Blebs promote cell survival by assembling oncogenic signalling hubs

Abstract

Most human cells require anchorage for survival. Cell-substrate adhesion activates diverse signalling pathways, without which cells undergo anoikis-a form of programmed cell death¹. Acquisition of anoikis resistance is a pivotal step in cancer disease progression, as metastasizing cells often lose firm attachment to surrounding tissue^2,3. In these poorly attached states, cells adopt rounded morphologies and form small hemispherical plasma membrane protrusions called blebs^4-11. Bleb function has been thoroughly investigated in the context of amoeboid migration, but it has been examined far less in other scenarios¹². Here we show by three-dimensional imaging and manipulation of cell morphological states that blebbing triggers the formation of plasma membrane-proximal signalling hubs that confer anoikis resistance. Specifically, in melanoma cells, blebbing generates plasma membrane contours that recruit curvature-sensing septin proteins as scaffolds for constitutively active mutant NRAS and effectors. These signalling hubs activate ERK and PI3K-well-established promoters of pro-survival pathways. Inhibition of blebs or septins has little effect on the survival of well-adhered cells, but in detached cells it causes NRAS mislocalization, reduced MAPK and PI3K activity, and ultimately, death. This unveils a morphological requirement for mutant NRAS to operate as an effective oncoprotein. Furthermore, whereas some BRAF-mutated melanoma cells do not rely on this survival pathway in a basal state, inhibition of BRAF and MEK strongly sensitizes them to both bleb and septin inhibition. Moreover, fibroblasts engineered to sustain blebbing acquire the same anoikis resistance as cancer cells even without harbouring oncogenic mutations. Thus, blebs are potent signalling organelles capable of integrating myriad cellular information flows into concerted cellular responses, in this case granting robust anoikis resistance.

Figure 1:. Bleb inhibition disrupts anoikis resistance in melanoma cells.

(A) Fraction of cell surface comprised of blebs for MV3 cells treated with different concentrations of WGA. Dashed lines separate quartiles. Dots represent individual cells. (B) Cell death as a function of WGA treatment at different doses for adhered and detached melanoma cells. Cells grown for 24 hours and assayed for cell death using ethidium homodimer staining. All treatment groups grown and assayed in simultaneous paired experiments, which were performed three times for each cell line (See Fig. S1 for individual experiments). Replicate data were normalized by subtracting negative control values from each treatment group. Error bars represent 95% confidence intervals. All 5 experimental groups were treated with the same dosages of 0, 2.5, 5, 10, and 20 μg/mL WGA. Cell counts for all replicates in ascending order of WGA dosage (see Table 1 for individual counts): MV3 Det. (470, 602, 479, 501, 562), MV3 Adh. (688, 569, 597, 426, 544), M498 Det. (290, 291, 159, 187, 133), M498 Adh. (405, 378, 347, 285, 316), A375 Det. (867, 797, 670, 847, 679), A375 Adh. (576, 453, 589, 578, 555), Prestressed A375 Det. (496, 493, 649, 317, 325), Prestressed A375 Adh. (495, 581, 448, 504, 475), MAPKi A375 Det. (170,201,240,214,264), MAPKi A375 Adh. (363,374,349,415,354). (C) Cell surface renderings of representative cells showing the spatial variation of intracellular mean curvature. (D) Cell death upon bleb inhibition using VitroGel coffins for adhered and detached melanoma cells. Cells grown for 24 hours in either integrin-binding VitroGel-RGD (adhered) or non-integrin-binding VitroGel (detached) and assayed for cell death using ethidium homodimer staining. Dots represent individual experiments. Cell counts for all replicates (see Table 1 for individual counts): MV3 (367, 389), A375 (461, 509).

Figure 2:. Bleb-generated plasma membrane curvature drives the assembly of cortical septin structures.

(A) Surface rendering of either the local intensity of the murine SEPT6-GFP probe within 1 μm of the cell surface or local intracellular mean curvature. Cells embedded in soft bovine collagen. (B) Directional correlation between blebs and septin localization in 17 MV3 cells and 2025 blebs. Cumulative correlative distribution is shown as orange solid line, randomized bleb localization control is shown as a dashed blue line, and zero correlation is shown as black dotted line. (C) Local intensity of septin signal as a function of local intracellular curvature, using same cells. The orange line indicates curvature/intensity relationship on the surface of blebs, the blue line on non-bleb surfaces, and the dashed purple line on the entirety of surfaces. Error bands represent 95% confidence intervals. (D) Probability distributions of local positive intracellular mean curvature for all cells in Fig. 1A with WGA dosage shown in μg/mL.(E) Local septin intensity and intracellular mean curvature as a function of distance from bleb edges using same cells as Fig. 2B. (F) Localization of the SEPT6-GFP probe in MV3 melanoma cells with diverse perturbations of bleb formation. Representative cells received treatment as follows: WGA (50 ug/ml), VitroGel, H1152 (1 μM), NSC668394 (10 μM). Maximum intensity projections (MIP) and single z-slices of 0.16 micron thickness. Arrowheads indicate septin accumulation in perturbed cells in regions with residual high-curvature. Cells embedded in soft bovine collagen. (G) Fraction of cortical voxels (within 0.96 μm of surface) in basal and bleb-inhibited MV3 cells with septin intensity higher than cytoplasmic mean intensity. WGA group treated at 10 μg/mL. Dashed lines separate quartiles. Basal/VitroGel tested with two sample t-test using pooled variance (p=0.013), normality tested with Shapiro-Wilk (p=0.2579 & 0.31), variance tested with two-tailed F test (p=0.448). Basal/WGA tested with Welch’s t-test (p<0.0001), normality tested with Shapiro-Wilk (p=0.4539 and 0.7941), variance tested with two-tailed F-test (p=0.0114) (H) Same data as Fig. 2G expressed as a function of cell surface fraction possessing intracellular curvature either above (right, κ > 0.4 μm⁻¹) or below (left,0 > κ > 0.4 μm⁻¹) septin recruiting threshold (R² p-values = 0.0032, 0.0054, & 0.000027). (I) Septin localization in representative MV3 melanoma cell expressing septin curvature-sensing mutant SEPT6(ΔAH)-GFP. (J) Time-lapse sequence showing changes in murine SEPT6-GFP probe localization over 14 minutes in maximum intensity projections of a representative MV3 cell. Individual time points shown in grayscale, left; and in a pseudo-color composite, right. Cell embedded in soft bovine collagen. (K) Time-lapse showing murine SEPT6-GFP probe accumulation during an individual blebbing event. Single z-slice of 0.16 μm thickness. Cell embedded in soft bovine collagen. (L) Overlaid temporal cross-correlation functions between SEPT6-GFP probe intensity and intracellular mean curvature magnitude for contigs comprised of positive (orange/red/yellow) or negative (blue/purple/green) intracellular mean curvature. Colors denote subdivision of datasets depending on dynamicity, as indicated in figure. Data from an MV3 cell embedded in soft bovine collagen imaged for 4 minutes with a stack acquisition rate of ~0.83 Hz. (M) SEPT6-GFP probe localization in representative MV3 melanoma cell expressing septin polymerization mutant SEPT2(33-306). (N) In red, local SEPT6-GFP intensity as a function of distance from bleb edges in MV3 cells expressing SEPT2(33-306), overlaid with SEPT6-GFP intensity curve from Fig 2E in blue.

Figure 3:. Septins are necessary for bleb-dependent anoikis resistance.

(A) Murine SEPT6-GFP probe localization in different melanoma cell lines and conditions. Maximum intensity projections (top) and single z-slices of 0.16 micron thickness (bottom) shown for representative cells. . Cells embedded in soft bovine collagen. (B) Cell death upon septin inhibition with 50 μM FCF for adhered and detached melanoma cells. Cells grown as in Fig 1B and treated with either FCF or EtOH control for 24 hours. Data were normalized by subtracting paired negative control values from each treatment group. Dots represent individual experiments. Cell counts for all, with control counts in parentheses (see Table 1 for individual counts): MV3 Adh.440(470), MV3 Det 644(688), M498 Adh. 298(290), M498 Det. 338(405), A375 Adh. 823(867), A375 Det. 562(576), Prestressed A375 Adh. 418(541), Prestressed A375 Det. 169(205), MAPKi A375 Adh. 231(233), MAPKi A375 Det. 516(702).

Figure 4:. Septins scaffold NRAS, promoting NRAS/MAPK and NRAS/PI3K survival signaling.

(A) Spearman correlation between NRAS and septin signal distributions on the surfaces of MV3 melanoma cells, compared to septin vs collagen negative control. White datapoints represent cells whose correlations were not significantly higher than analyses performed on the same surfaces with one signal randomly scrambled. Septin and NRAS signal distributions from the median cell in the dataset shown on the right. (B) Observed NRAS-GFP % enrichment at the cortex (voxels within 0.96 μm of surface) of individual unperturbed and septin-inhibited MV3 cells. Basal/FCF tested with Welch’s T-test (p<0.0001), normality tested with Shapiro-Wilk (p=0.2191 & 0.2198), variance tested with two-tailed F test (p=0.0054). Basal/SEPT2(33-306) tested with two sample t-test using pooled variance (p=0.0101), normality tested with Shapiro-Wilk (p=0.2191 & 0.7246), variance tested with two-tailed F test (p=0.768).Dashed lines separate quartiles. (C) Earth Mover’s Distance (EMD) between observed NRAS-GFP and homogenous distribution of equivalent signal as measured for each unperturbed or septin-inhibited MV3 cell. Control/FCF tested with Mann-Whitney U test (p= 0.000023 ), normality tested with Shapiro-Wilk (p= 0.07679 and 0.9351). Control/SEPT2(33-306) Mann-Whitney U test (p= 0.0029 ), normality tested with Shapiro-Wilk (p= 0.07679 and 0.03938). (D) Cell death upon expression of dominant negative NRAS(S17N) for adhered and detached MV3 cells. Cells grown as in Fig 1B for 24 hours. Data were normalized by subtracting paired negative control values from each treatment group. Dots represent individual experiments. Cell counts for all replicates, with control counts in parentheses (see Table 1 for individual counts): Adh. 1105(1039), Det. 799(734). (E) Effect of septin inhibition upon ERK activation levels in adhered and detached MV3 cells. Percent change in ERK activation between unperturbed and septin inhibited MV3 cells measured by ERK-nKTR-GFP biosensor. Cells grown as in Fig 1B for 12 hours. Dots represent individual experiments. Cell counts for all replicates (see Table 1 for individual counts): Adh. Control 234, Adh. FCF 236, Det. Control 405, Det. FCF 339. (F) Effect of detachment upon ERK activation levels in unperturbed, septin-inhibited, and bleb-inhibited MV3 cells. Percent change in ERK activation between attached and detached MV3 cells measured by ERK-nKTR-GFP biosensor. Cells grown as in Fig 1B for 12 hours. Dots represent individual experiments. Cell counts for all replicates (see Table 1 for individual counts): Adh. Control 276, Det. Control 294, Adh. SEPT2(33-306) 274, Det. SEPT2(33-306) 232, Adh. VitroGel 240, Det. VitroGel 205. (G) Spearman correlation between NRAS/Akt-PH and Septin/Akt-PH signal distributions on the surfaces of MV3 melanoma cells, compared to septin vs collagen negative control. White datapoints represent cells whose correlations were not significantly higher than analyses performed on the same surfaces with one signal randomly scrambled. Septin and Akt-PH signal distributions from the median cell in the dataset shown above. (H) Effect of septin or bleb inhibition upon PI3K activity in individual MV3 cells. PI3K activity measured with the PI3K biosensor Akt-PH-GFP and expressed as fraction of pixels brighter than cytosolic intensity in normalized sum intensity projections. Median cells from all groups shown on the right.

Figure 5:. Disrupting bleb attenuation yields anoikis resistance to fibroblasts.

(A) Recently detached MEF cells expressing SEPT6-GFP, imaged either before or after bleb attenuation. Maximum intensity projections (above) and single z- slices of 0.16 micron thickness (below) shown for representative cells. Cells embedded in soft bovine collagen. (B) Fraction of MEF cells showing blebby morphologies 90, 120, and 180 minutes after detachment from substrate. Orange datapoints indicate control MEF cells, green datapoints indicate MEFs expressing DYN2(K44A) in paired, same-day experiments (solid or dashed lines indicate paired data). Representative cells after 180 minutes of detachment shown to the right. Total cell counts for each timepoint, with control group in parenthesis (see Table 1 for individual counts): 90 min 115(180), 120 min 37(90), 180 min 91(126). (C) Additional caspase activation after 4 hours in detached MEF cells compared to paired adhered cells. Orange datapoints indicate control MEF cells, while green datapoints indicate MEFs expressing DYN2(K44A). Red datapoints indicate DYN2(K44A)-expressing cells treated w/ 10 μg/mL WGA and purple indicates DYN2(K44A)-expressing cells treated w/ 50 μM FCF. Caspase activity measured using CellEvent Caspase-3/7 biosensor. Total cell counts for each condition, with adhered group in parenthesis (see Table 1 for individual counts): Control 164(168), DYN2(K44A) 187(178), DYN2(K44A)+WGA 160(183), DYN2(K44A)+FCF 196(137).