Mechanistic insights into the cellular effects of a novel FN1 variant associated with a spondylometaphyseal dysplasia

2.1 Whole exome sequencing

For the affected patient and his parents, we performed whole exome sequencing using standard procedures. In brief, following DNA extraction from whole blood, sequencing was implemented on exon targets isolated by capture with the Agilent SureSelect Human All Exon V4 (50 Mb) Kit (Agilent Technologies, Santa Clara, CA, USA). DNA was sheared, ligated to adaptors, and purified for polymerase chain reaction (PCR) amplification, and products were captured according to the manufacturer's instructions. Sequences (100-bp paired-end) were generated on a HiSeq2500 (Illumina, San Diego, California, USA). Read alignment and variant calling were performed with DNAnexus (Palo Alto, California, USA) using default parameters with the human genome assembly hg19 (GRCh37) as reference. Probability calculations performed by 10 in silico prediction programs suggest that the identified FN1 mutation could have deleterious effects on protein function (Figure S2A in Appendix S1).

2.2 Mass spectrometry analysis of FN

Blood was drawn by standard methods, and anti-coagulated with ethylenediaminetetraacetic acid (EDTA). Plasma was separated by centrifugation and used to isolate plasma FN using gelatin-Sepharose affinity chromatography.19 FN was purified by a similar procedure from culture media conditioned by cells for 4 days. FN samples were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) under reducing conditions, bands were visualized with Coomassie Brilliant Blue dye, and excised gel bands were submitted for mass spectrometry (MS). Protein in gel bands was digested using 1.5-μg trypsin (Promega, Madison, WI, USA).20 Samples were dried completely in a speedvac and resuspended with 15 μL of 0.1% formic acid pH 3. 5 μL was injected per run using an Easy-nLC 1200 UPLC system (Thermo Scientific, Waltham, MA, USA). Samples were loaded directly onto a 40-cm-long 75-μm-inner-diameter nano capillary column packed with 1.9-μm C18-AQ (Dr. Maisch, Ammerbusch, Germany) mated to metal emitter (Thermo Scientific) in line with Thermo Fusion Lumos. The mass spectrometer was operated in data dependent mode with the 120 000-resolution MS1 scan (400-1800 m/z) in the Orbitrap followed by up to 20 MS/MS scans in the ion trap. Dynamic exclusion list was invoked to exclude previously sequenced peptides for 120 seconds if sequenced within the last 30 seconds.

Raw files were searched using MSAmanada,21 Sequest HT,22 and Byonic23 algorithms and validated using the Percolator algorithm24 within the Proteome Discoverer 2.1 suite (Thermo Scientific). 10 ppm MS1 and 0.6 Da MS2 mass tolerances were specified. Caramidomethylation and propionamidation of cysteine and oxidation of methionine were specified as dynamic modifications. The resulting msf file was used to construct a spectral library (percolator peptide q-value > 0.95) and extract MS1 ion chromatographs in Skyline for analysis.25, 26 Peptide abundance was calculated as the sum of areas-under-the-curve of MS1 chromatographs (intensity of peptide signal vs retention time) for peptides containing either the WT C amino acid and the modifications described, or variant W amino acid and modifications. The abundance of C97W peptides was adjusted for the relative ionization efficiency of C97W compared to WT peptides, as determined by relative abundance of common peptides from recombinant proteins. The percent of mutant FN was calculated as the adjusted abundance of C97W peptides divided by the sum of adjusted C97W and WT peptide abundances in that sample (Figure S2D in Appendix S1).

2.3 Isolation and culture of dermal fibroblasts

A skin biopsy was obtained from the affected patient, and FN^C97W/+ dermal fibroblasts were isolated from the biopsy by standard procedures. FN^C97W/+ cells were maintained in minimal essential medium (MEM) with 20% fetal bovine serum (FBS, Hyclone Logan, UT, USA). GM03349 dermal fibroblasts were obtained from the NIGMS Human Genetic Cell Repository at the Coriell Institute for Medical Research (Camden, NJ, USA). These cells were isolated from an apparently healthy 10-year-old male and were maintained in MEM with 15% FBS and non-essential amino acids. For experiments, cells were seeded in media prepared with serum depleted of FN by gelatin-Sepharose chromatography.19

2.4 Immunofluorescence staining and imaging

Cells grown on coverslips were fixed in 3.7% formaldehyde in PBS and, as noted for certain experiments, cells were permeabilized by treatment with 0.5% TritonX-100. FN was visualized using R184 rabbit anti-FN antiserum (raised against III_1-6 domains of FN) and Alexa Fluor 488- or Alexa Fluor 568-conjugated goat anti-rabbit IgG in 2% ovalbumin in phosphate-buffered saline (PBS). Co-staining with BiP was performed sequentially, and visualized with an anti-BiP antibody27 and Alexa Fluor 488-goat anti-mouse IgG. Matrix assembly was measured as the total mean fluorescence value for a field of cells. For each condition, 10 images were taken across two biological replicates, minimum value was subtracted out as background signal, and Student's t test was applied for statistical significance. Representative images were adjusted equivalently using FIJI.

2.5 Immunoblotting

Cells were lysed in SDS buffer (4% SDS, 20-mM Tris-HCl pH 8.8, 2-mM EDTA, and protease inhibitor cocktail [Roche Indianapolis, IN, USA]). For cell lysates, total protein concentration was measured by BCA assay (bicinchoninic acid assay, Pierce Chemical Co., Rockford, IL, USA), and an equal amount of total protein was loaded. Equal protein loading was confirmed by probing blots with an anti-GAPDH antibody (Cell Signaling Technology, Danvers, MA, USA). For culture media samples, cells were tryspinized and counted, and loading volume was adjusted for cell number. Normalized samples were reduced and separated by SDS-PAGE, transferred to nitrocellulose, then detected by immunoblotting using mouse anti-human FN hybridoma supernatant (HFN7.1; Developmental Studies Hybridoma Bank, University of Iowa, Iowa City, Iowa) and HRP (horseradish peroxidase)-conjugated goat anti-mouse IgG (Pierce Chemical Co.). To biotinylate ECM proteins, cells were incubated in freshly prepared 1 mg/mL sulfo-hydroxysuccinimidobiotin (Pierce Chemical Co.) in PBS for 30 minutes, followed by quenching with five washes in 25-mM Tris-buffered saline, pH 7.45. Cells were lysed in SDS buffer, reduced, separated, and transferred as described, and biotinylated proteins were detected using HRP-conjugated Streptavidin. FN is the most abundant protein in fibroblast ECM in culture,28 such that it is the most prominent band detected by this analysis. Blots were developed using ECL (enhanced chemiluminescence substrate, Thermo Scientific), films were scanned, and bands within the linear range were quantified using FIJI.29 Samples from at least three biological replicates were analyzed, and Student's t test was applied for statistical significance.

3.1 Case report

The affected male is the fifth of seven children born to healthy, unrelated parents, with no family history of skeletal dysplasia. Mild intrauterine growth restriction was noted during pregnancy, which was otherwise normal. He was born by spontaneous vaginal delivery after 38 weeks of gestation, birth weight 2300 g, below third percentile. There were no developmental problems in early infancy. Around the age of 1 year, short stature was noted. He started independent walking at the age of 1 year and 3 months. Subsequently, there was progressive asymmetric bowing of the legs and he developed a waddling gait. When he first presented to us at age 4, his height was 87.5 cm, below third percentile.

Radiographic studies began at age 2 years 10 months and showed severe progressive kyphoscoliosis (Figure 1A) and bilateral genu varum (Figure 2A). Major vertebral changes were later noted, including multiple ovoid vertebral bodies (Figure 1B), anterior vertebral wedging, and platyspondyly, affecting the cervical and thoracolumbar spine. Pectus carinatum was noted, as a part of the thoracic spine deformity (Figure 1C). Peripheral metaphyseal bone fragments, or corner fractures, were noted on the radii at age 2 years 10 months (Figure 2C, left), and in the knee (Figure 2A and B, left). Corner fractures resolved spontaneously by skeletal maturity (Figure 2B and C right). Avascular necrosis of the capitulum was noted incidentally, starting with capitulum surface irregularities and ultimately resulting in total bilateral capitulum necrosis and intra articular loose body formation (Figure 2D). The patient exhibits painless although limited elbow extension (Figure 2D) in the presence of a full range of flexion and prosupination.

The patient underwent multiple surgeries to address long bone dysplasia, and in all cases the post-operative course, bone formation following osteotomy, and healing processes were uneventful. This observation suggests that subperiosteal and endosteal bone formation remains unaffected. At age 6 years, he underwent bilateral corrective osteotomy and lengthening of both his tibiae, resulting in significant correction to the bow-leggedness on the right and complete correction on the left. At age 11 years 10 months, a tension band plate was inserted in the proximal right tibia and removed 18 months later to resolve residual tibial varus. At age 16 years 5 months, his height was below first percentile for age. Both tibiae were lengthened again, and he gained 7 cm in height, although his final height is 143 cm at skeletal maturity, below first percentile. No pathologic fractures were reported, suggesting that the bone is healthy once formed, and that remodeling and the endosteum are unaffected.

Joints of the hips, ankles, shoulders, feet, and hands were not affected and continued to develop with a normal radiological appearance, as did the skull, small tubular bones (phalanges) of the fingers and toes, and pelvis (Figure S1 in Appendix S1). Intramembranous bone formation appears to be relatively unaffected, as shown in the relatively normal head circumference (25th percentile in circumference at age 13 years). Shoulders, wrists and hands have a full range of motion.

3.2 Identification and characterization of the FN1 mutation

Whole exome sequencing of affected patient RNA yielded 50.5 million mapped reads with 85-fold mean coverage. Following reads alignment, variant calling, and filtration, 29 remaining variants were heterozygous, and all but one were inherited from one of the parents. The remaining variant was Hg19.Chr2: 216298171G>C, NM_212482.2. c.291C>G p.Cys97Trp in FN1 (MIM 135600). We verified the finding by Sanger sequencing in the affected patient, and confirmed its absence from the samples obtained from his parents. The variant was neither present in any of the 138 600 individuals whose exome analyses were deposited at gnomAD nor within an in-house database comprising more than 2000 exomes. This variant resides in exon 3 of 46, and replaces a highly conserved cysteine at position 97 of 2478 with a tryptophan (FN^C97W). In WT FN, C97 forms a disulfide bond with C125.30

To determine the effects of FN^C97W on cell functions, primary dermal fibroblasts were derived from the affected patient (FN^C97W/+ cells). Genomic DNA and total RNA were extracted from these cells and used as templates for PCR or reverse transcription-PCR amplification, respectively. Sequencing of PCR products confirmed that FN^C97W/+ cells are heterozygous for the mutation at the genomic and mRNA levels (Figure S2B,C in Appendix S1). To determine the proportions of secreted WT and mutant FN's, FN isolated from FN^C97W/+ cell culture medium was analyzed by MS. Peptides containing the single mutant amino acid change were detectable, but were significantly less abundant than WT peptides (~88% WT, 12% mutant; Figure S2D in Appendix S1).

MS was also used to determine the abundance of mutant FN relative to WT in vivo. Plasma FN was isolated from the blood of the patient. Plasma FN is produced and secreted by hepatocytes and thus is a source that is distinct from dermal cells. Peptides containing the single mutant amino acid change were detectable, but were significantly less abundant than WT peptides (~77% WT, 23% mutant; Figure S2D in Appendix S1).

3.3 Effects of the FN^C97W mutation on FN matrix assembly

The FN^C97W mutation resides in the N-terminal assembly domain that mediates FN-FN interactions essential for FN fibril formation.11 To compare the impact of this mutation on the phenotype of FN^C97W/+ cells, we used age- and gender-matched FN^+/+ dermal fibroblasts (GM03349; Coriell Repository). FN^+/+ and FN^C97W/+ cells show similar morphologies (Figure S3A in Appendix S1) and grow at similar rates (Figure S3B in Appendix S1). We compared these cells for the ability to assemble an FN matrix using indirect immunofluorescence staining. FN^C97W/+ cell matrix had visibly fewer fibrils than the matrix assembled by WT cells (Figure 3A). Quantitative analysis of the mean fluorescence intensity indicates that the FN^C97W/+ cell matrix is only 70% of the FN matrix of the WT control. To measure total FN protein in the matrix, ECM proteins were labeled in situ with a membrane-impermeable biotin cross-linking agent and FN levels were then quantified in total cell lysates (Figure 3B). FN at ~250 kDa is the predominant ECM protein produced by cultured fibroblasts,28 and this band exactly co-migrates with FN by immunoblot of identical samples run side-by-side on the same blot (data not shown). Quantification of FN levels showed that FN in FN^+/+ cell ECM is increased at least 4.5-fold compared to FN in the FN^C97W/+ ECM. These ECM analyses show that mutant cells have a defect in FN matrix assembly. This assembly defect is not due to deficient cell assembly mechanisms because FN^C97W/+ cells assemble normal levels of FN matrix when provided with an exogenous supply of WT FN (data not shown).

To investigate the mechanism that limits matrix assembly, we analyzed the distribution of mutant vs WT FN subunits by FN^C97W/+ cells. FN^+/+ cells secreted at least 3.8-fold higher levels of FN into the culture medium (Figure 4A). Combined with the MS analyses of secreted FN, these results show that the mutant protein is secreted but at significantly lower levels than the WT FN, leading to an overall lower level of total FN in the cell culture medium.

Mutations in ECM proteins can affect their ability to transit through the secretory pathway.9 Therefore, we compared the amounts of intracellular FN in WT and mutant cells. Cell lysates were prepared with subconfluent cells that had not assembled a FN matrix, such that the majority of FN was intracellular, not assembled into an ECM. Immunoblots of these lysates with anti-FN antibodies show dramatically higher levels of FN in the mutant lysates than in WT cells (Figure 4B). Mass spectrometric analysis of the intracellular FN fraction of FN^C97W/+ cells showed that, in contrast to FN in cell culture medium and plasma, the vast majority of the intracellular FN contained the mutant peptide (~95% mutant, 5% WT; Figure S2D in Appendix S1). To visualize differences in intracellular FN, subconfluent cells were permeabilized and immunostained. FN^C97W/+ cells show a different pattern of intracellular FN from FN^+/+ cells (Figure 4C). The intensity of intracellular FN staining is dramatically higher in mutant cells than in WT, which shows weaker, more diffuse intracellular staining. In addition, WT cells have extracellular FN fibrils that are visible between adjacent cells, but fibrils were not detected in subconfluent FN^C97W/+ cell cultures (Figure 4C, compare top left with top right). Co-staining with antibodies against BiP, an endoplasmic reticulum (ER) resident protein,27 shows colocalization of FN with BiP throughout FN^C97W/+ cells (Figure 4C), indicating that all portions of the ER contain FN^C97W/+ and suggesting a deficiency in FN secretion by mutant cells. In contrast, FN^+/+ cells had less FN within the ER and regions that stained heavily for BiP but not FN (Figure 4C). When viewed at higher magnification, large intracellular inclusions containing FN are clearly visible in FN^C97W/+ cells, whereas FN-positive structures within FN^+/+ cells were small and vesicular (Figure 4D). This phenotype is similar to the FN staining pattern observed when cells were treated with inhibitors of secretion.31 Together, these results show that mutant FN is not efficiently transported through the secretory pathway and that this preferential retention leads to an accumulation of mutant FN subunits in vesicular structures. Reduced FN secretion is probably to contribute to the observed reduction in FN matrix assembly by FN^C97W/+ cells.