THU610 Distribution Of Cortisol In Human Serum In Vitro: Role Of Competitive Ligand-protein Interactions In Women On Oral Contraceptives

Abstract Disclosure: R.I. Dorin: None. I. Perogamvros: None. C.R. Qualls: None. Background: The relationship between serum concentrations of total (XTotF) and free (XF) cortisol (test tube equilibrium at 37°C) is influenced by many factors. These include concentrations and affinities of serum binding proteins (BP), such as CBG and albumin (A). Among women taking oral contraceptives (OC), XF was found to be higher than predicted using ligand-BP association parameters developed in healthy volunteers (HV). We hypothesized that ligand(s) (XP) competing for XF binding to CBG may contribute to altered relationships between XTotF and XF observed in OC relative to HV. Methods: Feldman’s system of iterative, non-linear equilibrium equations1 were developed in (nxm) matrix format. Data included measured XTotF and XF at each time point (0-480 min) after 20 mg hydrocortisone2. Groups were stratified by (i) mode of administration (iv vs. po) and (ii) condition (HV, n=13 vs. OC, n=12). The matrix was interrupted to obtain iterative equilibrium solutions for parameters of interest using (1x2) minimal model (MM) and (2x2) ligand competition model (LCM). Solutions were optimized by minimizing least squares differences between measured and model predicted XF. Albumin-bound cortisol was parameterized as a constant ratio of XF (N) to simplify the matrix problem. Comparisons and interactions within the 2x2x2 design (delivery, group, and model) were analyzed by ANOVA. Results: In OC, LCM provided significantly better fit to measured XF relative to MM (13 vs. 19% error, P=0.003). MM solutions for CBG-cortisol affinities were significantly higher in HV (KD=26.6) vs. OC (KD=71.9 nmol/L), but LCM yielded similar affinities in both groups (KD = 23.9 in HV and 16.1 nmol/L in OC, interaction P<0.001). LCM solutions for concentrations and CBG-binding affinities of XP were significantly increased in OCP vs. HV (both P<0.001). LCM solutions for HV yielded higher N for po (3.1±1.2) vs. iv (1.4±0.4) (P=0.007). Limitations: Model solutions depend on the precision and accuracy of measurements made in clinical samples. Although the competitor (XP) is treated as a single compound in LCM, it may represent the combined effects (e.g., harmonic means) of multiple ligands. Conclusions: (i) matrix notation was useful to development of non-linear, iterative equations for competitive ligand-BP interactions, which were applicable to both forward and inverse solutions of the matrix problems (ii) LCM provided a better fit to experimental data for OC compared to MM, providing (indirect) evidence that competing ligand(s) (XP) influence the distribution of cortisol in serum samples obtained from women on OC, (iii) inter- and intra-subject variability for N was observed for both models, suggesting that assignment of a single population value for cortisol-albumin binding constant may be unrealistic. References: 1Feldman H. et al (1972) Analytical Biochemistry 45:530-566 2Perogamvros et al (2011) Clin Endo 74:30-36 Presentation: Thursday, June 15, 2023