Aurora B is a mitotic kinase that is essential for cell division. Because it is mutated or overexpressed in a range of cancer types, it has been suggested as a novel therapeutic target. Currently chemical inhibitors against Aurora B are in various phases of clinical trials for treatment of solid tumors and leukemia. Information regarding the molecular requirements for the reported phenotypes of Aurora B inhibition such as cell cycle arrest, activation of the tumor suppressor p53 and its target p21 are not well understood. In this study, I investigated the requirements for p21 induction after Aurora B inhibition. I found that p38 is phosphorylated and activated when Aurora B is inhibited. Experiments with chemical inhibitors against p38 indicate that p38 is required for p21 induction and cell cycle arrest in response to Aurora B inhibition. p53 induction after impairment of Aurora B function and the recruitment of p53 to its binding site in the p21 gene promoter occur independently of p38 signaling. Instead, I found that p38 is required for the enrichment of the elongating RNA Polymerase II in the coding region of the p21 gene. Furthermore, p38 is required for formation of the full-length p21 mRNA transcript. These data indicate that p38 promotes the transcriptional elongation of p21 gene in response to Aurora B inhibition. In further experiments I could show that the p21 causes cell cycle arrest due to a decrease in E2F-dependent transcription by promoting the dephosphorylation of the retinoblastoma protein. Using synchronized cells I could show that the induction of p21 in response to Aurora B inhibition requires transition through an aberrant mitosis and does not occur in cells that are arrested in interphase. Interestingly, p38, p53 and p21 are already induced by partial inhibition of Aurora B, which results in aneuploidy but not in cytokinesis failure and in tetraploidy. This supports the notion that activation of p38-p53-p21 signaling correlates with aneuploidy but not with tetraploidy or binucleation. Partial inhibition of Aurora B also leads to increased generation of reactive oxygen species (ROS), which are required for the activation of p38, p21 and cell cycle arrest. Based on these observations I propose the following model: Inhibition of Aurora B leads to chromosome missegregation resulting in aneuploidy. This results in increased generation of ROS (reactive oxygen species) possibly through proteotoxic stress caused by an imbalance of protein synthesis in aneuploid cells. ROS triggers the activation of p38, which then stimulates the transcriptional elongation of p21 resulting in cell cycle arrest. Aneuploidy, proteotoxic stress and oxidative stress are hallmarks of cancer cells. Based on my results reported in this study, I suggest that the combination of Aurora B inhibitors with drugs that specifically target aneuploid cells might be a novel strategy for cancer therapy, as this is a lethal combination for proliferation of cancer cells.