Tuberculosis (TB) is second only to COVID-19 as the most lethal infectious disease in the world, claiming 1.6 million lives per year. Novel, potent, and safe drug combinations are needed to shorten treatment durations, improve outcomes, and combat resistance against available therapeutics. While the TB drug development pipeline is rich with new and repurposed drug classes with novel mechanisms of action and novel drug combinations, there is no consensus on how best to prioritize multidrug regimens for clinical trials and no modus operandi for nonclinical-clinical translation of TB therapeutics. An increasingly applied solution for the accurate translation and integration of nonclinical findings and emerging clinical data is through model-based approaches. The work presented in this dissertation aims to develop data-driven models integrating a compendium of experimental data to predict clinical efficacy and to contribute toward development of model-based translation methods. In the first half of this dissertation, site-of-action models to predict lesion-centric pharmacokinetics (PK) and pharmacodynamics (PD) in patients were developed for three drugs. Kanamycin and amikacin do not reach therapeutic concentration at the site-of-action, highlighting their limited clinical utility. Clarithromycin accumulates in all tissues compared to plasma; however, its lack of bactericidal activity limits its utility in nontuberculous mycobacteria pulmonary disease (NTM-PD). The approach establishes a platform for future lesion penetration investigations for both TB and NTM-PD. In the second half of this work, semi-mechanistic models based on murine PK-PD data were developed to predict early bactericidal activity in active TB. Our predictive tool, built using nine approved drugs, can predict first-in-patient trials and de-risks entrance into clinical development. The approach was extended to compare efficacy of rifapentine alone and in combination in patients with latent TB infection. Our results suggest that ultra-short course therapy of six weeks with rifapentine is equivalent to approved latent TB regimens. The findings in this dissertation demonstrate model-based translational approaches to predict PK-PD at the site of action and to translate murine drug response in active and latent TB. The model-based tools developed and implemented in this dissertation contribute to regimen development – one of the greatest challenges in TB clinical development – and establish a basis for nonclinical-clinical translation in TB drug development.

Malaria remains a leading cause of morbidity and mortality worldwide, with young children at particular risk for severe, or complicated disease. The artemisinin-based combination therapies (ACTs) are the first-line treatment for uncomplicated (i.e., non-severe) malaria, but optimal dosing regimens are yet unclear for the most vulnerable populations, such as young children and those with potential drug-drug interactions for treatment of co-morbidities, like HIV. Suboptimal dosing not only diminishes clinical responses, but also exposes parasites to subtherapeutic levels of antimalarials, which may select for resistant parasites. For example, resistance to artemether-lumefantrine (AL), a commonly prescribed ACT, is an emerging problem in Africa. Thus, the goal of this dissertation work was to identify ways to optimize dosing in these populations in order to protect the utility of ACTs.

The specific aims of this work were to evaluate the pharmacologic and clinical benefits of extending AL malaria treatment duration from the standard 3-day (6 dose) regimen to a 5-day regimen (10 dose) in young HIV-uninfected children (Aim 1, Chapter 2) and HIV-infected children on efavirenz (EFV)-based antiretroviral therapy (ART) (Aim 2, Chapter 3). As EFV has potent metabolic inductive effects shown to result in lower AL exposure, in Aim 2 an extended regimen was evaluated as an attempt to compensate for this drug-drug interaction. The overall goal of the study was to determine the impact of extended duration treatment on the pharmacokinetics (PK) and pharmacodynamics (PD) of AL. Samples collected as part of a prospective, randomized, open-label phase 4 trial of uncomplicated malaria pediatric patients in Uganda with and without HIV co-infection were used to quantify (1) artemether, (2) active metabolite, dihydroartemisinin (DHA), (3) lumefantrine (LF), and (4) active metabolite, desbutyl-lumefantrine (DBL) plasma concentrations. Children received malaria treatment with standard 3-day (6-dose) or extended 5-day (10-dose) AL-based therapy. A combination of intensive and population PK sampling approaches was utilized to determine drug exposure. PK estimates were then linked to clinical outcomes (e.g., recurrent parasitemia) and assessed over a 42-day follow-up period for the PD analyses.

Overall, our analyses demonstrated that extended (5-day) duration of AL-based antimalarial therapy significantly increased AL and AL-metabolite exposure in HIV-uninfected children. These findings included statistically significant higher exposure of the artemisinins and long acting LF and DBL with extended 5-day therapy. In particular, higher LF concentrations were associated with a 46% and 42% lower risk of recurrent parasitemia at 28 and 42 days, respectively. In addition, weight-based associations were observed for LF among HIV-uninfected children; participants in the lowest weight-band had the lowest AL plasma concentrations. Among HIV-infected children, 5 days of AL successfully compensated for the EFV-driven cytochrome P450 (CYP) 3A4 induction effect and resulted in LF exposure that was equivalent to that seen in HIV-uninfected children given 3 days of AL.

Collectively, this dissertation contributes to the limited PK/PD data for AL-based treatment regimens for uncomplicated malaria in children. Here, we studied an underrepresented, pediatric population to highlight the necessity for optimized dosing regimens (for example, among HIV-coinfected and low weight children). Given that this population makes up the majority of malaria-related morbidity and death, efforts to optimize AL-based treatment regimen dosing in these patients are likely to lead to improved malaria clinical outcomes worldwide.

ABCG2 encodes for a multidrug efflux transporter called the mitoxantrone resistance protein (MXR, BCRP) that mediates the efflux of substrates out of the cell and is important in detoxification. The present study was focused on the expression and function of MXR amino acid variants, activity of the ABCG2 promoter and promoter variants, characterization of ABCG2 locus cis-regulatory elements and variant enhancers, and examination of DNA methylation around ABCG2. MXR expression, localization and activity were characterized using whole cells and inside-out vesicles. The Q141K variant had reduced expression. MXR I206L had increased efflux of pheophorbide A and both V12M and D620N had increased ATPase activity. The activity of ABCG2 regulatory elements was tested in in vitro and in vivo luciferase assays. Two promoter SNPs (rs76656413 and rs59370292) had decreased in vivo liver enhancer activity. Six regions with in vivo liver enhancer activity and several enhancer SNPs (rs9999111, rs12508471, rs72873421, rs149713212 and rs2725263) with altered activity in vivo were identified. Association of these SNPs with ABCG2, PPM1K or PDK2 expression in different tissues was detected. In vitro assays were used to identify nuclear receptor response elements. The ABCG2 promoter responded to multiple nuclear receptor ligands, and the promoter SNP rs66664036 had a significantly increased response to 17β-estradiol. Nine rifampin, six 17β-estradiol and three dexamethasone responsive regions were identified. Enhancer SNP rs12508471 had decreased response to 17β-estradiol and increased response to dexamethasone, while rs573519157 had an increased and rs190754327 had a decreased response to 17β-estradiol. Finally, methylation of CpG islands in the ABCG2 locus was correlated with the expression of ABCG2 in human liver and kidney tissues. There was no correlation of whole CpG island methylation with ABCG2 expression. However, a CpG site within CpG4 correlated with ABCG2 expression in the kidney, and part or all of select CpG islands had significantly lower methylation in liver than in kidney. The genetic and epigenetic regulation of the ABCG2 gene locus described in this dissertation may contribute to clinical variation in ABCG2 expression.

ATP-binding cassette (ABC) transporters are a superfamily of membrane proteins responsible for the import and export of a variety of xenobiotics and endogenous compounds. Multidrug resistance protein 4 (MRP4) is an ABC transporter expressed in a number of human tissues. MRP4 has a number of genetic variants, but the impact of many of these variants is unknown. First, we used Western blotting and fluorescence microscopy to examine the effect of a number of genetic variants on MRP4 expression and localization, identifying a number of variants with reduced expression. Next, we developed a membrane vesicle assay to examine the effect of phosphodiesterase inhibitors on MRP4 transport of the model substrate leukotriene C4.

The xenobiotic ABC transporters are membrane bound proteins that efflux substrates coupled with ATP hydrolysis. These transporters have a broad substrate specificity and tissue distribution in excretory and/or barrier sites, supporting their role in the distribution and elimination of xenobiotics. Many factors regulate the expression and function of xenobiotic ABC transporters, and a wealth of data suggests genetic variation may be a factor. The overall hypothesis investigated in this dissertation is that genetic variation affects clinical phenotypes via modulation of xenobiotic ABC transporter expression and function. Xenobiotic ABC polymorphisms were identified in healthy populations of ethnically diverse individuals. Seven amino acid changing variants and two haplotypes of P-glycoprotein (P-gp), encoded by ABCB1, were tested for their in vitro effects on P-gp expression and function. The Asn21Asp, Arg669Cys, Ala893Ser, Ala893Thr, Ser1141Thr and Val1251Ile variants, and the Asp21/1236T/Ser893/3435T haplotype showed altered intracellular accumulation of calcein-AM and/or bodipy-FL-paclitaxel. In a substrate-dependent manner, certain P-gp variants showed less sensitivity to cyclosporin A inhibition. ABCB1 3'-untranslated region (UTR) variants were investigated for their effects on ABCB1 mRNA stability. Computational methods predicted the impact of 3'-UTR variants on ABCB1 mRNA stability and a cell-based assay measured the mRNA half-life of ABCB1 reference and +89A>T, +146G>A and +193G>A 3'-UTR variants. The mRNA half-life for the 3'-UTR variants was similar to the reference half-life of 9.4 h. Colon cancer patients (n = 33) demonstrated variable ABCB1 and ABCC1-ABCC3 mRNA expression in matched normal and tumor colon tissue. A preliminary association analysis identified possible trends between specific ABC genotypes and mRNA expression. ABCB1, ABCC1 and ABCC3 showed altered mRNA expression in tumor tissue compared to adjacent healthy tissue, suggesting differential regulation in normal and tumor tissue. The functional relevance of promoter region polymorphisms in ABCB1, ABCC2-ABCC6 and ABCG2 was predicted using two computational methods. The predictions correctly identified three out of five ABCC2 promoter variants (-1450A>G, -1193A>G and -920A>G) that were previously shown to have decreased transcriptional activity, illustrating the potential utility of these computational predictions in guiding future studies. In summary, the results of this dissertation research suggest xenobiotic ABC polymorphisms can alter gene expression and transport function.

Since the emergence of the HIV epidemic it has been recognized that complications to HIV infection and variations in drug response and toxicity are influenced by patient genetics. Identification of genetic predictors of HIV infection complications and variation in drug response and toxicity will lead to better treatment options for patients and reduce HIV-related mortality and morbidity. This dissertation contains research that uses candidate gene and genome-wide approaches to identify and characterize novel genetic predictors of nevirapine pharmacokinetics, nucleoside reverse transcriptase inhibitor-induced peripheral neuropathy and HIV-induced peripheral neuropathy. This research demonstrates that nevirapine pharmacokinetic properties are heritable in European and African patients and characterizes the significant effects of CYP2B6 516G>T, CYP2B6 983T>C and ABCC10 rs2125739 on nevirapine Cmin concentrations in a Ugandan HIV+ population. It also highlights the importance of considering all three polymorphisms for prediction of nevirapine Cmin. This dissertation also explores the genetic predictors of NRTI-SN using whole genome and candidate gene approaches in a Ugandan HIV+ population. A polymorphism in VAMP4, rs188298690, was identified in the whole genome study and bioinformatic analyses found that this marker is in an active regulatory region and also a population specific eQTL locus. The candidate gene analysis found that polymorphisms in SLC28A1 and ABCC4 are predictive of the development of NRTI-SN. Finally, this dissertation describes research to identify genetic predictors of HIV-SN. Several polymorphisms in the FOLH1 region were identified in a whole genome study and bioinformatic analyses support a role for these polymorphisms in determining FOLH1 expression. Analysis of the top FOLH1 polymorphism in additional samples showed a trend towards significance and a meta-analysis of the discovery and replication cohorts had improved statistical significance. The research obtained in this dissertation increases the understanding of the role of genetic variation in determining antiviral pharmacokinetics and toxicity and in complications to HIV infection.

Anticancer therapies are commonly prescribed for breast cancer and other solid tumors. While cytotoxic agents are therapeutically efficacious in treatment of many cancers, they often lead to significant toxicities. One of the main non-hematological dose-limiting toxicities for cytotoxic chemotherapies is sensory peripheral neuropathy. Currently, there are no effective therapeutic strategies to prevent or treat chemotherapy-induced peripheral neuropathy (CIPN) due to the lack of understanding in the mechanisms underlying patient susceptibility. This dissertation presents a pharmacogenetic approach to aid in the identification of critical genetic networks that may explain the development of CIPN. With the use of genome-wide association studies conducted from clinical samples of breast cancer patients treated with microtubule-targeting agents, we highlighted the role of sphingosine lipid signaling in microtubule-targeting agent (MTA)-induced peripheral neuropathy (PN). Functional studies using human sensory neurons derived from induced pluripotent stem cells show that inhibiting such pathway attenuates paclitaxel-induced neuronal injury. Additional genes also involved in Rho GTPase signaling were further revealed from a next-generation exome sequencing study on patients treated with MTA chemotherapy. Future efforts are needed to explore how disruption in this signaling pathway develops on the molecular level and subsequently leads to manifestation of CIPN. While these studies have focused on revealing insights of the pathophysiology underlying CIPN, we also investigated the use of early patient-reported symptoms and genome-wide data for the prediction of dose-limiting neuropathy events. We show that early patient-reported outcomes can be used as a tool to predict risk for dose-limiting toxicity events. Future studies aimed at determining whether other types of genomic and clinical information may help improve such predictions. Together, the work presented herein provide novel insights to mechanisms underlying the development of CIPN in efforts to discover novel strategies for CIPN prevention and treatment.

Childhood tuberculosis (TB) affects one million children annually, and in 2019, 194,000 children under 15 years of age died from TB. Malnourished, HIV-positive, and young children have an increased risk of TB disease progression, severe forms of TB, and poor treatment outcomes. Historically, children with TB have been treated as small adults. This may contribute to poorer outcomes due to pharmacokinetic differences between adults and children that result in suboptimal drug exposure. The aims of this work were to apply modeling and simulation to optimize pediatric dosing of important drugs/regimens for drug-susceptible TB, drug-resistant TB, and latent TB infection. Current treatment guidelines for drug-susceptible TB endorsed by the World Health Organization (WHO) recommend dosing children by body weight alone, which may lead to systematic underdosing and worse treatment outcomes in underweight children. A systematic review and meta-analysis was performed to assess rifampicin pharmacokinetics in children. At the current WHO recommendations, rifampicin exposures were much lower in children compared to adults. Younger and HIV-positive children trended toward lower exposures compared to older and HIV-negative children, respectively, but data reporting by these groups were too sparse to draw definitive conclusions. In a separate study, an integrated pharmacologic-epidemiologic model was developed to predict childhood TB outcomes in a real-world population of children under 5 years of age from high burden TB countries. An alternative dosing method, which dosed children by ideal body weight instead of actual body weight, was predicted to reduce unfavorable child outcomes by at least 33%, with major improvements in the youngest children and those who were malnourished. The findings from both analyses support higher doses of rifampicin than currently recommended. Moxifloxacin is a high-priority drug for drug-resistant TB treatment per WHO guidelines. However, doses are not optimized for children due to a lack of pharmacokinetic and safety data, especially in young children. To address this need, the population pharmacokinetics and relationship with QT-interval prolongation were characterized in a cohort of 85 children with TB (median age 4.6 years). Optimal moxifloxacin doses were found to be 10-50% higher than current WHO recommendations, depending on child weight. The risk of QT-interval prolongation was low during the study, but requires further assessment at higher doses, especially with coadministration of other QT-prolonging anti-TB agents such as clofazimine and bedaquiline. Tuberculosis disease can be prevented in those with latent infection with anti-tuberculosis therapy. Novel short-course rifapentine-based therapies are an appealing, non-inferior alternative to the standard 9-months of isoniazid. However, rifapentine pharmacokinetic data in children are lacking. To inform on optimal rifapentine dosing in children, the pharmacokinetics and autoinduction profile were first characterized in adults. The adult model informed the model structure for pediatric simulations as well as pharmacokinetic targets for different regimens. Then, optimized and pragmatic weight band dosing was proposed for the 3-month once weekly regimen (3HP) for labelled use and for the experimental regimens being evaluated in pediatric clinical trials. Lastly, prior knowledge of rifapentine (and rifamycins, in general) pharmacokinetics were leveraged to inform the design of a pediatric PK study evaluating daily rifapentine for TB prevention. Collectively, this dissertation research contributes to the prevention and treatment of TB in children by applying model-based approaches to optimize the dosing of key drugs in current and novel regimens. There is a focus on vulnerable child populations that are typically underrepresented in clinical trials and underserved by standard weight-based dosing practices but represent a large burden of the mortality (e.g., malnourished, HIV-coinfected, young).

ATP binding cassette (ABC) transporters are a family of proteins whose activity is vital to cell detoxification, protection against xenobiotics and oxidative stress, and maintenance of homeostasis of endogenous compounds. Subtle changes in endogenous expression level of these transporters can have significant clinical implications. In particular, the impact of such variation on the role of ABC transporters in transport of pharmaceutical agents is of interest; inter-individual differences in expression of ABC transporters can result in changes in exposure to pharmaceutical agents or their metabolites, leading to altered drug efficacy or drug-induced toxicities. Variation in gene expression can be caused by a number of factors that modulate the "normal" activity during transcription or translation. In this study, mechanisms that regulate mRNA or protein expression of ATP-binding cassette transporters were characterized in human tissues. First, the impact of DNA sequence variation on mRNA expression across individuals was evaluated in the human kidney. Several expression quantitative trait loci (eQTLs) were identified for ABC transporters in the kidney, and one eQTL for ABCG2 (BCRP) was validated in an in vitro reporter-gene assay. Next, the role of alternative splicing in regulation of transporter expression was explored using transcriptome sequencing data from multiple individuals in several human tissues. Examples of splicing events that were found to modulate transporter expression across individuals include alternate 5 prime untranslated regions (5' UTRs) in the genes ABCC5 and ABCA8. Further, a splicing event in the gene ABCC6 was identified that produces a premature termination codon, triggering the nonsense mediated decay process; this event may be responsible for regulating inter-tissue expression of ABCC6. Finally, transcription factor regulators of ABC transporters were identified by searching for transcription factor binding motif enrichment in sets of genes co-expressed with ABC transporters in several human tissues. A number of potential transcriptional regulators of transporters were identified.