A biomimetic approach to organic solvent-free microencapsulation of proteins based on the self-healing capacity of poly (DL)-lactic-co-glycolic acid (PLGA) microspheres containing glycosaminoglycan-like biopolymers (BPs), was examined. To screen BPs, aqueous solutions of BP [high molecular weight dextran sulfate (HDS), low molecular weight dextran sulfate (LDS), chondroitin sulfate (CS), heparin (HP), hyaluronic acid (HA), chitosan (CH)] and model protein lysozyme (LYZ) were combined in different molar and mass ratios, at 37 °C and pH 7. The BP-PLGA microspheres (20–63 µm) were prepared by a double water-oil-water emulsion method with a range of BP content, and trehalose and MgCO3 to control microclimate pH and to create percolating pores for protein. Biomimetic active self-encapsulation (ASE) of proteins [LYZ, vascular endothelial growth factor165 (VEGF) and fibroblast growth factor (FgF-20)] was accomplished by incubating blank BP-PLGA microspheres in low concentration protein solutions at ~24 °C, for 48 h. Pore closure was induced at 42.5 °C under mild agitation for 42 h. Formulation parameters of BP-PLGA microspheres and loading conditions were studied to optimize protein loading and subsequent release. LDS and HP were found to bind >95% LYZ at BP:LYZ >0.125 w/w, whereas HDS and CS bound > 80% LYZ at BP:LYZ of 0.25–1 and < 0.33, respectively. HA-PLGA microspheres were found to be not ideal for obtaining high protein loading (>2% w/w of LYZ). Sulfated BP-PLGA microspheres were capable of loading LYZ (~2–7 % w/w), VEGF (~ 4% w/w), and FgF-20 (~2% w/w) with high efficiency. Protein loading was found to be dependent on the loading solution concentration, with higher protein loading obtained at higher loading solution concentration within the range investigated. Loading also increased with content of sulfated BP in microspheres. Release kinetics of proteins was evaluated in-vitro with complete release media replacement. Rate and extent of release were found to depend upon volume of release (with non-sink conditions observed < 5ml release volume for ~18mg loaded BP-PLGA microspheres), ionic strength of release media and loading solution concentration. HDS-PLGA formulations were identified as having ideal loading and release characteristics. These optimal microspheres released ~ 73–80 % of the encapsulated LYZ over 60 days, with > 90 % of protein being enzymatically active. Nearly 72% of immunoreactive VEGF was similarly released over 42 days, without significant losses in heparin binding affinity in the release medium.