Composite membranes are fabricated from hydroxyapatite (HAP) and gelatin for immunoisolation of cells. The films were fabricated by crosslinking 5 wt%, 10 wt%, and 20 wt% gelatin with 1 wt% glutaraldehyde (GA) in the presence of HAP. Fourier transform infrared spectroscopy analysis confirms imide bond formation between GA and gelatin, while the crystal structure of HAP powder remains unchanged from X-ray diffraction analysis. The degree of crosslinking depends on crosslinking time and gelatin concentration. For 5% and 10% gelatin, the degree of crosslinking levels off at 90% within 48 h. From scanning electron microscopy micrographs, the microstructure of the composite membrane depends on the amount of gelatin used in the crosslinking reaction. The mechanical strength of the composite membrane could be enhanced by increasing the gelatin concentration. BET analysis indicates that pore size of the micropores on the surface HAP/gelatin agglomerates decreases with increasing gelatin concentration. However, the macropore, through which diffusion of molecules occurs, is larger at higher gelatin concentrations. The permeability coefficients of different molecules through a HAP/gelatin composite membrane increase with increasing gelatin concentration and is inversely correlated with the molecular weight of the molecule. For immunoisolation of cells, the diffusion of large molecules stimulated by the immune system can be rejected by a chamber constructed from the HAP/gelatin membrane. Insulinoma cells were encapsulated in alginate-poly-L-lysine-alginate microcapsules and enclosed in a HAP/gelatin chamber. The chamber did not impair the viability and function of insulinoma cells and cells can secrete insulin in response to glucose concentration change. The chamber is therefore useful for the physiologically controlled secretion of insulin in response to the blood glucose level. Intraperitoneal transplantation of the chamber into streptozotocin-induced diabetic SD rats could maintain normal blood glucose levels in test animals for up to 60 days without immunosuppression. (c) 2012 Elsevier B.V. All rights reserved.