Polyglutamine-Expanded Ataxin-3: A Target Engagement Marker for Spinocerebellar Ataxia Type 3 in Peripheral Blood
Relevant conflicts of interest/financial disclosures: K.K. and J.P. are members of the company Evotec, which generated and validated the SMC ataxin-3 immunoassay described in this study. Both were blinded for genotypes and clinical data of all participants at each time and did not influence any project hypothesis or outcome measures. B.W. is supported by ZonMW, Hersenstichting, uniQure, and Gossweiler Foundation; has served on the scientific advisory board of uniQure; and collaborates within a research consortium with Vico Therapeutics. P.G. received funding from Reata Pharmaceutical, Vico Therapeutic, and Triplet Pharmaceutical. She has served on the advisory board of Triplet, Vico, and Reata. T.K. received personal fees from Roche, UCB, uniQure, and Vico Therapeutics, all activities outside the submitted work.
Funding agencies: This project is supported by the EU Joint Programme—Neurodegenerative Disease Research (JPND) through the following funding organizations under the aegis of JPND: Germany, Federal Ministry of Education and Research (BMBF; funding codes 01ED1602A/B); Netherlands, The Netherlands Organisation for Health Research and Development; Portugal, Foundation for Science and Technology (FCT, grant number JPCOFUND/0001/2015), and Regional Fund for Science and Technology of the Azores; and United Kingdom, Medical Research Council. This project has received funding from the European Union's Horizon 2020 research and innovation program under grant agreement number 643417. In addition, support has been received by the BIONIC project (number 733050822, which has been made possible by ZonMW as part of “Memorabel,” the research and innovation program for dementia, as part of the Dutch national “Deltaplan for Dementia”: zonmw.nl/dementiaresearch), the CAFÉ project (the National Institutes of Health, USA, grant number 5R01NS104147-02), and a grant from the Selfridges Group Foundation (NR170024). The BIONIC project is a consortium of Radboudumc, LUMC, ADX Neurosciences, and Rhode Island University.
Abstract
Background
Spinocerebellar ataxia type 3 is a rare neurodegenerative disease caused by a CAG repeat expansion in the ataxin-3 gene. Although no curative therapy is yet available, preclinical gene-silencing approaches to reduce polyglutamine (polyQ) toxicity demonstrate promising results. In view of upcoming clinical trials, quantitative and easily accessible molecular markers are of critical importance as pharmacodynamic and particularly as target engagement markers.
Objective
We aimed at developing an ultrasensitive immunoassay to measure specifically polyQ-expanded ataxin-3 in plasma and cerebrospinal fluid (CSF).
Methods
Using the novel single molecule counting ataxin-3 immunoassay, we analyzed cross-sectional and longitudinal patient biomaterials.
Results
Statistical analyses revealed a correlation with clinical parameters and a stability of polyQ-expanded ataxin-3 during conversion from the pre-ataxic to the ataxic phases.
Conclusions
The novel immunoassay is able to quantify polyQ-expanded ataxin-3 in plasma and CSF, whereas ataxin-3 levels in plasma correlate with disease severity. Longitudinal analyses demonstrated a high stability of polyQ-expanded ataxin-3 over a short period. © 2021 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society
Spinocerebellar ataxias are a heterogeneous group of dominantly inherited, progressive diseases. The most common among them globally is spinocerebellar ataxia type 3 (SCA3), also known as Machado–Joseph disease, a multisystem disorder characterized by the degeneration of spinocerebellar tracts, dentate nucleus, brainstem nuclei, and basal ganglia. It is caused by an unstable expansion of a polyglutamine (polyQ)-encoding CAG repeat in the ATXN3 gene resulting in the expression of an abnormally elongated ataxin-3 protein that is considered to be the major cause of neurodegeneration in SCA3. Currently, there is no treatment, but new approaches aiming to silence the disease gene are close to clinical trials.1
To demonstrate target engagement in such trials, the availability of ultrasensitive quantitative immunoassays to measure the concentration of polyQ-expanded ataxin-3 in body fluids is mandatory. Recently, an immunoassay that detected polyQ-expanded ataxin-3 in body fluids, including cerebrospinal fluid (CSF), and discriminated between SCA3 patients and healthy controls was developed.2 In addition, an assay based on time-resolved fluorescence energy transfer (TR-FRET) was reported, which was capable of measuring ataxin-3 concentrations in peripheral blood mononuclear cells (PBMCs) but failed to detect ataxin-3 in body fluids.3
Here, we report a novel single molecule counting (SMC) ataxin-3 immunoassay to specifically measure polyQ-expanded ataxin-3 in plasma and CSF. Using this assay, we found strong correlations between plasma polyQ-expanded ataxin-3 concentrations and clinical parameters. In a longitudinal study, we observe a high stability of polyQ-expanded ataxin-3 in pre-ataxic and ataxic mutation carriers.
Patients and Methods
Ethics and Consent to Participate
The study was approved by the local committees of all participating centers. Informed and written consent was obtained from all study participants at enrollment.
Study Participants
Blood and CSF samples were obtained from participants of the European Spinocerebellar Ataxia Type-3/Machado–Joseph Disease Initiative (ESMI) cohort. Biosamples were collected under highly standardized protocols at all participating centers. Details are provided in Table 1 and Appendix S1.
Demographic information | Controls | Pre-ataxic SCA3 | Ataxic SCA3 |
---|---|---|---|
Exploratory cohort | |||
Sample size (female) | 9 (44%) | ND | 10 (50%) |
Age (y) | 45 (39.5–58) | ND | 53 (36.7–56.5) |
Reported age at onset, AAO (y) | NA | NA | 47 (27.5–49.5) |
SARA score | 0.5 (0–1) | ND | 10.5 (7.5–15.5) |
Repeat count (long allele) | NA | ND | 67.5 (64.7–69) |
Validation cohort | |||
Sample size (female) | 15 (46.6%) | 11 (54.5%) | 45 (51.1%) |
Age (y) | 43.5 (21.1–68.2) | 35 (21–42) | 49.8 (40.2–61) |
Predicted/reported AAO (y) | NA | 44 (36.5–49.5) | 39.1 (13–68) |
SARA score | 0.2 (0–2) | 0.7 (0–1.5) | 15.3 (4–34.5) |
Repeat count (long allele) | NA | 68 (62–71) | 69 (58–73) |
CSF cohort | |||
Sample size (female) | 18 (55.5%) | 5 (80%) | 12 (66.6%) |
Age (y) | 45 (34–52) | 37 (35.5–41) | 47.2 (39.2–57.7) |
Reported AAO (y) | NA | NA | 42 (29–45) |
SARA score | 0 (0–1.2) | 1 (1–1.5) | 10.2 (8–17.5) |
Repeat count (long allele) | NA | 69.5 (69–70.7) | 69.5 (66.5–70.7) |
Longitudinal cohort | |||
Sample size (female) | 4 (50%) | 5 (80%) | 28 (53.5%) |
Participants for whom disease status changed | 0 | 3 | 2 |
Age (y) at baseline | 44.5 (23.5–65.9) | 36 (26–38.5) | 51.5 (43.7–60.2) |
Reported AAO (y): BL and FUP | NA | BL: NA FUP: 31 (26–37) | 41 (33–47.5) |
SARA score: BL and FUP | BL: 0.5 (0–1.7) FUP: 0 (0–1.5) |
BL: 1 (0.5–1) FUP: 3.5 (1–6.2) |
BL: 12.5 (6.5–17.5) FUP: 11.7 (9–19.5) |
Repeat count (long allele) | NA | 69 (65.5–70.5) | 70 (65.5–71.7) |
Time span between BL and FUP | 14.1 (11.5–15.2) | 13.2 (12.0–14.4) | 13.0 (11.3–17.7) |
Assay comparison | Meso Scale (Prudencio et al 2020) | SMC (described here) |
---|---|---|
Antibodies used | Atxn3 clone 1H9 and polyQ 3B5H10 | Atxn3 clone 1H9 and polyQ MW1 |
Applicable in | Plasma and CSF | Plasma and CSF |
Dynamic range | pg/mL | pg/mL |
LoD, LLoQ, and ULoQ | Not provided in the publication | LOD, 0.07 pg/mL; LLoQ, 0.253 pg/mL; ULoQ, 427.69 pg/mL |
AUC plasma | ||
cntrs vs. mut carrier | 1.0 | 1.0 |
Pre- vs. ataxic | 0.70 | 0.78 |
AUC CSF | ||
cntrs vs. mut carrier | 1.0 | 1.0 |
Pre- vs. ataxic | 0.89 | 0.58 |
Correlation with clinical data: CSF | None | None |
Plasma | AAO (P = 0.020) | AAO (P = 0.0003) |
Gait (P = 0.030) | SARA (P = 0.0202) | |
Not corrected for age or polyQ lengths | Corrected for age and polyQ lengths | |
Correlation between CSF and plasma polyQ-atxn3 | None (P = 0.82) | None (P = 0.45) |
Longitudinal data | ND | High stability in 1-year follow-up |
- Data are reported as median and interquartile range.
- Abbreviations: SMC, single molecule counting; SCA3, spinocerebellar ataxia type 3; ND, not determined; AAO, age at ataxia onset; NA, not applicable; SARA, Scale for the Assessment and Rating of Ataxia; CSF, cerebrospinal fluid; BL, baseline; FUP, follow-up visit; LOD, limit of detection; LLoQ, lower limit of quantification; ULoQ, upper limit of quantification; AUC, area under the curve (determination of discrimination efficiency between different genotypes); cntrs, healthy controls; mut carrier, mutation carriers including pre-ataxic and ataxic SCA3 mutation carriers; pre-, pre-ataxic; polyQ, polyglutamine.
Age at ataxia onset (AAO) in ataxic mutation carriers was defined as the reported age at onset of gait difficulties. In the pre-ataxic mutation carriers, predicted AAO was calculated based on age at recruitment and CAG repeat length.4 The Scale for the Assessment and Rating of Ataxia (SARA) was used to assess the severity of ataxia.5 Mutation carriers were classified as either pre-ataxic (SARA ˂ 3 points) or ataxic (SARA ≥ 3). The CAG repeat length of the expanded allele was determined using PCR-based fragment-length analysis.
Ataxin-3 SMC Assay
The assay employs SMC technology that provides ultrasensitivity and a wide linear detection.6 Specific detection of polyQ-expanded ataxin-3 occurs by bead-based immunoreaction with antibody combination 1H9 and MW1.7 Epitope-binding sites are shown in Figure 1A. Biomaterials were subsequently measured using the SMCxPro platform. All assays were performed by operators blinded to the genotype and clinical state of the participant. Detailed assay description is presented in Appendix S1.

Statistical Analysis
Analyte distribution was tested for normality using Shapiro–Wilk test. Nonparametric group analyses were performed using two-sided Mann-Whitney U test, with Bonferroni correction for multiple comparison. For linear correlation we used partial Spearman correlation. Data were adjusted for age and CAG repeat length. CAG repeat adjustment was included as MW1 antibody can bind 16 polyQ repeats with increasing intensity for longer repeats.7 Both age and expanded CAG repeat length were identified as covariable in our data set and as an independent modifier of SCA3 disease severity.8, 9 Multivariate analyses revealed that age at onset, disease duration, and sex did not correlate with our data sets. Therefore, all statistical analyses were corrected only for age and expanded CAG repeat length. Correlation analyses of plasma and CSF ataxin-3 levels were performed on z-transformed data sets. Effect sizes (r) were calculated as Cohen's d. Intraclass variation (ICC) was performed to analyze the stability of the analyte ataxin-3 at the longitudinal study design. To test the quality of classification of the cohort into healthy controls and mutation carriers, we calculated receiver operating characteristic (ROC) curves and determined the area under the curve (AUC).
Data are presented as median and interquartile range. Statistical significance is demonstrated by P-values (≤0.01 [**], ≤0.001 [***]).
All statistical and graphical evaluations were performed with GraphPad prism 8.0. For linear regression, we used IBM SPSS Statistics version 27.
Results
SMC Immunoassay Quantifies PolyQ-Expanded Ataxin-3 with High Specificity and Sensitivity
SMC immunoassay validation using human recombinant ataxin-3 with normal (15Q) or elongated (62Q) polyQ length demonstrated a high specificity for polyQ-expanded ataxin-3 over normal ataxin-3 (Fig. 1B). Determination of LOD (limit of detection, 0.07 pg/mL), lower limit of quantification (LLoQ, 0.252 pg/mL), and upper limit of quantification (ULoQ, 427.695 pg/mL) showed a low picomolar detection threshold and broad dynamic range (Fig. 1C). Spiking in recombinant polyQ-expanded ataxin-3 protein in human CSF or plasma from control subjects revealed a signal recovering rate of more than 99.5% and demonstrated the capacity of the assay to measure reference analytes in real human biomaterials. In the exploratory cohort, polyQ-expanded ataxin-3 plasma concentrations in SCA3 mutation carriers ranged from 18 to 87 pg/mL (59.63 pg/mL [48.93–74.37]), whereas polyQ-expanded ataxin-3 was not detectable in healthy controls (1.15 pg/mL [0.27–6.0]) (Fig. 1D).
PolyQ-Expanded Ataxin-3 Is Quantifiable in Mutation Carriers
In plasma and CSF samples from the validation cohort, polyQ-expanded ataxin-3 was quantifiable (plasma: 72.25 pg/mL [52.34–100.3], P < 0.0001, r = 0.84; CSF: 5.48 pg/mL [4.85–6.77], P < 0.0001, r = 0.869) in SCA3 mutation carriers, whereas concentrations were below the detection threshold in healthy controls (plasma: 0.14 pg/mL [0.1–0.4]; CSF: 0.11 pg/mL [0.08–0.15]; Fig. 1E,F). PolyQ-expanded ataxin-3 concentrations were higher in plasma samples of ataxic than those of pre-ataxic mutation carriers (83.30 pg/mL [55.38–106.6] vs. 53.80 pg/mL [40.28–63.37]; P = 0.009, r = 0.50) (Fig. 1E). In particular, patients with a more severe disease presented with higher ataxin-3 levels (SARA < 10: 75.80 pg/mL [41.03–95.65], SARA ≥ 10: 86.34 pg/mL [60.45–122.2]). CSF concentrations of polyQ-expanded ataxin-3 did not differ between ataxic and pre-ataxic mutation carriers. Correlation analysis failed to reveal an association between CSF and plasma ataxin-3 levels (R = 0.210, P = 0.45). No sex-specific differences of polyQ-expanded ataxin-3 protein levels were detected (Fig. 1D,F). The level of polyQ-expanded ataxin-3 in plasma and CSF perfectly discriminated between mutation carriers and healthy controls with AUC values of 1.00 in the ROC analysis. Plasma polyQ-expanded ataxin-3 showed a good discrimination ability comparing pre-ataxic and ataxic mutation carriers (AUC = 0.78) but failed for CSF (AUC = 0.58) (Fig. S1A–D, Appendix S1).
Plasma PolyQ-Expanded Ataxin-3 Level Correlates with Clinical Parameters and Remain Stable over a 1-Year Period
Plasma polyQ-expanded ataxin-3 were positively correlated with SARA (R = 0.5026, P = 0.020) (Fig. 1H), whereas it was negatively correlated with AAO (R = −0.6041, P < 0.001) (Fig. 1G): these results, however, were not replicable in CSF samples. Cross-sectional analyses of plasma polyQ-expanded ataxin-3 protein level relative to time to predicted/reported years from ataxia onset revealed a positive linear correlation (R = 0.3747, P = 0.005), demonstrating that the polyQ-expanded ataxin-3 protein levels are higher at a later stage of the disease (Fig. 1I). Longitudinal measurements of 33 mutation carriers over a 1-year period revealed a high stability of polyQ-expanded ataxin-3 in SCA3 mutation carriers, including three mutation carriers that converted from the pre-ataxic to the ataxic stage (ICC = 0.848 [0.693–0.925]) (Fig. 1J). Only at later disease stages were higher protein levels observed after a 1-year period (Fig. 1K).
Discussion
To demonstrate target engagement in future trials that aim at silencing the SCA3 disease gene, the availability of an ultrasensitive, quantitative immunoassay to measure the concentration of polyQ-expanded ataxin-3 in body fluids is mandatory.1
Here, we report on the successful generation and validation of a new ultrasensitive and quantitative immunoassay to specifically measure low concentrations (pg/mL) of polyQ-expanded ataxin-3 in human biofluids like blood plasma and CSF. Our SMC immunoassay perfectly discriminated between healthy controls and SCA3 mutation carriers, yielding discrimination values like a recent published ataxin-3-specific mesoscale assay.2 In addition, our assay allowed for a discrimination between pre-ataxic and ataxic mutation carriers in plasma. Moreover, polyQ-expanded ataxin-3 plasma levels correlated with the clinical features of the disease, namely SARA, suggesting that our assay might indeed quantify polyQ-expanded ataxin-3 in a way that reflects the severity of ataxia of SCA3. These findings extend our pilot study where we used a TR-FRET-based technique to quantify ataxin-3 in PBMCs,3 by demonstrating that polyQ-expanded ataxin-3 protein serves as a biomarker even in plasma and possibly CSF.
We did not find an association of polyQ-expanded ataxin-3 levels in plasma and CSF. Therefore, the pool of polyQ-expanded ataxin-3 in CSF differs from that in peripheral blood and blood cells, as reported earlier.2 This notion is further supported by the observation that the levels of polyQ-expanded ataxin-3 were >10 times higher in plasma as in CSF.
So far, neither our SMCTM nor the mesoscale-based immunoassay (comparison of both immunoassays in Table 1) showed any association between CSF polyQ-expanded ataxin-3 protein levels and clinical features of the disease. This could be explained by the assumption that CSF polyQ-expanded ataxin-3 represents a—rather disease-stage-independent—trait biomarker of the disease, as demonstrated for the respective key proteins of other neurodegenerative diseases, for example, C9orf72 dipeptides in ALS/FTD.10 Alternatively, it might be due to the sample size and composition of our SCA3 subject group. The total number of CSF samples, in general, was low. This applied even more to CSF from patients in the later stages of the disease, in whom we found higher plasma concentrations of polyQ-expanded ataxin-3. The lack of correlation of CSF polyQ-expanded ataxin-3 concentrations and disease severity, however, does not call into question the potential usefulness of CSF polyQ-expanded ataxin-3 that might serve as a target engagement biomarker in future interventional trials that investigate gene-silencing approaches.
Our longitudinal analyses revealed a high stability of plasma-derived polyQ-expanded ataxin-3 protein levels over a period of 1 year. If confirmed in a larger cohort and longer time period, the high degrees of stability of this biomarker would allow to reduce sample sizes in trials that include polyQ-expanded ataxin-3 as one of the endpoints.
In conclusion, our novel SMC immunoassay is able to quantify polyQ-expanded ataxin-3 in plasma and CSF while ataxin-3 levels in plasma correlating with disease severity. First longitudinal analyses demonstrated a high stability of polyQ-expanded ataxin-3 over a period of 1 year. Therefore, this immunoassay has the potential to support the clinical development of therapeutic drugs in SCA3, allowing to determine the levels of polyQ-expanded ataxin-3 as a target engagement biomarker in human biofluids, especially peripheral blood.
Acknowledgments
We thank Jonasz J. Weber for providing recombinant ataxin-3 protein as well as Alexandra Grenzendorf, Yvonne Schelling, and Melanie Kraft for excellent technical assistance. Open access funding enabled and organized by Projekt DEAL.
Open Research
Data Availability Statement
All data are available within the manuscript and the supplementary material
K.K. and J.P. are members of the company Evotec, which generated and validated the SMC ataxin-3 immunoassay described in this study. Both were blinded for genotypes and clinical data of all participants at each time and did not influence any project hypothesis or outcome measures. J.H.-S. received support from the National Ataxia Foundation, Center of Rare Diseases Medical Faculty, and the excellence program Athene of the University of Tübingen. K.K. reports no disclosures.. J.P. reports no disclosures. J.F. receives funding from the National Ataxia Foundation and is a fellow of the Hertie Academy for Clinical Neuroscience. M.M.S. receives support from the National Ataxia Foundation. H.H. receives support from the intramural fortüne program of the Medical Faculty of the University of Tübingen (grant no. 2554-0-0) and the Deutsche Forschungsgemeinschaft (DFG, HE 8803/1-1). H.J. reports no disclosures. K.R. has received grants from the German Federal Ministry of Education and Research (BMBF 01GQ1402, 01DN18022), the German Research Foundation (IRTG 2150), and Alzheimer Forschung Initiative e.V. (NL-18002CB) and honoraria for presentations or advisory boards from Biogen and Roche.H.G.-M. has received funding from AtaxiaUK and CureSCA3. M.R. received grants from Fundação para a Ciência e Tecnologia, Lisboa Portugal (CEECIND/03018/2018). J.G. reports no disclosures.J.I. reports no disclosures. K.M.S. has been receiving support from the UMEA Clinician Scientist Academy of the medical faculty of the University Duisburg-Essen (sponsored by the German Research Foundation) since January 2021. J.V. reports no disclosures. M.M.V. reports no disclosures. P.G. received funding from JPND/MRC MR/N028767/1. L.P.A. received the European Regional Development Fund through the Regional Operational Program Center 2020, Competitiveness Factors Operational Program (COMPETE 2020), and National Funds through FCT (Foundation for Science and Technology)—projects UID/NEU/04539/2020, BrainHealth2020 projects (CENTRO-01-0145-FEDER-000008), ViraVector (CENTRO-01-0145-FEDER-022095), SpreadSilencing POCI-01-0145-FEDER-029716, private funding from PTC Therapeutics. M.L. receives support from Fundo Regional para a Ciência e Tecnologia (support to the Azores participation in the ESMI network).B.W. receives research support from Radboudumc, ZonMW, Hersenstichting, and Gossweiler Foundation and royalties from BSL—Springer Nature and is on the scientific advisory board of uniQure.L.S. received funding from the German Ministry of Education and Research (BMBF) to the Treat-Ion project under the frame of E-Rare (grant no. 01GM1907A) and to the TreatHSP project under the frame of E-Rare (grant no. 01GM1905A), the European Commission to European Reference Network for Rare Neurological Diseases (ERN-RND) registry (grant no. 947588), the Innovationsfond of the G-BA to ZSE-DUO (grant no. 01NVF17031), and TNAMSE (grant no. 01NVF16024) as well as the German Ministry of Health (BMG) to the LeukoExpert project (grant no. ZMVI1-2520DAT94E). T.K. has received research support from the Bundesministerium für Bildung und Forschung (BMBF), the Bundesministerium für Gesundheit (BMG), and the National Institutes of Health (NIH). He has received consulting fees from Roche, UCB, uniQure, and Vico Therapeutics. M.S. received consultancy honoraria from Janssen Pharmaceuticals, Ionis Pharmaceuticals, and Orphazyme Pharmaceuticals, all unrelated to this work. O.R. and M.S. received grant 779257 “Solve-RD” from the Horizon 2020 research and innovation program.
J.H.-S.: design and conceptualization of the study, acquisition of data, statistical analysis of data, and drafting and revision of the manuscript. K.K. and J.P.: development of the assay, acquisition of data, drafting the manuscript, and revision of the manuscript. J.F.: subject recruitment, analysis of data, and revision of the manuscript. M.M.S., H.H., H.J., K.R., H.G.-M., M.R., J.G., J.I., K.M.S., J.V., M.M.V., P.G., L.P.A., M.L., B.W., and T.K.: subject recruitment, acquisition of data, and revision of the manuscript. L.S. and M.S.: subject recruitment, analysis of data, and revision of the manuscript. O.R.: design and conceptualization of the study and drafting and revision of the manuscript. Additional ESMI study group members are listed in Table S2 and participated in subject recruitment, acquisition of data, and revision of the manuscript. All authors read and approved the final manuscript.