- Title
- African population prevalent genetic variations of dihydropyrimidine dehydrogenase as the 5-flourouracil cancer drug metabolizing enzyme: computational approaches towards pharmacogenomics studies
- Creator
- Tendwa, Maureen Bilinga
- ThesisAdvisor
- Tatan Bishop, O.
- ThesisAdvisor
- Musyoka, T.
- Subject
- Uncatalogued
- Date
- 2023-10-13
- Type
- Academic theses
- Type
- Doctoral theses
- Type
- text
- Identifier
- http://hdl.handle.net/10962/432263
- Identifier
- vital:72856
- Identifier
- DOI 10.21504/10962/432270
- Description
- In an era of newly emerging cases of non-communicable diseases such as cancer, research is vital for both the medical and economic well-being of humanity. Pharmacogenomics has laidthegroundworkfor the identification of potential genes in cancer progression and treatment outcome investigations. Researchers are increasingly discovering heterogeneity in the efficacy and toxicity responses of drugmetabolizing enzymes (DMEs) in diverse patient populations receiving anti-cancer therapy. DMEs comprise of Phase I (Cytochrome P450s) and Phase II (glutathione-S-transferases (GSTs), UDP-glucuronosyltransferases (UGTs), and dihydropyrimidine dehydrogenases (DPD)enzymes. The main cause of disparity in DME treatment outcomes is genetic variation,which causes missense mutations leading to structural and kinetic properties of the enzyme. These modifications have a deleterious impact on the pharmacodynamics and pharmacokinetics of drugs through multiple mechanisms. Presently, most cancer medicines are manufacturedin developed countries based on the genetic background of non-African subpopulations. Thus, these drugs may not be optimally effective or can cause adverse side effects. Even though heterogeneity in toxicity and efficacy of these drugs has been observed in African descent, the basis of this population variance remains partially understood. For instance,a deficiencyof DPD, the first-rate limiting metabolizing enzyme in the pyrimidinepathway, causes severe toxicity when exposed to 5-fluorouracil (5-FU) chemotherapy. However, minimum studies have been conducted to unravel itsmolecular mechanismwhich may unravel the observed drug treatment outcomes.The aim of this pharmacogenomics study was to determine the underlying mechanism by which DPD missense mutations, which are associated with an African ancestry subpopulation, provoke dysfunctional 5-FU metabolism, resulting in drug toxicity. This knowledge will be critical in designing drug modulators to aid in the restoration of DPD function, a hallmark of precision medicine. Therefore, in the first part of the research we identified and reviewed the general role of Phase I and Phase II cancer drug metabolizing enzymes. We then used World Health Organization (WHO) essential medicine and drug.com to authenticate the usage of 5-FU as an anti-cancer treatment agent. The 3D structure and chemical structure of the agent was then downloaded from the Drug bank. Subsequently, Human Mutation Analysis - Variant Analysis PORtal (HUMA) and Mendelian Inheritance in Man (OMIM) were used to obtain data on DPD non-synonymous genetic variants. Additionally, the aggregate information of DPD missense mutations and their relation to human health were extracted from ClinVar and Pharmacogenomics Knowledge Base (PharmKGB). This information, along with additional data from single nucleotide polymorphisms (dbSNP), 1000 Genomes Project and Exome Sequencing Project (ESP MAF) considering variants classified based on their minor allele frequency (MAF) of 0.001, as well as research articles, consolidated information on missense mutations associated with African subpopulations. Finally, the wild type (WT) and detected mutation sequences were obtained from the Universal Protein Resources database (UniProt). However, because the 3D structure of human DPD was missing, the dimeric wild type (WT) human 3-dimensional (3D) structure was modeled via MODELLER using the pig’s structure as a template. PRIMO, HHpred, and the Protein Data Bank (PDB) were all used to locate the suitable template. As a result, six clinical (C29R, M166V, Y186C, S534N, I543V, and D949V) and thirteen non-clinical (S201R, K259E, D342N, D432N, S492L, R592Q, A664S, G674D, A721T, V732G, T768K, R886C, and L993R) mutations were discovered. Using AMBER tools, we then determined accurate force field parameters for each monomer of DPD protein's Fe2+ centers. Following the creation of each mutation model structure in Discovery Studio, the resulting AMBER force field parameters were inferred. For each model structure, a drug free (inactive/open-conformation) and drug bound (active/closed-conformation) model structure was created (WT and mutations). The model structures were validated using the consensus of three validation programs, namely ERRAT, PROCHECK, and ProSA. Similarly, the impact on structural functionalities was predicted by consensus from Variant Analysis Porta (VAPOR) web server, which include three support vector machines (SVM)-based tools; PhD-SNP, MUpro, and I-Mutation. After protonation in the H++ web server, the six clinical and thirteen non-clinical (six active site and seven non-active site) mutations identified were then exposed to 600 ns molecular dynamic (MD) simulation. The non-clinical data was divided into two categories to better understand the impact of the mutation based on its position in the protein: six catalytic-domain (R592Q, A664S, G674D, A721T, V732G, and T768K) and seven remote (S201R, K259E, D342N, D432N, S492L, R886C, and L993R) missense mutations. The post-MD analysis was done using the typical existing computational global investigations [RMSD, all versus all RMSD, RMSF, RG, hydrogen bonds (H-bonds) and dynamic cross correlation (DCC)]. In addition, we used in silico tools newly developed within the Research Unit in Bioinformatics (RUBI) group, such as comparative essential dynamics (ED)-principal component analysis and dynamic residue network (DRN) multi-metric [betweenness centrality (BC), closeness centrality (CC), degree of centrality (DC), eigen-centrality (EC) and Katz centrality (KC)] analysis algorithms. From the analysis, it was observed that the loop regions of the mutation proteins had increased loop flexibility, particularly around the catalytic loop, which could account for the enhanced asymmetric behavior of the mutation’s monomers compared to the WT. Notably, the A664S mutant showed relatively lower fluctuations, deviating from the observed heightened flexibility in other mutants. A general decrease in hydrogen bonds was observed in the 5-FU binding environment of the mutations compared to the WT. In particular, 5-FU contact analysis of the WT versus the mutation revealed a reduction in contact between core 5-FU binding residues and catalytic residues Cys671 and Ser670, which form hydrogen bonds that initiate DPD catalytic action. Additionally, BC was used to quantify the importance of a protein residue based on how often it acted as a bridge along the shortest paths between other residues. It reflected the potential control or influence a residue may have over communication between different parts of a protein structure. DC assesses the number of connections or interactions a residue had with other residues in the protein, indicating its overall connectivity within the structure. In both drug free and drug bound state, DPD data from the active site hubs' BC and DC revealed a dimeric asymmetric communication pathway per monomer involving a cluster of newly introduced hubs ensemble along the oxidoreduction conduit from NADPH to 5-FU. The two BC communication pathways were located more on the interior of the oxidoreduction conduit, while the two DC communication pathways were located on the exterior. In both cases, one pathway dominated the other. Partially lost function reported in mutation systems could be credited to the compensation communication response to the catalytic site via the least compromised routes. Similar patterns were observed in allosteric communication pathways to the active site induced by remote mutations. Mutations may have destabilized the active-loop and 5-FU binding environment, resulting in a compensatory mechanism seen by the addition of new hubs to the communication network. Surprisingly, EC hubs in the WT were found within the catalytic site domain, indicating that the region is important in 5-FU metabolism. EC measured the importance of a residue by considering both its own degree of connectivity and the degrees of connectivity with its neighboring residues, highlighting its significance in information flow and communication. Herein, EC hubs in mutant systems were found to lose this importance, with active site domain mutations suffering the most. This could explain why non-clinical catalytic domain mutations R592Q, A664S, and G674D, as well as clinical catalytic domain mutations S534N and I543V, experienced drug exit in one of their monomers during simulation. In contrast, there was no 5-FU exit in the non-clinical remote domain. Additionally, aside from the active site, KC hubs were also found around the cofactors, indicating that these components were equally important in DPD overall function. KC combines the concepts of both degree centrality and eigen-centrality, it incorporated both direct and indirect interactions to evaluate the importance of a residue, assigning higher centrality to residues that have connections to other highly central residues. Hence, providing a more comprehensive measure of influence within the protein network. More importantly, CC is known to measure how efficiently a residue can interact with other residues in the protein, considering the shortest path lengths. It indicates the proximity of a residue to others, suggesting its potential for information transfer or functional integration. CC revealed that the majority of persistent hubs were found within the protein-cores known as cold-spots. Overall, this study highlighted the communication pathways triggered by active site domain mutations, as well as the allosteric communication pathways triggered by each remote mutation in both drug free and drug bound states of the DPD enzyme. Both clinical and non-clinical mutations revealed each protein's adaptive compensation mechanism, which results in partial function loss. In each case, the communication network of the different monomers changed from inactive to activated DPD protein. Cold-spot areas were discovered to contain key persistent residues involved in protein function and stability. These areas have been proposed as potential targets for new or repurposed pharmacological modulators that can restore enzyme function. In the pursuit of precision medicine, it also lays the groundwork for detecting and explaining the molecular mechanisms of other drug metabolizing enzymes related to the African-descent subpopulation.
- Description
- Thesis (PhD) -- Faculty of Science, Biochemistry and Microbiology, 2023
- Format
- computer, online resource, application/pdf, 1 online resource (471 pages), pdf
- Publisher
- Rhodes University, Faculty of Science, Biochemistry and Microbiology
- Language
- English
- Rights
- Tendwa, Maureen Bilinga
- Rights
- Use of this resource is governed by the terms and conditions of the Creative Commons "Attribution-NonCommercial-ShareAlike" License (http://creativecommons.org/licenses/by-nc-sa/2.0/)
- Hits: 218
- Visitors: 220
- Downloads: 11
Thumbnail | File | Description | Size | Format | |||
---|---|---|---|---|---|---|---|
View Details | SOURCE1 | TENDWA-PHD-TR23-252.pdf | 20 MB | Adobe Acrobat PDF | View Details |