System-level deficiencies in Aurora B control in cancers

Yusuke Abe,T. Hirota

DOI: https://doi.org/10.1080/15384101.2016.1185850

2016-05-10

Cell Cycle

Abstract:Defects in chromosome segregation machinery give rise to generation of aneuploid cells containing abnormal number of chromosomes. The chromosomal passenger complex (CPC) composed of 4 core subunits, Aurora B kinase, INCENP, Survivin and Borealin/Dasra ensures faithful chromosome segregation primarily by releasing incorrect kinetochore-microtubule attachments. Earlier studies have found that a partial inhibition of Aurora B activity by a low-dose of inhibitor treatment already induces chromosome segregation errors, whereas more potent inhibition is needed to affect Aurora B-mediated histone H3 phosphorylation, chromosome assembly and movement and cytokinesis. These results have implied that fidelity of chromosome segregation depends on full activity of Aurora B. Although cancer cells, in general, have elevated rates of chromosome missegregation, molecular basis underlying such pathological conditions has been a longstanding question. The heterochromatin protein 1 (HP1) associates with CPC through INCENP and thereby is recruited to centromeres in mitosis. The significance of this HP1-CPC association was unclear until recently, when an allosteric effect of HP1 on Aurora B was found. HP1 doubled the substrate-processing rate of Aurora B in vitro, and this enhancement of the kinase activity is required for Aurora B to phosphorylate kinetochore substrates in cells. Remarkably, the fraction of HP1-bound CPC was significantly decreased in all the cancer cell lines tested, which led to the conclusion that insufficient Aurora B activity is a widespread feature of cancers. A corollary of these findings was that HP1-mediated full Aurora B activity supports the fail-safe chromosome segregation and indeed, disruption of HP1 binding to the CPC in non-transformed cells increased the rate of chromosome missegregation, without detectably affecting other mitotic events regulated by Aurora B. Deprivation of HP1 from the CPC lowered Aurora B activity and perturbed the phosphorylation of kinetochore substrates, but it also impaired the enrichment of Aurora B at centromeres. These results are consistent with the notion that CPC loading onto centromeres is regulated by multiple positive feedback loops, involving Aurora B activity itself. For instance, Aurora B enhances centromeric accumulation of the CPC by regulating Haspin kinase-mediated histone H3-phospho-Thr3, which in turn recruits the CPC at centromeres. Moreover, because Aurora B is known to promote chromatin association of condensin I, which confers physical stability to centromeres, accumulation of Aurora B activity at centromeres must be important to maintain the structural integrity at that site. Thus the “scaffolds” for CPC may become unstructured without an allosteric effect of HP1. Based on these observations, a reasonable view would be that the proper centromeric localization and function of the CPC are controlled by self-regulating mechanisms of Aurora B (Fig. 1). The finding that non-transformed cells have a robust mechanism to correct attachment errors by increasing Aurora B recruitment to misaligned chromosomes may underscore the importance of this systemlevel control. In a wide range of cancer cells, defined localization of Aurora B at centromeres is noticeably impaired to various extents, and it is instead remained undetached from chromosome arms (unpublished observations, Abe Y and Hirota T). In light of positive feedback loops controlling localization of Aurora B, this ill-defined centromeric localization can be reasonably explained by the finding that the activity of Aurora B is low by decreased HP1 binding in cancer cells. It is noteworthy to point out the interlinked positive feedback loop between HP1 and the CPC. As the association of HP1 with the CPC is stabilized by Aurora B-mediated phosphorylation of HP1 through an incompletely understood mechanism, the stable binding of HP1 to the CPC will keep the activity of Aurora B in a high level state (Fig. 1). Thus, the existence of system level controls allows us to predict that the activity of Aurora B is settled down to a steady state with lower levels, which renders cancer cells predisposed to chromosomes segregation errors in mitosis. The results that over-expression of HP1 did not facilitate formation of HP1-CPC at all well supports the notion that cancer cells suffer from the system-level deficiency of Aurora B, the condition reflected by an inherent limitation of HP1-binding ability. What limits the binding capacity of HP1 to the CPC in cancer cells is an important following question. It cannot be