Background
Optimizing rifapentine (RPT) dosing is critical for improving tuberculosis (TB) treatment response. Study 31/A5349 (NCT02410772) compared two 4-month RPT-containing regimens — one with and one without moxifloxacin (MFX) — to the standard 6-month regimen. A key determinant of unfavorable TB outcomes in this trial was RPT exposure, with large interindividual variability. We hypothesized that host genetics contributed to variability in clearance (CL) of RPT and its metabolite, 25-O-desacetyl-RPT (des-RPT) and evaluated an arylacetamide deacetylase (AADAC) variant rs1803155 reported to affect RPT exposure.
Methods
Participants received 1200 mg RPT daily, with plasma samples collected for RPT and des-RPT assays between weeks 2 and 8. We imputed genetic variants from low-coverage whole-genome sequence data using GLIMPSE2. Associations with RPT and des-RPT CL were studied genome-wide, adjusting for 8 genetic principal components and NAT2 acetylator status. We assessed the impact of selected variants on RPT CL by pharmacokinetic (PK) modeling using NONMEM.
Results
A total of 1094 participants had both PK and genetic data (560 in the RPT/MFX arm, 534 in the RPT arm), with a median age of 30 years (IQR 24, 41) and median weight of 53 kg (IQR 48, 59). There were 29% female, 75% Black, and 12% Asian participants. AADAC rs1803155 was associated with RPT CL in all participants (p=1.1E-11), the RPT/MFX arm (p=1.1E-9), and the RPT arm (p=4.3E-4). The GG genotype was associated with more rapid RPT CL and lower exposure. PK modeling showed that RPT CL increased by 0.266 L/h with AADAC GG and 0.059 L/h with AG compared to AA. AADAC rs1803155 was not associated with des-RPT CL. UGT1A1 rs887829 was associated with both RPT and des-RPT CL in all participants (p=6.3E-19 and p=3.7E-26), the RPT/MFX arm (p=4.5E-16 and p=6.2E-14), and the RPT arm (p=6.8E-6 and p=1.5E-14, respectively). Despite genome-wide significant associations of the UGT1A1 poor metabolizer T allele with more CL of both RPT and des-RPT CL, UGT1A1 only impacted des-RPT CL in PK modeling.
Conclusions
In participants receiving 1200 mg RPT daily, the AADAC rs1803155 G allele was associated with increased RPT CL but not des-RPT at genome-wide significance, especially with concomitant MFX, and improved prediction of RPT CL in PK modeling. This helps to explain interindividual variability of RPT PK. Host genetics should be considered when evaluating the efficacy and safety of RPT-containing regimens.