Determination of plasma protein binding of dalbavancin

Abstract

Objectives

Dalbavancin is a lipoglycopeptide with a long half-life, making it a promising treatment for infections requiring prolonged therapy, such as complicated Staphylococcus aureus bacteraemia. Free drug concentration is a critical consideration with prolonged treatment, since free concentration–time profiles may best correlate with therapeutic effect. In support of future clinical trials, we aimed to develop a reliable and reproducible assay for measuring free dalbavancin concentrations.

Methods

The ultracentrifugation technique was used to determine free dalbavancin concentrations in plasma at two concentrations (50 and 200 mg/L) in duplicate. Centrifuge tubes and pipette tips were treated for 24 h before use with Tween 80 to assess adsorption. Dalbavancin concentrations were analysed from the plasma samples (total) and middle layer samples (free) by LC/MS/MS with isotopically labelled internal standard. Warfarin served as a positive control with known high protein binding.

Results

Measurement of free dalbavancin was sensitive to adsorption onto plastic. Treatment of tubes and pipette tips with ≥2% Tween 80 effectively prevented drug loss during protein binding experiments. By the ultracentrifugation method, dalbavancin’s protein binding was estimated to be approximately 99%.

Conclusions

Dalbavancin has very high protein binding. Given dalbavancin’s high protein binding, accurate measurement of free dalbavancin concentrations should be a key consideration in future exposure–response studies, especially clinical trials. Future investigations should confirm if the active fraction is best predicted by the free or total fraction.

Introduction

Staphylococcus aureus remains a leading cause of bacteraemia and endocarditis, with mortality rates ranging from 15% to 50%.^1–3 Complicated S. aureus bacteraemia generally requires 4–6 weeks of IV therapy delivered by a peripherally inserted central catheter (PICC), which confers an additional risk of catheter-associated complications, such as infection or thrombosis.⁴ Dalbavancin is a semi-synthetic lipoglycopeptide active against Gram-positive pathogens, including S. aureus. Dalbavancin’s long half-life makes it a promising alternative treatment option, with the potential to achieve therapeutic drug levels for 4–6 weeks using a two-dose regimen, averting the need for PICC placement.^5–7

A critical consideration in prolonged therapy, especially for infections with a high rate of metastatic complications, such as S. aureus bacteraemia, is the unbound or free concentration–time profile. In order to be clinically effective, antibiotics need to diffuse from the plasma across capillary membranes into the target tissue where infection resides, such as skin and soft tissues, bone, joints or lung. Consequently, free plasma concentration may better correlate with tissue penetration and therapeutic effect than total plasma concentration.^8,9

Dalbavancin’s plasma protein binding remains poorly studied and has been reported to range between 90% and 99%, as summarized in Table 1.^10–12 Various methods have been used to estimate dalbavancin’s protein binding, given that non-specific binding to experimental equipment hinders assessment by ultrafiltration and dialysis methods.^12,13 Additionally, practical limitations to achievable radiochemical purity can affect quantitation by radiolabelling methods. We sought to develop and validate a reliable assay for free dalbavancin concentration in plasma to better link drug concentrations with patient outcomes in future clinical trials.

Table 1.

Comparison of protein binding estimates for dalbavancin in previous studies

Reference	Method	Plasma source	Reported dalbavancin protein binding (%)
Current study	ultracentrifugation	human	99
Andes and Craig10	MIC testing in mouse serum	mouse	98.4
Candiani et al.11	isothermal titration microcalorimetry	rat	>98
Product IND12	equilibrium dialysis	rat; dog; human	93.4–94.6; 90.4–94.1; 90.3–94.0

IND, investigational new drug.

Methods

Ultracentrifugation was used to determine free dalbavancin concentrations in human plasma at two concentrations (50 and 200 mg/L) in duplicate, as it is the preferred method to quantify in vitro plasma protein binding for ‘sticky’ compounds like dalbavancin.¹³ These dalbavancin concentrations were analysed as they approximated the post-distributional maximum total concentration on day 1 and minimum concentration on days 14–21 post-dalbavancin 1500 mg IV × 1.¹⁴ Samples were centrifuged at 400 000 g (106 000 RPM in a TLA-120.1 rotor) for 4 h at 37°C. Dalbavancin concentrations were analysed from the plasma samples (total drug) and middle layer samples (free drug) by LC/MS/MS with isotopically labelled internal standard (see the assay reports available as Supplementary data at JAC Online). To mitigate drug loss from non-specific binding to plastic, we pretreated test tubes and pipette tips for 24 h with Tween 80. A range of Tween 80 concentrations from 0% to 6% was assessed. Warfarin served as a positive control with known high protein binding and was assessed at concentrations of 5 and 50 mg/L.^13,15

The linearity of dalbavancin in K₂EDTA human plasma was demonstrated over the calibration range 0.5–500 mg/L. Intra-day precision [coefficient of variation (CV)] ranged from 1.21% to 2.73% and the accuracy [relative error (RE)] ranged from −3.62% to 9.80%, while inter-day precision (CV) ranged from 1.25% to 3.86% and the accuracy (RE) ranged from −5.68% to 5.27%. Inter-day precision and accuracy of total dalbavancin concentrations of 1, 5, 25, 100, 250, 400 and 500 mg/L were tested in five different K₂EDTA human single donor plasma lots, while six plasma lots were used for the lower limit of quantification concentration of 0.5 mg/L. The linearity of dalbavancin in human K₂EDTA plasma/protein-free human K₂EDTA plasma (1:1) was demonstrated over the calibration range 0.05–50 mg/L. Intra-day precision (CV) ranged from 0.46% to 19.09% and accuracy ranged from −7.67% to 11.00%. Inter-day precision (CV) ranged from 0.59% to 15.33% and accuracy (RE) ranged from −5.74% to 9.78%. Inter-day precision and accuracy of dalbavancin concentrations of 0.05, 0.1, 0.25, 1, 2.5, 10, 20, 40 and 50 mg/L were tested in three different single donor lots of human K₂EDTA plasma/protein-free human K₂EDTA plasma. Single donor lots of K₂EDTA human plasma for development of dalbavancin LC/MS/MS procedures for quantification of dalbavancin were obtained from BioIVT (https://bioivt.com/normal-human-plasma).

Results

Warfarin’s free fraction was 0.61% (95% CI = 0.59%–0.63%) at 5 mg/L and 0.73% (95% CI = 0.71%–0.75%) at 50 mg/L, consistent with previously reported protein binding of >99%.¹⁵ The addition of 0%–2% Tween 80 did not affect protein binding results for warfarin (Figure S1, available as Supplementary data at JAC Online).

Measurement of dalbavancin was sensitive to adsorption onto plastic, as the estimated percentage of free dalbavancin was statistically significantly different in the presence or absence of Tween 80 (delta = 0.11%, 95% CI = 0.05%–0.17%, P = 0.001). By pretreating tubes and pipette tips with 2% Tween 80, we were able to mitigate error due to non-specific binding of dalbavancin to plastic, a common issue with highly protein bound drugs. Addition of Tween 80 increased the measured percentage of free dalbavancin in an asymptotic manner, which we modelled using non-linear least squares regression (R, version 3.6.0). Yield of additional measurable free drug plateaued at ≥2% Tween 80 (Figure 1). In protein binding experiments using tubes and pipette tips pretreated with 2% Tween 80, the free fraction of dalbavancin was 0.96% (95% CI = 0.94%–0.98%) at 50 mg/L and 1.11% (95% CI = 1.08%–1.1%3) at 200 mg/L.

Figure 1.

Percentage of free dalbavancin versus various concentrations of Tween 80 for pretreatment of tubes. (a) 50 mg/L dalbavancin. (b) 200 mg/L dalbavancin.

Discussion

With an estimated protein binding of approximately 99% spanning concentrations of 50–200 mg/L, dalbavancin is a highly protein bound drug. Our results align with the study by Andes and Craig,¹⁰ which assessed the protein binding of dalbavancin by examining the impact of 95% mouse serum on the activity of dalbavancin in vitro.¹⁶ The increases in MIC observed in the presence of 95% mouse serum were consistent with the study by Candiani et al.¹¹ In the presence of 50% bovine serum, the MICs of dalbavancin increased 4 to 64 times for the Staphylococci strains tested, which is congruent with >98% protein binding. In previous equilibrium dialysis with radiolabelled dalbavancin, protein binding was only estimated to be 90%–94%. However, one of the major limitations to the measurement of protein binding with radiolabelled drug material is the radiolabelled drug material’s radiochemical purity, which is typically 97%–99%, for ¹²C-labelled drug. The extent of binding cannot be quantitatively measured beyond the radiochemical purity, resulting in potential underestimation of protein binding for highly protein bound drugs like dalbavancin.^9,17 Furthermore, the extent of protein binding of radiolabelled drugs is quantified by scintillation counting and it is well documented that sticky compounds like dalbavancin adhere to equilibrium dialysis membranes. Consequently, there is the potential that scintillation counting may overquantify the amount of radiolabelled drug in the buffer receiver chamber (corresponding to free drug) by counting radiolabelled drug that is bound to the membrane, again potentially resulting in underestimation of protein binding. Between the use of Tween 80 to prevent non-specific binding to plastic in laboratory equipment and avoidance of issues with quantification using radiolabelling, our rather simple method mitigates some of the common sources of error in accurate measurement of free drug concentrations. While the difference in percentage of free dalbavancin with or without the addition of Tween appears small on the absolute scale, a 0.11% difference represents a relatively large relative source of error when the total percentage of free dalbavancin is <1%. Accurate measurement of free drug concentrations should be a key consideration in any future dose–response studies, especially clinical trials. Given the potential differences in albumin and other circulating proteins between infected and uninfected subjects, future studies are needed to confirm our protein binding estimates in patients with infections as they may vary compared with our observed findings.¹⁸ Future studies are also required to determine whether therapeutic effect is best predicted by free or total drug concentration as this remains unverified for dalbavancin.

Funding

Supported by NIAID/NIH grant UM1AI104681. The funder had no role in the design, execution or analysis of the research presented.

Transparency declarations

Allan Xu is the Lab Director and a minority owner of Keystone Bioanalytical. All other authors: none to declare.

Disclaimer

Content is solely the authors’ responsibility and does not represent official NIH views.

Supplementary data

Assay reports and Figure S1 are available as Supplementary data at JAC Online.