• Open Access

Increased mechanosensitivity and nuclear stiffness in Hutchinson–Gilford progeria cells: effects of farnesyltransferase inhibitors

Authors

  • Valerie L. R. M. Verstraeten,

    1. Cardiovascular Division, Department of Medicine, Brigham & Women's Hospital/Harvard Medical School, Boston, MA 02115, USA
    2. Department of Dermatology, University Hospital Maastricht, 6202 AZ Maastricht, The Netherlands
    3. School for Oncology and Developmental Biology, Maastricht University, 6200 MD Maastricht, The Netherlands
    Search for more papers by this author
  • Julie Y. Ji,

    1. Cardiovascular Division, Department of Medicine, Brigham & Women's Hospital/Harvard Medical School, Boston, MA 02115, USA
    Search for more papers by this author
  • Kiersten S. Cummings,

    1. Cardiovascular Division, Department of Medicine, Brigham & Women's Hospital/Harvard Medical School, Boston, MA 02115, USA
    Search for more papers by this author
  • Richard T. Lee,

    1. Cardiovascular Division, Department of Medicine, Brigham & Women's Hospital/Harvard Medical School, Boston, MA 02115, USA
    Search for more papers by this author
  • Jan Lammerding

    1. Cardiovascular Division, Department of Medicine, Brigham & Women's Hospital/Harvard Medical School, Boston, MA 02115, USA
    Search for more papers by this author

  • Valerie L. R. M. Verstraeten and Julie Y. Ji contributed equally to this work.


Jan Lammerding, PhD, Cardiovascular Division, Partners Research Facility, Room 283, 65 Landsdowne St, Cambridge, MA 02139, USA. Tel.: 617 768 8273; fax: 617 768 8280; e-mail: jlammerding@rics.bwh.harvard.edu

Summary

Hutchinson–Gilford progeria syndrome (HGPS), reportedly a model for normal aging, is a genetic disorder in children marked by dramatic signs suggestive for premature aging. It is usually caused by de novo mutations in the nuclear envelope protein lamin A. Lamins are essential to maintaining nuclear integrity, and loss of lamin A/C results in increased cellular sensitivity to mechanical strain and defective mechanotransduction signaling. Since increased mechanical sensitivity in vascular cells could contribute to loss of smooth muscle cells and the development of arteriosclerosis – the leading cause of death in HGPS patients – we investigated the effect of mechanical stress on cells from HGPS patients. We found that skin fibroblasts from HGPS patients developed progressively stiffer nuclei with increasing passage number. Importantly, fibroblasts from HGPS patients had decreased viability and increased apoptosis under repetitive mechanical strain, as well as attenuated wound healing, and these defects preceded changes in nuclear stiffness. Treating fibroblasts with farnesyltransferase inhibitors restored nuclear stiffness in HGPS cells and accelerated the wound healing response in HGPS and healthy control cells by increasing the directional persistence of migrating cells. However, farnesyltransferase inhibitors did not improve cellular sensitivity to mechanical strain. These data suggest that increased mechanical sensitivity in HGPS cells is unrelated to changes in nuclear stiffness and that increased biomechanical sensitivity could provide a potential mechanism for the progressive loss of vascular smooth muscle cells under physiological strain in HGPS patients.

Introduction

Hutchinson–Gilford progeria syndrome (HGPS) is a rare progeroid disorder affecting approximately 1 in 4 million newborns. Affected children appear normal at birth, but fail to thrive shortly thereafter and die in their early teens. Clinical characteristics include alopecia, beaked nose, sclerodermatous skin, dwarfism, lipodystrophy, osteoporosis and acro-osteolysis (Hennekam, 2006). HGPS is typically caused by autosomal dominant de novo mutations in the LMNA gene which encodes the nuclear intermediate filament proteins lamin A and C (De Sandre-Giovannoli et al., 2003; Eriksson et al., 2003). Lamins are the main components of the lamina, a filamentous protein meshwork underlying the inner nuclear membrane. Lamins provide structural support to the nucleus and have also been ascribed a role in transcriptional regulation (Broers et al., 2006; Verstraeten et al., 2007).

Mature lamin A is derived from its precursor prelamin A, which contains a C-terminal CAAX motif that prompts farnesylation of the cysteine residue by a protein farnesyltransferase. Subsequently, the last three amino acids (i.e. -AAX) are cleaved by the zinc metalloproteinase 24 (ZMPSTE24) or the endoprotease Ras-converting enzyme 1 (RCE1), and the farnesylated cysteine is subsequently methylated by isoprenylcysteine carboxyl methyltransferase. Finally, the end-standing C-terminal 15 amino acids, including the farnesyl cysteine methyl ester, are cleaved by ZMPSTE24, resulting in mature lamin A (Broers et al., 2006).

In the majority of HGPS patients, a heterozygous c.1824C→T (p.G608G) mutation in LMNA partially activates a cryptic splice site in exon 11 resulting in a truncated prelamin A protein (progerin), lacking 50 amino acids near the C-terminus. The deletion does not affect the CAAX motif, and therefore, the mutant protein undergoes normal farnesylation, -AAX cleavage and methylation. However, as progerin lacks the second ZMPSTE24 cleavage site, it remains farnesylated. The expression of the farnesylated mutant progerin and its accumulation at the nuclear envelope leads to grossly abnormal nuclear shape and compromised nuclear integrity (Goldman et al., 2004; Scaffidi & Misteli, 2005). Interestingly, accumulation of progerin has also been demonstrated in cells from normally aged healthy individuals (Scaffidi & Misteli, 2006), suggesting that HGPS could serve as a model for normal aging.

The most common cause of death in HGPS children (> 90% of cases) is myocardial infarction or stroke resulting from progressive arteriosclerotic disease. Post-mortem studies have shown significant and progressive loss of vascular smooth muscle cells (VSMC) in the medial layer of major arteries, and their replacement by fibrous material (Stehbens et al., 1999, 2001; Capell et al., 2007). Since increased mechanical sensitivity in vascular cells could contribute to loss of smooth muscle cells and the development of arteriosclerosis, we studied nuclear mechanics in HGPS cells, investigated the effect of mechanical stress, and hypothesized that HGPS cells would reveal increased cellular sensitivity upon strain. Previous studies have shown that treatment of patient cells with farnesyltransferase inhibitors (FTI) can prevent progerin from accumulating at the nuclear envelope and improve nuclear shape (Capell et al., 2005; Glynn & Glover, 2005; Mallampalli et al., 2005; Toth et al., 2005; Yang et al., 2005; Moulson et al., 2007). Therefore, we hypothesized that FTI treatment could restore nuclear mechanics and cellular sensitivity to strain.

Here, we studied nuclear mechanics and cellular sensitivity to mechanical strain in fibroblasts from HGPS patients carrying the typical G608G mutation. We found that patient fibroblasts developed stiffer nuclei with increasing passage number. More importantly, HGPS fibroblasts had decreased viability and increased apoptosis under repetitive mechanical strain, as well as attenuated wound healing responses compared to cells from healthy controls. Treatment of patient cells with FTI restored nuclear stiffness and improved the cellular wound healing response.

Results

HGPS cells have increased nuclear stiffness with increasing passage

Skin fibroblasts from HGPS patients exhibit increasingly abnormal nuclear shape with increasing passage in culture (Fig. 1A), presumably caused by accumulation of farnesylated progerin at the nuclear envelope (Goldman et al., 2004). To examine if these changes in nuclear shape reflect altered mechanical properties of the nucleus, we subjected skin fibroblasts from HGPS patients and healthy controls at various passages to uniform, biaxial strain. We found that HGPS cells at low and intermediate passages were indistinguishable from healthy controls, but that normalized nuclear strain was significantly decreased in late-passage HGPS cells (Fig. 1B), indicating a progressive increase in nuclear stiffness in the HGPS cells with increasing passage (Fig. 1C). In contrast, nuclear stiffness remained unchanged in healthy control cells. To address the possibility that the decrease in induced nuclear strain in late-passage HGPS cells could be caused by changes in the force transmission to the nucleus, we examined cytoskeletal structure by fluorescence staining for F-actin and microtubules and by magnetic bead microrheology (Bausch et al., 1998). Late-passage HGPS cells had no overt defects in cytoskeletal organization (Fig. 1D), and magnetic bead displacements were normal in these cells (Fig. 1E), confirming that the decrease in normalized nuclear strain was caused by increased nuclear stiffness and not changes in cytoskeletal structure. These findings are consistent with recently reported abnormal nuclear mechanical properties in HGPS patient fibroblasts subjected to micropipette aspiration and osmotic swelling (Dahl et al., 2006). Earlier studies on cells from HGPS patients suggested that progerin levels could increase with increasing passage (Goldman et al., 2004; McClintock et al., 2006), but these data were not quantified in multiple cell lines and one of the Western blots lacked loading controls. To test if the observed progressive changes in nuclear stiffness and nuclear shape abnormalities could be due to increased cellular levels of progerin, we quantified expression of lamin A, progerin, and lamin C normalized to actin loading controls in four HGPS patient fibroblast lines collected at various passages for each cell line. In our experiments, progerin levels remained constant over the range of passages (#10 through #25) analyzed (Fig. 1F). These findings suggest that the previously reported accumulation of progerin at the nuclear envelope likely reflects re-localization of progerin to the nuclear envelope rather than an increase in protein levels, resulting in increasingly abnormally shaped and stiffer nuclei with increasing passage number.

Figure 1.

Increased nuclear stiffness in late-passage Hutchinson–Gilford progeria syndrome (HGPS) fibroblasts. (A) Nuclear shape was analyzed in HGPS fibroblasts and control cells at a wide range of passages. Late-passage HGPS had significantly fewer normally shaped nuclei than early-passage HGPS cells and passage-matched controls (**P < 0.01 vs. early-passage HGPS cells and late-passage controls) (B) The nuclear stiffness in early- and late-passage HGPS cells and healthy control cells was evaluated by cellular strain application. The extent of induced nuclear deformations was expressed as a ratio of nuclear strain to applied membrane strain (normalized nuclear strain). Lower values indicate increased nuclear stiffness. (C) Increase in nuclear stiffness with increasing passage number seen in one representative cell strain (AG01972). The solid line depicts the linear regression, with dashed lines marking the 95% confidence interval. The slope was –8.8 ± 2.3 × 10−3/passage. (D) Fluorescence labeling of F-actin (green) and microtubules (red) indicating normal cytoskeletal structure in HGPS cells. Scale bar 20 µm. (E) Cytoskeletal stiffness measured by magnetic bead microrheology. Bead displacement amplitude (left) and residual displacement (right) after sinusoidal force application were normal in HGPS cells. (F) Western blot analysis of lamin A, lamin C, and progerin of one representative HGPS cell line at various passages (top). Protein levels normalized to actin loading controls compared between early- and late-passage cells from four independent HGPS cell lines analyzed by densitometry (bottom).

HGPS cells have increased sensitivity to mechanical strain

Post-mortem studies in HGPS patients revealed a dramatic loss of vascular smooth muscle cells in the aorta and other large blood vessels (Stehbens et al., 1999, 2001). Genetically engineered mice carrying a bacterial artificial chromosome with a human progeria mutation (LMNA G608G) showed a similar vascular phenotype, but did not die prematurely (Varga et al., 2006). The aforementioned suggests that vascular smooth muscle cells are especially sensitive to hemodynamic stress and vessel expansion in HGPS patients (Stehbens et al., 1999, 2001; Capell et al., 2007). To examine if increased nuclear stiffness could render cells more sensitive to mechanical strain, we subjected HGPS and healthy control fibroblasts to 24 h of repetitive, biaxial strain. Following strain application, cells from HGPS patients had significantly larger fractions of propidium iodide-positive cells than unstrained HGPS cells and strained healthy controls, indicating increased cellular sensitivity to mechanical strain in HGPS cells (Fig. 2A). Importantly, similar results were seen in both early- and late-passage HGPS cells, i.e. even before the onset of increased nuclear stiffness. To analyze the increased rate of strain-induced cell death in HGPS fibroblasts in more detail, we determined the fraction of apoptotic cells in aliquots from the same samples. Strained HGPS fibroblasts had larger fractions of cells with sub-G1 DNA content than unstrained HGPS cells and strained healthy controls (Fig. 2B), thus closely mirroring the trend observed in the propidium iodide-uptake assay.

Figure 2.

Hutchinson–Gilford progeria syndrome (HGPS) cells are more sensitive to mechanical strain. (A) HGPS fibroblasts subjected to 24 h of repetitive, biaxial strain had increased fractions of propidium iodide-positive dead cells (*P < 0.05 vs. unstrained controls). (B) Fraction of apoptotic (sub-G1) cells from the same experiments (**P < 0.01 vs. unstrained controls). (C) Representative results of flow cytometry DNA content analysis of HGPS and control fibroblasts subjected to 24 h of cyclic, biaxial strain and unstrained controls. (D) HGPS cells lacked the strain-induced cell proliferation seen in healthy control fibroblasts (***P < 0.001 vs. unstrained controls).

HGPS cells have impaired cell cycle activation by biomechanical strain

In normal cells, strain application can induce mechanosensitive genes and proliferative signals, resulting in cell cycle activation and proliferation (Sedding et al., 2003). Accordingly, we found that the percentage of cells in S/G2 phase significantly increased in response to strain application in healthy controls (Fig. 2C,D). In contrast, cells from HGPS patients mostly lacked this cell cycle activation response, and the fraction of cells in S/G2 remained unchanged following strain application (Fig. 2C,D). These findings are consistent with recent reports that progerin can delay progression through mitosis and impair Rb-mediated cell cycle transition into S-phase (Cao et al., 2007; Dechat et al., 2007).

Wound healing is impaired in HGPS cells

Increased sensitivity to mechanical strain in vascular cells – especially in combination with defects in cell cycle activation and progression – could contribute to the progressive loss of smooth muscle cells and the development of arteriosclerosis seen in HGPS patients (Capell et al., 2007). To gain further insight into cellular functions relevant to HGPS, we studied cell migration and proliferation in an in vitro wound healing model (Fig. 3A). Wound closure was significantly impaired in HGPS fibroblasts (Fig. 3B). Reintroduction of serum led to complete wound closure within 2–3 days in HGPS cells (Data not shown), demonstrating that HGPS cells had maintained the potential to proliferate and had not simply senesced prematurely.

Figure 3.

Impaired wound healing in Hutchinson–Gilford progeria syndrome (HGPS) fibroblasts. (A) Representative images of the wound area at 0, 24 h, and 144 h after wound creation for HGPS cells and healthy controls show the slower wound closure in HGPS cells. Scale bar 100 µm. (B) The mean open wound area as a percentage of the initial wound area at different time points indicates a significantly delayed wound closure for HGPS fibroblasts compared to healthy controls (*P < 0.05 vs. controls; **P < 0.01 vs. controls).

Farnesyltransferase inhibitor treatment improves nuclear shape and stiffness

FTI treatment, which interferes with prelamin A processing (Fig. 4A) and decreases the level of farnesylated progerin, was shown to reduce the frequency of abnormally shaped nuclei (Capell et al., 2005; Glynn & Glover, 2005; Mallampalli et al., 2005; Toth et al., 2005; Yang et al., 2005; Moulson et al., 2007). Consistent with these reports, we found a small but significant reduction in the fraction of abnormally shaped nuclei in mid- and late-passage HGPS cells treated with FTIs (Fig. 4B). In healthy control cells, FTI treatment had no effect on nuclear shape. To examine whether FTI treatment could also restore changes in nuclear stiffness in HGPS fibroblasts, we carried out nuclear strain experiments on late-passage HGPS cells and passage-matched healthy controls treated with FTI or vehicle (dimethyl sulfoxide) alone. After 3 days of treatment, nuclear stiffness in HGPS cells improved significantly and reached values comparable to healthy control cells (Fig. 4C). Importantly, FTI treatment did not alter nuclear stiffness in healthy control fibroblasts. These experiments were confirmed with a second FTI, kindly provided by Loren Fong (Toth et al., 2005).

Figure 4.

Farnesyltransferase inhibitor (FTI) treatment improved nuclear mechanics in Hutchinson–Gilford progeria syndrome (HGPS) cells. (A) Efficacy of FTI treatment was confirmed by Western blot analysis of total cell lysates as described previously (Toth et al., 2005). FTI treatment resulted in accumulation of prelamin A in the FTI-treated HGPS and control samples (asterisk). (B) FTI treatment reduced the frequency of abnormally shaped nuclei in HGPS cells (*P < 0.05 vs. vehicle-treated cells). (C) FTI treatment restored nuclear stiffness in HGPS cells but had no effect on healthy control fibroblasts (*P < 0.05, **P < 0.01). (D) FTI-treatment had no effect on the fraction of dead (propidium iodide-positive) cells in response to 24 h of cyclic, biaxial strain application (10% strain at 1Hz).

Farnesyltransferase inhibitor treatment does not rescue cellular sensitivity to mechanical strain

After demonstrating that FTI treatment successfully restored nuclear stiffness in late-passage HGPS fibroblasts, we tested whether FTI treatment could also rescue the increased mechanical sensitivity in HGPS cells. However, we found no significant difference in strain-induced cell death (Fig. 4D) or apoptosis (Data not shown) between cells treated with FTI or vehicle alone. These data suggest that increased sensitivity to mechanical strain in the HGPS fibroblasts is independent of changes in nuclear stiffness, consistent with our above findings that HGPS cells have increased mechanical sensitivity even before changes in nuclear stiffness occur.

Farnesyltransferase inhibitor treatment improves the cellular wound healing response

Genetically engineered mice expressing a progerin construct develop severe premature aging symptoms that can be ameliorated with FTI treatment (Yang et al., 2006), indicating that FTIs improve cellular function in the treated animals. While we could not detect any improvements in our cell viability assay, we found that FTI treatment dramatically accelerated the wound healing process (Fig. 5A,B). Interestingly, FTI treatment also resulted in improved wound healing in healthy control cells (Fig. 5B), indicating that the observed effect was not specific to cells with prelamin A processing defects. To further address if the improved wound healing in response to FTI treatment was a primary effect on lamin A, we conducted wound healing experiments on one mouse embryo fibroblast cell line from mice lacking lamins A and C (Lmna−/–) (Sullivan et al., 1999), one cell line of mice lacking lamin A but not lamin C (lamin C-only, LmnaLCO/LCO) (Fong et al., 2006b) and a fibroblast cell line derived from wild-type littermates of the Lmna−/– mice (Fig. 5C). Lamin C-only cells and wild-type mouse embryo fibroblasts treated with 10 µm FTI closed the wound faster than cells treated with vehicle alone, indicating that the improved wound healing effect was independent of blocking lamin A/progerin farnesylation. In contrast to an earlier report (Lee et al., 2007), we found that lamin A/C-deficient cells had dramatically faster wound closure than lamin C-only and wild-type cells, and FTI treatment had no further effect on the wound closure in these cells. The difference between our findings and those of Lee et al. could possibly be due to the fact that each immortalized cell line develops multiple mutations over time and can, therefore, although exhibiting the same genotype, behave differently in terms of growth and morphology. Since wound closure can be a combination of cell migration and proliferation, we evaluated these functions in separate assays. BrdU-labeling revealed increased DNA synthesis in cells at the wound edge after wounding, whereas little or no DNA synthesis could be detected away from the wound (Fig. 5D). Additional experiments indicated that FTI treatment can result in impaired cell cycle progression in cells grown in full medium as evidenced by accumulation of cells in S/G2 (Supplementary Fig. S1). To assess the effect of FTI treatment on cell migration, we tracked single, serum-starved cells and computed the average migration speed and directional persistence time (Supplementary Fig. S2). We found that FTI treatment did not alter cell migration speed, but increased the persistence time in both HGPS and healthy control cells (Fig. 5E). It is also noteworthy that HGPS cells treated with vehicle alone had significantly lower migration speeds than comparable healthy controls (Fig. 5E). These data suggest that FTIs exert a beneficial effect on cell migration into the wound by increasing the directional persistence in cells.

Figure 5.

Farnesyltransferase inhibitor (FTI) treatment improved wound healing in Hutchinson–Gilford progeria syndrome (HGPS) and healthy control cells. (A) Representative images of the wound area at 0, 24 h, and 144 h after wound creation for HGPS cells treated with FTI or vehicle alone. Scale bar 100 µm. (B) Left, representative examples of the residual open area for HGPS (filled symbols) and healthy control cells (open symbols) treated with FTI (triangle) or vehicle alone (circle) indicate that FTI treatment accelerates wound closure. Right, comparison of the residual open area after 24 h and 144 h for HGPS and healthy control cells treated with FTI or vehicle alone (*P < 0.05, **P < 0.01, ***P < 0.001). (C) Wound healing experiments on wild-type (Lmna+/+, left), Lamin C–only (LmnaLCO/LCO, center), and lamin A/C-deficient (Lmna−/–, right) mouse embryo fibroblasts treated with 10 µm FTI or vehicle alone (*P < 0.05, **P < 0.01, ***P < 0.001). (D) BrdU-labeling revealed increased DNA synthesis at the wound edge within the first 24 h after wounding. Scale bar 200 µm. (E) Left, HGPS cells had significantly lower migration speeds compared to healthy control fibroblasts (**P < 0.01). FTI treatment had no effect on migration speed. Right, FTI treatment increased the directional persistence time of migration in both HGPS and control fibroblasts.

Discussion

Arteriosclerosis is the leading cause of death in HGPS patients (Al-Shali & Hegele, 2004), and ultrastructural analysis of vascular tissue from HGPS patients and from genetically engineered mice expressing the human LMNA G608G mutation reveals extensive loss of vascular smooth muscle cells and, thus, potentially an unusual susceptibility to hemodynamic stress (Stehbens et al., 1999, 2001; Varga et al., 2006). In addition, vascular endothelial cells appear unable to preserve intimal integrity after injury (Baker et al., 1981). As increased cellular sensitivity to mechanical stress and increased nuclear fragility had previously been reported for lamin A/C-deficient cells (Broers et al., 2004; Lammerding et al., 2004), we explored if similar cellular defects could contribute to the pathogenesis of HGPS. Here, we show that HGPS fibroblasts develop progressively stiffer nuclei with increasing passage number and are more sensitive to mechanical strain. Since the overall expression of progerin remained unchanged with increasing passage number, the increased nuclear stiffness could result from progressive accumulation of progerin at the nuclear envelope over time. At the nuclear periphery progerin can interact with wild-type lamins A and C and impair their mobility, causing abnormal nuclear mechanics and a reduced ability to rearrange the nuclear lamina under mechanical stress (Dahl et al., 2006). The concept that accumulation of farnesylated progerin at the nuclear envelope (and not an increase in progerin expression) leads to stiffer nuclei is further supported by our finding that FTI treatment successfully restored nuclear stiffness in late-passage HGPS cells. In the light of recent studies that report accumulation of progerin in cells from elderly unaffected individuals (Scaffidi & Misteli, 2006; McClintock et al., 2007), the conflicting findings regarding progerin accumulation with increasing passage in cultured HGPS cells, and the lack of in vivo measurements of changes in progerin levels in the progression of HGPS, the issue of increases in progerin levels and progerin accumulation at the nuclear envelope and its consequences on aging deserves a careful re-evaluation.

In addition to altered nuclear mechanics, we found that HGPS cells had significantly higher fractions of necrotic and apoptotic cells upon repetitive strain than healthy control fibroblasts. Importantly, increased nuclear stiffness could not fully explain the increased sensitivity of HGPS cells to mechanical strain. Early- and late-passage HGPS cells showed a similar fraction of dead cells in response to repetitive mechanical strain, even before changes in nuclear stiffness became apparent; and while FTI treatment rescued nuclear stiffness, it had no effect on cellular sensitivity to mechanical strain. Next to higher rates of cell death under mechanical stress, insufficient regeneration or repair of affected blood vessels could further contribute to the pathogenesis of arteriosclerosis in HGPS. Our results indicate that HGPS cells had significantly impaired cell cycle activation in response to strain stimulation. Furthermore, an in vitro wound healing assay showed an impaired ability of patient fibroblasts to migrate into and close the wound, regardless of their passage number. McClintock et al. (2006) previously reported that HGPS cells with the highest expression levels of progerin had impaired cell migration, and impaired cell cycle activation could further contribute to the slower wound closure in HGPS cells.

Thus, one of the primary findings of this study is that HGPS cells have increased sensitivity to mechanical stress and are further characterized by impaired cell cycle activation and migration. Mechanical weakening of vascular endothelial and smooth muscle cells could be the initiating pathological event that leads to arteriosclerosis (Davies, 1995; Mounkes & Stewart, 2004) or at least add to other cellular defects seen in HGPS such as genetic instability and premature cellular senescence (Mounkes & Stewart, 2004; Corso et al., 2005; Liu et al., 2005; Varela et al., 2005; Kudlow et al., 2007). Since vascular smooth muscle cells (VSCM) and endothelial cells were shown to be the primary targets of progerin build-up (McClintock et al., 2006) and are constantly exposed to fluid shear stress and strain in the vessel wall, increased cellular sensitivity to mechanical stress and impaired cell cycle activation may predominantly affect vascular cells and, therefore, contribute to arteriosclerosis. The importance of normal nuclear structure in the physiological response to shear stress is further supported by the recent finding that nuclei of endothelial cells subjected to fluid shear stress show persistent changes in nuclear shape and stiffness, acting as a stress-bearing organelle (Dahl et al., 2004; Deguchi et al., 2005). Whereas the increased mechanosensitivity and impaired wound healing in HGPS cells preceded gross changes in nuclear stiffness, we cannot rule out that ultrastructural changes at the nuclear periphery by altered interaction with progerin are ultimately responsible for these defective cellular functions.

Several groups have recently reported that FTI treatment leads to improved nuclear shape in cells expressing progerin (Capell et al., 2005; Glynn & Glover, 2005; Mallampalli et al., 2005; Toth et al., 2005; Yang et al., 2005; Moulson et al., 2007). However, while FTI treatment appears very effective at improving nuclear shape abnormalities in HGPS fibroblasts, it has remained unclear if normalization of nuclear morphology translates into a general rescue of cellular function and the reversal of tissue-specific disease phenotypes seen in HGPS. Here, we confirmed the improvement in nuclear shape and furthermore found that FTI treatment could normalize nuclear stiffness in HGPS cells. And while FTI treatment did not improve cellular sensitivity to strain, it dramatically improved the cellular wound healing response. Surprisingly, however, this effect was independent of the accumulation of progerin as FTI similarly improved wound healing both in HGPS and healthy control fibroblasts. Our experiments on cells lacking lamin A strongly suggest that some of the beneficial effects of FTI treatment are caused by targeting farnesylated proteins other than lamin A/progerin. Previous reports demonstrate an inhibitory effect of FTI on cell proliferation and migration at high doses (Kouchi et al., 1999; Kusama et al., 2003, 2006; Desrosiers et al., 2005). Our data, using lower doses of FTI, support the inhibitory effect of FTI on proliferation. We detected a significantly improved wound healing response as early as 3 h after wounding in FTI-treated cells (Data not shown), suggesting a beneficial effect of FTI on cell migration. Consistent with this hypothesis, we found an increased directional persistence in the migration of FTI-treated healthy control and HGPS fibroblasts, whereas FTI treatment did not alter migration speed. This increase in persistence time may result from the effect of FTI on other isoprenylated proteins such as Rho, Rac, and Ras (Kusama et al., 2003, 2006). In particular, a previous report indicated that a decrease in Rac activity can switch the migration pattern of cells from random to a directionally more persistent way of migration (Pankov et al., 2005), which is consistent with our observations.

In conclusion, we show that HGPS nuclei become progressively stiffer with increasing passage and that increased cellular sensitivity upon prolonged strain, impaired migration, and defective induction of proliferation needed for vessel regeneration might contribute to the loss of VSMC and the development of arteriosclerosis in HGPS patients. Furthermore, increased mechanical sensitivity and defective wound healing could also contribute to skin abnormalities and loss of bone and muscle tissue in HGPS patients. Moreover, bone fractures in HGPS knock-in mice and ZMPSTE24-deficient mice do not heal. FTI treatment effectively normalized nuclear stiffness and improved the wound healing response, which may contribute to the reduction of bone fractures in mouse models for premature aging treated with FTI (Fong et al., 2006a; Yang et al., 2006). However, FTI could not restore cellular viability and the proliferative capacity upon strain.

Experimental procedures

Cell culture and FTI treatment

Human dermal fibroblasts from HGPS patients [AG01972 (14-year-old female), AG06917 (3-year-old male), AG11513 (8-year-old female), HGADFN127 (~4-year-old female), HGADFN143 (~9-year-old male)] and apparently healthy controls [GM00038 (9-year-old female), GM00499 (8-year-old male), GM00498 (3-year-old male), GM01651 (13-year-old female), GM08398 (8-year-old male), AG09602 (92-year-old female), GM02037 (13-year-old male), GM00498 (3-year-old male)] were obtained from the Progeria Research Foundation and the Coriell Cell Repository. Results were pooled from at least two cell strains for each group. Fibroblasts were maintained at 37 °C in Earle's minimal essential medium (EMEM, Invitrogen, Eugene, OR, USA) supplemented with 15% fetal bovine serum (HyClone, Fisher Scientific, Ottawa, Ontario, Canada), and penicillin/streptomycin (Invitrogen). In case of serum starvation, cells were grown in EMEM supplemented with insulin, transferrin, and selenium (ITS, Sigma, St. Louis, MO, USA) and penicillin/streptomycin. For FTI treatment, cells were treated daily with 5 or 10 µm of FTI L744832 (Biomol) or equal volume of vehicle (dimethyl sulfoxide; final concentration 0.025%) for at least 72 h as described previously (Capell et al., 2005; Liu et al., 2006), with little difference between the two FTI doses.

Nuclear strain experiments

Experiments were performed as described previously (Lammerding et al., 2004, 2005), but without serum starvation. In brief, cells were cultured on fibronectin-coated silicone membranes for at least 3 days, incubated with Hoechst 33342 nuclear stain (Molecular Probes, Eugene, OR, USA) and subsequently kept in HBSS with calcium and magnesium. Uniform biaxial strain (~5%) was applied to the membranes with a custom-made strain device. Membrane and nuclear strain were computed based on phase-contrast and fluorescence images acquired before, during, and after strain application with custom image-analysis algorithms. Normalized nuclear strain was defined as the ratio of nuclear strain to membrane strain to compensate for small variations in applied membrane strain. Cells damaged during the experiments were excluded from the analysis.

Magnetic bead microrheology

Experiments were performed as described previously (Lammerding et al., 2004). In brief, cells were incubated for 40 min with paramagnetic beads (Invitrogen) coated with an antibody against β1-integrins (CD29; Fitzgerald Industries, Concord, MA, USA). To minimize nuclear effects, only beads attached more than 5 µm from the nucleus were selected for analysis. A sinusoidal force (1 Hz, amplitude 0.6 nN, offset 0.6 nN) was applied through a magnetic trap, and bead displacements were quantified with custom-written image-processing software.

Cell viability and apoptosis assays

Experiments to measure cell viability were carried out as described previously (Lammerding et al., 2004, 2005). Cells were plated on fibronectin-coated silicone membranes and maintained in full media. Following 24 h of cyclic, biaxial strain (10% strain at 1 Hz), cells were incubated with propidium iodide. Adherent and detached cells from each sample were collected and divided into two equal parts. One part was analyzed for cell death (propidium iodide uptake) by flow cytometry (Cytomics FC 500; Beckman Coulter, Fullerton, CA, USA), while cells in the other part were fixed in 80% ethanol, treated with ribonuclease A (Sigma), stained with propidium iodide, and analyzed for DNA content by flow cytometry (Walker et al., 1993).

Western blot analysis of nuclear proteins

Cells were lysed in RIPA buffer supplemented with 300 mm NaCl, 1 mm dithiothreitol, 0.5 mm phenylmethylsulfonyl fluoride, and protease inhibitor cocktail (all Sigma) or alternatively in 9 m urea buffer supplemented with 10 mm Tris-HCl, 0.2% 2-mercapto ethanol, 10 µm EDTA, 1 mm dithiothreitol, 0.5 mm phenylmethylsulfonyl fluoride, and protease inhibitor cocktail. Protein extracts were size-fractionated by sodium dodecylsulfate–polyacrylamide gel electrophoresis and blots were probed with primary antibody against lamin A/C (Santa Cruz Biotechnology sc-6215, Santa Cruz, CA, USA) and reprobed for actin (Sigma, A-2066) as loading control. The lamin A/C antibody recognizes lamins A and C, prelamin A, and progerin (Yang et al., 2005).

Wound healing assay

Confluent cell monolayers were serum starved for 3 days before creating an approximately 700-µm-wide wound with a 200-µL-micropipette tip. Cells were rinsed with EMEM + ITS, replaced with fresh media and returned to the incubator. Images were taken along the wound at ×10 magnification immediately after wounding and every 24 h for up to 6 days. The remaining wound area was quantified based on 4–10 sections at each time point with ImagePro (Media Cybernetics, Bethesda, MD, USA) and normalized to the initial wound area. For FTI treatment, cells were treated with 10 µm FTI for 3 days prior to wounding and treatment was continued for 6 days after wounding. In order to assess the level of proliferation induced upon wounding, bromodeoxyuridine (BrdU, Sigma) was added to the culture medium for 24 h. Subsequently, cells were fixed and stained with the anti-BrdU antibody (Roche #11170376001, Indianapolis, IN, USA) and scored by a blinded observer.

Cell migration assay

Cells plated at low cell density were serum-starved and treated with 10 µm of FTI or vehicle for 3 days prior to the experiments. Cells were placed on a temperature-controlled stage and images were acquired at ×10 magnification every 5 min for up to 12 h. Cell movement was tracked using custom-written software based on the nuclear position. Cells that divided or interacted with other cells were excluded from the analysis. Cell speed was calculated as the total path length divided by the observation time. The persistence time was computed from a random walk model fitted to the average mean square displacement of all cells from the same group as described previously (Harms et al., 2005).

Immunofluorescence microscopy

Cells were fixed in 4% paraformaldehyde, and permeabilized with 0.1% Triton-X 100. After blocking, cells were incubated with primary antibodies (antilamin A/C, Santa Cruz Biotechnology; anti-β-tubulin, Sigma) or OregonGreen488 phalloidin (Molecular Probes), followed by incubation with secondary antibodies conjugated to Cy3 (Sigma) or AlexaFluor 488 (Invitrogen). Cell nuclei were labeled with Hoechst 33342 (Invitrogen). Samples were imaged on an Olympus IX-70 microscope, Center Valley, PA, USA at ×40 (NA 1.15) and scored by a trained observer blinded for genotype and treatment.

Statistical analysis

Experiments were performed at least three independent times. Statistical analyses were performed with the PRISM 3.0 and INSTAT software (GraphPad, San Diego, CA, USA) by unpaired Student's t-test (allowing for different variance) with Welch's correction. For all experiments, a two-tailed P-value of < 0.05 was considered significant. All data are expressed as mean ± SEM.

Image acquisition and manipulation

Phase contrast and fluorescence images were acquired as described above using a CoolSNAP HQ digital CCD-camera (Photometrics, Roper Scientific, Tuscon, AZ, USA) mounted on an Olympus IX70 inverted microscope using IPlab 4.0 (BD Biosciences, Billerica, MA, USA) image acquisition software. Nuclear shape experiments were imaged with an Olympus LCPlanF ×20 phase contrast objective (NA 0.40), wound healing experiments with an Olympus UPlanFl ×10 phase contrast objective (NA 0.30) and nuclear strain experiments were conducted with an Olympus LCPlanFl ×60 objective (NA 0.70). Experiments on live cells were carried out in serum starved medium at 37 °C. Immunofluorescence samples were imaged with an Olympus UApo/340 ×40 water immersion objective (NA 1.15) at room temperature. For Western blot analysis, films were digitized on an Epson Perfection 2450 scanner using linear intensity settings. Digital images were processed using Adobe Photoshop (version 6.0) by adjusting the linear image intensity display range and fluorescence grayscale images were colorized in Adobe Photoshop or IPLab by selecting a colorplane (RGB) appropriate for the chromophore.

Acknowledgments

This work was supported by National Institutes of Health grants HL082792 and NS059348, the American Heart Association grant 0635359N, the Progeria Research Foundation, an National Research Service Award (NRSA) postdoctoral fellowship HL079862 (to J.Y.J.), a research grant from the University Hospital Maastricht, Maastricht, the Netherlands (to V.L.R.M.V.), and fellowships from the Netherlands Genomics Initiative and the Sint Annadal Stichting, Maastricht, the Netherlands (to V.L.R.M.V.). The authors thank Lana A. Peckham for her assistance in the nuclear shape analysis.

Ancillary