Streamlined alpha-synuclein RT-QuIC assay for various biospecimens in Parkinson’s disease and dementia with Lewy bodies

All human samples were received as de-identified postmortem cases originated from the respective sites and collected under the guidelines of local ethics committees. Unless otherwise specified, frozen brain tissue and CSF samples for the development of αSyn RT-QuIC assay were obtained from the NIH NeuroBioBank (NBB), a centralized network of biorepositories. Brain tissue samples included 3 cases each of PD and DLB with confirmed neuropathology and 3 cases of NS controls archived at NBB. A collection of 214 CSF samples from NBB included cases of PD (n = 88), DLB (n = 58) and NS controls (n = 68) including those from the neurologically normal (n = 23), amyotrophic lateral sclerosis (ALS, n = 9), multiple sclerosis (MS, n = 6), Alzheimer’s disease (AD, n = 7), Pick’s disease (n = 10), corticobasal degeneration (CBD, n = 4), and progressive supranuclear palsy (PSP, n = 9). The demographic data of CSF cases examined for diagnostic validation by RT-QuIC analyses are listed in Table 1. Multiple tissue panels available from the same cadavers included a case of neuropathologically confirmed PD (case number 6164) and a NS control affected by AD (case number 5687) with samples of the brain, scalp skin, and CSF obtained from NBB (used for Fig. 4a), a case of neuropathologically confirmed PD (case number 2054) and a NS control without any neurological disease (case number 2074) with samples of scalp skin, submandibular gland (SMG), sigmoid colon, and CSF obtained from the Banner Sun Health Research Institute (BSHRI, Sun City, Arizona, USA) (used for Fig. 4b), and a case of neuropathologically confirmed PD (case number 2052) and a NS control without any neurological disease (case number 2078) with samples of scalp skin, SMG, sigmoid colon obtained from BSHRI (used for Fig. 5).

All human biospecimens were received on dry ice and were stored at -80 °C before sample preparation. Frozen brain samples were lysed in ice-cold lysis buffer [Dulbecco’s phosphate-buffered saline (PBS) supplemented with 1% Triton X-100, 150 mM NaCl, 5 mM EDTA, and Roche mini-cOmplete™ proteinase inhibitors (Roche Diagnostics, Indianapolis, IN, USA)], followed by homogenization in the presence of zirconia beads (1 mm) in a mini-Beadbeater-16 device (BioSpec Products, Bartlesville, OK, USA) for 5 cycles of 1 min beating and 3 min cooling at 4 °C. Following a brief centrifugation for 5 min at 500 × g, a 10% brain homogenate (w/v) was prepared and stored at − 80 °C until use. For all peripheral tissues, previous studies included a 4-h enzymatic digestion step during tissue extraction [20, 23]. This step was eliminated in our protocol without noticeable changes in extraction efficiency. Frozen samples of scalp skin were thawed, and washed in ice-cold PBS at least three times until no visible blood remained. The skin tissue was minced with a razor blade, followed by homogenization in lysis buffer in the presence of zirconia beads as above. A 10% skin homogenate (w/v) was prepared and stored at -80 °C until use. Frozen samples of SMG and sigmoid colon were prepared in the same manner as the skin, as described above. A 10% homogenate (w/v) of SMG or colon was prepared and stored at -80 °C until use. Frozen samples of CSF were thawed and used immediately before assay without further processing.

All reagents used for αSyn RT-QuIC assay were obtained from commercial sources. Sodium phosphate (0.5 M, pH 8.0) was from Boston BioProducts (Ashland, MA, USA), sodium chloride (5 M) was from Invitrogen (Carlsbad, CA, USA), thioflavin T (ThT) was from Sigma (St. Louis, MO, USA). HPLC-grade water from Thermo Fisher (Waltham, MA, USA) was used to prepare solutions. Aliquots of individual stock solutions were stored at −20 °C until use. Lyophilized recombinant human αSyn (rec-Syn) were purchased from rPeptide (Watkinsville, GA, USA) with catalog number S-1001–02 and lot numbers 082517AS (lot 1) and 111317AS (lot 2), and stored at −20 °C until use.

For tissue samples, an initial tissue homogenate [10% (w/v), defined as 10^–1 dilution] as prepared above was subjected to serial tenfold dilutions with sample diluent containing PBS supplemented with Gibco 1 × N2 supplement (Thermo Fisher, Waltham, MA, USA). RT-QuIC reactions were performed in Nunc black 96-well plates with optical flat bottom (Thermo Fisher, Waltham, MA, USA). Each well was preloaded with six 0.8 mm low-binding silica beads (OPS Diagnostics, Lebanon, NJ, USA). Lyophilized rec-Syn (rPeptide) was reconstituted in HPLC-grade water to 1 mg/ml and filtered through Amicon 100 kDa filters (Millipore, Burlington, MA, USA) by centrifugation for 10 min at 4 °C. To measure αSyn RT-QuIC seeding activity, 2 μl of diluted tissue homogenate was added to individual wells containing 98 μl of RT-QuIC reaction mixture composed of 40 mM NaPO4 (pH 8.0), 170 mM NaCl, 20 µM ThT, 0.1 mg/ml rec-Syn. The plates were sealed with Nunc clear sealing film (Thermo Fisher, Waltham, MA, USA) and incubated at 42 °C in a BMG FLUOstar Omega plate reader (BMG Labtech, Cary, NC, USA) with cycles of 1 min shaking (400 rpm, double orbital) and 1 min rest throughout the assay. ThT fluorescence (450 nm excitation and 480 nm emission; bottom read) was recorded every 45 min for 60 h. For CSF samples, 2 μl of neat CSF or CSF serially diluted with sample diluent was added to individual wells containing 98 μl of RT-QuIC reaction mixture supplemented with 0.0005% SDS. Four replicate reactions were made for each sample. The plate reader was set up with optic gain setting (~ 1900) that would give a negative control baseline of ThT fluorescence around 15,000 relative fluorescence units (rfu). The maximal fluorescence response of the plate reader was fixed at 260,000 rfu. Raw data with ThT fluorescence intensity in rfu were normalized to a percentage of the maximal fluorescence at 260,000 rfu (defined as 100%). Positive RT-QuIC reactivity of individual wells is defined as enhanced ThT fluorescence above a predefined threshold within 60 h. This threshold was calculated as the average background fluorescence plus 5 standard deviations, equal to ~ 30,000 rfu or ~ 11% of maximum ThT response in our experiments. For RT-QuIC analyses, data from four technical replicates were examined for a given biospecimen. A biospecimen was considered positive when at least two out of the four replicates displayed positive ThT reactivity above the threshold. For each positive sample, the average fluorescence intensity from the positive replicates was calculated and plotted against time. For each negative sample, the average fluorescence intensity from the negative replicates was calculated and plotted against time. The lag phase of RT-QuIC reactions was defined as the time (h) to reach the threshold as defined above. Protein aggregation rate (PAR) was calculated as the inverse of lag-phase (1/h). For Spearman-Kärber analyses of end-point dilution RT-QuIC experiments, tenfold serial dilutions of a biospecimen were prepared. Each diluted sample was used to seed RT-QuIC reactions in quadruplicate. ThT fluorescence from individual replicates for each dilution were plotted against time. The dilution series continued until RT-QuIC reactivity reached ≤ 1 out 4 replicates. The seeding dose giving ThT positivity in 50% of replicate wells (SD₅₀) based on the Spearman-Kärber equations was calculated as described [24, 25].

The statistical comparisons were performed using t-tests and one-way ANOVA for comparing two or more groups, respectively (GraphPad Prism Software Version 9.0.0). Data were expressed as means ± SD (standard deviation) or SE (standard error of the mean). A two-sided type I error level of 0.05 was adopted. The statistically significant differences were expressed as * p < 0.05, ** p < 0.01, *** p < 0.005, and **** p < 0.001 or p < 0.0001, as indicated. The number of biological replicates was expressed as “n”, and 4 technical replicates was used for each biological sample.

Our streamlined RT-QuIC assay protocol was developed based on the second-generation platform of αSyn RT-QuIC as described by Caughey and colleagues [15], with the following modifications. We acquired all assay reagents from commercial sources, including the monomeric rec-Syn protein, the substrate present in vast excess for in vitro conversion into amyloid fibrils by seeding with minute quantities of αSyn^D aggregates from a clinical sample to be examined. This approach eliminated the need for the laborious and lengthy (~ 5–7 days) in-house production and purification of rec-Syn by highly specialized personnel, and minimized the variability in the quality of rec-Syn produced [15]. Following a screen of rec-Syn products from several sources, we found that a commercial product of monomeric human wild-type rec-Syn (rPeptide) was most suitable for this RT-QuIC assay (see Methods section). Two separate lots of the commercial rec-Syn protein were successfully used for the RT-QuIC assay with ~ 300 reaction plates over a two-year period. Accordingly, our simplified αSyn RT-QuIC assay involved minimal hands-on time (~ 1 h) to rapidly assemble all reagents for RT-QuIC reactions in a 96-well plate, sufficient for testing 24 samples in quadruplicate. Moreover, we markedly shortened the preanalytical processing of peripheral tissues by eliminating a 4-h enzymatic digestion step used in previous studies [20, 23]. In addition, the ThT concentration was increased from 10 µM to 20 µM to minimize self-quenching of ThT fluorescence [26]. Finally, we found that a single streamlined RT-QuIC assay protocol worked equally well for several different types of tissues including the brain, skin, salivary gland, and colon.

To verify that our streamlined RT-QuIC assay is able to detect the seeding activity of αSyn^D, we first examined brain homogenate (BH) prepared from neuropathologically confirmed cases of PD (n = 3), DLB (n = 3), and from NS controls (n = 3). The RT-QuIC seeding activity of αSyn^D aggregates in biospecimens was measured as time-dependent increases in ThT fluorescence intensity above background threshold [14, 15]. As expected, αSyn^D seeding activity was readily detected using our streamlined RT-QuIC assay. RT-QuIC reactions seeded by 2 µl of BH from cases of PD and DLB serially diluted to 10^–3 through 10^–8 showed varying degrees of ThT responses within 60 h (Fig. 1a). In contrast, no seeding activity was observed with BH from NS controls at dilutions of 10^–3 through 10^–6 (Fig. 1a). RT-QuIC reactions seeded by BH showed a lag phase ranging from 10 to 50 h, with DLB cases displaying shorter lag phase than PD cases at respective dilutions (Fig. 1b). To better characterize the RT-QuIC kinetics, protein aggregation rate (PAR), defined as the inverse of the lag phase (1/h), was used for measuring the rate of seeding activity at each dilution [27]. A dose-dependent increase in PAR was observed in BH-seeded RT-QuIC reactions from both PD and DLB cases (Fig. 1c). Moreover, PAR was significantly higher in DLB than PD across several orders of magnitudes in BH dilutions (Fig. 1c), consistent with higher levels of αSyn^D aggregates known to be present in brains of DLB patients as compared to those with PD [28]. Our results are in agreement with those in the published study of the second-generation αSyn RT-QuIC assay using in-house generated rec-Syn preparations [15], with similar RT-QuIC kinetics (within 60 h) at comparable dilutions of BH from PD and DLB. Taken together, these results confirmed the feasibility of using commercially available rec-Syn for an ultrasensitive RT-QuIC assay of αSyn^D in million-fold diluted BH of PD and DLB.

After we successfully detected αSyn^D in highly diluted BH of PD and DLB cases, we next evaluated the sensitivity of our RT-QuIC assay for postmortem CSF samples. As reported previously, addition of a small amount of SDS (0.0015%) is necessary to bring the CSF-seeded RT-QuIC reactions to completion within 60 h [15]. In our protocol, the level of SDS was further reduced to 0.0005% (5 ppm) (see Methods section). CSF samples were acquired from neuropathologically confirmed cases of PD (n = 4), DLB (n = 4), as well as NS controls (n = 2), and 2 µl of CSF either undiluted or serially diluted to 10^–1-10^–3, equivalent to 2–0.002 µl of CSF, respectively, was used to seed RT-QuIC reactions. Seeding activity was detectable in PD and DLB cases but not in controls at all CSF levels, and remarkably, even in as little as 0.002 µl of CSF (Fig. 2a). The lag phase varied from 12 to 45 h in CSF-seeded RT-QuIC reactions, with a shorter lag phase in DLB than PD in RT-QuIC reactions seeded by 0.02–2 µl CSF (Fig. 2b). Accordingly, PAR was significantly higher in DLB than in PD at the respective CSF levels (Fig. 2c). Interestingly, for both PD and DLB, PAR increased when seeding went from 0.002 µl to 0.02 µl of CSF, but plateaued at 0.2 µl CSF and even decreased at 2 µl of undiluted neat CSF (Fig. 2b). Thus, non-linear RT-QuIC responses seemed to occur at higher seeding doses of CSF. This is likely due to the presence of putative inhibitors in undiluted samples that may suppress initial RT-QuIC reactions which subsequently rebound upon dilutions, consistent with the previous observations as reported [29, 30]. In summary, our streamlined RT-QuIC assay enabled ultrasensitive detection of αSyn^D in PD and DLB cases using only nanoliters of CSF.

To further evaluate the diagnostic utility of the streamlined αSyn RT-QuIC assay, we acquired a large number of postmortem CSF samples from neuropathologically confirmed cases of PD, DLB, and NS controls from the NIH NeuroBioBank (NBB), a centralized network of biorepositories. All CSF samples were evaluated at the 0.2 µl seeding level (2 µl of tenfold dilution of neat CSF) for RT-QuIC reactions in quadruplicate with a throughput of 24 samples (including positive and negative controls) in a single 96-well plate per assay. We first examined the RT-QuIC profile in a set of 40 cases of PD and 40 NS controls. CSF samples from PD cases displayed positive RT-QuIC responses within ~ 18–50 h (Fig. 3a). For comparison, we tested another set of 30 cases each for DLB and NS controls. RT-QuIC of CSF samples from DLB cases exhibited strong ThT responses within ~ 15–30 h (Fig. 3b). NS controls showed negative reactivity below the threshold within 60 h. The overall RT-QuIC kinetics of CSF samples from PD cases exhibited a slower and broader RT-QuIC reactivity as compared to that of DLB (Fig. 3a vs. Figure 3b), as observed in a previous study of a small cohort of CSF samples from PD and DLB patients [15], suggesting a differential pattern of RT-QuIC kinetics between CSF samples from PD and DLB patients.

In total, our RT-QuIC assay examined 214 postmortem CSF samples from 88 cases of PD, 58 cases of DLB, and 68 cases of NS controls [neurologically normal (n = 23), ALS (n = 9), MS (n = 6), AD (n = 7), Pick’s disease (n = 10), CBD (n = 4), and PSP (n = 9)]. Analysis of the CSF seeding activity at the end of the 60 h RT-QuIC assay revealed an excellent analytical performance for both PD and DLB (Fig. 3c). Positive seeding activity was observed in 86 out of 88 CSF samples in the PD group and 57 out of 58 CSF samples in the DLB group, yielding a sensitivity of 98% for both PD and DLB. None of the 68 NS control CSF samples showed positive seeding activity, resulting in a specificity of 100%. In comparison, the overall CSF seeding activity was significantly higher in DLB than PD (p < 0.05), and the combined PD and DLB group was well above NS controls (p < 0.0001). Taken together, these results confirm that CSF is a useful specimen for validating the RT-QuIC assay as a robust tool for the diagnosis of PD and DLB.

Finally, we examined whether our streamlined RT-QuIC assay can be used for different types of specimens without the need to modify assay conditions. Several types of biospecimens were collected from the same PD cadavers, subjected to the same processing and serial dilution procedure, and tested under the same RT-QuIC assay conditions. Spearman-Kärber endpoint dilution analysis was used to determine the seeding dose at which 50% of replicate reactions were positive, termed SD₅₀ [24]. In a neuropathologically confirmed PD case from whom scalp skin was available in addition to the brain and CSF, we performed parallel RT-QuIC reactions seeded with the respective specimens in tenfold dilution series (from 10^–1 through 10^–8). As shown in Fig. 4a, αSyn^D seeding activity displayed dose-dependent titration in serially diluted PD samples of BH, skin homogenate, and CSF, respectively. The Spearman-Kärber analysis of this PD case yielded SD₅₀ values at 2.8 × 10⁶/mg for the brain, 5.0 × 10⁴/mg for the skin, and 2.8 × 10³/µl for CSF. No seeding activity was detected in corresponding biospecimens from a control cadaver affected by AD (Fig. 4a). To explore more broadly the tissue distribution of αSyn^D, another case of neuropathologically confirmed PD was examined from which scalp skin, submandibular salivary gland (SMG), and sigmoid colon were available in addition to CSF. Using the same RT-QuIC assay protocol, αSyn^D seeding activity was examined in serially diluted tissue homogenates of skin, SMG, and colon, and in serially diluted CSF (Fig. 4b). SD₅₀ values obtained from Spearman-Kärber analysis of this PD case were 5.0 × 10⁴/mg for skin, 8.9 × 10⁵/mg for SMG, 8.9 × 10⁴/mg for colon, and 2.8 × 10²/µl for CSF. In contrast, no seeding activity was found in corresponding specimens from a neurologically normal control case (Fig. 4b). Therefore, the αSyn^D seeding activity in PD peripheral tissues of the skin, SMG, and colon was relatively high, ranging from ~ 10⁴ to ~ 10⁵ per mg of tissues, which was only 1 to 2 orders of magnitude lower than that in the brain (~ 10⁶/mg). The seeding activity in PD CSF was in the range of 10²–10³ per µl. For brain and CSF, the previously reported SD₅₀ values in PD and DLB were in the range of 10⁵–10⁶/mg and 4–54/µl, respectively [15]. Thus, our streamlined RT-QuIC assay provides a single unified protocol for robust tracking of αSyn^D seeding activity across multiple biospecimens from the same patients affected by PD and other synucleinopathies.

To evaluate the reliability of our streamlined RT-QuIC protocol, we analyzed tissue homogenates from a third case of neuropathologically confirmed PD using different batches of rec-Syn substrate. Distinct patterns of RT-QuIC kinetics from reactions seeded by PD tissue homogenates from the skin, SMG, and colon were consistently reproduced when performed with two different lots of rec-Syn (Fig. 5). In contrast, only negative reactivity was observed in corresponding tissue homogenates from a neurologically normal control case, confirming the specificity of the RT-QuIC reactions (Fig. 5). Therefore, our streamlined RT-QuIC assay enables reliable and robust analysis of αSyn^D seeding activity in various patient-derived specimens.