Evaluating the stability of colistin and colistin methanesulphonate in human plasma under different conditions of storage

Abstract

Objectives

The purpose of this study was to assess the stability of colistin and colistin methanesulphonate (CMS) in human plasma under storage conditions typically used in clinical pharmacokinetic (PK) and PK/pharmacodynamic (PD) investigations.

Methods

Human plasma (pH adjusted to 7.4) containing colistin (2 mg/L) or CMS (2 or 30 mg/L) was stored at −20, −70 or −80°C for 6–12 months. At periodic intervals, the concentrations of colistin in colistin-spiked samples, and of CMS and formed colistin in CMS-spiked samples, were analysed (n = 3 replicates at each time) by HPLC.

Results

The time course of colistin concentrations in colistin-spiked plasma showed a substantially better stability at −80 and −70°C than at −20°C. With regard to CMS-spiked plasma of 2 and 30 mg/L stored at −80 and −70°C, no quantifiable colistin formed over a 4 month period. However, the plasma spiked to 2 mg/L stored at −20°C showed a substantial concentration of colistin (∼0.4 mg/L) within 2 months. At all three storage temperatures, the stability of CMS was substantially better for the plasma spiked to contain 30 mg/L as compared with 2 mg/L.

Conclusions

The results of our long-term stability study have significant implications for those involved in conducting clinical PK and PK/PD studies with CMS/colistin.

Introduction

Colistin (polymyxin E) is a cationic polypeptide antibiotic with a narrow spectrum of antimicrobial activity.¹ Increasingly, it is considered a last resort in the treatment of severe infections caused by multidrug-resistant Gram-negative organisms. Colistin is administered parenterally and by inhalation in the form of an inactive and less toxic prodrug, colistin methanesulphonate (CMS).¹ Antibacterial activity relies upon generation of colistin from CMS. Although CMS/colistin was approved for clinical use in the 1950s, it was never subjected to contemporary drug development procedures and so a thorough understanding of its appropriate use based upon solid pharmacokinetic/pharmacodynamic (PK/PD) principles has not been achieved. Early (and even a number of recent²) studies to elucidate the PK of ‘colistin’ following administration of CMS used microbiological assays, which are incapable of differentiating the colistin present in a biological sample at the time of its collection from that formed during the incubation period of the assay.¹ Recent HPLC and liquid chromatography–mass spectrometry methods^3–5 for separate quantification of CMS and colistin have afforded the opportunity to investigate the PK of the administered prodrug, CMS, and of the colistin generated from it in vivo. It is very important to recognize, however, that CMS is unstable in an aqueous environment and in solution it undergoes conversion into colistin.^6,7 While the in vivo conversion of CMS into colistin is required for antimicrobial activity, further conversion in vitro during storage of biological samples for analysis from clinical PK studies will at best confound interpretation of the results and at worst render them meaningless. It is important, therefore, that plasma samples collected in clinical PK studies are stored in a manner that minimizes conversion of CMS into colistin. This study was designed to assess the stability of colistin and CMS in human plasma under storage conditions (−20, −70 or −80°C) typically used in clinical PK investigations.

Materials and methods

Chemicals and materials

Colistin sulphate and 9-fluorenylmethyl chloroformate (FMOC-Cl) were purchased from Sigma-Aldrich (St Louis, MO, USA). CMS sodium (Coly-Mycin M^® Parenteral) was purchased from Pfizer Pty Ltd (West Ryde, NSW, Australia). Aqueous solutions of CMS and colistin for spiking human plasma were freshly prepared. Human plasma from healthy subjects was obtained from the Australian Red Cross Blood Service (South Bank, Victoria, Australia) and was shown by HPLC analysis⁶ to not contain CMS or colistin.

Stability of colistin and CMS in human plasma

The pH of human plasma was adjusted to 7.4 with phosphoric acid (85%, w/w). Separate aliquots (40 mL) of plasma were spiked with colistin sulphate (2 mg/L, corresponding to ∼1.7 mg/L colistin) or CMS sodium (2 or 30 mg/L, corresponding to ∼1.9 or 28.0 mg/L CMS), concentrations that are clinically relevant.^5,8,9 Spiking solutions of CMS had undetectable concentrations of colistin. Concentrations of colistin and CMS were determined in three replicates on the day of preparation. Aliquots (0.6 mL) of the spiked plasma were transferred immediately to 2 mL plastic tubes (BD, catalogue number 366664) that were promptly placed into storage at −20, −70 or −80°C. In an additional study, aliquots were placed in 1.5 mL tubes (Eppendorf, North Ryde, NSW, Australia) and stored at −70°C. Concentrations of colistin in the colistin-spiked plasma, and of CMS and formed colistin in the CMS-spiked plasma, were determined after 1, 2, 3, 4, 5, 6, 8, 10 and 12 months of storage (samples stored at −70°C were studied to 6 months). At each analysis time, three replicate tubes were removed from storage for each analyte and thawed at room temperature. Sample preparation for HPLC analysis commenced immediately after thawing.

HPLC analyses

Concentrations of colistin in colistin-spiked and CMS-spiked plasma were determined using a modified HPLC method;¹⁰ plasma was deproteinized with acetonitrile only. Good linearity was achieved (r² ≥ 0.997) for colistin over a calibration range of 0.10–4.00 mg/L. Inter-day analysis of quality control samples (QCs) of 0.30 and 3.50 mg/L (n = 30 samples of each over 10 assay days across 12 months) returned results of 0.27 ± 0.023 and 3.26 ± 0.057 mg/L, respectively. The limit of quantification (LOQ) was 0.10 mg/L. CMS concentrations in CMS-spiked plasma were quantified as described previously,^6,7 with the following modifications: plasma deproteinized with acetonitrile only; use of an Onyx^® C₁₈ column (2 µm, 100 × 4.6 mm ID) (Phenomenex, Torrance, CA, USA); and a mobile phase of acetonitrile/tetrahydrofuran/water (5:2.5:2.5, v/v/v). Good linearity was achieved (r² ≥ 0.993) for CMS between 0.50 and 16.0 mg/L; where appropriate, the stability samples were diluted such that the CMS concentrations were within the calibration range. Inter-day analysis of QCs of 1.00 and 12.0 mg/L (n = 30 samples as above) returned results of 0.98 ± 0.080 and 11.7 ± 0.36 mg/L, respectively. The LOQ was 0.50 mg/L.

Results

The time course of colistin concentrations in colistin-spiked plasma stored at −20 and −80°C is shown in Figure 1. Stability of colistin was substantially better at −80°C than at −20°C. At the lower of these two temperatures, the colistin concentration in plasma remained within 7% of the initial concentration for up to 6–8 months of storage. In contrast, at −20°C this same level of stability was retained for 1 month only and by 8 months the colistin concentration had declined by ∼35%. In general, the stability of colistin in plasma stored at −70°C (data not shown) was similar to that at −80°C. The ratio of the chromatographic peak area of colistin A to colistin B remained constant over the 12 month storage period.

Figure 1.

Time course of colistin, spiked to 2 mg/L colistin (sulphate), in human plasma stored at −20 or −80°C. Data represent means ± SD (n = 3).

With regard to CMS-spiked samples, the key outcome measures were the time course of CMS concentration and that of formed colistin (Figure 2). For CMS-spiked samples of 2 and 30 mg/L stored at −80°C, CMS concentrations remained within 10% of the initial concentration and there was no quantifiable colistin formed over a 4 month storage period. Similar stability was observed at −70°C (data not shown). At −20°C, >26% degradation of CMS was observed for the 2 mg/L samples within 2 months and a substantial concentration of colistin (∼0.4 mg/L) was already evident. At both −20 and −80°C, the stability was substantially better for the plasma spiked to contain 30 mg/L as compared with 2 mg/L (Figure 2).

Figure 2.

Time course of CMS and formed colistin in human plasma stored at −20 or −80°C. (a) Plasma spiked with 2 mg/L CMS (sodium). (b) Plasma spiked with 30 mg/L CMS (sodium). Data represent means ± SD (n = 3).

Discussion

In clinical PK and PK/PD studies, it is crucial to understand the stability of the administered inactive prodrug, CMS, and the formed colistin in biological samples during storage prior to analysis. Even a small degree of ongoing conversion of the prodrug into its active form during sample storage has the potential to substantially confound the interpretation of data from clinical PK and PK/PD investigations. To the best of our knowledge, this is the first study reporting the long-term stability (i.e. up to 12 months) of CMS and colistin in human plasma stored at sub-zero temperatures often used for storage of samples from clinical studies. We carefully selected the concentrations for the CMS- and colistin-spiked plasma samples to reflect the relative concentrations that occur with currently administered dosage regimens of CMS (CMS peak plasma concentrations up to ∼30 mg/L and trough concentrations ∼2 mg/L, with generated plasma colistin concentrations of 1–5 mg/L).⁹ We spiked CMS and colistin separately into plasma to facilitate interpretation of the stability data for each compound.

It is clearly evident from our study that storage of plasma at −20°C is not acceptable, unless samples are analysed within 1 month of collection. Even with plasma stored at −80 or −70°C, samples must be analysed within 4 months of collection to avoid substantial conversion of CMS into colistin and the degradation of both entities. Any formed colistin was not quantifiable at −80 or −70°C for up to 4 months in samples containing 2 or 30 mg/L CMS. Given that the LOQ for colistin was 0.10 mg/L and taking account of the molecular weights of CMS and colistin, at least 6.2% and 0.50% of CMS at 2 and 30 mg/L, respectively, would need to have converted into colistin for it to be quantifiable; however, these figures are likely to be slight underestimates as any formed colistin would also be subject to degradation (Figure 1). Our study demonstrates the importance of conducting stability studies across the range of clinically relevant concentrations of CMS. The degradation of CMS and the formation of colistin were very substantially lower in the plasma spiked to contain 30 mg/L CMS than for that containing 2 mg/L CMS. At any given storage temperature, if the same percentage conversion of CMS into colistin occurred at both CMS concentrations, the formed colistin should have been evident earlier and at higher concentrations for the 30 mg/L CMS sample as compared with the 2 mg/L CMS sample; in fact the converse was true (Figure 2), indicating that the stability of CMS in stored plasma was highly concentration dependent. Concentration-dependent stability of CMS in other aqueous media (e.g. solutions for administration to patients) has been observed previously and the mechanism is yet to be fully elucidated.⁷ The results of our long-term stability study have significant implications for those involved in conducting clinical PK and PK/PD studies with CMS/colistin.

Funding

This work was supported by the National Institute of Allergy and Infectious Diseases at the National Insitutes of Health (award numbers R01AI079330 and R01AI070896). J. L. is an Australian National Health and Medical Research Council R. Douglas Wright Research Fellow.

Transparency declarations

None to declare.

Disclaimer

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Allergy and Infectious Diseases or the National Institutes of Health.