Danofloxacin, a veterinary fluoroquinolone antimicrobial drug, is actively secreted into milk by an as yet unknown mechanism. One of the main determinants of active drug secretion into milk is the transporter (BCRP/ABCG2). The main purpose was to determine whether danofloxacin is an in vitro substrate for Bcrp1/BCRP and to assess its involvement in danofloxacin secretion into milk. In addition, the role of potential drug-drug interactions in this process was assessed using ivermectin. Danofloxacin was transported in vitro by Bcrp1/BCRP, and ivermectin efficiently blocked this transport. Experiments with Bcrp1(-/-) mice showed no evidence of the involvement of Bcrp1 in plasma pharmacokinetics of danofloxacin. However, the milk concentration and milk-to-plasma ratio of danofloxacin were almost twofold higher in wild-type compared with Bcrp1(-/-) mice. The in vivo interaction with ivermectin was studied in sheep after co-administration of danofloxacin (1.25 mg/kg, i.m.) and ivermectin (0.2 mg/kg, s.c.). Ivermectin had no significant effect on the plasma levels of danofloxacin but significantly decreased danofloxacin concentrations in milk by almost 40%. Concomitant administration of multiple drugs, often used in veterinary therapy, may not only affect their pharmacological activity but also their secretion into milk, because of potential drug-drug interactions mediated by BCRP.

The ATP binding cassette (ABC) transporters ABCG2 and ABCB1 perform ATP hydrolysis-dependent efflux of structurally highly diverse compounds, collectively called allocrites. Whereas much is known about allocrite-ABCB1 interactions, the chemical nature and strength of ABCG2-allocrite interactions have not yet been assessed. We quantified and characterized interactions of allocrite with ABCG2 and ABCB1 using a set of 39 diverse compounds. We also investigated potential allocrite binding sites based on available transporter structures and structural models. We demonstrate that ABCG2 binds its allocrites from the lipid membrane, despite their hydrophilicity. Hence, binding of allocrite to both transporters is a two-step process, starting with a lipid-water partitioning step, driven mainly by hydrophobic interactions, followed by a transporter binding step in the lipid membrane. We show that binding of allocrite to both transporters increases with the number of hydrogen bond acceptors in allocrites. Scrutinizing the transporter translocation pathways revealed ample hydrogen bond donors for allocrite binding. Importantly, the hydrogen bond donor strength is, on average, higher in ABCG2 than in ABCB1, which explains the higher measured affinity of allocrite for ABCG2. π-π stacking and π-cation interactions play additional roles in binding of allocrite to ABCG2 and ABCB1. With this analysis, we demonstrate that these membrane-mediated weak electrostatic interactions between transporters and allocrites allow for transporter promiscuity toward allocrites. The different sensitivities of the transporters to allocrites' charge and amphiphilicity provide transporter specificity. In addition, we show that the different hydrogen bond donor strengths in the two transporters allow for affinity tuning.

Colocalized in membrane barriers, the ABC transporters ABCB1 and ABCG2 strongly contribute to multidrug resistance (MDR). Here we investigate the as yet unknown mechanisms of activation and inhibition of ABCG2. For this purpose we measured the ATPase activity of ABCG2 and ABCB1 as a function of allocrite concentration using a calibration set of 30 diverse compounds and a validation set of 23 compounds. We demonstrate that ABCG2 is activated at low and inhibited at high allocrite concentrations, yielding bell-shaped activity curves. With an ATP regeneration assay we prove that the inhibitory part is indeed due to a decrease in activity because of high allocrite load in the transporter. However, inhibition is only observed if the membrane solubility of allocrites is sufficiently high. The concentrations of half-maximum activation and inhibition are at least 10-fold lower for ABCG2 than for ABCB1. Because ABCG2 binds its allocrites with higher affinity than ABCB1, it can extract hydrophilic, nonamphiphilic, and highly charged compounds out of the lipid membrane, typically exhibiting low lipid-water partition coefficients, but is inhibited by hydrophobic, amphiphilic, and moderately charged compounds, with high lipid-water partition coefficients. In contrast, ABCB1 is barely interacting with hydrophilic compounds, but is activated by hydrophobic compounds. We show that hydrophobicity, amphiphilicity, and charge have a dual role; they predict, on the one hand, allocrites' lipid-water partition coefficient and, on the other hand, the transporters' preference for the chemical nature of allocrites. Parameters reflecting hydrophobicity, amphiphilicity, and charge are therefore sufficient for differentiating between allocrites, activators, and inhibitors of ABCB1 and ABCG2.