Background:  There has been limited success in pharmacogenetic studies on lopinavir (LPV) pharmacokinetics because only few candidate genes and a limited number of allelic variants have been examined in a few underpowered studies. A change in paradigm emerges from the availability of the HapMap, the wealth of data on less common genetic polymorphisms, gene–drug interaction, and new genotyping technology. To identify potentially novel associations of functional and putative functional variants in genes encoding for ADME (absorption, distribution, metabolism, and excretion) proteins and LPV pharmacokinetics, we performed an ADME-pathway-genes case-control association study.

Methods:  Literature and Web resources were screen intensively to select ADME pathway genes with proven (experimental evidence of interaction), putative (interaction inferred by e.g. metabolites or regulatory networks), or with unknown, but with a potentially relevant, role. For each gene we defined single-nucleotide polymorphisms (SNP) with proven (experimental evidence of functional effect), putative (functional effect inferred by bioinformatics tools), and tagSNP from HapMap. The study population (n = 638) was selected from Caucasians of the Swiss HIV Cohort Study receiving LPV—117 low LPV clearance cases and 90 high clearance controls. Genotyping was performed by a 1536-plex customized GoldenGate assay on an Illumina BeadArray station. Differences in allele frequencies between cases and controls were assessed.

Results:  A total of 1383 SNP (90% of attempted) in 116 ADME genes were successfully genotyped, including:  117 proven, 265 putative functional SNP, and 1001 tagSNP. We detected a statistically significant enrichment trend of low p-values through the groups of unknown, putative, and proven LPV-interacting genes. Enrichment was observed preferentially in CYP3A and SLCO1A2-1B1 locus, known to interact with the study drug. Several SNP were associated with LPV phenotype, including a proven functional polymorphism in SLCO1B1 (rs4149056) that reached study-wide significance (p-value = 2.16E-5).

Conclusions:  Our approach, using a high-throughput customized selection of ADME-pathway genetic variants, allowed the detection of study-wide significant genetic associations with LPV clearance. Interestingly, genes known for a proven role in LPV ADME-pathway, presented a significant enrichment trend of low p-value. The general approach of ADME-pathway pharmacogenetics should be generalized to other drugs.