Dysregulation of protein activity, caused by alterations in protein sequence, expression, or localization, is associated with numerous diseases. In order to control the activity of harmful protein entities, affinity ligands such as proteins, oligonucleotides or small molecules can be engineered to specifically interact with them to modulate their function. In this thesis, non-immunoglobulin based affinity proteins known as affibody molecules are used to functionally inhibit proteins important for signaling through pathways that are overactive in different cancers.   In Paper I and Paper II, affibody molecules with high affinity for the receptor tyrosine kinases HER2 or EGFR are expressed in the secretory compartments of model cancer cell lines SKOV3 or A431 using a retrovirus-based gene delivery system. Equipping the affinity proteins with an ER retention tag, the affibody molecules together with their target protein are retained in the secretory compartments as shown by confocal fluorescence imaging. Flow cytometric analysis showed a 60 % or 80 % downregulation of surface located HER2 or EGFR in these cell lines, respectively. A significant decreased in proliferation rate of the cells was also observed, which for EGFR retention could be correlated with inhibition of phosphorylation in the kinase domain. In Paper III, novel affibody molecules interacting with the hormone binding site of the insulin growth factor-1 receptor were generated. One variant had high (1.2 nM) affinity for the receptor and could be used for immunofluorescence analysis and for receptor pull-out from cell lysates. Addition of this affibody molecule to MCF-7 cells had a dose dependent growth inhibitory effect on the cells. In Paper IV, novel affibody molecules against the intracellular oncoproteins H-Ras and Raf-1 were selected and characterized, and they proved to be specific for their target proteins. Mapping experiments showed that the affibody molecules selected against H-Ras interacted at over-lapping epitopes not affecting the interaction between Ras and Raf. In contrast, the predominant variant isolated during selection against Raf-1 could completely inhibit the Ras/Raf interaction in a real-time biospecific interaction analysis.   Taken together, the affibody molecules presented here and the strategies by which they are used to interfere with cancer related proteins and pathways may be valuable tools for further investigation of these systems and may possibly also be used to generate molecules suitable for cancer therapy. / QC 20100818