Analysis of microtubule growth dynamics arising from altered actin network structure and contractility in breast tumor cells

Abstract

The periphery of epithelial cells is shaped by opposing cytoskeletal physical forces generated predominately by two dynamic force generating systems-growing microtubule ends push against the boundary from the cell center, and the actin cortex contracts the attached plasma membrane. Here we investigate how changes to the structure and dynamics of the actin cortex alter the dynamics of microtubules. Current drugs target actin polymerization and contraction to reduce cell division and invasiveness; however, the impacts on microtubule dynamics remain incompletely understood. Using human MCF-7 breast tumor cells expressing GFP-tagged microtubule end-binding-protein-1 (EB1) and coexpression of cytoplasmic fluorescent protein mCherry, we map the trajectories of growing microtubule ends and cytoplasmic boundary respectively. Based on EB1 tracks and cytoplasmic boundary outlines, we calculate the speed, distance from cytoplasmic boundary, and straightness of microtubule growth. Actin depolymerization with Latrunculin-A reduces EB1 growth speed as well as allows the trajectories to extend beyond the cytoplasmic boundary. Blebbistatin, a direct myosin-II inhibitor, reduced EB1 speed and yielded less straight EB1 trajectories. Inhibiting signaling upstream of myosin-II contractility via the Rho-kinase inhibitor, Y-27632, altered EB1 dynamics differently from Blebbistatin. These results indicate that reduced actin cortex integrity can induce distinct alterations in microtubule dynamics. Given recent findings that tumor stem cell characteristics are increased by drugs which reduce actin contractility or stabilize microtubules, it remains important to clearly define how cytoskeletal drugs alter the interactions between these two filament systems in tumor cells.

Figure 1

Opposing forces of microtubule expansion and actin cortex contraction. In a normal, adherent epithelial cell, there is a cortex of cross-linked actin filaments beneath the plasma membrane that is under contraction mediated by the non-muscle myosin-II motor protein (red arrows). Microtubules are nucleated at the microtubule-organizing center (MTOC) and grow outward from the cell center toward the plasma membrane (green arrows). Microtubule end-binding protein-1 (EB1) binds preferentially to the growing GTP-capped microtubule plus end and forms comets projecting from the cell center. Microtubules grow outward to reach the network of cortical actin filaments, which have an estimated mesh size of 100 nm. Interaction of EB1 with cortical proteins, such as APC, also mediate capture of microtubules. Latrunculin-A depolymerizes actin, reducing the barrier to microtubule expansion. Phosphorylation of myosin-II by the Rho-kinase (ROCK) increases contractility, but can be inhibited by the small molecule ROCK inhibitor, Y-27632. Similarly, Blebbistatin binds directly to myosin-II and reduces contractility.

Figure 2

Actin disruption promotes microtubule extension beyond the cytoplasmic boundary. Human MCF-7 breast cancer cells were treated with vehicle control (0.5% DMSO), 5 μM Latrunculin-A, 25 μM Blebbistatin or 10 μM of Y-27632 for 30 min, fixed with formaldehyde and fluorescently stained for actin localization (red), microtubules (green) or DNA (blue). Confocal microscopy images are shown for each channel and condition, along with an overlay of all channels (right column; scalebar = 10 μm).

Figure 3

Tracking algorithm identifies dynamic EB1-GFP tips and trajectories. (A) Example of EB1 tips found. (B) Example of tips tracked. (C) Individual trajectories: (i) normal trajectory in cell bulk. (ii) Trajectory in cell bulk treated with 25 μM blebbistatin. (iii) Trajectory in cell bulk treated with 5 μM latrunculin. (iv) Trajectory in cell bulk treated with Y27632. (v) Normal trajectory near cell edge. (vi) Trajectory in near cell edge treated with 25 μM blebbistatin. (vii) Trajectory near cell edge treated with 5 μM latrunculin. (viii) Trajectory near cell edge treated with Y27632. (D) Trajectory overlays for normal cell where blue represents earlier time frames and red represents later time frames (scalebar = 10 μm). (E) Trajectory overlays for cell treated with 25 μM blebbistatin. (F) Trajectory overlays for cell treated with 5 μM Latrunculin-A. (G) Trajectory overlays for cell treated with Y-27632 (10 μM).

Figure 4

Analysis of localization and distribution of dynamic microtubule ends. (A) Cell body boundary from mean mCherry image taken across 50 frames (left). (B) Mean distance from the boundary for all 4 conditions (bottom left). Control (n = 19) treated cells (0.1% DMSO) show an average distance from cell body boundary of measured 1.7 ± 0.2 μm. Cells treated with 5 μM Latrunculin-A (n = 18) show an average distance of 0. 16 ± 0.2 μm which is a significant decrease when compared to control (anova p < 0.0001 and ks-test p = 1.0178 × 10⁻⁶). Cells treated with 25 μM Blebbistatin (n = 16) have an average distance of 1.5 ± 0.2 μm comparable to controls with no significant difference (anova p = 0.9397 and ks-test p = 0.0528). Cells treated with 10 μM Y-27632 (n = 19) have an average tip distance of 0.8 ± 0.3 μm which is significantly smaller than control (anova p = 0.0263 and ks-test p = 0.0181). Additionally, microtubule tip distance from the boundary for latrunculin treated cells was significantly less than blebbistatin treated cells (anova p = 0.0004 and ks-test p = 6.4616 × 10⁻⁶). (C) Percentage of particles outside the boundary (left), near the cell body boundary (within 10% of mCherry-defined cell body boundary) and in the cell bulk (right).

Figure 5

Average speed (μm s⁻¹) for all 4 conditions. (A) Average speed across all regions: control (black, n = 24) at 0.075 ± 0.002 μm s⁻¹, 5 μM Latrunculin-A (purple, n = 18), at 0.065 ± 0.004 μm s⁻¹, 25 μM Blebbistatin (red, n = 22) at 0.053 ± 0.002 μm s⁻¹, and 10 μM Y-27632 (yellow, n = 25) at 0.077 ± 0.003 μm s⁻¹. (B) Break down of speed for all 4 conditions broken into 3 regions from left to right: outside the cell body boundary, near the cell body boundary (the 10% of points inside the cell body closest to the boundary) and within the cell bulk (remaining area inside the cell body): speeds for control (black) were 0.057 ± 0.003 μm s⁻¹, 0.065 ± 0.003 μm s⁻¹, and 0.084 ± 0.003 μm s⁻¹ respectively; speeds for 5 μM Latrunculin-A (purple), were 0.048 ± 0.003 μm s⁻¹, 0.058 ± 0.004 μm s⁻¹, and 0.071 ± 0.005 μm s⁻¹ respectively; speeds for 25 μM Blebbistatin (red) were 0.037 ± 0.003 μm s⁻¹, 0.046 ± 0.002 μm s⁻¹, and 0.056 ± 0.003 μm s⁻¹ respectively; and speeds for 10 μM Y-27632 (yellow) were 0.059 ± 0.004 μm s⁻¹, 0.062 ± 0.002 μm s⁻¹, and 0.088 ± 0.003 μm s⁻¹ respectively. Error bars represent SEM.

Figure 6

Orientation autocorrelation as a function of time for control (black squares), Blebbistatin (red squares), Latrunculin-A (purple squares), and Y-27632 (yellow squares).

Figure 7

Orientation autocorrelation as a function of space for control (black squares), Blebbistatin (red squares), Latrunculin-A (purple squares), and Y-27632 (yellow squares).