Novel pathogenic variants in CUBN uncouple proteinuria from renal function

Genomic DNA was obtained from whole blood using the QIAamp DNA Mini Kit (180134, Qiagen). 97 genes related to kidney disease were kept as a gene capture strategy, using the GenCap custom enrichment kit (MyGenostics Inc) following the manufacturer’s protocol. The enriched libraries were sequenced using an Illumina HiSeq 2000 sequencer (Illumina), which running for paired-end reads of 150 bp. The clean reads were aligned to the reference human genome (hg19) with Short Oligonucleotide Analysis Package (SOAP) aligner software (SOAP2.21; soap.genomics.org.cn/soapsnp. html). Afterwards, single nucleotide polymorphisms (SNPs) were annotated with the SOAPsnp program, and the deletions and insertions (InDels) were detected using Genome Analysis Toolkit software 3.7. Low-quality variations were filtered out using a quality score ≥ 20 and MAF ≤ 0.01 and the schematic of screening workflow were shown in Additional file 1: Fig. S1. All variants were verified by Sanger sequencing.

Vector pSPL3, known as the exon trapping vector, was carried out for minigene assay. Briefly, exon 44 of CUBN and its adjacent intron 43 and 44, was PCR-amplified using the following primers: forward 5′-accagaattctggagctcgagATTCATCTAT CAGAAACATGATATATT-3′ and reverse 5′-accagaattctggagctcgagCAATGAGAATAGATAAATGGTCTGGCA-3′. XhoI and NheI were chosen as restriction sites. The PCR products were inserted into the vector pSPL3 following the standard process with ClonExpress II one step cloning kit (C112, Vazyme). The mutant type was constructed according to the procedure by Mut Express II Fast Mutagenesis Kit V2 (C214, Vazyme). Wild-type and mutant types were transfected into HEK293T cells with lipofectamine 3000 (Life Technologies). RNA was harvested using the Steadypure Quick RNA Extraction Kit (AG21023, Accurate biology) at 24 h after transfection. Then cDNA synthesis was performed with HiScript III 1st strand cDNA Synthesis Kit (R312, Vazyme). Subsequently, cDNA was PCR-amplified using the following pSPL3 specific primers:SD6-5′-TCTGAGTCACCTGGACAACC-3′ and SA2-5′-ATCTCAGTGGTATTTGTGAGC- 3′. The PCR fragments were identified by Sanger sequencing to evaluate the alternative splicing.

Phylogenetic analysis of CUBN was used an online tool, the interactive tree of life (https://​itol.​embl.​de). Schematic of the Cubilin protein domains was analyzed by the SMART (a Simple Modular Architecture Research Tool). ClustalW multiple sequence alignment of Cubilin protein in several species was achieved by Clustal Omega (https://​www.​ebi.​ac.​uk/​Tools/​msa/​clustalo/​). Crystal structures of CUB domain were analyzed by the SWISS-MODEL (https://​swissmodel.​expasy.​org).

Electron microscopic sample handling and detection were performed by the electron microscopic core lab of Chongqing Medical University. TEM images were analyzed using Image Pro plus 6.0. Four glomeruli were randomly selected and ten electron micrographs were taken in each glomerulus.

Kidney biopsies and 293 T cells fixed in neutral buffered formalin were embedded in paraffin or optimal cutting temperature compound by using standard procedures. Frozen and paraffin sections were stained with immunofluorescence, respectively. Immunofluorescent staining and images were obtained by a Nikon A1R Meta confocal microscope. Cover slips were observed.

The antibodies used were list below: anti-Cubilin-C-terminal antibody (1:500, ab191073, Abcam), Rat Cubilin(CUBN) polyclonal antibody (1:100, 31010, Bicell Scientific), anti-Synaptop-odin antibody (1:50, 21064-1-AP, Proteintech), anti-Wilms Tumor Protein antibody (1:50, ab89901, Abcam), anti-COL4A3 antibody (1:100, Kingmed, Guangzhou, China), anti-COL4A5 antibody (1:100, Kingmed, Guangzhou, China), anti-Amnionless antibody (1:10, sc-365384, Santa Cruz), anti-Megalin Antibody (1:30, CD7D5, Novus Biologicals), goat polyclonal secondary antibody to mouse Alexa fluor 488 (1:400, ab150113, Abcam), goat polyclonal secondary antibody to rabbit Alexa fluor 555 (1:400, ab150078, Abcam), rabbit monoclonal to HA tag (1:500, ab236632, Abcam), goat polyclonal secondary antibody to rabbit Alexa fluor 647 (1: 400, ab150079, Abcam), goat anti-mouse Alexa fluor 568 (1: 400, ab175473, Abcam), DAPI (1:1000, C1002, Beyotime).

293 T cells were cultured in DMEM supplemented with 10% (v/v) FBS (Hyclone, 10100147) and 1% (v/v) penicillin/streptomycin (Beyotime, C0222) at 37 ℃and 5% CO² in a humidified atmosphere and passaged every 2-3 days.

Male BALB/c mice (20–22 g per mouse) and male C57BL6 mice (20–25 g per mouse) was kept under pathogen-free conditions at the Laboratory Animal Centre institution, Children’s Hospital of Chongqing Medical University (Chongqing, P.R China). After adaptive feeding for one week, BALB/c mice were injected by adriamycin (11 mg/kg, Meilunbio) through tail vein, and male c57 mice were administered intraperitoneally with Lipopolysaccharides (LPS, 12 mg/kg, Sigma-Aldrich). Control groups were received an equal volume of saline. BALB/c mice were anesthetized and sacrificed using isofluorane and euthanized by cervical dislocation at the fourth week after tail vein injection, while c57 mice were sacrificed in the same way at the 24th hour after intraperitoneal LPS injection. Kidney tissues were excised and fixed in 4% paraformaldehyde, embedded in paraffin. Paraffin-embedded sections were used to analyze the co-localization between Cubilin and Amn according to standard protocol. The experiment was approved by the Animal Ethics Committee of the Children’s Hospital of Chongqing Medical University (No. CHCMU-IACUC20220629011).

The plasmids pLVX-IRES-ZsGreen1 and pCMV-HA-N were digested by EcoR I, respectively, and a 10,869 bp of human CUBN gene, a 1359 bp of human AMN gene, linearized pLVX-IRES-ZsGreen1 and pCMV-HA-N were purified. Then, AMN and pLVX-IRES-ZsGreen1, CUBN and pCMV-HA-N were linked utilizing In-Fusion Cloning (Vazyme, ClonExpress II One Step Cloning Kit, C112) to generate shuttle recombinant plasmids pLVX-IRES-ZsGreen1-AMN and pCMV-HA-N-CUBN. The shuttle plasmids were identified by Sanger sequencing analysis. Site-directed mutagenesis of CUBN were performed using Mut Express MultiS Fast Mutagenesis Kit V2(Vazyme, C215) and also identified by Sanger sequencing analysis.

The day prior to transfection, the cells were seeded into 24-well plates at 1 × 10⁵ cells/well. The cells were transfected using Lipofectamine3000 (ThermoFisher, USA) according to the manufacturer’s instructions with 500 ng of respective plasmid DNA per well. After 6–7 h, the medium was exchanged with fresh medium.

Protein extracts were prepared and incubated with anti-bodies against HA or IgG for 24 h on a rotating wheel. Then, Pierce Protein A/G Magnetic Beads (ThermoFisher, USA) were added and incubated for another 24 h. After the beads were boiled, the precipitated proteins were separated by SDS-PAGE and transferred to PVDF membranes for further analysis. For western blotting, cell samples were extracted and quantified then boiled at 95 ℃, 10 min. Protein sample was separated on a 6% sodium dodecyl sulfate polyacrylamide gel electrophoresis gel then transferred on a polyvinylidene fluoride (PVDF) membrane. Incubating primary antibodies overnight at 4 ℃, with specific primary antibodies against HA (1:1000, ab236632, Abcam), Cubilin (1:1000, ab191073, Abcam), anti-Amnionless antibody (1:500, sc-365384, Santa Cruz) and β-tubulin (AB0039, 1:2000, Abways) in Tris-Buffered Saline Tween-20 (TBST) containing 5% skim milk. After washed for 3 times with TBST, the membranes were incubated for 1 h at room temperature with a respective IgG-HRP labeled second antibody (1:10,000) in TBST containing 5% skim milk. Antigens were revealed using a chemiluminescence assay (Pierce ECL Western Blotting Substrate, 32,209, ThermoFisher, USA) and quantification of bands was achieved by densitometry using the Image J software.

HEK 293 T cells were grown in 24-well plates containing coverslips (14 mm diameter) and cultured overnight. Then cells were treated with plasmid as described for transient transfection. Coverslips were washed with PBS twice and fixed in 4% paraformaldehyde for 15 min. Then coverslips were blocked with Duolink Blocking Solution for 60 min at 37 °C. The primary antibody HA and AMN, diluted in blocking solution, was added to the coverslips and incubated overnight at 4 °C. Then coverslips were washed with 1 × Wash Buffer A and subsequently incubated with Duolink^® PLA Probe (Duolink^® In Situ PLA^® Probe Anti-Rabbit PLUS, DUO92002, Duolink^® In Situ PLA^® Probe Anti-Mouse MINUS, DUO92004) for 60 min at 37 °C. The subsequent steps of ligation and amplification were performed according to the manufacturer's instructions (DUO92013, Sigma). Finally, coverslips were covered with Duolink In Situ Mounting Medium with DAPI (DUO82040, sigma). Images were obtained using Nikon A1 confocal microscope.

All data included in this study are available upon request by contact with the corresponding author.

The proband-1 was an eight-year-old Chinese boy with unexplained recurrent proteinuria since the age of two, and the proband-2 also had unknown proteinuria. Furthermore, most of the treatments for lowering proteinuria remained ineffective. Based on early onset and unclear proteinuria etiology, genetic kidney disease was clinically suspected. Therefore, ES was performed to screen for strong candidate genes underlying susceptibility loci for the recurrent proteinuria and the schematic diagram of the workflow for screening pathogenic mutations was shown in Additional file 1: Fig. S1. ES performed in proband 1 showed variants of unknown significance in the following genes: ANLN (associated with FSGS), NM_018685, hemizygous c.2748 + 6 T > C (splicing); CLCNKB (associated with Bartter Syndrome Type 3 and Type 4B clinically featured by hypokalaemic metabolic alkalosis) hemizygous c.118delA (NM_001165945) and c.782-10_782-8delCCT (splicing, NM_000085) [18‐20]. Given these variants exhibiting symptoms that does not fall within the above clinical phenotypes of proband 1, we excluded the possibility. Finally, it was found that the variants (NM_001081, c.4397G > A (p.C1466Y) and c.6796C > T (p.R2266X)) in CUBN may be responsible for unknown proteinuria. Meanwhile, analysis of ES on proband 2 showed that only 2 suspicious variants (NM_001081, c.5153_5154delCT (p.S1718X) and c.6821 + 3A > G (splicing)) in CUBN were highly correlated with the above clinical phenotypes. To this end, we identified novel compound heterozygous CUBN variants with biallelic, likely pathogenic variants, segregating with proteinuria (Table 1 and Additional file 2: Table S1). All the variants were not listed in the clinvar database, and the minor allele frequency (MAF) was less than 0.01 (Table 1).

Subsequently, the variations were verified by Sanger sequencing (Fig. 1A). Phylogenetic CUBN analysis in different mammalian species showed the gene originated from a common ancestry root, and evolutionary diverged into three major clades suggesting interspecies CUBN conservation was high (Fig. 1B). In light of the genetic predisposition, Sanger sequencing of parents of the probands, who exhibited normal phenotypes, confirmed the variants were inherited from them (Fig. 1C). Interestingly, the sister of proband-2 carried the same variants as the proband and presented with proteinuria as well. Three-dimensional structural models showed that both variants potentially affected folding and function of CUB9, CUB11, CUB16 domains (Fig. 1D). Furthermore, Cubilin functional domain analysis identified the heterozygous C1466Y and S1718X variants were near the back of vitB12-binding domain, whereas the p.R2266X variant and c.6821 + 3A > G variants were in the C-terminal CUB16 region, four of which were shown to have proteinuria without vitB12 malabsorption (Fig. 1E and Table 2).

To assess effects of the variants on Cubilin dysfunction respectively, we first performed an amino acid conservation analysis on the missense variants site (p.C1466Y). Using ClustalW multiple sequence alignments, C1466Y was localized to an evolutionary, highly conserved region in the Cubilin CUB9 domain (Fig. 2A). As the minigene technology was confirmed as a reliable tool to functionally assay potential splicing variants, we checked the c.6821 + 3A > G variants to dissect alternative splicing effects using pSPL3 vector in HEK-293 cells. The splice variant showed skipping of exon 44 to generate premature termination codon (Fig. 2B and C).

To further verify variants associations with protein expression and functions, we constructed recombinant plasmids expressing Cubilin (wild type), Cubilin (p.C1466Y) and the truncated Cubilin (p.S1718X, p.R2266X and p.I2654X), and transiently transfected these into Human Embryonic Kidney 293 T cells. We then performed western blotting to assess Cubilin expression. The mutant CUBN (c.4397G > A) M1 plasmid expressed normal Cubilin, similar to Cubilin (wild type). However, the mutant CUBN (c.6796C > T) M2 disrupted the C-terminus, and truncated Cubilin expression (Fig. 2D). It was the same with the mutant CUBN (c.5153_5154delCT, p.S1718X) M3 and CUBN (c.6821 + 3A > G, p.I2654X) M4 (Fig. 2D).

Furthermore, renal biopsies from a patient (similar age to the proband) with clinically diagnosed non-hereditary MCD, and adjacent healthy kidney tissue from healthy kidney donor who received living-related donor nephrectomy as healthy control (HC) were used to assess in vivo Cubilin expression by immunohistochemistry (IHC). We observed a marked decline in Cubilin expression in the proband’s kidney biopsy when compared with HC and MCD samples (Fig. 2E). Therefore, different variants after vitB12-binding domain in CUBN all appeared to be associated with dysfunctional Cubilin.

To investigate if proteinuria was caused by glomerular damage, percutaneous renal biopsies from the two probands were analyzed by hematoxylin & eosin (H&E), periodic acid–Schiff (PAS), periodic acid-silver metheramine (PASM), and Masson staining. Thankfully, we observed no glomerular and segmental sclerosis in both probands (Fig. 3A). Next, we performed immunofluorescence (IF) using Synaptopodin (Synpo) (Fig. 3B) (a podocyte marker and actin-associated protein in podocyte foot processes) and a Wilms' tumor-1 (WT-1) podocyte nuclear marker (Fig. 3C). IF with these markers indicated no damage or podocyte loss. Furthermore, transmission electron microscopy (TEM) identified multifocal gaps in the GBM, with no electron-dense deposits (Fig. 3D). Also, glomeruli did not present with podocyte foot broadening or effacement (Fig. 3D). Further IF staining was performed to confirm GBM scaffold structural integrity; COL4A3 and COL4A5 are glomerular filtration function markers, and IF staining with these markers showed no structural changes in the GBM (Fig. 3D).

As indicated, CUBN variants caused no pathological structural changes in the glomerulus. To address the cause of proteinuria, we tracked renal function and measured serum autoantibodies and urine protein excretion over 24 h (Tables 3 and Additional file 2: Table S2). Of the autoantibodies associated with clinically relevant renal autoimmune disease, membranous nephropathy (MN), lupus nephritis (LN), and anti-neutrophil cytoplasmic autoantibody (ANCA)-associated vasculitis (AAV) [21] have been detected. These studies suggested neither of the probands had autoimmune dysfunction. We also observed that the urinary albumin-to-creatinine ratio (UACR) was significantly high (Table 3). However, in spite of these high levels, no obvious abnormalities in total urinary protein excretion were identified (Table 3). Of note, α1-microglobulin and transferrin, which are urine biomarkers of kidney tubule injury and dysfunction [22, 23], were significantly increased, while β2-microglobulin was normal (Table 3).

According to the literature, Cubilin is an important receptor, possibly mediating albumin, transferrin and α1-microglobulin endocytosis into renal tubules [24‐26]. When we further considered the probands’ clinical observations, we hypothesized the Cubilin variants may have been involved in renal tubule dysfunction and aberrant renal tubule protein reabsorption, possibly accounting for the chronic isolated proteinuria. To establish a relationship between genetic variants and chronic proteinuria, IF was used to assess Cubilin, Megalin, and AMN co-expression. Interestingly, the variants did not induce aberrant Megalin expression or localization when compared with the HC (Fig. 4A). However, decreased Cubilin expression (caused by the variants) was accompanied by aberrant cytoplasmic localization of AMN in renal tubule membranes (Fig. 4B).

From these results, we hypothesized if different levels of GFB dysfunction may have affected Cubilin and AMN colocalization. To this end, we first explored the colocalization of Cubilin and AMN in kidney biopsies in MCD and FSGS that were identified as non-hereditary nephropathy of GFB damage with similar age. Further TEM was used to investigate podocyte damage, with varying degrees of segmental podocyte foot process broadening, fusion, and effacement in MCD and FSGS, being consistent with the characteristics of the diseases (Fig. 4C). Surprisingly, neither Cubilin nor AMN showed aberrant localization (Fig. 4D). Additionally, in vivo experiments further confirmed that the reduced expression of renal Cubilin did not cause abnormal cytoplasmic localization of AMN in the mice model of LPS-induced acute kidney injury and Adriamycin-induced nephropathy with GFB impairment (Fig. 4E and F). These observations suggested that GFB dysfunction may not have been the cause of aberrant cytoplasmic localization of AMN.

To elucidate the colocalization relationship between Cubilin and AMN, HEK293T cells were co-transfected with eukaryotic transient-expression plasmids (AMN with CUBN, M1-CUBN, M2-CUBN, M3-CUBN and M4-CUBN). Cubilin was localized to the cell membrane, and AMN co-localized with Cubilin under normal physiological conditions (Fig. 5A). However, the four Cubilin variants all caused aberrant AMN localization, and retained it to the cytoplasm (Fig. 5A). Subsequently, western blot was employed to further investigate the expression of AMN and the results indicated that the variants of Cubilin may only cause abnormal localization of AMN by affecting the protein interaction with AMN, but does not affect its expression (Fig. 5B). To this end, we further tested the interaction between the variants of Cubilin and AMN using co-immunoprecipitation (co-IP) and Duolink proximity ligation assay (Fig. 5C and D). The results further demonstrated that the variants of Cubilin reduced the ability to interact with AMN.