Protein-loaded dextran microspheres were prepared by a water-in-water emulsion technique. With this technique, an aqueous solution of methacrylated dextran (dex-MA) is emulsified in an aqueous solution of poly(ethylene glycol) (PEG). Subsequently, the dispersed dex-MA phase is crosslinked by radical polymerization of the dextran-bound methacryloyl groups. This method renders microspheres with a hydrogel character of which the crosslink density can be controlled by the water content and the degree of substitution of the dex-MA (DS, the number of methacrylates per 100 glucopyranose residues). If an IgG solution was added to the dex-MA/PEG aqueous system prior to the polymerization reaction, the protein could be encapsulated in the dextran microspheres with a high yield (88-98%). The release of IgG was studied as a function of the water content, the DS and the degradation rate of the microspheres. The microspheres were rendered degradable by co-encapsulation of an endo-dextranase. Non-degrading microspheres mainly showed a burst release, which decreased with increasing crosslink density. By either a low water content (50%, w/w, or lower) or a high DS (DS 13), it was possible to reduce the burst release to about 10%, meaning that almost complete entrapment of the protein could be achieved. The release of IgG from degrading microspheres was predominantly dependent on the DS and the amount of encapsulated dextranase. No differences in release of IgG from microspheres with and without dextranase were observed at high DS (DS 13). This was ascribed to the inability of the enzyme to degrade these microspheres. On the other hand, the entrapped protein was completely released from enzymatically degrading microspheres with a DS 4. Moreover, the release rate of IgG was proportional to the degradation rate of these microspheres (depending on the amount of co-encapsulated dextranase). Interestingly, an almost zero-order release was observed from these microspheres for periods up to 30 days.