Transfer of ANS-Like Drugs from Micellar Drug Delivery Systems to Albumin Is Highly Favorable and Protected from Competition with Surfactant by "Reserved" Binding Sites

Micellar drug delivery systems (MDDS) for the intravenous administration of poorly soluble drugs have great advantages over alternative formulations in terms of the safety of their excipients, storage stability, and straightforward production. A classic example is mixed micelles of glycocholate (GC) and lecithin, both endogenous substances in human blood. What limits the use of MDDS is the complexity of the transitions after injection. In particular, as the MDDS disintegrate partially or completely after injection, the drug has to be transferred safely to endogenous carriers in the blood, such as human serum albumin (HSA). If this transfer is compromised, the drug might precipitate─a process that needs to be excluded under all circumstances. The key question of this paper is whether the high local concentration of GC at the moment and site of MDDS dissolution might transiently saturate HSA binding sites and, hence, endanger quick drug transfer. To address this question, we have used a new approach, which is time-resolved fluorescence spectroscopy of the single tryptophan in HSA, Trp-214, to characterize the competitive binding of GC and the drug substitute anilinonaphthalenesulfonate (ANS) to HSA. Time-resolved fluorescence of Trp-214 showed important advantages over established methods for tackling this problem. ANS has been the standard "model drug" to study albumin binding for decades, given its structural similarity to the class of naphthalene-containing acidic drugs and the fact that it is displaced from HSA by numerous drugs (which presumably bind to the same sites). Our complex global fit uses the critical approximation that the average lifetimes behave similarly to a single lifetime, but the resulting errors are found to be moderate and the results provide a convincing explanation of the, at first glance, counterintuitive behavior. Accordingly, and largely in line with the literature, we observed two types of sites binding ANS at HSA: 3 type A, rather peripheral, and 2 type B, likely more central sites. The latter quench Trp-214 by Förster Resonance Energy Transfer (FRET) with a rate constant of ≈0.4 ns-1 per ANS. Adding millimolar concentrations of GC displaces ANS from the A sites but not from B sites. At incomplete ANS saturation, this causes a GC-induced translocation of ANS from A to the more FRET-active B sites. This leads to the apparent paradox that the partial displacement of ANS from HSA increases its quenching effect on Trp-214. The most important conclusion is that (ANS-like) drugs cannot be displaced from the type-B sites, and consequently, drug transfer to these sites is not impaired by competitive binding of GC in the vicinity of a dissolving micelle. The second conclusion is that for unbound GC above the CMC (9 mM), ANS equilibrates between HSA and GC micelles but with a strong preference for free sites on HSA. That means that even persisting micelles would lose their cargo readily once exposed to HSA. For all MDDS sharing this property, targeted drug delivery approaches involving them as the nanocarrier would be pointless.

Extending the Pseudo-Phase Model of Detergent-Lipid Dispersions by a Detergent-Binding Protein

Mixed micellar drug delivery systems for poorly soluble active pharmaceutical ingredients (APIs) are easy to produce and long-term stable, because they represent equilibrium structures. However, their fate after intravenous injection is still largely unknown. Once injected into the bloodstream, they can potentially convert to vesicles or disappear altogether, with both API and excipients being picked up by blood components. Our study aimed at reducing the gap between the good, quantitative understanding of aqueous glycocholate (GC)-lecithin dispersions alone and the highly complex situation in the blood. To this end, we extended the pseudophase model previously established for lipid-detergent dispersions to include the detergent-binding protein albumin as another component. The model predicted a quaternary phase diagram with planar phase boundaries defined by key parameters of the ternary subsystems, which were then determined by isothermal titration calorimetry. They include the aqueous GC concentration upon bilayer-micelle coexistence, 5.2 mM, the GC-to-lipid mole ratios in coexisting bilayers (R_e^sat = 0.2) and micelles (R_e^sol = 0.7), as well as the capacity of the albumin to bind 0.1 GC molecules with a dissociation constant of K_D = 0.1 mM and 6 GC molecules with K_D = 0.7 mM. Subsequent measurements in the quaternary system showed phase boundaries in good agreement with the model predictions. In addition, the critical micelle concentration of GC shows a minimal value (midpoint of transition) of 9.1 mM at the temperature of 24 °C where the demicellization enthalpy is zero. The demicellization process is accompanied by a heat capacity change of 29 cal/mol K. The model improves the understanding of the mixed micellar drug delivery systems. The success of the approach encourages including even more blood components, like lipoproteins, to a quantitative treatment.