Glycosaminoglycans (GAGs) are of interest for biomedical applications because of their ability to retain proteins (e.g. growth factors) involved in cell-to-cell signaling processes. In this study, the potential of GAG-based microgels for protein delivery and their protein release kinetics upon encapsulation in hydrogel scaffolds were investigated. Monodisperse hyaluronic acid methacrylate (HAMA) and chondroitin sulfate methacrylate (CSMA) micro-hydrogel spheres (diameters 500-700 μm), were used to study the absorption of a cationic model protein (lysozyme), microgel (de)swelling, intra-gel lysozyme distribution and its diffusion coefficient in the microgels dispersed in buffers (pH 7.4) of varying ionic strengths. Upon incubation in 20 mM buffer, lysozyme was absorbed up to 3 and 4 mg mg⁻¹ dry microspheres for HAMA and CSMA microgels respectively, with loading efficiencies up to 100%. Binding stoichiometries of disaccharide : lysozyme (10.2 : 1 and 7.5 : 1 for HAMA and CSMA, respectively) were similar to those for GAG-lysozyme complex coacervates based on soluble GAGs found in literature. Complex coacervates inside GAG microgels were also formed in buffers of higher ionic strengths as opposed to GAG-lysozyme systems based on soluble GAGs, likely due to increased local anionic charge density in the GAG networks. Binding of cationic lysozyme to the negatively charged microgel networks resulted in deswelling up to a factor 2 in diameter. Lysozyme release from the microgels was dependent on the ionic strength of the buffer and on the number of anionic groups per disaccharide, (1 for HAMA versus 2 for CSMA). Lysozyme diffusion coefficients of 0.027 in HAMA and <0.006 μm² s⁻¹ in CSMA microgels were found in 170 mM buffer (duration of release 14 and 28 days respectively). Fluorescence Recovery After Photobleaching (FRAP) measurements yielded similar trends, although lysozyme diffusion was likely altered due to the negative charges introduced to the protein through the FITC-labeling resulting in weaker protein-matrix interactions. Finally, lysozyme-loaded CSMA microgels were embedded into a thermosensitive hydrogel scaffold. These composite systems showed complete lysozyme release in ∼58 days as opposed to only 3 days for GAG-free scaffolds. In conclusion, covalently crosslinked methacrylated GAG hydrogels have potential as controlled release depots for cationic proteins in tissue engineering applications.