A microfibril assembly assay identifies different mechanisms of dominance underlying Marfan syndrome, stiff skin syndrome and acromelic dysplasias

Abstract

Fibrillin-1 is the major component of the 10-12 nm diameter extracellular matrix microfibrils. The majority of mutations affecting the human fibrillin-1 gene, FBN1, result in Marfan syndrome (MFS), a common connective tissue disorder characterised by tall stature, ocular and cardiovascular defects. Recently, stiff skin syndrome (SSS) and a group of syndromes known collectively as the acromelic dysplasias, which typically result in short stature, skin thickening and joint stiffness, have been linked to FBN1 mutations that affect specific domains of the fibrillin-1 protein. Despite their apparent phenotypic differences, dysregulation of transforming growth factor β (TGFβ) is a common factor in all of these disorders. Using a newly developed assay to track the secretion and incorporation of full-length, GFP-tagged fibrillin-1 into the extracellular matrix, we investigated whether or not there were differences in the secretion and microfibril assembly profiles of fibrillin-1 variants containing substitutions associated with MFS, SSS or the acromelic dysplasias. We show that substitutions in fibrillin-1 domains TB4 and TB5 that cause SSS and the acromelic dysplasias do not prevent fibrillin-1 from being secreted or assembled into microfibrils, whereas MFS-associated substitutions in these domains result in a loss of recombinant protein in the culture medium and no association with microfibrils. These results suggest fundamental differences in the dominant pathogenic mechanisms underlying MFS, SSS and the acromelic dysplasias, which give rise to TGFβ dysregulation associated with these diseases.

Figure 1.

Fibrillin-1 domain organisation and mutation sites in domains TB4 and TB5. (A) The fibrillin-1 domain organisation is dominated by calcium-binding EGF-like domains (white) interspersed with transforming growth factor β-binding protein-like (TB; blue) and hybrid (diagonal stripes) domains. Other regions include the fibrillin unique N-terminal (FUN; purple) and non-calcium-binding EGF-like (grey) domains, a proline-rich region (orange) and a conserved 2Cys domain (yellow). N- and C-terminal propeptides (black) are processed by furin before microfibril assembly. (B) Structure of the cbEGF22-TB4-cbEGF23 fragment of fibrillin-1 (pdb 1UZJ) (6), with RGD integrin binding site indicated, showing the sites affected by SSS substitutions C1564S and W1570C (cyan spheres, left panel) and MFS substitution C1564Y (magenta spheres, right panel). In essence, residues important for stabilising TB domain fold are affected. (C) Homology model of the cbEGF24-TB5-cbEGF25 region of fibrillin-1, showing sites affected by geleophysic and acromicric dysplasias (light orange and dark orange spheres, respectively, left panel) and MFS substitutions (magenta spheres, right panel). All are predicted to affect the fold of the domain. The cbEGF24-TB5-cbEGF25 model was created with Modeller software using the cbEGF22-TB4-cbEGF23 structure as a template (10). Figures were rendered using PyMOL (Schrödinger, LLC) to show cbEGF domains (green), TB domains (blue), calcium ions (red) and disulphide bonds (yellow).

Figure 2.

Secretion profiles of GFP-Fbn constructs with disease-associated substitutions. Mutations associated with MFS, SSS, geleophysic dysplasia (GD) or acromicric dysplasia (AD) in domains TB4 and TB5 were engineered into a GFP-tagged fibrillin-1 construct (36) and the resulting constructs used to transiently transfect HEK293T cells. After 3 days in culture, samples of the medium and cell lysates were analysed by western blotting following separation on a reducing 6% SDS-PAGE gel, using an anti-GFP antibody. SSS, GD and AD mutants were detected in the culture medium, in contrast to the MFS mutants. Empty vector (pcDNA) and the wild-type construct (GFP-Fbn) were used as negative and positive controls. Cell lysate samples showed that the lack of recombinant material in the media of the MFS mutants was not due to a loss of protein expression.

Figure 3.

Microfibril incorporation of mutant GFP-Fbn constructs. FS2 fibroblasts were co-cultured for 5 days with HEK293T cells transiently transfected to express GFP-Fbn (WT; Panel A) or GFP-Fbn variants associated with SSS (Panels B and C), geleophysic dysplasia (GD; Panels D–F), acromicric dysplasia (AD; Panels G–I) or MFS (Panels J–L). Co-cultures were then fixed and stained using an anti-GFP antibody as described previously (36). SSS-associated mutants C1564S and W1570C, as well as the geleophysic dysplasia and acromicric dysplasia mutants, produced recombinant microfibril networks that were indistinguishable from the wild type. Co-cultures expressing the MFS mutants showed no recombinant microfibril staining, consistent with the lack of recombinant material observed in medium samples by western blotting (Fig. 2). Bar = 100 μm. Staining for endogenous fibrillin-1, expressed by the FS2 cells, is shown in Supplementary Material, Figure S2.

Figure 4.

Construction of the NterPro-cbEGF18-26 fibrillin-1 fragments and fibroblast secretion assays. (A) NterPro-cbEGF18-26 is a fusion of the N-terminal region of fibrillin-1, up to the proline-rich domain (orange), with the cbEGF18-26 region, which encompasses domains TB4 and TB5. The smaller size of this construct allows it to be distinguished from full-length fibrillin-1, following reducing SDS-PAGE, in western blots of medium from stably transfected fibroblasts when developed with an antibody raised again the proline-rich domain. (B) Secretion profiles of TB4 and TB5 domain-mutant constructs associated with MFS, SSS, acromicric dysplasia (AD) or geleophysic dysplasia (GD). Medium and cell lysate samples from untransfected fibroblasts (MSU-1.1) and fibroblasts transfected with the wild-type control construct (WT) were run as controls, allowing the identification of the recombinant construct (fusion). The full-length fibrillin-1 band (Fbn-1) acts as an endogenously expressed loading control. In cell lysates, a band migrating at ∼230 kDa (*) is likely to be due to a degradation product of the endogenously expressed fibrillin-1.

Figure 5.

Model of the mechanisms leading to different dominantly inherited fibrillinopathies associated with fibrillin-1 domains TB4 and TB5. (A) Wild-type fibrillin-1 is secreted from cells and assembled to produce a microfibril network that influences cell signalling through integrins and cell-surface HSPGs. Normal matrix deposition of growth factors (green and yellow), and growth factor activation, is mediated by interactions between cells and microfibrils (curved arrows). (B) When one of the alleles carries a pathogenic mutation that results in intracellular retention of the mutant protein, fewer functional microfibrils are produced in all tissues. Altered tissue dynamics, and possibly a reduction in growth factor binding sites in the matrix, lead to a general increase in growth factor activation and the development of MFS, but with microfibrils maintaining normal interactions with integrins and cell-surface HSPGs. (C) Alternatively, the mutant proteins could be secreted and assembled into structurally normal microfibrils, with defective interactions between microfibrils and integrins and/or cell-surface HSPGs being affected by altered binding site presentation. The resulting alterations in signalling lead to a secondary upregulation in the production of extracellular matrix. In these cases, the resulting phenotype is limited to tissues expressing the cell-surface proteins most sensitive to the altered binding sites on the microfibril, as seen in SSS and the acromelic dysplasias.