Author Response: Nuclear Fascin Regulates Cancer Cell Survival

Article Figures and data Abstract Editor's evaluation Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Fascin is an important regulator of F-actin bundling leading to enhanced filopodia assembly. Fascin is also overexpressed in most solid tumours where it supports invasion through control of F-actin structures at the periphery and nuclear envelope. Recently, fascin has been identified in the nucleus of a broad range of cell types but the contributions of nuclear fascin to cancer cell behaviour remain unknown. Here, we demonstrate that fascin bundles F-actin within the nucleus to support chromatin organisation and efficient DDR. Fascin associates directly with phosphorylated Histone H3 leading to regulated levels of nuclear fascin to support these phenotypes. Forcing nuclear fascin accumulation through the expression of nuclear-targeted fascin-specific nanobodies or inhibition of Histone H3 kinases results in enhanced and sustained nuclear F-actin bundling leading to reduced invasion, viability, and nuclear fascin-specific/driven apoptosis. These findings represent an additional important route through which fascin can support tumourigenesis and provide insight into potential pathways for targeted fascin-dependent cancer cell killing. Editor's evaluation This paper significantly extends previous work suggesting a role for fascin in the nucleus, with the authors concluding that it contributes to multiple aspects of cancer cell regulation and behaviour. The authors used a combination of methodologies, including biochemistry, excellent cell and actin imaging, controlled nanobody-mediated targeting of fascin to the nucleus (when endogenous fascin has been suppressed), proteomics, cancer cell biological assays, and high-content phenotypic screening to identify potential regulators, and function, of nuclear fascin, and the consequences of maintaining too high levels of fascin in the nucleus. Fascin is an important protein in cancer cell behaviour, and this work provides novel information on dynamic active transport in and out of the nucleus, on its role in nuclear actin bundling, its binding to histone-H3, and its contribution to the DNA damage response (DDR; monitored by gammaH2AX foci accumulation), chromatin compaction, cell migration, and cancer cell invasion in an in vitro assay; moreover, dynamic regulation of nuclear fascin is important because too much triggers apoptosis. https://doi.org/10.7554/eLife.79283.sa0 Decision letter Reviews on Sciety eLife's review process Introduction Fascin is an F-actin-binding protein that promotes the parallel bundling of actin filaments (Vignjevic et al., 2006a). These actin bundles can take the form of filopodia (extending beyond the plasma membrane), or microspikes (within lamellae of migrating cells or neuronal growth cones) and are involved in numerous biological processes and pathologies (Hashimoto et al., 2011; Mattila and Lappalainen, 2008; Wood and Martin, 2002), including promoting cell migration (Adams, 2004) and during embryonic development (Zanet et al., 2009). Importantly, fascin expression is very low or absent in normal adult epithelia, and a dramatic up-regulation at both gene and protein levels has been reported in the majority of human carcinomas studied to date (Hashimoto et al., 2005). Thus, fascin is emerging as a key prognostic marker and a potential therapeutic target for metastatic disease. The migration and invasion of carcinoma cells are highly coordinated processes that depend largely on both alterations to cell-cell and cell-extracellular matrix (ECM) adhesion and to the signalling events responsible for organisation of the actin cytoskeleton (Guo and Giancotti, 2004). Cancer cells migrating in 3D ECM and in tissue assemble membrane protrusions and ECM-degrading adhesions termed invadopodia to enable tunnelling through matrix (Condeelis et al., 2005; Friedl and Wolf, 2003; Li et al., 2010). Across a range of cell types, loss of fascin function results in reduced migration and invasion in vivo (Chen et al., 2010; Hashimoto et al., 2007; Jayo et al., 2012; Kim et al., 2009; Zanet et al., 2012). Fascin control of filopodia is in part coordinated spatially by formins that nucleate assembly of F-actin structures (Pfisterer et al., 2020). Fascin expression promotes increased cytoskeletal dynamics, migratory capacity, and thereby the potential for metastasis. Fascin is made up of four β-trefoil (βt) repeats separated by short flexible linker regions. Biochemical and modelling data have identified N- and C-terminal actin-binding sites in fascin that enable efficient bundling activity (Jansen et al., 2011). We and other groups have previously demonstrated that protein kinase C (α/γ) phosphorylates the highly conserved S39 site within βt1 of fascin and this negatively regulates actin bundling, cell protrusion assembly, and migration (Adams et al., 1999; Anilkumar et al., 2003; Hashimoto et al., 2007; Parsons and Adams, 2008; Vignjevic et al., 2006b). We also identified S274 as a second conserved phosphorylated site that lies within the C-terminal actin-binding site that contributes to actin bundling and plays an additional role in the control of microtubule dynamics (Villari et al., 2015; Zanet et al., 2012). Furthermore, we have shown that fascin is localised to the nuclear envelope (NE) where it binds directly to the nucleo-cytoplasmic linker protein nesprin-2 and regulates nuclear plasticity leading to enhanced cell invasion (Jayo et al., 2016). Additionally, we have demonstrated that fascin is located within the nucleus where it can bundle F-actin (Groen et al., 2015), but the role for nuclear-localised fascin and the potential impact on cancer cell behaviour remain unknown. Moreover, the general and distinct mechanisms that control these dynamic changes in fascin-binding partners, localisation, and function remain poorly understood. The nucleus contains high levels of actin, both monomeric and filamentous (Pederson, 2008), with actin filaments being shorter in length than those found in the cytoplasm. Our data have shown that fascin is required for endogenous nuclear actin bundles to form, and depletion of fascin from Drosophila nurse cells increases the size and number of nucleoli, suggesting a role in maintaining nuclear actin organisation and compartments (Groen et al., 2015). Several other actin-binding proteins, whose functions have been extensively investigated in the cytoplasm, are also present in the nucleus, including cofilin (suggested to be important for actin monomer import and accumulation in the nucleus) and profilin (partly responsible for actin nuclear export; Falahzadeh et al., 2015). Transient actin filaments, detected in nuclei upon serum stimulation or cell spreading, are reported to regulate MAL (MRTF or MKL1) transcription factor activation (Baarlink et al., 2013; Plessner et al., 2015). However, their effects on general transcription and chromatin remodelling are poorly understood. As high amounts of fascin are only present in certain stages of Drosophila follicle development and in growing human cancer cells (Groen et al., 2015), it is plausible that fascin import and export into and out of the nucleus is also tightly regulated depending on cell cycle, stress, or environmental influences. Thus, the identification of regulatory signals controlling dynamic movement of fascin within different subcellular compartments represents an important and unexplored area that may lead to new therapeutic avenues in cancer treatment. In the present study, we demonstrate that fascin is actively transported into the nucleus where it supports post-mitotic F-actin bundling, efficient DNA damage response (DDR), and cancer cell survival. Using high-content phenotypic imaging, we identify key upstream regulatory pathways that promote nuclear fascin and F-actin levels and demonstrate that forcing sustained nuclear fascin reduces cell invasion and promotes apoptosis. We further uncover roles for upstream kinases controlling Histone H3 phosphorylation in mediating fascin-dependent nuclear-DNA tethering and in controlling subsequent apoptotic responses. Our findings highlight the importance of dynamic nuclear fascin recruitment for tumour cell survival and provide new targets to explore for targeted cancer cell killing. Results Nuclear fascin contributes to efficient nuclear F-actin bundling We have previously reported that fascin is localised to the nucleus in a range of cell types in vitro and in vivo (Groen et al., 2015). We further verified localisation of endogenous fascin to the nucleus in two human cancer cell lines, HeLa and MDA-MB-231 (used throughout this study), and validated specificity of staining using stable fascin knockdown (KD) cells in each case (Figure 1—figure supplement 1A). We then analysed the sequence of fascin to determine whether any potential nuclear localisation (NLS) or nuclear export signal (NES) sequences existed that may mediate this nuclear transport. Our in silico analysis identified putative bipartite NLS and NES sequences (Figure 1—figure supplement 1B). We mutated these sequences and performed biochemical fractionation analysis that demonstrated a clear nuclear accumulation of the NES mutant but not the NLS mutant (Figure 1A; Figure 1—source data 1), indicating that fascin is actively transported into and out of the nucleus. Moreover, treatment of cells with leptomycin B, which inhibits nuclear export through blocking exportin 1, also resulted in accumulation of nuclear fascin (Figure 1—figure supplement 1C). It is notable that another group have identified the same NLS in fascin but suggested that removal of this sequence did not alter F-actin binding (Saad et al., 2016). In our hands, however, deletion or mutation of any of the NLS residues led to loss of F-actin binding (not shown), which is perhaps unsurprising given the very close proximity to the crucial S39 residue required for actin bundling. While this proximity to S39 is interesting in terms of fascin regulation, we chose not to use the NLS mutants for further functional analysis due to our concerns regarding loss of function. Figure 1 with 1 supplement see all Download asset Open asset Nuclear fascin contributes to F-actin bundling. (A) Representative western blot of fascin knockdown (KD) HeLa cells expressing specified GFP-fascin constructs subjected to biochemical fractionation. Nuclear and cytoplasmic compartments probed for GFP-fascin (80 kDa), Lamin A/C (69/62 kda) and GAPDH (36 kDa). Representative of three independent experiments. (B) Representative confocal images of nuclei of fascin KD HeLa cells co-expressing specified GFP-fascin constructs (green) and actin-NLS-FLAG construct, fixed and stained for FLAG (magenta) and F-actin (phalloidin). Scale bars are 10 µm. (C) Representative stills from time-lapse confocal movies of fascin KD HeLa cells co-expressing GFP or GFP-fascin (top panels) and iRFP-nAC nuclear F-actin probe (bottom panels) pre- and post-cytokinesis. Arrowheads point to dividing or daughter cells. Scale bars are 10 µm. (D) Quantification of duration of nuclear F-actin filaments in cells as in (C). (E) Organisation of nuclear F-actin in synchronised cells, 10 hr after release. For (D) and (E), N=89–100 cells/condition, pooled from three independent experiments. Graphs shows min/max and mean of dataset. ***=p < 0.001, ****=p < 0.0001. Figure 1—source data 1 Figure 1A full western blots. https://cdn.elifesciences.org/articles/79283/elife-79283-fig1-data1-v1.tiff Download elife-79283-fig1-data1-v1.tiff To investigate nuclear fascin effects on F-actin assembly, we depleted fascin from HeLa cells and stained with an N-terminal anti-actin antibody (AC15) that recognises nuclear actin (Miyamoto et al., 2011). Images demonstrated a striking loss of endogenous nuclear F-actin in fascin-depleted cells compared to controls (Figure 1—figure supplement 1D). We next sought to determine whether nuclear fascin induces nuclear F-actin bundles. Expression of actin fused to an NLS and FLAG epitope tag (actin-NLS) revealed long, crosslinked F-actin bundles decorated with endogenous fascin in control cells (Figure 1—figure supplement 1E). Moreover, GFP-fascin expressed in fascin KD HeLa cells also colocalised with nuclear actin-NLS-tagged filaments and assembly of these filaments was markedly reduced in fascin KD cells within the same field of view (Figure 1—figure supplement 1F). Further confirmation of the requirement for fascin-dependent nuclear F-actin assembly was demonstrated by data showing that S39A fascin, which constitutively bundles F-actin, could induce nuclear F-actin bundles, but the non-bundling mutant S39D could not (Figure 1B). Ratios of WT, S39A, and S39D fascin in nuclear vs. cytoplasmic fractions were very similar (Figure 1A), indicating the differences in nuclear F-actin assembly were not due to variation in nuclear abundance of these proteins. These data demonstrate that fascin is required for efficient nuclear F-actin bundling and this activity requires the N-terminal actin-binding site. We next assessed how fascin contributed to dynamic assembly of nuclear F-actin. Previous studies have shown nuclear F-actin assembles in nuclei in re-spreading, post-mitotic cells to support nuclear expansion in early G1 (Baarlink et al., 2017). To assess whether fascin play a role in this process, we expressed a nanobody that recognises F-actin tagged with both an NLS and iRFP (iRFP-nAC) in HeLa cells stably depleted of fascin and re-expressing GFP-fascin or GFP only. Live imaging of GFP-fascin cells identified that fascin transiently localised to the nucleus following cytokinesis. In agreement with previous reports, nuclear F-actin was also found to transiently assemble post-mitosis in these cells. However, F-actin assembly was visibly impaired in fascin-depleted cells (Figure 1C). Quantification confirmed a significant reduction in the duration of nuclear F-actin assembly post-mitosis in fascin KD cells (Figure 1D). To quantify F-actin structure, we developed an automated analysis pipeline that enables segmentation and analysis of the structure of nuclear F-actin marked by iRFP-nAC. Higher values from this automated analysis indicate more highly organised and contiguous nuclear F-actin; lower values indicate short or no nuclear actin filaments. Quantification of iRFP-nAC from post-mitotic cells revealed a significant reduction in nuclear F-actin organisation in fascin-depleted cells (Figure 1E). This data collectively demonstrates a requirement for nuclear fascin to support efficient nuclear F-actin assembly. Nuclear fascin directly associates with Histone H3 To explore potential nuclear fascin-binding partners other than F-actin, we adapted previously characterised fascin-specific GFP- or mCherry-tagged expressible nanobodies (Van Audenhove et al., 2014) by inserting NLS or NES sequences, which enabled movement of endogenous fascin to the nucleus or cytoplasm, respectively. GFP or mCherry alone and GFP- or mCherry-fascin nanobody (Nb2) were used as controls. Notably, Nb2 has previously been shown not to affect fascin function when expressed in cells (Van Audenhove et al., 2014). We verified successful expression and expected relocation of endogenous fascin in cells expressing NLS and NES mCherry-Nb2 plasmids, with no change seen in endogenous fascin localisation in controls (Figure 2A). Biochemical fractionation experiments further confirmed the enrichment of nuclear fascin in cells expressing Nb2-NLS (Figure 2B and C; Figure 2—source data 1). Cells expressing GFP-Nb2 plasmids were subjected to GFP-Trap to immunoprecipitate Nb2, and selected associated complexes were then identified by proteomic analysis. The resulting data were analysed and prioritised based on those hits with highest numbers of unique peptides identified in Nb2-NLS but not Nb2-NES complexes. Data revealed actin as the top hit across all Nb2 samples, with Histones H3 and H4 the most highly represented proteins in the Nb2-NLS samples only (Figure 2—source data 2). We chose to pursue Histone H3 for further validation to determine whether this protein is a nuclear-binding partner for fascin. Repeat GFP-Trap experiments verified the association of Histone H3 with Nb2-NLS but not Nb2-NES (Figure 2D; Figure 2—source data 3) and co-staining of cells for endogenous proteins revealed partial colocalisation within the nucleus (Figure 2E). To determine whether the fascin-Histone H3 association was direct, we co-expressed GFP-Histone H3 and mCherry-Nb2-NLS in cells and analysed binding using fluorescence resonance energy transfer (FRET). A range of FRET efficiencies were seen within a population of cells (Figure 2F and G), demonstrating that fascin and Histone H3 can directly associate and suggesting Histone H3 may act as a nuclear tether for fascin, contributing to fascin-dependent functions within this compartment. To further explore the relationship between fascin-F-actin bundling and fascin-Histone H3 binding, we co-expressed RFP-WT, S39A, and S39D fascin with GFP-Nb2-NLS followed by GFP-Trap to enrich for nuclear fascin species. Subsequent probing of these complexes revealed a loss of Histone H3 binding to S39A fascin compared to WT or S39D fascin, indicating that fascin-dependent F-actin bundling reduces the interaction of fascin with Histone H3 (Figure 2H; Figure 2—source data 4). Taken together these findings demonstrate that nuclear fascin binds to Histone H3 when not associated with F-actin and suggests a dynamic exchange of fascin within the nucleus that may act to coordinate nuclear organisation. Figure 2 Download asset Open asset Nuclear fascin directly associates with Histone H3. (A) Representative confocal images of fascin knockdown (KD) MDA-MB-231 cells expressing GFP-fascin (green) and specified mCherry-Nb2 constructs (magenta) fixed and stained with DAPI (blue). Scale bars are 10 µm. (B) Representative western blot of fascin KD MDA-MB-231 cells expressing GFP or GFP-fascin (80 kDa) subjected to biochemical fractionation. Nuclear and cytoplasmic compartments probed for GAPDH (36 kDa) and Lamin A/C (69/62 kDa) . (C) Quantification of data from (B) from four independent experiments. (D) Representative western blot of HeLa cells expressing GFP or GFP-Nb2-nuclear localisation signal (NLS)/nuclear export signal (NES), subjected to GFP-Trap and probed for Histone H3 (15 kDa), GFP (25 kDa and ~50 kDa Nbs). Input shown on right; GAPDH as loading control (36 kDa). Representative of four independent experiments. (E) Representative confocal images of HeLa nuclei fixed and stained for fascin (green) and Histone H3 (magenta). Scale bar is 10 µm. (F) Representative images of HeLa cells expressing GFP-Histone H3 (top panels) or GFP-Histone H3 and mCherry-Nb2-NLS (bottom panels). Donor and acceptor channels shown; lifetime shown in far-right panels. Scale bars are 10 µm. (G) Quantification of fluorescence resonance energy transfer (FRET) efficiency from data as in (F) plus mCherry and mCherry-Nb2-NES acceptor controls. N=35 cells, from four independent experiments. Graph shows min/max and mean of dataset; each point represents a single cell. (H) Representative western blot of HeLa cells expressing RFP-fascin WT, S39A or S39D (80 kDa), and GFP-Nb2-NLS (~50 kDa), subjected to GFP-Trap and probed for specified proteins (Histone H3 is 15 kDa). Input shown on right. Representative of four independent experiments. Figure 2—source data 1 Figure 2B full western blots. https://cdn.elifesciences.org/articles/79283/elife-79283-fig2-data1-v1.tiff Download elife-79283-fig2-data1-v1.tiff Figure 2—source data 2 Table detailing proteins identified by mass spectrometry associated with Nb2-NLS or Nb2-NES immunoprecipitated from HeLa cells. Peptide counts per identified protein shown for each protein/sample. Nuclear fascin (Nb2-NLS) enriched samples shown in tab 1. Full dataset shown in tab 2. Proteomics reporting details shown in tab 3. https://cdn.elifesciences.org/articles/79283/elife-79283-fig2-data2-v1.xlsx Download elife-79283-fig2-data2-v1.xlsx Figure 2—source data 3 Figure 2D full western blots. https://cdn.elifesciences.org/articles/79283/elife-79283-fig2-data3-v1.tiff Download elife-79283-fig2-data3-v1.tiff Figure 2—source data 4 Figure 2H full western blots. https://cdn.elifesciences.org/articles/79283/elife-79283-fig2-data4-v1.tiff Download elife-79283-fig2-data4-v1.tiff Nuclear fascin promotes efficient DDR Assembly of nuclear F-actin occurs following replication stress and induction of DNA damage, and acts under these settings to regulate nuclear volume, chromatin organisation, and oxidation status, driving efficient DNA repair (Belin et al., 2015; Lamm et al., 2020). To determine whether nuclear fascin contributed to these functions, fascin KD HeLa cells expressing GFP or GFP-fascin were treated with neocarzinostatin (NCS), an ionising radiation mimetic, to directly induce DNA double-strand breaks (DSB), and levels of the key DDR factor γH2AX were then analysed by immunostaining (Figure 3A). Quantification of these data revealed a significant reduction in the immediate recruitment of γH2AX to damaged DNA foci in fascin-depleted cells (GFP) compared to GFP-fascin rescued cells at 30 and 60 min post-NCS treatment (Figure 3B). Interestingly, γH2AX levels significantly reduced in fascin expressing cells 120 min post-NCS addition upon resolution of DSB; however, fascin KD cells remained at a significantly higher level indicating sustained DDR in these cells (Figure 3B). When treated with the common chemotherapy drug cisplatin, which also induces DSB, fascin KD cells also exhibited reduced γH2AX levels compared to cells expressing WT fascin (Figure 3C). The reduced initial response to NCS in fascin-depleted cells was further verified in biochemically fractionated cells indicating enhanced γH2AX in nuclear fractions of NCS-treated GFP-fascin cells compared to fascin KD cells (Figure 3D; Figure 3—source data 1). Further analysis of GFP-fascin rescued cells demonstrated an increase in nuclear fascin and F-actin upon NCS treatment (Figure 3E). As DDR both depend upon and trigger changes in chromatin compaction status, we next analysed this using Histone H2B-H2B FRET as previously characterised by others (Llères et al., 2009) in control and fascin KD cells treated with NCS. Lifetime images (Figure 3F) and FRET efficiency calculations (Figure 3G) revealed a significant increase in FRET efficiency in fascin expressing cells upon NCS treatment indicating enhanced chromatin compaction (Figure 3G). However, fascin KD cells showed basally enhanced chromatin compaction with no significant change upon NCS treatment (Figure 3G). These data demonstrate that fascin is required for chromatin organisation and efficient DDR. Figure 3 Download asset Open asset Nuclear fascin promotes efficient DNA damage response. (A) Representative confocal images of fascin knockdown (KD) HeLa cells expressing GFP or GFP-fascin (green) fixed and stained for ɣH2AX (magenta) before (0 min) or after (30 min) treatment with 0.5 µg/ml neocarzinostatin (NCS). Scale bars are 10 µm. (B) Quantification of ɣH2AX levels in cells from data as in (A). N=300–350 cells/condition, pooled from three independent experiments. (C) Quantification of ɣH2AX levels in cells treated with 5 µM cisplatin for 18 hr. N=240–300 cells/condition, pooled from three independent experiments. (D) Representative western blot of fascin KD HeLa cells expressing GFP (25 kDa) or GFP-fascin (FSN; 80 kDa) subjected to biochemical fractionation. Nuclear and cytoplasmic compartments probed for ɣH2AX (~15 kDa), Actin (42 kDa) and Lamic A/C (69/62 kDa). Representative of three independent experiments. (E) Representative confocal images of fascin KD HeLa cells expressing GFP-fascin (green) fixed and stained for ɣH2AX (cyan) and phalloidin (magenta) before (0 min) or after (60 min) treatment with NCS. Scale bars are 10 µm. (F) Representative images of WT or fascin KD HeLa cells expressing GFP-Histone H2B (left panels) or GFP-Histone H2B and mCherry-Histone H2B, with or without 30 min NCS treatment. Donor and acceptor channels shown; lifetime shown in bottom panels. Scale bars are 10 µm. (G) Quantification of fluorescence resonance energy transfer (FRET) efficiency from data as in (F). N=35 cells, pooled from three independent experiments. All graphs show min/max and mean of dataset; for figures in (B), (C), and (G), data is shown as violin plot with each point representing a single cell and mean shown in blue. **=p < 0.01, ***=p < 0.001, ****=p < 0.0001. Figure 3—source data 1 Figure 3D full western blots. https://cdn.elifesciences.org/articles/79283/elife-79283-fig3-data1-v1.tiff Download elife-79283-fig3-data1-v1.tiff Sustained nuclear fascin reduces cell invasion and viability To explore the functional consequences of enhanced and sustained nuclear fascin, we generated a doxycycline-inducible lentiviral version of the mCherry-tagged fascin-specific nanobodies coupled to NLS or NES sequences (as in Figure 2). This enabled precise triggering of expression of Nb2 in cells for sustained periods of time. We performed experiments in MDA-MB-231 cells as they display enhanced migratory and invasive capacity compared to HeLa and therefore represented a better model to assess these functions. Given that fascin is a canonical F-actin bundler known to promote filopodia formation, we evaluated filopodia assembly in live cells expressing Nb2 proteins for 48 hr. Filopodia were significantly reduced in Nb2-NLS expressing cells compared to controls, with no other changes seen in Nb2 or Nb2-NES expressing cells (Figure 4A and B). Similarly, cells expressing Nb2-NLS showed reduced migration speed in cells on 2D surfaces (Figure 4C) and reduced invasion into 3D collagen gels (Figure 4D and E; note higher nuclei numbers in 0 µm Z-section images of the Nb2-NLS samples indicating fewer cells able to invade) compared to control cells. Notably, Nb2-NLS expressing cells exhibited a similar reduction in invasion to fascin KD cells (Figure 4E), indicating that relocation of fascin to the nucleus prevents fascin function in the cytoplasm. Figure 4 Download asset Open asset Sustained nuclear fascin reduces cell invasion. (A) Representative confocal images of fascin knockdown (KD) MDA-MB-231 cells expressing GFP-fascin (green) and Nb2 constructs (magenta). Zoom region of filopodia shown below each. Scale bars are 10 µm. (B) Quantification of filopodia number/cell from data as in (A) from 30 cells, pooled from three independent experiments. (C) Quantification of 2D migration speed of cells as in (A) from 16 hr time-lapse movies. (D) Representative images of confocal Z-stacks from inverted invasion assays from cells as in (A), fixed after 48 hr invasion and stained for DAPI (shown). Fascin KD cells additionally shown (bottom row). (E) Quantification of invasion from data as in (D). Three wells imaged per experiment (3 fields of view/well), pooled from three independent experiments. All graphs show min/max and mean of dataset. **=p < 0.01, ***=p < 0.001, ****=p < 0.0001. To determine longer-term consequences of forced nuclear fascin on cancer cell viability, we assessed proliferation over 96 hr post-Nb2 expression induction. Proliferation was significantly reduced in Nb2-NLS expressing cells compared to controls from 48 hr time periods onwards (Figure 5A) and this was coupled with a significant reduction in cell viability (Figure 5B). Further analysis of DNA content 48 hr post-Nb2 induction revealed that a significant proportion of cells expressing Nb2-NLS accumulated in the G0/G1 phase of the cell cycle relative to controls. Nb2-NLS expressing cells also exhibited reduced levels of S10 phosphorylated Histone H3 after 48 hr (Figure 5D) and enhanced cleaved caspase 3 by 72 hr post-induction (Figure 5E; Figure 5—source data 1). These findings collectively indicate that enhanced and sustained nuclear fascin specifically induces cell cycle arrest resulting in reduced viability and enhanced apoptosis. Figure 5 Download asset Open asset Forced and sustained nuclear fascin reduces cancer cell viability. (A) Quantification of proliferation of MDA-MB-231 cells expressing specified mCherry-Nb2 constructs over 96 hr. Cell nuclei counted from three replicate wells per condition per experiment. Mean ± SEM are shown. Representative of three independent experiments. (B) Quantification of cell viability using MTT assays of MDA-MB-231 cells expressing specified mCherry-Nb2 constructs over 96 hr. Three replicate wells per condition per experiment. Min/max and mean are shown. Representative of three independent experiments. (C) Quantification of cell cycle stage of MDA-MB-231 cells expressing specified mCherry-Nb2 constructs after 48 hr using FACS analysis of PI-stained cells. Three samples per condition were analysed. Mean ± SEM are shown. Representative of three independent experiments. (D) Quantification of pS10-Histone H3 staining from confocal images of MDA-MB-231 cells expressing specified mCherry-Nb2 constructs after 48 hr. N=90 cells analysed per condition, pooled from three independent experiments. Graph shows min/max and mean of dataset. ***=p < 0.001, ****=p < 0.0001. (E) Representative western blot of MDA-MB-231 cells expressing specified mCherry-Nb2 constructs after 72 hr probed for cleaved caspase 3 (~32 kDa) and GAPDH (36 kDa). Staurosporine (1 µm) treated cells (24 hr) were used as a positive control. Figure 5—source data 1 Figure 5E full western blots. https:/