Main

Parkinson’s disease (PD) is the second most prevalent neurodegenerative disorder1, affecting up to 1.2 million people in the United States, with 90,000 new cases annually1,2,3. PD is a progressive disorder characterized clinically by bradykinesia, rigidity, resting tremor and gait disturbances with postural instability1,3. Pathologically, initiation and progression of disease is linked to the accumulation of α-synuclein in fibrillar aggregates and/or inclusion bodies internalized by neurons, often resulting in the Lewy pathology4. This process, in turn, drives cell death by mitophagy, Parthanatos and other mechanisms5,6,7, resulting in the progressive degeneration of dopaminergic (DA) neurons in the substantia nigra pars compacta, with a corresponding reduction in striatal dopamine. Although motor dysfunction is driven by DA neuron loss, PD pathology becomes widespread as the disease progresses, affecting non-dopaminergic neurons, including serotoninergic, cholinergic and adrenergic neurons as well as nerve cells in the olfactory system, cerebral hemisphere, brain stem, spinal cord and peripheral autonomic nervous system.

c-Abl plays a fundamental role in the biology of all mature cells and tissues, acting as a sensor of cellular stresses throughout the body. C-Abl can recognize abnormalities or toxicities that may arise in a mature cell or tissue, including neurons, and responds by initiating biochemical cascades driving multiple mechanisms of cell death5,6,7. Inhibition of c-Abl blocks these death cascades, as recently shown for risvodetinib, a potent, selective inhibitor of c-Abl in models of inherited or sporadic PD8. Once-daily oral risvodetinib blocked neurodegeneration in murine models, with a concomitant reduction or clearance of α-synuclein pathology, partial to complete restoration of central nervous system (CNS) function, arrested disease progression and suppressed neuroinflammation8. The benefit of treatment with risvodetinib may not be restricted to PD, as demonstrated from analysis of postmortem human brain tissue from patients with multiple system atrophy and in a murine model of multiple system atrophy9,10. Thus, inhibiting c-Abl may have a therapeutic benefit across multiple synucleinopathies.

Favorable early clinical evaluation of risvodetinib in older healthy adults and in participants with treated or untreated PD led to the 201 Trial11, a 12-week, double-blind placebo-controlled dose-escalation study of risvodetinib in participants with early untreated PD. The 12-week trial duration was guided by multidose chronic toxicology studies in which most adverse events that could be clinically meaningful were observed by 13 weeks8. As the first extended human dosing study, the primary end point for the 201 Trial was the evaluation of safety and tolerability. A total of 14 secondary clinical assessments evaluated motor and nonmotor features referable to the central, autonomic and enteric nervous systems. An exploratory evaluation assessed the impact of risvodetinib on the underlying α-synuclein pathology of PD, measured by quantification of phosphorylated-Ser129-α-synuclein deposition in cutaneous neurons of the skin using confocal microscopy12,13,14.

Results

Participants

A total of 220 prospective participants were screened and 137 were enrolled between 22 August 2022 and 2 October 2024 and randomly assigned to once-daily 50 mg, 100 mg or 200 mg risvodetinib or to an equivalent placebo dose (CONSORT diagram; Fig. 1). The mean age of participants was 63.7 ± 7.6 years; 62% of risvodetinib-treated participants were male compared to 54% of placebo-treated controls. The average time from diagnosis was 13.5 ± 14.8 months for placebo-treated and 12.7 ± 17.9 months for risvodetinib-treated participants. Overall, 67% of treated participants and 71% of placebo controls had a Hoehn & Yahr score of 2.0 at screening; no participant had a Hoehn & Yahr score >2.5 (Table 1). The trial was temporarily paused after 11 participants had begun treatment due to a 60-day clinical hold to allow the US Food and Drug Administration (FDA) to review chronic toxicology and pharmacokinetic data with the sponsor. When the trial was restarted, the 11 pre-hold participants were included in the safety analysis set, resulting in 137 participants contributing to the primary end point (CONSORT diagram; Fig. 1). Pre-hold participant data were not included in the secondary efficacy analysis, as their assessments were performed days or weeks after the last administered dose.

Fig. 1: Flow diagram of the 201 Trial evaluating risvodetinib in untreated PD.
figure 1

The flow chart depicts the process of patient screening, randomization and how many participants completed the final analysis in the study. ‘Study paused’ represents the number of participants that were randomized before the 60-day clinical hold. ‘Withdrew consent’ refers to participants who withdrew from the study before end of treatment post-clinical hold.

Full size image
Table 1 Demographics of 201 Trial participants
Full size table

Primary outcomes

A total of 126 participants entered the trial after the pause, and 120 completed the 12-week dosing regimen; there was a 95% completion rate with 99% dosing compliance among participants. No deaths and three severe adverse events (SAEs) occurred during the study that were unrelated to risvodetinib. One from the development of sepsis resulting from an untreated cut to a participant’s hand for a participant on the 200-mg dose; one involved a case of progressive diverticulitis for a participant administered placebo and a third related to the repair of a broken femur following a fall that required hospitalization for a participant administered the 50-mg dose. Across the 137 participants in the safety set, the fraction of participants who experienced at least one treatment-emergent adverse event (TEAE) was 54%, 62%, 64% and 60% in the 50 mg, 100 mg, 200 mg and placebo groups, respectively (Table 2 and Supplementary Table 1). The average number of adverse events per person was similar for treated participants at any dose and for participants given placebo (Table 2 and Supplementary Table 1). TEAEs common to this drug class, such as nausea, diarrhea, vomiting, peripheral edema and cardiovascular toxicities, were minimal or not observed and not materially different from controls (Table 2, with further details in Supplementary Table 1). Falling, reported as a TEAE, occurred in 3% of risvodetinib-treated participants compared to 14% of placebo-treated controls (Supplementary Table 1). Similar to observations made in the phase 1/1b trials11, there were asymptomatic sporadic elevations of lipase and/or amylase for participants in the 201 Trial. Overall, 23 of these elevations occurred in actively treated participants with five events reported as adverse and in eight placebo-treated participants with one incidence reported as adverse, but there was no indication of pancreatitis or other organ injury. Thorough ophthalmology exams were performed at baseline and at end of treatment because long-term toxicology studies in rats, but not monkeys, suggested a potential risk for damage to the retina8. No TEAE of blurred vision was reported and no ophthalmology exam following end of treatment detected any injury or changes to the retina or other eye structures in any trial participant relative to baseline.

Table 2 Summary of TEAEs by treatment arm1
Full size table

Secondary outcomes

The 201 Trial was not powered to detect a significant difference between treatment with risvodetinib and placebo. On entry, participants had clinically manifest motor and nonmotor symptoms of early PD, with Movement Disorder Society revision to the Unified PD Rating Scale (MDS-UPDRS) scores of 5.0 ± 3.8, 6.7 ± 4.4 and 24.3 ± 9.6 for Parts I, II and III, respectively. Participants were not somnolent (Epworth Sleepiness Scale (ESS) of 3.9 ± 2.9). Participants also did not have significant lower gastrointestinal (GI) dysfunction (Complete Satisfaction of Bowel Movement (CSBM) score of 6.1 ± 3.3) nor had significant upper GI dysfunction (baseline score near 0 for the Patient Assessment of Gastrointestinal Disorders-Symptom Severity Index (PAGI-SYM)). GI Quality of Life using the Patient Assessment of Constipation (PAC) and the PAGI Quality of Life (PAC-QOL and PAGI-QOL, respectively) also indicated that participants had only mild GI dysfunction on enrollment. Both participants and clinicians rated disease severity as mild at baseline (Patient Global Impression and Clinician Global Impression of Severity (PGI-S and CGI-S, respectively) of 2.0 ± 0.5 and 2.6 ± 0.7, respectively).

Secondary efficacy outcomes from the modified intent-to-treat (mITT) population are summarized in Table 3. The primary hierarchical secondary end point of the MDS-UPDRS Parts II + III did not reach statistical significance. A nominally significant improvement was observed for participants administered the 100-mg dose in the MDS-UPDRS Part II and for the Schwab and England Activities of Daily Living (SEADL) scale for participants administered the 50-mg dose (Table 3). One participant during the 12-week treatment period initiated anti-parkinsonian medication during their participation following 8 weeks of treatment at the 200-mg dose despite no worsening of MDS-UPDRS Part II or Part III assessments. Three participants initiated anti-parkinsonian medication during the 2-week follow-up period after their 12-week dose. Two of these participants had been treated with placebo and one of these participants was treated at the 200-mg dose during the 12-week treatment period.

Table 3 Analysis of secondary end points by a mixed model of repeated measures1
Full size table

Exploratory outcomes

Voluntary skin biopsy and lumbar puncture were planned for the 201 Trial to evaluate the effect of treatment on the underlying protein pathology of PD. Just four participants agreed to lumbar puncture and sampling of cerebrospinal fluid (CSF); thus, no analysis was performed beyond measures of the risvodetinib concentration at steady-state in the CSF. A second exploratory measure was performed using skin biopsy to evaluate the potential treatment effect of risvodetinib on pathological α-synuclein deposits in cutaneous neurons12,13,14.

Pathological α-synuclein can be detected by immunofluorescence to evaluate the distribution of synuclein pathology in animal models and human tissues8,9,10,12,13,14,15,16,17,18,19,20,21,22,23. In the 201 Trial, longitudinal samples collected at baseline and at end of treatment (12 weeks) from three biopsy locations included 10 sample sets from placebo controls, 10 sample sets from participants treated with 50 mg, 5 sample sets from participants treated with 100 mg and 11 sample sets from participants treated with 200 mg once daily. From these 36 sample sets, 6 sample sets lacked synuclein deposition at baseline and after 12 weeks of treatment. Figure 2 shows a representative example of the data at baseline and after 12 weeks of treatment along with a plot of the fractional change in α-synuclein aggregate deposition by treatment group (Fig. 2b).

Fig. 2: Longitudinal changes in α-synuclein pathology in cutaneous neurons in placebo and treated conditions.
figure 2

a, Cross-sectional representative fluorescence intensity map of α-synuclein aggregates at each of three biopsy locations. Faint outlines in each intensity plot represent the shape of the tissue sections observed in the confocal microscope. Numbers to the right of each intensity map are the summed fluorescent intensities observed within each biopsy sample. b, Plot of the fractional change in α-synuclein pathology deposited in cutaneous neurons. Biopsy was voluntary and approximately 40% of randomized participants (n = 51) agreed to have skin biopsies taken. From those, only 71% of the participants (n = 36) had complete data at baseline and after 12 weeks of treatment, with 10 placebo controls, 10 at 50-mg, 5 at 100-mg and 11 at 200-mg doses, assigned randomly. From the 36 complete sets, 4 participants in the 50-mg group and 1 participant in each of the placebo and 100-mg groups had no detectable α-synuclein deposition at either baseline or 12 weeks. The fractional change represents the difference in the summed intensities across the three biopsy locations measured at baseline and 12 weeks relative to the baseline summed intensities. Shown are the median values and the 25th and 75th quartiles below and above the median line, respectively. Treatment groups were individually compared to the placebo group using a one-way analysis of variance followed by Dunnett multiple comparisons test using Prism v.10. Differences between any treatment group and placebo were not found to be statistically significant. All data above the lower limit of detection were utilized unless the observed baseline signal occurred in just a single cutaneous neuron. If a single neuron gave rise to the observed signal at baseline, the participant data were excluded as it could not be known whether the same neuron could be sampled at the next longitudinal time point.

Source data

Full size image

Of the ten placebo-treated controls, six (60%) either showed no change or a fractional increase in synuclein deposition across three biopsy locations at 12 weeks. Two control participants showed reductions of 2% to 5% and two others showed substantial reductions of 45% to 85% relative to baseline levels. Treated participants showed fewer increases in α-synuclein pathology. Instead, the proportion of participants in which α-synuclein deposition was reduced increased with increasing dose, with reduced pathological α-synuclein detected in 30%, 60% and 64% of participants treated with the 50 mg, 100 mg or 200 mg doses, respectively (Fig. 2); however, the differences in α-synuclein deposition between any treatment group and placebo was not statistically significant.

Discussion

The 201 Trial was designed to evaluate the safety and tolerability of once-daily risvodetinib, to gain insight into potential functional benefit(s) and to simultaneously evaluate the impact of risvodetinib on the underlying biology of PD. The safety and tolerability profile of risvodetinib was similar to placebo for every organ system, an unexpected profile for a protein kinase inhibitor. This profile was observed despite risvodetinib reaching a high steady-state plasma exposure, which for the 200-mg dose is twofold to 200-fold higher than the steady-state exposure for the approved anticancer c-Abl inhibitors imatinib, dasatinib, nilotinib, bosutinib and ponatinib8,11. The selectivity of risvodetinib for the nonreceptor Abl kinases8 likely accounts for the favorable TEAE profile relative to other drugs in this class24 and compared to all other approved protein kinase inhibitors25. Although it was unexpected to see the difference in the reporting of falls as an adverse event between risvodetinib and placebo treatment, a formal assessment of falling in future trials will be needed to establish whether risvodetinib is having a direct impact on the frequency of falling in PD.

Secondary clinical assessments showed little change over 12 weeks, as was expected for a trial of this short duration, and no conclusion regarding clinical efficacy could be drawn from the observation of nominal significance for measures of the activities of daily living. Risvodetinib did not worsen clinical features, although there was a possible increase in motor examination scores.

Genetic, biochemical and neuropathological studies in humans26 and model27 systems have established the likelihood of α-synuclein aggregation as a pathological feature of PD, but may not be the only protein aggregates contributing to disease28. Although α-synuclein aggregates come in many forms and manifest alone26,29 or in higher-order complexes7,30,31, any successful disease-modifying therapeutic should reduce or eliminate α-synuclein pathology as a consequence of treatment. Immunotherapies targeting extra-neuronal α-synuclein aggregates32,33, small molecule disaggregators34,35,36 and intracellular effectors are among many efforts37 that have been tried to clear α-synuclein pathology in PD, all with no success. Because c-Abl is activated by α-synuclein aggregate internalization8,15,16,17,18,19,20,21,22,23, and, in model systems, c-Abl inhibition drove clearance of α-synuclein pathology8, it was hypothesized that c-Abl inhibition might succeed in reducing the α-synuclein pathology in human PD and would do so by reactivating endogenous processes of protein clearance5.

Using immunofluorescence of pathological α-synuclein to monitor the change in α-synuclein pathology yielded a clinical measure that treatment was reducing the underlying α-synuclein pathology in patients who volunteered for skin biopsy. The observations herein are consistent with our previous animal model studies that demonstrated the intrinsic capacity to clear aggregate deposition exists in neurons inside and outside of the brain5,8. As Fig. 2b indicates, there is a wide variability in rates of deposition of synuclein pathology seen in the placebo-treated participants, as well as variability in the reduction of synuclein pathology seen in participants treated with risvodetinib. We believe this variability is a reflection of the heterogeneous population of enrolled participants, where each participant entered the trial at a different point on their path of disease progression and each participant has a varying amount of pathological α-synuclein deposition in cutaneous neurons on entry. Many physiological differences between individual participants could contribute to this variability, so it is not surprising that different rates of response are seen with risvodetinib treatment. Given the small dataset in each group, it is not surprising that the comparison between any risvodetinib treatment group and placebo treatment did not reach statistical significance. Qualitatively, risvodetinib treatment seems to reduce synuclein pathology from the skin for the majority of treated participants who volunteered for skin biopsy. The outliers in the treatment groups, particularly one each in the 100-mg and 200-mg dosing groups, could indicate treatment nonresponders, or perhaps could reflect a subtherapeutic exposure to risvodetinib during the trial for these participants. A number of factors can contribute to a variable pharmacokinetic exposure to c-Abl inhibitors across trial participants38, which is now recognized as a common problem often requiring dose optimization experiments to ensure therapeutic exposures across the human population39.

The incomplete clearance of α-synuclein pathology at 12 weeks suggests that skin biopsy may be a suitable method to monitor the effect of treatment on synuclein pathology in the clinical trial setting. Longer duration studies will be needed to evaluate whether the protein pathology of PD can be stably reduced, or even eliminated, by risvodetinib treatment. Further, it must be demonstrated whether monitoring the pathology of disease by skin biopsy can act as a predictive biomarker of treatment success. At our present state of knowledge, it is not possible to correlate the quantity of pathological aggregate deposited in cutaneous neurons with the disease state or with the rate of disease progression. It is also not known whether a peripheral treatment effect on synuclein pathology will inform on the focal pathology in the substantia nigra pars compacta or the global brain pathology in humans, despite the correlation seen in disease models8.

There have now been four c-Abl inhibitors evaluated as potential disease-modifying therapies in phase 2 clinical trials of PD. Nilotinib, the anticancer c-Abl inhibitor, was the first of these to be studied, has known cardiovascular and other toxicities incompatible in patients with PD, had limited activity in animal model studies and limited oral penetration into the CNS and was not clinically active in PD40,41,42,43. Similarly, vodobatinib, a dual inhibitor of Abl and SRC kinases, while reported to be active in preclinical models and reaching apparent therapeutic concentrations in human CSF44,45,46, proved to have significant toxicities and was not clinically active47. Radotinib has been shown to be active in animal model studies48 and is presently being evaluated in an ongoing clinical trial (NCT04691661), although its human side effect profile is not encouraging as therapeutic doses reach dose-limiting toxicities and the ongoing trial is exploring doses at one-half the therapeutic dose49. Risvodetinib, by contrast, has an excellent safety profile and may reduce the underlying pathology of PD. A definitive measure of efficacy is now warranted for risvodetinib.

Methods

Ethics statement

The protocol and recruitment materials were approved by the institutional review boards or ethics committees at 32 study sites across the United States. For 27 sites, Advarra, a central Institutional Review Board was used; at five university-based sites, Advarra as well as the local institutional review board of the academic institution was used. The 201 Trial was conducted according to the principles of the Declaration of Helsinki and Good Clinical Practice guidelines. All participants provided written informed consent before participation and for use of tissue or fluid samples for subsequent data analyses. Participants were reimbursed for costs of travel to and from a clinical site but were not otherwise compensated for their participation. The trial was registered with https://clinicaltrials.gov under registry number NCT05424276.

Study design

The 201 Trial was a 12-week, randomized, double-blind, placebo-controlled dose-ranging clinical trial with doses of 50 mg, 100 mg or 200 mg risvodetinib or placebo in participants with untreated PD, followed by a safety assessment in the absence of treatment at week 14. The trial was designed to primarily assess safety and tolerability. The study randomized 137 participants at 32 clinical sites across the United States using two randomization schemes. First, participants were randomized across the 50-mg, 100-mg or placebo groups in a 1:1:1 randomization scheme. Once five people were randomized to each of the 50-mg, 100-mg or placebo groups, additional participants were randomized to the 50-mg, 100-mg, 200-mg or placebo groups in a 5:5:5:6 randomization scheme. This two phase randomization was performed to accommodate the delayed FDA review of pharmacokinetic data at the 200 mg dose in healthy volunteers. At the end of the trial, all treatment groups were evenly distributed across all four trial arms.

Participants

Prospective participants were between 30 and 80 years old, had a diagnosis of PD consistent with UK Brain Bank and MDS Research criteria that included bradykinesia with sequence effect and motor asymmetry. Participants had to be <3.0 on the Hoehn & Yahr staging scale and have a Montreal Cognitive Assessment score ≥24. Key exclusion criteria included diagnosis of atypical parkinsonism or a high likelihood of needing anti-parkinsonian medication during the 12-week period of the study. Prospective participants could not have clinically significant orthostatic hypotension or hallucinations requiring antipsychotic medications nor could prospective participants have been treated with dopaminergic agents for more than 28 days and within 28 days before screening. They could not have been exposed to monoamine-oxidase B inhibitors in the 90-day period before medical screening nor be treated with these inhibitors for more than 28 days. See Supplementary Protocol for additional information.

Outcomes

The primary end points were safety and tolerability assessed descriptively by the incidence and temporal profile of TEAEs, the incidence of TEAEs leading to withdrawal of the study drug, the incidence of SAEs and the proportion of randomized participants who discontinued their assigned dose.

Secondary efficacy end points, in hierarchical order, were the sum of MDS-UPDRS Parts II and III, Parkinson’s Disease Questionnaire (PDQ-39) Summary Index, Patient Global Impression-Severity (PGI-S), Clinician Global Impression of Severity (CGI-S), MDS-UPDRS Part II, MDS-UPDRS Part III, MDS-UPDRS Part I, Non-Motor Symptom Scale (NMSS), Complete Spontaneous Bowel Movement Score (CSBM), ESS, SEADL Scale, Patient Assessment of Upper Gastrointestinal Disorders Severity Index (PAGI-SYM), Patient Assessment of Constipation Quality of Life (PAC-QoL) and the Patient Assessment of Gastrointestinal Disorders Severity Quality of Life (PAGI-QoL).

Skin biopsy collection was optional and conducted following written participant consent. Three-millimeter punch biopsies were obtained from three anatomical sites: the lateral distal leg, lateral distal thigh and posterior cervical regions as previously described12,13,14. All three sites were sampled at each longitudinal time point and in proximity to the previous biopsy. Skin biopsy specimens were immediately fixed in Zamboni solution (2% paraformaldehyde-lysine-periodate) then cryoprotected (20% glycerol and 20% 0.4 M Sorensen buffer). Specimens were subsequently cryo-sectioned into six to eight 50-μm-thick sections.

Analysis of secondary clinical measures

The secondary efficacy end points were defined as the change from baseline at week 12 and were evaluated in hierarchical order (by end point and within each end point, by dose). Descriptive statistics for each continuous end point were presented as change from baseline by treatment group and visit. The differences between the dose levels versus placebo at week 12 in the continuous secondary end points were estimated using mixed models for repeated measures (MMRM) using the mITT dataset. In this study, treatment group sizes within the mITT analysis set ranged from 29 to 32 participants. A typical threshold for reliance on the Central Limit Theorem is approximately 20–30 observations, thus our sample sizes were generally supportive of parametric methods. Further, the use of a longitudinal model such as the MMRM also appropriately accounts for within-subject correlations and accommodates incomplete data without the need for ad hoc imputation, making it preferable to alternative methods that either discard longitudinal information or require stronger assumptions to address any missing observations. The MMRM included the observed changes from baseline from all post-baseline visits as the response values. Missing data associated with efficacy end points were not imputed but were analyzed as missing in the model. The treatment group, visit and the interaction between the treatment group and visit were the fixed factors in the analysis. The baseline value was used as a covariate in the model. The least squared mean (LSM) for each treatment group, s.e.m. and LSM difference between different dose levels and placebo along with the P value and 95% confidence intervals (CIs) were also calculated.

Analysis of exploratory biomarkers

For each biopsy, a minimum of six sections underwent dual-label immunofluorescent staining for simultaneous covisualization of total nerve fibers using protein gene product 9.5 (anti-rabbit, Cederlane: polyclonal 1:1,000 dilution) and α-synuclein deposition using phosphorylated-Ser129 α-synuclein antibody (anti-mouse, Wako: monoclonal: 1:2,000 dilution); phosphorylated-Ser129 α-synuclein is a well-established marker of α-synuclein pathology in PD8. Only intra-axonal deposits of phosphorylated α-synuclein, confirmed by colocalization within nerve fibers, were identified as pathological deposits, as previously described12,13,14. Each staining batch included positive and negative control samples. Additionally, intra-epidermal nerve fiber density was calculated for each biopsy site as an internal control using standard techniques to confirm stain success14.

Slide scanning was performed using a confocal slide scanning microscope (SLIDEVIEW VS200, Evident) with a ×20 objective (UPlanXApo, NA 0.80). Images were acquired at 2-μm intervals throughout the tissue sections and processed using extended focal imaging before review and analysis. Image analysis (NerValence, CND Life Sciences and Oncotopix, Visiopharm, Hørsholm v.2023.09.3.15043) was employed to objectively identify and measure cutaneous axons and colocalized phosphorylated α-synuclein deposits. All slides and processed images underwent review by an independent pathological reviewer blinded to treatment, disease state and visit sequence. Manual confirmation of all synuclein deposits was completed until concordance between image analysis and manual pathological review for phosphorylated α-synuclein was 100%.

The burden of phosphorylated α-synuclein deposition was quantified as µm2 of synuclein deposits per mm2 of analyzed tissue. With a minimum of six tissue sections studied per biopsy, the total tissue area analyzed ranged from 33–55 mm2 per biopsy. Measurements from each biopsy site were reported independently, with the total α-synuclein burden calculated as the sum of the three biopsy sites per participant per visit.

To assess potential treatment impact, the total phosphorylated-Ser129 α-synuclein burden at baseline was compared to an identical measure after 12 weeks of treatment. Samples from 36 participants were complete at both baseline and the 12-week time point. Samples were excluded if the baseline signal was detected in only a single axon in the tissue sample as this could result in a sampling error if the same axon is not biopsied at the 12-week time point. Samples were also excluded if the baseline and 12-week samples at any given biopsy location contained unequal numbers of axons in the tissue sample. The fractional change in phosphorylated-Ser129 α-synuclein deposition area intensities between the two time points at all three biopsy locations was calculated to represent the potential treatment effect for each participant. No interpretation was made concerning the absolute signal intensities at any tissue location and a statistical analysis of the fractional change is not warranted for this exploratory end point. Neither the distribution of fluorescent aggregates within a biopsy location nor the magnitude of the fractional change over time was considered to carry information on disease progression or disease status.

Statistics and reproducibility

A sample size of 30 participants for each treatment group was considered sufficient to adequately assess safety and tolerability over 12 weeks as the first extended human dosing study with risvodetinib. The safety analysis set was defined as all randomized participants that received at least one dose of study drug, while the mITT dataset included all randomized participants who had a valid baseline MDS-UPDRS, received at least one dose of the study drug, and had at least one post-baseline evaluation of the MDS-UPDRS assessment. The mITT dataset excluded participants who had been enrolled before the FDA imposed a clinical hold because their last secondary end point assessments occurred days or weeks after their last administered dose.

To enroll the trial, screening data and medical records were reviewed by an Enrollment Authorization Committee composed of an independent outside group of Movement Disorder Specialist physicians who reviewed screening data and confirmed eligibility and suitability of participants enrollment. Those selected were enrolled and randomized to one of three active arms or a placebo arm using an interactive web response system (IWRS). The IWRS is a secure, web-based tool used in clinical trials to manage how participants are assigned to different treatment groups. It ensures that randomization is accurate, unbiased, and kept confidential throughout the study. The system also tracks study drug supply in real time, reducing the risk of errors. The 201 Trial employed an IWRS system developed and validated by Medidata. See Supplementary Protocol for additional details for screening and enrollment.

The safety and tolerability of risvodetinib were assessed using descriptive statistics for each treatment group. TEAEs were defined as those that start or worsen after the first dose of study intervention until the safety follow-up visit is completed. Safety assessments included TEAEs tabulated by treatment group as well as system organ class (SOC) and preferred term (PT). Furthermore, the incidence of TEAEs was reported as the number (percent) of participants with TEAEs within SOC and PT. Participants were counted only once within an SOC and PT, even if the participant experienced more than one TEAE within a specific SOC and PT. Inferential statistical analysis was not performed on safety data.

The secondary efficacy end points (as changes from baseline at week 12) were evaluated in hierarchical order using an MMRM (Supplementary Protocol). The MMRM included the observed change from baseline from week 4, week 8 and week 12 as the response values, baseline value as a covariate and treatment group, visit, and treatment group by visit interaction term as fixed factors. End points were tested in hierarchical order using a two-sided α = 0.05 as the cutoff. The estimand was defined using participants in the mITT population evaluating the change from baseline to week 12 for each secondary end point. Intercurrent events in this analysis included whether a participant prematurely discontinued study treatment or initiated medications to control motor features of PD. The data were summarized by calculating the differences of the LSM change from baseline at week 12 by dose versus placebo. The LSM estimate for each treatment group, s.e.m., the differences in LSM and the difference between different dose levels and placebo were computed along with the 95% CI. Only the LSM values were used for statistical comparison. Categorical end points, such as PGI-S and CGI-S, were analyzed using the GLIMMIX procedure for binomial data with the logit link. The model included the observed binomial values from all post-baseline visits as the response values and the treatment group, visit and the interaction between the treatment group and visit as fixed factors. The baseline value was used as a covariate in the model. ‘Improved’ was defined as at least a whole number decrease in score for PGI-S and CGI-S. Model-based LSM estimates, P values and associated 95% CIs were generated for each treatment group at each post-baseline visit which were used for statistical comparison.

For the exploratory analysis of the treatment effect on α-synuclein aggregate deposition following skin biopsy, treatment groups were individually compared to the placebo group using a one-way analysis of variance followed by Dunnett multiple comparisons test using Prism v.10 (www.graphpad.com).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.