Mouse Shh-LIGHT2 cells (kind gift of A. McGee, Imperial College) (Taipale et al., 2000), Luzp1^–/– cells, + LUZP1 (Bozal-Basterra et al., 2020) and human HEK 293FT (Invitrogen), were cultured at 37°C and 5% CO₂ in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS, Gibco) and 1% penicillin/streptomycin (Gibco).

HEK 293FT cell LUZP1 locus was targeted by CRISPR-Cas9 to generate 293^LUZP1KO cells. Two high-scoring sgRNAs were selected¹ to target near the initiation codon (sg2: 5′-CTTAAATCGCAGGTGGCGGT_TGG-3′; sg3: 5′-CTTCAA TCTTCAGTACCCGC_TGG-3′). These sequences were cloned into px459 2.0 (Addgene #62988; kind gift of F. Zhang, MIT), for expressing both sgRNAs and Cas9 with puromycin selection. Transfections were performed in HEK 293FT cells with Lipofectamine 3000 (Thermo). Twenty-four hours after transfection, transient puromycin selection (0.5 μg/ml) was applied for 48 h to enrich for transfected cells. Cells were plated at clonal density, and well-isolated clones were picked and propagated individually. Loss-of-function mutations were confirmed by PCR-sequencing (Bozal-Basterra et al., 2020).

Cells were lysed in cold RIPA buffer (Cell Signaling Technology) supplemented with 1x protease inhibitor cocktail (Roche), and also in some cases with PhosphoStop 1x (Roche). Lysates were kept on ice for 30 min vortexing every 5 min and then cleared by centrifugation (25,000 × g, 20 min, 4°C). Supernatants were collected and protein content was quantified by BCA protein quantification assay (Pierce). After SDS-PAGE and transfer to nitrocellulose membranes, blocking in 5% milk, or in 5% BSA (Bovine Serum Albumin, Fraction V, Sigma) in PBT (1x PBS, 0.1% Tween-20) was performed. In general, primary antibodies were incubated overnight at 4°C and secondary antibodies for 1 h at room temperature (RT). Antibodies used: anti-vinculin (Sigma, 1:1,000), anti-cofilin and anti-phospho-cofilin (Cell Signaling Technology, 1:1,000), anti-Rac1 and anti-pRac1 (Cell Signaling Technology, 1:1,000), [anti-Actr3 (Machesky et al., 1997), 1:1,000], anti-GAPDH (Glyceraldehyde-3-Phosphate Dehydrogenase; Proteintech, 1:1,000) and anti-actin (Sigma, 1:1,000). Secondary antibodies were anti-mouse or anti-rabbit HRP-conjugates (Jackson Immunoresearch). Proteins were detected using Clarity ECL (BioRad) or Super Signal West Femto (Pierce). Quantification of bands was performed using ImageJ software and normalized against GAPDH or actin levels. At least three independent blots were quantified per experiment.

Shh-LIGHT2 cells and HEK 293FT cells were seeded on 11 mm coverslips (15,000–25,000 cells per well; 24-well plate). After washing once with cold 1x PBS, cells were fixed with methanol 100% for 10 min at −20°C or with 4% PFA supplemented with 0.1% Triton X-100 in PBS for 15 min at RT. Then, coverslips were washed 3 times with 1x PBS. Blocking was performed for 1 h at 37°C in blocking buffer (2% fetal calf serum, 1% BSA in 1x PBS). Primary antibodies were incubated overnight at 4°C and cells were washed with 1x PBS 3 times. We used antibodies anti-gamma-tubulin (Proteintech, 1:160), rabbit anti-Cleaved Caspase-3 (Cell Signaling Technology 9661S, 1:200) and anti-vinculin (Sigma hVIN-1, 1:200).

Donkey anti-mouse or anti-rabbit secondary antibodies (Jackson Immunoresearch) conjugated to Alexa 488 or Alexa 568 (1:200) and Alexa 568-conjugated phalloidin (Invitrogen 1:500), were incubated for 1 h at 37°C, followed by nuclear staining with DAPI (10 min, 300 ng/ml in PBS; Sigma). Fluorescence imaging was performed using an upright fluorescent microscope (Axioimager D1, Zeiss).

Cells were seeded in 96-well plates to be approximately 80% confluent the day of the EdU (5-ethynyl-2 deoxyuridine) labeling. Cell cultures were incubated 30 min with 10 μM EdU (Sigma, #900584) at 37°C in their own culture medium. Cells were washed and then fixed 15 min in 4% (w/v) formaldehyde solution in phosphate buffered saline (PBS). Following 3x PBS washes, fixed cells were permeabilized with 0.2% Triton X-100 in PBS for 30 min at room temperature. After washing with PBS, cells were directly incubated 30 min at room temperature (RT) in click reaction buffer. For 1 ml of click reaction buffer, following amounts of the different components are mixed in 680 μl milliQ water: 100 μl 1M Tris-HCl pH = 8; 20 μl 100 mM CuSO4; 0.5 μl Alexa Fluor^TM 647 Azide 1 μg/μl (AA648; Invitrogen) and 200 μl 0.5 M Ascorbic Acid. Finally, cells were washed 3x with PBS and incubated in 0.5 μg/ml DAPI-containing PBS for at least 30 minutes or until imaging.

For quantitative image-based cytometry images (QIBC), EdU labeled cells were obtained in an automated manner with the ScanR acquisition software controlling a motorized Olympus IX-83 wide-field microscope. Images from 3 technical replicates in 8 independent experiments were then processed using the ScanR image analysis software and analyzed with TIBCO Spotfire software.

To evaluate apoptosis, Shh-LIGHT2 cells were washed with 1x PBS and then stained with Annexin V (BD Biosciences) and DRAQ7 (Biostatus Ltd.). Data from 4 biological replicates were collected on a Fluorescence-activated Cell Sorting (FACS) Canto (BD Biosciences).

5 × 10³ cells were plated in triplicate in 12-well plates. Twenty-four hours later, the cells were considered day 0 (t₀) and were fixed in formalin 10% for 15 min. The same procedure was performed after 3 and 6 days. Cell viability was measured by staining with crystal violet (0.1% in 20% methanol) for 45 min at RT. After washing 3 times with water, all samples were air dried. The precipitate was solubilized in 10% acetic acid for 20 min at RT, and the absorbance was measured at 595 nm. For each timepoint, 4 biological replicates were measured.

Shh-LIGHT2 control, Luzp1^–/– mutant cells and + LUZP1 cells were grown in 24-well plates and a scratch wound was performed using a 20 μl pipette tip. Subsequently, medium was changed to remove detached cells. Pictures were taken at three different positions per sample in three technical replicates and at least eight biological replicates were analyzed of each. The scratch width was measured using ImageJ Fiji.

Filopodia were quantified by staining cells with Alexa 594-conjugated wheat germ agglutinin (WGA) and using FiloQuant, a plugin for the ImageJ software (Figure 2F and Supplementary Figure S1A; Jacquemet et al., 2017). The average number of filopodia of individual cells in 7 biological replicates was pooled together to perform statistical analysis.

Both WT and Luzp1^–/– Shh-LIGHT2 cells were suspended in DMEM medium plus 5% Methyl cellulose (Sigma) at 14,0000 cells/ml. Cell spheroids were subsequently formed by serial pipetting of 25 μl into the lid of a 10 cm dish (700 cells/spheroid) and incubated in an inverted position. After 48 h, cell spheroids were embedded into a volume of 300 μl of 2.3 mg/ml bovine collagen type I matrix (Advanced) and transferred to individual wells of a 24-well plate. Four hours later, each well was filled with complete media. Collective cell invasion was monitored using a Nikon Eclipse TS100 Live Imaging microscope. Images were taken just after adding complete medium to the collagen-embedded cells (t = 0 h) and 18 h later (t = 18 h). The area of each individual spheroid was measured in 3 technical replicates and in at least 3 biological replicates using ImageJ analysis program. The fold change in invasive area was determined by dividing the final area (at 18 h) by the initial area (at 0 h) in each cell type.

For transwell assay (24-well format), 2 × 10⁴ WT, Luzp1^–/– and + LUZP1 Shh-LIGHT2 cells were seeded in the upper chamber in serum-free medium (0.5 ml; inserts 6.5 mm, 8 μm pore size; Corning Costar). The lower chamber was loaded with 1 ml medium supplemented with 10% FBS. After 6 h of incubation at 37°C with 5% CO₂, the migrated cells in the membrane were stained by DAPI. Images were obtained in an automated manner with the ScanR acquisition software controlling a motorized Olympus IX-83 wide-field microscope. Images from 6 independent experiments were then processed using the ScanR image analysis software.

Patient copy number and mRNA expression information was obtained from cBioPortal (Cerami et al., 2012; Gao et al., 2013), Cancertool (Cortazar et al., 2018), and TCGA Copy Number Portal².

Statistical analysis was performed using GraphPad 6.0 software. Data were analyzed by Shapiro-Wilk normality test and Levene’s test of variance. We used two tailed unpaired Student’s t-test or Mann Whitney-U tests for comparing two groups, One-way ANOVA or Kruskall-Wallis and the corresponding post hoc tests for more than two groups and two-way ANOVA to compare more than one variable in more than two groups. P-values were represented by asterisks as follows: ^∗P < 0.05; ^∗∗P < 0.01; ^∗∗∗P < 0.001; ^****P < 0.0001. Differences were considered significant when P < 0.05.

The process of cellular transformation from normal to malignant cellular behavior derives from the acquisition of genomic aberrations and the development of cancer hallmarks, such as activation of invasive and migratory phenotypes and resistance to apoptosis (Hanahan and Weinberg, 2011). We first aimed to characterize the genomic alterations of LUZP1 reported in various cancer types using a publicly available webtool (cBioPortal) (Cerami et al., 2012; Gao et al., 2013). Interestingly, we found genomic aberrations in LUZP1 at high frequency in cancer specimens, reaching an alteration frequency of almost 80% in cholangiocarcinoma and greater than 40% in 13 out of the 54 cancer types that we analyzed (Figure 1A). Importantly, shallow deletions (most possibly heterozygous deletions according to cBioPortal) were the most prevalent copy number alterations event in a large fraction of cancers (Figure 1A, aquamarine). We next evaluated whether copy number aberrations would influence LUZP1 gene expression across several cancer types based on the TGCA Pancancer Atlas database (cBioPortal). As predicted, shallow and deep deletions (most probably homozygous deletions, according to cBioPortal) were associated to lower LUZP1 mRNA expression compared to diploid, gain or amplification events in cholangiocarcinoma (Figure 1B), breast cancer (Figure 1C) and prostate cancer (Figure 1D), indicating that genomic events are, at least in part, responsible for reduced LUZP1 expression associated to cancer. In line with the genomic analysis and the gene dosage-mRNA expression association, using Cancertool (Cortazar et al., 2018) we found the mRNA levels of LUZP1 to be significantly reduced in prostate cancer (Figures 1F,G), and non-significantly in breast cancer (Figure 1E). From our analysis, LUZP1 emerges as a potential cancer-relevant gene, exhibiting genomic and transcriptional perturbation pattern suggestive of a tumor suppressive function.

Based on the genetic alterations of LUZP1 in cancer specimens as mentioned above, we analyzed intrinsic cellular features altered in cancer, such as cell migration and invasion in vitro, to see how they might be affected with loss of Luzp1. Using CRISPR/Cas9 gene editing directed to exon 1 of murine Luzp1, we previously generated Shh-LIGHT2 mouse embryonic fibroblasts (Taipale et al., 2000) null for Luzp1 (Luzp1^–/– cells), and additionally rescued the same cells by expression of human LUZP1-YFP fusion (+ LUZP1 cells) (Bozal-Basterra et al., 2020). Interestingly, Luzp1^–/– cells elicited a remarkable increase in migratory capacity compared to WT cells, as shown by wound healing assays (Figures 2A,B). This phenotype was suppressed in + LUZP1 rescue cells. We further confirmed the heightened migratory capacity of Luzp1^–/– cells by Boyden chamber or transwell assay (Figures 2C,D).

To migrate, a cell must coordinate a number of different inputs into appropriate cellular responses. Vinculin and phalloidin staining was performed in migrating cells to examine focal adhesions and visualize cell shape, respectively. We observed that Luzp1^–/– cells form more lamellipodia than WT at the leading edge of migrating cells (Figure 2E). Filopodia and filopodia-like protrusions are prominent features of migrating cells in vitro (Petrie and Yamada, 2012; Jacquemet et al., 2013; Paul et al., 2015). In concordance with increased migration, Luzp1^–/– cells displayed more filopodia than WT cells (Figure 2F and Supplementary Figure S1A). To further characterize the regulation of invasive properties by LUZP1, we generated spheroids to measure the invasive growth into 3D collagen matrix. The results show that the Luzp1^–/– cells showed higher invasive capacity than WT cells (Figures 2G,H). Taken together, these data revealed that Luzp1 loss leads to an increase in cell migration and invasion.

An increase in cell migration and invasion could be influenced by a differential rate of cell proliferation in Luzp1^–/– cells. To check this, we analyzed cell viability in WT, Luzp1^–/– and + LUZP1 cells using crystal violet assay. Surprisingly, Luzp1^–/– cells exhibited a significant reduction in cell numbers at day 3 and 6 after seeding compared to WT cells (Figure 3A). + LUZP1 cells partially rescued the cell viability impairment (Figure 3A, green line). In addition, EdU labeling and QIBC analysis revealed significant changes in cell cycle (Figure 3B), showing fewer Luzp1^–/– cells in G0/G1 phases and more Luzp1^–/– cells in S phase compared to WT. To determine the level of apoptosis occurring within each population, we performed FACS analysis of cells co-stained for Annexin V and DRAQ7. We detected more apoptotic cells among the Luzp1^–/– cell population (Figure 3C and Supplementary Figure S1B). In agreement with these results, we performed immunofluorescence staining for cleaved caspase-3 (CC3), a marker of mid-stage apoptosis. More apoptotic cells were detected among the Luzp1^–/– cell population compared to WT cells (Figures 3D,E). Our results showed a global reduction in cell viability rate. The overall increase in proliferation in Luzp1^–/– cells by EdU labeling might be counterbalanced by an increase in apoptosis in Luzp1^–/– cells compared to WT. Whether or not these phenotypes are independently linked to LUZP1 function, or one is a consequence of the other, remains to be determined.

The regulation of cell division, migration and invasion is influenced by the size and number of intracellular structures (Ogden et al., 2013; Bell and Lammerding, 2016). We noted a striking alteration in nuclear size, which was significantly increased in Luzp1^–/– cells and suppressed in + LUZP1 rescue cells (Figures 4A,B). As LUZP1 localizes to centrosomes and plays a role in ciliogenesis, (Bozal-Basterra et al., 2020; Goncalves et al., 2020), we examined centrosomes from Luzp1^–/– cells throughout the cell cycle and noticed an additional phenotype: Luzp1 depletion resulted in the heightened incidence of cells with more than four centrioles (Figure 4C). This phenotype was suppressed in + LUZP1 rescue cells. The presence of multiple centrioles was further verified in 293^LUZP1–KO cells (Figure 4D). Taken together, these results point to a regulatory role for LUZP1 in centrosome duplication, cytokinesis, or both, resulting in larger nuclei and multiple centrioles.

Based on the association of LUZP1 to F-actin and the reduction of actin filaments that we had previously observed in Luzp1^–/– cells (Bozal-Basterra et al., 2020), as well as the reported interaction of LUZP1 with Arp2 (Hein et al., 2015; Goncalves et al., 2020), we wondered whether other regulators of actin polymerization might be altered. Using western blot, we observed a decrease in total ACTR3 levels (Figure 5A,B), that was accompanied by an increase in phosphorylated cofilin in Luzp1^–/– cells (Figures 5C,D). Activation of Rac1 GTPase activity leads to changes in actin polymerization mediated by cofilin and the ARP2/3 complex, but in terms of Rac1 activation by phosphorylation (pRAC1:RAC1 ratio), we did not find significant differences between Luzp1^–/– and WT cells (Figures 5E,F). Taken together, these data point to multiple roles for Luzp1 in actin cytoskeleton dynamics.

Although actin stress fibers contribute to cell shape and adhesion, their exact role in cell migration/invasion has been debated. Stress fibers are absent from several highly motile cells, such as leukocytes (Valerius et al., 1981) and amoeba of Dictyostelium discoideum (Rubino et al., 1984). These observations, together with the relative lack of stress fibers in cells grown in three-dimensional matrices have led to the suggestion that they are not essential for cell migration (Burridge et al., 1988). Indeed, it is possible that, under certain conditions, stress fibers might inhibit motility (Badley et al., 1980; Herman et al., 1981; Cramer et al., 1997; Kemp and Brieher, 2018). These studies match our observations that Luzp1^–/– cells, which contain less stress fibers than WT cells (Bozal-Basterra et al., 2020), are more motile than WT.

Many cellular proteins are involved in the tight regulation of actin assembly, which directly influences cell migration and invasion. Actin-related proteins act synergistically to maintain a pool of unpolymerized actin monomers, nucleate, elongate and cap actin filaments, promote dissociation of Pi from ADP-Pi-subunits, sever actin filaments, and crosslink filaments into higher order structures. The Rho GTPases have been studied in association with their roles in the regulation of cell division, migration and invasion, mainly via actin filament organization (Aspenstrom et al., 2004). For instance, RhoD (Ras Homolog Family Member D) binds FILIP1, which binds FLNA and has a crucial function in cell migration in the brain (Nagano et al., 2004; Gad et al., 2012). Interestingly, the domain of FILIP1 that binds RhoD (nucleotides 431–778) has more than 40% homology with the equivalent region of LUZP1. Furthermore, a third related protein, FILIP1L, also inhibits tumor cell migration and invasion in colorectal cancer models (Park et al., 2016). Whether the roles of LUZP1, FILIP1, and FILIP1L are distinct or redundant remains undefined. Future mechanistic studies on how LUZP1 and related proteins regulate actin dynamics will be necessary to understand their roles in these diverse actin-driven cellular processes.

During mitosis the actin cytoskeleton must rearrange and localize to the contractile ring during cytokinesis (Heng and Koh, 2010). This recruitment of F-actin and actin regulatory proteins to the cell cortex during mitosis is essential for the interaction between astral microtubules and cortical actin, which is believed to be important in regulating mitotic spindle orientation (Rankin and Wordeman, 2010). The role of actin and its regulatory proteins in these processes ranges from regulating centrosome separation to proper spindle assembly and orientation and to elongate kinetochore microtubules (Firat-Karalar and Welch, 2011). We observed that Luzp1^–/– cells exhibited multiple centrioles, bigger nuclei, decreased cell viability and increased apoptosis. The increased apoptosis seen in our study matches the elevated apoptosis reported in the neuroepithelium of the NTD Luzp1 KO mouse hindbrain, which displays NTD (Hsu et al., 2008). Whether some or all of these phenotypes are due to the reduced actin cytoskeleton seen in Luzp1^–/– cells, or due to actin-independent roles for LUZP1, remains to be determined.

Moreover, it was previously found that LUZP1 localizes not only to centrioles and actin cytoskeleton, but also to the midbody in dividing cells (Bozal-Basterra et al., 2020; Goncalves et al., 2020). The midbody is formed at the intercellular bridge in the last phase of cytokinesis and contains crucial proteins for the abscission between the dividing cells (D’Avino and Capalbo, 2016). While one explanation for the multiple centrioles in Luzp1^–/– cells could be failed cytokinesis, we did not observe an increase in cells with multiple nuclei, and QIBC analysis did not reveal increased polyploidy in cells, even though nuclei were increased in size. The timing of abscission can influence karyoplasmic ratios, which are thought to be linked to metastatic properties (Rizzotto and Schirmer, 2017). Understanding the role of LUZP1 in the midbody, namely in timing and dynamics of contractile ring formation and contraction, as well as abscission, would be of major interest.

Actin exists as monomers (G-actin) and filamentous polymers (F-actin) and the maintenance of the balance between G-actin vs. F-actin is important for physiological functions including cell locomotion, cytokinesis, maintenance of cell shape and muscle contraction (Stricker et al., 2010). Two important actors control the polymerization and depolymerization of the actin filaments in cells: the ARP2/3 complex and cofilin. They work synergistically in such a way that the newly polymerized filaments from cofilin-generated barbed ends are ATP-rich filaments that promote the nucleation and branching activity of the ARP2/3 complex (DesMarais et al., 2004). Biochemical studies have shown that LUZP1 is an actin cross-linking protein (Wang and Nakamura, 2019a). As it has been proposed for EPLIN (a LUZP1 interactor) (Song et al., 2002; Maul et al., 2003), one hypothesis is that LUZP1 and EPLIN could regulate actin polymerization by influencing both assembly (nucleation, especially of ARP2/3 branched structures) and disassembly (stability) of F-actin, but this awaits detailed biochemical studies. We observed a reduction in ACTR3 in Luzp1^–/– cells compared to WT cells, suggesting that nucleation of branched actin might be diminished. In contrast, we observed an increase in phosphorylated (or inactive) cofilin levels, suggesting that Rho GTPase signaling might be activated upon loss of Luzp1. While normally this leads to increased stress fibers (Ridley and Hall, 1992), if the actin filaments are unable to become bundled due to absence of Luzp1, then phalloidin-positive stress fibers will be reduced, as we observed in Luzp1^–/– cells. As suggested by others (Liu et al., 2007), the elevated cofilin phosphorylation levels in Luzp1^–/– cells may reflect a compensatory response to weakened stress fibers.

In this work, we demonstrate that heterozygous loss of LUZP1 is frequent in different cancer types. As observed in Luzp1^–/– cells, a switch from proliferation to migration/invasion is a common event in the context of cancer (Mehlen and Puisieux, 2006). However, the increase in apoptosis was puzzling. One possibility is that the cytoskeletal alterations and the multiple centrioles could lead to genomic instability. This phenomenon would be deleterious in benign cells, but tolerated in cancer cells. The heterozygous loss of LUZP1 observed in tumor samples could be the consequence of the balance between the advantage of decreasing LUZP1 levels enough to promote invasion and the counterselection of complete loss to avoid genomic instability or cell division defects leading to apoptosis. Another perspective would be that complete LUZP1 loss might be counterselected in cancer due to the antiproliferative and proapoptotic effects, and this explains the frequency of heterozygous losses.

In summary, our study demonstrates that LUZP1 controls proliferative and invasive features in cancer, thus providing a feasible explanation for its frequent copy number aberrations in various cancer types.