Acidic fibroblast growth factor (aFGF) lacks a classical signal sequence for secretion via the exocytic pathway but yet has to be released from cells in order to interact with high affinity receptors on the cell surface. To study the release process, we have expressed human aFGF in HeLa cells using a T7 RNA polymerase-driven vaccinia virus system. The high level of expression in combination with an efficient antibody allowed us to analyze the release of aFGF by pulse-chase experiments, and to immunolocalize the protein in transfected cells. In the absence of heparin, only negligible amounts of aFGF were detected in the medium during a IS hr chase period. However, if heparin was present during the chase, readily detectable amounts (about 10–20% of total) of aFGF were found in the medium during the 15 hr chase. Extracellular aFGF was first detected at 8 hr and increased during the chase. Concomitantly, only small amounts of lactate dehydrogenase activity, used as a cytoplasmic marker, was released from the cells. Further analyses indicated that heparin both stabilized the protein from degradation and prevented the binding of released aFGF to extracellular heparan-sulfate proteoglycans. Thus, both factors contributed to the increased recovery of aFGF in the presence of heparin. The slow and inefficient release of aFGF is consistent with our previous results obtained in insect cells expressing aFGF to a very high level, as well as with those obtained by others in cultured cells producing FGF.Immunolocalization using an affinity purified antibody made against native aFGF, showed strong fluorescence in the nuclei in most cells, while staining in the cytoplasm was usually weaker and varied between cells. The nuclear localization was confirmed by subcellular fractionation and immunoblot analysis. At an early time point following transfection (4 hr), aFGF was preferentially localized to the nuclei, while the distribution of the protein between cytoplasm and nuclei was about equal at later time points (12 hr). Thus, we conclude that aFGF is capable of efficiently entering the nucleus and apparently becoming trapped there.