Endostatin, a 20 KDa fragment of collagen XVIII, was shown to have an inhibitory effect on angiogenesis and can potentially be used as a tumor growth suppressor. To obtain the amount needed for testing, the protein was successfully cloned and expressed in Pichia pastoris. At the end of the fermentation process, the concentration of the endostatin in the culture was 50 mg per liter, accompanied by 400 gr per liter (wet weight) of biomass. Before the protein can be captured and purified on a packed bed of heparin-Sepharose, the biomass must be removed. Because of the high biomass concentration, conventional biomass removal techniques like centrifugation or filtration are inefficient and cumbersome. Therefore, the expanded-bed adsorption technique was chosen as an alternative approach. An efficient procedure for the initial recovery and purification of the endostatin was developed. The process utilized a cation- exchanger resin instead of a heparin-based affinity resin, because its dynamic capacity was higher, even though it was affected by the high linear flow on the expanded bed. After adjusting the conductivity, pH and biomass concentration, the complete broth was pumped directly on the expanded-bed matrix (Streamline SP XL). Though the yields of protein are similar, the expanded-bed approach is superior to the packed-bed method for several reasons. The expanded-bed process was shorter (only 8 hours compared to 16 hours for the packed bed), it is cheaper, and the product has higher specific activity (29% compared with 18%). Endostatin produced by the expanded-bed adsorption method showed the expected bioactivity and is currently being tested for its potential as a tumor suppressor.