Cells and prions: A license to replicate

1. Prion diseases: infectious neurodegenerative disorders in humans and animals 

Prion diseases, also designated as transmissible spongiform encephalopathies (TSEs), are infectious and invariably fatal neurodegenerative disorders affecting humans and animals [1], [2], [3]. These include kuru, Creutzfeldt–Jakob disease (CJD), Gerstmann–Sträussler–Scheinker disease (GSS) and fatal familial insomnia (FFI) in humans, as well as scrapie of sheep and goats, bovine spongiform encephalopathy (BSE) or “mad cow disease” of cattle and chronic wasting disease (CWD) of deer and elk [1], [2], [3].

The last century has witnessed a tremendous series of tragic episodes related to prion diseases [2]. Until the middle of the 20th century kuru decimated Papua New Guinea aborigines devoted to ritualistic consumption of central nervous system (CNS)-derived tissues [4]. The employment of prion-contaminated gonadotropins, corneal transplantations, dura mater grafts and surgical instruments has resulted in more than 250 cases of iatrogenic CJD in the 1970s with drastic implications for clinical practice [2]. Starting from 1986, the epidemic of BSE has affected more than 280000 cattle, provoking a food crisis with unprecedented, worldwide economic consequences. Transmission of the BSE agent to humans has been regarded as the cause of a new clinico-pathological entity, termed variant CJD (vCJD), which was first described in 1996 [5], [6] and has so far caused death in 212 individuals (http://www.cjd.ed.ac.uk/vcjdworld.htm). The rising incidence of vCJD in the late 1990s and the possibility that millions of people have been exposed to BSE-contaminated meat have evoked fears of an upcoming pandemic.

2. The versatile prion 

The beginning of the present century has brought the encouraging news of a stabilizing or subsiding incidence of vCJD, but novel alarming facts are indicating that prions are much more versatile than previously thought. The occurrence of horizontal transmission of vCJD by blood transfusions [7], [8] has indicated that vCJD prions can recycle among subclinical humans. This trouble was not necessarily predicted by mouse models [9], [10] and poses a new, worldwide challenge for health authorities in terms of donor deferral criteria and blood quality surveillance, while a specific and sensitive test for the detection of prions in human blood is still pending [11]. Only recently, the Health Protection Agency of the United Kingdom has reported a case of a hemophilic patient who, most likely, has acquired vCJD prions through Factor VIII preparation derived from plasma donated by a “preclinical” vCJD patient (http://www.hpa.org.uk). Moreover, recent reports from work with kuru patients in New Guinea indicate that heterozygosity at polymorphic codon 129 (129M/V) of PRNP, the gene encoding the human prion protein, could lead to incubation periods exceeding 50years [12].

These findings demonstrate that we need to learn more about horizontal prion transmission in humans, how to detect subclinical carriers and if or how those can become a potential risk for public health. Future observations will have to show whether any of the above mentioned will impact on the epidemiology of vCJD.

Moreover, a recrudescence of scrapie outbreaks among European sheep flocks has been described in the last years and new cases of BSE and vCJD have been reported in countries and continents previously deemed prion-free [13]. In this regard it is important to mention that is unknown by today whether scrapie prions can pose a potential health risk to humans.

In addition, an atypical BSE strain – called bovine amyloidotic spongiform encephalopathy (BASE) – has been identified in different countries displaying biochemical properties similar to a subgroup of sporadic CJD (sCJD) [14]. These data questioned whether a temporarily defined increase in sCJD cases (e.g. in Switzerland) was a result of BASE transmission to humans [15]. Furthermore, the spread of CWD of cervids in North America and the potential dietary-exposure of millions of individuals to prion-contaminated venison have raised the still unsolved question of human susceptibility to CWD prions [16].

Recent inventions and refinements in the technologies for detection of the disease-associated prion protein have significantly extended and shifted our knowledge about prion tropism [17]. The general belief that sCJD prions are rather confined to the CNS when compared to vCJD prions, which also colonize peripheral organs, was first shaken by investigating extraneural organs of sCJD patients using novel or advanced techniques [18], [19], [20]. Disease-associated prion proteins have been detected in the olfactory epithelium [21], as well as in spleen and muscle tissue of sCJD victims [19]. Moreover, prion infectivity was demonstrated in saliva, milk, blood and muscles of TSE affected individuals [22], [23], [24], [25] and, as substantiated below, in chronically inflamed organs (e.g. mammary gland, liver and kidney) [26], [27], [28]. Scrapie infection in nephritic mice leads to urinary prion secretion (also termed prionuria), even at subclinical stage [29]. Further, moderate intestinal inflammation at the time of prion exposure increases the susceptibility to orally administered prions [30]. Taken together, these observations indicate that environmental factors (e.g. inflammation) can change the prion tropism to organs hitherto believed prion-free [17], [28], [31]. Nevertheless, peripheral prion accumulation was reported to occur also in excretory organs and body fluids under non-inflammatory conditions, however to much less degree [25], [32]. In summary, these data question the current risk assessment of high-infectivity organs in humans and animals, so far mainly including the CNS and the lymphoreticular system (LRS).

Although prion diseases have always been regarded as a “sui generis” class of infectious maladies, recent findings have challenged this notion. Data from murine models of Alzheimer’s disease indicate that exogenous Aβ amyloid is capable of inducing cerebral Aβ amyloidosis with disease phenotypes that are governed by both the host and the inoculum, features reminiscent of prion strains [33]. These findings are fascinating and support the hypothesis that pathogenic mechanisms operating in prion diseases might be shared by other neurodegenerative disorders [34], [35].

Despite substantial progress in our understanding of prion pathogenesis, a number of crucial questions still remain unanswered. These include the exact nature of the infectious agent, the cellular and molecular mechanisms of central and peripheral prion replication, the defining traits of prion strains, the determinants of prion toxicity, the routes of horizontal transmission and the mechanistic basis of species barrier.

Here, we review the current state of art in prion biology and prion immunology. In addition we examine the currently available tools to detect prions and summarize the immunological interventional strategies against prion diseases explored so far.

3. The infectious agent 

The etiologic agent of TSEs is termed prion (proteinaceous infectious only) and displays unconventional properties, such as resistance to UV irradiation, exposure to high pressures or temperatures and formaldehyde treatment. According to the protein-only hypothesis, it is devoid of informational nucleic acid and coincides with scrapie prion protein (PrPSc), an abnormal isoform of PrPC, which is capable of converting PrPC into a likeliness of itself [36].

Despite considerable efforts, no posttranslational chemical modifications that might discriminate between PrPC and PrPSc were found, indicating that solely conformational changes distinguish the two PrP isoforms [37]. The fine structure of PrPC is known at the atomic level [38]. On the other hand, no high-resolution structure is available for PrPSc, but limited data from low-resolution structural methods are compatible with a significantly different conformation of PrPSc with respect to PrPC [39], [40].

4. Prion strains 

The prion strains phenomenon adds a further level of complexity to the question of protein structure [41]. Prion strains are TSE isolates or sources of prion infectivity that, upon inoculation into genetically identical hosts, cause disease with consistent characteristics, including incubation time, lesion profiles within the CNS and even tropism for extracerebral cell types [41], [42], [43]. To accommodate the existence of different prion strains within the frame of the protein-only hypothesis, one should postulate that PrPSc must exist in various distinct pathological conformations, each one able to impart its own conformation onto PrPC, thus culminating in distinct disease characteristics [41], [44].

5. Detection of prions: from mice to test-tubes 

In this context, detection of bona fide prions is achievable only through determination of prion infectivity. This is classically performed by bioassay in which serial dilutions of the test material are inoculated into experimental animals, and the dilution at which 50% of the animals contract the disease (termed ID50) is determined [45]. However, this system suffers from inaccuracy, is time-consuming and is limited by the requirement of scores of animals.

These limitations are partly overcome by a recently developed assay based on prion-susceptible cell lines, termed scrapie cell endpoint assay [46]. This method combines the sensitivity and intrinsic biological validity of the bioassay (i.e. direct measurement of infectivity) with the speed, convenience and amenability to high-throughput automation of an in vitro system. Extension of the susceptibility of neuronal cell lines used in this assay, currently restricted to various murine prion strains, could pave the way for a sensitive test to detect prion infectivity, with both medical and veterinarian applications.

As an alternative to prion infectivity assessment, various biochemical or biophysical peculiarities of PrPSc, presumably stemming from its differential conformation with respect to PrPC, can be operationally used as surrogate markers for prion infection.

The first reliable surrogate marker of prion infectivity is its partial resistance against proteolytic degradation. Incubation with 50μg/mL of proteinase K (PK) at 37°C for 2h does not degrade the carboxyl-proximal domain of the disease-associated prion protein, nor decrease the infectious titer of the prion preparation [47]. More than 25years after it was first described, the detection of PK-resistant prion protein still remains the gold standard for the biochemical diagnosis of prion diseases and forms the basis of all the marketed BSE tests.

Biochemical detection of PK-resistant prion protein can be performed in combination with histological techniques. In particular, blotting either paraffin-embedded tissue (PET blots) [48] or cryosections (histoblots) [49] onto nitrocellulose membrane, followed by digestion with PK and decoration with anti-PrP antibodies enables the detection of PK-resistant prion protein in situ, thus providing precious information about topographic distribution of the pathologic protein in disease tissues.

PK-resistant prion protein is not always easily detectable in tissues or body fluids from affected hosts. In 2001 a significant step towards a more sensitive method of detection of PrPSc has been achieved [18]. Selective precipitation of PrPSc from tissue homogenates with sodium phosphotungstic acid results in the concentration of PrPSc from large volumes of test material. Combination of this preparative step with immunoblot detection results in increased sensitivity of Western blot analysis by up to three orders of magnitude and has been extensively exploited to detect PrPSc in extraneural sites of TSE victims [18], [19]. In the same year another milestone for sensitive detection of the pathological prion protein is represented by the so-called protein misfolding cyclic amplification (PMCA) [50]. Analogously to a PCR reaction, PMCA generates PrPSc from monomeric substrates and a small amount of template PrPSc. Amplification of PrPSc is achieved through cycles of sonication, which disrupts the PrPSc aggregates, and elongation, that entails incubation with the monomeric substrate which is recruited by the misfolded prion protein, and is accompanied by an increase in prion infectivity. This technology succeeded to detect amplifiable PrPSc in blood of prion-infected hamsters in the presymptomatic phase of the disease [51], suggesting that PMCA might prove to become a sensitive non-invasive method for early diagnosis of prion diseases. However, the fact that PMCA can generate de novo prions from non-infectious material [52] emphasizes that the technique may be prone to false positives, thus questioning its usefulness as a diagnostic tool [11]. Besides this limitation, PMCA could also prove instrumental in expanding our understanding of prion replication mechanisms [11].

An additional strategy to detect the disease-associated protein relies on the assumption that, upon conversion to PrPSc, at least some epitopes of PrPC are buried in the aggregates. The measurement of the differential binding of anti-PrP antibodies to native vs. denaturated prion protein, which constitutes the basis of the conformation-dependent immunoassay, provides a sensitive assay to detect PrPSc and a good tool to investigate the existence of conformational differences between distinct strains of prions [53].

6. The physiological function of PrPC: PrPC and the CNS 

PrPC is tethered to the external surface of cells by a glycosyl phosphatidyl inositol (GPI) anchor and is enriched in detergent-resistant microdomains of cellular membrane termed lipid rafts. It undergoes facultative N-linked glycosylation at two sites, which results in un-, mono- or diglycosylated moieties.

The expression pattern of PrPC is broad, developmentally regulated and includes the nervous system (with high PrPC levels in synaptic membranes of neurons and on astrocytes), secondary lymphoid organs, skeletal muscle, kidney and heart. PrPC is highly conserved among mammals, and paralogues thereof are present in birds, reptiles, amphibians and possibly in fish. No naturally occurring Prnp-null alleles have ever been observed in any mammalian species. These observations suggest a broad and conserved function for the protein [2].

Despite intensive investigations and the availability of Prnp0/0 mice since 1992, the physiological function of PrPC has not been clearly identified. Prnp ablation per se does not elicit neurodegeneration [54], even when induced postnatally [55]. Therefore, prion pathology is unlikely to be the result of a loss of PrPC function. However, PrPSc conversion might alter the physiological function of PrPC and confer a toxic dominant function. This could include altered signal transduction, enzymatic activity, and change in substrate specificity or protein binding properties. In this scenario, elucidating the physiological function of PrPC may be instrumental to deciphering the mechanisms of prion pathogenesis and eventually to devising tailored interventional strategies. A non-exhaustive list of putative functions of PrPC includes signal transduction, regulation of circadian rhythm, copper binding, proliferation of neural precursors, processing of sensory information by the olfactory system, cellular iron uptake and transport, pro-apoptotic or anti-apoptotic function and others [56].

7. PrPC, prion replication and neurodegeneration 

The only universally acknowledged function of PrPC is to replicate prions and mediate their toxicity: Prnp0/0 mice are resistant to prion infection [54]. Brain tissue devoid of PrPC is not damaged by exogenous PrPSc, as demonstrated by refined brain grafting experiments of PrPC overexpressing brain tissue into Prnp0/0 brains [57]. Surprisingly, depleting endogenous neuronal PrPC in prion-infected mice reverses early spongiosis and prevents neuronal loss and progression to clinical disease, despite the accumulation of extraneuronal PrPSc to levels normally found in terminally sick wild-type (wt) animals [55]. Moreover, scrapie-infected transgenic mice exclusively expressing a monomeric, soluble secreted form of PrPC that lacks the GPI anchor do not develop overt prion disease, while PrPSc accumulate in their brains in form of amyloid plaques [58]. Collectively, these findings indicate that PrPSc is per se innocuous and that prion replication avails itself of membrane-bound PrPC on neurons to elicit neurotoxicity [59].

Recently, Strittmatter and colleagues have provided evidence that PrPC is a high-affinity cell-surface receptor for soluble synthetic Aβ oligomers on neurons and might play a central role in the pathophysiological process of Alzheimer’s disease [60]. This finding is intriguing and future investigations will have to elucidate whether PrPC binds to naturally occurring Aβ oligomers in vivo. In addition, the clinical relevance and the therapeutic implications of this discovery have to be clarified [61].

8. PrPC and the immune system 

With regard to the immune system, PrPC is expressed in T and B lymphocytes, natural killer (NK) cells, platelets, erythrocytes, monocytes, dendritic cells (DCs) and follicular dendritic cells (FDCs), albeit with significant differences across species and among states of maturation and subsets of immune cells [17]. PrPC has been implicated in T lymphocyte development, activation and in the interaction of T lymphocytes and DCs [62], [63]. Overexpression of PrPC in transgenic mice has been shown to alter T cell development in the thymus via local generation of an antioxidant milieu [64], but recent work of Zabel et al. has indicated that this phenotype might be caused by an insertional mutation of the Prnp transgene [65].

Genetic or pharmacological ablation or PrPC in macrophages results in an increased rate of phagocytosis of various apoptotic cells, indicating that PrPC is a negative regulator of phagocytosis [66]. PrPC has also been shown to promote the swimming internalization of Brucella abortus into macrophages through the interaction with bacterial Hsp60 [67], [68]. However, these data are highly controversial and one publication could not reproduce the finding by Watarai et al. [69]. Furthermore, another study indicated that PrPC is expressed on the surface of hematopoietic stem cells and supports their self-renewal suggesting that PrPC might be a critical survival factor for hematopoietic stem cells [70].

For those functions ascribed to PrPC on the basis of Prnp0/0 mice, major concerns arise about the possibility of a genetic artifact [71]. For example, alleles in linkage disequilibrium with the deleted gene (which is constantly selected for by investigators during breeding) could be actually responsible for the described phenotype. Future investigations with Prnp0/0 mice will have to face this and other caveats, and devise genetically stringent experiments to assess whether the presumed phenotypes are a primary effect of Prnp deletion and how this is related to the physiological function of PrPC.

9. Prions and the immune system: a fatal tête à tête 

Although prion diseases are neurodegenerative disorders it has been established already a long time ago that infectious prions do not only accumulate in the CNS but can also colonize extraneural organs of infected individuals. Interestingly, in contrast to the central nervous tissue prions are not believed to induce tissue damage in secondary lymphoid organs, although some prion related aberrations were reported [72].

It was already shown in the early days of prion immunology that upon peripheral prion exposure secondary lymphoid organs appear to play an important role in the development of prion disease: genetic asplenia or splenectomy of mice prior or after peripheral prion challenge prolongs the life span of scrapie-infected mice. In contrast, thymectomy or genetic athymia had no significant effect [73] suggesting a dispensable role for T-lymphocytes in prion pathogenesis. Further experiments investigated the cascade of events upon prion exposure in more detail: splenectomy after intraperitoneal (i.p.) prion inoculation revealed that peripheral prion pathogenesis becomes independent of the spleen once prions have reached the spinal cord [74]. However, a splenic replication phase is not obligatory in all rodent TSE models and might be strain-dependent [43], [75]. For example Tateishi and coworkers have found no effect of splenectomy on incubation times for the Fukuoka-1 strain, a mouse-adapted GSS prion [76], [77]. This is one of several hints that different prion strains appear to differ in their tissue tropisms: they are either lymphotropic, indicating that they colonize lymphoid organs right after peripheral infection before invading the CNS, or neurotrophic – meaning that they can invade the brain without replicating in the LRS [41], [43], [44].

Before the invention of fluorescence-activated cell sorting, the cellular distribution of prions in the LRS was studied by separation of splenocytes into various subpopulations based on buoyant density of different adherence of cells to plastic surface [78]. In these early experiments, it was reported that cells with relatively high specific infectivity had a density characteristic of lymphoblasts, myeloblasts and macrophages. Enrichment of macrophages did not enhance scrapie infectivity. Already in these (relatively crude) fractionation attempts it became clear that the stromal compartment contained ∼10 times more prion infectivity than the pulp [79]. Sublethal doses of gamma irradiation, which affect mitotically active cells of hematopoietic origin but not resident post-mitotic cells, failed to alter the incubation period of the disease [80]. These findings suggested that scrapie replication or accumulation in the LRS largely depends on radioresistant, post-mitotic cells localized within the stromal compartment. However, these experiments did not address the exact cellular and molecular preconditions needed for efficient prion transport from the site of entry (e.g. gut) to the LRS and prion invasion into the CNS.

10. Neuroinvasion proper 

The use of Prnp0/0 mice [54] and of bone marrow reconstitution as a technique to generate chimeric mice with stromal and hematopoietic compartments of different genotypes revealed a first surprise: PrPC itself is involved in transporting prion infectivity from peripheral sites to the CNS. Titration experiments indicated that adoptive transfer with wt bone marrow into Prnp0/0 mice reconstitutes the capability of the spleen to accumulate prions of the mouse-adapted Rocky Mountain Laboratory (RML) scrapie strain [81], [82], [83]. These results were taken to suggest that PrPC-expressing hematopoietic cells transport prions from the entry site to the LRS, where prions are efficiently replicated. However, the elemental compartment for prion neuroinvasion appears to be non-hematopoietic, since it cannot be adopted by bone marrow reconstitution [81], [82], [84].

In conclusion, after peripheral exposure – may this be by ingestion or by peripheral infection – prion pathogenesis can be regarded as a dynamic process that can be split spatially and temporally [3]: (1) infection and peripheral prion replication, (2) prion neuroinvasion and (3) progressive, fatal neurodegeneration.

But what are the exact cellular and molecular mechanisms underlying those three major stages? Astoundingly, B-lymphocytes were demonstrated to be of crucial importance for prion accumulation in the LRS as well as for neuroinvasion [84]. However, PrPC expression on B-lymphocytes is not required [85], [86]. This combined with the fact that a stromal compartment was hypothesized to be the essential mediator of neuroinvasion, indicated that B-lymphocytes themselves are unlike to represent a major replicative unit for prions. Instead, B-lymphocyte involvement in peripheral pathogenesis was believed to be indirect, such as supporting the development or maintenance of a prion replicating cell type that should express PrPC, be – most likely – of stromal origin and in close proximity to B-lymphocytes.

11. FDCs: Cells with the license to replicate prions 

The cell fulfilling these criteria the most was the FDC. FDCs accumulate PrPSc following scrapie infection [87], express high levels of PrPC and are localized in close proximity to B-lymphocytes within splenic and lymph nodal B-cell follicles. Interestingly, FDC development and maintenance was demonstrated to depend on tumor necrosis factor (TNF) superfamily members lymphotoxin α and β (LTα and LTβ), which are cytokines mainly produced by B-lymphocytes, T-lymphocytes and NK cells [88].

Indeed, blockade of LT signaling by administration of a soluble LTβ receptor protein fused to a Fcγ portion (LTβR-Ig) ablates mature FDCs and significantly impairs peripheral prion pathogenesis (Fig. 1) [89]. In line, mice deficient in LT signaling (ltα−/− or ltβr−/− mice) are largely resistant to peripherally administered prions [84]. These data for the first time directly identified a cell type to be involved in peripheral prion pathogenesis in vivo.

View Large Image Download to PowerPoint

Fig. 1. Impact of LTβR signaling on peripheral prion replication competence in secondary lymphoid organs, chronic lymphocytic inflammation (tertiary lymphoid organs) and granulomas. (A) Peripheral prion replication in secondary or tertiary lymphoid organs is mainly accomplished by PrPC+/LTβR+/Mfge8+ follicular dendritic cells (FDCs). Which can be found in splenic white pulp follicles, nodal B-cell follicles or within B-cell clusters of inflammatory foci. In contrast, replication of PrPSc and prion infectivity in granulomas can occur in the absence of FDCs and is most likely accomplished by mesenchymal progenitor cells (MPCs), which are PrPC+/LTβR+/Mfge8−/PDGFRαβ+. (B) Block of LTβR-signaling by repetitive injections of LTβR-Ig dedifferentiates or depletes FDCs in secondary lymphoid organs and leads to a drastic decrease in PrPSc load and prion infectivity. In granulomas, this is different. Block of LTβR signaling by injection of LTβR-Ig does not deplete or dedifferentiate PrPC+ MPCs. Still, prion replication competence is abolished by LTβR-Ig treatment in granulomas, indicating a functional link between LTβR-signaling and prion replication competence in PrPC+ MPCs.

FDCs are still poorly described cells, characterized by the expression of Mfge8 [90]. It is known that FDCs support the maintenance of the lymphoid microarchitecture, trap immune complexes by Fcγ receptors or by binding opsonized antigens to the CD21/CD35 complement receptors. Indeed, various studies have demonstrated that the complement components expressed on or bound by FDCs are relevant to prion pathogenesis: mice genetically engineered to lack complement factors [91] or mice depleted of the C3 complement component [92] exhibited enhanced resistance to peripheral prion inoculation. Further, CD21/35 expressed on FDCs was demonstrated to be involved in targeting prions to FDCs and expediting neuroinvasion following peripheral prion exposure [93].

12. FDCs and peripheral nerves: a long-distance relationship 

It is counterintuitive that FDCs which are sessile cells and of stromal origin, localized in germinal centers of the LRS, should be involved in transporting prions to the CNS. It therefore became crucial to find the prion entry site to the CNS within secondary lymphoid organs. Sympathetic nerves, which innervate lymphoid organs to high degree [94] were hypothesized as a possible entry port. Therefore, many independent studies have focused on the role of the sympathetic nervous system in neuroinvasion from secondary lymphoid organs to the CNS. Indeed, these studies indicated that the autonomic nervous system might be responsible for this process [79], [95], [96], [97]. In line, sympathectomy delays the onset of experimental scrapie in mice, while sympathetic hyperinnervation enhances splenic prion replication and neuroinvasion upon peripheral exposure [98], [99]. This suggested that innervation of secondary lymphoid organs might be indeed the rate limiting step to neuroinvasion [98].

However, there is no direct physical synapse between FDCs localized in germinal centers and sympathetic nerve endings [100]. So how do prions march from hot spots of prion replication to peripheral nerve endings and how do changes in the relative distance of these two poles affect prion neuroinvasion? Experimental evidence pointed to the fact that the distance between FDCs and splenic nerves indeed affects the velocity of neuroinvasion [101]. FDC positioning was manipulated by ablation of the CXCR5 chemokine receptor, which directs B-lymphocytes towards specific micro-compartments [102]. In this model the distance between germinal center associated FDCs and peripheral nerve endings induced superimposition of FDCs and peripheral nerves [101], [102]. CXCR5 deficiency did not affect any aspect of prion pathogenesis within the CNS. However, although velocity of neuroinvasion was similar in CXCR5−/− and wt mice following peripheral administration of high prion titers, delivery of smaller titers resulted in a dose dependent increase in incubation periods in wt when compared to CXCR5−/− mice. Measurement of the kinetics of prion infectivity titers in the thoracic spinal cord showed that increased velocity of prion entry into the CNS of CXCR5−/− mice is due to FDCs juxtaposed to highly innervated, splenic arterioles. This was validated by the prolongation of incubation periods in CXCR5-/- mice depleted for mature FDCs [101].

This study also raised the possibility that the spread of infection to peripheral nerves occurs more rapidly in lymphoid tissues where FDCs are in near proximity to nerves, such as Peyer’s patches [103]. Indeed, it was demonstrated that FDCs are crucial to disease progression for only a very short time window after oral scrapie challenge: if FDCs are depleted in this time window, mice will not succumb to scrapie [104].

Although these data have been extremely informative, the exact mechanism of prion transport from FDCs to peripheral nerves is unknown. Prions could be (1) transported by various cell types leaving germinal centers towards nerve terminals, (2) incorporated by murine budding retroviruses, (3) released in FDC-derived exosomes or (4) passively diffuse from the site of replication to the site of peripheral innervation. Germinal center B-cells [105] as well as DCs, although previously implicated in direct prion transport to the CNS [106], were so far excluded as candidates for the active intrasplenic prion transport [107]. Future experiments will have to reveal the exact mechanisms of intrasplenic prion transport, which is still one of the key questions in the field of prion immunology.

13. Prion replication in inflamed organs: a paradigm to study prion replication competence? 

Because lymphoid infectivity is found in most prion diseases (e.g. sCJD, vCJD, scrapie and CWD) and proinflammatory cytokines and immune cells are involved in lymphoid prion replication [87], [89], [101], [108], [109], [110], it was of interest to test whether chronic inflammatory conditions affect peripheral prion pathogenesis. Chronic inflammation in non-lymphoid organs induced prion replication at sites that were previously believed to be prion-free [26], [27] including liver, pancreas, kidney or mammary gland. This finding, which was initially described in transgenic mice experimentally inoculated with prions, could be reproduced in free ranging, naturally infected sheep or experimentally infected whit-tail deer [26], [111].

All of the inflammatory disorders described above shared the presence of extra follicular structures, the so-called tertiary follicles, which invariably contain FDCs. It was therefore speculated that prion replication in inflammatory foci depends on FDCs: Indeed, ltβr−/− or ltα−/− mice, which display chronic hepatitis without FDCs, did not replicate prions [27].

14. LTβR signaling: from structure to function 

Hence, from these and other experiments with ltβr−/−, ltα−/− or ltβ−/− mice [109] LT signaling was regarded as a mere prerequisite for FDC development and maintenance [112] rather than directly linked to enabling peripheral prion replication competence.

However, recently published work of various groups in vitro and in vivo suggested a more nuanced situation. While crucially dependent on LT signaling in all conditions investigated, extraneuronal prion replication was found to occur in the absence of immunohistochemically recognizable FDCs in various paradigms: these included fibroblast or muscle cell lines as well as muscles of a sCJD patient with myositis (see also Table 1) [24], [113], [114], [115], [116].

Table 1. List of cultured cells capable of replicating prions. Black rectangles indicate permissiveness of cells or cell lines to replicate a particular prion strain.

301C: a mouse-adapted bovine spongiform encephalopathy prion strain; FU: Fukuoka-1, a mouse-adapted Gerstmann–Sträussler–Schinker disease prion strain; SY: a mouse-adapted sporadic Creutzfeld–Jakob disease prion strain; CJD: a patient-derived Creutzfeld–Jakob disease prion isolate; CWD: chronic wasting disease prion isolate; PrPC: cellular prion protein. See [133], [134], [135], [136], [137], [138], [139], [140], [141], [142], [143], [144], [145], [146], [147], [148], [149], [150], [151], [152], [153], [154], [155], [156], [157].

It was therefore hypothesized that extraneuronal prion replication could potentially take place in inflammatory disorders devoid of FDCs. Therefore, we studied prion replication in subcutaneous granulomas [117], a very common form of chronic inflammation, expressing PrPC and LTβR, yet lacking FDCs and Mfge8 expression. After i.p. prion inoculation, Prnp+/+ granulomas, but neither Prnp0/0 granulomas nor healthy Prnp+/+ skin, accumulated prion infectivity and PrPSc long before clinical disease. Reciprocal bone marrow transfers between Prnp+/+ and Prnp0/0 mice revealed that prion accumulation in granulomas depended on PrPC-expressing stromal cells. We also took into account that infectivity found in granulomas may represent “spill-over” from lymphoid organs such as spleen. However, granulomas of Prnp+/+→Prnp0/0 chimeric mice completely lacked any prion infectivity, although their spleens displayed prion infectivity titers similar to those of wt mice [81], [82], [83]. Homogenates of skin located in immediate vicinity of granulomas lacked prion infectivity in both Prnp+/+ and Prnp0/0 mice at subclinical stage, in agreement with a report that skin may contain prions only at late stages of disease [118]. These results identified granulomas as previously unrecognized, clinically silent reservoirs of prion infectivity.

Further, the nature of the underlying cell type enabling prion replication in granulomas was investigated by flow cytometry. We identified a CD45 negative cell type, expressing high levels of PrPC and LTβR as well as markers characterizing mesenchymal progenitor cells (MPCs) (e.g. platelet derived growth factor receptor α and β). Whether this cell type resembles a FDC related stromal cell or even a pre-FDC still remains to be determined.

Administration of LTβR-Ig drastically reduced prion infectivity of granulomas, although no significant changes in PrPC expression level or frequency of any of the cell populations investigated (e.g. lymphocytes, macrophages, MPCs) could be detected. Therefore, lack of prion replication competence in granulomas does not depend on LTβR-Ig induced depletion of particular cell populations but is rather related to the depletion of LT signaling itself (see also Fig. 1). This is corroborated by the mRNA downregulation of known LT target genes (e.g. CCL2) in LTβR-Ig treated granulomas [117].

Therefore, besides being an important prerequisite for maintenance of lymphoid microarchitecture and FDC neogenesis, LT signaling is very likely to enable peripheral prion replication competence on stromal cell types that are histogenetically different from FDCs. These data corroborate previous results that peripheral prion replication can occur in the absence of FDCs or PrPC expressing FDCs in vivo [82], [109]. Whether these results can also be extended to other prion diseased species and different prion strains is not known and should be investigated in the future.

PrpC has been demonstrated to be necessary but not sufficient for peripheral prion replication [9], indicating the ultimate requirement of additional factors. Future experiments may show whether PrPC expression in combination with LT signaling would enable prion replication competence or whether other, yet to be identified, factors are necessary. In addition, it will be of highest interest to investigate which LT target genes might be involved in allowing prion replication competence and how the accrued knowledge can be translated to understand prion replication competence of neurons.

15. Manipulating the immune system to prevent or treat prion diseases 

Proverbial sturdiness of prions, lack of a beneficial defensive response of hosts upon prion infection and limited understanding of the physiological functions of PrPC hamper the development of efficient interventional strategies against prion diseases.

On the other hand, in recent years accrued knowledge about the role of the immune system in prion pathogenesis has encouraged investigators to explore the feasibility of various interventional strategies against prion diseases [119]. Repeated administration of cytidyl guanosyl containing oligodeoxynucleotides (CpG-ODN) – mimicking unmethylated bacterial DNA and stimulating the innate immune system through Toll-like receptor 9 (TLR9) – was able to extend the survival of peripherally prion-inoculated mice. This observation suggested a potential role of CpG-ODN based regimens in post-exposure prophylaxis against prion diseases [120]. However, this finding was difficult to reconcile with the knowledge that various types of immune deficiencies confer resistance to peripherally administered prions and with the unaltered prion pathogenesis in myd88−/− mice, in which TLR9 signaling is impaired [121]. Subsequent investigations showed that repeated CpG-ODN administration results in immunosuppression and lymphoid follicle destruction, a fact that could per se well explain the alleged anti-prion properties of this regimen [122].

In 1988 Gabizon et al. found that in vitro exposure of infectious hamster brain homogenate to anti-PrP antisera resulted in the reduction of the infectivity titer [123]. Later, anti-PrP antibodies proved to efficiently inhibit the formation of protease-resistant PrP in cell-free systems [124] and to suppress prion replication in cultured cells [125], [126].

While these data constitute a rational for the development of anti-prions immunotherapy, achievement of active immunization against PrP is hampered by essential tolerance of the mammalian immune system to PrPC. In an ingenious model, transgenic expression of an immunoglobulin μ chain containing the epitope-interacting region of a high-affinity anti-PrP antibody circumvented the tolerance and resulted in build-up of anti-PrPC titers and prevention of prion pathogenesis upon i.p. prion inoculation [127]. Sigurdsson et al. succeeded to induce active immunization in wt mice with recombinant prion protein, achieving a modest therapeutic effect [128]. The observation that antibodies generated against bacterially expressed PrP often display low affinity towards native cell-surface PrPC [129] might explain the limited therapeutic results obtained with this approach. Several others studies have shown successful circumvention of immunological tolerance to PrP with development of a protective B-cell mediated response against prions, including a recent report of effective mucosal vaccination protecting against oral prion infection through oral administration of PrP expressed in an attenuated Salmonella vector [130].

An obvious and more practical alternative to vaccination aimed at achieving protective anti-prion immune response is passive immunization. Passive transfer of anti-PrP antibodies shortly after peripheral prion inoculation succeeded to delay the onset of prion disease [131]. However, no beneficial effect was seen when the antibodies were administered at onset of clinical signs, suggesting that passive immunization might be a good candidate for post-exposure prophylaxis rather than for therapy of TSEs.

Given the potential of antibodies to interfere with prion pathogenesis it was disappointing that intracranial delivery of specific anti-PrP antibodies has resulted in rapid and extensive apoptosis in hippocampal and cerebellar neurons [132]. Data in this study suggested that this occurs possibly through cross-linking of PrPC, assumed to trigger an abnormal signaling pathway [132]. These findings emphasize the need for scrupulous in vivo safety studies before the feasibility of prion immunoprophylaxis trials in humans can be considered.

The relevance of LT signaling in peripheral prion replication has indicated inhibition of this pathway as a promising strategy against prion diseases. In particular, treatment with LTβR-Ig might represent a plausible candidate for early post-exposure prophylaxis, namely iatrogenic or occupational exposure (e.g. blood transfusion, medical accidents). Interestingly, LTβR-Ig (baminercept α) proteins have already entered clinical trials as a treatment for rheumatoid arthritis and preliminary results concerning the safety of this drug are encouraging [88]. Additionally, other immunosuppressive reagents (e.g. corticosteroids) could be efficiently used to prevent peripheral prion accumulation or replication.

16. Prion diseases: future perspectives 

The last years of prion research have accumulated an enormous wealth of data widening our horizon about how prions accomplish peripheral replication, horizontal transmission and neurodegeneration. Although prion diseases represent the only class of neurodegenerative disorders in which scientists have at least a grasp on how neurotoxicity is exerted, our knowledge is not sufficient to halt or reverse the pathological process in the brain. On the positive side it is worth mentioning that we have started to understand the details of how and where prions can be replicated and horizontally transmitted. This should pave the way for preventing pandemic outbreaks of known, potentially human pathogenic prion strains (e.g. BSE) in the future. However, as prion research has indicated in the past, new prion strains are on the rise (e.g. CWD) that potentially might represent a health risk to humans. Therefore, in vivo models to identify potentially human pathogenic prion strains are urgently needed. Although the BSE crisis has been resolved and the numbers of vCJD cases are decreasing, we have little knowledge about the distribution of prion infectivity within the general population and for how long this prion infectivity can be conveyed in subclinical carriers leading to unprecedented human to human transmissions.

The above indicates that we are just at the beginning of a journey into the details of the prion universe, that will lead us to new surprises, challenges and serendipities in the future – most likely also highly useful for understanding other neurodegenerative and infectious disorders.