JC is a Director and JC and JDFW are shareholders and consultants of D-Gen Limited, an academic spin-out company working in the field of prion disease diagnosis, decontamination, and therapeutics. D-Gen markets the ICSM35 antibody used in this study. SB carries out a large-scale screening study for the detection of abnormal prion protein in archival appendix samples. This study was commissioned in a competitive tender by the UK Health Protection Agency. The other authors declare no conflicts of interest.
Experimental confirmation that variant Creutzfeldt–Jakob disease (vCJD) 1 is caused by the same prion strain as that causing bovine spongiform encephalopathy (BSE) in cattle 2–5 led to global concern that human exposure to BSE prions posed a significant threat to public health 6–8. While the risk of new dietary exposure to BSE prions in the UK is now remote 9, the majority of the UK population may have been exposed during the late 1980s and early 1990s. While the number of recorded clinical cases (∼200) has been relatively low, and epidemiological modelling argued against a large number of additional cases 10, it is increasingly clear that sub-clinical carrier states of prion infection may occur, particularly on crossing a species barrier and with low-dose exposure 11–15. Modelling cannot estimate the number of sub-clinically infected carriers, and the discrepancy between numbers of clinical cases and prevalence estimates from population screening remains unexplained 16–19. Incubation periods in human prion infections, even in the absence of a species barrier, may exceed five decades 6, 20, 21, and multiple genetic loci in addition to the PrP gene are known to influence prion incubation periods in mice 22–25 and susceptibility in humans 26, 27. Asymptomatically infected carriers appear to have transmitted vCJD prion infection and disease to others via blood transfusion 28–30 or blood products 31. Other iatrogenic routes, notably from contaminated surgical instruments, present a potential risk 6, 8, 32.
In the absence of a blood test for human prion disease, the distinct phenotype of vCJD has provided the basis for investigating infection prevalence within the UK population. The peripheral pathogenesis of vCJD differs significantly from that of classical CJD, inherited prion disease, and kuru, with disease-related PrP (PrPSc) being uniformly and prominently detected in a range of lymphoreticular tissues at autopsy 33–37. To date, antemortem tonsil biopsy has shown 100% sensitivity and specificity for diagnosis of clinical vCJD 30, 33, 34, and the fact that lymphoreticular prion infection is not a feature of iatrogenic CJD 33, 36 or kuru 21, 38 argues that the distinct pathogenesis of vCJD relates to the effect of prion strain rather than to a peripheral route of infection 21, 38, 39. Because lymphoreticular colonization is thought to precede neuroinvasion in vCJD, and indeed has been detected in archived surgical samples removed prior to the development of overt clinical symptoms 40, 41, anonymous screening of lymphoreticular tissues removed during routine surgery appears to offer a practical means to estimate the population prevalence of asymptomatic vCJD infection. However, the sensitivity and specificity of such testing during the pre-clinical incubation period or in chronic carriers (who may have a distinct pathogenesis which conceivably might not involve the same degree of lymphoreticular colonization) are unknown. Although peripheral tissue involvement in recently identified cases of variable protease-sensitive prionopathy 42–45 has not yet been defined, the apparent rarity of this condition would not be expected to confound estimates of UK vCJD prevalence.
To date, three prevalence studies have been reported 41, 46, 47. The first study used immunohistochemistry to retrospectively examine 11 247 appendicectomy and 1427 tonsillectomy specimens that had been archived as formalin-fixed tissue blocks 41. Positive lymphoreticular accumulation of abnormal PrP was found in three appendixes from 10 278 specimens in a patient birth cohort of 1961–1985, giving a population prevalence in this age group of 292 (95% confidence interval 60–853) per million 41, 47. Subsequent studies have prospectively examined only tonsil as this tissue can be readily obtained from large numbers of routine tonsillectomy procedures. Tonsil appeared to be a more sensitive reporter of clinical vCJD infection than appendix in one study 48, although not in another 36. The first prospective pilot scale study of 2000 specimens found no evidence for tonsillar accumulation of abnormal PrP using high sensitivity methods 46. More recently, a much larger national study of 63 007 tonsils reported no positive specimens using enzyme immunoassays to detect disease-related PrP 47. To confirm the reliability of this result, all specimens in the 1961–1985 birth cohort (9160 tonsils) plus controls were tested by immunohistochemistry 19, 49. Three specimens showed some initial reactivity, but for two this was concluded to be non-specific. For the third, one strongly positive follicle was identified on two slides from adjacent sections using different anti-PrP antibodies. However, subsequent re-sampling of this specimen revealed no further reactivity by either immunohistochemistry or immunoblotting 19, 49. Notably, an incidence of either zero or one positive in the 1961–1985 birth cohort in these large-scale studies of tonsil does not significantly alter the infection prevalence estimated by the original study of appendix specimens 19, 47, 49.
Inevitably, the failure to identify an unequivocally positive sample in more recent prospective studies has meant that the positive appendix specimens identified by Hilton et al have been the subject of considerable debate 17–19, 41. Central to this is whether the identified immuno-reactivity truly reflects authentic prion infection. Interpretation has been hampered by the inability to do confirmatory immunoblotting of formalin-fixed tissue and by the fact that two of the three positive specimens that could be analysed were from individuals with a PRNP codon 129 valine homozygous genotype 50. Polymorphism at residue 129 of human PrP [encoding either methionine (M) or valine (V)] powerfully affects susceptibility to human prion diseases, with residue 129 acting to restrict the propagation of particular prion strains through conformational selection 6, 51, 52 and heterozygosity conferring resistance by inhibiting homologous protein–protein interactions 52–54. To date, all patients with neuropathologically confirmed vCJD have been PRNP codon 129 methionine homozygotes 7, 26 and have a remarkably uniform and distinct neuropathological phenotype defined by the presence of abundant florid PrP plaques 1 and the propagation of type 4 PrPSc2, 55 in the brain.
Transgenic modelling has provided a molecular explanation for these findings by showing that human PrP 129 valine is unable to propagate the vCJD prion strain 51 and that BSE- or vCJD-challenged human PrP 129V transgenic mice propagate novel prion strains that have not yet been identified in humans 3, 39, 51. It is unclear whether humans of the PRNP 129VV genotype following infection with BSE prions would develop a clinical disease and if so, what clinicopathological phenotype would result 7, 8, 51. Thus, in the absence of any natural history of BSE prion infection in humans with a PRNP codon 129VV genotype, no positive control tissue exists for comparison with the appendix specimens identified in the prevalence screen. A clinical vCJD case in a PRNP codon 129 MV heterozygote has been reported, but no tissue biopsy or autopsy was performed 56. In the present study, we have therefore sought to investigate these specimens further by attempting to transmit prion infection from formalin-fixed tissue to transgenic mice overexpressing human PrP. Owing to the paucity of tissue available, it was not possible to inoculate different lines of transgenic mice expressing either human PrP 129 methionine or valine. Based on our experience of the transmission properties of vCJD and BSE prions, we chose to use Tg45 mice homozygous for 129 methionine 5. These mice faithfully propagate the molecular and neuropathological phenotype of vCJD following primary challenge with either vCJD prions 5, 57 or BSE prions 5 and are known to be sensitive to infection when inoculated with vCJD peripheral tissue containing PrPSc at a concentration 104.7-fold lower than in brain 57. Importantly, the novel prion strain (associated with type 5 PrPSc) that is generated in vCJD-challenged human PrP 129V transgenic mice also transmits infection efficiently to transgenic mice expressing human PrP 129M 51.
Materials and methods
Storage and biochemical analysis of human tissue samples and transmission studies to mice were performed with consent from patients or relatives under approval from the Local Research Ethics Committee of UCL Institute of Neurology/National Hospital for Neurology and Neurosurgery and the code of practice specified in the Human Tissue Authority licence held by UCL Institute of Neurology. Work with mice was performed under licence granted by the UK Home Office and conformed to University College London institutional guidelines. All procedures were carried out in microbiological containment level 3 facilities with strict adherence to safety protocols.
Preparation of inocula
Frozen brain (frontal cortex) and appendix (full thickness tissue) from a single neuropathologically confirmed vCJD patient 48 were prepared as 10% w/v homogenates in Dulbecco's sterile phosphate-buffered saline lacking Ca2+ and Mg2+ ions (D-PBS) by either serial passage through needles of decreasing diameter (brain) or the use of a Duall tissue grinder (appendix). PrPSc-positive 10% w/v homogenates 48 were diluted to 1% w/v with D-PBS and passed through a 25-gauge syringe needle prior to storage at −70 °C. Formalin-fixed, paraffin-embedded (FFPE) tissues that had not been exposed to formic acid were each available as three tissue sections each of approximate dimensions 4 µm × 0.5 cm × 1 cm mounted on glass slides. FFPE tissues for investigation comprised vCJD brain and vCJD spleen and two of the three abnormal PrP-positive appendix specimens (both PRNP codon 129 valine homozygous genotype) identified in the prevalence screen by Hilton et al41. Tissue was recovered by sequential treatment, firstly to remove paraffin by immersion in two changes of xylene for 5 min, followed by sequential immersion for 5 min in graded ethanol (100% × 2 and 70% × 1), after which tissue was re-hydrated by immersion in D-PBS for 5 min. Each of the cognate tissue sections was scraped from the glass slides and pooled in 300 µl of D-PBS and passaged through 16-, 19-, 21-, and 23-gauge syringe needles prior to storage at −70 °C. We estimate that this procedure produced tissue homogenates of a concentration of ∼0.2% w/v.
Transgenic mice homozygous for a human PrP 129 methionine (M) transgene array and murine PrP null alleles (Prnpo/o) designated Tg(HuPrP129M+/+Prnpo/o)-45 (129 MM Tg45 mice) have been described previously 5, 57, 58. Inocula (30 µl of tissue homogenate in D-PBS as described above) were injected intra-cerebrally into groups of ten 129 MM Tg45 mice as described elsewhere 5 and thereafter examined daily and were killed if they exhibited signs of distress or once a clinical diagnosis of prion disease was established. Brains from inoculated mice were analysed by high sensitivity immunoblotting and by neuropathological examination.
Transgenic mouse brain was analysed with anti-PrP monoclonal antibody ICSM 35 (D-Gen Ltd, London, UK), using a Ventana automated immunohistochemical staining machine (Ventana Medical Systems Inc, Tucson, AZ, USA) as described previously 38, 59. Briefly, tissue was fixed in 10% buffered formal saline followed by incubation in 98% formic acid for 1 h. Following further washing for 24 h in 10% buffered formal saline, tissue samples were processed and paraffin wax-embedded. Sections were cut at a nominal thickness of 4 µm, treated with 98% formic acid for 5 min, and then boiled in a low ionic strength buffer (2.1 mM Tris, 1.3 mM EDTA, 1.1 mM sodium citrate, pH 7.8) for 20 min. Abnormal PrP accumulation was examined using anti-PrP monoclonal ICSM 35 followed by a biotinylated anti-mouse IgG secondary antibody (iView Biotinylated Ig, Ventana Medical Systems Inc) and an avidin–biotin horseradish peroxidase conjugate (iView SA-HRP, Ventana Medical Systems Inc) before development with 3′,3-diaminobenzidine tetrachloride as the chromogen (iView DAB, Ventana Medical Systems Inc). Haematoxylin was used as the counterstain. Haematoxylin and eosin staining of serial sections was performed using conventional methods 59. Appropriate controls were used throughout.
Human or transgenic mouse tissues prepared as 10% w/v homogenates in D-PBS were analysed by proteinase K digestion (50 or 100 µg/ml final protease concentration, 1 h, 37 °C) and immunoblotting with anti-PrP monoclonal antibody 3F4 60 using high sensitivity enhanced chemiluminescence as described previously 34, 59. For analysis of PrP glycoforms, blots were developed in chemifluorescent substrate (AttoPhos; Promega, Madison, WI, USA) and visualized on a Storm 840 phosphoimager (GE Healthcare, Little Chalfont, UK) using ImageQuaNT software (GE Healthcare) 34, 61. Sodium phosphotungstic acid precipitation of PrPSc from tissue homogenate 62 was performed as described previously 34, 59.
Previously we have reported the comparison of PrPSc in brain and appendix from the same vCJD patient 48. PrPSc in appendix was at a concentration ∼0.5% of that present in brain 48 and was distinct from type 4 PrPSc in brain, being characterized by a greater predominance of di-glycosylated PrP (Figure 1A) closely similar to that seen in tonsil and other lymphoreticular tissues in vCJD 33, 34, 36, 37, 57. These PrPSc-positive tissue homogenates were used as positive controls for transmission studies in comparison to FFPE tissue specimens. Consistent with our previous findings 5, 57, intra-cerebral challenge of 129 MM Tg45 transgenic mice with vCJD brain homogenate resulted in efficient transmission of prion infection, with nine of ten inoculated mice scored as affected (Table 1). Four mice developed clinical prion disease with a mean incubation period of 396 ± 22 days, while five other recipients were sub-clinically infected (Table 1). The single mouse scored as non-affected was culled, due to inter-current illness at 280 days post-inoculation. All nine infected mice had detectable PrPSc in brain that showed a predominance of di-glycosylated PrP, and accurate molecular strain typing was possible with four samples, which confirmed the propagation of type 4 PrPSc (Figure 1B). In two of these brains, the presence of abundant florid PrP plaques was demonstrated by immunohistochemistry (Figure 2).
Table 1. Summary of transmissions of human tissues to 129 MM Tg45 transgenic mice
Mice were inoculated with 30 μl of either 1% w/v vCJD brain or vCJD appendix homogenate, or ∼0.2% w/v formalin-fixed, paraffin-embedded (FFPE) tissue homogenate.
Total attack rate is defined as the total number of both clinically affected and sub-clinically infected mice as a proportion of the number of inoculated mice. Sub-clinical prion infection was assessed by analysis for PrPSc (by either direct immunoblotting or after sodium phosphotungstic acid precipitation of 250 μl of 10% w/v brain homogenate) and also by immunohistochemical examination of brain.
Mice were culled at 360, 442, 442, 610, 610, 610, 610, and 610 days post-inoculation.
Mice were culled at 406, 451, 561, 610, 610, 610, 610, and 610 days post-inoculation.
In comparison, and consistent with a lower concentration of PrPSc in the inoculum, vCJD appendix transmitted infection to 129 MM Tg45 mice with lower efficiency. Four of seven inoculated mice were infected, but only a single mouse developed clinical prion disease at 457 days post-inoculation (Table 1). The three mice scored as non-affected were culled between 314 and 395 days post-inoculation. All four infected mice had detectable PrPSc in brain that showed a predominance of di-glycosylated PrP, and three of these samples had PrPSc at sufficient concentration to permit accurate molecular strain typing. This analysis showed the propagation of type 4 PrPSc with no statistically significant difference in the PrPSc glycoform ratio from that seen in mice inoculated with vCJD brain (Table 2). All three vCJD appendix-inoculated 129 MM Tg45 mice propagating high levels of type 4 PrPSc showed the presence of florid PrP plaques in the brain by immunohistochemistry (Figure 2). Thus, despite a difference in the glycoform ratio of PrPSc between vCJD brain and appendix (Figure 1A), congruent transmission of the vCJD prion strain was seen in human PrP 129M transgenic mice characterized by the propagation of type 4 PrPSc and the occurrence of florid PrP plaques in the brain.
Table 2. Glycoform analysis of type 4 PrPSc propagated in the brains of 129 MM Tg45 mice after transmission of vCJD brain or appendix
Relative proportion of each PrP glycoform (mean ± SD %) in type 4 PrPSc propagated in the brains of 129 MM Tg45 transgenic mice inoculated with vCJD brain (n = 4 mice) or appendix (n = 3 mice).
Unpaired two-tailed t-test.
55 ± 2
53 ± 3
30 ± 1
32 ± 3
15 ± 2
15 ± 1
While transmission of sporadic Creutzfeldt–Jakob disease prions from formalin-fixed human brain tissue to primates has been reported 63, 64, to date there have been no data concerning the effect of formaldehyde on vCJD prions. We found that FFPE vCJD brain transmitted infection to five of nine inoculated 129 MM Tg45 mice without an apparent change in prion strain properties. Three inoculated mice developed clinical prion disease, with a mean incubation period of 580 ± 21 days, while two other recipients were sub-clinically infected (Table 1). The four mice scored as non-affected were culled between 540 and 610 days post-inoculation. All five infected mice had positive PrP immunohistochemistry, and in three brains prominent florid PrP plaques were observed (Figure 2). Accurate molecular strain typing was possible with two brains, which confirmed the propagation of type 4 PrPSc (Figure 1C).
Although the strain properties of vCJD prions are preserved after exposure to formaldehyde, the unknown level of prion infectivity present in the starting vCJD brain tissue prevents us from commenting on how formalin fixation and subsequent processing may have affected vCJD prion titre. However, the reduced attack rate and more prolonged mean incubation period observed with the FFPE vCJD brain compared with the control vCJD brain homogenate are consistent with a significantly reduced prion titre. In this regard, formalin fixation of prion-infected hamster brain followed by histological processing in the absence of exposure to formic acid has been reported to reduce prion infectivity between 2 and 3 logs 65, 66.
In contrast to FFPE vCJD brain, FFPE vCJD spleen and both of the FFPE abnormal PrP-positive appendix specimens failed to transmit prion infection to 129 MM Tg45 mice (Table 1). Exhaustive examination of brains from mice at up to 610 days post-inoculation failed to reveal the presence of PrPSc after sodium phosphotungstic acid precipitation or the presence of abnormal PrP deposition by immunohistochemistry (data not shown).
Current UK Department of Health risk assessment and risk management are based to a large extent on the results of the Hilton et al study 17–19. Fundamental to this is whether the three positives identified by PrP immunohistochemistry are ‘true positives’ indicating vCJD prion infection. In this study, we have attempted to transmit prion infectivity from FFPE tissue sections derived from two abnormal PrP-positive appendix specimens identified in the retrospective vCJD prevalence screen of Hilton et al41. While transgenic mice expressing human PrP 129M reported vCJD prion infection after inoculation with frozen vCJD brain or appendix and also FFPE vCJD brain, no prion transmission was observed from FFPE vCJD spleen or the FFPE appendix specimens. However, these negative findings cannot be interpreted, as the levels of infectivity in the positive control lymphoreticular tissue following formalin fixation were below the detection limit of our assay. It is entirely possible, therefore, that the Hilton et al positives were indeed prion-infected and that a lower titre in these pre-clinical tissues and/or the effect of formalin fixation precluded detection by our bioassay. Unless more sensitive assays for vCJD prions become available, verification of the estimated UK prevalence of vCJD infection will have to come from future prevalence studies 17–19, 47.
Although the present findings are not informative with regard to vCJD prevalence, two results are of some importance. Firstly, we have demonstrated that the strain properties of vCJD prions do not appear to be affected by tissue fixation with formaldehyde or by exposure to other chemicals that are commonly used during histological processing. Secondly, we have shown that prions associated with distinct PrPSc glycotypes in vCJD brain (type 4) 2 or lymphoreticular tissues (type 4t) 33, 34 do not transmit disease with distinct prion strain properties. These findings are in agreement with two other recent studies that have also shown congruent prion strain properties after transmission of brain or lymphoid tissue from sheep scrapie to bank voles 67 or from vCJD to wild-type mice 68. Prion strains may comprise an ensemble or quasi-species maintained under selection pressure in a host and the populations of which may therefore differ in different tissues 52, 69. According to this general model 52, even if a distinct prion strain were propagating in lymphoreticular as opposed to CNS tissue, it would revert to the brain-specific strain type on passage in transgenic mouse brain. Thus, based on the current findings and other published data 30, 70 it appears that molecular and neuropathological examination will not be able to differentiate between primary vCJD cases resulting from exposure to BSE prions and secondary cases resulting from iatrogenic exposure to vCJD-infected peripheral tissues.
We thank all patients and their families for generously consenting to the use of human tissues in this research. This work was funded by the Medical Research Council. Some of the work was undertaken at University College London Hospital NHS Foundation Trust, which received a proportion of funding from the Department of Health's NIHR Biomedical Research Centres funding scheme.
Author contribution statement
JDFW and JC coordinated the design and operation of the study, and drafted the manuscript. JWI and DAH identified the positive appendix specimens, provided formalin-fixed tissues, and contributed to the design of the study. EAA contributed to the study design and generated transgenic mice. JDFW, ID-M, and SJ performed the biochemical analysis. JL, CO'M, CP, and SB carried out neuropathological analysis. All authors contributed to and approved the final version of the manuscript. JC had full access to all the data in the study and had final responsibility for the decision to submit for publication.