Detection of Prions in Brain Homogenates and CSF Samples Using a Second-Generation RT-QuIC Assay: A Useful Tool for Retrospective Analysis of Archived Samples

The possibilities for diagnosing prion diseases have shifted significantly over the last 10 years. The RT-QuIC assay option has been added for neuropsychiatric symptoms, supporting biomarkers and final post-mortem confirmation. Samples of brain homogenates used for final diagnosis, archived for many years, provide the possibility for retrospective studies. We used a second-generation RT-QuIC assay to detect seeding activity in different types of sporadic and genetic prion diseases in archival brain homogenates and post-mortem CSF samples that were 2 to 15 years old. Together, we tested 92 archival brain homogenates: 39 with definite prion disease, 28 with definite other neurological disease, and 25 with no signs of neurological disorders. The sensitivity and specificity of the assay were 97.4% and 100%, respectively. Differences were observed in gCJD E200K, compared to the sporadic CJD group. In 52 post-mortem CSF samples—24 with definite prion disease and 28 controls—we detected the inhibition of seeding reaction due to high protein content. Diluting the samples eliminated such inhibition and led to 95.8% sensitivity and 100% specificity of the assay. In conclusion, we proved the reliability of archived brain homogenates and post-mortem CSF samples for retrospective analysis by RT-QuIC after long-term storage, without changed reactivity.


Introduction
Prion diseases, despite being rare, are invariably fatal. Sporadic Creutzfeldt-Jakob disease (sCJD), the most common human prion disease, belongs to the group of transmissible spongiform encephalopathies (TSEs) [1]. Aside from sporadic etiology, TSEs may have a genetic predisposition background, with mutations in the prion protein gene making the protein prone to misfolding and aggregation. Other genetic prion diseases include genetic CJD (gCJD), Gerstmann-Sträussler-Scheinker disease (GSS), and Fatal familial insomnia (FFI) [2]. Misfolding of the prion protein occurs at the beginning of the disease, regardless of the origin. This event is common to all prion diseases, and does not play a role if it arises stochastically (sporadic form), from a genetic predisposition, or enters the body from the outside (transmissible/iatrogenic form). Misfolded prion protein (PrP Sc ) represents the only specific TSE marker, and its detection is used for final confirmation of the disease by immunohistochemistry and western blot in post-mortem brain samples. Clinical diagnosis of TSEs is based mainly on characteristic neuropsychiatric symptoms, supported by results of EEG tests, MRI scans, and diagnostic biomarkers, such as the level of 14-3-3 and total Tau protein in CSF [3]. The ante-mortem options for direct detection of prions are limited, due to low levels of PrP Sc in peripheral tissues. However, the implementation

Establishment of the Second-Generation RT-QuIC Assay
We purified and refolded the rSHa PrP(90-231) substrate at the quality and quantity required for the second-generation RT-QuIC assay. The yield of the recombinant protein was 45-60 mg per purification. SDS-PAGE with Coomassie staining and immunoblotting confirmed the correct size of rSHa PrP (90-231):~17 kDa (not shown). The quality control assay showed a robust positive RT-QuIC response for 5 × 10 −6 -5 × 10 −9 dilutions of control sCJD MM1 BH and lack of spontaneous aggregation. To compare buffers for BHs, we prepared BH samples in buffer used for RT-QuIC (100 mM Tris-Cl, pH 8.0; 150 mM NaCl; 13 mM EDTA; 12 mM sodium deoxicolate; 0.5% IGEPAL CA-630) and in buffer used for western blot in our laboratory (see methods). The RT-QuIC assay results showed that the buffer used for western blot can be used for RT-QuIC, and the type of detergent and the absence of EDTA did not affect the assay. We tested BH dilutions from 5 × 10 −5 to 5 × 10 −12 , which have been used in RT-QuIC studies with similar results (data not shown). The calculated thresholds for BH and CSF samples estimated using OND controls were 65,413 AU and 33,166 AU, respectively. The BH threshold calculated using corneal donors was 67,672 AU, almost identical to the OND BH threshold ( Figure 1A,B). The calculated threshold for CSF diluted OND samples (76,436 AU) was more than twice higher than for undiluted CSF (not shown). As a standard positive control, we used freshly prepared sCJD MM1 BH in dilutions of 5 × 10 −6 up to 5 × 10 −10 . The seeding activity in the positive control was regularly detected in a 5 × 10 −9 dilution ( Figure 1C), corresponding to a final dilution of BH in the reaction mix of 10 −11 . Figure 1. Establishment of the RT-QuIC assay threshold for post-mortem brain homogenates: (A) The results of the analysis of the control brain homogenate samples (dilution 5 × 10 −6 ) of patients with other neurological diseases (OND, n = 28). The red line represents the arithmetic mean of ThT fluorescence (expressed as arbitrary units AU) at each time point, along with the standard deviation (black bars). The threshold was calculated as the arithmetic mean of maximal ThT fluorescence + 5 SD; (B) Analysis of control brain homogenate samples (dilution 5 × 10 −6 ) of corneal donors (CD, n = 25); (C) Typical results of the second-generation RT-QuIC assay with serial dilutions of control sCJD brain homogenate (sCJD MM1). The assay gave a positive result for the brain homogenate diluted by 5 × 10 −9 .

Analysis of Archived Brain Homogenates Using RT-QuIC
Seeding activity was detected in all archived BH samples with confirmed TSE, except for one case of sCJD VV2. No control OND or corneal donors BH gave positive results ( Figure 1A,B). The mean maximal ThT fluorescence of glycotype 1 sCJD BHs was significantly higher (207063 AU vs. 179126 AU, p < 0.05; Figure 2D), and their time to threshold value at a dilution of 5 × 10 −7 was shorter (p < 0.05; Figure 2E) than the respective values of glycotype 2 sCJD BHs. Differences within the 129 polymorphism group in maximal ThT fluorescence and time to threshold values were not evaluated, due to the small number of analyzed samples in the individual groups. The samples from the sCJD group were clearly identifiable as positive in all four replicates at 5 × 10 −6 sample dilution, and were positive at a dilution of BH up to 5 × 10 −8 or 5 × 10 −9 ( Figure 2E). The results of the analysis of the control brain homogenate samples (dilution 5 × 10 −6 ) of patients with other neurological diseases (OND, n = 28). The red line represents the arithmetic mean of ThT fluorescence (expressed as arbitrary units AU) at each time point, along with the standard deviation (black bars). The threshold was calculated as the arithmetic mean of maximal ThT fluorescence + 5 SD; (B) Analysis of control brain homogenate samples (dilution 5 × 10 −6 ) of corneal donors (CD, n = 25); (C) Typical results of the second-generation RT-QuIC assay with serial dilutions of control sCJD brain homogenate (sCJD MM1). The assay gave a positive result for the brain homogenate diluted by 5 × 10 −9 .

Analysis of Archived Brain Homogenates Using RT-QuIC
Seeding activity was detected in all archived BH samples with confirmed TSE, except for one case of sCJD VV2. No control OND or corneal donors BH gave positive results ( Figure 1A,B). The mean maximal ThT fluorescence of glycotype 1 sCJD BHs was significantly higher (207063 AU vs. 179126 AU, p < 0.05; Figure 2D), and their time to threshold value at a dilution of 5 × 10 −7 was shorter (p < 0.05; Figure 2E) than the respective values of glycotype 2 sCJD BHs. Differences within the 129 polymorphism group in maximal ThT fluorescence and time to threshold values were not evaluated, due to the small number of analyzed samples in the individual groups. The samples from the sCJD group were clearly identifiable as positive in all four replicates at 5 × 10 −6 sample dilution, and were positive at a dilution of BH up to 5 × 10 −8 or 5 × 10 −9 ( Figure 2E).
The one sCJD VV2 case gave negative results in three independent RT-QuIC assays utilizing two different archived BH aliquots. The negative results of the RT-QuIC assay of the case were unequivocal with characteristic spongiform changes and presence of proteinase K-resistant prion protein, visualized both by immunohistochemistry and western blot at the time of sampling (not shown). RT-QuIC analysis of the corresponding archived CSF samples confirmed the presence of prions ( Figure 3B); however, the shape of the curve representing the ThT fluorescence signal was unusual, suggesting either the potential presence of inhibitors or spontaneous aggregation. The 10× dilution of the CSF sample led to clear positivity with a typical curve shape ( Figure 3B). To demonstrate that the negative RT-QuIC results were likely caused by BH degradation during long-term storage, we prepared a fresh BH from frozen brain tissue. Fresh BH showed clear RT-QuIC positivity in the dilutions of 5 × 10 −6 and 5 × 10 −7 ( Figure 3A). The max ThT fluorescence was lower than in other tested sCJD BHs, but clearly above the threshold.
The VPSPr BH analyzed by RT-QuIC revealed clear positivity in all four wells, but the max ThT fluorescence at a dilution of 5 × 10 −6 (92753 AU) was notably lower than in the other TSE cases ( Figure 4) and the time to threshold was markedly longer (Figure 2A).
The RT-QuIC analysis of all gCJD E200K-archived BHs (n = 15) led to well-shaped curves and short time to threshold values ( Figure 2C). The E200K BHs produced significantly higher max ThT fluorescence than sCJD BHs at dilutions 5 × 10 −8 and 5 × 10 −9 ( Figure 5A). The mean time to threshold was also significantly shorter for E200K BHs than for sCJD BHs (p < 0.0001, Figure 5B  The one sCJD VV2 case gave negative results in three independent RT-QuIC assays utilizing two different archived BH aliquots. The negative results of the RT-QuIC assay of the case were unequivocal with characteristic spongiform changes and presence of proteinase K-resistant prion protein, visualized both by immunohistochemistry and western blot at the time of sampling (not shown). RT-QuIC analysis of the corresponding archived CSF samples confirmed the presence of prions ( Figure 3B); however, the shape of the curve representing the ThT fluorescence signal was unusual, suggesting either the potential presence of inhibitors or spontaneous aggregation. The 10× dilution of the CSF sample led to clear positivity with a typical curve shape ( Figure 3B). To demonstrate that the negative RT-QuIC results were likely caused by BH degradation during long-term storage, we prepared a fresh BH from frozen brain tissue. Fresh BH showed clear RT-QuIC positivity in the dilutions of 5 × 10 −6 and 5 × 10 −7 ( Figure 3A). The max ThT fluorescence was lower than in other tested sCJD BHs, but clearly above the threshold. Taken together, the analysis of the archived post-mortem BH samples of patients with a definite diagnosis of TSE (n = 39), OND patients (n = 28), and normal corneal donors (n = 25) led to 100% specificity and 97.4% sensitivity of the second-generation RT-QuIC assay.

Analysis of Archived CSF Samples with the RT-QuIC Assay
We analyzed three archived CSF samples of each sCJD type (MM1, MV1, VV1, MM2, MV2, and VV2) and three of gCJD E200K and GSS P102L (n = 24). Control CSF samples of OND patients (n = 28) were used for determination of the threshold (33,166 AU). The RT-QuIC assay with undiluted CSF samples of TSE patients provided unconvincing results. Of the 24 samples, only 13 samples gave a typical response curve ( Figure 6A). The VPSPr BH analyzed by RT-QuIC revealed clear positivity in all four wells, but the max ThT fluorescence at a dilution of 5 × 10 −6 (92753 AU) was notably lower than in the other TSE cases ( Figure 4) and the time to threshold was markedly longer (Figure 2A).
The RT-QuIC analysis of all gCJD E200K-archived BHs (n = 15) led to well-shaped curves and short time to threshold values ( Figure 2C). The E200K BHs produced significantly higher max ThT fluorescence than sCJD BHs at dilutions 5 × 10 −8 and 5 × 10 −9 (Figure 5A). The mean time to threshold was also significantly shorter for E200K BHs than for sCJD BHs (p < 0.0001, Figure 5B). The codon 129 polymorphism (MM, MV, VV) in the E200K group did not have any significant effect on max ThT fluorescence or time to threshold values (not shown). Archived BH samples of GSS P102L patients (n = 3) were all positive by RT-QuIC, with max ThT fluorescence values similar to those of the sCJD or E200K groups (Figure 4). Two additional BHs of gCJD cases with mutations D178N (MV2) and R208H (VV2) were also clearly positive ( Figures 2C and 4).    Average of the max ThT fluorescence of the archived brain homogenates in all patient groups. Average fluorescence maxima from quadruplicates of all RT-QuIC-tested BHs diluted by 10 −6 . Bars represent the arithmetic mean of the group and error bars represent the standard deviation. The threshold is indicated by a dashed line. All control samples are below the threshold of the assay. One case of sCJD VV2 was diagnosed as negative by RT-QuIC. The VPSPr case had lower max ThT fluorescence; however, it was clearly above the threshold.

Analysis of Archived CSF Samples with the RT-QuIC Assay
We analyzed three archived CSF samples of each sCJD type (MM1, MV1, VV1, MM2, MV2, and VV2) and three of gCJD E200K and GSS P102L (n = 24). Control CSF samples of OND patients (n = 28) were used for determination of the threshold (33,166 AU). The RT-QuIC assay with undiluted CSF samples of TSE patients provided unconvincing results. Of the 24 samples, only 13 samples gave a typical response curve ( Figure  6A).
The rest of the samples had a low max ThT fluorescence value with an unusually long time to threshold, and five samples were negative ( Figure 7A,B). The sensitivity of the assay with undiluted archived CSF samples was 79.2%. The protein content of archived post-mortem CSF samples was heterogeneous and notably higher than the physiological CSF protein concentration (3.8 ± 2.3 g/L vs. 0.2-0.4 g/L), suggesting the possible presence of RT-QuIC inhibitors. Repetition of the analysis with 10 times diluted CSF samples provided much more convincing results. The threshold, calculated using diluted OND CSF samples, increased to 76,436 AU. Of the 24 TSE CSF samples, 23 produced typical positive curves ( Figure 6B). The max ThT fluorescence of the samples with previous poor response substantially increased and the time to threshold dramatically shortened ( Figure 7C,D). Only one GSS sample was assessed as negative, with response just below the threshold. The sensitivity of the assay with diluted archived CSF samples increased to 95.8%. The dependence of RT-QuIC results on CSF protein concentration is shown in Figure 8. The samples with a protein content above 2.5 g/L exhibited inhibitory effects in the assay. The dilution of the CSF samples prevented RT-QuIC inhibition without compromising the assay sensitivity.  The rest of the samples had a low max ThT fluorescence value with an unusually long time to threshold, and five samples were negative ( Figure 7A,B). The sensitivity of the assay with undiluted archived CSF samples was 79.2%. The protein content of archived post-mortem CSF samples was heterogeneous and notably higher than the physiological CSF protein concentration (3.8 ± 2.3 g/L vs. 0.2-0.4 g/L), suggesting the possible presence of RT-QuIC inhibitors. Repetition of the analysis with 10 times diluted CSF samples provided much more convincing results. The threshold, calculated using diluted OND CSF samples, increased to 76,436 AU. Of the 24 TSE CSF samples, 23 produced typical positive curves ( Figure 6B). The max ThT fluorescence of the samples with previous poor response substantially increased and the time to threshold dramatically shortened ( Figure 7C,D). Only one GSS sample was assessed as negative, with response just below the threshold. The sensitivity of the assay with diluted archived CSF samples increased to 95.8%. The dependence of RT-QuIC results on CSF protein concentration is shown in Figure 8. The samples with a protein content above 2.5 g/L exhibited inhibitory effects in the assay. The dilution of the CSF samples prevented RT-QuIC inhibition without compromising the assay sensitivity.

Discussion
The second-generation RT-QuIC, also known as the IQ assay (Improved QuIC), is a version of the assay with increased sensitivity and significantly shorter time of detection [10]. Almost all human studies utilizing second-generation RT-QuIC have used CSF as the analyte [8][9][10][11][15][16][17][18][19][20][21]. In our retrospective study, we evaluated the second-generation RT-QuIC assay for the analysis of archival post-mortem BH and CSF samples within a TSE patient cohort typical of the Czech Republic [22,23]. In addition to samples of different types of sCJD, we included a large group of the most common gCJD in the Czech Republic, E200K. Other studies including gCJD with mutations D178N or R208H and GSS with P102L mutation are comparatively rare [24]. The archived BH samples were originally prepared for confirmation of the presence of PK-resistant PrP Sc by western blot. The samples were stored as 10% BH in lysis buffer containing detergents at −80 • C for 2 to 15 years. To estimate the diagnostic accuracy of the assay, we selected two different control groups, in order to prevent misinterpretation of the results. A group of patients with other neurodegenerative disorders was used as the most relevant differential diagnostic group and healthy corneal donors served as a tool for the identification of potential OND false positivity; however, both control groups provided similar data, with nearly identical calculated threshold values and without false positive results. This is an important observation, suggesting that, in the case of archival BH samples, negative specimens have the same background whether they come from healthy controls or neuropathologically proven neurodegeneration. All tested BH samples with definite diagnosis of TSE, except for one sCJD VV2 case, were RT-QuIC positive. Two different aliquots of the sCJD VV2 archived BH were repeatedly RT-QuIC negative, but the archived CSF gave a positive result. The false-negative BH was stored for 7 years; however, other much older samples tested positive, with no apparent effect of storage time on the test results. The archival BH was positive for the presence of PK-resistant PrP Sc by western blot in 2013, but not in 2021, suggesting that the negative RT-QuIC result was, indeed, caused by sample degradation. Freshly prepared sCJD VV2 BH showed positivity both in RT-QuIC and western blot. The RT-QuIC assay of 92 post-mortem archival BH samples showed 97.4% sensitivity and 100% specificity, comparable to previously published results [12]. Our data suggest that BHs originally prepared for definitive post-mortem diagnosis and frozen for several years are generally suitable for RT-QuIC retrospective studies; however, the presence of one negative sample confirmed that long-term storage of BH, in the form of detergent lysate, may occasionally lead to a loss of RT-QuIC seeding activity.
We observed higher variability, lower max ThT fluorescence, and longer time to threshold in the type-2 sCJD compared to type-1 sCJD BH samples, suggesting that the PrP Sc glycotype may affect the seeding potential. However, no effect of the sCJD glycotype alone on the seeding potential of CSF has been previously reported [13]. On the other hand, the higher seeding ability of the gCJD E200K samples, in comparison to sCJD BH samples, was in agreement with published CSF results [25]. Interestingly, the difference between the gCJD E200K and sCJD groups was not evident at the initial 5 × 10 −6 dilution and manifested at higher BHs dilutions, especially at a dilution of 5 × 10 −8 .
A comparison of the time to threshold parameter demonstrated the difference between the gCJD E200K and sCJD groups at all dilutions. The positivity manifested much faster in the gCJD E200K group and the variability was very low compared to the sCJD group. Our data suggest that the seeding activity of gCJD E200K present in archived post-mortem BHs was higher than in sCJD samples; however, the cause of this difference has to be elucidated. The greatest variability of results was observed in the GSS group. This phenomenon has been described in the literature and was also evident in our archival samples [14,26]. On the positive side, all archived GSS BHs tested positive, even after long-term storage. All tested TSE samples with a positive result had maximal ThT fluorescence values safely above the threshold, except for the case of VPSPr; however, even this sample was clearly above the threshold [25].
In most of the cases included in our BH study, we also had archival post-mortem CSF samples. The main problem with the analysis of post-mortem CSF samples was the excessive protein concentration connected with inhibition of the RT-QuIC assay and false negative results. The high concentration of proteins present in post-mortem CSF has already been reported [27]. Simple dilution of the CSF samples allowed for detection of their seeding activity in almost all TSE samples, with 95.8% sensitivity and 100% specificity. A possible inhibitory effect has been described in a CSF sample with a total protein concentration higher than 0.45 g/L [28]. Tze How Mok et al. showed, using post-mortem CSF samples, that only 70 of 79 sCJD samples and 9 of 20 various inherited prion diseases were positive, using undiluted post-mortem CSF in RT-QuIC with bank vole prion protein as substrate [29]. The lower sensitivity could be due to the high protein concentration in the post-mortem samples. We did not observe inhibition of seeding up to the concentration of 2.5 g/L of total protein. Above this value, there was a significant reduction in the sensitivity of the assay. In our assay, the threshold determined from the diluted CSF negative controls was twice as high as the threshold from the undiluted controls. This phenomenon is important to consider when testing CSF samples, where excessive amounts of protein can lead to a false negative result. Furthermore, when the threshold for diluted samples is calculated using undiluted high protein negative controls, false positive results can occur. It must be noted that, despite great care and proper technique, the post-mortem sampling of ventricular CSF is inherently associated with a risk of CSF contamination by brain tissue. Although the level of contamination is likely to be negligible, its effect on RT-QuIC results cannot be ruled out, and the 100% sensitivity reached in our post mortem CSF study should be interpreted with caution.
To sum up the most important result of our validation study, we proved that archival post-mortem BHs or CSF samples can be easily analyzed by RT-QuIC, regardless of the storage duration and the detergents present in the storage buffer, without interfering with the assay. The second-generation RT-QuIC assay with shortened rSHa PrP(90-231) substrate utilized in our study was able to detect seeding activity in all types of human TSEs tested, including rare gCJD (D178N, R208H), GSS P102L, and VPSPr cases. Even though the number of samples in the experimental groups was relatively low, the observed differences in the max ThT fluorescence and time to threshold among the diverse TSEs suggest that further refinement of the RT-QuIC assay may lead to the development of diagnostic laboratory tools capable of distinguishing specific types of human prion diseases.

Ethics Statement
The study was reviewed and approved by the Ethics Committee of the Institute of Clinical and Experimental Medicine and Thomayer Hospital in Prague, Czech Republic (approval no. G-17-06-28).

Prion Disease Samples
All samples were provided by the National Reference Laboratory for Diagnostics of Human TSE/CJD, Thomayer University Hospital, Prague, Czech Republic. From sporadic patients, we selected three different cases from each sCJD type, according to the codon 129 polymorphism and glycotype (MM1, MV1, VV1, MM2, MV2, VV2).