Ambroxol for the Treatment of Patients With Parkinson Disease With and Without Glucocerebrosidase Gene Mutations

Key Points Question Does ambroxol cross the blood-brain barrier, and what are the biochemical changes associated with ambroxol therapy in patients with Parkinson disease with and without glucocerebrosidase gene mutations? Findings In this open-label clinical trial of 17 patients with Parkinson disease, ambroxol crossed the blood-brain barrier and bound to the β-glucocerebrosidase enzyme, and it increased β-glucocerebrosidase enzyme protein levels and cerebrospinal fluid α-synuclein levels in patients both with and without glucocerebrosidase gene mutations. Meaning Ambroxol therapy has potential for study as a neuroprotective compound for the treatment of patients with Parkinson disease both with and without glucocerebrosidase gene mutations.


Male or female;
2. Age ≥ 40 and ≤ 80 years of age; 3. Confirmed diagnosis of Parkinson disease at any time; and Hoehn and Yahr criteria, confirmed staged between I -III, inclusive; 4. Able and willing to provide informed consent prior to any study related assessments and procedures at screening visit 1; 5. Capable of complying with all study procedures, including fasting lumbar puncture; 6. Willing to provide a blood sample for screening genomic for Parkinson Disease related DNA analysis and/or consent to Investigators obtaining and using participants previous DNA results if applicable; 7. Willing and able to self-administer oral ambroxol medication, from day 1 to 186 (at 60 mg TID (day 1-7), 120 mg TID (day [8][9][10][11][12][13][14], 180 mg TID (day 15-21), 300 mg TID (day 22-28) and 420 mg TID (day 29-186)); 8. Able to travel to the participating study site; Contraceptive Methods with a Failure Rate of < 1%: Oral contraceptive, either combined or progestogen alone; Injectable progestogen; Implants of levonorgestrel; Estrogenic vaginal ring; Percutaneous contraceptive patches; Intrauterine device (IUD) or intrauterine system (IUS) that meets the <1% failure rate as stated in the product label; Please note: All male and female participants of child bearing potential must agree with their partners to use double-barrier birth control or abstinence while participating in the study and for 2 weeks following the last dose of the study drug.
Participants may continue to take PD medications including glutamate antagonists, 3. Exposure to more than three investigational medicinal products within 12 months prior to the first dose in the current study; 4. Confirmed dysphagia that would preclude self-administration of ambroxol up to 7 tablets TID for the duration of day 1 to day 186);

Liquid chromatography mass spectrometry of GCase and glucosylceramide
To 600 µL CSF, 3 pmol heavy labelled GCase peptide, NFVDSPIIVDITK (Genscript, USA), were added as internal standard. The samples were freeze dried and trypsin digested as previously described 1 . Sample clean-up was performed using C18 cartridges (Biotage, Sweden) which were washed with two 1 mL aliquots of 70% acetonitrile, 0.1% trifluoroacetic acid (TFA) and primed with two 1 mL aliquots of 0.1% TFA before the sample was loaded.
The flow-through was re-applied and the bound peptides washed with one 1 mL aliquot of 0.1% TFA. The peptides were eluted with 500 µL 70% acetonitrile, 0.1% TFA and solvents were evaporated using a SpeedVac. Before analysis, the peptides were re-constituted in 120 µL 3% acetonitrile, 0.1% TFA.
Mass spectral analysis was performed as previously described 2 . 5 µL of digest were injected and peptides separated on a Waters Acquity UPLC system coupled to a Xevo TQ-S mass were added. The samples were incubated in room temperature for 60 minutes before centrifugation at +4 °C, 5000g for 10 minutes. 300 µL of the organic phase were transferred to a glass vial and solvents were evaporated under nitrogen. Before analysis, the samples were re-constituted in 50 µL methanol.
Analysis was performed using a Waters Acquity Liquid Chromatography Quaternary Solvent Manager system coupled to a Waters Xevo TQ-S mass spectrometer. 5 µL of sample were injected and separated on a Waters Acquity UPLC BEH C8 column, 1.7 µm, 2.1 x 50 mm with a VanGuard pre-column of the same chemistry. The mobile phase consisted of A: water, 0.1% formic acid and B: methanol, 0.1% formic acid. The gradient profile lasted for 6.5 minutes and was initially set to 50% B for 0.2 minutes, then linearly increased to 100% B over 1.8 minutes. The column was washed with 100% B for 1 minute, then returning to equilibrate at initial conditions before the next injection. The flow rate was 0.5 mL min -1 .
Glucosylceramides were detected using multiple reaction monitoring in positive mode, see below for transitions. The data were integrated using an in-house script written in Python.
Analyte responses were normalised to internal standard response before concentrations were calculated in the average of three technical replicates using a calibration curve created from glucosylceramide standard (#1522, Matreya, USA) ranging from 0 to 0.5 ng µL -1 .
Quality control samples consisting of pooled extract were run at least every seventh sample and showed a coefficient of variance of 11%.

Liquid chromatography mass spectrometry of ambroxol
Liquid chromatography mass spectrometry quantification of ambroxol was performed by Laboratories of the Government Chemist (LGC), Teddington, UK. The achieved limit of detection/quantification was 0.5ng/ml and 1.0ng/ml respectively, the latter being in excess of specified limit of detection (included in the tender and protocol) of 20 ng/ml.
Details of the optimisation assays are provided in supplementary materials section 8.

Alpha synuclein and tau measurement
Total A-SYN concentration was measured using a commercially available enzyme-linked immunosorbent assay (ELISA) from Covance (Covance, Dedham, MA, USA). Tau concentration was measured using a commercially available INNOTEST ELISA (Fujirebio, Ghent, Belgium). 2.

RESEARCH QUESTIONS ANDHYPOTHESIS
The primary research question in this proof-of-concept and safety study is whether after escalation of the dose in a safety study, ambroxol crosses the BBB in sufficient amounts to register a significant increase in cerebrospinal fluid (CSF) concentrations of the drug. We predict a concomitant change in GCaseactivity. There is, however, no pre-existing evidence for what constitutes a clinically meaningful change in either ambroxol levels or GCase activity. Therefore interpretation will be primarily through the size of effect and the width of the confidence intervals Descriptive summaries shall be provided for all other outcomes by each visit and by GBA mutation, andfor baseline variables. Descriptive summaries shall also bepresented for the pair of primary outcomes by GBAmutation.

Primary outcomes
The results for two primary outcomes of interest shall be reported: • Change in Ambroxol concentration in CSF from baseline to the six month visit • Change in GCase enzyme activity in CSF from baseline to six month visit

Secondary outcomes
This excludes free text, qualitative data, and meta-data such as the date of assessment and assessor details.
For each visit, where collected: • Weight • Vital signs: heart rate, blood pressure, respiration rate, temperature

FOLLOW-UP AND PARTICIPANT FLOW
Outcomes and fixed baseline patient characteristics will be reported from screening at visits 1 and 2.All outcome variables will be reported at subsequent treatment phase visits 3, 4 5,6 and at the termination visit 7. The endpoint for the primary outcome is reported at visit 6.
Patient numbers and progression, along with the counts and severity of adverse events shall be presented in a flow diagram for the key stages of the trial at the screening visits, the telephone visits following dose escalation, and the treatment-phase visits recording outcomes.

DATA PRESENTATION
Summary statistics for the primary, baseline and secondary outcomes will be presented as appropriate point estimates with their standard deviations and the number of observations, or as counts with percentages, where the data are categorical. Additionally, summary statistics of the primary outcomes will be presented by GBA genotype, but no inference shall be made about the significance of the difference between genotypes.
For descriptive purposes only, 95% confidence intervals will be presented for the laboratory results for the enzyme activity, biomarker and antibody panels by GBA genotype, and shall not be used to determine the success of the study.

STATISTICAL ANALYSIS
Following the data lock, the following analyses will be undertaken by the Trial Statistician, and the results acted upon in accordance with the decision rule outlined in this statistical analysis plan.

Overall strategy
The sample size for the study was selected by the Chief Investigator but there was no formal sample size calculation or statistical consideration. The main focus of all analyses will be on the interpretation of 95% confidence intervals. The results from formal significance testing will be presented as supplementary information only, as indication of the degree of uncertainty.

Descriptive analysis
A summary of baseline characteristics and the outcomes at a baseline and at subsequent treatment-phase visits for the whole cohort will be presented as point estimates with their standard deviations, or as counts and percentages where the data are categorical. Additionally the 95% confidence intervals for the antibody, biomarker and enzyme activity panels shall be reported. No formal significance testing will be conducted, apart from the primary analysis.

Primary analysis
Primary analysis will be based on interpretation of the confidence intervals, arbitrarily covering 95% of each distribution of the mean changes in CSF ambroxol concentration and GCase enzyme activity, modelled according to at-distribution of degrees of freedom based on the sample size. Since only an increase in ambroxol is of interest (or possible, if participants not taking ambroxol), then the lower 95% confidence bound will be presented for the change in ambroxol. A two-sided 95% confidence interval will be presented for GCase activity. The distribution of the changes from baseline to visit six in ambroxol concentration and GCaseenzyme activity shall be inspected graphically and transformed, if necessary, to fit a Normal distribution before calculating the confidence intervals. The confidence intervals will then be interpreted relative to the transformed null value, which might not be zero, if an offset has been applied.
To account for the quantification limit of the instrument measuring ambroxol concentrations, if the mean concentration at visit six is below the1.0ng/ml limit then CSFambroxol will be deemed to have not changed, regardless of the change from baseline to visit six.

Further descriptive analysis
The individual response in GCase activity to potential changes in CSF ambroxol concentrations will be further explored through a bivariate plot of the changes in each. It is unknown whether there will be a systematic difference between patients in CSF ambroxol to observe any potential relationship, having administered the same ambroxol dose to each. However, the Pearson correlation coefficient will be presented along with its 95% confidence intervals having transformed both variables, if such an observed relationship is non-linear.

ADVERSE EVENTS
Safety data and adverse events will be listed descriptively and include details of the event.

MODEL CHECKING ANDVALIDATION
All analyses will be undertaken using STATA v14.2.
Before undergoing the primary analysis, decisions about whether to transform the primary outcome (e.g. natural log) will be made based on plots of its distribution, and appropriate statistics reported for skewness if necessary. Checks will be undertaken to assess the robustness of the primary analysis based on the normality of the residuals.

. Scatter Plot of Change in CSF GCase Activity Against Change in CSF Ambroxol
Dashed line represents no change in CSF GCase activity. We considered the distribution of samples for the primary outcomes. In the case of both CSF GCase and CSF ambroxol the data appeared to be normally distributed (see supplementary figures 1 and 2). We identified one outlier who had marked fall in CSF GCase activity with a minimal rise in CSF ambroxol concentration (circled in red in supplementary figure 3). We considered whether this could be due to blood contamination, however at both timepoints no red blood cells were seen on CSF microscopy. In the absence of a convincing scientific rationale to exclude the subject, they were included in both primary and secondary analyses.    83 (17)