Monopolar HeLa cells fail to generate a single polarity axis for cytokinesis when actin filaments are disrupted.

Splitting mitotic cells in two is not the one-way signaling road it once seemed, based on evidence from Hu et al. The group identifies a positive feedback loop that creates a furrow at cytokinesis. The findings also throw a wrench in the well-accepted dogma of microtubule dynamic instability.

Hu and colleagues devised a monopolar HeLa system to determine how the cytokinetic furrow is created. Monopolar spindles are useful because they can be forced into cytokinesis synchronously, but the chromosomes don't get in the way as they would in bipolar cells. In the round, monopolar HeLa cells, the spindle and its chromosomes were organized in a radially symmetric manner. Gradually, however, this symmetry waned, and microtubules and cortical furrow components began to polarize toward one side of the cell.

The de novo creation of this asymmetry implies that a positive feedback loop exists. Previous models proposed that a one-way signal is sent from spindle microtubules to the cortex to create a furrow. But the new results indicate that the cortex must talk back to microtubules, to stabilize them and further ensure polarization. Polarization required microtubules, Aurora B kinase activity, actin, myosin, and RhoA, which activates cortical contraction.

Because the monopolar spindles lacked a spindle midzone, where plus ends of microtubules from opposite poles overlap, their plus ends were easily identified. The group was stunned to note that these microtubule tips appeared nondynamic and terminated evenly at a distance from the growing, polarizing cortex. Their behavior contradicts popular dynamic instability models and suggests that plus ends may be capped in monopolar spindles and perhaps even at the midzone of bipolar spindles.

Within the gap between the plus ends and the expanding cortex, Aurora B colocalized with actin filaments. In bipolar spindles, Aurora B shifts from the kinetochore to microtubule plus ends. The new results suggest that it then travels along actin filaments to signal from the spindle to the cortex and possibly back again.

Hu, C.-K., et al.
J. Cell Biol.