Review Article

The Binding of Drugs by Plasma Proteins

Many drugs bind to serum transport proteins, which can affect both drug distribution and activity in the body. α₁-Acid glycoprotein (AGP) is a key transport protein for basic and neutral drugs. Both elevated levels and altered glycosylation patterns of AGP have been seen in clinical conditions such as systemic lupus erythematosus (SLE). This study developed, optimized, and used the method of ultrafast affinity extraction (UAE) to examine whether these changes in AGP are associated with changes in the binding by some drugs to this transport protein. This approach used affinity microcolumns to capture and measure, in serum, the free fractions of several drugs known to bind AGP. These measurements were made with pooled normal control serum and serum samples from individuals with SLE. Immunoaffinity chromatography was used to obtain the content of AGP and HSA in these samples, and CE was used to examine the glycoform pattern for AGP in each serum sample. The free drug fractions measured for normal control serum ranged from 3.5 to 29.1%, in agreement with the results of ultrafiltration, and provided binding constants of ~10⁵–10⁶ M⁻¹ for the given drugs with AGP at 37⁰C. Analysis of a screening set of SLE serum samples by UAE gave decreased free fractions (relative change, 12-55%) vs normal serum when spiked with the same types and amounts of drugs. These changes were related in some cases to AGP content, with some SLE samples having AGP levels 1.3- to 2.1-fold above the upper end of the normal range. In other cases, the changes in free fractions appeared to be linked to alterations in the glycoforms and binding constants of AGP, with some affinities differing by 1.2- to 1.5-fold vs normal AGP. This approach can be employed with other solute-protein systems and to investigate binding by other drugs or transport proteins directly in clinical samples.

The focus of this work is pointing out the different behavior of two structurally related Pt^(II) complexes, the anionic cyclometalated NBu₄[(Bzq)Pt(Thio)], 1 and the neutral [(Phen)Pt(Thio)], 2, (Bzq = benzo[h]quinoline, Phen = 1,10-phenantroline, Thio = 1,2-benzenedithiolate), on the interaction with human serum albumin (HSA), a key drug-delivery protein in the bloodstream. Being very limited the number of anionic Pt^(II) complexes reported to date, this is a pioneering example of report on a protein-ligand interaction involving a negatively charged platinum compound. The study was carried out by using fluorescence spectroscopy, differential scanning calorimetry and molecular docking simulations. The results revealed a strong binding affinity between the anionic compound and the protein, whereas a weak/moderate binding interaction was highlighted for the neutral one. Comparative studies with site specific ligands (warfarin and ibuprofen), allowed us to identify the protein binding sites of the two compounds.

The work aims to shed light on the relevance of the charge in designing new drugs with a favorable binding affinity for HSA, which strongly contributes to influence their pharmacological and toxicological profile.

We recently reported that aripiprazole binds strongly to human albumin. In continuing our investigations, we investigated the mechanism responsible for the binding and the related interactions of aripiprazole with α1-acid glycoprotein (AGP). The extrinsic Cotton effects for the binding of aripiprazole and its derivatives to AGP were generated, but the magnitudes of the induced circular dichroism intensities did not correlate with those for the binding affinities. It therefore appears that the binding mode of aripiprazole with AGP is somewhat complicated, compared with that of albumin. Isothermal titration calorimetry data obtained for the binding of aripiprazole with AGP were different from that for albumin systems in that the 3 driving reactions, entropy-driven, enthalpy-driven, and the entropy-enthalpy mixed type, were all found for the AGP system, but not albumin. Moreover, the weak binding mode of aripiprazole with the 2 proteins were supported by a molecular docking model analysis. The concentration of albumin in plasma is about 50 times higher than those of AGP, but AGP levels in plasma are increased by about 10 times under inflammatory disease. Therefore, the involvement of these 2 plasma proteins should be considered in more depth for understanding the pharmacokinetics of aripiprazole.

With this issue of the Journal of Pharmaceutical Sciences, we celebrate the nearly 6 decades of contributions to mechanistic-based modeling and computational pharmaceutical sciences. Along with its predecessor, The Journal of the American Pharmaceutical Association: Scientific Edition first published in 1911, JPharmSci has been a leader in the advancement of pharmaceutical sciences beginning with its inaugural edition in 1961. As one of the first scientific journals focusing on pharmaceutical sciences, JPharmSci has established a reputation for publishing high-quality research articles using computational methods and mechanism-based modeling. The journal’s publication record is remarkable. With over 15,000 articles, 3000 notes, and more than 650 reviews from industry, academia, and regulatory agencies around the world, JPharmSci has truly been the leader in advancing pharmaceutical sciences.

The unbound fraction of a drug in plasma can profoundly influence both its pharmacokinetics and pharmacodynamics. For most drugs, the unbound fraction is relatively constant across the clinically relevant range of concentrations in a given individual. Nonlinear plasma protein binding involving saturation of binding sites results in increasing unbound fraction as the concentration of the drug increases, a phenomenon that is consistent with the law of mass action and is well recognized. Not widely appreciated is that some drugs undergo atypical concentration-dependent binding to plasma proteins, whereby the unbound fraction decreases with increasing concentration. In this article we review the drugs for which atypical nonlinear plasma protein binding has been reported. For each drug, the evidence for the phenomenon is presented and the proposed mechanism discussed. Also reviewed are the potential implications of atypical nonlinearity in plasma protein binding. Highlighted is the importance of understanding the relationship between unbound fraction and plasma drug concentration during the preclinical and early clinical stages of drug development, and during the routine clinical use of a drug especially if therapeutic drug monitoring is used to assist in optimization of the dosing regimen. The lesson is that the unexpected concentration-dependent behavior that has been observed for a number of drugs should be expected to occur for some other drugs.

Cardiotoxicity assessment using human induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs) forms a key component of the Comprehensive in Vitro Proarrhythmia Assay (CiPA). A potentially impactful factor on iPSC-CM testing is the presence of serum in the experimental media. Generally, serum-free media is used to most accurately reproduce “free” drug concentration. However, caution is needed; drug solubility and cardiomyocyte electrophysiology could be affected by media formulation, potentially impacting interpretation of drug-induced effects.

Effects of 25 drugs on properties of spontaneous field potentials in iPSC-CMs were assayed using a high-throughput microelectrode array (MEA) in two media formulations: serum-containing and serum-free. Comparative analysis was conducted on rate-corrected field potential duration (FPDc) and prevalence of arrhythmic events. Further MEA experiments were conducted, varying percentages of serum as well as carbon substrate components. Comparative LC-MS/MS analysis was done on two compounds to evaluate drug concentrations.

In serum-free media, 9 drugs prolonged FPDc. In serum-containing, 11 drugs prolonged FPDc. Eighteen drugs induced arrhythmias, 8 of these induced arrhythmias at lower concentrations in serum-containing media. At the highest non-arrhythmic concentrations, 13 of 25 drugs exhibited significant differences in FPDc prolongation/shortening between the media. Increasing fractions of serum in media yielded higher FPDc measurements. LC-MS/MS analysis of moxifloxacin and quinidine showed higher concentrations in serum-containing media.

The present study highlights media formulation as an important consideration for cardiac safety testing with iPSC-CMs. Results described here suggest that media formulation influences both compound availability and baseline electrophysiological properties. Special attention should be paid to media for future iPSC-CM assays.