Colistimethate Acidic Hydrolysis Revisited: Arrhenius Equation Modeling Using UPLC-QToF MS

Coacervation in Slow Motion: Kinetics of Complex Micelle Formation Induced by the Hydrolysis of an Antibiotic Prodrug

Colistin methanesulfonate (CMS) is the less-toxic prodrug of highly nephrotoxic colistin. To develop and understand highly necessary new antibiotic formulations, the hydrolysis of CMS to colistin must be better understood. Herein, with the addition of poly(ethylene oxide)-b-poly(methacrylic acid) (PEO-b-PMAA) to CMS, we show that we can follow the hydrolysis kinetics, employing small-angle X-ray scattering (SAXS) through complex coacervation. During this hydrolysis, hydroxy methanesulfonate (HMS) groups from CMS are cleaved, while the newly formed cationic amino groups complex with the anionic charge from the PMAA block. As the hydrolysis of HMS groups is slow, we can follow the complex coacervation process by the gradual formation of complex micelles containing activated antibiotics. Combining mass spectrometry (MS) with SAXS, we quantify the hydrolysis as a function of pH. Upon modeling the kinetic pathways, we found that complexation only happens after complete hydrolysis into colistin and that the process is accelerated under acidic conditions. At pH = 5.0, effective charge switching was identified as the slowest step in the CMS conversion, constituting the rate-limiting step in colistin formation.

Simple and robust LC-MS/MS method for quantification of colistin methanesulfonate and colistin in human plasma for therapeutic drug monitoring

A rapid, simple, and robust LC-MS/MS method was developed and validated for the quantitation of colistin and colistin methanesulfonate (CMS) in human plasma. The method also prevented overestimation of colistin concentration by establishing the stability of CMS under sample preparation conditions, including blood and plasma storage conditions. Polymyxin B1 was used as an internal standard, and positive-ion electrospray ionization in multiple reaction monitoring mode was used for quantification. Chromatographic separation was achieved using a Zorbax eclipse C18 column (3.5 µm, 2.1 mm i.d. × 100 mm), with a flow rate of 0.5 mL/min, 5 μL injection volume, and gradient elution with a mixture of acetonitrile-water (containing 0.1 % trifluoroacetic acid). The method had a quantifiable range of 0.043-8.61 and 0.057-11.39 μg/mL for colistin A and B in human plasma, respectively, under a total runtime of 6.0 min. Further, it demonstrated appropriate extraction efficiency, no significant interference from co-eluting endogenous compounds, and satisfactory intraday and interday precision and accuracy. The proposed procedure for sample preparation successfully addressed the issue of CMS instability, consequently diminishing the probability of overestimating the concentration of colistin. Therefore, this simple and robust LC-MS/MS method for CMS and colistin quantification in human plasma is a valuable tool for clinicians to accurately monitor colistin treatment in patients with infections caused by multidrug-resistant (MDR) Gram-negative bacteria.

HPLC Determination of Colistin in Human Urine Using Alkaline Mobile Phase Combined with Post-Column Derivatization: Validation Using Accuracy Profiles

In this study, the development, validation, and application of a new liquid chromatography post-column derivatization method for the determination of Colistin in human urine samples is demonstrated. Separation of Colistin was performed using a core–shell C₁₈ analytical column in an alkaline medium in order (i) to be compatible with the o-phthalaldehyde-based post-column derivatization reaction and (ii) to obtain better retention of the analyte. The Colistin derivative was detected spectrofluorometrically (λ_ext/λ_em = 340/460 nm) after post-column derivatization with o-phthalaldehyde and N-acetyl cysteine. The post-column derivatization parameters were optimized using the Box–Behnken experimental design, and the method was validated using the total error concept. The β-expectation tolerance intervals did not exceed the acceptance criteria of ±15%, meaning that 95% of future results would be included in the defined bias limits. The limit of detection of the method was adequate corresponding to 100 nmol·L⁻¹. The mean analytical bias (expressed as relative error) in the spiking levels was suitable, being in the range of −2.8 to +2.5% for both compounds with the percentage relative standard deviation lower than 3.4% in all cases. The proposed analytical method was satisfactorily applied to the analysis of the drug in human urine samples.