Lamin A and Prelamin A Counteract Migration of Osteosarcoma Cells

A type lamins are fundamental components of the nuclear lamina. Changes in lamin A expression correlate with malignant transformation in several cancers. However, the role of lamin A has not been explored in osteosarcoma (OS). Here, we wanted to investigate the role of lamin A in normal osteoblasts (OBs) and OS cells. Thus, we studied the expression of lamin A/C in OS cells compared to OBs and evaluated the effects of lamin A overexpression in OS cell lines. We show that, while lamin A expression increases during osteoblast differentiation, all examined OS cell lines express lower lamin A levels relative to differentiated OBs. The condition of low LMNA expression confers to OS cells a significant increase in migration potential, while overexpression of lamin A reduces migration ability of OS cells. Moreover, overexpression of unprocessable prelamin A also reduces cell migration. In agreement with the latter finding, OS cells which accumulate the highest prelamin A levels upon inhibition of lamin A maturation by statins, had significantly reduced migration ability. Importantly, OS cells subjected to statin treatment underwent apoptotic cell death in a RAS-independent, lamin A-dependent manner. Our results show that pro-apoptotic effects of statins and statin inhibitory effect on OS cell migration are comparable to those obtained by prelamin A accumulation and further suggest that modulation of lamin A expression and post-translational processing can be a tool to decrease migration potential in OS cells.


Introduction
Osteosarcoma (OS) is the most common primary bone tumor in children and adolescents and therefore has an important social impact despite its rarity [1]. OS displays a high degree of aggressiveness and tendency to metastasize [2]. Surgical resection combined with chemotherapy is the most effective therapeutic strategy against OS [3] and this multidisciplinary approach has improved Adipose-derived Stem Cells (ADSCs) were purchased from Thermo Fisher Scientific and cultured in Complete MesenPro RS medium (Gibco, Thermo Fisher Scientific, Waltham, MA, USA), in a 37 • C incubator with 5% CO 2 , as described in the manufacturer's protocol. ADSCs were differentiated toward the osteogenic lineage by seeding 5 × 10 3 cells/cm 2 and replacing the culture medium after 2 days with Complete STEMPRO Osteogenesis Differentiation Medium (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) [27].
Primary osteoblasts isolated from healthy donors were obtained from BioLaM biobank (IOR CE approval 0018259-01-13, issued on 05/09/2016) and maintained under standard cell culture conditions of 5% CO 2 and 95% humidity and were used at passages 3-5. For differentiation experiments, primary osteoblasts were cultured with osteogenic medium containing 10% FBS, 1% penicillin and streptomycin, 50 µg/mL of ascorbic acid, and 10 mM β-glycerophosphate. The medium was changed every 3 days.
For Alizarin Red Staining, cells were fixed with neutral buffered formalin (10%) for 10 min. The fixed cells were rinsed with PBS and stained with 0.5% Alizarin Red staining solution (Sigma-Aldrich, St. Louis, MO, USA) (pH 4.2) for 10 min at room temperature. After being washed with ddH 2 O on a shaking platform, the mineralized nodules were imaged by a scanner.
LMNA silencing in MG63 cell line, was done by employing siRNAs duplexes specific for human LMNA (siRNA LMNA) purchased from Thermo Fisher Scientific (Waltham, MA, USA). Scrambled duplexes were used as control (silencer select negative control: 4390843). An amount of 0.15 × 10 6 MG63 cells was plated in 6-well plates. At approximately 50% confluence, cells were transfected with the annealed siRNA-LMNA and siRNA-scrambled (siRNA final concentration 10 nM) using lipofectamine 3000. The experiments were performed after 48 h from transfection.
Overexpression of unprocessable prelamin A (LA-C661M) was obtained as reported elsewhere [28]. Biochemical analyses were performed 24 h after transfection. Mevinolin (Sigma Aldrich, Saint Louis, MO, USA), an inhibitor of the hydroxymethyl-glutaryl-synthase that impairs farnesylation of prelamin A, was added to cell culture medium at 5 or 10 µM and cells were harvested after 24, 48 or 72 h.

Cell Proliferation
To assess cellular growth, cells were seeded in 6-wells in DMEM plus 10% FBS. After 24 h, increasing concentrations of mevinolin were added, and the cells were exposed to the drug for up to 72 h. Growth curves were generated by plotting direct count of cells. Viable cells were counted by a hemocytometer using 0.2% Trypan Blue.

Wound Healing Assay
Cells were seeded in 6-well plates and the wound healing assay was performed. A reproducible longitudinal scratch in the monolayer was made the following day using sterile micropipette tips. The process of wound closure was monitored at 0 and 24 h by photographing the central field of the scratches under an inverted light microscopy (Olympus CKX41, Olympus Corp, Tokyo, Japan) mounted with a digital camera (C-7070 Wide Zoom, Olympus) at 10× magnification. The pictured field was standardized each time against a horizontal line drawn on the base of the plate passing through the center of each well. Morphometric analysis of cell migration was performed using a Cells 2020, 9, 774 4 of 15 computerized image analysis system (Qwin, 3.0 software, Leica Microsystem Imaging Solution, Ltd., Wetzlar, Germany). A region that included the artificial scratch and the adjacent cell monolayer was selected as the standard region of interest (ROI). The wound healing effect was calculated as (1-Ax/A0) %, where A0 and Ax represented the empty scratch area at 0 and 24 h, respectively.

Annexin V-FITC/PI Staining
Apoptosis analysis was determined through the Annexin V-FITC Apoptosis Detection kit (eBioscience, Thermo Fisher Scientific, Waltham, MA, USA), according to the manufacturer's instructions, using an FC500 flow cytometer (Beckman Coulter, Brea, CA, USA) equipped with the appropriate software (version 2.2, CXP, Beckman Coulter). At least 10.000 events were acquired.

qRT-PCR
Total RNA was extracted using the RNeasy Mini Kit (Qiagen, Venlo, The Netherlands), according to the manufacturer's instructions, and 1 µg of total RNA was reverse transcribed using High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific, Waltham, MA, USA). Gene expression was assessed using the TaqMan ® Gene Expression Master Mix, using the 7300 RT-PCR system (Applied Biosystems, Foster City, CA, USA). Results were normalized to the level of the ubiquitously expressed RNA 18S ribosomal 1 gene (RN18S), and were expressed as 2 −∆Ct (∆Ct = CT gene of interest − CT internal control) to compare the relative gene expression among samples, and as 2 −∆∆Ct (∆∆Ct = ((CT gene of interest − CT internal control) sample − (CT gene of interest − CT internal control) universal)) to compare gene expression of the treated cell lines with that of untreated control [29]. The alkaline phosphatase (ALPL) probe (Hs.PT.56a.40555206) was from Roche Diagnostics (Basel, Switzerland), and the LMNA probe (TaqMan Hs.01108900_g1) was from Thermo Fisher Scientific (Waltham, MA, USA). The RUNX2 (Taqman Hs.00231692_m1) probe was from Thermo Fisher Scientific (Waltham, MA, USA).

Immunofluorescence Microscopy
To determine the subcellular localization of lamin A/C and prelamin A, OS cells were seeded in 12-well plates and the cells were grown on coverslips, fixed with methanol for 8 min at −20 • C. Cells were then blocked with 3% BSA-containing PBS for 1 h. Antibodies diluted in 3% BSA-containing PBS were applied overnight at 4 • C, and revealed by using FITC-conjugated secondary antibodies diluted 1:200 (incubated for 1 h at RT). Samples were mounted with a DAPI-containing anti-fade reagent (Molecular Probes, Thermo Fisher Scientific, Waltham, MA, USA) and observed with a Nikon Eclipse Ni epifluorescence microscope. The images captured with NIS-Elements 4.3 AR software were processed using Photoshop CS4 (Adobe Systems, Inc., San Jose, CA, USA). The following antibodies were used: anti-prelamin A (Merck Millipore, Burlington, MASS, USA), diluted 1:500, anti-lamin A/C (#4777 Cell Signaling Tecnhology, Danvers, MA, USA), diluted 1:100.

Statistical Analysis
Statistical analyses were performed using Student's t test for LMNA overexpression/silencing experiments, for cell growth and apoptosis studies, and for migration experiments, while one-way ANOVA and Dunnett's multiple comparison test was employed to evaluate differences in the expression levels of proteins or genes analyzed. Correlations were established according to Spearman's rank test. At least three different experiments were performed and data plotted represent mean ± SD (standard deviation) with a significance level of *p < 0.05, **p < 0.01, ***p < 0.001 (GraphPad Prism Software, Version 5, San Diego, CA, USA).

Lamin A/C Expression Increases in Differentiating Osteoblasts
To evaluate lamin A/C expression in differentiating osteoblasts from healthy donors, we cultured cells in differentiation medium up to 21 days. Morphological changes and formation of hydroxyapatite crystals in cell cultures were observed starting from day 7 of culture in differentiation medium ( Figure 1a). Osteoblast differentiation was confirmed by alizarin red staining (Figure 1b), as well as by expression levels of ALPL and RUNX2 genes, well-established osteoblast differentiation markers, which were strongly upregulated, with a peak at day 7 and day 14, respectively ( Figure 1c). The LMNA gene was significantly upregulated during osteoblast differentiation ( Figure 1d). Moreover, lamin A/C protein levels were significantly increased in osteoblasts at 21 days of differentiation, relative to undifferentiated cells (nd) (Figure 1e).
These data were confirmed in adipose tissue-derived pluripotent stem cells (ADSCs), which can differentiate into different cell types, including osteoblasts [27,30]. Figure 1f shows that ALPL gene expression significantly increased in a time-dependent manner during ADSC osteogenic differentiation. Moreover, protein levels of osteogenic differentiation markers Osteoblast-specific Transcription Factor Osterix (OSX) and Osteocalcin were significantly increased in ADSCs at 14 and 21 days in differentiating medium ( Figure 1g). During ADSC osteogenic differentiation, LMNA gene expression was significantly increased (Figure 1h). Moreover, the level of lamin A/C protein was decreased in ADSC cells at 7 days in osteogenic differentiation medium and increased at day 14 and 21 ( Figure 1i).

Lamin A/C Expression is Reduced in Osteosarcoma Cells
We evaluated lamin A/C protein levels in several OS cell lines by western blot analysis. Relative to OBs from healthy donors differentiated for 21 days, all OS cell lines displayed significantly lower lamin A expression levels (Figure 2a,b).
However, as compared to other OS cell lines, MG63 showed higher protein levels of lamin A (Figure 2a,b). The lowest lamin A/C levels were determined in HOS cells, and in the 143B Ras-mutated HOS-derived cell line [31,32] (Figure 2a,b). These data were also confirmed in other OS cell lines obtained in our laboratory (Supplementary Figure S1).
We have also evaluated lamin A/C localization and we found that lamin A/C localized at the nuclear periphery ( Figure 2c). Morphological abnormalities were observed both in osteoblasts and Statistical analyses were performed using Student's t test and asterisks indicate statistically significant differences with respect to nd, **p < 0.01, ***p < 0.001. (d) LMNA gene expression analysis by qRT-PCR in nd versus differentiated osteoblasts (OBs). Asterisks indicate statistically significant differences with respect to nd, *p < 0.05, ***p < 0.001 (e) Western blotting analysis of lamin A/C protein expression in nd and differentiated OBs. Densitometric analysis is shown as mean values ± SD of three different experiments. Statistical analyses were performed using Student's t test and asterisks indicate statistically significant differences with respect to nd, ***p < 0.001. (f) Increased ALPL gene expression by qRT-PCR analysis. Statistical analyses were performed using Student's t test and asterisks indicate statistically significant differences with respect to non-differentiated adipose-derived stem cells (ADSCs) (nd), **p < 0.01, ***p < 0.001. (g) Protein expression analysis of Osterix (OSX) and Osteocalcin by western blotting. Densitometric analysis is reported, showing significantly increased levels of OSX and Osteocalcin compared to non-differentiated ADSCs (nd). Statistical analyses were performed using Student's t test. Asterisks indicate statistically significant differences with respect to nondifferentiated ADSCs (nd), **p < 0.01, ***p < 0.001. (h) LMNA gene expression analysis by qRT-PCR in ADSCs versus differentiated OBs. Asterisks indicate statistically significant differences with respect to non-differentiated ADSCs (nd), **p < 0.01. (i) Western blotting analysis of lamin A/C protein expression in non-differentiated ADSCs (nd) and differentiated OBs. Densitometric analysis is . Asterisks indicate statistically significant differences with respect to nd, *p < 0.05, ***p < 0.001 (e) Western blotting analysis of lamin A/C protein expression in nd and differentiated OBs. Densitometric analysis is shown as mean values ± SD of three different experiments. Statistical analyses were performed using Student's t test and asterisks indicate statistically significant differences with respect to nd, ***p < 0.001. (f) Increased ALPL gene expression by qRT-PCR analysis. Statistical analyses were performed using Student's t test and asterisks indicate statistically significant differences with respect to non-differentiated adipose-derived stem cells (ADSCs) (nd), **p < 0.01, ***p < 0.001. (g) Protein expression analysis of Osterix (OSX) and Osteocalcin by western blotting. Densitometric analysis is reported, showing significantly increased levels of OSX and Osteocalcin compared to non-differentiated ADSCs (nd). Statistical analyses were performed using Student's t test.
to OBs from healthy donors differentiated for 21 days, all OS cell lines displayed significantly lower lamin A expression levels (Figure 2a,b). However, as compared to other OS cell lines, MG63 showed higher protein levels of lamin A (Figure 2a,b). The lowest lamin A/C levels were determined in HOS cells, and in the 143B Rasmutated HOS-derived cell line [31,32] (Figure 2a,b). These data were also confirmed in other OS cell lines obtained in our laboratory (Supplementary Figure S1). We have also evaluated lamin A/C localization and we found that lamin A/C localized at the nuclear periphery ( Figure 2c). Morphological abnormalities were observed both in osteoblasts and osteosarcoma cells, although honeycomb structures, which are typical of some laminopathic cells, were only detected in OS nuclei (Figure 2c,f).

Migration and Proliferation Ability are Increased in OS Cells with Low LMNA Expression
To gain insight into the role of lamin A in OS cells, we used 143B and HOS cell lines, which have the lowest lamin A/C levels and we expressed a lamin A/GFP fusion protein. After 24 h of transfection, 143B and HOS cells expressed lamin A/GFP (Figure 3a).
Then, we wanted to analyze the effect of lamin A overexpression on cellular proliferation. Empty GFP and lamin A/GFP-transfected cells were seeded at the same density and cell growth was assessed by trypan blue exclusion assay and cell count. Overexpression of lamin A clearly slowed down proliferation of 143B cells 24 and 48 h after transfection with lamin A/GFP vector (Figure 3b).
Given that OS is a highly metastatic tumor and lamin A is involved in cellular migration [11], we tested migration potential of lamin A-overexpressing cells. Lamin A/GFP-transfected cells displayed a significantly lower ability to migrate in a wound healing assay as compared to cells only expressing GFP (Figure 3c). These results indicated that lamin A reduces the migratory ability in OS cells. To confirm the role of lamin A in cellular migration, we silenced the LMNA gene in MG63 cells and performed a wound healing assay, demonstrating that the silenced cells migrated faster compared to scramble (Figure 3d,e).
Based on the latter results, we performed a wound healing assay on five OS cell lines. 143B and HOS cells were able to almost completely repair the wound after 24 h. Of note, these cells expressed the lowest lamin A/C levels among tumor cell lines here examined. On the contrary, MG63 cells, expressing the highest lamin A amount among OS cell lines, failed to totally repair the wound in the same time frame (Figure 3f,g). Thus, OS cell lines with the highest lamin A levels showed significantly reduced wound healing ability, suggesting an inverse correlation between lamin A/C protein amount and cellular migration. We performed a Spearman's Rank correlation test, obtaining an r value = −0.7 (p < 0.05). This confirmed the existence of an inverse correlation between lamin A expression levels and cellular migration ability in the OS samples analyzed.  (Figure 3a). Then, we wanted to analyze the effect of lamin A overexpression on cellular proliferation. Empty GFP and lamin A/GFP-transfected cells were seeded at the same density and cell growth was assessed

Prelamin A Accumulation Reduces Migration Potential of OS Cells
We then wondered if an unbalance in lamin A maturation could reduce migration potential of OS cells. To test this hypothesis, we expressed an unprocessable prelamin A sequence (LA-C661M), which causes accumulation of non-farnesylated prelamin A, in 143B and HOS cells, which are the OS cell lines with the lowest lamin A expression. Western blotting analysis showed prelamin A accumulation in both cell lines (Figure 4a). Accumulation of prelamin A induced a strong decrease in migration ability in 143B and HOS cells (Figure 4b).

Mevinolin Effects in OS Cells Are Related to Prelamin A Accumulation
Concentration-dependent viability was assessed upon mevinolin treatment in OS cells. After 24 or 48 h of treatment, only 143B and HOS cells showed a significant decrease in cell viability (around 25% when mevinolin was employed at 10 μM concentration) (Figure 5a). Mevinolin is a drug known to reduce production of farnesyl-pyrophosphate. As a consequence, mevinolin blocks prelamin A farnesylation, which is required for further protein processing, thus impairing the whole process of lamin A maturation and causing accumulation of non-farnesylated prelamin A [33,34]. OS cell lines were treated with mevinolin and accumulation of prelamin A was obtained, as shown by western blot analysis (Figure 4c). Immunofluorescence staining of prelamin A confirmed this result and further showed typical nuclear morphological changes associated with prelamin A accumulation [24,28,35], consisting of intranuclear clusters and invaginations ( Figure 4d). As expected, treatment with mevinolin caused a reduction in migration ability in all OS cells. This effect was mostly evident in MG63 cells, which accumulate the highest prelamin A amount, while milder effects of statins on cellular migration were observed in the other cell lines (Figure 4e). Consistent with these observations, signaling effectors of cellular migration and invasion, like MMP9 (matrix metallopeptidase 9) and Cathepsin K, were decreased in statin-treated MG63 OS cells and, to a lesser extent in the other OS cell lines (Figure 4f,g). Finally, ROCK2 (Rho-associated protein kinase 2) kinase, which is a crucial intracellular mediator regulating osteosarcoma migration [36,37], was significantly downmodulated in MG63 OS cells (Figure 4f,g).
These findings indicated a major role for prelamin A in reducing cellular migration upon mevinolin administration.

Mevinolin Effects in OS Cells Are Related to Prelamin A Accumulation
Concentration-dependent viability was assessed upon mevinolin treatment in OS cells. After 24 or 48 h of treatment, only 143B and HOS cells showed a significant decrease in cell viability (around 25% when mevinolin was employed at 10 µM concentration) (Figure 5a). Interestingly, although 72 h mevinolin treatment reduced cell viability in all OS cell lines, the most significant effect of mevinolin was observed in MG63 cells, expressing the highest prelamin A levels ( Figure 5a). Moreover, after long-term mevinolin treatment, the highest increase in the percentage of apoptotic cells was observed in MG63 cell culture (Figure 5b).
It is well-known that cellular proliferation and survival, especially in tumor cells, is also related to Ras, whose activity is dependent on farnesylation [38]. Since 143B cells carry a Ras mutation [31,32] we reasoned that mevinolin-induced reduction of cellular viability could be also related to Ras defarnesylation. Interestingly, while impairment of Ras farnesylation by mevinolin was evident in 143B cells, the highest induction of apoptotic markers, including cleaved Caspase 3 and PARP (Poly (ADPribose) polymerase), was observed in mevinolin-treated MG63 cells (Figure 5c). These results suggest that, although effects of mevinolin on cellular viability and apoptosis are in part due to Ras inhibition, the most efficient reduction of cell survival is associated with prelamin A accumulation.

Discussion
Although an increasing number of reports recently implicated lamins in human cancers, their role in OS, the most common primary malignant neoplasia of the skeletal system [39], has been Interestingly, although 72 h mevinolin treatment reduced cell viability in all OS cell lines, the most significant effect of mevinolin was observed in MG63 cells, expressing the highest prelamin A levels ( Figure 5a). Moreover, after long-term mevinolin treatment, the highest increase in the percentage of apoptotic cells was observed in MG63 cell culture (Figure 5b).
It is well-known that cellular proliferation and survival, especially in tumor cells, is also related to Ras, whose activity is dependent on farnesylation [38]. Since 143B cells carry a Ras mutation [31,32] we reasoned that mevinolin-induced reduction of cellular viability could be also related to Ras de-farnesylation. Interestingly, while impairment of Ras farnesylation by mevinolin was evident in 143B cells, the highest induction of apoptotic markers, including cleaved Caspase 3 and PARP (Poly (ADP-ribose) polymerase), was observed in mevinolin-treated MG63 cells (Figure 5c). These results suggest that, although effects of mevinolin on cellular viability and apoptosis are in part due to Ras inhibition, the most efficient reduction of cell survival is associated with prelamin A accumulation.

Discussion
Although an increasing number of reports recently implicated lamins in human cancers, their role in OS, the most common primary malignant neoplasia of the skeletal system [39], has been elusive.
It was previously reported that lamin A is upregulated during OB differentiation of human bone marrow stromal cells, suggesting a critical role in osteogenesis [40]. Here, we firstly examined lamin A physiological expression in ADSCs and in differentiated OBs derived from ADSC cells or from healthy donors. Our findings revealed that lamin A expression directly correlates with the acquisition of differentiated status, as demonstrated in many cell types [41]. Further, our results are consistent with previous evidence demonstrating that lamin A/C knockdown suppresses OB differentiation [40,42].
In agreement with these considerations is also the finding that all OS cells here examined show lower lamin A expression levels relative to differentiated OBs. On the other hand, this observation also suggests that lamin A loss could play a role in OS tumorigenesis. Consistent with this hypothesis, we were able to show that elevated lamin A levels directly correlate with reduced migration and proliferation of OS cells and overexpression of lamin A in 143B OS cells, which express very low lamin A levels, is sufficient to significantly diminish cellular migration.
Based on these results, we tested the effects of mevinolin, a statin that interferes with processing of prelamin A, the lamin A precursor protein, which undergoes a complex post-translational processing requiring farnesylation of its C-terminus end [24]. Mevinolin inhibits HMG-CoA (3-hydroxy-3-methylglutaryl-coenzyme A) reductase by competitively binding a portion of the HMG-CoA binding site, thereby blocking the synthesis of isoprenoid lipids in the mevalonate pathway [43]. This condition reduces availability of the farnesyl moiety to many cellular constituents, among which prelamin A and Ras are examples. Statins have been extensively tested in humans, as they are widely used in hypercholesterolemia treatment. Moreover, previous observations in human and murine OS cells showed that statins sensitize OS cells to anticancer drugs and synergistically act with chemotherapeutic agents to reduce cell invasiveness [44,45]. Here, we focused on the role of prelamin A in statin-dependent effects. Although we confirm that mevinolin effects are related to Ras de-farnesylation and inhibition, our data show that the highest efficacy of mevinolin treatment is obtained in OS cells that accumulate the highest prelamin A levels. This supports the view previously suggested in a study performed in laminopathic mice [46], that prelamin A accumulation per se reduces tumor cell migration and metastatic potential. A correlation between accumulation of prelamin A and reduced cancer cell migration has also been demonstrated in pancreatic cancer cells using diverse drugs that inhibit lamin A maturation [47]. In addition, our results show that the first effect of statins in OS cells is to reduce cellular migration, while effects on cellular viability appear later on. This is not unexpected, provided the largely demonstrated mutual relationship between lamin A and prelamin A accumulation at the nuclear lamina and changes in cytoskeleton-dependent mechanisms [33,48,49].
Low lamin A expression could cause an increase of cellular migration due to effects of lamin A absence on mechanosignaling, as recently demonstrated [50]. The reason why prelamin A accumulation reduces OS cell migration deserves further investigation. A possibility could be that changes in LMNA levels affect linking of nucleoskeleton and cytoskeleton (LINC) proteins and LINC-dependent effects on mechanosignaling and cell migration. It is known that a prelamin A increase elicits high SUN1 levels [33,51,52]. This could impact on nucleo-cytoskeleton interplay. For instance, a recently published paper shows that low SUN1 levels favor migration of bone-marrow derived mesenchymal stem cells [53].
Although further studies in OS patient cohorts are necessary to validate the clinical relevance of our observations, the data presented here support for the first time a role of lamin A in the regulation of proliferation, differentiation and migration capabilities of OS cells and support the possible use of statins as adjuvant agent in the treatment of OS patients.