Kicking the cell to see whether determinism drops out.

Abstract

Andrew Moore Editor-in-Chief You might remember this from physics classes at school: you have two metal sheets with regularly spaced holes, but the periodicity of spacing differs between the sheets. Now superimpose the sheets and hold them up to the light. If the holes are relatively closely spaced, and the difference in periodicity is not very large, you'll see the so-called Moiré effect, a pattern of dots with periodicity and shape at a larger scale than either of the individual ones. The individual sheets both have deterministic, rather than random, hole spacing. Now turn the experiment around: you are presented with a pattern of dots, and are told that at least one component of the pattern is a deterministic – i.e. regular – array of holes. Now you need to find (or make) a second sheet with an array of holes that reproduces a similar pattern when superimposed. Will it be a “deterministic” second sheet, or one with random hole-spacing? The analogy is symbolizes the “kicked cell cycle” model presented by Mizrahi et al. in this issue 1. This tests the proposition that there are two oscillatory processes in cells that influence cell division, and that because, overwhelmingly, they are not of the same periodicity (and hence phase), the synchronization of the cell cycle between single cell divisions is not well conserved, but that a pattern of synchronization conservation appears across generations, i.e. horizontally in the cell-division dendrogram: crucial in this system is the comparison between the mother cell – at the top of the dendrogram – and those cells at two levels further down, the so-called “cousin” cells. You could say that this is the periodicity of the cellular Moiré pattern, to re-use the analogy. Basically, if cousin cells are in phase with each other, but poorly correlated in phase with mother cells, there seems to be a second deterministic component in the system, but with different periodicity from the cell cycle. Essentially, at every division point in the dendrogram, the cell cycle and the other factor are at a different phase relative to each other, and horizontally across the dendrogram at any one point, this phase discrepancy is the same: hence giving cell-cycle synchrony amongst cousin cells. The supposition that the “other” factor is also an oscillator is almost implicit in the observation that phase synchrony appears horizontally across the dendrogram. Variation between generations arises because the “other” factor influences the cell cycle duration, and is different in strength at each cell division point in a lineage. In the vast majority of cells, the periodicity of division is not identical to the periodicity of the cell's circadian clock, and hence at each division, cell cycle duration is “kicked” to a variable degree. This is, in fact, the basis on which the “kicked cell cycle” model makes the circadian clock “drop out” as the most likely deterministic factor that produces the seeming stochasticity of cell division in cell cultures, rapidly leading to a culture of desynchronized cells: apparent desynchronization in cell cultures is, hence, not a result of stochasticity, but of two deterministic events of different periodicity, one of which influences the other! We just can't see the cellular Moiré pattern because the cells are jumbled up. Mizrahi et al. generalize this “cousin-mother inequality” to make a model that could be used for identifying other types of determinism in cell physiology by looking at the values of the given variable vertically and horizontally in a dividing cell population: for example, size control, gene expression, behavioral diversity and organelle inheritance. Wherever we see such cousin-mother inequality, we should suspect an underlying deterministic process, even if at a population level stochasticity seems to be the order of the day – though, as Mizrahi et al. admit, there is currently a technical difficulty in clonally tracking sister and cousin cells. There is also – as the authors fleetingly point out – the possibility that unequal distribution of very sparsely represented cytoplasmic or nuclear components during division produces truly stochastic patterns in a dividing population. And so a final thought occurs to me: are not all cell divisions unequal in terms of the balance of degraded cellular components that that each cell “receives”? Is not one cell slightly “younger” than the other? Could this represent a challenge for the “kicked cell cycle” model? Or could the “kicked cell cycle” model be a useful tool for investigating this very effect, revealing which systematic inequalities of division influence perceived stochasticity at population level. Andrew Moore Editor-in-Chief