Determination of colistin in luminal and parietal intestinal matrices of chicken by ultra‐high‐performance liquid chromatography‐tandem mass spectrometry

Abstract Justification for continued use of colistin in veterinary medicine, for example medicated water, relies on pharmacokinetic/pharmacodynamic (PK/PD) studies that require accurate measurement of colistin content in the digestive tract. A method for the detection and quantification of colistin in poultry intestinal material was developed and validated. Colistin is not absorbed after oral administration, and the biophase is the gastrointestinal tract. Extraction of colistin from the matrix was achieved using solid‐phase extraction with a methanol:water (1:2; v/v) solution. Polymyxin B was used as an internal standard. Colistin A and colistin B, the main components of colistin, were separated, detected and measured using ultra‐high‐performance liquid chromatography coupled with tandem mass spectrometry (UHPLC‐MS/MS). The method was validated for linearity/quadraticity between 1.1 (LOQ) and 56.7 mg/kg. Mean accuracy was between 82.7% and 107.7% with inter‐ and intra‐day precision lower than 13.3% and 15% respectively. Freeze–thaw, long‐term and bench storage were validated. Incurred samples following colistin treatment in poultry at the approved clinical dose of 75,000 IU/kg in drinking water and oral gavage were quantifiable and in line with expected intestinal transit times. The method is considered appropriately accurate and precise for the purposes of pharmacokinetic analysis in the gastrointestinal tract.

raises concerns that use in livestock production may accelerate resistance selection and dissemination in animals and humans (Liu et al., 2016;Shen et al., 2016;Walsh & Wu, 2016).
Safeguarding colistin as a last-line antibiotic requires enhanced understanding of digestive pharmacokinetics (PK) in livestock, specifically the transit of colistin through the luminal intestinal content (LIC), residual binding to parietal intestinal content (PIC) and delayed excretion due to differential rates of emptying in luminal caecal content (LCC) and parietal caecal content (PCC) (Clench & Mathias, 1995;Guyonnet et al., 2010).
Colistin contains multiple compounds: predominantly colistin A (polymyxin E 1 ) and colistin B (polymyxin E 2 ), the proportions of which vary dependent on supplier and batch, making quantification difficult (Brink et al., 2014). Previous methods include microbiological assays (Sato et al., 1972), immunological assays (Kitagawa et al., 1985) and, more recently, high-performance liquid chromatography (HPLC) coupled with tandem mass spectrometry (MS/MS) (Cangemi et al., 2016;Chepyala et al., 2015;Fu et al., 2018). These methods quantified colistin in human plasma following intravenous administration, which may not be suitable for more complex matrices, like gastrointestinal (GI) content.  Chromatographic analyses were performed on an Acquity ultraperformance liquid chromatography system with a BEH C 18 separation column (1.7 µm particle size, 2.1 × 50 mm) (Waters). The column and autosampler were maintained, respectively, at 50°C and 10°C, and the injection volume was 20 µl. Mobile phases consisted of 0.1% formic acid in water (solvent A) and 0.1% formic acid in acetonitrile (solvent B). The flow rate and solvent gradient varied according to Table S1.

Measurement of colistin in plasma is
The UHPLC system was coupled to a Xevo TQS-Micro triple quadrupole mass spectrometer (Waters). The mass spectrometer was operated with positive electrospray ionization (ESI) in multiple reaction monitoring (MRM) mode (full method in supplementary materials, Table S2). Colistin concentration was calculated as the ratio of the sum of peak areas of colistin A and B over the internal standard polymyxin B 1 peak area. Calibration curves were obtained by least-squares quadratic regression with a weighting factor (1/x²) and excluding the origin. The correlation coefficients R² of the calibration curves were above 0.99 for the 5 validation days, and regression was assessed by ANOVA (Table S3). Specificity was acceptable, with negligible carry over of 0.02 mg/kg, far below LOQ (1.1 mg/kg). Accuracy and precision at the LOQ, within run (RSDr) and between-run (RSDR) precision and accuracy were acceptable (Tables S4 and S5). Colistin spiked samples showed acceptable stability at 4°C, long-term frozen storage, stability during analysis and multiple freeze/thaw cycles, indicating that storage up to fifteen weeks was achievable and that freeze/ thawing had no significant impact on recovery (Tables S6-S8).  (Guyonnet et al., 2010;Sato et al., 1972). Although UHPLC-MS/MS methods provide more precise measurements, the preprocess purification and deproteination results in total colistin measurements, which require further analysis of the protein binding fraction to account for 'free' and unbound colistin.
Although absorption of colistin is negligible, impact of protein binding/binding to materials within the digesta may limit 'free' colistin, the subsequent antimicrobial efficacy, and how the dose is related to the pharmacokinetics in the GI tract. Guyonnet et al. (2010) demonstrated that for pig gut liquor, the apparent ratio between antimicrobial effect and colistin as measured by HPLC was 0.8:1. However, this may be different in chicken intestinal matrix due to differences in digesta, which cannot be accounted for without further study. Varying constitution of the intestinal matrix, due to dietary conditions, may impact on the accuracy of this method. A secondary study (not reported here) successfully used this method to quantify colistin in LIC with older birds (35 days old) fed a grower feed (complete flour-based feed), but further validation is needed to explore the robustness of this technique between different feeding profiles and laboratories.
Results from samples tested here show that the method is suitable to quantify colistin for developing a digestive PK profile.
Compared with the profile published by Sato et al. (1972), which showed high concentrations within the small intestine at eight hours, our study shows a more rapid elimination, with colistin levels below the LOQ within four hours of dosing cessation. This is likely related to physiological differences in gut transit time between the 6-month-old layer hens and 16-day-old broiler chicks, and impacted by methodological differences between reporting total colistin via UHPLC-MS/MS and 'free' colistin using a microbiological method.
Determination of colistin pharmacokinetics is vital for designing efficacious treatments. This paper describes a UHPLC-MS/MS method that is specific, accurate, precise and suitable for quantifying colistin in chicken intestinal matrices. Its limit of quantification was validated at 1.1 mg/kg, corresponding to the lower end of typical MIC values for pathogenic E. coli. This method is suitable for optimizing PK data and future PK/PD predictions and informing colistin usage.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

A N I M A L WE LFA R E A N D E TH I C S S TATE M E NT
This study was reviewed and approved by Royal Veterinary College ethics and welfare committee in accordance with ASPA (1986) legislation (PPL number: PCCBD6D98).

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available in the supplementary material of this article.