GUVac formation induced by F-actin disruption in matrix-deprived mammary epithelial cells. (A) Representative differential interference contrast (DIC) microscopy images of suspended MCF-10A cells showing entotic vacuole or GUVac formation after culture for 24 h. Nuclei were stained with Hoechst 33342. Scale bars: 20 μm (main panels) or 10 μm (inset). (B) Percentage of suspended MCF-10A cells showing entotic vacuole or GUVac formation after 24 h (n = 772). (C) Representative bright field microscopy images of MCF-10A cells suspended with F-actin cytoskeleton inhibitors. Scale bar: 15 μm (D) Percentage of suspended MCF-10A cells showing GUVac formation after exposure to the indicated drugs for the indicated times. DMSO: 0 h (n = 512), 6 h (n = 723), 12 h (n = 690), 24 h (n = 690). LatB: 6 h (n = 634), 12 h (n = 693), 24 h (n = 428). Cytochalasin D: 6 h (n = 613), 12 h (n = 618), 24 h (n = 464). Y-27632: 6 h (n = 448), 12 h (n = 601), 24 h (n = 660). (E) Percentage of suspended HMEpiCs showing GUVac formation after incubation with DMSO or LatB for the indicated times. DMSO: 6 h (n = 722), 12 h (n = 647), 24 h (n = 417). LatB: 6 h (n = 1225), 12 h (n = 1505), 24 h (n = 1335). (F) Percentage of suspended MCF-10A cells showing GUVac formation after incubation with the indicated drugs for 18 h. DMSO (n = 1087), LatB (n = 1634), jasplakinolide (n = 2578), Rho inhibitor (n = 839), Y-27632 (n = 1322), EHT 1864 (n = 905), ML 141 (n = 997), nocodazole (n = 1342), blebbistatin (n = 520). (G) Representative DIC microscopy images with phalloidin staining show disruption of the actin cytoskeleton and the GUVac formation in SpvB-expressing cells after suspension culture for 24 h. Scale bar: 10 μm (H) Percentage of GUVac formation in control or SpvB-expressing MCF-10A cells. (I) Representative DIC microscopy images after suspension culture of indicated cell lines for 24hours with LatB. Scale bar: 10 μm (J) Representative DIC images of suspended MCF-10A cells treated with DMSO or LatB for the indicated times. White arrowheads indicate the accumulation of vacuoles. Nuclei were stained with Hoechst 33342. Scale bar, 20 μm. (K) Representative fluorescence and DIC time-lapse images of MCF-10A cells expressing mCherry-H2B and membrane-targeted EGFP obtained at the indicated times after the onset of LatB treatment. Times are presented as hour:minute:second (h:m:s). Scale bars, 10 μm. All quantitative data are means ± SD. The n values represent the total number of cells quantified for two (D) or three (B, E, F and H) independent experiments. **P < 0.01 (two-tailed unpaired t test) for the indicated comparison (F and H) or versus the corresponding value for DMSO (E).

Macropinocytosis-like process contributes to GUVac formation. (A) Representative TEM images of suspended MCF-10A cells treated with DMSO or LatB for the indicated times. Scale bars, 5 µm (main panels) or 500 nm (inset). (B) Number of inwardly curved plasma membrane structures with a diameter of >500 nm detected by TEM in individual suspended MCF-10A cells treated with DMSO or LatB for 6 h. Each data point represents an individual cell analyzed. (C) Representative DIC images of suspended MCF-10A cells treated with EIPA or LatB for 18 h. Nuclei were stained with Hoechst 33342. Scale bar, 20 μm. (D) Percentage of cells showing GUVac formation in experiments as in (C). DMSO (n = 1489), EIPA (n = 1256), LatB (n = 1482), LatB + EIPA (n = 1159). (E) Representative DIC and FITC-dextran fluorescence images of suspended MCF-10A cells treated with LatB or EIPA for 18 h. The cells were exposed to FITC–dextran (70 kDa) at 1 mg/ml. Scale bar, 20 μm. (F) Percentage of cells showing FITC-dextran uptake in experiments as in (E). DMSO (n = 797), LatB (n = 1534), LatB + EIPA (n = 821). (G) Representative super-resolution SIM time-lapse images of suspended MCF-10A cells showing the plasma membrane labeled by CellMask obtained at the indicated times after DMSO or LatB treatment. Arrowheads indicated plasma membrane invagination and vesicle formation. Scale bars, 5 μm (main panels) or 1 μm (inset). (H) Representative spinning disk confocal time-lapse images of suspended MCF-10A cells showing CellMask-labeled plasma membrane and 70 kDa dextran obtained at the indicated times after DMSO or LatB treatment. Scale bar, 5 μm. (I) Schematic showing the differences among conventional endocytosis, conventional macropinocytosis, and unconventional macropinocytosis-like process with regard to the size of endocytic cups and reliance on actin polymerization. Schematic was created using Biorender. All quantitative data are means ± SD. The n values represent the total number of cells examined in three independent experiments (D and F). *P < 0.05, ****P < 0.0001 by one-way ANOVA (B) or one-way ANOVA with Tukey’s multiple comparisons (D and F).

Recruitment of septin to the fluctuating plasma membrane drives macropinocytosis-like phenomenon. (A) Representative super-resolution SIM fluorescence images (maximum intensity projection) of MCF-10A cells expressing membrane-targeted EGFP and mCherry– septin 6 treated with indicated drugs for 6 h. Scale bars, 5 μm (main panels) or 1 μm (inset). (B) Number of assembled mCherry-septin 6 structures per cell quantified in experiments as in (A). DMSO (n = 24), LatB (n = 47), LatB + FCF (n = 41). Each data point represents an individual cell analyzed from two independent experiments. (C) Percentage of suspended MCF-10A cells showing GUVac formation after incubation with the indicated drugs for 18 h. LatB (n = 783), LatB + FCF (n = 1016). (D) Representative super-resolution SIM fluorescence images (maximum intensity projection) of MCF-10A cells expressing membrane-targeted EGFP and either mCherry-septin 6 WT or a mutant lacking the AH domain (ΔAH), which were suspended in the presence of LatB for 3 hours. Scale bars, 5 μm (main panels) or 1 μm (inset). (E) Percentage of cells showing GUVac formation for MCF-10A cells that had been transfected with control or dynamin 2 siRNAs and treated with LatB in suspension culture for 18 h (left panel). siControl (n = 746), siDNM2-1 (n = 822), siDNM2-2 (n = 712). Immunoblot analysis of dynamin 2 in MCF-10A cells transfected with control or two independent dynamin 2 siRNAs (right panel). (F) Percentage of cells showing GUVac formation for MCF-10A cells expressing WT or dominant negative mutant (K44A) forms of dynamin 2 that had been suspended in the presence of LatB for 18. WT (n = 607), K44A (n = 737). (G) Percentage of MCF-10A cells showing GUVac formation after suspension in the presence of the indicated inhibitors and incubation for 18 h. DMSO (n = 1423), LatB (n = 1151), LatB + Dynasore (n = 1407), LatB + Dynole 34-2 (n = 947), LatB + OctMAB (n = 699), LatB + MitMAB (n = 1404). All quantitative data are means ± SD. The n values represent the total number of cells examined in two (B) or three independent experiments (C and E–G). **P < 0.01, ***P < 0.001, ****P < 0.0001 by one-way ANOVA with Tukey’s multiple comparisons (B, E, and G) or two-tailed unpaired t test (C and F).

PI(3)P is required for vacuole fusion during GUVac formation. (A) Percentage of suspended MCF-10A cells showing GUVac formation after treatment with the indicated drugs for 18 h. DMSO (n = 1122), LatB (n = 1386), LatB + VPS34-IN1 (n = 989), LatB + LY294002 (n = 876), LatB + GDC-0941 (n = 787), LatB + TGX-221 (n = 801). (B) Percentage of cells showing GUVac formation for MCF-10A cells transfected with the indicated siRNAs and suspended in the presence of LatB for 24 h. siRNAs: control (n = 456), p110α (n = 718), p110β (n = 709), PI3K-C2α (n = 421), VPS34 (n = 295). (C) Percentage of cells showing GUVac formation for nontargeting control (NTC), PI3K-C2α KO, or VPS34 KO MCF-10A cells suspended in the presence of DMSO or LatB for 24 h. NTC/DMSO (n = 746), NTC/LatB (n = 730), PI3K-C2α KO/LatB (n = 824), VPS34 KO/LatB (n = 939). (D) Representative TEM images of suspended WT, PI3K-C2α KO, or VPS34 KO MCF-10A cells treated with LatB for the indicated times. Scale bars, 2 µm. (E) Representative super-resolution SIM fluorescence images of MCF-10A cells expressing membrane-targeted EGFP that had been suspended in the presence of the indicated drugs for the indicated times. The nucleus was denoted by the letter “N” and outlined with a dashed line. Arrows indicated vacuoles. Scale bars, 5 μm. (F) Maximum diameter of vacuoles in each cell for experiments as in (E). LatB: 3 h (n = 29), 6 h (n = 39), 12 h (n = 36). LatB + VPS34-IN1: 3 h (n = 47), 6 h (n = 43), 12 h (n = 39). (G) Representative super-resolution SIM fluorescence images of MCF-10A cells expressing membrane-targeted EGFP and either mCherry-2xFYVE or mCherry–PH-TAPP1 that had been suspended in the presence of DMSO or LatB for 6 h. All quantitative data are means ± SD. The n values represent the total number of cells examined in three independent experiments (A-C, and F). *P < 0.05, **P < 0.01, ****P < 0.0001; ns, not significant by one-way ANOVA with Tukey’s multiple comparisons (A-C) or two-tailed unpaired t test (F).

GUVac formation promotes cell survival in altered actin and matrix environments. (A–C) MCF-10A cells were suspended in the presence of LatB or EDTA for the indicated times and then immunostained for cleaved caspase-3. Nuclei were stained with Hoechst 33342. Representative DIC and fluorescence images at 24 h (scale bar, 20 μm) (A), the percentage of cells positive for GUVac formation at 24 h (B), and the time course for the percentage of cells positive for cleaved caspase-3 (C) are shown. DMSO: 0 h (n = 1030), 6 h (n = 1742), 12 h (n = 1423), 24 h (n = 1273). LatB: 6 h (n = 1479), 12 h (n = 1508), 24 h (n = 1443). EDTA: 6 h (n = 1976), 12 h (n = 1578), 24 h (n = 1296). (D–F) MCF-10A cells were suspended with EDTA and in the presence of DMSO or LatB for the indicated times and then stained with trypan blue. Representative bright-field images (scale bar, 20 μm) (D) as well as the percentage of cells positive for GUVac formation (E) and the percentage of trypan blue– positive cells (F) are shown. DMSO: 24 h (n = 1003), 48 h (n = 797), 72 h (n = 943). LatB: 24 h (n = 1473), 48 h (n = 2100), 72 h (n = 1350). (G) Percentage of trypan blue–positive cells for suspended MCF-10A cells that either were simultaneously treated with both LatB and the indicated concentrations of sFasL for 24 h (a condition under which cell death signaling precedes GUVac formation) or were treated with LatB for 24 h and then incubated in the presence of sFasL for 24 h (a condition under which GUVac formation precedes cell death signaling). GUVac(–): sFasL at 0 ng/ml (n = 1193), 100 ng/ml (n = 1441), or 500 ng/ml (n = 994). GUVac(+): sFasL at 0 ng/ml (n = 1240), 100 ng/ml (n = 1418), or 500 ng/ml (n = 1529). (H and I) Nontargeting control (NTC), PI3K-C2α KO, and VPS34 KO MCF-10A cells were suspended with EDTA in the presence of DMSO or LatB for 48 h, after which the percentage of cells showing GUVac formation (H) and the percentage of trypan blue–positive cells (I) were determined. NTC: DMSO (n = 927), LatB (n = 1369). PI3K-C2α KO: DMSO (n = 627), LatB (n = 883). VPS34 KO: DMSO (n = 1026), LatB (n = 885). (J) Time-lapse bright-field images of cells with GUVacs after matrix reattachment. MCF-10A cells were suspended in the presence of LatB for 24 h, washed to eliminate LatB, and then cultured in an adhesive confocal dish for 14 h before imaging for the indicated times (hour:minute). Scale bar, 10 μm. All quantitative data are means ± SD. The n values represent the total number of cells examined in three (B, C, and G–I) or four (E and F) independent experiments. **P < 0.01, ****P < 0.0001; ns, not significant (two-tailed unpaired t test).

(A) Representative DIC images of suspended HMEpiCs treated with DMSO or LatB for 0, 6, 12, or 24 h. Nuclei were stained with Hoechst 33342. Scale bar, 10 μm. (B) Representative DIC images of suspended cells of the indicated cell lines after treatment with LatB for 24 h. Nuclei were stained with Hoechst 33342. Scale bar, 10 μm. (C) Representative fluorescence images for F-actin (phalloidin staining) or nuclei (Hoechst 33342 staining) as well as DIC images of adherent MCF-10A cells treated with DMSO, LatB, or cytochalasin D (CytoD) for 12 h. Scale bar, 10 μm. (D) Percentage of suspended MCF-10A cells showing GUVac formation after incubation for 18 h with DMSO (n = 523), LatB (n = 571), or LatB plus 2.5% Matrigel (n = 587). Data are means ± SD for the indicated numbers of cells examined in three independent experiments. **P < 0.01 (two-tailed unpaired t test).

(A) Representative super-resolution SIM time-lapse images of suspended MCF-10A cells showing CellMask-labeled plasma membrane obtained at the indicated times after DMSO or LatB treatment. Arrowheads indicated vesicle formation at the plasma membrane. Scale bar, 1 μm. (B) Representative super-resolution SIM time-lapse images of suspended MCF-10A cells showing CellMask-labeled plasma membrane and 70 kDa dextran obtained at 43 min after DMSO or LatB treatment. Scale bars, 5 μm (main panels) or 1 μm (inset).

(A) Quantitative RT-PCR analysis of relative p110α mRNA abundance in MCF-10A cells transfected with control or p110α siRNAs. Data are means ± SD from three independent experiments. ***P < 0.001 (two-tailed unpaired t test). (B) Quantitative RT-PCR analysis of relative p110β mRNA abundance in MCF-10A cells transfected with control or p110β siRNAs. Data are means ± SD from three independent experiments. ****P < 0.0001 (two-tailed unpaired t test). (C) Immunoblot analysis of PI3K-C2α and VPS34 in MCF-10A cells transfected with the indicated siRNAs. β-actin was examined as a loading control. (D) Immunoblot analysis of PI3K-C2α in nontargeting control (NTC) and PI3K-C2α KO MCF-10A cell clones. (E) Immunoblot analysis of VPS34 in VPS34 KO MCF-10A cell clones. Clones 23, 26, and 29 were selected as VPS34 KO cell lines for further study.

Immunoblot analysis of survival signaling in MCF-10A cells suspended with EDTA and in the presence of DMSO or LatB for the indicated times.

Hypothetical model for the formation of GUVac that enhances the anoikis resistance. In matrix-attached cells, cortical actin and extracellular matrix (ECM) attachment suppresses plasma membrane (PM) fluctuation and maintains membrane tension. However, in matrix-detached cells with disrupted cortical actin, the PM exhibits micron-scale membrane fluctuations. This promotes the recruitment of septin to the micrometer-sized inwardly curved plasma membrane via its amphipathic helix (AH) domain and the subsequent recruitment further accelerates the PM invagination. As a result of this macropinocytosis-like process, followed by dynamin-mediated pinching off, vacuoles accumulate within the cell and gradually coalesce, a process facilitated by the synthesis of PI(3)P, involving the activity of VPS34 and PI3K-C2α. Eventually, these processes lead to the formation of GUVac. Cells that possess GUVac demonstrate resistance to anoikis. The schematic was generated using Biorender.