Analytical and Functional Determination of Polymyxin B Protein Binding in Serum

Abstract

To enhance our understanding of the pharmacological properties of polymyxin B, serum protein binding for polymyxin B1, B2, and B3 and for isoleucine-polymyxin B1 was evaluated. Using equilibrium dialysis and ultrafiltration, comparable protein binding was found in all 4 components of polymyxin B (92% to 99%). Protein binding in human serum was further assessed using a functional assay, the results of which were in general agreement with previous findings (approximately 90%).

TEXT

Polymyxin B has grown increasingly important in tackling Gram-negative infections that are refractory to treatment by first-line agents. Polymyxin B is a basic polypeptide antibiotic commercially available as a mixture of closely related analogs, the most prominent being polymyxin B1, B2, and B3 and isoleucine-polymyxin B1. In spite of its widespread use, there are gaps in our understanding of its pharmacokinetics and pharmacodynamics. Particularly, the degree of polymyxin B protein binding remains unclear. A previous study reported the protein binding to be 96.9 to 98.4% for polymyxin B1 (1). However, a more recent study using equilibrium dialysis reported a median protein binding value of 58% (2). The extent of protein binding affects the volume of distribution and clearance of polymyxin B and, thus, its overall efficacy (3). As a result, reliable estimates of polymyxin B protein binding are vital to achieving an optimal clinical drug exposure. The objective of this study is to provide additional insights into the degree of polymyxin B protein binding in human, rat, and mouse sera.

Polymyxin B sulfate (USP) powder was purchased from Sigma-Aldrich (St. Louis, MO). A stock solution of polymyxin B in sterile water was prepared, aliquoted, and stored at −80°C. Prior to each assay, an aliquot of the drug was thawed and diluted to the desired concentrations. Sterile human, rat, and mouse sera (Equitech-Bio, Inc., Kerrville, TX) were aliquoted into 6-ml aliquots and stored at −20°C. Prior to use, sera were thawed overnight in a refrigerator (4°C) and centrifuged at 10,000 × g for 15 min to remove aggregated lipids.

Pseudomonas aeruginosa strain ATCC 27853 (American Type Culture Collection, Manassas, VA) and Acinetobacter baumannii strain ATCC BAA 747 were used in the study. The bacteria were stored at −80°C in Protect storage vials (Key Scientific Products, Round Rock, TX). Fresh isolates were subcultured twice on 5% blood agar plates (Hardy Diagnostics, Santa Maria, CA) for 24 h at 35°C prior to each experiment. The MICs of polymyxin B were previously determined to be 2 μg/ml for P. aeruginosa and 0.5 μg/ml for A. baumannii.

Equilibrium dialysis bags were purchased from Spectrum Laboratories (Rancho Dominguez, CA). An Amicon Centrifree device for ultrafiltration was purchased from Millipore Corporation (Bedford, MA). Human, rat, and mouse serum samples were spiked with polymyxin B at concentrations of 1.25, 2.5, 5, and 10 μg/ml and incubated for 2 h at 37°C. Spiked plasma samples (1 ml) were added to the equilibrium dialysis bag, and phosphate-buffered saline (PBS) was used as the receiver medium. Samples were obtained in triplicate for each concentration of polymyxin B and assayed by a validated liquid chromatography-tandem mass spectrometry (LC-MS/MS) method (4). The lower limit of quantification (LLOQ) was 12.5 ng/ml for all major components of polymyxin B.

Spiked serum samples were also assessed by ultrafiltration. After incubation at 37°C for 2 h, aliquots (1 ml, in triplicate) of each concentration of spiked serum were added to Centrifree ultrafiltration tubes. The ultrafiltrate was obtained by centrifugation of the serum samples at 2,000 × g for 30 min. The concentration of unbound drug in the filtrate was determined in triplicate using the same LC-MS/MS method.

The in vitro growth rates of P. aeruginosa and A. baumannii were determined individually in full-strength, 0.5-strength, and 0.25-strength cation-adjusted Mueller-Hinton broth (CA-MHB) and compared to the growth profile in human serum in order to normalize the experimental conditions. One to two fresh colonies of each strain were inoculated in CA-MHB until reaching late-log-phase growth. Based on the absorbance at 630 nm, the bacterial suspension was diluted to achieve an inoculum of approximately 5 × 10⁵ CFU/ml. The experiment was conducted in a shaker water bath at 37°C. Serial samples of 250 μl were obtained over 6 h in duplicate. Quantitative culture was used to determine colony counts by plating 10× serial dilutions of the samples (100 μl) onto Mueller-Hinton agar plates. Colony counts were enumerated after incubation at 35°C in a humidified incubator for up to 24 h.

Time-kill studies were conducted with different polymyxin B concentrations in human serum and CA-MHB. Four clinically achievable concentrations of polymyxin B were selected for each strain (1). For P. aeruginosa, concentrations of 0 (control), 1, 5, and 10 μg/ml were used. For A. baumannii, 0 (control), 0.5, 2.5, and 5 μg/ml were used. The inocula were prepared as described above. Serial samples were obtained in duplicate from each flask to characterize the total bacterial population over 4 h. Prior to quantitative culture of the bacteria, the bacterial samples were centrifuged at 10,000 × g and 4°C for 15 min and reconstituted with sterile normal saline to their original volumes to minimize drug carryover. The lower limit of detection was 2.3 log CFU/ml.

Using the equilibrium dialysis method, only polymyxin B1 (the most abundant component in the USP mixture) was detectable in the receiver buffer. As shown in Table 1, the serum protein binding rates in different sera were similar, and the difference between the two methods was <4%. Furthermore, protein binding remained relatively consistent across different drug concentrations. In contrast, all major polymyxin B components were detected using ultrafiltration. Using the average values for the 4 drug concentrations examined, the protein binding rates of polymyxin B1, B2, and B3 and isoleucine-polymyxin B1 in human serum were 96.4%, 91.6%, 95.4%, and 94.7%, respectively. These findings agreed with previously published results reporting protein binding of 96.9% to 98.4% for polymyxin B1 (1). The corresponding values in rat and mouse sera ranged from 93.5% to 97.7% (data not shown).

TABLE 1.

Polymyxin B1 protein binding in various sera by different methods

Polymyxin B concn (μg/ml)	% binding (mean ± SD) by serum type and method
	Human		Rat		Mouse
	Equilibrium dialysis	Ultrafiltration	Equilibrium dialysis	Ultrafiltration	Equilibrium dialysis	Ultrafiltration
1.25	96.3 ± 1.4	96.8 ± 1.5	98.9 ± 3.3	98.1 ± 0.4	99.7 ± 1.0	98.7 ± 0.8
2.5	98.9 ± 0.3	97.1 ± 0.8	97.5 ± 0.2	97.2 ± 1.1	99.1 ± 0.2	97.8 ± 1.4
5	99.2 ± 0.2	96.3 ± 0.4	98.2 ± 0.8	96.0 ± 1.2	99.6 ± 0.04	97.1 ± 1.5
10	99.6 ± 0.03	95.4 ± 0.2	96.7 ± 0.8	92.5 ± 0.8	99.2 ± 0.5	96.9 ± 1.6
Mean	98.5	96.4	97.9	96.0	99.4	97.6

For both bacteria, the growth profiles in different media revealed that bacterial growth in human serum was most comparable to that with half-strength CA-MHB (data not shown). Consequently, half-strength CA-MHB was used in the time-kill studies. As shown in Fig. 1A, the killing profile of P. aeruginosa associated with 10 μg/ml polymyxin B in serum was more comparable to that for 1 μg/ml than 5 μg/ml polymyxin B in CA-MHB. Similarly, in Fig. 1B, the killing activity against A. baumannii by 5 μg/ml polymyxin B in serum was roughly equivalent to that with 0.5 μg/ml polymyxin B in CA-MHB, and it was distinctly different from that for 2.5 μg/ml polymyxin B in CA-MHB. If only the free (unbound) fraction of drug is pharmacologically active, protein binding of approximately 90% would be derived in both scenarios.

FIG 1.

Time-killing studies of polymyxin B in 0.5-strength CA-MHB and human serum for P. aeruginosa ATCC 27853 (A) and A. baumannii ATCC BAA 747 (B). The data are shown as means ± standard deviations (SD); MHB, Mueller-Hinton broth.

In summary, we demonstrated that polymyxin B protein binding in serum was consistent across three mammalian species. Among the individual components, protein binding was also comparable, ranging from 92% to 99% using equilibrium dialysis and ultrafiltration. Furthermore, using time-kill studies, the bactericidal activity of polymyxin B was reasonably well correlated with protein binding of approximately 90%.