Disease-drug-drug interaction in drug repurposing of intravenous imatinib for COVID-19 ARDS
MM Said (1,2), JR Schippers (3,4), L Atmowihardjo (3,4), Y Li (1), MS van der Plas (1), RAA Mathot (1), HJ Bogaard (3,4), LDJ Bos (3,4) , J Aman (3,4), EL Swart (1,2,5), IH Bartelink (1,2)
(1) Department of Pharmacy and Clinical Pharmacology, Amsterdam UMC, location VUmc, Amsterdam, The Netherlands. (2) Cancer Center Amsterdam, Amsterdam, The Netherlands. (3) Department of Pulmonary Medicine, Amsterdam UMC, location VUmc, Amsterdam, The Netherlands. (4) Amsterdam Cardiovascular Sciences, Amsterdam, The Netherlands. (5) Amsterdam Institute for Infection and Immunity, Amsterdam, The Netherlands.
Introduction:
In CounterCOVID, the treatment of coronavirus disease 2019 (COVID-19) patients resulted in improved survival rates and reduced the duration of mechanical ventilation [1]. Pharmacokinetic (PK) analysis indicated increased total imatinib exposure due to increased alpha-1-acid glycoprotein (AAG) levels when compared to chronic myeloid leukemia (CML) and gastrointestinal stromal tumor (GIST) patients after oral imatinib treatment [2]. An intravenous (IV) formulation may reduce PK variability compared to oral among patients in the intensive care unit (ICU) and was tested against a placebo in mechanically ventilated patients with COVID-19 ARDS in the InventCOVID trial [3]. During InventCOVID, patients received IL-6 receptor (IL-6R) inhibitors as standard of care. We hypothesize that total imatinib exposure is reduced and clearance of unbound imatinib is increased due to increased disease severity and when co-administered with IL-6R inhibitors, while decreasing PK variability in InventCOVID due to IV administration. Subsequently, the prior AAG-PK model may underpredict total and overpredict unbound concentrations.
Objectives:
- To evaluate the predictive performance of the previously developed AAG-PK model.
- To identify additional covariates on imatinib PK if model evaluation suggested that the InventCOVID data was not adequately described.
Methods:
The AAG-PK model was based on the hospitalized and ICU patients from the CounterCOVID and CML/GIST outpatient cohort receiving 100-800 mg oral imatinib once daily was evaluated using InventCOVID data. InventCOVID patients received 200 mg IV imatinib twice a day. Model evaluation was performed using prediction-corrected visual predictive check (pcVPC) simulations and mean prediction errors (PE). Further model development was performed if the mean PE was > 30%. Other covariates examined were body weight, age, gender, albumin, alanine transaminase, aspartate aminotransferase, estimated GFR, IL-6, concomitant treatment and ICU stay.
Results:
160 total and 109 unbound plasma concentration samples from 32 patients were collected for PK analysis. At baseline, InventCOVID patients were similar in median age (P=0.565) and weight (P=0.573) to CounterCOVID patients. InventCOVID patients had higher baseline WHO score (7 vs 4, P=<0.0001) and 100% vs. 19.8% received mechanical ventilation. InventCOVID patients had significantly higher IL-6 (P=<0.001) and lower AAG levels (P=0.007) as 90.6% of patients received an IL-6R inhibitor. In the InventCOVID trial, total imatinib steady state concentrations (Css) was significantly lower compared to CounterCOVID (842, coefficient of variation (CV) 65.7% vs 1600 ng/mL, CV 66.3%, P=<0.001) while unbound concentrations increased (48.6, CV 31.4% vs 40.2 ng/mL, CV 53.3%, P=<0.001). Median free fraction of imatinib at steady state in InventCOVID trial was significantly increased (5.74%) compared to CounterCOVID (2.70%, P=<0.0001) and previous CML/GIST patients (3.97%, P=<0.0001).
The AAG-PK model overpredicted total steady state concentrations (Css) in INVENT with PE at 83.9% ± SD 48.5% and underpredicted unbound concentrations at-11.4% ± SD 32.3%. A new parameter estimate for the AAG dissociation constant (KD) in InventCOVID population improved the prediction of all data (ΔOFV -50.9). KD was significantly higher in the InventCOVID compared to CounterCOVID/CML/GIST populations (702, CV 4% vs 336 ng/mL, CV 1%). Age, body weight and IL-6 inhibitors and WHO scorewere predictive of unbound clearance.
Conclusions: To our knowledge, this is the first study to describe a potential interaction between COVID-19 ARDS, imatinib and IL-6R inhibitors. In the InventCOVID trial, either disease severity and/or the use of IL-6R inhibitors resulted in a decrease in AAG levels compared to CounterCOVID, thereby lowering total imatinib exposure. IL-6 signal inhibition may have resulted in a decreased affinity for AAG, and reversed IL-6-mediated CYP enzyme suppression. The decreased variability and slight increase in Css of unbound imatinib in InventCOVID compared to CounterCOVID could be due to bypassing absorption route. For drugs that have low hepatic extraction and high protein binding, free fraction and metabolism might be modified due to pathophysiological conditions and concomitant treatment. The developed model could be used to explore and optimize dose-exposure-efficacy and –toxicity relationships.
References: [1] Aman, J., et al., Imatinib in patients with severe COVID-19: a randomised, double-blind, placebo-controlled, clinical trial. Lancet Respir Med, 2021. 9(9): p. 957-968.
[2] Bartelink, I.H., et al., Elevated acute phase proteins affect pharmacokinetics in COVID-19 trials: Lessons from the CounterCOVID - imatinib study. CPT Pharmacometrics Syst Pharmacol, 2021. 10(12): p. 1497-1511.
[3] Atmowihardjo, L., et al., The INVENT COVID trial: a structured protocol for a randomized controlled trial investigating the efficacy and safety of intravenous imatinib mesylate (Impentri(R)) in subjects with acute respiratory distress syndrome induced by COVID-19. Trials, 2022. 23(1): p. 158.
[4] Lee, E.B., et al., Disease-Drug Interaction of Sarilumab and Simvastatin in Patients with Rheumatoid Arthritis. Clin Pharmacokinet, 2017. 56(6): p. 607-615.