Julia Parlow: Diffusion of model drugs in extracellular matrix mimetic hydrogels: Towards an in vitro model for subcutaneous injection

Abstract

Subcutaneous injection is the preferred administration route for degradation-sensitive therapeutic peptides and proteins because it is patient-friendly and can provide sufficient bioavailability. However, both bioavailability and absorption rate vary considerably and are difficult to predict. As yet, no existing in vitro model can reliably estimate the pharmacokinetics of new drug compounds, highlighting the need for methods that offer mechanistic insight into transport processes under physiologically relevant conditions.

This thesis investigates the transport properties of model drugs with varying physicochemical properties in hydrogels that mimic the extracellular matrix of subcutaneous tissue. Partitioning and diffusion coefficients were measured in gels and in solution using confocal laser scanning microscopy and fluorescence recovery after photobleaching (FRAP). Experimental results were compared with theoretical models describing obstruction and electrostatic interactions, and the experimentally derived parameters were incorporated into a physiologically based pharmacokinetic (PBPK) model to predict subcutaneous absorption rate and bioavailability.

The results confirm that FRAP enables reliable quantification of local diffusion coefficients in heterogeneous gels. Combined with information on distribution within the gel and gel–solution partitioning, the method provides insight into interactions of the compounds with the gel matrix. Composite collagen–hyaluronic acid gels proved to be the most physiologically relevant, offering appropriate interaction sites along with reproducibility and stability.

Diffusivities in solution depended not only on molecular weight but also on molecular shape and oligomerization. In gels, near-neutral compounds generally showed reduced partitioning and diffusion, whereas highly positively charged peptides and proteins were enriched in polymer-dense regions. This enrichment substantially decreased diffusivity due to obstruction and electrostatic binding. The electrostatic interactions were decreased in the presence of human serum albumin for compounds containing a high-affinity albumin-binding domain, effectively facilitating their transport. Albumin also altered the apparent hydrodynamic size of some molecules by solubilizing oligomers or forming larger complexes.

Finally, it was demonstrated that diffusion and partitioning data from collagen–hyaluronic acid gels correlated with in vivo absorption rates of therapeutic proteins and could be used to improve PBPK model predictions. Overall, the presented method offers a valuable characterization tool that can facilitate the design of new drug candidates with predictable pharmacokinetic profiles.