Review article
Amyloid and prions: some biochemical investigations of cerebral amyloidosis in mice

Introduction

Historically, a prion was defined as a “proteinaceous infectious particle” whose “main or sole component” was the prion protein (PrP) [31] and the name was original­ly essentially a synonym for the self-replicating agent which causes the progressive, transmissible spongiform encephalopathies (TSEs) of mammals. Infectious and transmissible (and contagious) can be used interchan­geably in lay language to describe the spread of a disease pathology or a disease-causing agent from one cell or organism to another, usually by direct contact, or via air or water. In science, in this context, transmissible is used generically to describe that property of movement and replication of a non-host pathogen but it can also apply to the simple propagation of disease or disease pathology in the absence of a self-replicating agent.

Prion-like transmission of protein aggregates or amyloid in several neurodegenerative diseases, such as Par­kinson’s disease, Huntington’s disease and Alzheimer’s disease, in addition to the transmissible spongiform encephalopathies (or prion diseases), has been proposed recently [9,12,15,16]. Induced diseases such as experimental encephalomyelitis [26] or systemic amyloidosis [39] are examples of transmissible but non-infectious diseases, and they are not normally considered as prion diseases. But can they be regarded as “prion-like”? Perhaps, but it is not informative to lump these diseases, as well as AD, HD and PD, together with TSEs as “prion-like”: we need a more stringent description than the histo­rical definition of “prion” to guide our understanding. This better definition was provided by Reed Wickner in 1996 [36], and subsequently refined further [37].

The original definition of a yeast prion highlighted the following properties:

• reversible curability;

• over-producing the (prion) protein increases the frequency of prion generation;

• the phenotype of (a) mutant (prion protein) resembles the phenotype of the prion; and it required

the biochemical proof of a prion by demonstrating infectivity from an altered version of a synthetic or recombinant form.

These properties of a “prion” were distilled from the genetics of yeast but, apart from “reversible curability” which seems a long way away in terms of mammalian prion disorders, the revised concept of a prion is very similar and has four tenets:

• a prion is a chromosomally-encoded protein that exists as a conformational isomer of one of two covalently identical forms;

• the prion form can induce the other (normal or cellular) form to a new copy of the prion form;

• in both mammals and yeast, over-producing the protein will increase the frequency of prion generation; and

• that the effects of the prion form can be reproduced in vivo by a suitably re-folded recombinant or synthe­tic copy of the protein.

Nowhere in this definition is there a need for the prion isoform to aggregate for its function, or for those aggregates to have a distinct morphology or biochemistry such as an amyloid fibril or resistance to proteases, respectively. Several naturally-occurring prion diseases, such as bovine spongiform encephalopathy and classical scrapie in sheep and their laboratory rodent models, are associated with protease-resistant prion protein (PrP) which can be visualised by light microscopy as amyloid plaques and by electron microscopy of tissue extracts or tissue sections as part of a fibrillar structure [19,20]. However, there are model systems of prion diseases which produce high titres of infectivity in the presence of little or no protease-resistant prion protein [1], and the abnormal prions of natural cases of atypical scrapie in sheep have a very different biochemistry to those causing classical scrapie and BSE [17]. Baskakov and Breydo considered what might make a protein infectious [2] and they and others finally provided evidence meeting Wick­ner’s fourth criterion in the definition of a prion al­most 30 years after Prusiner’s initial hypothesis (for a review see [27]), and protease-resistance is not a necessary feature of synthetic prions [11].

Prion diseases differ from other amyloid-associated protein misfolding diseases (e.g. Alzheimer’s) because they are naturally transmitted between individuals and involve spread of protein conformational changes between tissues. Factors underlying these features of prion diseases are poorly understood although, of all protein misfolding disorders, only prion diseas­es involve the misfolding of a glycosyl-phosphatidyl-inositol (GPI)-anchored protein and GPI anchoring has been proposed as the key factor which modulates the propagation and spread of prion aggregates [25]. Using a GPI-anchored version of the amyloidogenic yeast protein Sup35NM (Sup35(GPI)) ex­pressed in neuronal cells, Speare et al. found treatment of cells with Sup35NM fibrils induced the GPI anchor-dependent formation of self-propagating, detergent-insoluble, protease-resis­tant, prion-like aggregates of Sup35(GPI) [34]. While it can be inferred from these model studies that a GPI-an­­chor may enhance the transmissibility and pathoge­nesis of prion diseases relative to other protein misfold­ing diseases, apparently contrary evidence showing the lack of a GPI-anchor on the mammalian prion protein does not prevent amyloid formation and amplification of the level of in­fectious particles was found in a Tg PrP GPI-minus mouse challenged intracerebrally with various strains of murine-passaged scrapie [10].

These studies, with apparently contradictory indications, focus on the amyloid or aggregate nature of some prions although as we have previously noted “amyloid” is not part of the definition of a prion so it is also relevant here to look at the definition of “amyloid”. The Wiktionary [21] defines amyloid as i) a waxy compound of protein and polysaccharides that is found deposited in tissues in amyloidosis; and ii) any of various starch-like substances. The name is derived from the Latin for starch (amylum), because of its compar­able iodine-binding properties, and pre-dates the realisation that this material is protein. Various histological stains have been used to provide an operational definition of amyloids and, most widely, its binding to the polyaromatic dye Congo Red and characteristic light-green birefringence when visualised under polarized light has been diagnostic. Many neurological diseases have amyloid deposits in the brain as their distinguishing feature and years before the characterisation of the mammalian prion protein by Prusiner et al., histologi­cal staining for amyloid plaques and vascular amyloid was a technique for the morphometric classification of different types or strains of experimental prion disease in mice [13,38]. More recently, a broader biophysical definition of amyloid has come to include any polypeptide which polymerizes to form a cross- structure (for a re­view of the structure and nomenclature of amyloids see [33]), in vivo, or in vitro.

Our interpretation of the current controversy about the prion-like nature of AD, HD and PD, and the non-TSE-like nature of some human and animal prion protein diseases [23,30], is that it could be clarified by a better operational definition of the prion form of the relevant proteins so that the definition covers both amyloid and non-amyloid components of prions and does not focus on “protease resistance” as its main operational criterion. As a first step towards that goal, we report here the detergent extractability and epitope occlusion of murine prion protein in the Fraser and Bruce experimental models of mouse prion cerebral amyloidosis.

Material and methods

All materials and methods used to produce the electron micrographs of murine amyloid in Fig. 1 are provided by Jeffrey et al. [23] and those used to determine the data presented for mouse survival time and amyloid score in Table I are to be found in the references in the Table footnotes. The materials and methods outlined below refer to those used to produce the data for Fig. 2 and the A’ score of Table I.

Animals and prion strains

Frozen brain from terminally-affected mice (VM/Dk or SV/Dk strains) infected intracerebrally with a 20 L dose of 10-2 (w/v)-inoculum of various prion strains (22L, 87V, 22A, ME7, 79A, 301V) were kindly supplied to us by Ms Angela Chong and Dr Robert Somerville, BBSRC Institute for Animal Health, Neuropathogenesis Unit, Edinburgh. 87V, 22A and 301V prions were produced by serial passage in Prn-pb mice and ME7, 79A and 22L prions were produced by serial passage in Prn-pa mice. Under these passage conditions, their characteristics of lesion profile and relative incubation period are stable and this encouraged us to compare our bioche­mical data with historical data in the literature on their amyloid content [5-8,14,28].

Biochemical procedures

Brain homogenisation and use of various detergents

For each prion strain/mouse strain combination, six whole mouse brains were processed and analysed individually as follows: brain (0.4 g) was homogenised at 20% (w/v) in 1.4M-Guanidinium chloride, 10mM-Tris-HCl, pH 7.4 at 4°C and then mixed with an equal volume of 1% (w/v) ZW3-14 detergent and allowed to stand for 15 minutes. The homogenate was spun at 8000 γ for 15 minutes, and pellet and supernatant separat­ed and sonicated in either 6M-GdnHCL, 1mM-Dithiothreitol (DTT), 0.5%-ZW3-14, pH 7.4 (pellet) or mixed with 3 volumes of 8M-GdnHCL and adjusted to 1 mM-Dithiothreitol (DTT), 0.5%-ZW3-14, pH 7.4 (superna­tant). The denatured pellet (P) or supernatant (S) was then diluted 1 : 50 into time-resolved fluoroimmu­noassay buffer for measurement of total PrP content. Values were averaged for all brains of a particular strain/mouse line model (n = 6) and the standard error of the mean was ± 15%. The efficiency of extraction and denaturation by other detergents (Triton-X100, sodium dodecyl sulphate, octyl glucoside and NP-40), and chaotropic agents (guanidinium with different counter anions; and urea) on the assay signal was also evaluated (data not presented; but see [18]).

Time-resolved fluoroimmunoassay

The time-resolved fluoroimmunoassay for abnormal prion protein has been described previously in principle [3,32] and used a standard capture antibody: detection antibody sandwich format. For this investigation, we absorbed monoclonal antibody (mAb) FH11 (produced at the Institute for Animal Health, Compton and provided by Dr Chris Birkett and Mrs Ruth Hennion) to NUNC low fluorescence Maxisorp microtitre plates by overnight incubation with 0.2 mL antibody (~ 1 mg/ mL in phosphate-buffered saline, pH7.4), aspiration, air drying and storage prior to use at 4°C. The mAb was raised against recombinant bovine PrP and cross reacts with the bovine sequence G49NRYPPQGGGGWG and to a les­s­­­­­er extent to the mouse sequence G89QGGGTHNQWNK (by Pepscan analysis, Birkett, unpublished). For detection mAb, we used mAb 6H4 purchased from Prionics SG, Zurich, Switzerland; this mAb has a wide species range (including mouse) and the epitope is in the C-terminal region of PrP (residues 145-160; Prionics SG, Zurich). This reagent was labelled with N-1(p-isothio­cyanatophenyl)-diethylene-triamine N1, N2, N3-tetraacetate chelated to europium and the conjugate characterised according to the manufacturer’s protocol (Perkin-Elmer, Beaconsfield, UK; Eu-labelling kit, Cat. No. 1244-302). Typically, stock solutions were prepared and stored at 20 g/mL immunoglobulin (11 Eu2+/Ig) in Tris-HCl buffer (pH 7.8) containing sodium azide (0.1%), sodium chloride (0.9%) and bovine serum albumin (0.1%). Before use, the stock was dilut­ed 1 : 100 v/v in Tween-40 assay buffer (Perkin-Elmer; Cat. No. 1244-106) and filtered through a 0.22 m filter. The final concentration of detecting antibody was about 200 ng/mL and 200 L (40 ng) was used per microtitre plate well.

Recombinant mouse prion protein (recPrP, sequence equivalent to the Prn-pa allotype [35]) was used to calibrate the fluorescence signal in terms of a concentration of prion protein; the protein was produced in E. coli using a dual-origin vector based on pMG165, purified and supplied to us by Dr Alan Bennett, IAH, Compton). Standard curves were produced using recPrP in concentrations of guanidinium chloride and detergent equivalent to those in the samples; equivalent dose-response curves were produced in this assay format using the mouse Prn-pb allotype (data not presented).

In brief, the assay protocol was as follows: dry coated microtitre plates were washed in assay buffer (Perkin-Elmer, Cat. No. 1244-106) containing 0.5%-Tween 40. Standard or sample (200 L) in Tween 40-assay buff­er was added to the wells and the plate incubated with shaking for 1h at room temperature. Subsequently, plates were washed 3 using Perkin-Elmer plate wash­er and tapped dry. Eu-labelled mAb (200 L) was added to each well and the plates incubated 1h at room temperature with shaking. The plates were then washed 6 and dried. Enhancement solution (200 L) was added to each well, incubated for 5 minutes and then the signal measured using the Perkin Elmer Victor II Plate fluorimeter. The concentration of recPrP equivalents in the detergent-extracted-brain pellet ([PrP-P]) and con­centration of recPrP equivalents in the detergent-extracted-brain supernatant ([PrP-S]) were then determined from the recPrP standard curve.

Results & Discussion

The mammalian prion proteins (PrP) of transm­issi­ble spongiform encephalopathies (TSE) are GPI-linked plasma membrane polypeptides of ~230 amino acids. More than half the molecule is tightly folded in a three -helices structure linked by short stretches of -sheet and the rest, containing a metal binding His repetitive structure, adopts a flexible random-coil conformation in solution. In the wild-type mouse, there are two alleles of the prion protein gene, Prn-pa and Prn-pb, and they encode proteins differing in two amino-acids, at codons 108 and 189; Prn-pa mice encode the allotype 108L/189T and Prn-pb encodes the allotype 108F/189V [35].

Most of the data in the literature which we have mined as part of our meta-analysis of mouse prion amyloid pre-dates the discovery of the prion protein and it is summarized in Table I, with our more recent biochemical investigations of the detergent solubility and extractability of abnormal prion protein from the brains of terminally-affected, scrapie-infected mice.

Fraser and Bruce used various strains of wild-type mice for their studies, and defined their strains of prions in terms of histological lesion profiles at set levels of the brains and the relative ranking of survival time of a panel of mice differing at their Sinc (scrapie incubation time) locus [4]. Our biochemical investigations used Sinc congenic mice: VM/Dk (Sincp7) and SV/Dk (Sincs7) [produced by back-crossing the Sincs7 allele of C57Bl/6 mice onto the VM/Dk background] [22].

In their studies on the aetiology of scrapie amyloid in mice, Fraser and Bruce relied on conventional dye stains such as Masson’s trichrome or haemato­xylin/eosin and their metric for amyloidosis was the number of amyloid structures per section visualized by light microscopy and the percentage of mice affected. They painstakingly evaluated the experimental parameters which affected their amyloid score and concluded that Sinc genotype and prion strain were the predominating factors (see references in Table I), although they also uncovered in their use of non-Sinc congenic mouse lines a contribution from other genes to their amyloid score [6]. Sinc congenic lines were later de­ve­loped by Dickinson and colleagues, and Prn-p shown to be encoded within the Sinc locus [22]. The Sincs7 mouse lines used by Fraser and Bruce are now known to encode the Prn-pa allotype, and their VM/Dk Sincp7 line encodes the Prn-pb allotype [29].

During TSE pathogenesis in Prn-p congenic mice, neural cell membrane structures contort, glial cells proliferate and at the end stages of disease there is widespread deposition of conformationally-altered aggregates of PrP and other cell membrane debris in the brain. Figure 1 shows labelling of cerebral amyloid plaques in the brain of a Prn-pb mouse terminally-affected by the 87V prion strain using gold-particle-tagged mAb P2 and electron microscopy; mAb P2 is specific of residues 60 to 80 of PrP [23]. Various pathological parameters such as vacuolation, degree of gliosis, amyloid deposition or distribution and type of PrPD im­mu­nostaining have been used to categorise prions and compliment their biological properties such as the relative survival time of mice differing in their Prn-p (Sinc/Prn-i) allelism.

In this study, we have expanded the categorisation of different types of prion in Prp-p congenic mice to include the ease with which detergents extract and solubilise abnormal prion protein (PrPD) from brain membranes and the amyloidic character of that PrPD as determined by guanidinium chloride treatment and a version of Safar’s conformation-dependent im­mu­noassay (CDI) [32]. Figure 2 plots the PrP signal after denaturation of an “operationally-insoluble” or “pellet” fraction of mouse brain infected with different prion strains ([PrP-P]) against the ratio of that signal and the PrP signal after denaturation of an “operationally-soluble” or “supernatant” fraction from the same brain extract ([PrP-P]/[PrP-S]). We found that the ratio of the amounts of PrP in a detergent soluble form to its amyloidic or “insoluble” form (the amyloid co-efficient, A’) varied with prion type or strain but was mostly independent of the Prn-p genotype of the congenic mouse host.

For a single prion strain, the ranking of its incubation periods in Prn-pa and Prn-pb mice, and the corresponding A’, correlated with [PrP-D]; similarly, Fraser and Bruce also reported their amyloid score correlated with the incubation period rather than the age at inoculation of mice for individual strains [8, Table I]. We found an exception in the example of the 22A prion where its [PrP-P] in Prn-pa mice differed considerably from that in Prn-pb and both this parameter, and A’, were inversely correlated to their respective incubation periods. This disparity in biochemical properties of 22A in these congenic mice mimics the disparity of the histochemical characteristics of 22A in Prn-pa (where it is re-categorised as 22F) and Prn-pb mice [4] and adds weight to the use of A’ as an independent biochemical parameter for categorising prion types.

Acknowledgements

We would like to thank all of our past colleagues at the BBSRC Institute for Animal Health, Compton and Edinburgh for their generosity of spirit and time, particularly Drs Bennett, Birkett, Somerville and Ms Angie Chong for their kind gifts of antisera, recombinant protein and scrapie-infected mouse brain tissue. This work benefited from BBSRC Core funding to the Institute for Animal Health, and Defra research grant SE1739.

References

 1. Barron RM, Campbell SL, King D, Bellon A, Chapman KE, Williamson RA, Manson JC. High titers of transmissible spongiform encephalopathy infectivity associated with extremely low levels of PrPSc in vivo. J Biol Chem 2007; 282: 35878-35886.

 2. Baskakov IV, Breydo L. Converting the prion protein: What makes the protein infectious. Biochim Biophys Acta 2007; 1772: 692-703.

 3. Bellon A, Seyfert-Brandt W, Lang W, Baron H, Gröner A, Vey M. Improved conformation-dependent immunoassay: suitability for human prion detection with enhanced sensitivity. J Gen Virol 2003; 84: 1921-1925.

 4. Bruce M, Fraser H, McBride P, Scott J, Dickinson A. The basis of strain variation in scrapie. In: Prusiner SB, Collinge J, Powell J, Anderton B (eds.). Prion Diseases of Humans and Animals, Ellis Horwood, New York 1992; pp. 497–508.

 5. Bruce ME. Serial studies on the development of cerebral amyloidosis and vacuolar degeneration in murine scrapie. J Comp Pathol 1981; 91: 589-597.

 6. Bruce ME, Dickinson AG. Genetic control of amyloid plaque production and incubation period in scrapie-infected mice. J Neuropathol Exp Neurol 1985; 44: 285-294.

 7. Bruce ME, Fraser H. Effect of route of infection on the frequency and distribution of cerebral amyloid plaques in scrapie mice. Neuropathol Appl Neurobiol 1981; 7: 289-298.

 8. Bruce ME, Fraser H. Effects of age on cerebral amyloid plaques in murine scrapie. Neuropathol Appl Neurobiol 1982; 8: 71-74.

 9. Brundin P, Melki R, Kopito R. Prion-like transmission of protein aggregates in neurodegenerative diseases. Nat Rev Mol Cell Biol 2010; 11: 301-307.

10. Chesebro B, Trifilo M, Race R, Meade-White K, Teng C, LaCasse R, Raymond L, Favara C, Baron G, Priola S., Caughey B, Masliah E, Oldstone M. Anchorless prion protein results in infectious amyloid disease without clinical scrapie. Science 2005; 308: 1435-1439.

11. Colby DW, Wain R, Baskakov IV, Legname G, Palmer CG, Ngu­yen HO, Lemus A, Cohen FE, DeArmond SJ, Prusiner SB. Protease-sensitive synthetic prions. PLoS Pathog 2010; 6: e1000736.

12. Cushman M, Johnson BS, King OD, Gitler AD, Shorter J. Prion-like disorders: blurring the divide between transmissibility and infectivity. J Cell Sci 2010; 123: 1191-1201.

13. Fraser H, Bruce M. Argyrophilic plaques in mice inoculated with scrapie from particular sources. Lancet 1973; 1: 617-618.

14. Fraser H, Bruce ME. Experimental control of cerebral amyloid in scrapie in mice. Prog Brain Res 1983; 59: 281-290.

15. Frost B, Diamond MI. The expanding realm of prion phenomena in neurodegenerative disease. Prion 2009; 3: 74-77.

16. Frost B, Diamond MI. Prion-like mechanisms in neurodegenerative diseases. Nat Rev Neurosci 2010; 11: 155-159.

17. Griffiths PC, Spiropoulos J, Lockey R, Tout AC, Jayasena D, Pla­ter JM, Chave A, Green RB, Simonini S, Thorne L, Dexter I, Balkema-Busch­mann A, Groschup MH, Berinque V, Le Dur A, Laude H, Hope J.

Characterization of atypical scrapie cases from Great Britain in transgenic ovine PrP mice. J Gen Virol 2010; 91: 2132-2138.

18. Hope J. The development of an ultra-sensitive time-resolved fluroimmunoassay for the prion protein (PrP). Department of Environment, Food and Rural Affairs. Final Report, SE1739, 1999.

19. Hope J, Morton LJD, Farquhar CF, Multhaup G, Beyreuther K,

Kimberlin RH. The major polypeptide of scrapie-associated fibrils (SAF) has the same size, charge distribution and N-terminal protein sequence as predicted for the normal brain protein (PrP). EMBO J 1986; 5: 2591-2597.

20. Hope J, Reekie LJ, Hunter N, Multhaup G, Beyreuther K, White H, Scott AC, Stack MJ, Dawson M, Wells GA. Fibrils from brains of cows with new cattle disease contain scrapie-associated protein. Nature 1988; 336: 390-392.

21. http://en.wiktionary.org/wiki/amyloid; accessed on 14 December 2011.

22. Hunter N, Dann JC, Bennett AD, Somerville RA, McConnell I, Hope J. Are Sinc and the PrP Gene Congruent? Evidence from PrP Gene Analysis in Sinc Congenic Mice. J Gen Virol 1992; 73: 2751-2755.

23. Jeffrey M, Goodsir CM, Bruce ME, McBride PA, Farquhar C. Morphogenesis of amyloid plaques in 87V murine scrapie. Neuropathol Appl Neurobiol 1994; 20: 535-542.

24. Jeffrey M, McGovern G, Chambers EV, King D, Gonzalez L, Manson JC, Ghetti B, Piccardo P, Barron RM. Mechanism of PrP-amyloid formation in mice without transmissible spongiform encephalopathy. Brain Pathol 2012; 22: 58-66.

25. Klingeborn M, Race B, Meade-White KD, Rosenke R, Striebel JF, Chesebro B. Crucial role for prion protein membrane anchoring in the neuroinvasion and neural spread of prion infection. J Virol 2011; 85: 1484-1494.

26. Mackay IR, Anderson WH. What’s in a name? Experimental encephalomyelitis: ‘allergic’ or ‘autoimmune’. J Neuroimmunol 2010; 223: 1-4.

27. Makarava N, Kovacs GG, Savtchenko R, Alexeeva I, Budka H, Rohwer RG, Baskakov IV. Genesis of mammalian prions: from non-infectious amyloid fibrils to a transmissible prion disease. PLoS Pathog 2011; 7: e1002419.

28. McBride PA, Bruce ME, Fraser H. Immunostaining of scrapie cerebral amyloid plaques with antisera raised to scrapie-associated fibrils (SAF). Neuropathol Appl Neurobiol 1988; 14: 325-336.

29. Moore RC, Hope J, McBride PA, McConnell I, Selfridge J, Mel­ton DW, Manson JC. Mice with gene targetted prion protein alterations show that Prnp, Sinc and Prni are congruent. Nat Genet 1998; 18: 118-125.

30. Piccardo P, Manson JC, King D, Ghetti B, Barron RM. Accumulation of prion protein in the brain that is not associated with transmissible disease. Proc Natl Acad Sci U S Am 2007; 104: 4712-4717.

31. Prusiner SB. Novel proteinaceous infectious particles cause scrapie. Science 1982; 216: 136-144.

32. Safar J, Wille H, Itrri V, Groth D, Serban H, Torchia M, Cohen FE, Prusiner SB. Eight prion strains have PrPSc molecules with different conformations. Nat Med 1998; 4: 1157-1165.

33. Sipe J, Benson M, Buxbaum J, Ikeda S-I, Merlini G, Saraiva M, Westermark P. Amyloid fibril protein nomenclature: 2010 Recommendations from the nomenclature committee of the International society of amyloidosis. Amyloid 2010; 17: 101-104.

34. Speare JO, Offerdahl DK, Hasenkrug A, Carmody AB, Baron GS. GPI anchoring facilitates propagation and spread of misfolded Sup35 aggregates in mammalian cells. EMBO J 2010; 29: 782-794.

35. Westaway D, Goodman PA, Mirenda CA, McKinley MP, Carlson GA, Prusiner SB. Distinct prion proteins in short and long scrapie incubation period mice. Cell 1987; 51: 651-662.

36. Wickner RB. Prions and RNA viruses of Saccharomyces cerevisiae. Ann Rev Genet 1996; 30: 109-139.

37. Wickner RB, Edskes HK, Ross ED, Pierce MM, Baxa U, Brachmann A, Shewmaker F. Prion genetics: new rules for a new kind of gene. Annu Rev Genet 2004; 38: 681-707.

38. Wisniewski HM, Bruce ME, Fraser H. Infectious etiology of neuritic (senile) plaques in mice. Science 1975; 190: 1108-1110.

39. Zhang B, Une Y, Fu X, Yan J, Ge F, Yao J, Sawashita J, Mori M, Tomozawa H, Kametani F. Higuchi K. Fecal transmission of AA amyloidosis in the cheetah contributes to high incidence of disease. Proc Natl Acad Sci U S A 2008; 105: 7263-7268.