The critical role of the activation of signal transducer and activator of transcription 3 (STAT3) in the growth and survival of human tumor cells was one of the subjects of this thesis. It was found that the stable incorporation of v-Src into MCF10A cells, a human immortalized breast epithelial cell line, induced constitutive phosphorylation of STAT3 in these cells. MCF10A-v-Src cells displayed growth factor independence for proliferation and survival, and anchorage independence. However, these cells did not form tumors in nude mice, indicating that they were not fully transformed. Furthermore, inhibition of activated STAT3 in 293 cells decreased growth and induced apoptosis in these cells. The same results were obtained in 293 cells stably transfected with tetracycline-inducible dominant negative (DN) STAT3, and in 293 cells transiently expressing STAT3 RNAi. Since STAT3 promises to be a good target for cancer treatment, it was attempted to develop inhibitors through rational design. Within STAT3, the SH2-domain was chosen as being the best target site. One limitation of this site turned out to be its similarity to family members in this region, and even to other SH2-domain containing proteins, such as Src, all of which present a highly conserved pocket to bind phosphotyrosines. A medium-throughput assay to measure STAT3 dimerization in vitro was established to determine the affinity of compounds to the SH2-domain of STAT3. Peptides of different length derived from the phosphotyrosine-containing motif of STAT3 (AAPY*LKTKF) were tested in the assay. Y*L was found to be the minimal sequence to block dimerization, and out of the tested peptides, PY*LKT had the highest affinity. This was consistent with the observations made on the crystal structure of STAT3, which indicated that besides the phosphotyrosine-binding site, position + 1 and + 3 contributed most importantly to binding. The relevance of position + 1 was further confirmed by substitution of Leu+ 1 by Ala in the peptide Y*LKT, which decreased the affinity more than ten times. Finally, one non-peptidic small molecule inhibitor identified by virtual ligand screening (VLS) proved to be a STAT3 dimerization inhibitor in vitro. This is the first small molecule inhibitor of STAT3 identified so far, and might serve to study structure activity relationships (SAR) to optimize the structure and find more potent and bioavailable compounds.