| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
INSERM EMI 0209, Hôpital Saint-Antoine, Paris, France
Correspondence: Pierre Aucouturier, INSERM EMI 0209, Hôpital Saint-Antoine, 184 rue du Faubourg Saint-Antoine, 75571, Paris Cedex 12, France. E-mail: aucouturier{at}necker.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONAN OVERVIEW ON PRION...THE LYMPHORETICULAR SYSTEM AS...IMMUNOINTERVENTION TARGETED AT...PERSPECTIVESREFERENCES |
|---|
Key Words: TSE • lymphoinvasion • neuroinvasion • anti-PrP immunity
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONAN OVERVIEW ON PRION...THE LYMPHORETICULAR SYSTEM AS...IMMUNOINTERVENTION TARGETED AT...PERSPECTIVESREFERENCES |
|---|
Although the field of TSEs has been neglected by immunologists for some time, mainly because prions do not elicit immune responses, this is not the case any longer. The finding that prions invade the lymphoreticular system long before the CNS opens prospects regarding preclinical diagnosis and prevention of brain colonization. At the same time, the recognition that the prion protein (PrP) is an essential or according to the "protein only" theory, even the sole component of infectivity raises new opportunities of immunointervention targeted against this unusual protein, provided that self-tolerance can be broken. This review will essentially focus on those two aspects of TSE pathogenesis, namely on what is presently known regarding lymphoinvasion by prions and on the recent developments about anti-PrP immunointervention. However, before going into this analysis, we will briefly describe what the TSEs are and the animal models used for studying their pathogenesis.
|
AN OVERVIEW ON PRION DISEASES AND THEIR ANIMAL MODELS |
|---|
|
TOPABSTRACTINTRODUCTIONAN OVERVIEW ON PRION...THE LYMPHORETICULAR SYSTEM AS...IMMUNOINTERVENTION TARGETED AT...PERSPECTIVESREFERENCES |
|---|
-helices and random coils, poor in ß-pleated sheets, soluble in detergents, and sensitive to PK digestion. TSEs may be classified, according to their etiology, as infectious, familial, or sporadic. A number of TSEs are acquired by infectious transmission, as a result of infrequent procedures or life habits or in apparently "natural" contexts [6 ]. Sheep scrapie has been known since the 18th century, and its transmissibility, including to goats, was demonstrated 100 years ago. Still, the precise modes of contamination within flocks remain obscure. Epizootics of TSEs were also described in mink [transmissible mink encephalopathy (TME)], in deer and elk [chronic wasting disease (CWD)], and in cow [bovine spongiform encephalopathy (BSE)]. Humans are susceptible to certain infectious forms of TSE. The most famous is the "new variant" form of Creutzfeldt-Jakob disease (CJD), transmitted through consumption of meat contaminated with the BSE agent. Other forms include the epidemics of Kuru, passaged by ritual cannibalism in the Fore tribe of New Guinea, and the iatrogenic CJD, related to injection of human growth hormone or to certain surgical practices.
Pure hereditary forms have been demonstrated in humans, such as the Gertsmann-Straüssler-Scheinker syndrome (GSS), certain cases of CJD, and fatal familial insomnia. In all described families, the disorders have been related to point mutations of the PRNP gene that encodes PrP. There is no evidence that hereditary TSEs naturally disseminate to other individuals, although brain extracts from cadavers can transmit disease into experimental animals [7 , 8 ].
Sporadic CJD (i.e., without evidence of transmission or familial origin) occurs with an incidence of approximately one per million per year in various populations around the world. It is plausible that similar sporadic cases exist in the animal kingdom, but they have remained unrecognized yet. As already mentioned above for familial forms, an essential and most original peculiarity of all TSEs, irrespective of their etiology as opposed to other neurodegenerative disorders, is their transmissibility through inoculation of pathological tissue into healthy recipients. Importantly, the infectious principle copurifies systematically with PrPSc, the pathological conformer of PrP. It has therefore been hypothesized that prion and PrPSc were the same entity, thus explaining why the host immune system does not react against an agent that is actually of self-origin. This contention is not unanimously accepted, and some specialists suggest that PrPSc might only be the reactional signature of the infected host to a short noncoding, nucleotidic element—a "virino"—deeply embedded in PrPSc and thus copurifying with it. Nevertheless, everybody agrees that, based on the fact that Prnp knockout mice are totally refractory to scrapie [9 ], the transition from PrPC to PrPSc is a critical event in the pathogenesis of prion diseases.
Experimental animal models of TSEs have been developed by adapting isolates first from sheep and goat scrapie and more recently from TME, BSE, and human GSS or CJD to mice, Syrian hamsters, rats, or to certain strains of monkeys [10 11 12 13 14 ]. The species barrier is defined by the observation that passage to a different species always results in a significantly longer incubation period or even in the absence of clinical disease. In the latter case, infectious agent can often be detected in the recipients’ brains late in life, which has led to question the concept of absolute resistance. After initiation by intracerebral inoculation of a high dose of natural isolate, serial passages in mice and in hamsters have given rise to stabilized strains, most originating from sheep scrapie and some of which were cloned by limiting dilutions in laboratory animals. Once stabilized and cloned in a given species, the characteristics of a given prion strain are remarkably stable throughout subsequent passages, a feature that is difficult to conciliate with the idea that all the information necessary for perpetuating such characteristics is enciphered exclusively in a PrPSc molecule. Prion strains differ among themselves in many respects [15 ]. They may have different incubation periods in a given host, different clinical manifestations, and different histopathological localizations and expression in the CNS. They may have different sensitivity to PK digestion [16 ], different electrophoretic patterns after PK digestion, and as shown below, different physiopathogenic features. In short, each prion strain determines its own disease profile, hence the pitfall of extending the conclusions from one model to another and to generalize too rapidly to all TSEs including those occurring under spontaneous conditions. With this important caveat in mind, we can now examine the set of data pinpointing the complicity of the lymphoreticular system in TSE pathogenesis.
|
THE LYMPHORETICULAR SYSTEM AS A TROJAN’S HORSE |
|---|
|
TOPABSTRACTINTRODUCTIONAN OVERVIEW ON PRION...THE LYMPHORETICULAR SYSTEM AS...IMMUNOINTERVENTION TARGETED AT...PERSPECTIVESREFERENCES |
|---|
The involvement of peripheral lymphoid tissues in prion propagation is supported by three main lines of evidence: 1) early infectivity of lymphoid tissues and progressive increase of titres well before infectivity can be demonstrated in the CNS; 2) accumulation of PrPSc as evidenced by Western blot or immunohistochemistry in defined cell types and areas of secondary lymphoid organs; and c) effect of targeted immune impairments on host susceptibility to peripheral inoculation.
Numerous studies have addressed the evolution of infectivity titers in different organs in mouse and hamster scrapie after inoculation by different routes. In all experimental models as well as in most natural TSEs, with a few remarkable exceptions such as BSE [17 ], prions accumulate and multiply; i.e., the infectivity titres increase progressively in lymphoid tissues, early in the preclinical phase, and long before they are detectable in the CNS [18 19 20 21 22 ]. Prion inoculation via peritoneal, cutaneous, or venous routes typically leads to lymph node and spleen invasion first, followed by that of intestine mucosa and spinal cord, and then the brain. After oral contamination, early accumulation of prions is seen first in the gut-associated lymphoid tissue, mainly the ileal Peyer’s patches, in natural or experimental scrapie [23 24 25 26 27 28 ], in sheep and mice infected with BSE [17 , 24 , 29 ], or in deer CWD [23 ]. Mesenteric lymph nodes are quickly involved, followed progressively by most other secondary lymphoid organs, the sympathetic nervous fibers and myenteric ganglia [30 , 31 ]. Thus, accumulation of PrPSc in the peripheral nerves occurs shortly after infection of lymphoid tissues, suggesting that prions could progress toward the CNS by retrograde axonal transport [27 ]. This was recently supported by observations that chemical or immunological suppression of the sympathetic innervation of lymphoid organs significantly delayed the onset of mouse experimental scrapie [32 ]. Of note, direct intracerebral inoculation also results in early infectivity of spleen and lymph nodes, but this is probably related to the entry of part of the inoculum into the vascular compartment at the time of injection. In sharp contrast with what happens in the brain, the spreading of prions through the lymphoid tissues is not associated with major morphological or functional alterations. Infected animals develop normal humoral and cell-mediated responses against conventional antigens or mitogenic lectins [2 , 33 , 34 ].
Studies on immunologically impaired recipients have given strong support to the concept of lymphoinvasion being a key event in the pathogenesis of infectious TSEs. Historical experiments, performed with the tools that existed at that time, had already shown that splenectomy [35
, 36
], young age [37
], or corticosteroid treatment [38
] resulted in prolonged incubation periods after peripheral challenge with scrapie agents. Conversely, mitogenic stimulation of lymphoid tissues enhanced susceptibility to scrapie [39
]. More decisive arguments have been obtained since then, owing to the use of immunological mutants and of genetically engineered strains of mice. Severe combined immunodeficiency (SCID) mice, which have a profound defect in B and T lymphocyte differentiation, are in the majority resistant to peripheral but not to intracerebral inoculation of prions [12
, 40
]. The few mice that become sick after intraperitoneal (i.p.) injection display prolonged incubation periods; these escapees are actually found in the C57BL/6 strain in which the SCID phenotype is particularly "leaky" [41
]. Resistance of SCID mice to scrapie infection is invariably reverted by immune reconstitution with normal spleen cells [42
]. More recently, the susceptibility to the mouse-stabilized RML strain, originating from a goat isolate, has been analyzed by Klein et al. [41
], using as recipients knockout lines of mice with targeted defects on T and B lymphocytes (RAG-1 and RAG-2°/°), on B cells (µMT), or on selected T cell subsets [IA°/°, ß2M°/°, T cell receptor (TCR)-°/°]. The authors found that mice that were devoid of B and T lymphocytes or of B lymphocytes only were significantly resistant to disease inoculation by peripheral route, and their sensitivity to intracerebral injection of RML was normal. Conversely, mice deficient in any of the T cell subsets were as susceptible to the peripheral route of inoculation as wild-type controls. It is interesting that in a significant proportion of clinically resistant mice, prions nevertheless reached the brain, and much higher doses of inoculum were even able to provoke scrapie in SCID, µMT, and RAG°/° mice [41
, 43
, 44
].
The difference of behavior between B and T cell-deficient mice suggested that not all subsets of the lymphoreticular system were equally implicated in the propagation of prions to the CNS. The following studies were thus aimed at identifying the most important cell types involved, mainly by targeted elimination. We will review the main results of this analytical approach, which although simple in its principle, has generated a great deal of confusion, owing to the diversity of experimental models and lack of standardization.
B lymphocytes
Although PrPSc had long been found to accumulate in follicular dendritic cells (FDCs), for a while B lymphocytes were thought to be the main actors of lymphoinvasion on the basis of the resistance of µMT mice, which have a specific defect in early B cell differentiation, to peripheral inoculation with the RML scrapie strain [41
]. The results were confirmed with other strains of mouse-stabilized prions but rather surprisingly not with FU, a strain originally derived from an isolate of human GSS [8
]. Manuelidis and co-workers [45
] indeed reported that B cell-deprived mice peripherically challenged with the FU strain were almost as permissive as wild-type controls. This is one of the many examples showing that different strains of prions may privilege different routes of neuroinvasion.
A next step in the comprehension of the role played by B cells in neuroinvasion came from elegant experiments in which µMT mice were immunologically reconstituted with hematopoietic precursors from Prnp knockout donors. Such PrP chimeric mice had a completely restored sensitivity to peripheral challenge by RML scrapie, showing that expression of physiological PrPC on B cells was not required for neuroinvasion [43 ]. As the replication of the agent is critically dependent on the presence of endogenous PrPC [46 ], it was concluded that B cells were probably not directly involved in this function. Brown et al. [47 ] confirmed, in the ME7 scrapie strain mouse model, the indirect role played by B cells. Using mouse chimeras, which expressed PrPC on their stromal cell compartment only, they showed that B cells deprived of PrPC still allowed infection, and they proposed that the contribution of this subset was related to its ability to induce the functional maturation of FDCs, which differentiate in the vicinity of B cells and whose normal function is to provide sustained delivery of antigen. Reciprocal experiments performed with the RML strain confirmed that PrPC expression on B cells only did not support amplification of infectivity [48 ]. If B cells do not directly support prion replication, they can still carry infectivity, as suggested in some experimental reports [49 , 50 ]. B cells might passively acquire infectivity through tight contacts with infected FDCs. Overall, there is no evidence that allows excluding them as participants in the neuroinvasion process.
T lymphocytes
T lymphocytes have not been considered major actors of lymphoinvasion, as Nude mice appear to have incubation periods similar to those of wild-type controls [51
] and as no clear-cut resistance to prion invasion was found in association with selective defects on T cell subsets [41
, 43
]. Moreover, when PrP expression was specifically directed to T lymphocytes by introducing a Prnp transgene under the control of the Lck promoter in Prnp knockout mice, replication of RML prions could not be detected in the thymus and the spleen [52
]. Although these results indicate clearly that T cells are dispensable and are not a site of prion replication, they do not exclude a possible contribution in association with other cell types. We have found high-infectivity titers close to those of DCs and about 12 times higher than those of B cells in the spleen T cell fraction of mice inoculated 10 weeks earlier with 139A scrapie strain (P. Aucouturier and Richard I. Carp, unpublished results). CD4+ T lymphocytes, which express relatively high levels of cell surface PrPC, particularly after activation (Pascaline Fontes and C. Carnaud, unpublished results), could capture infectious PrPSc and transport it to sites of neuroinvasion. How prions are transferred from peripheral sites of accumulation to the CNS is still unclear, and circulating cells such as T lymphocytes might be a missing link.
FDCs
FDCs have long been demonstrated as major sites of prion accumulation in the germinal centers of spleen, lymph nodes, and mucosa-associated lymphoid tissue following experimental or natural TSE contamination [12
, 23
, 26
27
28
, 53
54
55
]. Furthermore, FDC networks in germinal centers of healthy animals are strongly decorated with PrPC, a feature that could account for their proneness to capture prions and to become infected [47
, 56
]. Additional indirect evidence supports their involvement in neuroinvasion: the maintained permissiveness to TSE of total body 
-irradiated mice [57
]; the susceptibility of PrP mismatched chimeras [47
]; the role of complement fractions C1q and C3 early in the infectious process, suggesting that FDCs could capture prions through their complement receptors [58
, 59
]; the resistance of SCID, RAG knockout, and µMT mice, which besides other obvious alterations, are all devoid of functional FDCs; and finally, the total dependency of this subset on B cells for its maturation.
To provide more direct evidence for FDC implication in the neuroinvasion process, several groups have exploited the fact that the differentiation and maintenance of FDC networks in germinal centers depend on cytokines such as tumor necrosis factor (TNF-) and lymphotoxins (LT) and ß, which are locally secreted by B cells [60
]. The genetic amputation of any of these cytokines or the blocking by gene invalidation or competition with soluble receptors of their signaling pathways through receptors TNFR-1, TNFR-2, and LTßR causes immediate impairment of FDCs together with complementary effects on the microarchitecture of secondary lymphoid structures. Paradoxically, all these models of invalidation, which were supposed to provide clear answers regarding FDC contribution, have generated a good deal of controversial and confusing results imputable to the fact that TNF or LT have pleiomorphic effects that extend beyond the mere disappearance of FDCs and to the fact that there have not been attempts at standardizing experimental procedures. Although some coherence emerges from data obtained in a given laboratory, there are striking discrepancies between groups using different strains and protocols. A good illustration of those limitations is found in a series of studies from the Aguzzi group [41
], who reported first that TNFR1 knockout mice were normally sensitive to peripheral challenge with RML scrapie in spite of their FDC deficiency. Shortly after, the same group reported that temporary abrogation of FDCs by inhibition of the LTßR signaling pathway with soluble receptor significantly prolonged the incubation period, a result that was independently confirmed by Mabbott et al. [61
] using the ME7 scrapie strain [62
]. In a recent publication, the group finds that LT, LTß, and LTßR knockout mice are resistant to peripheral challenge with RML, and TNFR-1, TNFR-2, and some TNF- knockout mice are almost normally susceptible [63
]. Another recent study, also using the RML strain but at much higher doses, partially confirms those results by showing that LTß knockout mice are totally refractory to peripheral infection, and LT knockout mice are partially permissive. Altogether, these reports suggest that even if prions do normally replicate in FDCs, they can also replicate in "extra-FDC" sites provided that other, still enigmatic cell subsets, probably associated with the microarchitecture of the secondary lymphoid follicles, are maintained. It would seem that, on the basis of the published results, inhibition of the LTßR pathway used by the biologically active LT/ß heterotrimer would affect more deeply such extra FDC sites of replication than the inhibition of the TNFR1/TNFR2 pathways used by TNF and LT homotrimers. However, again, this is counting without the disease characteristics of each individual prion strain. Using the same LTß knockout mice infected with the mouse-adapted FU strain of human GSS origin, Manuelidis et al. [64
] reported that FDC-deficient mice were fully permissive to peripheral exposure, with only a small difference in incubation periods as compared with wild-type mice. Thus, the FU strain seems to use different pathways of lymphoinvasion that seem to bypass B cells, FDCs, or any other cell type that would require the presence of LTß.
In summary, several lines of evidence support the involvement of FDCs as sites of prion concentration and replication, and the identification of FDCs as a potential target for therapeutic or prophylactic interventions is justified. Conversely, FDCs cannot account for the entire process of neuroinvasion. It is worth noting that FDCs form immobile networks with no evident connection with the nervous system. This suggests that other cells should be involved in carrying prions from sites of exposure to germinal centers and then to sites of neuroinvasion. Microscopic and ultrastructural observations suggest that FDCs shed PrPSc [65 ], which seems to be captured by other cell types. We have evoked a possible passage to T or B lymphocytes, but other subsets closely resembling macrophages or DCs have also been described and implicated.
Macrophages and DCs
Macrophages of the lymphoid follicles have been shown to contain PrPSc at early stages of TSE infections and at terminal disease [21
, 26
, 27
, 65
]. Accumulation was found in tingible body macrophages, a subset which seems to be specialized in phagocytic activity in the germinal centers [65
]. In another study performed on naturally infected sheep, PrPSc deposits were associated first on the CD68+ cells of the dome area in Peyer’s patches before being detected on FDCs [26
]. Finally, cells resembling macrophages were found to contain PrPSc in FDC-deficient mice infected with scrapie [61
, 63
]. Thus, macrophages could provide the alternative sites of prion accumulation and replication in the absence of functional FDCs. They could also have a specific and ambiguous role in upstream or downstream FDC involvement in TSE pathogenesis. Macrophages might propagate the infectious agent but also contribute to its clearance, as suggested by in vitro and in vivo experiments [66
, 67
].
DCs have been curiously ignored by most pathologists or have been erroneously assimilated to FDCs [12 , 68 ] in spite of their completely different nature. Contrary to FDCs whose lineage is ambiguous, DCs belong to a hematopoietic lineage [69 ]. They are phenotypically close to monocytes and macrophages with which they share a number of surface markers such as CD11b, F4/80, and CD68. DCs express high levels of PrPC, which increase during maturation in parallel with surface major histocompatibility complex (MHC) class II molecules and costimulatory molecules [70 ]. Their trafficking properties and the fact that they can retain endocytosed particles without degradation, for longer periods of time than macrophages, make them obvious candidates for disseminating prions through tissues and body fluids. A first direct indication of DC implication in the transport of prions from the gut to the mesenteric lymph nodes comes from experiments by Huang et al. [71 ]. The authors have demonstrated, by immunostaining, the presence of PrPSc on DCs shortly after oral infection of rats with ME7 scrapie. As DCs can sample antigens from the gut, after transcytosis in M cells [72 ] or directly in the lumen itself [73 ], they may be essential transporters of prions from the site of exposure to the lymphoid sites of replication.
We have specifically addressed the role of DCs in the propagation of prions toward the sites of neuroinvasion after their initial accumulation in lymphoid tissues [50
]. The model consisted of inoculating mice i.p. with 139A prions, collecting their spleens 8–10 weeks later, and testing fractionated B and DC subsets as lysates for determining infectivity titres in an animal bioassay or as live cells for adoptive transfer into RAG-1°/° mice. The RAG-1°/° mice have no T cells, no B cells, no FDCs, and no lymphoid follicles and are resistant to peripheral challenge with prions. High-infectivity titers were noticed in the DC lysate, but it is most interesting that injection of live, infected DCs induced scrapie with a reasonable incubation period in every RAG°/° recipient. It was verified that germinal centers had not been restored in those mice, thus excluding a contamination of the DC inoculum by lymphocyte precursors, and that PrPSc had not accumulated in the spleen. Thus, DCs can propagate prions directly from the periphery to the CNS in the absence of any additional lymphoid elements and with no need of PrP replication in such sites. As a result of their unique migration properties, they could be essential factors of prion spreading to the sites of neuroinvasion by bringing prions to the vicinity of neurone endings or at the hematomeningeal interface. More direct evidence for DC implication will come from experiments where these cells will be functionally deleted or altered through conditional knockout or via the manipulation of chemokines and chemokine receptors. A crude attempt at evaluating the impact of DC elimination on the progression of mouse scrapie has been recently reported [74
]. Deletion was achieved by viral infection with a particular strain of lymphochoriomeningitis virus, which targets the DCs and suppresses their functional activity. After inoculation of high doses of RML prions, it was observed that the disappearance of DCs had not modified disease progression, but these conclusions need to be confirmed in more controlled and refined experimental settings. Finally, it should be stressed again that the various knockouts, which result in the abrogation of FDCs, notably LT°/° and LTß°/°, may impact on other cell types that express TNF and LT receptors. DCs, for instance, are sensitive to LT at the stage of hematopoiesis and of functional maturation [75
]. Thus, what has been attributed to a lack of FDCs might in fact be a result of a functional impairment of DCs also. The enigmatic cells bearing macrophage markers [63
] and responsible for prion storing in lymphoid tissues of mice devoid of FDCs might well be DCs.
In summary, the large corpus of experiments that has been reviewed demonstrates clearly that with a few possible exceptions such as BSE infection in cows, lymphoid structures play a key role in the stages that precede neuroinvasion. Mice with profound immune defects display evident resistance to peripheral prion challenge. Conversely, it seems difficult at the present time to ascribe the entire invasion process to a single cell type. Several subsets, separately or more probably in cooperation, likely contribute to the sampling, concentration, replication, and dissemination of TSE agents toward the CNS. This is further complicated by the fact that every prion strain seems to have its own specific pathway. Under such conditions, extrapolations from one model to another become extremely risky. Finally, one may raise the question of the real purposes of lymphoinvasion in the pathogenesis of TSEs. One obvious logical purpose is to transport prions from where they enter to sites of neuroinvasion. Another function implicitly assumed but not demonstrated is the necessity for prions to replicate and to reach a critical mass before they can invade the brain. We would like to suggest a third hypothesis, namely the possibility that qualitative modifications of TSE agents occur in cells of the lymphoreticular system. Certain observations, such as differences of glycoform patterns between PrPSc extracted from spleen and brain of mice infected by the 139A scrapie strain, would be compatible with such views [76 ].
|
IMMUNOINTERVENTION TARGETED AT PrP |
|---|
|
TOPABSTRACTINTRODUCTIONAN OVERVIEW ON PRION...THE LYMPHORETICULAR SYSTEM AS...IMMUNOINTERVENTION TARGETED AT...PERSPECTIVESREFERENCES |
|---|
Although self-tolerance to PrP remains a real obstacle to vaccination attempts, several groups have nervertheless started to evaluate, through in vitro models or by using transgenic mice, the potential benefits of an immune response.
A first encouraging indication has come from two recently published studies showing that anti-PrP antibodies added to the culture medium could cure infected cell lines by preventing the conversion of PrPC into PrPSc [81 , 82 ]. In both studies, the lines or clones were derived from N2a, a murine neuroblastoma of A/J strain origin, which constitutively expresses PrPC and can be infected with relative ease by coculture with infectious brain extracts from scrapie mice. Infected N2a cells can continuously produce PK-resistant PrPSc and remain infectious for mice even after multiple in vitro passages [83 ]. They are excellent models for studying the cellular and molecular events involved in PrP conversion and for screening potential inhibitors of the reaction [84 ].
The monoclonal antibodies (mAb) that prevent in vitro cellular conversion of PrPC into PrPSc are not specific for one particular isoform, as they bind equally to both. Conversely, not every anti-PrP antibody has the capacity to block conversion with the same efficacy. As shown in both studies, the binding to a region of the PrP molecule between residues 132 and 156, corresponding to the first helix of PrPC, appears to be critical. How these antibodies inhibit conversion is still not totally understood. They could possibly hinder physical contact between the two conformers or prevent the docking of an auxiliary cofactor catalyzing transconformation. Alternatively, it is possible that the antibodies redirect PrPC traffic and sequester the protein in a subcellular compartment where conversion cannot take place, as it is inaccessible to PrPSc, or the particular physicochemical conditions do not favor conversion. Finally, a less exciting explanation with doubtful in vivo relevance might be that antibodies enhance selective pressure against infected cells and let uninfected cells over-grow.
The question of whether antibodies can also prevent conversion in vivo has actually been addressed recently [85
]. The authors have produced a transgenic mouse expressing the rearranged VH domain of anti-PrP antibody 6H4, one of the mAb that blocks conversion in vitro. The transgenic line develops a B cell repertoire that is heterogeneous because of the free endogenous rearrangements of 
and 
light chains but is still biased toward PrP specificities as a result of the constraint imposed by the rearranged 6H4 heavy chain. 6H4 µ-Chain transgenic mice produce spontaneous anti-PrP antibodies of immunoglobulin M isotype, reactive in enzyme-linked immunosorbent assay, and Western blot assays. PrPSc replication and scrapie infectivity have been followed in tissues of these mice after i.p. inoculation of RML prions. Compared with wild-type mice, replication of prions is significantly delayed in the 6H4 µ-chain transgenics. There is also considerably less accumulation of PrPSc in the spleen and in the brain. Thus, endogenously produced anti-PrP antibodies can inhibit scrapie progression in vivo. However, the report does not mention whether clinical disease is effectively delayed and if so, for how long. Although the mechanisms of inhibition remain to be clarified, the results are encouraging, as they suggest first, that antibodies can antagonize prion conversion in vivo; second, that B cell clones directed at self-PrP are not necessarily doomed to clonal deletion or peripheral anergy; and third, that anti-PrP autoantibodies do not cause apparent manifestations of autoimmunity. This latter issue must be considered with a certain degree of caution after the report of Souan et al. [86
] showing that Lewis rats immunized with immunogenic peptides from homologous PrP may develop severe skin inflammation.
Cell-mediated immunity against PrP has by far not been as intensively explored as humoral immunity. Yet, there are good reasons to believe that T cells recognize processed epitopes of PrPC or PrPSc, as anti-PrP antibodies produced under heterologous conditions or in Prnp°/° mice are subject to isotypic switching, an event that requires the cooperation of antigen-specific helper T cells. Souan et al. [86 , 87 ] have addressed the question of the T cell-mediated response against PrP in two recent studies. In the more detailed study, the authors have immunized normal wild-type mice of three independent strains, NOD, C57BL/6, and A/J, with a library of synthetic peptides covering almost the entire sequence of mouse PrP. After a few cycles of in vitro stimulation with peptides and antigen-presenting cells (APC), they were able to indentify two epitopic motifs that recall CD4+ T cells in the three strains of mice, despite their different MHC haplotypes. One peptide spans from residue 131 to 150, the very same region that is critical in the blocking of PrP conversion by antibodies; the other is at the C-terminus between residues 211 and 230. Finally, the authors show that in compatible A/J mice, which have been immunized with immunogenic PrP peptides and have received a transplant of N2a-infected cells, the amount of PrPSc recovered from the growing tumors is significantly lower than in tumors growing in naive controls. Thus, an ongoing anti-PrP immune response seems able to antagonize the in vivo conversion of PrP. Like the report by Aguzzi and co-workers [85 ], this study is encouraging in several respects. It suggests that, although rather indirectly, anti-PrP immunity might slow down PrP conversion and that T cell tolerance can be overcome in normal mice by active immunization with synthetic peptides with no adverse manifestation of autoimmunity. However, several questions remain to be answered: the effect of peptide immunization on the natural disease and the respective contributions of cell-mediated and humoral immunity, as antibodies are produced concomitantly by B cells as a consequence of helper T cell activation. Starting from experiments made in Prnp knockout mice not tolerant to PrP, our own observations confirm that PrP is normally processed and presented to T cells by APC and that T cell lines can be generated against a few dominant epitopes residing in the same region-spaning residues, 142 and 186. However, at variance with the results of Souan et al. [87 ], we found that wild-type C57/B6 mice are rather refractory to immunization with PrP peptides. Relatively few clones emerge from in vitro cultures and require numerous iterative cycles of restimulation in the presence of antigen and interleukin-2. Some epitopes are the same as those identified with T cells from Prnp knockout mice; others are not, suggesting that the T cell repertoires in mice expressing or not expressing PrP are not totally overlapping. We anticipate that, on the basis of our results, breaking tolerance in normal individuals will probably require sophisticated strategies similar to those used for immunization against poorly immunogenic tumors. Tumor immunology has pionneered in this field using strategies based on strong bacterial adjuvants, modified peptides, antigen-loaded DCs, and T cell clones redirected by transfection of high-affinity TCRs. Similar strategies will probably have to be developed for breaking tolerance and antagonizing prion propagation in infected hosts.
|
PERSPECTIVES |
|---|
|
TOPABSTRACTINTRODUCTIONAN OVERVIEW ON PRION...THE LYMPHORETICULAR SYSTEM AS...IMMUNOINTERVENTION TARGETED AT...PERSPECTIVESREFERENCES |
|---|
The other important question is obviously to identify the cell(s) that is involved in lymphoinvasion and to define the precise events, i.e., accumulation, replication, and transport, which take place in such subsets. Good cell candidates have been identified, such as FDCs, DCs, and macrophages, possibly assisted by auxiliary subsets such as B and T cells, which do not seem to support replication but might carry and disseminate the infectious agent. Here too, the available experimental data suggest that the relative importance of a given subset may depend on the prion strain. Differences reported between RML, ME7, and FU models of peripheral inoculation give a good illustration of the importance of this parameter. It will thus be essential to understand when, where, how, and which prions are propagated before developing intervention strategies that should be sufficiently focused so as not to jeopardize the immune status of the host.
The recent and still preliminary studies on anti-PrP immunity raise the hope that immune effectors could constitute a line of defense against prion invasion. The way such effectors might work is still unclear. Antobodies seem able to prevent the conversion of endogenous PrPC into PrPSc, a process that is essential for the replication of the infectious agent and the development of histopathological lesions in the CNS. Other immune effector cells such as proinflammatory CD4+ T cells or CD8+ cytotoxic T lymphocytes might also be effective in preventing prion spread by eliminating cell reservoirs of infectious agent. Actors of innate immunity, complement, macrophages, and natural killer cells will have to be considered as well as agents facilitating invasion but also as possible contributors to host defenses. The implication of complement in the capture of prions by FDCs or DCs and that of macrophages in the clearing of infectivity in the spleen illustrates this possibility.
Finally, it is possible that the main obstacle to vaccinotherapy will come from the tolerance of the immune system to prions and more specifically to PrP. Some recent publications suggest that this may not be a problem, provided that the adequate peptides of PrP are identified. Our personal experience suggests that tolerance to PrP is relatively strong in normal individuals and resists usual strategies of peptide vaccination. This is obviously an issue of extreme importance if one day we want to effectively prevent TSE with immunological tools.
|
ACKNOWLEDGEMENTS |
|---|
Received June 4, 2002; revised July 23, 2002; accepted July 25, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONAN OVERVIEW ON PRION...THE LYMPHORETICULAR SYSTEM AS...IMMUNOINTERVENTION TARGETED AT...PERSPECTIVESREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. Levavasseur, P. Metharom, G. Dorban, H. Nakano, T. Kakiuchi, C. Carnaud, P. Sarradin, and P. Aucouturier Experimental scrapie in 'plt' mice: an assessment of the role of dendritic-cell migration in the pathogenesis of prion diseases J. Gen. Virol., August 1, 2007; 88(8): 2353 - 2360. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Martinez del Hoyo, M. Lopez-Bravo, P. Metharom, C. Ardavin, and P. Aucouturier Prion Protein Expression by Mouse Dendritic Cells Is Restricted to the Nonplasmacytoid Subsets and Correlates with the Maturation State J. Immunol., November 1, 2006; 177(9): 6137 - 6142. [Abstract] [Full Text] [PDF]
|
|
