AUTHOR=Corona Giuseppe , Di Gregorio Emanuela , Buonadonna Angela , Lombardi Davide , Scalone Simona , Steffan Agostino , Miolo Gianmaria 

TITLE=Pharmacometabolomics of trabectedin in metastatic soft tissue sarcoma patients

JOURNAL=Frontiers in Pharmacology

VOLUME=Volume 14 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/pharmacology/articles/10.3389/fphar.2023.1212634

DOI=10.3389/fphar.2023.1212634

ISSN=1663-9812

ABSTRACT=Trabectedin is an anti-cancer drug commonly used for the treatment of patients with metastatic soft tissue sarcoma (mSTS). Despite its recognized efficacy, significant variability in pharmacological response has been observed among mSTS patients. To address this issue, this pharmacometabolomics study aimed to identify pre-dose plasma metabolomics signatures that can explain individual variations in trabectedin pharmacokinetics and overall clinical response to treatment. In this study, 40 mSTS patients treated with trabectedin administered by 24h-intravenous infusion at a dose of 1.5 mg/m 2 trabectedin as a single agent were enrolled. The patients' baseline plasma metabolomics profiles, which included derivatives of amino acids and bile acids, were analyzed using multiple reaction monitoring LC-MS/MS together with their pharmacokinetics profile of trabectedin. Multivariate partial least squares (PLS) regression and univariate statistical analyses were utilized to identify correlations between baseline metabolite concentrations and trabectedin pharmacokinetics, while PLS-discriminant analysis (DA) was employed to evaluate associations with clinical response. The multiple regression model, derived from the correlation between the AUC of trabectedin and pre-dose metabolomics, exhibited the best performance by incorporating cystathionine, hemoglobin, taurocholic acid, citrulline, and the phenylalanine/tyrosine ratio. This model demonstrated a bias of 4.6% and a precision of 17.4% in predicting drug AUC, effectively accounting for up to 70% of the inter-individual pharmacokinetic variability.The best performing multiple regression model included cystathionine, hemoglobin, taurocholic acid, citrulline, and phenylalanine/tyrosine ratio, exhibiting a bias of 4.6% and precision of 17.4% for predicting drug AUC, and explaining up to 70% of inter-individual pharmacokinetic variability. Through the use of PLS-DA, cystathionine and hemoglobin were identified as specific metabolic signatures that effectively distinguish patients with stable disease from those with progressive disease.The findings from this study provide compelling evidence to support the utilization of pre-dose metabolomics in uncovering the underlying causes of pharmacokinetic variability of trabectedin, as well as facilitating the identification of patients who are most likely to benefit from this treatment.