Edinburgh Napier University

  Nowadays immobilized artificial membranes (IAM) are commonly used for modelling drug
transport through biological lipid layers. In most published studies a compound’s lipophilicity
measurements are performed in isocratic conditions where a low percentage of organic
modifier is added into the mobile phase. However, in the current study we utilized a previously
validated fast-gradient high-performance liquid chromatography (HPLC) method to obtain a
chromatographic hydrophobicity index (CHIIAM) data. IAM phospholipophilicity values were
measured for 20 structurally unrelated drug-like molecules and compared with apparent
pulmonary permeability values obtained from Calu-3 cell lines (log Papp) taken from the
literature. The relationships between logarithmic CHIIAM (CHI log D), log Papp data and
calculated n-octanol/water partition and distribution coefficients (clog PN, clog D7.4) were
investigated.
CHI log D [1] was moderately related to clog PN values and poorly related to clog D7.4
,
suggesting that phospholipid partitioning supports both hydrophobic/lipophilic forces as well as
electrostatic interactions occurring between ionized drugs and charged phospholipid moieties.
The computed Δ/Δ' log kw
IAM.DD2 [2] component accounting for these excess electrostatic
interactions was compared with in vitro permeability values, and an inverse linear relationship
was found for the given dataset. Thus, it can be concluded that electrostatic interactions can
act as ‘trapping forces’, hindering membrane permeation.
A significant correlation (r2=0.85) was observed between log Papp and CHI log D values for
compounds with molecular mass> 300 Da for which paracellular diffusion is a negligible
transport route. Phospholipophilicity predictions for small hydrophilic compounds (MW< 300
Da) and drug transported by membrane transporters were not accurate.
The strength of IAM HPLC predictive power was evaluated by statistical best optimized model
generated in VEGA ZZ based on CHIIAM values and another 28 calculated static and
conformational molecular descriptors. A high degree of accuracy (r2=0.8, q2=0.55) was found
between predicted and experimental permeability models.
Literature:
[1] Valkó, K., Bevan, C., & Reynolds, D. (1997). Chromatographic Hydrophobicity Index by Fast-Gradient
RP-HPLC: A High-Throughput Alternative to log P/log D. Analytical Chemistry, 69(11), 2022–2029. doi:
10.1021/ac961242d
[2] Grumetto, L., & Russo, G. (2021). cΔlog kwIAM: can we afford estimation of small molecules’ bloodbrain barrier passage based upon in silico phospholipophilicity? ADMET and DMPK, 9(4), 267–281. doi:
10.5599/admet.1034