Protein abundances of drug transporters and metabolizing enzymes were quantified in ten identical human liver samples in six laboratories using their respective in-house methods. A series of phase I-metabolizing enzymes (CYP1A2, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, CYP3A4, and CYP3A5), phase II-metabolizing enzymes (UGT1A1, UGT1A3, UGT2B7, and UGT2B15), uptake transporters (SLCO1A2, SLCO1B1, SLCO1B3, SCLO2B1, and SLC22A1), and efflux transporters (ABCB2, ABCC2, ABCC3, ABCG2, and SLC47A1) were included in the analysis. The quantifications were performed using five targeted proteomics methodologies. All five used stable isotope labeled peptide standards. The sixth laboratory used a global, label free approach. Three laboratories studied proteins in the whole cell lysate while the other three analyzed proteins in enriched membrane fractions. All samples were analyzed as technical duplicates at two independent occasions. Protein abundance levels obtained with the different techniques were compared. Enrichment factors were used to re-calculate abundance levels of proteins from membrane fractions to protein levels in whole cell. The inter-individual spread from the liver samples was analyzed as well.
In general the mean values for the drug metabolizing enzymes were in good agreement between the different laboratories. The protein levels of the majority of the CYP-enzymes displayed a three-fold difference. The exception was the lower expressed CYP2B6 that showed more than ten-fold difference between the laboratories. The expression levels of UGT2B7 and UGT2B15 were in the same range as the majority of the CYP-enzymes and also differed three-fold. A higher variability in reported protein levels was displayed for the lower abundant UGT1A1 and UGT1A3. The mean values of both SLC- and ABC-transporting proteins were in the same range as the low abundant enzymes. In agreement to what was seen for the low abundant drug metabolizing enzymes, the variability of the drug transporting proteins was more than ten-fold between the laboratories. The protein levels obtained from the enriched membrane fractions (n=2) differed more compared to the protein levels analyzed in whole cell lysates (n=3). The technical repeats correlated better for the targeted approaches compared to the global approach.
Despite large differences in methodologies, the highly abundant proteins were in good agreement. A higher variability was seen for the lower abundant proteins. Protein levels varied significantly between the different membrane enrichment methods. A higher variability between the technical duplicates was seen in the results from the label free, global approach. Further analysis investigating the impact of enrichment procedures, peptide quality and sensitivity of quantification will provide additional information on the causes of the variability.