In prion diseases, such as variant Creutzfeldt-Jakob disease normal cellular prion protein (PrPC), a largely α-helical structure is converted to an abnormal conformational isoform (PrPSc) that shows an increase in β-sheet content. Similarly, the recombinant form of PrPC (rα-PrP) can be converted to a conformation dominated by β-sheet (rβ-PrP) by reduction and mild acidification in vitro, a process that may mimic in vivo conversion following PrPC internalization during recycling. Despite PrPSc accumulation and prion propagation in the lymphoreticular system before detectable neuroinvasion, no Ab response to PrP has been detected, probably due to immune tolerance. To investigate how the immune system may respond to α- and β-PrP, we immunized Prnp0/0 mice that are not tolerant of PrP with rα-PrP and rβ-PrP. In this study, we show that although T cells stimulated by these differently folded conformers PrP recognize similar immunodominant epitopes (residues 111–130 and 191–210) the cytokine profile in response to rα- and rβ-PrP was different. Challenge with rα-PrP elicited a strong response of IL-5 and IL-10, whereas rβ-PrP led to an early increased production of IFN-γ. In addition, immunization with rα-PrP led to production of predominantly IgG1 isotype Ab in the sera, whereas after immunization with rβ-PrP, IgG2b was significantly produced. Thus, both humoral and cellular responses to these differently folded isoforms of the same protein are different, indicating a possible involvement of Th1 and Th2 pathway activation. These differences may be exploitable diagnostically and therapeutically for prion diseases, such as variant Creutzfeldt-Jakob disease.
The prion diseases or transmissible spongiform encephalopathies are a group of fatal, transmissible, neurodegenerative conditions, which include Creutzfeldt-Jakob disease, Gerstmann-Sträussler-Scheinker disease, and kuru in humans, and scrapie and bovine spongiform encephalopathy in animals. The transmissibility of human prion disease has become an intense public health issue because of evidence that bovine spongiform encephalopathy-like prions have infected humans (1, 2, 3). A common feature of all prion diseases is the posttranslational conversion of the normal cellular prion protein (PrPC),3 a host-encoded GPI-anchored sialoglycoprotein (4), α-helical in structure and expressed on most cells including B and T cells (5), into an abnormal isoform termed PrPSc (6), structurally dominated by β-sheet.
Neither amino acid sequencing nor systematic studies of known covalent posttranslational modifications have shown any consistent differences between PrPC and PrPSc (7). The abnormal isoform, PrPSc, is isolated from tissue as a highly aggregated and detergent insoluble polymer, which has markedly increased resistance to proteolysis. PrPC can be completely digested by the serine protease, proteinase K (PK), while in marked contrast, PK removes between 90 and 100 amino acids from the N terminus of abnormal isoform PrPSc, depending upon prion strain type, leaving a large protease-resistant C-terminal fragment (8). At low pH, following reduction of the native disulphide bond, the folded C-terminal domain of human prion protein is shown to exist as a soluble material with β-sheet structure (9). It is possible that the in vitro conversion of α-PrP to β-PrP, caused by reduction and mild acidification (10), might occur in vivo, following PrPC internalization within the cell during recycling (11), and this could underlie prion propagation (9).
Accumulation of PrPSc has been found in the lymphoreticular system in human variant Creutzfeldt-Jakob disease (12, 13). In scrapie-affected mice, prions are associated with splenic B and T cells, but not (14) or very little (15, 16) with circulating PBMCs. In mice, infectious prions are detectable in the spleen and lymph nodes long before they can be detected in the brain (17) (A. Khalili-Shirazi and S. H. Hawke, unpublished observation), and PrPSc is accumulated in splenic follicular dendritic cells (FDC) (18, 19, 20). An intact immune system is essential for prion propagation. FDCs are required for neuroinvasion and the development of clinical scrapie, and in turn, B cells are required for maturation of FDCs and formation of germinal centers (21). In addition, PrPSc has also been reported to be associated with macrophages and dendritic cells (22, 23, 24, 25, 26).
Despite the involvement of the immune system in PrPSc replication and neuroinvasion, no Ab response to PrP has been detected in prion disease. The lack of detectable immunity in prion infection is thought to be largely due to immune tolerance because of the widespread expression of PrPC in the hemopoietic system (27, 28, 29). Furthermore, attempts to produce PrPSc-specific Abs by immunization of Prnp0/0 mice (that are not tolerant to PrP) with PrPSc have been uniformly unsuccessful, although it has recently been reported that Abs raised to the repeat motif of PrP, tyrosine-tyrosine-arginine, can specifically recognize the pathological isoform of the PrP (30). We had previously immunized Prnp0/0 mice with highly purified human recombinant (r)α- and rβ-PrP (31) to obtain a panel of mAbs that strongly immunoprecipitate native PrPC and PrPSc.4 In this study, we have characterized the T cell-mediated immune responses to rα- and rβ-PrP in Prnp0/0 mice and show that the different conformations of the same protein are able to induce a qualitatively different type of response. Such information may provide insights into both the biology of prion diseases and the potential use of PrP conformers in therapeutic vaccination strategies.
Materials and Methods
We used Prnp0/0 mice (27) backcrossed onto an FVB/N mouse background to study the difference between the immune responses of the two different conformers of rα-PrP and rβ-PrP. These mice were housed under specific pathogen-free conditions. Animal care conformed to national and institutional guidelines.
Ag-rα-PrP, -rβ-PrP, and -PrP peptides
Human and mouse rα-PrP91–231 and rβ-PrP91–231 used as Ag were synthesized, as already described (9, 31), rα-PrP91–231 being a soluble α-helical monomer and PK-sensitive, whereas rβ-PrP91–231 is a soluble monomer, rich in β-sheet, and relatively PK-resistant. We used 91–231 rPrP, rather than full-length protein because it more readily forms a stable β-form. Also, native PrPC is completely digested by PK, while depending upon prion strain type, between 90 and 100 amino acids from the N terminus of abnormal isoform PrPSc are removed leaving a protease-resistant core (8). Such protease-resistant fractions retain infectivity and transgenic mice expressing N-terminally truncated PrP remain able to support prion propagation (32), arguing that the 91–231 construct is the most appropriate study object.
For epitope mapping we used 20-mer peptides of human PrP sequence 91–231, with 10 residues overlap and 15-mer peptides with 13 residues overlap. Peptides were made by solid phase stepwise synthesis using the Fmoc N-terminal protection strategy, using the Applied Biosystems 433 model automated synthesizer (Advanced Biotechnology Centre, Imperial College London). After synthesis, peptides were cleaved from the solid phase and fully deprotected by standard methods. Purification was performed by reverse phase HPLC. Products were analyzed by reverse phase HPLC and MALDI mass spectrometry.
For each arm of the study for T cell experiments, three 4- to 6-wk-old Prnp0/0 mice (27) backcrossed onto on FVB/N background were used. These mice have been found to respond well to PrP Ag. The mice were immunized s.c./intradermally at the base of the tail, with 100 μg of human rα- and rβ-PrP91–231 (α- and β-PrP) in CFA. Animals immunized with PBS in CFA and OVA in CFA were used as negative controls. Inguinal lymph nodes from the immunized mice were excised after 10 days and pooled. From each pool, a single cell suspension was made in RPMI 1640 (Life Technologies, BD Biosciences), supplemented with penicillin (100 IU/ml), streptomycin (100 μg/ml), glutamine (2 mM), 2-ME 0.5 mM, and FCS (10%) to analyze the T cell response in vitro.
For sera Ab production, 6-wk-old Prnp0/0 mice were immunized s.c. at the base of the tail with 100 μg of either mouse rα- or rβ-PrP91–231 in CFA, and 3 wk later, boosted s.c. at the base of the tail with 100 μg of rα- or rβ-PrP91–231 in IFA. The serum from each mouse was tested at day 28, at 1/100 dilution, for responses to rα- or rβ-PrP91–231 by direct ELISA, coating the plates with rPrP91–231 at 10 μg/ml in carbonate bicarbonate buffer (pH 8.6). The Ab IgG isotypes were also determined, using a mouse Ig isotyping ELISA kit (BD Biosciences, BD Pharmingen).
For the production of mAbs, Prnp0/0 mice were immunized s.c. at the base of the tail with 100 μg of Ag (human rα- or rβ-PrP91–231) in CFA and boosted on days 21 and 42 with 100 μg of the similar Ag in IFA, then finally boosted i.p. on day 50 with 50 μg of the Ag in PBS. Three days after final boosting, single cell suspensions of mouse splenocytes were prepared and fused with the NS0 myeloma cell line, selecting hybridomas using conventional methodology. Detailed methods for screening of these mAbs have been described elsewhere.4
Proliferation and cytokine assays
For proliferation assays, T cells derived from the lymph nodes of mice immunized with either rα- or rβ-PrP Ag or control mice were used. The study with rα- and rβ-PrP as Ag were repeated four times and with PBS twice. The T cells were derived from a pool of lymph nodes from three mice and tested in triplicate, presenting the results as stimulation index (SI) ± SE. The cells proliferated to ∼9000 cpm and a SI of 2.5–4 was considered significant. For proliferation assays, cells were plated in round-bottom 96-well plates (Falcon), at 4 × 105 cells per 0.2-ml well, and for cytokine assays in 24-well plates, at 1–2 × 106 cells per 2-ml well. Human rα- and rβ-PrP were sterile filtered through a 0.2-μm filter (Nalgene), titrated in culture for T cell proliferation, and 50 μg/ml was found to be the optimal concentration. Human PrP peptides were also sterile filtered and used at 10 μg/ml. Where appropriate, OVA at 50 μg/ml was used as negative control Ag. Cells were maintained in a standard 72-h proliferation assay and pulsed with [3H]thymidine (0.5 μCi/well) (Amersham Biosciences) in the last 18 h prior to harvesting (TomTec cell harvester) and determining cpm, using a beta counter (Microbeta counter; PerkinElmer).
The production of cytokines from cells in lymph nodes were measured in vitro after 24, 48, and 72 h. IL-2, IL-4, IL-5, IL-10, TGF-β, and IFN-γ were analyzed by sandwich ELISA, using standard protocols, including a standard curve for each plate (BD Biosciences, BD Pharmingen).
Characterization of anti-PrP mAbs
As previously described, mAbs ICSM4, -10, and -18 were obtained upon immunization with rα-PrP91–231 while ICSM35 and -37 with rβ-PrP91–231.4 ICSM18 and -35 were tested for staining native PrP, in a Rocky Mountain Laboratory isolate of Chandler mouse scrapie prion (RML) (33) -infected neuroblastoma cell line (N2a) (33, 34). The cells were grown in culture on multichamber slides (Sigma-Aldrich) until semiconfluent. To label cell surface PrP, they were incubated at 4°C with anti-PrP mAbs, before fixation with paraformaldehyde. For detection, FITC-conjugated anti-mouse secondary Ab was used.
Direct ELISA method was used to detect binding of ICSM18 and -35 to rα- and rβ-PrP91–231 at 20–10 μg/ml, using anti-mouse IgG (Fab-specific) HRP (Sigma-Aldrich) as secondary Ab and o-phenylenediamine (Sigma-Aldrich) as chromogen. Direct ELISA for detection of native PrP proved difficult, despite using different brain homogenate concentrations, plate manufactures, coating-buffers (carbonate bicarbonate (pH 9.5), Tris (pH 8.5), and PBS (pH 7.4)), blocking solutions (normal goat serum, normal mouse serum, FCS, BSA, and soya milk), and temperatures (4°, 23°, and 37°C), so an indirect sandwich ELISA technique was devised. ICSM18 and ICSM35 were individually used to coat the ELISA plate and after washing, normal or scrapie-affected brain homogenate was added as Ag, and PrP was detected with biotinylated (B) mAb ICSM18B or ICSM35B, using a different mAbs for capture and detection. Finally streptavidin-HRP was added and o-phenylenediamine was used as chromogen. The brain homogenates were tested at 0.5–0.1%, the mAb used for PrP capture coating the ELISA plate was tested at IgG concentrations of 2–10 μg/ml, and the biotinylated mAb used for PrP detection was tested at IgG concentrations of 1 and 0.5 μg/ml. Both ICSM18 and 35 worked well for capturing and for detecting native PrP.
The isotypes of these mAbs were determined using a panel of anti-mouse Ig subclass-specific mAbs (Pierce) and their immunodominant epitopes were defined by direct ELISA, using 20-mer synthetic peptides (10–20 μg/ml), spanning the PrP sequence 91–231 overlapping by 10 residues. These epitopes were further refined by direct peptide ELISA, using 15-mer peptides overlapping by 13 residues. The accuracy of the immunodominant epitopes found was reconfirmed by competition and inhibition ELISA, where the peptides were used to inhibit binding to recombinant rα- and rβ-PrP. For these studies, ELISA plates were coated with rα- and rβ-PrP and the mAbs were tested either by preincubating them with the peptides to inhibit binding to rα- and rβ-PrP, or the peptides were added simultaneously as the mAb to ELISA plates, to compete for the binding of the mAbs to rα- and rβ-PrP. Similar results were obtained, and the latter method was mainly conducted. ELISA plates were coated with rα- or rβ-PrP (20–10 μg/ml) in carbonate/bicarbonate buffer (pH 9.5), to which each mAb IgG (1 μg/ml) and peptide (10–50 μg/ml) were added. The ELISA was then conducted by conventional methods as described above.
Data were analyzed for statistical significance using ANOVA and two-tailed t test.
Both rα- and rβ-PrP are recognized by T cells
Immunization with either rα- or rβ-PrP evoked clear proliferative T cell responses to both rα- or rβ-PrP Ags (Fig. 1), unlike T cells from immunizations with PBS-CFA (Fig. 1) or OVA-CFA (data not shown). There was no significant difference between the in vitro recognition of rα- and β-PrP by T cells derived from either rα- or rβ-PrP immunization (p = 0.316, p = 0.154, two-tailed t test). The viability and Ag-recognition pathway of all the T cells was tested by their response to stimulation with anti-CD3 Ab, where a mean SI of 7–12 was observed (data not shown).
T cell epitope recognition of rα- and rβ-PrP is similar
T cells from rα- and rβ-PrP immunizations showed no significant difference in their recognition of PrP peptides (Fig. 2, A and B). T cells from lymph nodes of both rα-PrP (Fig. 2,A) and rβ-PrP (Fig. 2,B) immunized mice recognized N-terminal residues within the sequence 111–130 and C-terminal residues within 191–210 (p = 0.005, ANOVA). In keeping with the previous result, T cells from PBS (or OVA)-immunized mice did not proliferate in response to PrP peptides (Fig. 2 C and data not shown). PrP residues 111–130 are thought to be within the structured part of PrP while residues 191–220 are just C-terminal to the glycosylation sites of PrP.
The intracellular proteolytic degradation of Ag by APC might generate different peptide ligands for interaction with MHC class II molecules. Because for both rα- and rβ-PrP immunizations we used Prnp0/0 mice of FVB/N background with identical MHC class II molecule, it is not surprising that we found a similar linear sequence of peptides being recognized by T cells from both groups.
The cytokines produced by rα- and rβ-PrP-derived T cells are different
The overall proliferation of T cells from Prnp0/0 mice, and in particular proliferation to PrP peptides was weak. Therefore, we checked for differences in cytokine profiles of lymph node cells from rα- and rβ-PrP immunized mice. Measuring cytokines, in cultures of lymph nodes cells challenged in vitro with rPrP, was most suitable after 24 h and 48 h. Although there was some overlap of the cytokines produced, T cells from lymph nodes of rβ-PrP-immunized mice were skewed more toward Th1 (producing more IFN-γ and sooner, and less IL-5 and IL-10), whereas those from rα-PrP immunization were skewed more toward Th2 (producing more IL-5 and IL-10, and less or later IFN-γ) (Fig. 3).
The cytokine levels were clearly affected by the rPrP isoform used for in vitro challenge, as well as for immunization. After 24-h stimulation with rα-PrP in vitro, T cells from rα-PrP-immunized mice produced only 340 pg/ml IFN-γ, whereas T cells from rβ-PrP immunization produced >2000 pg/ml IFN-γ (Fig. 3,A). After 48-h in vitro stimulation of cells, T cells from both rα- and rβ-PrP immunizations produced a large amount of IFN-γ (>2000 pg/ml). A small amount of IFN-γ was also produced after 48 h, by unstimulated T cells from rβ-PrP immunization, and from PBS immunization stimulated in vitro with rβ-PrP (Fig. 3 A).
Th2-induced cytokines like IL-5 and IL-10 were also investigated. IL-5 production was substantially higher after in vitro stimulation with rα-PrP compared with rβ-PrP, by T cells from both rα-PrP (p = 0.001, t test) and rβ-PrP (p = 0.0108, t test) immunizations. T cells from rα-PrP-immunized mice produced 227 pg/ml IL-5 in response to rα-PrP but only 90 pg/ml in response to rβ-PrP, and T cells from rβ-PrP-immunized mice produced 122 pg/ml in response to rα-PrP and only 16 pg/ml in response to rβ-PrP (Fig. 3,B). The T cells from rα-PrP-immunized mice also produced significantly more IL-5 in response to in vitro stimulation with rα-PrP, than those from rβ-PrP immunization (p = 0.0147, t test). No IL-5 was produced by T cells from PBS-immunized mice, regardless of using rα- or rβ-PrP for the in vitro challenges (Fig. 3,B). The IL-10 responses showed that it was produced, slightly more by T cells from rα-PrP than rβ-PrP-immunized mice (Fig. 3 C).
Not surprisingly, in view of the results shown in Figs. 1 and 2, there was no statistical difference in the production of IL-2 in response to either rα- or rβ-PrP from both control groups (Fig. 3,D), and small amounts of IL-2 was also produced by T cells from PBS-immunized mice, after in vitro challenge with rα- or rβ-PrP (Fig. 3 D). Production of IL-4 and TGF-β are not shown, as they could not be convincingly measured.
PrP conformation influences the Ab Ig isotypes produced in the sera
Total IgG Ab levels were similar in mice immunized with rα- and rβ-PrP and their sera Abs were positive to rα- and rβ-PrP Ags by ELISA, while no Ab to rα- and rβ-PrP was present in the serum of the unimmunized mouse (Fig. 4,A). However, a significant difference was observed in the Ig isotypes produced in the sera of mice immunized with rα-PrP, as compared with those immunized with rβ-PrP. Immunization with rα-PrP evoked predominantly Th2 isotype Abs of IgG1 isotype, whereas rβ-PrP evoked mainly Th1 IgG2b in the serum (Fig. 4 B).
Significantly higher levels of IgG1 Abs were produced in the sera of rα-PrP-immunized mice, compared with those produced in sera of mice immunized with rβ-PrP (p < 0.005, two-tailed t test). In contrast, IgG2b Ab was produced at higher levels in the sera of rβ-PrP-immunized mice, when compared with rα-PrP-immunized mice (p < 0.005, t test). There was no significant difference between the levels of the other Ig isotypes in the sera of these two groups of mice. Because the mice were bred under specific pathogen-free conditions, it is not surprising that no Ab IgG isotypes and only very low levels of IgM were detected in the sera of the unimmunized mice (Fig. 4,B). In addition, different anti-PrP Ab IgG isotypes were found in the sera of mice immunized with rα-PrP and rβ-PrP and tested by ELISA against either rα-PrP (data not shown) or rβ-PrP (Fig. 4,C). IgG1 anti-PrP Ab was produced in sera of mice immunized with rα-PrP and IgG2b Ab produced after rβ-PrP immunization (Fig. 4 C). These findings confirm that despite the weak T cell proliferation responses, the α- and β-isoforms of PrP are recognized differently by the immune system.
PrP conformation affects the IgG isotype and the specificity of the mAbs produced
A significant difference was also observed in the specificity and IgG isotype of the mAbs raised against α- and β-isoforms of PrP. The results obtained for the mAbs produced to rα- and rβ-PrP further demonstrate that PrP in the different conformations (α or β) give rise to a different type of cellular response, skewed more toward Th1 for rβ-PrP and Th2 for rα-PrP. This reconfirms the data already obtained for T cells and the sera Ab Ig isotypes. ICSM4, -10, -18, -35, and -37 mAbs, raised to rα- and β-PrP, had different specificities and IgG isotypes. Those produced against rα-PrP (ICSM4, -10, and -18) were of IgG1 isotype, whereas rβ-PrP-derived mAbs (ICSM35 and -37) were IgG2b and IgG2a.4 Furthermore, IgG isotypes were also different in the 40 mAbs we have produced to rα-PrP and 41 mAbs to rβ-PrP. The isotypes of the mAbs produced from rα-PrP immunizations were 92.5% IgG1 and 7.5% IgG2 (2.5% IgG2a and 5% IgG2b), whereas the isotypes of those produced to rβ-PrP immunizations were 53.6% IgG1, 41.5% IgG2 (12.2% IgG2a, 29.3% IgG2b), and 4.9% IgM (manuscript in preparation).
Both ICSM18 and ICSM35 (raised against rα- and rβ-PrP respectively), were equally suitable for capturing or detecting native PrP from normal and RML-infected scrapie-affected mouse brain homogenates by sandwich ELISA (Fig. 5,A). As previously described ICSM4, -10, and -18 (raised to rα-PrP) and ICSM35 and 37 (raised to rβ-PrP) were all positive to denatured PrPC and PK-treated PrPSc in a Western blot, but they recognized different PrP glycoforms.4 The Western blots of ICSM18 and ICSM35 showed that they both recognize all the PrP glycoforms (Fig. 5,B). In contrast, ICSM4 only recognized unglycosylated PrPC and PrPSc, while ICSM18 recognized di-, mono-, and unglycosylated PrPC and PrPSc, by Western blot using aliquots of the same normal and scrapie-affected mouse brain homogenates (Fig. 5 B). ICSM4 and -10 also differentially immunoprecipitated PrP glycoforms from PrPC, while anti-β-PrP mAbs ICSM35 and -37 did not.4 ICSM18 and ICSM35 were chosen, as strongly binding representatives of PrP-specific Th2 response for α-PrP and Th1 response for β-PrP.
We have labeled PrP on RML-infected N2a cells with ICSM18 and ICSM35 by immunofluorescence (Fig. 6). Although both ICSM18 and ICSM35 specifically labeled PrP on these cells, staining of PrP with ICSM18 proved to be stronger and more uniform (Fig. 6,A) than ICSM35 (Fig. 6,B). ICSM18 has a high affinity for native PrPC and a weak affinity for native PrPSc, while ICSM35 has a higher affinity for native PrPSc than for native PrPC by immunoprecipitation.4 This may explain the strong immunofluorescence detected by ICSM18 and the weaker immunostaining detected by ICSM35. The weak staining of ICSM35 may also indicate that native PrPSc is mainly intracellular. No nonspecific binding of PrP was observed using negative control mAb (Fig. 6,C), or no primary Ab, second layer alone (Fig. 6 D).
Detailed epitope mapping, showed striking differences for the rα- and rβ-PrP-generated mAbs. Approximately similar epitopes for rβ-PrP-derived mAbs ICSM35 and -37 were found that were different from the rα-PrP-derived mAb ICSM18 (Fig. 7). ICSM18 was specific for the linear sequence 143–153 (Fig. 7,A), ICSM35 and ICSM37 for sequences 93–105 and 97–105, respectively (Fig. 7,B). ICSM4 and -10 were negative in the peptide ELISA and could not be epitope mapped. The immunodominant epitopes of ICSM18 (Fig. 7,C) and ICSM35 (Fig. 7 D) were reconfirmed by peptide competition ELISA, using the peptides as competitors for binding of the mAbs to rα- and rβ-PrP Ags.
These data show that rα- and rβ-PrP generate diverse immune responses, cellular and humoral, and mAbs could be raised with high affinity for native PrPC (like ICSM18) or for native PrPSc (like ICSM35).
In this study, we have demonstrated that immunization of mice with either α- or β-PrP evokes a different cytokine milieu and Ab isotype profile. Usually CD4+ T cells recognize linear processed peptides presented in the groove of MHC class II molecules, whereas B cells recognize both conformational and/or linear epitopes of an Ag. When we immunized PrP-deficient mice (of FVB/N MHC H-2q haplotype) with either rα- or rβ-PrP, comparable T cell proliferation against similar immunodominant epitopes was observed. Although α- and β-PrP do not differ in amino acid sequence, they elicited qualitatively different cellular and humoral response. Although T cells from rα- and rβ-PrP immunizations recognized both rα- and rβ-PrP and peptide residues 111–130 and 191–210; by contrast the mAbs reflecting the activity of B cells could partially discriminate between PrPC and PrPSc, and their immunodominant peptides were different: 93–105 for rβ-PrP immunizations (ICSM35 and -37) and 143–153 for rα-PrP (ICSM18).
We have observed that the T cell proliferation, particularly against PrP peptides, was weak. One possible explanation for this could be the absence of PrP in the cells of Prnp0/0 mice. Lack of PrPC on T lymphocytes could lead to generally low activation levels, as we found that even the response to soluble anti-CD3 Ab was very modest. Indeed, PrPC may have an important role in T cell activation. PrPC is significantly up-regulated on mouse T cells upon challenge with concanavalin A (Con-A), but in Prnp0/0 mice, T cell response to Con-A is significantly reduced to 50–80%, suggesting that PrPC is a lymphocyte surface molecule that participates in T cell activation (35). In human T lymphocytes, PrPC expression is also increased upon challenge with Con-A and reduced by the addition of anti-PrP polyclonal Ab (36). PrPC expression on normal human T lymphocytes is also increased after challenge with either Con-A or anti-CD3 Ab and inhibited by addition of anti-PrP mAbs, and in addition, CD45RO+ memory T cells consistently expressed more PrPC than CD45RA+ naive T lymphocytes (37). A recent report suggests PrPC to be a component of the multimolecular signaling complex, within microdomains involved in T cell activation (38). The absence of PrPC on T lymphocytes of Prnp0/0 mice may therefore be responsible for the weak proliferative response we have observed in this study, but apparently it does not affect their cytokine production. The proliferation response was particularly weak after in vitro challenge with the peptides, but in a preliminary study we have found that splenocyte T cells from rα-PrP-immunized mice produce IL-5 in response to in vitro challenge with peptide sequence 191–210, and T cells from both rα- and rβ-PrP immunizations produce IFN-γ in response to this peptide (S. H. Hawke and Z. Sattar, unpublished data). Sera Abs were also positive to this epitope after immunization of Prnp0/0 mice with rPrP (A. Khalili-Shirazi and L. Summers, unpublished observation).
The processing of different proteins by different APCs, such as dendritic cells and macrophages, leads to activation of different signaling pathways leading to the production of different cytokines (39). T cell activation leads to a CD4+ or CD8+ T cell response, CD4+ initiating either a Th1 or Th2 immunity, each triggering production of a series of different cytokines and IgG isotypes (40, 41, 42). Th1 type response leads to activation of macrophages and is characterized by production of IFN-γ, whereas Th2 leads to activation of different APCs, and production of IL-4, IL-5, and IL-10 (43, 44). This dichotomy might be interpreted as the result of different T cell activation. Indeed, it has been reported that Th2 cells provide B cell help and stimulate the production of noncomplement fixing IgG1 Abs, while Th1 cells induce B cells to switch to IgG2 and IgG3 (45).
In this study, we have shown that despite some overlap, the immune response to rα-PrP was skewed more toward Th2, inducing T cells to produce increased amounts of IL-5 and IL-10, and B cells to produce polyclonal IgG1 Abs in the sera, and IgG1 mAbs such as ICSM4, -10, and -18. In contrast, the immune response to rβ-PrP was skewed more toward Th1, with a high and early production of IFN-γ, less IL-5 and IL-10 cytokines, and production of serum IgG2b isotype Abs and IgG2b and IgG2a mAbs, like ICSM35 and -37. These data indicate that different APCs may be involved in processing of rα- and rβ-PrP. When rα-PrP is used as the immunogen, its greater sensitivity to proteolytic cleavage might mean that it is more easily processed for presentation to B cells and may involve different APCs from those that process rβ-PrP. In comparison with rα-PrP, rβ-PrP is rich in β-sheet and partially PK-resistant (9). It is possible that rα-PrP is mainly processed by B cells and dendritic cells and much less by macrophages, whereas β-PrP is mainly processed by macrophages and less by B cells or dendritic cells. Immune tolerance to PrP in Prnp+/+ mice can reportedly be overcome by improved immunization techniques, using mouse rPrP (46) and dimeric mouse rPrP (47), leading to production of anti-PrP polyclonal Abs. In the course of prion inoculation, activated T cells have been detected in the brain and spleen of mice affected with scrapie (48). Macrophages appear to have an important role in neutralizing infectivity (23) and an association between dendritic cells and PrPSc has been reported (26, 49). Our findings may provide an insight into how PrPC and PrPSc might be differentially processed by APCs, such as dendritic cells and macrophages. The data presented in this study on the immune responses to rα- and rβ-PrP might reflect those to PrPC and PrPSc, if there was no tolerance to PrP or if it could be overcome by vaccination. A transient, but undetected, immune response may also occur early in the course of prion infection, when conversion of PrPC to PrPSc has just begun and before formation of large PrPSc deposits and onset of clinical signs. These findings may therefore be relevant to the understanding of the molecular mechanisms underlying prion diseases in humans and help in the design of an effective therapeutic immunization or vaccination approach.
J. Collinge and G. S. Jackson are consultants and shareholders in D-Gen Ltd., and J. Collinge is also a Director.
We thank Daljit Bhelt for technical help in making the rPrP, the Central Biomedical Services staff of Imperial College London at Charing Cross campus, and Ray Young for graphics.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was funded by the United Kingdom Medical Research Council.
Abbreviations used in this paper: PrPC, cellular prion protein; PrPSc, abnormal isoform prion protein; PK, proteinase K; FDC, follicular dendritic cells; r, recombinant; RML, Rocky Mountain Laboratory isolate of Chandler mouse scrapie prion; SI, stimulation index; B, biotinylated; Con-A, concanavalin A.
A. Khalili-Shirazi, L. Summers, J. Linehan, G. Mallinson, D. Anstee, S. Hawke, G. S. Jackson, and J. Collinge. PrP glycoforms are associated in a strain-specific ratio in native PrPSc. Submitted for publication.