Dunnigan-type familial partial lipodystrophy (FPLD), characterized by an abnormal body fat redistribution with insulin resistance, is caused by missense heterozygous mutations in A-type lamins (lamins A and C). A- and B-type lamins are ubiquitous intermediate filament proteins that polymerize at the inner face of the nuclear envelope. We have analyzed primary cultures of skin fibroblasts from three patients harboring R482Q or R482W mutations. These cells were euploid and able to cycle and divide. A subpopulation of these cells had abnormal blebbing nuclei with A-type lamins forming a peripheral meshwork, which was frequently disorganized. Inner nuclear membrane protein emerin, an A-type lamin-binding protein, strictly colocalized with this abnormal meshwork. Cells from lipodystrophic patients often had other nuclear envelope defects, mainly consisting of nuclear envelope herniations that were deficient in B-type lamins, nuclear pore complexes, lamina-associated protein 2 beta, and chromatin. The mechanical properties of nuclear envelopes were altered, as judged from the extensive deformations observed in nuclei from heat-shocked cells, and from the low stringency of extraction of their components. These structural nuclear alterations were caused by the lamins A/C mutations, as the same changes were introduced in human control fibroblasts by ectopic expression of R482W mutated lamin A.
The familial partial lipodystrophy of the Dunnigan type (FPLD) is a rare autosomal dominant disease characterized by a post-pubertal regression of the subcutaneous fat from limbs and trunk contrasting with its accumulation in face and neck. This lipodystrophy is associated with insulin resistance and hypertriglyceridemia, which could be secondary to adipose tissue involution. Missense heterozygous mutations in the LMNA gene, affecting the C-terminal domain of lamins A and C, have been recently shown to be responsible for the disease (Cao and Hegele, 2000; Shackleton et al., 2000; Speckman et al., 2000; Vigouroux et al., 2000). However, the link between a mutation in lamins A and C (LaA/C), ubiquitous nuclear envelope (NE) proteins and the selective regression of adipose tissue is elusive. Furthermore, other alterations in LMNA, widely distributed all along the gene, result in skeletal or cardiac muscular diseases, or both (Bonne et al., 1999; Fatkin et al., 1999; Muchir et al., 2000).
Lamins A and C are major components of the nuclear lamina. They are members of the intermediate filaments protein family, with a similar primary and secondary structure (McKeon et al., 1986; Fisher et al., 1986). The LMNA gene generates lamin A and lamin C by alternative RNA splicing (Lin and Worman, 1993). These proteins are identical for their first 566 amino acids, which encompass the N-terminal head, the central rod helicoidal domain, and most of the tail domain. Thus the R482Q/W mutations, which are the most frequent mutations in FPLD, affect both A and C lamins. Lamins B1 and B2 are the other major components of the lamin family of proteins and are coded by different genes (Stuurman et al., 1998; Worman and Courvalin, 2000). A- and B-type lamins polymerize in various ratios to form the nuclear lamina, a protein network that is located between inner nuclear membrane and chromatin. Lamin genes are differentially expressed during development and cell maturation; B-type lamins are constitutive and A-type lamins are preferentially expressed in differentiated nonproliferating cells (Stewart and Burke, 1987; Guilly et al., 1987; Guilly et al., 1990).
A specific set of nuclear integral proteins interacts with lamina and could mediate its attachment to the inner nuclear membrane (INM) (Worman and Courvalin, 2000). Among these proteins, the lamin B receptor (LBR) and lamina associated protein 2 β (LAP2β) interact more specifically with B-type lamins (Furukawa et al., 1995; Worman et al., 1988), whereas emerin preferentially binds to A-type lamins (Fairley et al., 1999; Sullivan et al., 1999; Clements et al., 2000). The lamina is tightly associated with nuclear pore complexes (NPCs), possibly involving a direct interaction between lamins and Nup153, a peripheral component of the NPCs (Smythe et al., 2000). Finally, the nuclear lamina also interacts with chromatin, through lamin sequences located in their carboxyl-terminal end, downstream from the rod domain (Taniura et al., 1995; Goldberg et al., 1999).
Several roles for the nuclear lamina have been proposed, including its involvement in nuclear structure, gene expression, cell-cycle progression, and DNA replication (Lourim and Lin, 1992; Liu et al., 2000; Moir et al., 2000a). To investigate the cellular alterations involved in FPLD, we have performed a morphological and biochemical analysis of cultured skin fibroblasts from patients with the LaA/C R482Q/W heterozygous mutations. We show here that a population of cells from FPLD patients and from control individuals expressing ectopic LaA R482W present dysmorphic nuclei or a disorganization of the lamina, or both. In addition, cells in FPLD patients can have an altered protein composition of the INM, a discontinuous repartition of the nuclear pore complexes, accompanied by an increased sensitivity to heat shock and to extraction by nonionic detergents.
MATERIALS AND METHODS
Patients and cells
All individuals gave their informed consent for cell studies, which were approved by an institutional review board. Three female patients (JM, LM and KN, who were 34, 52 and 31 years old, respectively) were affected with FPLD secondary to heterozygous substitutions in lamins A and C (R482Q for JM, R482W for LM and KN) (Vigouroux et al., 2000) (C.V., unpublished). Subcutaneous fibroblasts from the FPLD patients (obtained by skin biopsy) and from two control individuals (25 and 38 year old females without any known disease who underwent cosmetic surgery) were cultured in DMEM medium containing 10% fetal calf serum for a maximum of 12 passages.
Fluorescence cytometric analyses
For FACS analysis, 105 trypsinized fibroblasts from control individuals and patients JM and LM were collected by centrifugation, resuspended in PBS and finally fixed in ethanol (70%) at 4°C. After RNase A (500 μg/ml) digestion and propidium iodide (50 μg/ml) staining, cells were analyzed on a FACS Calibur (Becton Dickinson) equipped with an argon laser emitting at 488 nm. Cells sorted according to the different phases of the cell cycle were observed by fluorescence microscopy and the percentage of abnormal nuclei in each fraction was calculated. These experiments were repeated four times. For cytofluorometric analysis of adherent cells, cells grown on coverslips were fixed in methanol, stained by propidium iodide as indicated above, and finally analyzed individually by DNA fluorescence using Meridian ACAS 570 interactive laser. Fluorescence of 100 cells was measured for each of the three patients and the two control individuals.
After a one day culture, cells from control individuals and patients JM and LM, seeded onto glass slides in Petri dishes, were mounted in observation microchambers known as the Rose chamber (Rose, 1954). A time-lapse unit generated one impulse per minute that controlled a 16 mm Arriflex camera. Phase contrast recordings were made over periods of 2 days. Between two successive frames, the light was turned off to avoid cellular damage due to light energy. After development, films were analyzed on a frame by frame projector and selected frames were enlarged and printed.
Antibodies and immunological methods
For conventional and confocal immunofluorescence microscopy, human fibroblasts grown on glass coverslips were fixed in methanol at –20°C, then processed as described (Buendia and Courvalin, 1997; Buendia et al., 1999). Monoclonal antibodies (mab) directed against A-type lamins, QE5 mab directed against nucleoporins p62, Nup153 and Nup214, and a mab specific for Nup153 were generous gifts from B. Burke (University of Calgary, Alberta, Canada) and R. Bastos (Institut Jacques Monod, CNRS, Paris). Mab anti-emerin (NCL-emerin clone 4G5) was purchased from Novocastra laboratories (Newcastle upon Tyne, UK), and mab anti-FLAG (M2) from Sigma (St Louis, MO). Rabbit antibodies directed against B-type lamins and LAP2β and human anti-lamin B antibodies have been previously described (Chaudhary and Courvalin, 1993; Buendia et al., 1999). Affinity-purified FITC-conjugated anti-rabbit antibodies and Texas Red-conjugated anti-mouse antibodies were purchased from Jackson ImmunoResearch laboratories (West Grove, PE). Given that the mean thickness of nuclei was statistically not different in the different populations of fibroblasts (z values estimated by observation at the confocal microscope), surface measurements were performed on two-dimensional images of DAPI-stained nuclei (58 control nuclei, 53 and 74 dysmorphic nuclei from patients JM and NK, respectively, and 59 eumorphic nuclei from patient NK) using manual outlining in Canvas 5.0. Intensity of the DAPI-staining was evaluated in 83 dysmorphic nuclei from patients JM and NK, in parallel in the bud and in the rest of each nucleus, using Metaview software. Immunoblotting analysis of cell extracts was performed as previously described (Buendia and Courvalin, 1997).
Heat-shock experiments were performed by transferring cells from 37°C to 45°C for 30 minutes. Cells were then either used immediately or transferred back to 37°C. For morphological examination, fibroblasts from the two control individuals and the three patients were grown on glass coverslips, washed for one second in PBS after heat shock, then immediately fixed in cold methanol (–20°C) for 10 minutes, and finally processed for immunofluorescence studies. After staining with anti-lamin A/C antibodies, 200-300 nuclei from patient and control fibroblasts were examined. Cellular viability after heat shock was evaluated at 0, 24 and 48 hours. Attached cultured cells, heated or unheated, from control individuals and patients LM and KN were trypsinized, collected by centrifugation, and counted. This experiment was repeated four times.
Monolayers of fibroblasts grown on Petri dishes were sequentially extracted as described (Fey et al., 1984). The first extraction was performed with 0.5% Triton-X100 to release soluble proteins (S1). This extraction was then repeated in the presence of 0.25 M (NH4)2SO4 to release cytoskeletal proteins (S2). Finally, cells were digested with DNase 1 and RNase A, then extracted again with 0.25 M (NH4)2SO4 to release chromatin material (S3). Insoluble final material (Ins.) contained nuclear matrix and intermediate filament proteins, including lamins as checked by immunofluorescence microscopy (data not shown). Proteins from the different soluble fractions were collected by precipitation with 10% TCA, then resuspended in one volume of SDS sample buffer (Laemmli, 1970). The insoluble material (Ins.) was directly resuspended in the same volume of SDS sample buffer. Whole-cell extracts were prepared by direct solubilization of fibroblasts in SDS sample buffer. Electrophoresis was performed on a 7.5% polyacrylamide gel, according to Laemmli (Laemmli, 1970).
Transfection of control human fibroblasts
Primary cultures of control human fibroblasts were grown in DMEM medium containing 15% fetal calf serum. cDNAs encoding wild-type or R482W-prelamin A were inserted into pSVK3 plasmids. For detection of expressed proteins, FLAG-epitopes were fused to the amino-termini of the constructs (Östlund et al., 2001). Cells grown at a 60% confluency were transfected in chamber slides using Lipofectamine PLUSTM (Life Technologies), following the manufacturer’s instructions. The cells were overlaid with the lipid-DNA complexes for 5 hours in serum-free medium, then grown in fresh complete medium with 15% serum for 24 hours. Cells were then fixed in methanol for 10 minutes at –20°C and processed for immunofluorescence studies. In these cells, the efficiency of transfection was low (2-4%).
Nuclei in some fibroblasts from patients with FPLD have a dysmorphic shape or a disorganization of A- and B-type lamin networks, or both
We have established primary cultures of skin fibroblasts from control individuals and from three patients with FPLD harboring the R482Q (patient JM) or R482W (patients LM and KN) heterozygous mutations in lamins A and C. Cells were fixed, DNA was stained with DAPI and A- and B-type lamins were labeled with specific antibodies, before conventional (Fig. 1A) and confocal (Fig. 1B) immunofluorescence microscopy. Several abnormalities were observed in nuclei from patients, which affected the shape of the nucleus, chromatin density and lamina organization. In control fibroblasts, nuclei were generally round or ovoid, whereas a subpopulation of fibroblasts from FPLD patients contained nuclei of irregular shape with occasional budding (dysmorphic nuclei) (Fig. 1A,B). A broad pattern of nuclear dysmorphies was observed in these lobulated nuclei, some herniations being connected to the nuclei by large necks (Fig. 1A,B; FPLD, arrowheads), whereas others showed a strong lamin staining at their base (Fig. 1B; FPLD, arrows), likely due to twisting of the NE in these areas. Although dysmorphic nuclei could also exist in the control cell population, they were rare and presented less-severe shape abnormalities (Fig. 1A).
Abnormalities of A- and B-type lamin networks were observed in nuclei from FPLD patients. As in control nuclei, A-type lamins were present at the whole nuclear periphery (Fig. 1A,B), but with a frequent disorganization of the lattice in the buds or at some nuclear poles (Fig. 1B). The honeycomb aspect of the lamin A/C staining was not seen in control fibroblasts, even in the rare dysmorphic nuclei present in this cell population (Fig. 1A, left panel, arrowheads). The B-type lamin network was also affected in some nuclei of FPLD patients. A loss of B-type lamins frequently occurred in nuclear buds or at nuclear poles (Fig. 1A,B; FPLD, arrowheads in green panels). When the loss in B-type lamins was only partial, the structure of B-type network was altered, with a honeycomb aspect (Fig. 1B; FPLD, asterisks).
DNA staining appeared homogeneous in control nuclei but heterogeneous in dysmorphic nuclei of FPLD patients, with a weaker labeling in buds (Fig. 1A). DAPI staining intensity was quantified in the buds versus the rest of the nuclei in a series of fibroblasts (∼80) from patients JM (LaA/C R482Q) and KN (LaA/C R482W). Results were similar in the two patients and showed that DAPI staining was significantly lower in the buds versus the rest of the nuclei (26±16%). These data suggested that, in the buds of FPLD nuclei, the chromatin was more decondensed.
Variability in the nuclear abnormalities observed in cells from FPLD patients
Although the type of nuclear abnormalities present in the fibroblasts from the three patients was similar, large variations in the frequency of these alterations were observed between patients and between cells from the same patient. First, only a subset of the cells had dysmorphic nuclei, its proportion increasing with cell passages. Patient JM (LaA/C R482Q) had the more prevalent dysmorphic nuclei, their percentage increasing from 10 to 22% between cell passages 4 and 10. Only 5-13% of nuclei were dysmorphic in fibroblasts from patients LM and KN (LaA/C R482W), and 2-4% in nuclei from control fibroblasts. JM dysmorphic nuclei were also larger (367±126 μm2) than control nuclei (192±79 μm2) and than the nuclei from the two other patients (203±79 μm2). Second, the frequency of the honeycomb aspect of lamin A/C staining was different among patients, varying from 1% (JM) to 15% or 50% (LM and KN, respectively) of fibroblasts with dysmorphic nuclei. This lamin A/C lattice disorganization was also observed in 3% of eumorphic nuclei from patients LM and KN (Fig. 1B, asterisks). Third, the loss in B-type lamins was more pronounced in nuclear buds or poles of fibroblasts from patient KN (80%) than in those from patient JM (30%).
Abnormal localization of INM proteins and nucleoporins in the NE of cells from FPLD patients
The consequences of lamina disorganization on the localization of emerin, LAP2β and nucleoporin Nup153 was checked by double immunofluorescence and confocal microscopy. In control nuclei, a similar nuclear peripheral localization was observed for both types of lamins, emerin, LAP2β and NPC markers (Fig. 2, top). Nuclear envelope abnormalities were identical in patients with the R482Q and W mutations in lamin A/C, with a conserved colocalization of these lamins and emerin even in areas where the lamin network was disorganized (Fig. 2, bottom, left column). In nuclear buds depleted of B-type lamins, (1) emerin was still present but with a honeycomb aspect, likely reflecting the structure of the underlying A-type lamin lattice (Fig. 2, bottom), (2) the signals for LAP2β and Nup153 were variable, ranging from normal to reduced to abolished (Fig. 2, bottom). The reduction of Nup153 signal was not specific for this nucleoporin, as similar data were obtained with mab QE5 which, in addition to Nup153, also recognizes nucleoporins p62 and Nup214 (data not shown).
Cell-cycle progression in cells from FPLD patients
Different methods were used to analyze the cell-cycle progression in fibroblasts from FPLD patients with the LaA/C R482Q/W mutations. First, exponentially growing control fibroblasts and fibroblasts from patients JM (R482Q) and LM (R482W) were studied by FACS analysis and compared. In these respective cell populations, the percentages of cells in G0/G1 (83.4±7.9 versus 82.9±3.1), S (7.6±4.3 versus 8.0±3.5), G2/M (7.5±3.8 versus 7.6±1.8) and apoptosis (2.4±1.6 versus 2.9±1.6) were not statistically different (Fig. 3A). Second, cells from FPLD patients were sorted according to their position in the cell cycle (G0/G1, S, and G2/M), assessed by the intensity of the propidium iodide labeling, then examined by fluorescence microscopy. The percentage of dysmorphic nuclei was identical in all cell fractions (data not shown). Third, dysmorphic nuclei were analyzed for their DNA content using cytofluorometric analysis of adherent cells (see Materials and Methods). The amount of DNA (between 2N and 4N) found in these nuclei was similar to that measured in normal cycling cells. Finally, living fibroblasts from control and patients JM (R482Q) and LM (R482W) were examined over a 36 hour time period by phase contrast and microcinematography. The fate of a cell from patient LM with a dysmorphic nucleus is presented in Fig. 3B. Pictures show that (1) nuclear dysmorphy was not transient, as it remained unchanged over a 28 hour time period, (2) the cell was able to divide (Fig. 3B, 32 h, 36 h). Altogether, data obtained by these four approaches showed that the large size or dysmorphy, or both, of nuclei in cells from FPLD patients were linked neither to an arrest at the G2/M transition phase, nor to polyploidy, nor to apoptosis.
FPLD fibroblasts are hypersensitive to heat shock
To evaluate the resistance to stress of the nuclear envelope, fibroblasts from the three patients bearing R482Q or W mutations in lamin A/C were submitted to heat-shock treatment (45°C for 30 minutes), then immediatly fixed. Nuclear shape was examined after labeling with antibodies directed against lamin A/C. Three serial confocal cell sections were registered (Fig. 4, top, middle, bottom), as well as their superposition (Fig. 4, Total). Nuclear shape and A-type lamin distribution in control nuclei were not modified by heat-shock treatment (Fig. 4, compare the two left columns), whereas extensive nuclear deformations were appearing in fibroblasts from FPLD patients. In unheated cells from control and patients, as well as in heated control cells, the overall nuclear periphery was entirely visible on the middle section, whereas three optical sections (i.e. 3×0.5 μm) were required to observe the total periphery of nuclei in heat-treated cells from FPLD patients (Fig. 4). These nuclear deformations were accompanied by the occurrence of pleats in the nuclear envelope, which appeared as strongly stained linear and branched structures (Fig. 4, arrowheads). This particular labeling was indicative of envelope folds rather than intranuclear structure because of its strict colocalization with emerin in double labeling experiments (data not shown). In some nuclear sections, well defined areas of the lamina appear unlabeled, suggesting the presence of a hole in the lamin meshwork (Fig. 4, arrows). This feature was also visible in nuclei from unheated cells from FPLD patients and from heated control fibroblasts (Fig. 4), and may correspond to the origin of nuclear invaginations, evaginations or transnuclear channels (Fricker et al., 1997). The honeycomb aspect of the lamina, rarely observed in nuclei from unheated cells from the patient with the LaA/C R482Q mutation, became prominent in several nuclei from the same cells after heating (Fig. 4, right column, asterisks). The survival rate of heat-shocked fibroblasts from control and patient KN (LaA/C R482W) was evaluated. Compared with unheated cells, the proportion of living cells 48 hours after heat shock was 42% for KN fibroblasts and 67% for control fibroblasts. Therefore, the post-heat-shock mortality was almost twice as high in fibroblasts from FPLD patient than in control cells, showing an hypersensitivity of nuclei to stress in FPLD.
NE proteins are easier to extract from nuclei of FPLD patients than from control nuclei
The relative amounts of A- and B-type lamins, emerin and LAP2β were compared by immunoblotting of whole-cell extracts from control individuals and from the three FPLD patients. Similar amounts of the four NE proteins were present in both types of cells (Fig. 5A). We sequentially extracted cells with Triton-X100, then Triton and salt, followed by chromatin digestion by DNase 1 and RNase A and a final extraction by salt. The material solubilized at each step (S1, S2 and S3, respectively) and the final insoluble material (Ins.) was then analyzed by immunoblotting. Fig. 5B shows that in control cells, A- and B-type lamins, emerin and LAP2β were not extracted under these three conditions and remained in the insoluble fraction. By contrast, in cells from FPLD patients, the four proteins were partially extracted roughly in parallel in S1 and S2 fractions, resulting in an insoluble depleted fraction. These results were similar in the three patients studied. Therefore, the LaA/C R482Q/W mutations do not affect the nuclear content in lamins, emerin and LAP2β but increase their extractibility. From the parallel extraction of A-type lamins and emerin, it is likely that the A-type lamin-emerin interactions were not impaired by these mutations.
Lamina disorganization can be induced by ectopic expression of LaA R482W in human control fibroblasts
To determine whether the ectopic expression of LaA R482W was sufficient to generate the lamina abnormalities observed in fibroblasts from FPLD patients, we transfected primary cultures of control human fibroblasts (which express LaA/C) with expression vectors carrying either FLAG-wild-type prelamin A or FLAG-R482W prelamin A. Prelamin A is a precursor form of mature lamin A, which is processed by endoproteolytic cleavage of the final 18 amino acids (Weber et al., 1989; Sinensky et al., 1994). Twenty nine hours after transfection, cells were processed for immunofluorescence with anti-FLAG and anti-lamin of the B-type antibodies. The distribution of lamins and chromatin was different in some fibroblasts expressing the R482W isoform (Fig. 6D-G) compared with both cells expressing ectopic wild-type lamin A (Fig. 6A-C) and untransfected fibroblasts (Fig. 6A-C, arrow). Ectopic wild-type lamin A was mostly located at the NE, where it colocalized with B-type lamin (Fig. 6A-C). By contrast, some cells expressing ectopic LaA R482W had dysmorphic budding nuclei, with a disorganized lamin A network with holes present in the buds and at some poles (Fig. 6E,H). These buds often also had reduced staining for lamin B (Fig. 6F,I), and reduced chromatin staining (Fig. 6D,G, arrowhead), similar to the phenotypes of cells from FPLD patients. Owing to the low efficiency of transfection of human fibroblasts (2-4%), it was impossible to quantify the percentage of dysmorphic nuclei in cells expressing ectopic lamin A. The prevalence of aberrant phenotypes following expression of ectopic WT or mutated lamin A has been measured in other cell systems (C. Favreau et al., unpublished).
The pathophysiological basis of the diseases associated with mutations in LMNA gene remains elusive. In an initial attempt to understand the cellular basis of FPLD, we have analyzed some of the nuclear modifications present in the skin fibroblasts from three patients bearing heterozygous R482Q or R482W mutations in lamins A and C.
Abnormalities of lamin networks in fibroblasts from FPLD patients
Immunofluorescence analysis revealed the presence in fibroblasts from FPLD patients of a subpopulation of cells with dysmorphic budding nuclei surrounded by a layer of lamins A and C. No aggregation of A-type lamins was observed in the cytoplasm or in the nucleoplasm. However, defects in the lamin A/C meshwork were often present, mainly in the buds, but also sometimes at poles of normally shaped nuclei. These alterations in lamin A/C distribution generated either a honeycomb structure or a few large holes. A striking additional modification of the lamina structure in this subpopulation of cells was the partial or total loss of B-type lamins around nuclear buds or at some nuclear poles, contrasting with their normal localization and density in the rest of the nuclei. Although nuclear abnormalities were qualitatively similar in the patients with the different mutations, some quantitative differences were apparent. The patient with the LaA/C R482Q mutation had numerous (10-25%) large dysmorphic nuclei, whereas nuclei from the patients bearing the R482W substitution had a more frequent disorganization of the A-type lamin network (5-50%). These differences in cellular disorders were of no prognostic value because the severity of the disease was similar in the three patients of our study. Owing to our lack of knowledge concerning A- and B-type lamin polymerization in vivo (Herrmann and Aebi, 2000), it was difficult to assess if the loss of B-type lamins in some membrane domains of FPLD nuclei was due to a decrease in the concentration of B-type lamins in lamin heteropolymers, or to a low density or disappearance of B-type homopolymers.
Inner nuclear membrane proteins, NPCs and chromatin abnormalities
In addition to lamina disorganization, other components of the NE were altered in this subset of fibroblasts from FPLD patients, including the INM, nuclear pore complexes and adjacent chromatin. Emerin strictly colocalized with A-type lamins, even when the lamina network was severely altered, suggesting that both molecules were remaining tightly associated. This observation is in agreement with (1) the lamin A-dependent localization of emerin in the nuclear envelope of LMNA(–/–) mice fibroblasts expressing either wild-type or R482W mutated lamin A (Sullivan et al., 1999; Raharjo et al., 2001), and (2) the ability of LaA R482Q to interact with emerin in vitro (Holt et al., 2001). LAP2β, which, like emerin is a member of the LEM family of INM proteins (Lin et al., 2000), has been reported to specifically bind lamin B1 (Foisner and Gerace, 1993). Therefore, one would have expected a parallel loss of both proteins in the nuclear envelope areas lacking B-type lamins. However, the loss of LAP2β in these membrane domains was less frequently severe than that of lamin B1. This result suggested that component(s) other than lamin B1 may also play a role in targeting LAP2β to the INM, or more trivially, that the antibody against LAP2β had a higher affinity for its antigen than the antibody against LB1.
Nuclear pore complexes are tightly associated with the lamina (Dwyer and Blobel, 1976), thus we were expecting an alteration in their distribution in dysmorphic nuclei from FPLD patients. As for LAP2β, the presence of nucleoporins Nup153, p62 and Nup214 in blebbing membranes was either normal, diminished or abolished, suggesting that some NE domains may be impoverished or depleted in NPCs. The requirement of an intact nuclear lamina for a normal NPC localization in the NE has been emphasized by other experimental data including, (1) the polar defects in NPCs observed in nuclei from LMNA(–/–) mice (Sullivan et al., 1999) and from mouse cells overexpressing lamin A mutated in the rod domain (Östlund et al., 2001), (2) the NE clustering of NPCs in Drosophila and C. elegans cells depleted of B-type lamins (Lenz-Bohme et al., 1997; Liu et al., 2000), or in cultured cells undergoing apoptotic cleavage of lamins (Lazebnik et al., 1993; Buendia et al., 1999).
The intensity of DAPI staining in nuclear buds was significantly weaker than that measured in other parts of the nuclei, suggesting that chromatin was decondensed in areas with disorganized NE domains. Despite these nuclear pleiotropic disorders, cell-cycle analysis by FACS and other methods showed that fibroblasts from FPLD patients, were euploid, normally cycling and not apoptotic, as control fibroblasts.
Ectopic expression of lamin A R482W provokes nuclear abnormalities similar to that observed in FPLD fibroblasts
By expressing LaA R482W in control human fibroblasts, we were able to generate alterations in nuclear shape and in A- and B-type lamin networks similar to that observed in fibroblasts from FPLD patients. This is a strong argument in favor of the involvement of mutated lamin A at the origin of the nuclear abnormalities observed in FPLD. Lamina disorganization can also be induced by expressing ectopic LaA R482W in mouse cells (C. Favreau et al., unpublished). Furthermore, the prevalence of aberrant nuclei in response to expression of mutated LaA/C appeared cell-type specific (Raharjo et al., 2001) (C. Favreau et al., unpublished), consistent with the tissue-sensitivity of human diseases linked to LMNA mutations.
The nuclear abnormalities in cells from FPLD patients reported here are reminiscent of those present in fibroblasts from LMNA(–/–) mice. Similarities include polar losses in B-type lamins, in LAP2β and in NPCs, as well as chromatin disorganization (Sullivan et al., 1999). Our finding that a subset (5-22%) of FPLD cells had nuclear structural defects is also similar to findings by Ognibene et al. (Ognibene et al., 1999), who found an abnormal distribution of lamins and chromatin in nuclear poles of a subset of fibroblasts from a patient with an emerin-null mutation.
Mechanical properties of NEs are altered in fibroblasts from FPLD patients
Putative functional consequences of the altered nuclear lamina structure were investigated by checking the resistance to heat shock of fibroblasts from FPLD patients. Compared with control fibroblasts, extensive modifications were observed in nuclei from patients, with the appearance in the NEs of extensive folds, invaginations, evaginations and transnuclear channels. Nuclear channels are common features present in a variety of living cells in culture (Fricker et al., 1997); however, those observed by immunofluorescence in heat-shocked cells from patients were of an unusual large size. Similar large channels were also observed in cells expressing ectopic lamin A or LBR (Broers et al., 1999; Ellenberg et al., 1997), showing that these particular NE alterations may occur in response to different modifications in NE composition. Finally, heat-shocked cells from FPLD patients had a twofold reduction in their survival rate compared with control cells. Therefore, the mechanical properties of the NEs, which play a crucial role under stress conditions (Krachmarov and Traub, 1993; Zhu et al., 1999), may be impaired by missense mutations in lamins A and C. In that respect, lamins would be similar to other members of the intermediate filament proteins family (Morley et al., 1995).
The modifications of the mechanical properties of the NEs in fibroblasts from FPLD patients were also attested by their low resistance to extraction by nonionic detergents and salt (Ellis et al., 1998; Fey et al., 1984; Gerace and Blobel, 1980). A significant fraction of A- and B-type lamins, emerin and LAP2β were solubilized from patients’ nuclei under conditions that do not remove these components from control nuclei. If this diminished resistance to extraction was restricted to dysmorphic nuclei or also affected nuclei with a normal shape is presently unknown. Finally, the parallel extraction of A-type lamins and emerin, together with their strict colocalization in immunofluorescent studies, further confirmed that the R482Q/W mutation does not affect the interaction of these two NE components.
In summary, our initial attempts to unravel the cellular pathophysiology of FPLD in cultured skin fibroblasts of three patients bearing the R482Q and R482W mutations in lamins A and C, has revealed abnormalities in the shape of a subset of nuclei, associated with an unusual NE organization, composition and fragility. Similar nuclear alterations probably occur in adipocytes from FPLD patients, impairing essential functions of adipose tissue and leading to lipodystrophy and its metabolic consequences. Further work on adipocytes will be necessary to understand the cellular specificity of this lamina disorganization. The evidence for mechanical interactions between nuclear scaffolding proteins, cytoskeletal filaments and integrins (Maniotis et al., 1997) may provide a clue to understanding the amplification at the tissue level of this intranuclear structural disorganization.
We acknowledge the patients whose cells have been used in this study. We are grateful to R. Piquemal, P. Valensi, M. C. Vantyghem, R. Levet, P. Levan, B. Lormeau and J. Magré for referring to us the patients and/or providing control and FPLD skin fibroblasts. We thank H. J. Worman and B. Burke for sharing their unpublished manuscripts. We thank B. Burke, R. Bastos and C. Johanet for the gift of antibodies, and H. J. Worman and C. Östlund for providing wild-type and mutated lamin A cDNAs. We thank G. Bonne, I. Callebaut, P. Collas, I. Duband-Goulet, I. Dunia, C. Favreau, P. Hossenlopp, A. Muchir, M. F. Portnoï, K. Schwartz, and S. Zinn-Justin for helpful discussion. We acknowledge P. Fontanges for help in picture acquisition, and M. Kornprobst, H. Kiefer and M. C. Gendron for cell-cycle analyses. This work was supported by INSERM and by grants from la Fondation pour la Recherche Médicale, l’Association pour la Recherche contre le Cancer, la Société Française d’Endocrinologie/bourse IPSEN/BIOTECH 2000, l’Association de Langue Française pour l’Etude du Diabète et des Maladies Métaboliques (ALFEDIAM)/Bourse Roche Diagnostics 2001 and l’Association Française de lutte contre les Myopathies.