Biomimetic Dispersive Solid-Phase Microextraction: A Novel Concept for High-Throughput Estimation of Human Oral Absorption of Organic Compounds

There is a quest for a novel in vitro analytical methodology that is properly validated for the prediction of human oral absorption and bioaccumulation of organic compounds with no need of animal models. The traditional log P parameter might not serve to predict bioparameters accurately inasmuch as it merely accounts for the hydrophobicity of the compound, but the actual interaction with the components of eukaryotic cells is neglected. This contribution proposes for the first time a novel biomimetic microextraction approach capitalized on immobilized phosphatidylcholine as a plasma membrane surrogate onto organic polymeric sorptive phases for the estimation of human intestinal effective permeability of a number of pharmaceuticals that are also deemed contaminants of emerging concern in environmental settings. A comprehensive exploration of the conformation of the lipid structure onto the surfaces is undertaken so as to discriminate the generation of either lipid monolayers or bilayers or the attachment of lipid nanovesicles. The experimentally obtained biomimetic extraction data is proven to be a superb parameter against other molecular descriptors for the development of reliable prediction models of human jejunum permeability with R2 = 0.76, but the incorporation of log D and the number of aromatic rings in multiple linear regression equations enabled improved correlations up to R2 = 0.88. This work is expected to open new avenues for expeditious in vitro screening methods for oral absorption of organic contaminants of emerging concern in human exposomics.


Table of content
Reagents and solutions (p.S3) Synthesis and characterization of liposomes (p.S5) HPLC analysis (p.S6) Figure S1.Analytical workflow for elucidation of the conformation of the PC following anchorage to the monolith (p.S7) Figure S2.Multicollinearity plot of the parameters listed in Table 1 (p.S8) Table S1.Non-standardized coefficients of all the prediction models evaluated for the prediction of the effective permeability across the human intestine (p.S9) References (p.S10)
A 100 mM phosphate-buffered saline (PBS) solution was prepared by dissolving 1.2 g of potassium dihydrogen phosphate, 7.2 g of disodium phosphate, 40 g of sodium chloride and 1.0 g of potassium chloride in 500 mL of water.Working solution (10 mM PBS, pH 7.4) was prepared by a 10-fold dilution of the stock solution in water.

Synthesis and characterization of LUVs
The synthesis of LUVs was performed using an extruder set from Avanti Polar Lipids, Inc. (Alabaster, Alabama).This set includes the necessary laboratory material to perform the preparation of LUVs including a mini-extruder, syringes (1,000 L), polycarbonate membranes (0.1 m, 19 mm diameter), filter supports (10 mm diameter) and a holder/heating block.
LUVs were prepared by lipid film hydration 1 followed by extrusion for unilamellar liposome formation 2 .To this end, a soybean-PC (Sb-PC) stock solution was prepared by weighing 0.38 g of PC (LIPOID S100, average molecular weight of 787 g mol -1 ) in chloroform in a round bottom flask.The solvent was removed in a rotary evaporator at 30 ºC under reduced pressure (290 mbar) for 2 h and protected from the light, followed by 2 h more under vacuum at room temperature to facilitate the quantitative removal of solvent traces while enabling a uniform dried lipid film on the flask bottom wall.
Afterwards, the lipid film was hydrated with a given volume of 10 mM PBS (pH 7.4) at room temperature to afford a lipid concentration of ca. 100 mM Sb-PC.The aqueous PC solution was vortexed for 1 min every 5 min throughout 1 h to facilitate the complete resuspension of the PC in the aqueous saline medium.The milky suspension of multilamellar vesicles (MLV) was stored at 4 ºC overnight for stabilization.The LUVs were formed by extruding the MLV containing solution through a 100 nm pore size polycarbonate filter 3 , which operational procedure was repeated 29 times.The characterization of LUVs was performed by dynamic light scattering (DLS) using a Zetasizer Nano ZS90 system (Malvern Panalytical, Malvern, UK).The evaluation of the anchored LUVs onto the porous organic polymer surfaces was performed by UV-Vis spectrophotometry (UV-Vis Carry 300 Bio) and scanning electron microscopy (SEM) (HITACHI S-3400N).

S4A 0
.1 M iron (III) thiocyanate solution was prepared by dissolving 1.35 g of iron(III) chloride hexahydrate (FeCl 3 •6H 2 O) and 1.52 g of ammonium thiocyanate (NH 4 SCN) in 50 mL of water.The solution was stored at room temperature.Stock solutions of CECs were prepared at a concentration level of 1000 mg L -1 in MeOH (PCT, CLP, GLP, KTP, DCF, FLV), water (RNT, DMI, CEX, MET, CTM), ACN (CAF, FUR), or 50/50 (v/v) acetic acid/MeOH (MBZ).Intermediate stock solutions containing all the analytes at 100 mg L -1 were prepared in water.The standard solutions were stored at 4 ºC pending use.Working standard solutions for the d-BMSPE studies were prepared daily by appropriate dilution with PBS solution from the concentrated/intermediate stock solutions.

Figure S2 .
Figure S2.Multicollinearity plot of the parameters listed in Table1.

Table S1 .
Non-standardized coefficients of all the prediction models evaluated for the prediction of the effective permeability across the human intestine Model obtained without MBZ Coefficients in bold for models with R 2 >0.7 *