Progeroid syndrome patients with ZMPSTE24 deficiency could benefit when treated with rapamycin and dimethylsulfoxide

Abstract

Patients with progeroid syndromes such as mandibuloacral dysplasia, type B (MADB) and restrictive dermopathy (RD) harbor mutations in zinc metalloproteinase (ZMPSTE24), an enzyme essential for posttranslational proteolysis of prelamin A to form mature lamin A. Dermal fibroblasts from these patients show increased nuclear dysmorphology and reduced proliferation; however, the efficacy of various pharmacological agents in reversing these cellular phenotypes remains unknown. In this study, fibroblasts from MADB patients exhibited marked nuclear abnormalities and reduced proliferation that improved upon treatment with rapamycin and dimethylsulfoxide but not with other agents, including farnesyl transferase inhibitors. Surprisingly, fibroblasts from an RD patient with a homozygous null mutation in ZMPSTE24, resulting in exclusive accumulation of prelamin A with no lamin A on immunoblotting of cellular lysate, exhibited few nuclear abnormalities and near-normal cellular proliferation. An unbiased proteomic analysis of the cellular lysate from RD fibroblasts revealed a lack of processing of vimentin, a cytoskeletal protein. Interestingly, the assembly of the vimentin microfibrils in MADB fibroblasts improved with rapamycin and dimethylsulfoxide. We conclude that rapamycin and dimethylsulfoxide are beneficial for improving nuclear morphology and cell proliferation of MADB fibroblasts. Data from a single RD patient's fibroblasts also suggest that prelamin A accumulation by itself might not be detrimental and requires additional alterations at the cellular level to manifest the phenotype.

Keywords: severe postnatal growth retardation.

Figure 1.

Cellular proliferation measurement of fibroblasts from patients with mandibuloacral dysplasia (MAD) and restrictive dermopathy (RD). Fibroblasts were passaged continually, and at the cell passages some cells were assayed for the incorporation of bromodeoxyuridine (BrdU) in the proliferating cells. The proliferation was observed in the unaffected control and in the unaffected parents (4700.1 and 4700.2) of MAD pedigree, whereas affected subjects (4700.3, 4700.4, 3300.3, and 3300.5) showed reduced incorporation of BrdU with increasing passages. The fibroblasts from the patient with RD (500.3) appeared to have a similar proliferation rate as the control fibroblasts used in this study.

Figure 2.

Immunoblot analysis of prelamin A and mature lamin A and C in the fibroblasts from mandibuloacral dysplasia, type B (MADB) and restrictive dermopathy (RD) patients. (A) The presence of prelamin A in the fibroblasts from the affected patients MAD4700.3 and 4700.4 at earlier passages 11–13 and MAD3300.3. Note in RD500.3 only prelamin A could be detected. Shown also are the immunoblots for the ZMPSTE24 protein. (B,C) The quantification of lamins in these samples as estimated by Image J. The ratio of prelamin A to lamin C remains similar in unaffected controls and in parents of affected patients, MAD4700.1 and 4700.2. This ratio tends to increase in affected patients and is several-fold higher in RD500.3 (B). The ratio of lamin A/lamin C is also altered in affected subjects. Comparison for the ratio between lamin A and lamin C between these groups was also statistically significant (C). Symbols represent individual quantifications of blots presented in B and C; n = 6; P < 0.001; (AU) arbitrary units.

Figure 3.

Nuclear morphology of dermal fibroblasts from healthy control, mandibuloacral dysplasia, type B (MADB) patients, MAD4700.3 and MAD4700.4 and restrictive dermopathy patient RD500.3, in response to various treatments. Shown are the representative nuclear images for each treatment. Approximately 25–50 nuclei for each treatment were observed. Details of the treatments are described in Methods.

Figure 4.

Nuclear morphology of dermal fibroblasts from affected mandibuloacral dysplasia (MAD) and restrictive dermopathy (RD) patients in response to rapamycin dissolved in dimethyl sulfoxide (DMSO) and DMSO-alone treatment. (A) The indirect immunofluorescence images of nuclei stained for lamin A and lamin C proteins with and without DMSO (1%) and rapamycin (1 µM) treatment for 14 d. In contrast to control fibroblasts, patients with MAD (4700.3, 4700.4, and 3300.3) show significant instances of abnormal nuclear morphology that improved upon treatment with DMSO and rapamycin. We observed very few fibroblasts with abnormal nuclear morphology from the patient with RD. (B) Similar improvements in nuclear morphology of the MAD4700.3 patient were observed when fibroblasts were treated with rapamycin prepared in ethanol. Again, no changes were noted with RD500.3. Approximately 200 nuclei were observed for each treatment and an asterisk (*) denotes statistical significance of P < 0.001 between various groups. (NT) Untreated.

Figure 5.

Electron micrograph images of nuclei in fibroblasts obtained from patients with mandibuloacral dysplasia (MAD) after treatment with dimethyl sulfoxide (DMSO) and a farnesyltransferase inhibitor, FTI-277. Marked with red triangles are the “vesicle-like structures,” which could represent nuclear invaginations not seen in the unaffected nuclei. Such “vesicle-like structures” are not seen in fibroblasts treated with DMSO. FTI-277 treatment failed to inhibit these features in the patients with MAD. Scale bars are shown in the images. All images were within 0.5–2 µm.

Figure 6.

Schematic of yeast proprotein a-factor, human prelamin A, and total fibroblast proteins analyzed by two-dimensional-fluorescence difference gel electrophoresis (2D-DIGE). (A) The various amino acids and the peptide bonds cleaved by yeast CAAX prenyl protease/CAAX prenyl protease 2 (Ste24/Rce1) and putative protease (Axl1). The generation of mature yeast a-factor is marked. (B) The partial carboxyl-terminal amino acids for human prelamin A and the proteolytic sites for zinc metalloprotease (ZMPSTE24). The CAAX motif is marked where C is cysteine, AA denotes any two aliphatic residues, and X is any amino acid. The amino acids are numbered from the amino terminus. For more details see the review by Michaelis and Barrowman (2012). (C) The prelamin A protein spots as identified by mass spectrometry detection. Also shown are the fold changes between various combinations of affected and unaffected subjects as well as within the MAD4700 pedigree. Only protein spot 48 showed a decreased molecular weight (MW) in most samples, which could represent lamin A or lamin C, and the several-fold increase in protein spots 32, 33, and 37 could represent prelamin A. These protein spots may represent other posttranslational modifications including lamin phosphorylation, as discussed in the text. (D) The protein identified as vimentin. Only protein spot 61 showed increased expression under various group comparisons, but protein spots 75, 80, and 82 also identified as vimentin were decreased and are of lower MW.

Figure 7.

Indirect immunofluorescence images of fibroblasts treated with dimethyl sulfoxide (DMSO) or rapamycin and stained for vimentin. The fibroblasts from MAD4700.3 had irregular vimentin cytoskeletal architecture compared to the unaffected control subject. This abnormal vimentin cytoskeletal architecture improved dramatically and significantly upon DMSO and rapamycin treatments. Similar observations were made in restrictive dermopathy RD500.3 fibroblasts. Red false coloring represents vimentin, and blue represents DAPI (4′,6-diamidino-2-phenylindole, dihydrochloride). All images were obtained under similar microscopic conditions; scale bar, 10 µm; (*) P < 0.001, (**) P = 0.001, (***) P = 0.009. (NT) Untreated.

Figure 8.

Immunoblot analysis of endogenous vimentin in the fibroblasts from MAD4700 pedigree, affected patients from MAD3300.3 and RD500.3, and exogenous expression of Flag-hVimentin-V5 tagged mammalian construct in HeLa cells and in RD500.3 fibroblasts. (A) Total cell lysate resolved on sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and probed with vimentin antibody showed vimentin (VIM I) and an additional faster migrating band, VIM II. VIM II was not detected in MAD4700.4 and RD500.3. (B) Upon protein titration from RD500.3, the level of VIM II was consistently less than that of the unaffected control. (C,D) Quantification of VIM I and VIM II. The ratio changes in RD500.3 in favor of more VIM II. Only representative blots are shown. Symbols represent individual quantifications of blots presented in A and B. The loading control protein GAPDH was determined for each blot. The ratios for VIM I/VIM II between various groups are all statistically significant. P < 0.01; (AU) arbitrary units. (E) Amino acid sequence of human vimentin. The sequences highlighted in yellow were consistently detected on mass spectroscopy analysis. The boxed sequence was not detected in three out of four protein spots, whereas that in green was detected only once, suggesting proteolytic cleavage in this region of the vimentin. (F) Schematic of vimentin construct showing the Flag and V5 epitopes in the amino and carboxy terminus of the protein, respectively. (G) Immunoblot of tagged vimentin expressed in HeLa cells that correctly process the prelamin A. Probing with V5 antibody detects a faster migrating band similar to RD500.3 fibroblasts probed with vimentin antibody (A,B), indicating the amino terminus cleavage of the vimentin protein. However, when the same protein blot is stripped and reprobed with Flag antibody, only one protein band was detected; the expected smaller protein band was not detected. (H) A similar observation was made when the tagged vimentin was expressed in RD500.3 fibroblasts, which carry the nonfunctional ZMPSTE24 enzyme.