Our goal was to deliver therapeutically active macromolecules into the cytosol of target cells. First, attempts were made to prepare virosomes that specifically interact with OVCAR-3 cells (human ovarian cancer cells). Detergent solubilized influenza virus envelopes were reconstituted forming virosomes. Cell specificity was introduced by incorporating PEG-derivatized lipids with mAB 323/A3 (Fab' fragments) connected to their distal PEG end. These cell-specific, modified virosomes maintained their fusogenic activity when lowering the pH. Most importantly, antibody-mediated binding was a prerequisite for low-pH induced membrane fusion. However, basically, there are two problems with this approach: (1) these virosomes are quite leaky and (2) virosomes can be expected to be immunogenic. A solution to tackle leakage and potential immunogenicity of these site-specific liposomal structures is to use immuno-PEG-liposomes with a pH-dependent fusogen inside the liposome. The system that we designed to test this concept consisted of (1) the fusogenic di-peptide dINF-7, (2) the monoclonal antibody 425 connected to the distal end of PEG-PE (for site specific binding and endosomal uptake), (3) diphtheria toxin chain A (DTA, as carrier-dependent active compound) and phosphatidylcholine/cholesterol as 'bilayer backbone'. A series of tests were performed to show that selective binding and pH-dependent destabilization of (endosomal) membranes indeed occurred. To test the cytotoxic activity of these DTA loaded liposomes, OVCAR-3 cells were used for testing. OVCAR-3 cells express the epidermal growth factor receptor, which is the ligand for antibody 425. In vitro, these site specific and fusogenic liposomes showed a remarkable, cell specific cytotoxic effect.