Bevacizumab-loaded nanoparticles (B-NP) were prepared by a desolvation process followed by freeze-drying, without any chemical, physical or enzymatic cross-linkage. Compared with typical HSA nanoparticles crosslinked with glutaraldehyde (B-NP-GLU), B-NP displayed a significantly higher mean size (310 nm vs. 180 nm) and a lower negative zeta potential (-15 mV vs. -36 mV). On the contrary, B-NP displayed a high payload of approximately 13% when measured by a specific ELISA, whereas B-NP-GLU presented a very low bevacizumab loading (0.1 mu g/mg). These results could be related to the inactivation of bevacizumab after reacting with glutaraldehyde. From B-NP, bevacizumab was released following an initial burst effect, proceeded by a continuous release of bevacizumab at a rate of 6 mu g/h. Cytotoxicity studies in ARPE cells were carried out at a single dose up to 72 h and with repeated doses over a 5-day period. Neither bevacizumab nor B-NP altered cell viability even when repeated doses were used. Finally, B-NP were labeled with Tc-99m and administered as eye drops in rats. Tc-99m-B-NP remained in the eye for at least 4 h while Tc-99m-HSA was rapidly drained from the administration point. In summary, HSA nanoparticles may be an appropriate candidate for ocular delivery of bevacizumab.