Human liver microsomes, or HLMs, play a key role in drug metabolism because they contain important cytochrome P450 (CYP) enzymes that are critical in oxidation and hydrolysis processes. We hypothesized that the variability in CYP enzyme concentration and activity among individuals will contribute to differences in metabolic capacity. In this study, my goal was to characterize HLMs from a bank of individual donors with different genetic backgrounds. I assessed their drug-metabolizing capabilities, with an emphasis on CYP-mediated pathways. I used bicinchoninic acid assays (BCA) to quantify protein content, carbon monoxide heme spectra (P450 spectra) to determine CYP enzyme concentrations, and a cytochrome c reduction assay to measure P450 reductase activity. For conducting these experiments, we used HLM samples from five different individuals from the UW Human Liver Bank. From the first subset of samples, I found that the ratio of CYP protein to total protein content varied across individuals with different clinical presentations. The differences among individuals in our findings emphasize the importance of HLM characterization in preclinical drug trials to understand the need for the personalization of pharmacological treatments. The implementation of such characterization can have a tremendous effect on predictions of drug responses and optimization of drug dosages, ultimately improving drug efficacy and safety.

Cytochrome P450 enzymes (CYPs) are a family of proteins that play an integral part in drug metabolism. An example is CYP3A4, a specialized protein within the CYP family, involved in the oxidation of small drugs and toxins, like cannabinoids, to allow for their removal from the body. Understanding this reaction is important in evaluating how different expression levels of CYP3A4 in different individuals affect the efficiency of drug metabolism. In this project, I studied drug binding to recombinant, purified, CYP3A4 protein. The binding assays of CYP3A4 allow evaluation of the quality of the purified protein. Ultraviolet-visible spectroscopy (UV-Vis) was used to quantify the protein, and spectral binding through drug titration was used to characterize the binding affinities of different substrates. Drug titration enables us to observe the amount of ligand bound to the protein at various concentrations, displayed as a binding curve from which we can determine the spectral binding constant (Ks). The CYP active site contains a heme group. Upon binding, the ligand replaces the water molecule that is originally docked in the active site. This transfer reaction shifts the heme spin state and appears on the UV-Vis spectra as an increasing absorbance at the 380 nm wavelength and a decreasing absorbance at 420 nm. Inhibitors would have the reverse spectrum. In vitro studies with purified protein have several benefits when investigating protein function, such as simplifying the experimental system and reducing the limitations of complicated sample preparations from living organisms. Having a well-defined assay to determine recombinant 3A4 protein quality will contribute to the value of further in vitro activity and pharmacokinetic studies with this protein. 

Liver fatty acid binding protein (FABP1) is highly expressed in the liver, kidney, and gut and is known for its role in binding endogenous lipids. FABP1 has also been shown to bind drugs and modulate metabolism in the liver. A high frequency single nucleotide polymorphism (SNP T94A) in FABP1 is shown to correlate with nonalcoholic fatty liver disease. We hypothesize that this SNP also affects drug binding. To evaluate drug-FABP1 binding, I measure equilibrium dissociation constants (K_ds) by fluorescent displacement assays for both FABP1 wild-type and T94A using two fluorescent probes, 11-(dansylamino)undecanoic acid (DAUDA) and 8-anilino-1-naphthalenesulfonic acid (ANS). FABP1 has a large binding pocket that can accommodate 2 ligands simultaneously in a ‘high affinity’ and ‘low affinity’ binding site. When DAUDA-FABP1 or ANS-FABP1 are titrated with a drug, a drug-FABP1-probe ternary complex is formed rather than the probe being fully displaced. This complicates data analysis and suggests that endogenous lipids may change the affinity of drugs for FABP1. Therefore, I use multiple fluorescent probes with different binding affinity to obtain drug K_d values. I use singular value decomposition (SVD) to isolate individual fluorescent components from the overall observed fluorescence spectra. I then estimate drug and probe K_ds for FABP1 T94A and T94T by fitting the fluorescent change due to binding to dynamic models in COPASI software. From forward and reverse titrations, DAUDA K_d for FABP1 wild-type was found to be 0.194 µM while ANS binds more weakly (K_d = 1.38 µM). From DAUDA displacement assays, diclofenac was found to have a K_d of 3.90 µM for wild-type FABP1 and 3.78 µM for T94A. I anticipate measurement of K_ds for 8 other drugs using both DAUDA and ANS in the coming months. The developed methods will enable evaluation of FABP1’s role in drug disposition.

The metabolite of vitamin A, retinoic acid (RA), plays a critical role in regulating cell differentiation in mammals. RA and its metabolites exist as different geometric isoforms (all-trans, 13-cis, 9-cis, 4-oxo-all-trans, 4-oxo-13-cis .). All-trans-RA is the biologically active isomer, and 13-cis-RA is found as the drug “Accutane” used to treat severe acne. Cellular retinoic acid binding proteins are evolutionarily conserved intracellular proteins that regulate RA tissue concentrations. The all-trans isomer is known to bind tightly to CRABP1 and 2, but little is known about the binding of the other four isomers and metabolites. No data on the binding affinities of the 4-oxo isomers is available. I hypothesize that the RA and 4-oxo-RA isomers that have not been extensively researched, have different binding affinities between the two CRABPs. To test this hypothesis, I use fluorescence spectroscopy coupled with single value decomposition (SVD) analysis and stopped-flow analysis to measure the equilibrium dissociation constant (Kd), association rate constant (kon), and dissociation rate constant (koff) for retinoid binding to CRABPs. My current data generated by the fluorescence spectroscopy method shows that binding affinities of the tested retinoids are comparable between CRABP 1 and 2, except for 13-cis-RA which bound CRABP2 significantly more tightly than CRABP1. (CRABP1 Kd  = 609 nM, CRABP2 Kd  = 70.5 nM) . All-trans-RA (atRA) has the tightest binding to both CRABP 1 and 2, (CRABP1 Kd  = 0.51 nM, CRABP2 Kd  = 0.73 nM), followed by 4-oxo-atRA, (0.39nM, 1.4 nM), 9-cis-RA, (61.5 nM, 96.2 nM), and finally 4-oxo-13-cis (779.5 nM, 743.6 nM) with the lowest binding affinity. These relationships will be further investigated using the stopped-flow method.

With the rising popularity of TikTok and its role in widespread information sharing, and concerns about political misinformation on the platform, studying political discourse through immersive ethnographic methods leading up to the 2024 US Presidential Election was imperative. This study, conducted over 12 weeks in the summer of 2024, aimed to uncover the rhetorical and ideological topics and trends salient to Black Republican and Democratic TikTok creators through content analysis. To capture the distinct feeds that a person interested in right-leaning or left-leaning content may see, two partisan personas were created on separate phones. The personas were developed from a seed list of known partisan creators and snowball sampling. A quantitative content analysis was conducted using LabelStudio software on a sample of 120 acquired videos from Black creators across the two research phones. The videos were coded for style, topic, and person of interest, alongside other inductive attributes that emerged during the coding process. Thematic analysis revealed key discursive themes around harm and blame, along with different tactics of evidence used by creators to further their points. We find a divide between left- and right-leaning creators with regard to the institutions and politicians they hold accountable for harms, and the ideologies they perceive being pushed by the oppositional party. This study shows how the affordances of the TikTok platform allowed for, and algorithmically rewarded, infighting within the Black community leading up to the election. Future studies may apply these methods of persona-enabled ethnographic data collection for conducting bipartisan investigation on other online communities, including but not limited to racial minority groups, in gaining a better understanding of prevalent issues within these communities, political or otherwise.