Geleophysic dysplasia: novel missense variants and insights into ADAMTSL2 intracellular trafficking

Geleophysic dysplasia (GPHYSD1, MIM231050; GPHYSD2, MIM614185; GPHYSD3, MIM617809) is an autosomal disorder characterized by short-limb dwarfism, brachydactyly, cardiac valvular disease, and laryngotracheal stenosis. Mutations in ADAMTSL2, FBN1, and LTBP3 genes are responsible for this condition. We found that three previously described cases of GPHYSD diagnosed clinically were homozygote or compound heterozygotes for five ADAMTSL2 variants, four of which not being previously reported. By electron microscopy, skin fibroblasts available in one case homozygote for an ADAMTSL2 variant showed a defective intracellular localization of mutant ADAMTSL2 protein that did not accumulate within lysosome-like intra-cytoplasmic inclusions. Moreover, this mutant ADAMTSL2 protein was less secreted in medium and resulted in increased SMAD2 phosphorylation in transfected HEK293 cells.

Three distinct approaches were used to investigate the functional and three-dimensional (3D) structural effect of the missense variants: (1) measurement of the evolutionary conservation in the genome of vertebrate and mammalian species (i.e., phyloP, phastCons, GERP++), (2) prediction of the impact of amino acid substitutions in protein function (i.e., SIFT, Polyphen2, LRT, MutationAssessor, MutationTaster, fathmm-MKL, PROVEAN, MetaSVM, MetaLR, REVEL), and (3) prediction of protein 3D destabilization (i.e., PoPMuSiCv3.1, CUPSAT, I-Mutant v3.0, MAESTRO, INPS-3D). A consensus interpretation was established according to the number of tools predicting a damaging effect versus a benign or tolerated one. The consensus was given if at least 70% (functional impact: 8 of 11; 3D stability: 4 of 5) of predictors agreed in the variant classification. Otherwise, the results were considered as inconclusive. Moreover, we included functional annotations (e.g., ligand binding sites, catalytic residues, post-translational modifications of proteins, residues in protein-protein interaction interfaces) retrieved by the Structure-PPi system [21]. Structure-PPi also considers residues in physical proximity (at a 5Å distance) to amino acid changes found in human diseases. The co-localizing and co-clustering of somatic mutations and germline variants onto protein 3D-structure have been applied to link rare variants with functional consequence [22].
A pCMV-SPORT6 plasmid bearing human ADAMTSL2 cDNA was purchased from Thermo Fisher Scientific (Waltham, MA, USA). Mutations were introduced using QuickChange XL site-directed mutagenesis kit (Agilent Technologies, Wilmington, DE, USA) according to the manufacturer's instructions and confirmed by direct DNA sequencing. HEK293 cells were transfected with wild-type and mutated ADAMTSL2 constructs using TransIT-LT1 transfection reagent (Mirus Bio, Madison, WI, USA) according to manufacturer's instructions. 72 h after transfection, media and cells were collected for Western blotting with anti-ADAMTSL2 (#GTX-102069, GeneTex, Irvine, CA, USA) and anti-pSMAD2 (#3108, Cell Signaling Technology, Danvers, MA, USA) antibodies. Anti-Calnexin (#SPA860, Assays Design, Ann Arbor, MI, USA) was used for normalization. Analysis of protein band intensities were performed by Quantity One basic software (Bio-Rad laboratories, Hercules, CA, USA).

Electron microscopy
Fibroblasts were tripsinized and resuspended in glutaraldehyde for electron microscopy (EM). For immuno-EM analysis of ADAMTSL2 distribution, fibroblasts were fixed with a mixture of 4% paraformaldehyde (PFA) and 0.05% glutaraldehyde for 10 min at room temperature, then washed once with 4% PFA to remove residual glutaraldehyde and fixed again with 4% PFA for 30 min at room temperature. Cells were then incubated with blocking/permeabilizing mixture (0.5% BSA, 0.1% saponin, 50 mM NH 4 Cl) for 30 min and next with the primary polyclonal antibody against ADAMTSL2 diluted 1:500 in blocking/permeabilizing solution. The following day, cells were washed and incubated with the secondary antibody, the anti-rabbit Fab fragment coupled to 1.4-nm gold particles (diluted 1:50 in blocking/ permeabilizing solution) for 2 h at room temperature. Specimens were post-fixed as previously described [23] and after dehydration they were embedded in epoxy resin and polymerized at 60°C for 72 h. 60 nm-thin sections were cut at the Leica EM UC7 microtome. EM images were acquired from thin sections using a FEI Tecnai-12 electron microscope equipped with a VELETTA CCD digital camera (FEI, Eindhoven, The Netherlands). Quantification of ADAMTSL2-associated gold particles was performed using iTEM software (Olympus SYS, Germany). Total number of gold particles was counted within the cell and gold particles in the Golgi area reported as percent of total.

Statistical analyses
Statistical significance was computed using the Student's two tail test. A p value < .05 was considered statistically significant.

ADAMTSL2 variants in three GPHYSD individuals and ultrastructure studies
Direct ADAMTSL2 sequencing was performed in three previously published cases. Their clinical features are presented in greater details elsewhere [18][19][20]. In all three cases, either homozygous or compound heterozygous ADAMTSL2 variants were identified ( Table 1). The variants found in subject 1 c.2431G > A (p.Gly811Arg, rs113994124) localized in the spacer domain and c.1942C > T (p.Arg648Cys, rs1198735320) localized in the TSR3 domain are annotated in dbSNP and 1000Genomes, whereas the remaining variants including c.1943G > C (p.Arg648Pro) and c.1966G > A (p.Gly656Ser) in subject 2 both affecting TSR3, and the homozygous c.886G > A (p.Gly296Arg) variant of subject 3 affecting the spacer domain were not previously reported but they are clustered with two or more other previously reported GPHYSD-related variants on 3D models of ADAMTSL2 (Fig. 1). All five variants are not reported in controls of the Table 1 Summary of ADAMTSL2 variants. Exome Aggregation Consortium (ExAC) and Genome Aggregation Database (gnomAD) databases. However, the region between exons 9 and 18 of ADAMTSL2 has poor coverage in ExAC and gnomAD databases and better coverage in 1000Genomes. All variants except p.Gly296Arg were predicted to have a damaging effect (impaired functional predictions and destabilization of the 3D structure) (Supplementary Table 1). The variant p.Gly296Arg was classified as inconclusive according to the functional impact predictions and corresponding 3D structural predictions. Nevertheless, all variants affected conserved positions in vertebrate and mammalian species, as shown by phyloP, phastCons, and GERP++. Intracellular inclusion bodies were previously reported in GPHYSD cases [2,24,25] and we recently found that storage occurs within lysosomes in GPHYSD cells carrying FBN1 mutation [26]. Primary fibroblasts from skin biopsy available only for subject 3 were analyzed by EM and showed lysosomal-like vesicles with lamellar structure appearance and electron-dense storage material (Fig. 2) suggesting that inclusions are a feature of GPHYSD cells either carrying ADAMTSL2 or FBN1 mutations [26].

ADAMTSL2 carrying the p.Gly296Arg variant fails to traffic through the Golgi complex but does not accumulate in lysosome-like vesicles
ADAMTSL2 is a glycoprotein that traffics through the Golgi complex [7]. To investigate whether mutated ADAMTSL2 is correctly targeted to the Golgi, we performed on GPHYSD fibroblasts from subject 3 and control fibroblasts a confocal immunofluorescence analysis for ADAMTSL2 and TGN46, a trans-Golgi network (TGN) marker [27]. Control fibroblasts showed a punctate cytosolic pattern of ADAMTSL2 signals and a distinctive ADAMTSL2 accumulation around the nuclei, corresponding to TGN46 positive cisternae (Fig. 3A). In contrast, GPHYSD fibroblasts showed reduced ADAMTSL2 staining with distribution in cytosolic clumps lacking TGN46 co-localization (Fig. 3A). To confirm that mutated ADAMTSL2 protein fails to traffic though the Golgi, we performed immuno-EM with anti-ADAMTSL2 antibody in GPHYSD fibroblasts. GPHYSD fibroblasts showed a diffuse cytoplasmic ADAMTSL2 signal in contrast to wild-type fibroblasts exhibiting distribution of ADAMTSL2 in the Golgi apparatus with gold particles concentrated at the TGN area ( Fig. 3B-C). Moreover, intracytoplasmic  inclusions in GPHYSD fibroblasts did not appear to contain ADAMTSL2positive signals (Fig. 3D).

The p.Gly296Arg ADAMTSL2 variant impairs protein secretion and results in increased TGF-β signaling
We next investigated the functional consequences of the p.Gly296Arg variant of subject 3 by transient transfection of HEK293 cells with a plasmid bearing the mutated human ADAMTSL2 cDNA. Conditioned media and cell lysates were analyzed by western blot using an anti-ADAMTSL2 antibody. Consistent with other previously reported mutations [3], reduced levels of ADAMTSL2 protein were detected in conditioned medium of cells transfected with the mutant ADAMTSL2 construct compared to control cells expressing the wild-type ADAMTSL2 (Fig. 4A). Moreover, lysates of HEK293 cells transfected with mutated ADAMTSL2 showed increased phosphorylated SMAD2 (pSMAD2) compared to cells transfected with wild-type ADAMTSL2 (Fig. 4B). Taken together, these findings show that the p.Gly296Arg variant results in poor secretion of ADAMTSL2 and increased TGF-β signaling in transfected HEK293 cells.

Discussion
In this study, we identified five novel ADAMTSL2 variants in three published cases of GPHYSD [18][19][20]. Storage material staining positively for periodic acid-Schiff (PAS) staining in GPHYSD patients, has previously suggested glycoprotein accumulation [2,24]. Inclusions have been detected in multiple tissues including skin, liver, bone cartilage, trachea, and heart, and in primary skin fibroblasts. These inclusions have been described as lysosomal-like vacuoles containing granular material, lamellar structures or electron-dense material [14,[28][29][30]. Similar inclusions were also found in bronchial epithelial cells of Adamtsl2 −/− mouse, thus suggesting glycogen storage [8]. Moreover, we previously found that inclusions are also present in fibroblasts of GPHYSD patients carrying FBN1 mutations and Myhre syndrome (MIM139210) patients carrying SMAD4 mutations [26], suggesting that a common pathway is responsible for the formation of such inclusions.
Here, we showed that lysosomal inclusions do not appear to contain mutant ADAMTSL2.
We also showed that ADAMTSL2 with the missense variant p.Gly296Arg fails to localize in the Golgi and its secretion is impaired in contrast to wild-type ADAMTSL2 that traffics trough the Golgi complex and is efficiently secreted. Therefore, a combination of ADAMTSL2 secretion and pSMAD2 analyses as performed in this study in cell lines might be useful for validation of the pathogenicity of ADAMTSL2 variants without the need of obtaining skin fibroblasts. Moreover, these cell assays may be useful to search for drugs increasing ADAMTSL2 secretion which might have therapeutic potential. However, further studies on ADAMTSL2 degradative pathways can define the fate of mutated protein and its consequence on TGF-β signaling.  Fig. 4. Mutated ADAMTSL2 is poorly secreted and activates TGF-β signaling. (A) Western blotting for ADAMTSL2 on conditioned media and cell lysates of HEK293 cells transfected with wild-type (WT) and mutant ADAMTSL2 constructs. Quantification of ADAMTSL2 in media over cell lysates normalized for calnexin (CNX) is shown (n = 6; t-test: **p < .01). (B) Western blot of SMAD2 phosphorylation (pSMAD2), a downstream effect of TGF-β pathway activation, in HEK293 cells transfected with mutant or wild-type ADAMTSL2. CNX is shown as loading control (n = 6; t-test: *p < .05).