This was a randomised, double-blind, placebo-controlled, two-way crossover study. After completing a screening period, eligible patients were randomised equally to one of two treatment sequences, each comprising two 28-day treatment periods separated by a 28-day washout, inhaling CHF6523 in one period, and placebo in the other, both twice daily.

All assessment timings were based on the morning dose. Pre-dose on Days 1 and 28 of each treatment period, blood samples were taken for pharmacokinetic analyses, followed by forced oscillation (including slow vital capacity manoeuvres for inspiratory capacity [IC] and vital capacity [VC]), spirometry (forced expiratory volume in 1 s [FEV1] and forced vital capacity [FVC]), and the COPD Assessment Test (CAT). Blood samples were also taken up to 12 h post-dose on Days 1 and 28 for pharmacokinetic analyses, with blood samples for biomarker assessment pre-dose on Day 1 and 2 h post-dose on Day 28, and spirometry pre-dose on Days 20 and 24. Sputum was induced prior to each treatment period for baseline determination of biomarkers, and post-dose on Day 20 for collection of phosphatidylinositol (3,4,5)-trisphosphate (PIP3), and on Days 24 and 28 for collection of biomarkers. Target engagement of a PI3Kδ inhibitor is demonstrated in terms of a reduction in the relative proportion of sputum PIP3 in the total amount of phosphatidylinositol (4,5)-bisphosphate (PIP2) plus PIP3.

Given the importance of sputum data in the study, there was a special focus on sputum sample quality. Eligible patients were spontaneous sputum producers, i.e., they were able to produce an adequate induced sputum sample of at least 300 mg with a cell viability of at least 70% (with less than 30% epithelial cells) at screening. In addition, sputum quality was evaluated centrally at the sputum laboratory of the Pulmonary Research Institute, Grosshansdorf, Germany. See the supplement for additional details.

Safety and tolerability were evaluated throughout the study in terms of the occurrence of adverse events, blood chemistry, haematology and urinalysis, vital signs, and 12-lead electrocardiograms. Because of the occurrence of cough in healthy volunteers in Parts 1 and 2, cough episodes were monitored by asking patients to grade episodes on a visual analogue scale (VAS), and to report any episodes in a daily diary. Throughout the study, patients were permitted to use salbutamol as rescue medication, but not within 6 h prior to any spirometry assessment – any rescue medication intake was also captured in the daily diary.

Pharmacokinetic assessments were performed under fasting conditions. On days when pharmacokinetic samples were taken, study drug was administered in the morning at the clinical site after a fast of ≥ 10 h (water was allowed), with food or fluid intake not allowed until 1 h post-dose, with the exception of 100 mL of water if the patient coughed as a result of inhalation. Starting from 1 h post-dose and for the following 6 h, patients drank ≥ 240 mL of water every 2 h, with breakfast and lunch served approximately 3 and 5 h after morning study drug, and, on Day 1, dinner at least 2 h before or 1 h after evening drug administration. No intake of alcohol, grapefruit, or xanthine/caffeine containing beverages or food was allowed from 48 h prior to each intake of study medication and until the last visit procedure, with strenuous exercise prohibited for 24 h before any spirometry measurement.

The study was approved by the independent ethics committees at each institution (listed in the supplement), was performed in accordance with the Declaration of Helsinki and Good Clinical Practice, and was registered at ClinicalTrials.gov (NCT04032535; 23rd July 2019). The protocol was amended three times after recruitment started, only one of which involved any substantial changes – permitting vaccination during the washout period between the two treatment periods, and specifying that IC and VC data were recorded from forced oscillation.

Participants

In addition to a requirement to be spontaneous sputum producers, eligible patients were diagnosed with COPD at least 12 months prior to inclusion, were current or ex-smokers with a smoking history of ≥ 10 pack-years, blood eosinophil count ≥ 150 cells/µL, post-bronchodilator FEV1 30–70% predicted and FEV1/FVC < 0.70, were symptomatic (CAT total score ≥ 10), and had been receiving inhaled triple therapy (ICS/LABA/LAMA) for ≥ 6 months prior to entry. The main exclusion criteria were a current diagnosis of asthma, a COPD exacerbation in the 6 weeks prior to entry, and the use of any other COPD maintenance therapy other than ICS/LABA/LAMA in the 6 months prior to entry. The full list of inclusion and exclusion criteria is in the supplement. All patients provided written informed consent prior to any study-related procedure.

Interventions

The treatments administered were CHF6523 2 mg twice daily or matching placebo, both via capsule-based dry-powder inhaler, with all patients continuing to take the ICS/LABA/LAMA that they were receiving on entry to the study as background COPD therapy for the duration of follow-up. The CHF6523 dose was selected on the basis of pharmacokinetic and pharmacodynamic modelling from healthy volunteer data in Parts 1 and 2 (it was slightly above the predicted therapeutic dose, which was a total daily dose of 3 mg). Patients were randomised according to a balanced block scheme according to a pre-established randomisation list prepared by the sponsor. Patients, investigators and their staff, monitors and the sponsor’s clinical team were all blinded to treatment.

Outcomes

The primary objective was to assess the safety and tolerability of CHF6523 in patients with COPD. The secondary objective was to investigate the pharmacokinetic profile of CHF6523 in plasma in terms of:

On Day 1: Area under the curve from 0 to 30 min and 0–12 h post-dose (AUC0–30 min and AUC0–12 h), maximum concentration (Cmax), and time to maximum concentration (tmax).

On Days 20 and 24: pre-dose concentration (Ctrough).

On Day 28: Ctrough, AUC0–30 min, AUC0–12 h, Cmax, minimum concentration (Cmin), tmax, time to minimum concentration (tmin), average concentration (Cav), accumulation ratios (Rac) for Cmax and AUC0–12 h, and the apparent total body clearance (calculated as dose/area under the curve from time 0 extrapolated to 12 h post-dose) at steady state (CL/Fss).

Exploratory endpoints included PIP3, cell counts and target biomarkers of inflammation in induced sputum, plasma biomarkers, airflow obstruction, hyperinflation and lung mechanics, symptoms, and to identify biomarker(s) of drug response using proteomics (Olink and mass-spectrometry) and transcriptomics (RNA-Seq). In addition, the number and severity of cough episodes, and the use of rescue medication (percentage of days with no use and average daily use) were evaluated as exploratory endpoints.

Sample size and statistical methods

The study was not formally powered, but a total of 42 evaluable patients was deemed adequate to assess the safety and tolerability of CHF6523. Assuming a drop-out of 30%, 60 patients were to be randomised.

Fold-change from baseline in cell counts, target biomarkers in sputum and blood, and PIP3 were log-transformed and analysed using an analysis of covariance (ANCOVA) model, with treatment, patient, and period as fixed effects, and baseline value as covariate. Changes from baseline in spirometry, forced oscillation and CAT endpoints were analysed using a similar ANCOVA model, but without log transformation. Rescue medication data were analysed using an analysis of variance model, including treatment, period and patient as fixed effects. Other data, including the proportion of days on which patients coughed and the cough VAS results, are presented descriptively only. Further details, including the transcriptomics, mass-spectrometry and Olink proteomics, bioinformatics and multi-omic data integration are in the supplement.

The safety set, used in the analysis of all safety variables, was all randomised patients who received at least one dose of study drug. The pharmacokinetic and pharmacodynamic sets were all patients in the safety set excluding those without any valid pharmacokinetic or pharmacodynamic measurements, respectively, or who had major protocol deviations that impacted the pharmacokinetic or pharmacodynamic assessments, respectively. Pharmacokinetic variables were analysed using the pharmacokinetic set; exploratory pharmacodynamic variables were analysed in the pharmacodynamic set. Given the cross-over design, the inclusion of patients in the analysis sets was defined on a per-period basis.