Search

Antiretroviral pharmacokinetics is one of the most important predictors of therapy efficacy and correlations between virological suppression and plasma concentrations have been described for all antiretroviral classes. Plasma concentrations are the result of absorption, distribution, metabolism and elimination (AD ME) processes which are mediated by numerous proteins in different tissues. Single nucleotide polymorphisms (SNPs) in genes encoding for individual proteins may be an important determinants drug exposure. The aim of this thesis was to characterise the role of some key SNPs determining antiretroviral plasma concentrations in order to improve the current understanding of dose-dependent efficacy and toxicity. A number of different strategies to investigate the role of genetic variability in antiretroviral exposure were developed in this thesis. The polymorphism 63396 C>T in the PXR gene, a nuclear factor regulating the expression of CYPs and transporter in hepatocytes, was proven to influence unboosted atazanavir clearance using a pharmacogenetic based population phannacokinetic model and patients with 63396TT genotype had a increment in ATV clearance of 17% (Chapter 2). In Chapter 3, the predictors of intracellular phannacokinetic of unboosted and boosted atazanavir were investigated and expression of SLC03Al, an uptake transporter expressed on PBMC was correlated with cellular accumulation of boosted and unboosted atazanavir (rho=0.706, p=0.010; rho = 0.830, P = 0.005, respectively). An in vitro in vivo extrapolation model for efavirenz was developed in Chapter 4 and the effect of CYP2B6 516 G>T on standard regimens or dose reduction was quantified. Efavirenz pharmacokinetics was predicted with good accuracy compared to available data; simulated Ctrough was 2119 ng/ml vs 1764 ng/ml, simulated Cmax was 3725 ng/ml vs 4063 ng/ml. The effect of 516G>T on simulated EFV clearance (GT = - 24% and TT = -58%) was comparable to previously published population phannacokinetic data (GT =-36%, TT =-66%). Maraviroc, a CCR5 inhibitor was identified as a substrate for SLC01B1 with a Km of 33.85 IlM and Vmax of 187.9 finol/oocyte/min and the inhibitory potential of concomitant protease inhibitors on this transporter was investigated using Xenopus laevis oocytes heterologous protein expression system (Chapter 5). Moreover the SNP SLCOIBl 521 T>C was correlated with maraviroc Ctrough in a cohort of HIV infected patients; MVC Ctrough in SLC01B1 521 heterozygote patients (TC) (n=8) was higher than in wild type homozygotes (TT) (n=22), 103 ng/ml (69 - 124) vs 46 ng/ml (28 - 66), p=0.003 (Chapter 6). The convergence of different factors under study permitted partial definition of the complex interplay in which drug-drug interactions and physical characteristics interact with genetic variability to define the phannacokinetic phenotype. These findings will help identify patients with a higher risk of achieving sub-therapeutic or potentially toxic plasma concentrations and will serve as a knowledge-base from which to grow and validate strategies for optimisation of therapy.

The WHO currently recommends tenofovir (TFV) disoproxil fumarate (TDF) or TFV alafenamide (TAF) as a preferred first-line antiretroviral therapy (ART) for the treatment and prevention of HIV. However, mild to moderate TFV-related renal toxicity is an uncommon but significant complication for TDF therapy. Efavirenz (EFV) is also approved for first line treatment, with the approval of a lower dose treatment, 600mg to 400mg, offering an opportunity for a greater number of patients to receive ART for the same amount of current public health expenditure, whilst enabling the maintenance of viral load suppression and partial mitigation of adverse events. However, EFV displays broad inter-patient and inter-population variability in its pharmacokinetics and displays a range of dose dependent adverse effects. In order to gain further understanding of the impact patient genetics has on response to ART, studies investigating the effect of single nucleotide polymorphisms (SNPs) within multi ethnic populations receiving either TFV or EFV were completed. Additional work was conducted to investigate potential drug-substrate interactions at the transporter level within proximal tubule cells (PTCs), and the ramifications this may have for diagnosis of renal impairment. This work was conducted through assessing the relationship between SNPs found within genes encoding transporters on the apical membrane of PTCs and a diagnosis of renal dysfunction, defined as kidney tubular dysfunction or chronic kidney disease within a Ghanaian population (Chapter 2). Additionally the association between pharmacogenetic variants linked to TDF metabolism or TFV excretion and TFV plasma and urine concentrations, in a majority Caucasian patient population, receiving TFV as part of their ART was investigated (Chapter 3). This thesis sought to assess whether SNPs in genes involved in EFV and nevirapine metabolism were linked to previously observed changes in the contraceptive subdermal implant levonorgestrels (LNG) pharmacokinetics, when prescribed alongside EFV, 19 within a Ugandan population (Chapter 6). In order to study the effect of transporter interactions at the site of TFV toxicity, the proximal tubule, a transiently transfected Human Embryonic Kidney 293 cell line overexpressing either multidrug and toxin extrusion protein (MATE) 1 or MATE2K was developed and assessed for correct functionality (Chapter 4). This cell line was then utilised for drug-substrate interaction studies between TAF and TFV and either the endogenous biomarker creatinine or the type 2 diabetes drug metformin (Chapter 5). These studies provided a greater understanding of the contribution pharmacogenetics provides to the interaction observed between EFV and LNG, and outlined a genetic association between TFV transporter SNPs and TFV plasma and urine concentrations. Utilisation of these findings in future pharmacogenetics studies would aid in the understanding of the impact of genetic variants in different populations, and the consequences this has for achieving sustained virological response to ART. The study of the effect of TAF and TFV on MATE1 and MATE2K transport of metformin and creatinine produced novel data on these interactions. Overall, the studies included within this thesis have clinical impact through further elucidating the potential mechanisms of toxicity or treatment failure within patients receiving ART.

The first-order degradation rate constant (kdeg) of drug metabolising enzymes (DMEs) is a known source of uncertainty in the prediction of time-dependent drug-drug interactions (DDIs) in physiologically-based pharmacokinetic (PBPK) modelling. There is a large disparity or paucity of published kdeg and related half-life (t1/2) values for DMEs. Physiologically-relevant kdeg values should ideally be derived in vivo to facilitate accurate DDI predictions. However, direct measurement of DME degradation in humans is not routinely possible and indirect measurements utilising changes in levels of specific endogenous substrates have only been described for a few DMEs. This thesis aims to develop an in vitro method of measuring DME protein degradation rates to improve the prediction accuracy of time-dependent DDIs. One in vitro approach of measuring protein degradation rates involves inhibiting de novo protein synthesis, followed by tracking residual protein or activity decline over time. Pharmacological protein synthesis inhibitor agents are commonly used for this purpose but may cause cytotoxicity. Four commonly used inhibitor agents were assessed for their capacity to inhibit protein synthesis without overt cytotoxicity. However, all selected compounds were too cytotoxic for subsequent use in kdeg studies. Small-interfering ribose nucleic acid (siRNA) can be added to in vitro systems to initiate gene-specific silencing by inhibiting messenger RNA (mRNA) translation. It was hypothesised that siRNA would inhibit de novo protein synthesis with less cytotoxicity owing to its specificity. CYP3A4 is the most widely studied cytochrome P450 (CYP) enzyme in terms of DDIs because of its well-recognised role in xenobiotic metabolism. Primary human hepatocytes were treated with CYP3A4-specific siRNA to suppress mRNA translation, followed by the tracking of enzyme activity and protein loss over time to derive kdeg. CYP3A4 kdeg was calculated at 0.019 (± 0.044) and 0.020 (± 0.0003) h-1 from protein and activity loss, respectively. These values were in good agreement with existing published values. The siRNA approach was subsequently applied to determine CYP2B6 kdeg. The CYP2B6 kdeg values derived from siRNA-treated hepatocytes were 0.081 (± 0.009) h-1 from protein loss and 0.058 (± 0.010) and 0.062 (± 0.006) h-1 from activity loss, which were assessed by different methods. The CYP2B6 kdeg values derived from untreated hepatocytes were similar to values in literature. This novel approach can now be used for other less well-characterised DMEs that are associated with time-dependent DDIs. Cellular protein abundance is a balance between synthesis and degradation. Dysregulation of the lysosomal or proteasomal protein degradation mechanisms affects steady-state protein levels and impacts on overall cellular functions. It was hypothesised that single nucleotide polymorphisms (SNPs) in the CYP3A4 protein degradation machinery could affect CYP3A4 protein abundance and downstream activity. Five SNPs were investigated for associations with plasma atazanair (ATV) concentrations, which was a surrogate measure for CYP3A4 activity. No associations were found and this was likely due to the lack of clear understanding of the specific mechanisms that commits CYP proteins for degradation. Further work in this field will identify targets that may be exploited in the future for more accurate measurements of DME kdeg. The data presented in this thesis enhances the understanding of methods used to study protein degradation and this can be applied to multiple fields of cellular research. Importantly, work herein has generated a novel approach to measuring kdeg of proteins that can be applied to other less-well characterised enzymes for better prediction of time-dependent DDIs.