On the origins of the mitotic shift in proliferating cell layers
- William T Gibson†1,
- Boris Y Rubinstein†2,
- Emily J Meyer2,
- James H Veldhuis3,
- G Wayne Brodland3,
- Radhika Nagpal4 and
- Matthew C Gibson2, 5Email author
© Gibson et al.; licensee BioMed Central Ltd. 2014
Received: 4 November 2013
Accepted: 5 May 2014
Published: 27 May 2014
During plant and animal development, monolayer cell sheets display a stereotyped distribution of polygonal cell shapes. In interphase cells these shapes range from quadrilaterals to decagons, with a robust average of six sides per cell. In contrast, the subset of cells in mitosis exhibits a distinct distribution with an average of seven sides. It remains unclear whether this ‘mitotic shift’ reflects a causal relationship between increased polygonal sidedness and increased division likelihood, or alternatively, a passive effect of local proliferation on cell shape.
We use a combination of probabilistic analysis and mathematical modeling to predict the geometry of mitotic polygonal cells in a proliferating cell layer. To test these predictions experimentally, we use Flp-Out stochastic labeling in the Drosophila wing disc to induce single cell clones, and confocal imaging to quantify the polygonal topologies of these clones as a function of cellular age. For a more generic test in an idealized cell layer, we model epithelial sheet proliferation in a finite element framework, which yields a computationally robust, emergent prediction of the mitotic cell shape distribution.
Using both mathematical and experimental approaches, we show that the mitotic shift derives primarily from passive, non-autonomous effects of mitoses in neighboring cells on each cell’s geometry over the course of the cell cycle. Computationally, we predict that interphase cells should passively gain sides over time, such that cells at more advanced stages of the cell cycle will tend to have a larger number of neighbors than those at earlier stages. Validating this prediction, experimental analysis of randomly labeled epithelial cells in the Drosophila wing disc demonstrates that labeled cells exhibit an age-dependent increase in polygonal sidedness. Reinforcing these data, finite element simulations of epithelial sheet proliferation demonstrate in a generic framework that passive side-gaining is sufficient to generate a mitotic shift.
Taken together, our results strongly suggest that the mitotic shift reflects a time-dependent accumulation of shared cellular interfaces over the course of the cell cycle. These results uncover fundamental constraints on the relationship between cell shape and cell division that should be general in adherent, polarized cell layers.
In a network of adherent cellular polygons, cell shape emerges both from autonomous and from non-autonomous effects of cell division. Mitosis alters cell geometry cell-autonomously by reducing the number of neighbors of a dividing cell (i.e., an octagon may divide into a pair of hexagons, resulting in “side loss”; Figure 1E). Simultaneously, mitosis acts cell non-autonomously by generating new neighbor interfaces for cells adjacent to the recent site of division, which results in “side gaining” (Figure 1F). Numerous theoretical and simulation studies, in combination with live-imaging experiments and clonal analysis in Drosophila, suggest that side gaining and side loss drive epithelial cell shape emergence via cell division and cell sorting, with cell division being the dominant influence [5, 10, 12, 19, 24, 27, 28].
Notably, despite the broad range of theoretically possible cell shape distributions [29–31], cell layers in many plant and animal species nevertheless converge on a conserved equilibrium distribution having approximately 25% pentagons, 45% hexagons, and 20% heptagons [9, 12, 32]. The form of the distribution is likely constrained entropically . Intriguingly, for both Drosophila (a representative animal model system) and Cucumis (a representative plant model system), the form of the mitotic cell shape distribution is nearly identical to the overall distribution, with the critical difference being that it is shifted by a single polygon class to have a heptagonal mean, in contrast to the hexagonal mean characteristic of the overall distribution (shown in Figures 1C-D). Hence, despite the independent evolutionary origins of plant and animal multicellularity , it appears that both are governed by fundamentally similar topological constraints.
Although the existence of the single-integer mitotic shift may imply a fundamental correlation between polygon class and division likelihood in proliferating cell layers, its cellular basis remains unclear. The chicken-egg nature of the problem centers on how to interpret the shift in terms of the mitotic cell cycle. For instance, one possibility is that increased cell sidedness promotes mitotic entry, although there is no functional evidence to support this view [28, 34]. An alternative interpretation is that over time, interphase cells simply gain sides as a passive consequence of adjacent mitotic events [5, 12, 24, 32, 35]. Under steady-state assumptions, for instance, a shifted (heptagonal) mean and mitotic distribution can be predicted algebraically [5, 35]. Hence, rather than indicating active cell-cycle regulation, the mitotic shift could reflect an emergent interaction between cell packing and heterogeneous proliferation. Here, in order to resolve this problem, we develop a novel mathematical framework to explicitly define the implications of non-autonomous side gaining for the mitotic cell shape distribution in cellular monolayers featuring tight cell adhesion and negligible rearrangements. Our computations predict that interphase cells should passively gain sides over time, such that cells that are more advanced in the cell cycle will tend to have a larger number of neighbors. This inference is borne out by experimental analysis of proliferating Drosophila epithelial cells as well as by finite element simulations of proliferating epithelia. We argue that the mitotic shift is likely to be a widespread geometrical feature of adherent, proliferating cellular monolayers in plants and animals.
Defining the logical relationship between polygon class and mitotic entry
The existence of the mitotic shift implies that within cell sheets, the probability of cell division F(N) must correlate with polygon class N. To show this, assume there is no such correlation, meaning that all polygon classes undergo mitosis with the same probability per unit time. Under these conditions, at steady-state, the fraction of N-sided polygonal mitotic cells would be identical to the fraction of N-sided polygonal cells overall (Additional file 1: Figure S1), contradicting the existence of the nearly identical single integer mitotic shifts observed in plant and animal cell layers [9, 12, 28]. Therefore, irrespective of the underlying mechanism, the mitotic shift implies that division probability and hence cell cycle state correlates with polygon class.
Similar reasoning leads to a second insight, which is that for tissues exhibiting a mitotic shift, cells cannot have perfectly synchronized cycles. For the case of perfectly synchronized cell cycles, all cells in the cell layer would divide simultaneously at each round of division. To show that this synchronized scenario cannot exist simultaneously with a mitotic shift, assume perfect mitotic synchrony in a proliferating cell layer (ie, a situation in which all cells divide simultaneously). Under these conditions, the distribution of mitotic and non-mitotic cells would be identical at steady-state, contradicting the existence of the shift. As a consequence of asynchronous proliferation, a time delay will necessarily exist between the divisions of neighboring cells. As a result, the average interphase cell will tend to gain additional cell-cell contacts from its apposed mitotic neighbors over the course of the cell cycle. Based on this logic, the intuitive expectation is that within cell sheets, asynchronous division will result in a positive correlation between polygon class and cell cycle state. Indeed, empirical data from previous studies has suggested that, on average, cells having more sides are more likely to undergo mitosis (according to multiple metrics, including metaphase marker staining, cell rounding, and cytokinesis [12, 24]).
A positive correlation between polygon class and cell cycle state is the default expectation, and a trend across diverse organisms
Equation (3) implies that the distribution P(N) achieves its maximum when the function F(N) crosses the average division rate P(D) from below. The fact that P(N) is uni-modal empirically indicates that F(N) crosses the value P(D) exactly once, meaning that all values on the right side of the crossing point stay above P(D), and all values on the left side of it stay below. Therefore, higher-order polygon classes tend to have a greater division probability than lower-order polygon classes. The uni-modal character of the shape distribution P(N) is observed in diverse plant and animal species (and also in simulations), suggesting that this reasoning may be general [9, 12, 28–30, 32, 36]. Consistent with the above analysis, when F(N) is assumed to have an exponential form (which is approximately true for the Drosophila wing disc epithelium  and for the epidermis of Cucumis), equation (3) implies that the distribution of polygonal cell shapes has the following form (with parameters p 1 and p 2 ; for 〈N〉 ≈ 6): , which is uni-modal.
Diverse organisms obey the constraint specified in equation ( 4)
Anacharis (leaf, abaxial)
Anacharis (leaf, adaxial)
Modeling the emergence of the mitotic shift in terms of cellular age
The distribution of neighboring cells surrounding dividing cells is approximately the same as the overall polygonal distribution of cells, P(N), which is at steady-state. This is approximately true empirically in the Drosophila wing disc .
Cell rearrangement can be neglected for the case of a single round of division. Multiple lines of evidence are consistent with this view [9, 12, 24], which simplifies analytical treatment, although it is straightforward to simulate a scenario in which rearrangements are present. Moreover, the results of the model can be directly compared with clone experiments in Drosophila to test for any potential role of rearrangement.
For analysis (including mathematical summations), we assume that all cells in the epithelium have between 4 and 9 sides, which is an empirical fact in both Drosophila and Cucumis, save for very rare 10-sided cells, which have negligible frequency.
We define the conditional probability Q w (m) that a dividing w-sided cell orients its cleavage plane so as to cleave its common interface with an m-sided neighbor. The computations below use the empirically measured values of a mean-field function Q(m), which is an average of the function Q w (m) over all possible w values . For comparison with a completely random cleavage plane, on average (denoted by angular brackets), the system has a probability of cleaving a common interface with an arbitrarily selected m-sided neighbor of a mitotic cell.
The side-gaining process as a binary tree
The algorithm to compute the mitotic polygonal cell shape distribution is written exclusively in terms of side-gaining events. Side-gaining is a direct consequence of neighbor cell mitosis, wherein the mitotic neighbor cleaves its common interface with the cell in question, thereby creating two edges where only one existed previously, and hence increasing the recipient cell’s polygon class by a single edge. Note that side gaining only occurs when the mitotic neighbor’s cleavage plane orients in a given cell’s direction; otherwise no such common interface is cleaved, and the cell’s polygon class remains unchanged. Side gaining is therefore a binary event; for each neighboring cell division, a cell either gains a single side or it does not.
For analysis, we assume that mitotic events occur stochastically. Using this approach, after k neighbor cell divisions, a polygonal cell can gain at minimum zero sides (if none of the cleavage planes point in its direction), and at most k sides (if all of the cleavage planes point in its direction). Side-gaining events are assumed to be independent.
where r is an index, and G(m,k,k) is the probability of gaining k sides after k divisions. For instance, the probability that a hexagon gains two sides after two neighbor cell divisions is just the probability Q() that a hexagon gains one side due to a neighboring division to become a heptagon, times the probability Q() that the newly-formed heptagon gains one side to become an octagon.
Visually, the function G(m,k,V) can be represented as a binary tree (Figure 2A-B). Figure 2A illustrates how multiple stochastic trajectories can lead to the same eventual side-gaining outcome. For instance, the chance that a hexagon gains zero, one, or two sides after two neighboring divisions is illustrated graphically in Figure 2B. In order to compute the chance of reaching a particular polygon class, the paths leading to that potential outcome must be added together. For instance, there is only one path leading to a hexagonal fate (Figure 2B, bottom line). By contrast, the heptagonal fate has two paths impinging on it, which must be summed to determine the chance of the hexagon transitioning to a heptagon.
The recursion relation has two terms because there are two ways to reach G(m,k,V) from the previous division step V-1. One way is to have gained k sides already after V-1 divisions, and then to gain no sides on the Vth division. This is equivalent to following a horizontal path on the binary tree (Figure 2A). The other way is to have gained k-1 sides after V-1 divisions, and to gain the kth side on the Vth division. This corresponds to taking one of the inclined paths on the binary tree. In this Markovian framework, it is then straightforward to compute the likelihood of each possible trajectory for the side-gaining dynamics of an m-sided cell.
The stochastic dynamics of neighbor division events
which is the average waiting time until the m-cell divides.
To construct p(J m ), we first generate all possible local neighborhoods that could surround each m-sided central cell, and then compute the expected total number of dividing neighbors for each such neighborhood, which is rounded to a whole number for purposes of substitution into G. The distribution p(J m ) can then be generated by assigning probability mass to each such value of J m using the multinomial distribution, which gives the chance of observing that particular combination of neighboring cells. Numerical evaluation of equation (17) agrees closely with a Monte Carlo computation using 105 stochastically generated local neighborhoods for each m-sided polygonal central cell (Additional file 2: Figure S2). We conclude that the weighted mean-field approximation (equation 16) closely approximates the direct computation (equation 17; see Addititional file 2: Figure S2), and provides an efficient method to compute the distribution P(N|D).
Predictions of the model
Passive side gaining drives an increase in polygonal sidedness in vivo
While SCC analysis is sufficient to detect a net gain in polygonal sidedness, it does not reveal a complete integer shift analogous to the one seen in the mitotic shift (compare Figures 4C and D). To address this discrepancy, we note that a positive correlation exists between division probability and polygonal cell shape (see previous sections). SCC analysis is therefore biased topologically, because it only considers labeled cells that have not yet undergone mitosis, which tend to have fewer cell-cell contacts (and were therefore, on average, at an earlier mitotic stage at the time of clone induction). TCC analysis suffers the opposite bias; it only considers labeled cells that have already divided, which show enrichment in cell-cell contacts. Assuming that cell cycle times are roughly asynchronous at the population level, the average pair of daughter cells in a TCC is expected to have divided at the six-hour time point, which is the midpoint of the 12 hour experiment. As the average mitotic cell has seven sides at mitosis, this means that on average, each of the daughter cells is expected to have had 5.5 sides at the six-hour time point. It is therefore notable that the experimentally measured average polygonal topology of two cell clones was 6.09 (Figure 4C). By this reasoning, daughter cells in TCC’s are expected to have gained at least 0.59 sides in a six-hour period. Note this gain in sidedness is approximately ½ of the theoretically expected value of 1 side per cell per cell cycle of 12 hours. Based on these findings, we postulate that side gaining is the primary topological transformation responsible for generating the mitotic shift in the Drosophila wing disc.
Passive side gaining drives an increase in polygonal sidedness in silico
By combining mathematical and experimental approaches, we have shown that the asynchrony of cell division plays a dominant role in generating the mitotic shift within a proliferating monolayer epithelium. Based on a minimal set of assumptions, for a given dependency between polygon class and division likelihood, we have developed an analytical framework to derive the distribution of mitotic cell shapes. Our mathematical analysis, experimental results, and finite element simulations together suggest that the mitotic shift is a topological phenomenon that is primarily a consequence of the correlation between autonomous cell cycle progression and non-autonomous side gaining. Some systems may rely on additional, redundant, shape-dependent cell division induction mechanisms [28, 34]. However, particularly in light of the fact that the mitotic shift is manifest in independently evolved forms of multicellular life , the most parsimonious conclusion is that such mechanisms are not required even if they cannot be completely ruled out.
To test for a subtle role of mechanical stress in generating the mitotic shift in the Drosophila wing disc, advances in live imaging may eventually permit precise tracking of both cellular age and of cellular geometry over time [24, 42–44]. Even if a small geometric influence could be detected, the divergent genetics and divergent mechanics of plant epidermis and animal epithelia make it unlikely that such a mechanism would be conserved across kingdoms. Hence, such a hypothesized influence would most likely be a feature of a particular tissue, not a general explanation for the shift. The framework developed here places strong quantitative limits on the possible contribution of cellular geometry (or correlative mechanical stress), while simultaneously demonstrating the dominance of the division process.
Looking forward, our results lead to several questions for future analysis. Perhaps the most critical would be to ask how the frequency of cell-cell rearrangement impacts the dynamics considered here, especially if there were topological biases in the rates of different cell-cell movements. Previous studies in Drosophila wing discs have reported variable rates of neighbor exchange events, ranging from negligibly low  to high . At one unlikely extreme, a very high degree of undirected neighbor exchange events could essentially erase the mitotic shift and drive the tissue towards a hexagonal topology. This is itself an argument against the existence of large-scale neighbor exchanges in the developing wing imaginal disc. At the other extreme, a low degree of neighbor exchange events, particularly if they favored particular topological transformations, could produce more subtle perturbations of the mitotic shift or the global distribution of cell shapes. Since cell movements within epithelia are likely to play a key role in different aspects of tissue morphogenesis, understanding their implications for both the overall topology and the mitotic shift could be a key avenue for future studies.
Numerical computations were performed in Mathematica 8.0 (Wolfram Research, Inc.). For parameter fitting, Mathematica’s FindFit function was used. A full description of the methods for implementing finite-element based simulations of epithelial proliferation (see Figure 5) and topological simulations of epithelial proliferation (see Additional file 1: Figure S1) can be found elsewhere . Simulation results presented in Figure 5 are based on runs that were replicated in triplicate, with each run containing at least 1050 cell divisions. Simulation results presented in Additional file 1: Figure S1 were also replicated in triplicate, with each run containing at least 80,000 cell divisions.
To visualize the septate junctions (Figure 1), we used a neuroglian-gfp exon trap line, which was described in a previous study (nrg-gfp; ).GFP-expressing clones (Figure 4) were induced in flies of the following genotype:
yw hs-flp122; Actin5c> > Gal4,UAS-GFP/+ with a 30-minute heat shock at 37C followed by a 12-hour recovery period prior to dissection.
Wing discs expressing marked clones (Figure 4) were stained with mouse anti-discs large (1:1000 dilution, DSHB) to mark the septate junctions.
Wing Disc sample preparation and Imaging
Wing discs were dissected from wandering 3rd instar larvae in Ringers’ solution, fixed in 4% paraformaldehyde in PBS, and mounted in 70% glycerol/PBS. Discs were imaged on a Leica SP5 with a 63× glycerol objective.
Image processing procedures
Single cell clones (SCC’s) and two cell clones (TCC’s) were imaged in multiple focal planes, and were displayed as two-color image stacks (one color for the Flp-Out GFP, and one color for anti-discs large or neuroglian-GFP) in Leica’s LAS AF imaging software for the SP5 confocal microscope. Analysis was performed by hand; cells having ambiguous polygonal topology were not counted. To control for the possibility of cell sorting, SCC’s and/or TCC’s were not considered for analysis unless they were separated by at least two cell diameters within the tissue. In order to control for boundary effects, cells located on tissue folds close to the anterior-posterior (AP) or dorsal-ventral (DV) compartment boundaries were not counted. To prevent mis-identification of SCC or TCC clones, we did not consider cells for scoring if the source of the GFP signal was ambiguous (for example, if the GFP source overlapped with another bright clone in a different focal plane).For display (non-analytical) purposes, Figure 1D shows mitotic cells that have been first inverted, and then subjected to a brightness threshold cutoff in Adobe Photoshop.
Sample sizes for single cell clone (SCC) analysis and two cell clone (TCC) analysis
Samples sizes used to compute each polygonal cell shape’s respective frequency for single cell clones (SCC’s) and two cell clones (TCC’s) are as follows: Single cell clones, (4, 1; 5, 7; 6, 32; 7, 44; 8, 15; 9, 0; 10, 0). Two cell clones, (4, 5; 5, 77; 6, 162; 7, 83; 8, 17; 9, 0; 10, 0).
Sample sizes and polygonal counts by organism
Sample sizes used to compute each polygonal cell shape’s respective frequency in Drosophila, Xenopus, Hydra, and Cucumis, have been previously described. For reference, these are as follows: Drosophila, (4, 64; 5, 606; 6, 993; 7, 437; 8, 69; 9,3). Xenopus, (3, 2; 4, 40; 5, 305; 6, 451; 7, 191; 8, 52; 9, 8; 10, 2), Hydra, (4, 16; 5, 159; 6, 278; 7, 125; 8, 23; 9, 1). Cucumis, (4, 20; 5, 251; 6, 474; 7, 224; 8, 30; 9, 1). Aggregate sample sizes and polygonal frequencies for Allium, Euonymus, Dryopteris, and Anacharis have been previously described . For reference, these are as follows: Allium (n = 500 cells), (4, 0.040; 5, 0.302; 6, 0.412; 7, 0.181; 8, 0.047; 9, 0.015). Euonymous (n = 200 cells), (4, 0.030; 5, 0.290; 6, 0.400; 7, 0.220; 8, 0.066; 9, 0). Dryopteris (n = 200 cells), (4, 0.040; 5, 0.260; 6, 0.425; 7, 0.205; 8, 0.050; 9, 0.020). Anacharis (leaf, abaxial, n = 200 cells), (4, 0.025; 5, 0.240; 6, 0.595; 7, 0.100; 8, 0.035; 9, 0). Anacharis (leaf, adaxial, n = 200 cells), (4, 0.035; 5, 0.255; 6, 0.570; 7, 0.130; 8, 0.010; 9, 0). Anacharis (bud, n = 200 cells), (4, 0.055; 5, 0.295; 6, 0.450; 7, 0.200; 8, 0; 9, 0).
Sample sizes for overall and mitotic cell shape distributions
Sample sizes used to compute each polygonal cell shape’s respective frequency for resting and mitotic cells, respectively, in both Drosophila[12, 24] and Cucumis, have been described previously. For reference, these are as follows: Drosophila (overall cell shape distribution), (4, 64; 5, 606; 6, 993; 7, 437; 8, 69; 9,3). Drosophila (mitotic cell shape distribution), (4, 0; 5, 13; 6, 100; 7, 212; 8, 80; 9, 13; 10, 3). Cucumis (overall cell shape distribution), (4, 20; 5, 251; 6, 474; 7, 224; 8, 30; 9, 1). Cucumis (mitotic cell shape distribution), (4, 0; 5, 16; 6, 255; 7, 478; 8, 224; 9, 26; 10,1).
We thank Norbert Perrimon for critical comments and discussion. We are grateful to the Stowers Institute and to HHMI for financial support. WTG is a fellow of the Jane Coffin Childs foundation for medical research.
- Bohn S, Pauchard L, Couder Y: Hierarchical crack pattern as formed by successive domain divisions. Phys Rev E. 2005, 71: 046214-View ArticleGoogle Scholar
- Bohn S, Platkiewicz J, Andreotti B, Adda-Bedia M, Couder Y: Hierarchical crack pattern as formed by successive domain divisions. II. From disordered to deterministic behavior. Phys Rev E. 2005, 71: 046215-View ArticleGoogle Scholar
- Korneta W, Mendiratta S, Menteiro J: Topological and geometrical properties of crack patterns produced by the thermal shock in ceramics. Phys Rev E. 1998, 57: 3142-10.1103/PhysRevE.57.3142.View ArticleGoogle Scholar
- Goehring L, Mahadevan L, Morris SW: Nonequilibrium scale selection mechanism for columnar jointing. Proc Natl Acad Sci. 2009, 106: 387-392. 10.1073/pnas.0805132106.PubMed CentralView ArticlePubMedGoogle Scholar
- Rivier N, Schliecker G, Dubertret B: The stationary state of epithelia. Acta Biotheor. 1995, 43: 403-423. 10.1007/BF00713562.View ArticlePubMedGoogle Scholar
- Weaire D, Rivier N: Soap, cells and statistics—random patterns in two dimensions. Contemp Phys. 1984, 25: 59-99. 10.1080/00107518408210979.View ArticleGoogle Scholar
- Glazier JA, Gross SP, Stavans J: Dynamics of two-dimensional soap froths. Phys Rev A. 1987, 36: 306-10.1103/PhysRevA.36.306.View ArticlePubMedGoogle Scholar
- Glazier JA, Anderson MP, Grest GS: Coarsening in the two-dimensional soap froth and the large-Q Potts model: a detailed comparison. Philos Mag B. 1990, 62: 615-645. 10.1080/13642819008215259.View ArticleGoogle Scholar
- Lewis FT: The Correlation Between Cell Division and the Shapes and Sizes of Prismatic Cells in the Epidermis of Cucumis. Anat Rec. 1928, 38: 341-376. 10.1002/ar.1090380305.View ArticleGoogle Scholar
- Dubertret B, Rivier N: The renewal of the epidermis: a topological mechanism. Biophys J. 1997, 73: 38-44. 10.1016/S0006-3495(97)78045-7.PubMed CentralView ArticlePubMedGoogle Scholar
- Miri M, Rivier N: Universality in two-dimensional cellular structures evolving by cell division and disappearance. Phys Rev E Stat Nonlin Soft Matter Phys. 2006, 73: 031101-View ArticlePubMedGoogle Scholar
- Gibson MC, Patel AB, Nagpal R, Perrimon N: The emergence of geometric order in proliferating metazoan epithelia. Nature. 2006, 442: 1038-1041. 10.1038/nature05014.View ArticlePubMedGoogle Scholar
- Corson F, Hamant O, Bohn S, Traas J, Boudaoud A, Couder Y: Turning a plant tissue into a living cell froth through isotropic growth. Proc Natl Acad Sci. 2009, 106: 8453-8458. 10.1073/pnas.0812493106.PubMed CentralView ArticlePubMedGoogle Scholar
- Rivier N, Lissowski A: On the correlation between sizes and shapes of cells in epithelial mosaics. J Phys A: Math Gen. 1982, 15: L143-L148. 10.1088/0305-4470/15/3/012.View ArticleGoogle Scholar
- Peshkin MA, Strandburg KJ, Rivier N: Entropic predictions for cellular networks. Phys Rev Lett. 1991, 67: 1803-1806. 10.1103/PhysRevLett.67.1803.View ArticlePubMedGoogle Scholar
- Dumais J: Can mechanics control pattern formation in plants?. Curr Opin Plant Biol. 2007, 10: 58-62. 10.1016/j.pbi.2006.11.014.View ArticlePubMedGoogle Scholar
- Hamant O, Heisler MG, Jönsson H, Krupinski P, Uyttewaal M, Bokov P, Corson F, Sahlin P, Boudaoud A, Meyerowitz EM, Couder Y, Traas J: Developmental patterning by mechanical signals in Arabidopsis. Science. 2008, 322: 1650-1655. 10.1126/science.1165594.View ArticlePubMedGoogle Scholar
- Lintilhac PM, Vesecky TB: Stress-induced alignment of division plane in plant tissues grown in vitro. Nature. 1984, 307: 363-364. 10.1038/307363a0.View ArticleGoogle Scholar
- Dubertret B, Aste T, Ohlenbusch HM, Rivier N: Two-dimensional froths and the dynamics of biological tissues. Phys Rev E. 1998, 58: 6368-6378. 10.1103/PhysRevE.58.6368.View ArticleGoogle Scholar
- Graustein WC: On the Average Number of Sides of Polygons of a Net. Ann Math, Second Series. 1931, 32: 149-153. 10.2307/1968421.View ArticleGoogle Scholar
- Desch CH: Second report to the Beilby Prize Committee of the Institute of Metals on the Solidification of metals from the liquid state. J Inst Met. 1919, 22: 241-263.Google Scholar
- Thompson DW: On growth and form. 1942, New York: Macmillan: Cambridge: University PressGoogle Scholar
- Shraiman BI: Mechanical feedback as a possible regulator of tissue growth. Proc Natl Acad Sci U S A. 2005, 102: 3318-3323. 10.1073/pnas.0404782102.PubMed CentralView ArticlePubMedGoogle Scholar
- Gibson WT, Veldhuis JH, Rubinstein B, Cartwright HN, Perrimon N, Brodland GW, Nagpal R, Gibson MC: Control of the mitotic cleavage plane by local epithelial topology. Cell. 2011, 144: 427-438. 10.1016/j.cell.2010.12.035.PubMed CentralView ArticlePubMedGoogle Scholar
- Quyn AJ, Appleton PL, Carey FA, Steele RJ, Barker N, Clevers H, Ridgway RA, Sansom OJ, Näthke IS: Spindle orientation bias in gut epithelial stem cell compartments is lost in precancerous tissue. Cell Stem Cell. 2010, 6: 175-181. 10.1016/j.stem.2009.12.007.View ArticlePubMedGoogle Scholar
- Li W, Kale A, Baker NE: Oriented cell division as a response to cell death and cell competition. Curr Biol. 2009, 19: 1821-1826. 10.1016/j.cub.2009.09.023.PubMed CentralView ArticlePubMedGoogle Scholar
- Farhadifar R, Roper JC, Aigouy B, Eaton S, Julicher F: The influence of cell mechanics, cell-cell interactions, and proliferation on epithelial packing. Curr Biol. 2007, 17: 2095-2104. 10.1016/j.cub.2007.11.049.View ArticlePubMedGoogle Scholar
- Aegerter-Wilmsen T, Smith AC, Christen AJ, Aegerter CM, Hafen E, Basler K: Exploring the effects of mechanical feedback on epithelial topology. Development. 2010, 137: 499-506. 10.1242/dev.041731.View ArticlePubMedGoogle Scholar
- Patel AB, Gibson WT, Gibson MC, Nagpal R: Modeling and inferring cleavage patterns in proliferating epithelia. PLoS Comput Biol. 2009, 5: e1000412-10.1371/journal.pcbi.1000412.PubMed CentralView ArticlePubMedGoogle Scholar
- Sahlin P, Hamant O, Jönsson H: Statistical Properties of Cell Topology and Geometry in a Tissue-Growth Model. Complex Sciences: First International Conference, Complex 2009, Shanghai, China, February 23–25, 2009 Revised Papers, Part 1: Springer Berlin Heidelberg. Edited by: Zhou J. 2009, 971-979.Google Scholar
- Sahlin P, Jonsson H: A modeling study on how cell division affects properties of epithelial tissues under isotropic growth. PLoS One. 2010, 5: e11750-10.1371/journal.pone.0011750.PubMed CentralView ArticlePubMedGoogle Scholar
- Korn RW, Spalding RM: The Geometry of Plant Epidermal Cells. New Phytol. 1973, 72: 1357-1365. 10.1111/j.1469-8137.1973.tb02114.x.View ArticleGoogle Scholar
- Knoll AH: The multiple origins of complex multicellularity. Annu Rev Earth Planet Sci. 2011, 39: 217-239. 10.1146/annurev.earth.031208.100209.View ArticleGoogle Scholar
- Lewis FT: The geometry of growth and cell division in epithelial mosaics. Am J Bot. 1943, 30: 766-776. 10.2307/2437550.View ArticleGoogle Scholar
- Dormer KJ: Fundamental tissue geometry for biologists. 1980, Cambridge, UK: Cambridge University Press, 149-Google Scholar
- Brodland GW, Veldhuis JH: Computer simulations of mitosis and interdependencies between mitosis orientation, cell shape and epithelia reshaping. J Biomech. 2002, 35: 673-681. 10.1016/S0021-9290(02)00006-4.View ArticlePubMedGoogle Scholar
- Struhl G, Basler K: Organizing activity of wingless protein in Drosophila. Cell. 1993, 72: 527-540. 10.1016/0092-8674(93)90072-X.View ArticlePubMedGoogle Scholar
- Milan M, Campuzano S, Garcia-Bellido A: Cell cycling and patterned cell proliferation in the wing primordium of Drosophila. Proc Natl Acad Sci U S A. 1996, 93: 640-645. 10.1073/pnas.93.2.640.PubMed CentralView ArticlePubMedGoogle Scholar
- Johnston LA, Prober DA, Edgar BA, Eisenman RN, Gallant P: Drosophila myc regulates cellular growth during development. Cell. 1999, 98: 779-790. 10.1016/S0092-8674(00)81512-3.View ArticlePubMedGoogle Scholar
- Datar SA, Jacobs HW, de la Cruz AFA, Lehner CF, Edgar BA: The Drosophila cyclin D–Cdk4 complex promotes cellular growth. EMBO J. 2000, 19: 4543-4554. 10.1093/emboj/19.17.4543.PubMed CentralView ArticlePubMedGoogle Scholar
- Gibson W, Gibson M: Cell topology, geometry, and morphogenesis in proliferating epithelia. Curr Top Dev Biol. 2009, 89: 87-114.View ArticlePubMedGoogle Scholar
- Bosveld F, Bonnet I, Guirao B, Tlili S, Wang Z, Petitalot A, Marchand R, Bardet PL, Marcq P, Graner F, Bellaïche Y: Mechanical control of morphogenesis by Fat/Dachsous/Four-jointed planar cell polarity pathway. Science. 2012, 336: 724-727. 10.1126/science.1221071.View ArticlePubMedGoogle Scholar
- Aldaz S, Escudero LM, Freeman M: Live imaging of Drosophila imaginal disc development. Proc Natl Acad Sci U S A. 2010, 107: 14217-14222. 10.1073/pnas.1008623107.PubMed CentralView ArticlePubMedGoogle Scholar
- Aigouy B, Farhadifar R, Staple DB, Sagner A, Roper JC, Jülicher F, Eaton S: Cell flow reorients the axis of planar polarity in the wing epithelium of Drosophila. Cell. 2010, 142: 773-786. 10.1016/j.cell.2010.07.042.View ArticlePubMedGoogle Scholar
- Zartman J, Restrepo S, Basler K: A high-throughput template for optimizing Drosophila organ culture with response-surface methods. Development. 2013, 140: 667-674. 10.1242/dev.088872.View ArticlePubMedGoogle Scholar
- Morin X, Daneman R, Zavortink M, Chia W: A protein trap strategy to detect GFP-tagged proteins expressed from their endogenous loci in Drosophila. Proc Natl Acad Sci. 2001, 98: 15050-15055. 10.1073/pnas.261408198.PubMed CentralView ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.