A Systems Approach to Vitamin A Homeostasis: Integrating Proteomics, Physiology, and Hormonal Modulation of Retinol, RBP4, and TTR

Abstract

Vitamin A (retinol) is an essential lipophilic micronutrient that plays critical roles in vision, immune function, development, and metabolic homeostasis. In circulation, retinol is bound to retinol binding protein 4 (RBP4), which in turn forms a complex with transthyretin (TTR). Formation of this higher molecular weight complex prevents glomerular filtration of RBP4. In the absence of nutritional deficiency or disease, circulating concentrations of retinol, RBP4, and TTR are thought to be relatively stable within an individual. The underlying sources of inter-individual variability in circulating retinol, RBP4, and TTR concentrations are not completely understood. Moreover, retinol, RBP4, and TTR are rarely quantified together in the same individual, limiting insight into how their concentrations or binding states may differ by individual or disease status, particularly in obesity and diabetes. This dissertation integrates proteomic quantification, clinical data, and kinetic modeling to investigate the sources of inter-individual variability in circulating retinol, RBP4, and TTR concentrations. To enable accurate quantitation, a targeted LC-MS/MS assay was developed and validated for absolute measurement of RBP4 and TTR in serum and plasma. The method allows for concurrent absolute quantification of RBP4 and TTR concentrations by using well characterized protein standards, robust surrogate peptides, stable isotope-labeled peptides for internal standards, and a process control to monitor digestion efficiency. Method accuracy across different biological matrices was confirmed with parallelism and matrix effect experiments. The assay was applied to samples from both healthy individuals and patients with diabetic kidney disease. Measured intra-individual variability in RBP4 and TTR concentrations was comparable to assay variance (6 – 8.5%), highlighting the low intra-individual variability of RBP4 and TTR concentrations in circulation. Compared to ELISA-based RBP4 and TTR measurements in samples from individuals with diabetic kidney disease, our LC-MS/MS assay showed overall agreement. Measurement of absolute concentrations of both RBP4 and TTR, combined with measurement of retinol concentrations, in serum enabled development and application of a kinetic binding model to predict complex formation between RBP4, TTR, and retinol using experimentally defined dissociation constants. The validated method was then applied to a carefully selected cohort of metabolically healthy individuals with a range of BMI (21–56 kg/m2), none of whom were diagnosed with diabetes, hepatic, or renal impairment. Prior studies have reported conflicting associations between serum RBP4 and BMI: some identified positive associations, while others reported no correlation. These discrepancies have been attributed to differences in study populations, including sex distribution, disease status, and critically, the analytical methods used to quantify RBP4. In the cohort studied here, serum RBP4 and TTR concentrations did not correlate with BMI. Instead, sex and age emerged as significant predictors of both serum TTR (p=0.02) and RBP4 (p=0.009) concentrations. A modest negative association between BMI and total serum retinol was observed (p=0.02). Using the developed kinetic binding model parameterized with experimentally derived dissociation constants, the majority of total RBP4 was predicted to be complexed with TTR, primarily as the ternary complex (retinol:RBP4:TTR). The model predicts that approximately 24% of total circulating RBP4 is unbound to retinol but remains complexed with TTR, highlighting RBP4:TTR as a distinct and previously underappreciated binding species. These findings reinforce the role of TTR in increasing RBP4 half-life by preventing glomerular filtration of RBP4. Results from the kinetic model further suggested that the sex differences in total RBP4 concentrations may be driven by differences in TTR concentrations between men and women. To explore hormonal regulation of sex differences, we examined the effects of estradiol and cortisol on retinoid and binding protein concentrations in two clinical studies. In postmenopausal women administered oral estradiol, linear mixed-effects modeling identified an inverse association between serum estradiol and RBP4 concentrations. In premenopausal women, oral hydrocortisone administration increased serum 13-cis-retinoic acid (13cisRA) concentrations, which were associated with cortisol exposure. These findings provide new clinical evidence that estradiol and cortisol may modulate circulating retinoids and their binding proteins, though further investigation is warranted. Collectively, this dissertation demonstrates the value of integrating quantitative proteomics with kinetic modeling to further our understanding of vitamin A homeostasis. These results highlight the importance of sex, age, and hormone exposure in modulating circulating retinoids and binding proteins and underscore how comorbidities such as hepatic or renal impairment can complicate interpretation of retinoid, RBP4, and TTR measurements. These findings also have implications for the utility of RBP4 as a biomarker, suggesting that concurrent measurement of retinol and TTR and inclusion of sex as a covariate should be considered in clinical studies investigating diagnostic and prognostic potential of RBP4 serum concentrations in disease states. Further studies on the relationships between retinol, RBP4, and TTR concentrations in populations with metabolic, hepatic, or renal disease are warranted. Together, this work contributes to a more clinically relevant understanding of vitamin A homeostasis.