Creutzfeldt–Jakob disease (CJD) is the most common human prion disease. It includes sporadic, genetic, and acquired forms. Sporadic CJD (sCJD), accounting for 85–90% of all cases, is clinically characterized by RPD with ataxia, myoclonus or other neurologic signs and, neuropathologically, by a widespread brain deposition of PrPSc leading to spongiform change, microglial activation, synaptic and neuronal loss, and astrocytic gliosis of variable severity and regional involvement (Zerr and Parchi, 2018). Despite these common features, the disease shows wide phenotypic variability, recognized since its early descriptions. Current sCJD classification comprises six major clinicopathological phenotypes that largely correlate at the molecular level with the genotype (methionine, M or valine, V) at the polymorphic codon 129 of the prion protein gene (PRNP) and two PrPSc forms (types 1 and 2) with a distinctive size (21 and 19 kDa) of their proteinase K-resistant core (Parchi et al., 1999a,2009, 2012). More precisely, each phenotypic variant or “subtype” of sCJD results from a specific codon 129 genotype/PrPSc type combination, with only two exceptions (Table 1). The MM(V)1 subtype includes both MM1 and MV1 cases, given that they are phenotypically indistinguishable. In contrast, the MM2 group comprises two distinctive histotypes with a different topographical distribution of lesions mainly affecting the cerebral cortex (MM2-Cortical or MM2C) or the thalamus (MM2-Thalamic or MM2T). The latter phenotype is also known as the sporadic form of fatal insomnia (Parchi et al., 1999b; Abu-Rumeileh et al., 2018). It is also noteworthy that about 35% of sCJD cases show the co-occurrence of PrPSc types 1 and 2, mainly affecting the subgroup of sCJD subjects carrying MM at codon 129 (Parchi et al., 2009). In these “mixed” subtypes, pathological features of two “pure” subtypes co-occur, with the predominant lesions reflecting the most prevalent PrPSc type in the brain (Parchi et al., 2009). Genetic CJD (gCJD), also including the familial form of fatal familial insomnia (FFI), a prion disease subtype with distinctive clinicopathological features, is the second most common human prion disease, accounting for 10–15% of CJD cases. Patients with gCJD carry point mutations or octapeptide repeat insertions (OPRI) in PRNP (Capellari et al., 2011). The clinicopathological variability of gCJD essentially reproduces that of sCJD since the disease phenotype is also primarily determined by the PrPSc types 1 and 2 in conjunction with the codon 129 genotype in the mutated allele (Baiardi et al., 2021). Nevertheless, there are a few significant exceptions to this general rule, given that a few mutations co-distribute with a phenotype that is divergent from that expected from the PRNP haplotype (i.e., genotype determined by the 129 polymorphism and the mutation) and PrPSc type combination (Baiardi et al., 2021). Iatrogenic CJD (iCJD) resulted from the human-to-human transmission of CJD through medical procedures (Brown et al., 2012). The principal sources of infectious prions have been contaminated dura mater graft, and human-derived pituitary growth hormone extracted from tissues obtained post-mortem. Others included contaminated neurosurgical instruments, human derive-gonadotropins, corneal grafts, and blood transfusions from patients with variant CJD (vCJD). Although, as in sCJD, the pathological phenotype of iCJD largely depends on the PRNP codon 129 genotype and the type of PrPSc accumulating in the brain, the phenotypic spectrum of iCJD, likewise gCJD, does not fully reproduce that of the sporadic form. Of note, while, to date, there have been no proven iCJD cases showing the MM2C, MM2T, or VV1 histotype, a “plaque-type” histotype linked to MM at codon 129 has been consistently observed in iCJD but not in sCJD (Shimizu et al., 1999; Hoshi et al., 2000; Kretzschmar et al., 2003; Yamada et al., 2009; Kobayashi et al., 2014).

Variant CJD is an acquired, zoonotic prion disease caused by oral exposure to food contaminated with material from bovine spongiform encephalopathy (BSE)-affected cattle. Since 1995, when the first identification of the disease in the United Kingdom occurred (Will et al., 1996), there have been approximately 230 individuals affected by vCJD, mainly in the United Kingdom (Will and Ironside, 2017). However, the disease has become exceedingly rare in recent years, following the decline of BSE in cattle. Pathological hallmarks of vCJD include the presence of widespread and numerous florid plaques, characterized by a central eosinophilic amyloid core and a pale periphery of radiant fibrils surrounded by a corona of vacuolation, and prion protein (PrP) accumulation as small cluster plaques (Ritchie and Ironside, 2017). Given the “peripheral” pathogenesis and the lymphotropism of the BSE strain, PrPSc is also present in relatively high amounts (i.e., in comparison to other CJD phenotypes) outside the CNS, especially in the lymphoreticular tissues (Wadsworth et al., 2001).

Variably protease-sensitive prionopathy (VPSPr) is the most recently recognized sporadic prion disease and owes its name to the peculiar properties of its PrPSc aggregates. Although VPSPr may affect subjects harboring each of the three genotypes at codon 129, it is most frequently linked to valine homozygosity (129VV) (Head et al., 2009, 2013; Zou et al., 2010; Assar et al., 2015). The clinical phenotype is often consistent with an unspecific, Alzheimer-like, or frontotemporal-like dementia. However, two patients with autopsy-confirmed VPSPr and a clinical diagnosis of amyotrophic lateral sclerosis have also been reported (Vicente-Pascual et al., 2018). Spongiform change is characterized by non-confluent vacuoles of intermediate size [i.e., larger than in typical sCJD MM(V)1, but smaller than in sCJD MM2C]. The lesion profile shows a predominant involvement of cerebral cortices, striatum, and thalamus, while the cerebellum is variably affected and the hippocampus relatively spared. PrP immunostaining reveals micro plaques in the molecular layer of the cerebellum surrounded by clusters of granular deposits (i.e., target-like pattern). Western blot analysis of PrPSc shows a ladder-like profile of at least 5 proteinase K-resistant fragments migrating in the 7–26 kDa range (Zou et al., 2010). The lack or the minimal representation of the diglycosylated PrPSc isoform is another distinctive feature of VPSPr. Furthermore, both the amount and the resistance of PrPSc to proteinase K digestion are lower in VPSPr than in sCJD (Saverioni et al., 2013). Proteinase K-resistance is also significantly influenced by the codon 129 genotype: it is lowest in VPSPr 129VV, highest in 129MM, and intermediate in 129MV (Zou et al., 2010; Baiardi et al., 2022).

Inherited PrP-amyloidosis encompasses a distinct group of genetic prion diseases characterized by the accumulation of an internal PrPSc fragment, truncated at both C- and N-termini, in the form of amyloid plaques or cerebral amyloid angiopathy (Ghetti et al., 2018). Depending on the causative mutation, the molecular mass of the abnormal PrPSc fragment may range from 6 to 11 kDa and is occasionally associated with other PrP truncated species with higher molecular mass. The broad phenotypic heterogeneity in this group, observed even within family members, depends on the several disease-associated genetic mutations, including point, stop-codon PRNP mutations, 7–12 OPRI, and duplication in the hydrophobic domain, the specific PRNP haplotype, and other yet undefined individual factors. Gerstmann–Sträussler–Scheinker (GSS) is the most prevalent inherited PrP-amyloidosis and the first genetic prion disease linked to a PRNP mutation (Hsiao et al., 1989). Clinically, GSS often manifests slowly progressive ataxia, pyramidal signs, and late dementia. The accumulation of amyloid PrP plaques in the CNS, especially in the cerebral and cerebellar cortices, represents its pathological hallmark (Ghetti et al., 2018). Spongiform change can be either absent (more often), limited to small foci or widespread. Finally, it is noteworthy that the GSS phenotype also comprises tau pathology in the form of neurofibrillary tangles and dystrophic neurites around the PrP plaques (Ghetti et al., 2018). The more recent discovery of premature truncating mutations in PRNP linked to familial prion diseases has significantly expanded the phenotypic spectrum of PrP-amyloidosis (Table 1). Besides GSS, PRNP truncating mutations have been linked to cerebral amyloid angiopathy (Ghetti et al., 2018) and, more recently, even found in patients with chronic diarrhea, progressive autonomic failure, and peripheral polyneuropathy in association with widespread abnormal PrP deposition in both peripheral organs and CNS (Matsuzono et al., 2013; Mead et al., 2013; Capellari et al., 2018).

The broad clinical heterogeneity at onset, often mimicking other RPDs, and the variable disease progression and survival times, sometimes overlapping with other more prevalent neurodegenerative dementias (e.g., Alzheimer’s disease), significantly challenge the early diagnosis and clinical management of patients with prion disease (Baiardi et al., 2018). Yet, a reliable early diagnostic formulation is crucial to rule out other rapidly progressive and potentially treatable neurological syndromes, reduce the risk of iatrogenic transmission, and improve epidemiological disease surveillance. Besides the search of mutations in PRNP, which represent the main diagnostic tool for familial prion disease, electroencephalographic findings, brain-derived cerebrospinal fluid (CSF) protein assays serving as surrogate markers for neuronal damage, and diffusion-weighted magnetic resonance imaging (DW-MRI), have provided the primary support for the clinical diagnosis of prion disease in the past 20 years. However, these biomarkers provide a suboptimal diagnostic accuracy and show a significant variability across diseases and their subtypes. Accordingly, the in vitam diagnosis of sCJD has become ‘‘probabilistic’’ (‘‘possible’’ or ‘‘probable’’), based on the evaluation of neurological symptoms and signs and the results of the above-mentioned ancillary tests¹ (Table 2). Periodic sharp-wave complexes (PSWCs) represent the most characteristic electroencephalographic anomaly detected in sCJD (Wieser et al., 2006). However, it is only found in about 65% of patients, resulting in a low diagnostic sensitivity (Steinhoff et al., 2004). Moreover, PSWCs may also be recorded in other RPDs (e.g., various encephalitis, hepatic encephalopathy, and lithium intoxication) (Karnaze and Bickford, 1984; Smith and Kocen, 1988; Steinhoff et al., 1996, 2004; Tschampa et al., 2001), resulting in a suboptimal specificity. As an additional limitation, PSWCs typically only show up during the clinical course of the disease, 2–3 months after onset (Hansen et al., 1998; Wieser et al., 2004), and may even disappear in the pre-terminal phase (Wieser et al., 2004). Hence, an electroencephalographic recording done too early or, less commonly, too late in the course of the disease may miss the PSWCs. Finally, PSWCs are commonly recorded only in the “typical” form of the disease [i.e., the MM(V)1 subtype], not in the other sCJD subtypes (Parchi et al., 1999a). Brain MRI has played a pivotal role in the clinical diagnosis of sCJD since the introduction of the Fluid Attenuated Inversion Recovery and DWI techniques. The step forward has significantly increased the MRI’s diagnostic accuracy reaching a 91–96% sensitivity and a 88–95% specificity (Shiga et al., 2004; Young et al., 2005; Vitali et al., 2011; Carswell et al., 2012). The DWI sequence has a higher sensitivity than Fluid Attenuated Inversion Recovery, particularly for the early detection of signal hyperintensity and its evaluation at the cortical level (Shiga et al., 2004; Tschampa et al., 2007; Vitali et al., 2011). The increased signal in DWI scans correlates with the underlying presence and degree of spongiform change (Manners et al., 2009). In general, typical findings in sCJD patients are signal hyperintensity in the deep gray matter (caudate nucleus and putamen, less commonly the thalamus), the cerebral cortex, or both (Zerr et al., 2009). As the disease progresses, the signal intensity usually increases and extends to additional brain regions (Ukisu et al., 2005; Eisenmenger et al., 2016). These signal abnormalities reflect the heterogeneity of the lesion profiles among sCJD subtypes in terms of spatial pattern and its evolution during disease progression (Baiardi et al., 2019a). The typical abnormal sCJD signal profile comprises an increased signal in Fluid Attenuated Inversion Recovery or DWI sequences in at least two cortical regions (temporal, parietal, or occipital) or at the level of caudate and putamen or in both areas. This profile has optimal diagnostic accuracy in the differential diagnosis with other RPDs and is part of the current diagnostic criteria for sCJD (Hermann et al., 2021). DW-MRI may disclose a typical, almost subtype-specific, lesion pattern in a subset of cases, including a high signal intensity in the basal ganglia and thalamus in the VV2 subtype, a prominent, often isolated, cortical signal hyperintensity in subjects MM2C or VV1, and a marked asymmetry of the signal hyperintensity in the MM(V)1 variant (Meissner et al., 2009; Bizzi et al., 2021). On the other hand, DW-MRI may be unrevealing in 5–10% of cases, depending on the disease subtype and the time from disease onset, and yields a variable percentage of false-positive scans in RPDs with non-prion etiologies (Zerr et al., 2009). In this regard, it is worth noting that CJD surveillance centers, likely as a consequence of improved technology and increased awareness of DWI hyperintensities, have been witnessing an increase in referrals due to MRI findings of isolated cortical ribboning in patients with seizures, hyponatremia, Wernicke’s encephalopathy or even ischemia, often in the absence of typical clinical features of CJD. The analysis of CSF offers essential support to the clinical diagnosis of sCJD and the exclusion of an inflammatory pathology of the CNS. Routine cytological and biochemical examinations of CSF from affected patients are typically free of significant abnormalities (Olsson, 1980). One-third of cases show a slight non-specific increase in total protein concentration, and, occasionally, a mild pleocytosis or oligoclonal bands are also present (Jacobi et al., 2005). Among the CSF markers of CJD, proteins 14-3-3 and total tau represent validated surrogate markers of neuronal damage (Hermann et al., 2021). The 14-3-3 protein assay exhibits a diagnostic sensitivity of 85–95% and a specificity of 74–96% (Zerr and Parchi, 2018). The finding of elevated concentrations of 14-3-3 in an appropriate clinical setting supports the diagnosis of CJD. However, given that this cytosolic protein is released into the CSF following non-specific neuronal damage (Satoh et al., 1999), other pathologies, including other neurodegenerative RPDs, metabolic encephalopathies, stroke, paraneoplastic syndromes, tumors, may also yield a positive test (Saiz et al., 1999; Chapman et al., 2000; Burkhard et al., 2001; Lattanzio et al., 2017). Patients with CJD usually show an increase in levels of 14-3-3 as the disease progresses, unlike those with acute neuronal damage, where the trend is the opposite (Sanchez-Juan et al., 2007). Of note, the assay demonstrated a lower sensitivity and higher incidence of false negatives for sCJD subtypes characterized by slower progression and longer disease duration (such as MM2 and MV2K) (Castellani et al., 2004; Sanchez-Juan et al., 2006; Lattanzio et al., 2017). High values of total tau (>1,200–1,300 pg/ml) have shown a slightly higher diagnostic accuracy for all subtypes of sCJD. However, likewise 14-3-3, the dosage of total tau is not exempt from false positives and negative readouts (Zerr and Parchi, 2018). Indeed, patients with Alzheimer’s disease and with inflammatory or neoplastic disorders of the CNS may also show high concentrations of total tau in CSF (Constantinescu et al., 2016; Olsson et al., 2016; Lehmann et al., 2019). Finally, similarly to 14-3-3, a reduced sensitivity has been demonstrated in the MM2 and MV2K subtypes, VPSPr, and early disease stages (Cohen et al., 2016; Lattanzio et al., 2017).

RT-QuIC is the first antemortem technique that can reliably detect misfolded PrP, the only available disease-specific biomarker for prion disease. The assay demonstrated high specificity and variable sensitivity for human prions, depending on the disease subtype and the applied protocol. RT-QuIC exploits the self-replicating (seeding) ability of pathogenic PrPSc as an amplification strategy to ultra-sensitively detect the limited amounts of the abnormal protein present in CSF and other tissues.

The development of RT-QuIC and other similar seeding assays (e.g., protein misfolding cyclic amplification) stem from the protein-only hypothesis of prion replication (Prusiner, 1998) and the demonstration that aggregated species of PrP can be generated in a cell-free assay (Kocisko et al., 1994). This seminal result was obtained by incubating an excess of partially denatured PrPSc, isolated from prion-infected hamsters, with 35S-labeled recombinant PrPC (rPrP) derived from uninfected hamster brains. The subsequent demonstration of proteinase K-resistant labeled PrP proved the de novo formation of misfolded species (Kocisko et al., 1994). Using cyclic sonication and bacterially expressed rPrP in vast excess to facilitate the conversion into its pathogenic counterpart and an automated system to guarantee reproducibility, Atarashi et al. (2007) developed the first QuIC assay. In this reaction, minute amounts of seed derived from brain homogenates were sufficient to trigger a significant aggregation of the excess of the rPrP in a reasonable time. By combining the QuIC cyclic amplification with the real-time detection of the products by thioflavin T (ThT), Wilham et al. (2010) later developed the currently used RT-QuIC set-up.

In the prion RT-QuIC assay, a patient’s biological sample containing PrPSc (seed) is mixed in 96-well plates with an excess of rPrP as a substrate, a standard buffer, and ThT, a fluorescence emitting dye with a strong affinity with protein amyloid structures (Wilham et al., 2010; Atarashi et al., 2011; Orrú et al., 2015a) (Figure 1). The plate is then incubated at various temperatures (from 42 to 55°C) depending on the specific protocol and subjected to cycles of double orbital shaking and rest inside a fluorescence microplate reader. While incubating, the specimen PrPSc seed induces a conformational change of the rPrP substrate, resulting in aggregation and amyloid fibril formation (Atarashi et al., 2011). The ThT fluorescence emission, recorded in real-time during the entire incubation period, allows the detection of these aggregates when formed. A typical sigmoidal curve of fluorescence indicates the positivity of the test. The amyloid formation follows an exponential increase after an initial lag phase of variable duration with only a background signal. Shaking, another critical factor for the reaction, promotes the fragmentation of aggregates and the formation of novel, free reactive seeds, accelerating the amyloid fibril formation. A decline in fluorescence is observed at the end of the assay, likely due to ThT binding to amyloid fibrils which leads to fluorescence self-quenching and fibril compaction (Lindberg et al., 2017) (Figure 1).

There is a significant influence of the disease subtype/prion strain on the outcome of the RT-QuIC, determining a variable diagnostic performance of the assay in human prion disease (see the section on disease diagnostics). Besides, pre-analytical factors, the recombinant substrate, and the tissue matrix may also affect the assay’s results.

The concentration of particular cellular or molecular components in the tested sample matrices affects the prion RT-QuIC performance. For example, an increased total protein concentration in CSF > 1.0 g/L or a white cell count >10 × 10^6/L, can lead to false-positive results (Green, 2019). In contrast, red cells CSF contamination has an opposite effect and may result in a potentially false negative outcome (Cramm et al., 2016; Foutz et al., 2017). An excessive volume of human CSF can also inhibit the assay (Xiao et al., 2018), probably due to the presence of unknown factors that hamper amyloid formation in the RT-QuIC reaction. A high concentration of brain homogenate has a similar effect (Wilham et al., 2010; Mori et al., 2016), probably due to the inhibition exerted by endogenous brain polar lipids (Hoover et al., 2017).

In contrast, other classic pre-analytical variables known to potentially affect biomarker performance in biofluids, such as sample storing at room temperature or 4°C for up to 8 days or repeated freeze-thaw cycles, seem to have minimal or no effects on the RT-QuIC outcome (Cramm et al., 2016). However, these findings should not be considered conclusive and await replication in confirmatory studies.

Recombinant prion proteins, expressed in Escherichia coli, represent the substrates of choice for RT-QuIC. The development and production of suitable rPrP substrates, rapidly converting to amyloid in the presence of prion seeds but not prone to self-aggregation, is mandatory for optimal performance. Early RT-QuIC assays employed the full-length Syrian hamster rPrP (Ha rPrP23-231) or the full-length human rPrP (Hu rPrP23-231) (Wilham et al., 2010; Atarashi et al., 2011) as substrates. These proteins were stable and performed well, although not optimally when seeded with brain or CSF from typical sCJD (McGuire et al., 2012) but showed much lower sensitivity with samples from vCJD (Peden et al., 2012) or other rare and atypical sCJD subtypes (Lattanzio et al., 2017). Therefore, there have been attempts to improve the sensitivity for these specimens using modified protocols (i.e., sodium dodecyl sulfate concentration in the reaction buffer, temperature of incubation, and shaking speed) (Orrú et al., 2016), and more reactive substrates such as truncated hamster rPrP (Ha rPrP90-231) (Orrú et al., 2015a) in the so-called improved QuIC (IQ-CSF). The use of the N-terminal truncated hamster rPrP as substrate associated with modified parameters, including the incubation temperature of 55°C, the shaking speed of 700 rpm, and the sodium dodecyl sulfate concentration of 0.002%, produced a shorter experimental time with an increase in sensitivity of the test for sCJD in comparison to the first-generation RT-QuIC assay. The reason why the Ha rPrP90-231 substrate produces a more rapid and sensitive response is unclear. One possibility is that the absence of the flexible N-terminal region (23–89) destabilizes the native rPrP conformation, allowing a more rapidly refolding into the amyloid conformation (Orrú et al., 2015a).

The introduction of the IQ-CSF protocol contributed to significant development of the assay by showing retained high specificity, increased sensitivity, and a reduced assay time compared to the previous QuIC (PQ) protocol (Table 3) (Orrú et al., 2015a,2020; Groveman et al., 2016; Park et al., 2016; Bongianni et al., 2017; Foutz et al., 2017; Franceschini et al., 2017; Rudge et al., 2018; Abu-Rumeileh et al., 2019; Fiorini et al., 2020; Rhoads et al., 2020; Vallabh et al., 2020; Xiao et al., 2020; Moško et al., 2021). In this regard, the IQ-CSF protocol showed a lower detection limit for seeding prions in a large patient cohort comprising the whole spectrum of human prions (Franceschini et al., 2017). Indeed, in this study, IQ-CSF correctly identified 81% of definite CJD cases tested negative by PQ-CSF, mainly including atypical and uncommon prion phenotypes such as VPSPr. Similarly, Rhoads and collaborators (Rhoads et al., 2020) obtained high sensitivity (90.3%) and specificity (98.5%) with IQ-CSF in a large study group, including 567 autopsy-verified cases (Table 3). Thanks to the large size of samples from patients with post-mortem histotyping or carrying PRNP mutations, the authors of the latter two studies identified several factors associated with the reduced sensitivity of RT-QuIC. Indeed, some atypical sporadic prion disease subtypes (i.e., VPSPr, sCJD VV1, MM2T, and MM2C), as well as some genetic prion diseases (i.e., FFI and GSS), are more likely to have false-negative RT-QuIC results (Franceschini et al., 2017; Rhoads et al., 2020).

Two additional substrates, the sheep-hamster chimeric rPrP and in particular the bank vole (Bv) rPrP23-230 (Cramm et al., 2015; Orrú et al., 2015b; Mok et al., 2021), demonstrated a diagnostic potential despite their high reactivity and self-aggregating predisposition. Bank voles and transgenic mice expressing Bv rPrP are highly susceptible to many prion strains (Nonno et al., 2006; Watts et al., 2014). Following a first successful attempt to convert BSE prions in the skin of 2 vCJD patients using Bv rPrP (Orrú et al., 2017), in a recent study, 1 out 2 CSF samples collected post-mortem from vCJD cases was also positively converted by Bv rPrP (Mok et al., 2021). However, the calculated sensitivity (88.6%) and specificity (91.2%) in the overall study cohort were below the mean of those reported in previous studies using traditional substrates (Table 3). The use of CSF samples obtained post-mortem represents a significant limitation of this study and makes the comparison with the results obtained with CSF collected in vitam difficult. Nevertheless, the results raise interest in using Bv rPrP substrate for prion strains that are not responsive to other conventional substrates, such as the BSE strain. It also emphasizes the need for further assay optimization and validation for routinely diagnostic uses to minimize false-positive results (Mok et al., 2021).

Finally, differences in the substrate production methods and refolding procedures may affect the performance of RT-QuIC (Cheng et al., 2018). In this study, the authors linked different rPrP production methods to their performance in RT-QuIC assay. They reported that Ha rPrP provided consistent results regardless of the investigated production procedures. The high tolerance of Ha rPrP to different production protocols may explain the robust and reproducible performance of RT-QuIC obtained in international studies in which various homemade Ha rPrP substrates were used (Cramm et al., 2016; McGuire et al., 2016; Orrú et al., 2020). However, given the effect that various rPrP production protocols have on RT-QuIC performance, caution is required when comparing inter-laboratory results.

Currently, CSF is the matrix of choice for the prion RT-QuIC. The high performance of the CSF assay, the fact that CSF represents the accessible fluid that is more proximal to the brain, where PrPSc predominantly accumulates, and the systematic collection of this biofluid in patients with RPDs for diagnostic purposes all contributed to this scenario.

However, despite the highly dominant tropism of prions for the CNS, it is well-established that PrPSc may also accumulate in extra-neural tissues. Given the relatively low amount of PrPSc deposition in peripheral tissues, the detection of prions outside the CNS requires the use of ultrasensitive techniques, such as enhanced chemiluminescence (Head et al., 2003; Zanusso et al., 2003; Favereaux et al., 2004; Orrú et al., 2017; Baiardi et al., 2019b), tissue precipitation with sodium phosphotungstic acid (Glatzel et al., 2003; Zanusso et al., 2003), or other concentration/purification protocols (Head et al., 2003; Orrú et al., 2017; Baiardi et al., 2019b). Using these strategies, several groups detected PrPSc in the olfactory and retinal epithelium, cranial and peripheral nerves, muscles, skin, and spleen of sCJD patients. Moreover, they demonstrated PrPSc accumulations in many peripheral tissues, including tonsils, spleen, lymph nodes, circulating lymphocytes, and thymus in patients with vCJD (Wadsworth et al., 2001; Notari et al., 2010). These studies, performed in the pre-RT-QuIC era, paved the way for the application of the assay to accessible peripheral tissues collected in vitam. Indeed, several studies confirmed the ability of RT-QuIC to amplify a meager amount of PrPSc in peripheral tissue homogenates [i.e., olfactory mucosa (OM), and skin obtained by nasal brushing and punch biopsy, respectively] (Table 4). In sCJD, one group reported a sensitivity ranging from 90.5 to 96.5% and a specificity of 100% for the OM RT-QuIC using the Ha rPrP23-231 substrate (Orrú et al., 2014; Bongianni et al., 2017; Fiorini et al., 2020), suggesting an excellent diagnostic accuracy. Skin also represents a valuable matrix for the RT-QuIC assay (Orrú et al., 2017; Mammana et al., 2020). One study demonstrated that the PrPSc signal from both sCJD and vCJD skin samples taken ex vivo could be efficiently amplified by RT-QuIC using Bv rPrP as a substrate (100% sensitivity when multiple sites are analyzed). In contrast, only the PrPSc from sCJD skin samples reacted positively with the Ha rPrP23-231 substrate (Orrú et al., 2017). Moreover, the results of this study indicated that prion seeding activity was approximately 1000-fold lower in the skin than in the brain of the same patient and that the skin contains prion seed concentrations similar to those found in nasal brushings (Orrú et al., 2017). In another study, the skin RT-QuIC performed on punch biopsies taken both ex vivo and in vitam confirmed the high diagnostic value of the assay (Mammana et al., 2020) (Table 4). However, the Bv rPrP (88.6% sensitivity and 100% specificity) showed a better performance compared to Ha rPrP23-231 (68.6% sensitivity and 100% specificity). Moreover, although very preliminary, evidence also suggests that the prion-seeding activity in the skin increases with disease progression (Mammana et al., 2020). Recently, Xiao et al. (2020) applied the IQ protocol to skin samples collected in vitam from patients with probable sCJD and controls. They obtained a good sensitivity but a suboptimal specificity (Table 4).

A study demonstrated positive post-mortem RT-QuIC results on multiple eye components, including the retina, optic nerve, extraocular muscle, choroid, lens, vitreous, and sclera of 11 sCJD patients (Orrú et al., 2018). The assessment of the diagnostic value of the analysis of routinely accessible eye tissue or fluid needs further studies and validation. Besides cranial nerves, prion seeding activity has been demonstrated in peripheral nerves (sural, sciatic, femoral, and common peroneal) collected post-mortem from one gCJD-V210I and 11 sCJD individuals (Baiardi et al., 2019b). Both studies used Ha rPrP90-231 as recombinant substrate and reported complete sensitivity and specificity of the RT-QuIC assay (Orrú et al., 2018; Baiardi et al., 2019b). The observation that in scrapie-infected hamsters and transgenic mice, skin samples reveal prion seeding activity before the onset of overt clinical signs of infection (Wang et al., 2019) suggests that skin punch biopsies could be relevant for preclinical diagnosis of human prion disease, particularly in subjects at genetic risk for the disease.

Besides the skin punches and the brushing of OM, digestive tract biopsies may provide an accessible tissue source for the detection of prion seeding activity. Satoh et al. (2019) recently demonstrated a 100% sensitivity in a few patients with sCJD (n = 4) and genetic prion disease (n = 2) using post-mortem tissue and multiple sites for each patient (Table 4). The RT-QuIC assay is far less efficient than other cell-free in vitro seeding assays, such as protein misfolding cyclic amplification and its adaptations, in detecting vCJD prions (Giaccone and Moda, 2020). Orrú et al. (2011) tried to improve the assay by adding an initial immunoprecipitation step to the RT-QuIC protocol and replenishing the rPrP (chimeric sheep-hamster) at 24 h. The amended assay, called enhanced QuIC, allowed PrPSc detection in human plasma samples spiked with vCJD brain homogenates up to 10¹⁴-fold dilutions, and showed significantly more sensitivity than the standard protocol (Orrú et al., 2011). However, the higher technical complexity of enhanced QuIC, compared with protein misfolding cyclic amplification, prevented its reproducibility, standardization, and diffusion.

Shortly after its application to clinical cohorts, the valuable performance of RT-QuIC in discriminating sCJD from non-CJD with exceptionally high specificity prompted the update of the diagnostic criteria for sCJD. Since January 2017, the positivity of RT-QuIC assay in CSF or other tissues in subjects presenting with a progressive neurological syndrome supports the clinical diagnosis of probable sCJD (Table 2), making no longer necessary the previous stringent clinical criteria for “possible” CJD (i.e., the association between dementia and multifocal neurological signs) as mandatory for accepting the diagnosis of probable disease based on the positivity of electroencephalography, CSF or MRI biomarkers. Indeed, the latter clinical scenario is uncommon at onset and during the early disease stage, especially in the “atypical” sCJD subtypes (all but the MM(V)1), which account for up to 40% of cases. Therefore, the association between positive RT-QuIC assay and a non-further specified “progressive neurological syndrome” allows to diagnose the disease earlier and more accurately even in cases presenting “unusual” neurological symptoms and/signs or isolated symptoms such as in the Heidenhain and the ataxic (VV2) variants (Baiardi et al., 2018).

Nevertheless, despite the addition of RT-QuIC, the novel criteria still include the diagnostic investigations of the previous criteria (Zerr et al., 2009), given their wider diffusion (i.e., at least DW-MRI, and electroencephalography are often available in secondary care centers). They presumably have an added diagnostic value in cases tested negative by RT-QuIC. However, to some extent, they also increase the false positive cases, given their incomplete specificity. Future studies should carefully address this issue and evaluate if the insufficient specificity of the other biomarkers exceeds the limit in sensitivity of the RT-QuIC.

The variable protocols used among laboratories, including the different recombinant substrates known to affect the test performance, represent another issue to address. Notably, current updated diagnostic criteria do not include any technical specification or detail in this regard.

Finally, the choice to keep the door open to matrices other than CSF for diagnostic RT-QuIC deserves a further comment. To date, CSF is by far the most studied matrix. CSF RT-QuIC using different recombinant substrates showed reproducible results among research groups and high inter-laboratory agreement in ring trials (McGuire et al., 2016; Orrú et al., 2020). Conversely, despite the promising initial results, RT-QuIC using alternative matrices, such as OM and skin, is still limited to a few specialized laboratories, confirmatory studies are lacking, and the performance of different recombinant substrates are largely unknown. For these reasons, RT-QuIC assays currently use CSF for diagnostic purposes, even if other specimens are allowed according to the official criteria (Hermann et al., 2021).

The vast phenotypic heterogeneity of human prion diseases has profound implications for the clinical diagnosis, including the results of ancillary investigations, which consistently show variable accuracy depending on the disease phenotype. Here we reviewed the results of studies applying the RT-QuIC on CSF or peripheral tissue samples collected in vitam or post-mortem in neuropathologically confirmed sCJD cases classified according to the disease histotype (Table 5). Since the description of mixed histotypes is scanty in most studies, we considered only those clearly stating the dominant histotype of mixed cases. Similarly, we only included MM2 subjects when we found a clear statement on the distinction between the cortical and thalamic subtypes.

The anamnestic positive family history or the presence of a pathogenic PRNP mutation in a patient with a progressive neuropsychiatric syndrome already supports the clinical diagnosis of genetic prion disease. However, this diagnosis can be complex sometimes due to the wide range of clinicopathological presentations, the very low incidence, and the possibility of a negative familial history even for mutations with high penetrance (Ladogana and Kovacs, 2018). In this context, the novel ultrasensitive seeding assays contributed to a more accurate recognition of patients with genetic prion disease. As for the sporadic form, the diagnostic value of RT-QuIC differs according to the mutation and clinicopathological phenotype. Overall, the diagnostic accuracy is high in gCJD and low in GSS and FFI cases. Finally, RT-QuIC might be of pivotal importance to determine when an eventual preventive therapy should be started in future preventive clinical trials in people at risk for genetic prion diseases (Hermann et al., 2020).