Rare heterozygous GDF6 variants in patients with renal anomalies

Although over 50 genes are known to cause renal malformation if mutated, the underlying genetic basis, most easily identified in syndromic cases, remains unsolved in most patients. In search of novel causative genes, whole-exome sequencing in a patient with renal, i.e., crossed fused renal ectopia, and extrarenal, i.e., skeletal, eye, and ear, malformations yielded a rare heterozygous variant in the GDF6 gene encoding growth differentiation factor 6, a member of the BMP family of ligands. Previously, GDF6 variants were reported to cause pleiotropic defects including skeletal, e.g., vertebral, carpal, tarsal fusions, and ocular, e.g., microphthalmia and coloboma, phenotypes. To assess the role of GDF6 in the pathogenesis of renal malformation, we performed targeted sequencing in 193 further patients identifying rare GDF6 variants in two cases with kidney hypodysplasia and extrarenal manifestations. During development, gdf6 was expressed in the pronephric tubule of Xenopus laevis, and Gdf6 expression was observed in the ureteric tree of the murine kidney by RNA in situ hybridization. CRISPR/Cas9-derived knockout of Gdf6 attenuated migration of murine IMCD3 cells, an effect rescued by expression of wild-type but not mutant GDF6, indicating affected variant function regarding a fundamental developmental process. Knockdown of gdf6 in Xenopus laevis resulted in impaired pronephros development. Altogether, we identified rare heterozygous GDF6 variants in 1.6% of all renal anomaly patients and 5.4% of renal anomaly patients additionally manifesting skeletal, ocular, or auricular abnormalities, adding renal hypodysplasia and fusion to the phenotype spectrum of GDF6 variant carriers and suggesting an involvement of GDF6 in nephrogenesis.

determine the expression of Gdf6 in the development of the murine urogenital system, RNA in situ hybridization analysis on sections of the kidney and the bladder of NMRI wildtype embryos from E11.5 to E18.5 was carried out following a standard protocol (Moorman et al. 2001). For each stage, at least three specimens were analyzed. Stained sections were documented using a Leica DM5000 microscope and a Leica DFC350FX digital camera (Leica Microsystems) and processed with Adobe Creative Cloud.

CRISPR/Cas9 genome engineering of mIMCD3 cells
A protocol by Ran et al. using the CRISPR/Cas9-mediated system for RNA-programmable genome editing (Ran et al. 2013) was modified to generate a Gdf6-knockout cellular model using mIMCD3 cells. Using a bioinformatic tool (http://www.crispor.tefor.net), the single guide RNA (sgRNA) target sequence 5'-AGT GAT CGT ATT AGC TGA CT-3' targeting exon 1 of Gdf6 was selected, and sense and antisense oligonucleotides were synthesized (Eurofins Genomics, Ebersberg, Germany) containing a 5'-CAC CG-3' cloning overhang at the 5'-end and a 5'-TGG-3' protospacer adjacent motif (PAM) site at the 3'-end. This sgRNA target sequence is predicted to have no exonic off-target binding regions. Complementary oligonucleotides were cloned into the pSpCas9(BB)-2A-GFP plasmid (Addgene plasmid #48138) containing the sgRNA scaffold and expression cassettes for Cas9 and GFP using a directional topoisomerase cloning protocol. Murine IMCD3 cells were transiently transfected with the resulting construct. GFP-positive cells were isolated 24 hours after transfection at the Cell Sorting Core Facility of Hannover Medical School using a MoFlo XDP cell sorter (Beckman-Coulter, Brea, MA, USA). DNA was extracted from 36 single cell clones after cell expansion using the innuPREP DNA Mini Kit (Analytik Jena, Jena, Germany), and sequence analysis of Gdf6 exon 1 was done using oligonucleotides listed in Supplementary Table 4. For allele-specific sequence analysis of clone 34, PCR amplicons of genomic DNA containing the Gdf6 exon 1 target locus were cloned into the pcDNA3.1 vector (Invitrogen, Carlsbad, CA, USA) using restriction enzyme-based cloning. After transformation, 20 Escherichia coli colonies were amplified, plasmid DNA was isolated using the NucleoSpin Plasmid Mini Kit (Macherey-Nagel, Düren, Germany), and Gdf6 exon 1 inserts were sequenced ( Supplementary Fig. 2). Consequences of nucleotide alterations were predicted using SnapGene Viewer (version 4.2.5; GSL Biotech, Chicago, IL, USA).

Cloning of GDF6 expression constructs and stable transfection of mIMCD3 cells
The human full-length GDF6 coding sequence was amplified by PCR from cDNA synthesized from total RNA of an individual without GDF6 variants using the SuperScript III First-Strand Synthesis Kit (Thermo Fisher Scientific). pUNO1-GDF6 expression constructs were generated by cloning of human GDF6 cDNA into the pUNO1-mcs vector (InvivoGen, San Diego, CA, USA) using the In-Fusion HD Cloning Kit (Takara Bio, Kusatsu, Japan) and the oligonucleotides listed in Supplementary  Supplementary Fig. 6).

Cell viability assay
Viability of mIMCD3 cells was measured using the CellTiter 96 AQueous One Solution Cell Proliferation Assay (MTS assay; Promega, Madison, WI, USA). Cells were seeded in 96-well plates at 1.2 x 10 4 cells per well and cultured for 24 h. After adding 20 µl of MTS solution and incubating at 37°C for 2 h, light absorbance at 490 nm was detected using the FLUOstar Omega Plate Reader (BMG Labtech, Ortenberg, Germany). Three independent experiments were performed per cell line and mean values were calculated.

Cell migration assay
Cell migration capacity was analyzed using a wound healing assay. Murine IMCD3 cells were cultured in 35 mm µ-Dishes with 3-well culture inserts (Ibidi, Martinsried, Germany). In each well, 3.0x10 4 cells were seeded and cultured for 24 h to reach full confluency. After removal of the culture insert, cells were gently washed with PBS (Merck) and treated with 30 µg/µl mitomycin C (Santa Cruz Biotechnology) in standard culture medium for the first 2 h to inhibit cell proliferation. Four images per well were documented initially and 8 h after removal of the culture inserts using a DM IL LED FLUO microscope (Leica Microsystems) equipped with an EC3 camera (Leica Microsystems). For each image, the area of the cell free gap was measured using Fiji/ImageJ (Schindelin et al. 2012;Schneider et al. 2012) and mean values per well were calculated. Cell migration was determined as the difference between mean gap area at time points 0 and 8 h. For each cell line, mean values were calculated from three independent experiments, and cell migration relative to mIMCD3 cells was determined.

Knockdown and functional rescue experiments in Xenopus laevis
A translation-blocking morpholino oligonucleotide (MO) for Xenopus laevis gdf6 (5'-GGC TCC TGT ATG TAT CCA TTA GCG G-3') was designed by and ordered from Gene Tools (Philomath, Oregon, USA). Full-length human GDF6 was cloned into a VF10 vector using MluI and NotI cloning sites and linearized for GDF6 mRNA synthesis using SalI. The T7 RNA polymerase (Roche) was used for in vitro transcription with the mMESSAGE mMACHINE Kit (Ambion, Kassel, Germany). A random control MO (5'-CCT CTT ACC TCA GTT ACA ATT TAT A-3') from Gene Tools was used as control for the gdf6 MO and memRFP mRNA (plasmid kindly provided by J. Wallingford, Austin, Texas, USA) as control for GDF6 mRNA.
Xenopus laevis embryos were transferred to 2% Ficoll dissolved in 0.3x Marc's Modified Ringer's solution (MMR) for injections (Sive et al. 2000). Subsequently, 13.6 ng gdf6 MO in migration and, thereby, the optimal duration of a wound healing assay using these cells was determined at intervals of 2 h for a total time period of 10h. After 8 h, the gap size was reduced by 50%. Therefore, this time point was chosen for subsequent analyses (see Fig. 3 and Supplementary Fig. 5). Scale bar: 150 µm. Migration of Gdf6 +/-mIMCD3 cell clone 30 was significantly reduced after 8 h compared to Gdf6 +/+ mIMCD3 cell clone 2, but lesser so than migration of Gdf6 -/-mIMCD3 cell clone 32; results are mean ± SD from three independent experiments. Scale bar: 150 µm. Supplementary Fig. 6 Murine IMCD3 cell clone 32 (Gdf6 -/-) was stably transfected with pUNO1 expression constructs harboring no insert (empty), human GDF6 wildtype or mutant cDNAs after two weeks of selection using the antibiotic blasticidin S. Primers had been designed to amplify a fragment encompassing sequences of the pUNO1 backbone (pUNO1, bottom row) or of pUNO1 backbone and parts of GDF6 cDNA insert (pUNO1-GDF6 cDNA insert, top row) (Supplementary Table 7). PCR amplification was performed using genomic