-Dalcq, Michel Goldberg, and Pierre-Marie Lledo for expert advice and critical reading of the manuscript, Karim Sebastien and Eddie Maranghi for animal care, Marie-Jésus Rojas for expert technical advice regarding cell cultures, and Jacques Christophe Cossec for the initial Western blot studies.
This work was supported by Bio-Rad, the Pasteur Institute, Paris (PTR 105), France, and GIS 2004 infections à prions. C.R.-B. was supported by a postdoctoral PTR fellowship from the Pasteur Institute.