Protein binding of glycopeptide antibiotics with diverse physical-chemical properties in mouse, rat, and human serum

Protein Binding in Translational Antimicrobial Development-Focus on Interspecies Differences

Background/Introduction: Plasma protein binding (PPB) continues to be a key aspect of antibiotic development and clinical use. PPB is essential to understand several properties of drug candidates, including antimicrobial activity, drug-drug interaction, drug clearance, volume of distribution, and therapeutic index. Focus areas of the review: In this review, we discuss the basics of PPB, including the main drug binding proteins i.e., Albumin and α-1-acid glycoprotein (AAG). Furthermore, we present the effects of PPB on the antimicrobial activity of antibiotics and the current role of PPB in in vitro pharmacodynamic (PD) models of antibiotics. Moreover, the effect of PPB on the PK/PD of antibiotics has been discussed in this review. A key aspect of this paper is a concise evaluation of PPB between animal species (dog, rat, mouse, rabbit and monkey) and humans. Our statistical analysis of the data available in the literature suggests a significant difference between antibiotic binding in humans and that of dogs or mice, with the majority of measurements from the pre-clinical species falling within five-fold of the human plasma value. Conversely, no significant difference in binding was found between humans and rats, rabbits, or monkeys. This information may be helpful for drug researchers to select the most relevant animal species in which the metabolism of a compound can be studied for extrapolating the results to humans. Furthermore, state-of-the-art methods for determining PPB such as equilibrium dialysis, ultracentrifugation, microdialysis, gel filtration, chromatographic methods and fluorescence spectroscopy are highlighted with their advantages and disadvantages.

Uncoupling of invasive bacterial mucosal immunogenicity from pathogenicity

There is the notion that infection with a virulent intestinal pathogen induces generally stronger mucosal adaptive immunity than the exposure to an avirulent strain. Whether the associated mucosal inflammation is important or redundant for effective induction of immunity is, however, still unclear. Here we use a model of auxotrophic Salmonella infection in germ-free mice to show that live bacterial virulence factor-driven immunogenicity can be uncoupled from inflammatory pathogenicity. Although live auxotrophic Salmonella no longer causes inflammation, its mucosal virulence factors remain the main drivers of protective mucosal immunity; virulence factor-deficient, like killed, bacteria show reduced efficacy. Assessing the involvement of innate pathogen sensing mechanisms, we show MYD88/TRIF, Caspase-1/Caspase-11 inflammasome, and NOD1/NOD2 nodosome signaling to be individually redundant. In colonized animals we show that microbiota metabolite cross-feeding may recover intestinal luminal colonization but not pathogenicity. Consequent immunoglobulin A immunity and microbial niche competition synergistically protect against Salmonella wild-type infection.

Evaluation of Tumor-Targeting Properties of an Antagonistic Bombesin Analogue RM26 Conjugated with a Non-Residualizing Radioiodine Label Comparison with a Radiometal-Labelled Counterpart

Radiolabelled antagonistic bombesin analogues are successfully used for targeting of gastrin-releasing peptide receptors (GRPR) that are overexpressed in prostate cancer. Internalization of antagonistic bombesin analogues is slow. We hypothesized that the use of a non-residualizing radioiodine label might not affect the tumour uptake but would reduce the retention in normal organs, where radiopharmaceutical would be internalized. To test this hypothesis, tyrosine was conjugated via diethylene glycol linker to N-terminus of an antagonistic bombesin analogue RM26 to form Tyr-PEG₂-RM26. [¹¹¹In]In-DOTA-PEG₂-RM26 was used as a control with a residualizing label. Tyr-PEG₂-RM26 was labelled with ¹²⁵I with 95% radiochemical purity and retained binding specificity to GRPR. The IC₅₀ values for Tyr-PEG₂-RM26 and DOTA-PEG₂-RM26 were 1.7 ± 0.3 nM and 3.3 ± 0.5 nM, respectively. The cellular processing of [¹²⁵I]I-Tyr-PEG₂-RM26 by PC-3 cells showed unusually fast internalization. Biodistribution showed that uptake in pancreas and tumour was GRPR-specific for both radioconjugates. Blood clearance of [¹²⁵I]I-Tyr-PEG₂-RM26 was appreciably slower and activity accumulation in all organs was significantly higher than for [¹¹¹In]In-DOTA-PEG₂-RM26. Tumor uptake of [¹¹¹In]In-DOTA-PEG₂-RM26 was significantly higher than for [¹²⁵I]I-Tyr-PEG₂-RM26, resulting in higher tumour-to-organ ratio for [¹¹¹In]In-DOTA-PEG₂-RM26 at studied time points. Incorporation of amino acids with hydrophilic side-chains next to tyrosine might overcome the problems associated with the use of tyrosine as a prosthetic group for radioiodination.

Development of a Physiologically Based Pharmacokinetic Modelling Approach to Predict the Pharmacokinetics of Vancomycin in Critically Ill Septic Patients

BACKGROUND AND OBJECTIVES

Sepsis is characterised by an excessive release of inflammatory mediators substantially affecting body composition and physiology, which can be further affected by intensive care management. Consequently, drug pharmacokinetics can be substantially altered. This study aimed to extend a whole-body physiologically based pharmacokinetic (PBPK) model for healthy adults based on disease-related physiological changes of critically ill septic patients and to evaluate the accuracy of this PBPK model using vancomycin as a clinically relevant drug.

METHODS

The literature was searched for relevant information on physiological changes in critically ill patients with sepsis, severe sepsis and septic shock. Consolidated information was incorporated into a validated PBPK vancomycin model for healthy adults. In addition, the model was further individualised based on patient data from a study including ten septic patients treated with intravenous vancomycin. Models were evaluated comparing predicted concentrations with observed patient concentration-time data.

RESULTS

The literature-based PBPK model correctly predicted pharmacokinetic changes and observed plasma concentrations especially for the distribution phase as a result of a consideration of interstitial water accumulation. Incorporation of disease-related changes improved the model prediction from 55 to 88% within a threshold of 30% variability of predicted vs. observed concentrations. In particular, the consideration of individualised creatinine clearance data, which were highly variable in this patient population, had an influence on model performance.

CONCLUSION

PBPK modelling incorporating literature data and individual patient data is able to correctly predict vancomycin pharmacokinetics in septic patients. This study therefore provides essential key parameters for further development of PBPK models and dose optimisation strategies in critically ill patients with sepsis.

Protective effect of inactive ingredients against nephrotoxicity of vancomycin hydrochloride in rats

A generic form of vancomycin for I.V. infusion (MEEK) is more soluble and stable than the brand-name form of vancomycin hydrochloride (VCM) due to the addition of two inactive ingredients: D-mannitol and Macrogol400 (PEG400). The aim of the present study was to compare the nephrotoxicity of MEEK with that of brand-name VCM (S-VCM) and to analyze the pharmacokinetics of these preparations. Following administration to rats at the clinical dose of 40 mg/kg, there was no difference between MEEK and S-VCM with regard to pharmacokinetics and effects on the kidneys, indicating that MEEK should be as effective as S-VCM. When administered at the nephrotoxic dose of 400 mg/kg, S-VCM caused impairment of renal function and kidney damage, and an increase of the plasma concentration due to decreased renal clearance was observed. In contrast, MEEK had virtually no effect on renal function or the kidneys and did not cause a marked change of renal clearance. These findings suggest that the inactive ingredients in MEEK play a role in reducing the nephrotoxicity of VCM.

In vitro studies of teicoplanin binding to rat tissues and erythrocytes

Teicoplanin is a large polar antibiotic with a large distribution volume (Vss = 1.2-2.8 l/kg) despite extensive binding to plasma albumin. To understand this observation, binding of 3H-teicoplanin to 10% rat tissue homogenates was determined in vitro by ultracentrifugation in the presence of teicoplanin (1-30 microg/ml). Binding and efflux from erythrocytes were also studied. The in vitro total-to-unbound tissue concentration ratio (Kpu) differed widely but for each tissue was concentration independent. Upon correction for the plasma unbound fraction, there was discrepancy between the in vitro and in vivo tissue-to-plasma concentration ratios (KP), the latter calculated from previously published tissue and plasma concentration-time data using the area method. Moreover, calculation of Vss from in vitro Kp (8.10 l/kg) overestimated the in vivo value. These results suggest that in vivo teicoplanin binds to cell membranes and enters some but not all cells, such as erythrocytes.

A closer look at vancomycin, teicoplanin, and antimicrobial resistance

The worldwide increase in the incidence of resistant Gram-positive infections has renewed interest in the glycopeptide class of antimicrobial agents. Two glycopeptides are available in many parts of the world--vancomycin and teicoplanin. These two agents appear to differ in several respects, including: potential for selecting microbial resistance, dosing convenience, safety, and efficacy in severe infection. Teicoplanin appears to have lower toxicity and greater convenience; however, its widespread acceptance has been plagued by concerns over antimicrobial resistance, efficacy, and appropriate dosing. A review of available studies suggests that teicoplanin, when dosed at 6 mg/kg/day, is better tolerated than vancomycin 15 mg/kg/q12h; however, at these doses, it appears to be somewhat less effective than vancomycin in serious Staphylococcus aureus infection, such as endocarditis. Although higher doses of teicoplanin, 12 mg/kg/day to 30 mg/kg/day, have been associated with efficacy comparable to that of vancomycin in serious S. aureus infections, such doses may eliminate some of the safety advantages conferred by lower teicoplanin doses. Teicoplanin has been associated with resistance among coagulase-negative staphylococci and the selection of resistance in S. aureus. There is some evidence that widespread use of teicoplanin might accelerate the development of S. aureus resistance to both teicoplanin and vancomycin. The selection of an appropriate glycopeptide in an individual patient should be based not only on convenience, but also on a determination of optimal efficacy, safety at an efficacious dose, and the potential for resistance.

Serum protein-binding characteristics of vancomycin

A synthesis of studies of serum protein binding of vancomycin and its reported abnormal binding in serum with very high concentrations of immunoglobulin A (IgA) suggests that this antibiotic may be bound to more than one serum protein. Using an ultrafiltration method for separating free from bound drug and high-performance liquid chromatography to measure drug concentration, we studied the binding characteristics of vancomycin for alpha-1 acid glycoprotein, IgG, IgM, IgA, and albumin. The results showed that vancomycin does not bind to alpha-1 acid glycoprotein, IgG, or IgM. Major binding to albumin and IgA occurs, and total drug binding to serum proteins can be fully explained by binding to these two proteins. We calculated an N (number of binding sites per molecule) of 1.3 +/- 0.4 and a K (association constant) of 3.3 x 10(5) +/- 6.3 x 10(4) M-1 (NK = 4.3 x 10(5) M-1) for binding to IgA, whereas the corresponding NK value for albumin was only 527.5 M-1, indicating that vancomycin preferentially binds to IgA. Very high concentrations of IgA in serum (i.e., grams per deciliter), such as in patients with IgA myeloma, may result in the paradox of high (total) concentrations of vancomycin in serum that may be clinically ineffective.

The influence of vancomycin concentration and the pH of plasma on vancomycin protein binding

A review of numerous studies of the protein binding of vancomycin suggests major discrepancies among their results. The reported percent protein binding of vancomycin varies from 0% to 98%. The influence of pH and concentration on the protein binding of vancomycin was investigated in this study. There was a significant difference (p < 0.001) in percent protein binding in vancomycin-spiked plasma samples across the pH range of 7.0-8.0. There was no significant difference (p > 0.05) in percent protein binding in vancomycin-spiked plasma samples across the concentration range of 2-80 mg/L. It is likely that some of the variation reported to date may be due to a lack of control of pH during the measurement of protein binding of vancomycin.

Pharmacokinetics of individual components of teicoplanin in man

Teicoplanin is a new antibiotic consisting of closely related glycopeptides. Following an iv bolus of 400 mg teicoplanin, the pharmacokinetics of the individual components A3-1, A2-1, A2-2, A2-3, A2-4, and A2-5 was studied in five healthy volunteers by HPLC. For each subject, plasma and urine data of the individual components were simultaneously fitted by a triexponential disposition model. No significant differences were observed between the components of the A2 group in the initial volume of distribution, 0.05-0.06 L/kg, and the half-life of the second disposition phase, 2.5-3.0 hr. Significant differences were found in the volume of distribution at steady state (Vss 0.42-0.92 L/kg), the half-lives of the first (0.18-0.26 hr) and the third (48.1-66.8 hr) disposition phases, the total clearance (CL 5.4-19.3 ml/hr per kg), the renal clearance (CLR 2.8-16.1 ml/hr per kg), and the percentage of the administered dose excreted in urine (Ae 53-85%). A highly significant correlation was found between the lipophilicity of the individual components increasing from A2-1 to A2-5, and the values of the kinetic parameters. As the lipophilicity increases the fraction unbound in plasma, Vss, CL, CLR, and Ae decrease, whereas the unbound steady state volume of distribution and the unbound nonrenal clearance increase. A modest degree of accumulation of each teicoplanin component in plasma is predicted to occur at steady state following repeated administration of teicoplanin given daily, with accumulation slightly higher for the more lipophilic components A2-4 and A2-5.

Protein binding of vancomycin in a patient with immunoglobulin A myeloma

Atypical vancomycin pharmacokinetics were observed in an immunoglobulin A myeloma patient. Drug concentrations in serum were extremely elevated, the elimination half-life was prolonged despite normal renal function, and the vancomycin therapy was ineffective. Extensive binding of vancomycin, presumably by high concentrations of an aberrant immunoglobulin A protein, may have accounted for these observations.

Effect of cardiopulmonary bypass on vancomycin and netilmicin disposition

The effect of cardiopulmonary bypass (CPB) on the disposition of vancomycin (15 mg/kg) and of netilmicin (3 mg/kg) was studied in 10 adults. The concentration-time profile of the drug in serum and renal clearance were characterized pre-CPB, during CPB, and post-CPB. Vancomycin and netilmicin exhibited initial decreases in mean concentrations in serum of 4.0 mg/liter (16.8%) and 2.2 mg/liter (29.1%), respectively, upon initiation of CPB. Netilmicin concentrations in serum rebounded to a mean of 0.6 mg/liter (15.4%) within 90 min on CPB and then continuously decreased. Vancomycin concentrations in serum demonstrated a rebound increase of 2.3 mg/liter (23.5%) at the end of CPB when the aorta was unclamped. Mean renal clearance throughout CPB was decreased for vancomycin (58.4 to 43.4 ml/min per m2) and netilmicin (53.4 to 31.5 ml/min per m2). The rebound in vancomycin concentration in serum strongly correlated with the length of time between unclamping the aorta and coming off CPB (r = 0.94), as well as with the increase in temperature upon rewarming (r = 0.92).

Vancomycin serum protein binding determination by ultrafiltration

Sixty-two serum concentrations were obtained from 12 infected patients enrolled in a vancomycin pharmacokinetic study. Both unbound and total serum vancomycin concentrations were measured using ultrafiltration and a commercial fluorescent polarization immunoassay. Ultrafiltrates were obtained by centrifugation at 1000 X g for ten minutes at room temperature and their assay indicated a range in protein binding from 7.9 to 71 percent. The mean protein binding (mean +/- SD) was 41.95 +/- 14.15 percent. No measurable adsorption of vancomycin onto the ultrafiltration membrane was noted. Orthogonal regression of unbound versus total vancomycin concentrations was described by the equation y = 0.597x-0.362 with a correlation coefficient of 0.948.