Received: 13 June 2020 | Revised: 13 July 2020 | Accepted: 4 August 2020 DOI: 10.1002/mgg3.1475 CLINICAL REPORT Novel compound heterozygous mutations of CLDN16 in a patient with familial hypomagnesemia with hypercalciuria and nephrocalcinosis Alejandro García-Castaño1 | Ana Perdomo-Ramirez2 | Mònica Vall-Palomar3 | Elena Ramos-Trujillo2 | Leire Madariaga1,4 | Gema Ariceta3,5,6 | 2 Felix Claverie-Martin 1 Biocruces Bizkaia Research Institute, Barakaldo, Bizkaia, Spain Abstract 2 Unidad de Investigación, Hospital Background: Familial hypomagnesemia with hypercalciuria and nephrocalcinosis Universitario Nuestra Señora de Candelaria, (FHHNC) is an autosomal recessive tubulopathy characterized by excessive uri- Santa Cruz de Tenerife, Spain nary wasting of magnesium and calcium, bilateral nephrocalcinosis, and progressive 3 Fisiopatologia Renal, Centro de chronic renal failure in childhood or adolescence. FHHNC is caused by mutations Investigaciones en Bioquímica y Biología Molecular (CIBBIM), Vall d’Hebron in CLDN16 and CLDN19, which encode the tight-junction proteins claudin-16 and Institut de Recerca (VHIR, Barcelona, claudin-19, respectively. Most of these mutations are missense mutations and large Spain 4 deletions are rare. Pediatric Nephrology Department, Cruces University Hospital, UPV/EHU, Barakaldo, Methods: We examined the clinical and biochemical features of a Spanish boy with Spain early onset of FHHNC symptoms. Exons and flanking intronic segments of CLDN16 5 Servicio de Nefrología Pediátrica, Hospital and CLDN19 were analyzed by direct sequencing. We developed a new assay based Universitari Vall d'Hebron, Barcelona, on Quantitative Multiplex PCR of Short Fluorescent Fragments (QMPSF) to inves- Spain 6 Departamento de Pediatría, Universitat tigate large CLDN16 deletions. Autònoma de Barcelona, Bellaterra Results: Genetic analysis revealed two novel compound heterozygous mutations of (Barcelona), Spain CLDN16, comprising a missense mutation, c.277G>A; p.(Ala93Thr), in one allele, Correspondence and a gross deletion that lacked exons 4 and 5,c.(840+25_?)del, in the other allele. Felix Claverie-Martin, Unidad de The patient inherited these variants from his mother and father, respectively. Investigación, Hospital Universitario Conclusions: Using direct sequencing and our QMPSF assay, we identified the ge- Nuestra Señora de Candelaria, Carretera del Rosario 145, 38010 Santa Cruz de Tenerife, netic cause of FHHNC in our patient. This QMPSF assay should facilitate the genetic Spain. diagnosis of FHHNC. Our study provided additional data on the genotypic spectrum Email: fclamar@gobiernodecanarias.org; of the CLDN16 gene. fclamar@gmail.com FUNDING INFORMATION KEYWORDS This work was supported by Grant claudin-16, CLDN16, deletion, hypomagnesemia, novel mutations, QMPSF PI17/00153 co-financed by the Instituto de Salud Carlos III (Spain) and the European Regional Development Fund “Another way to build Europe”. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2020 The Authors. Molecular Genetics & Genomic Medicine published by Wiley Periodicals LLC. Mol Genet Genomic Med. 2020;8:e1475.  wileyonlinelibrary.com/journal/mgg3 | 1 of 10 https://doi.org/10.1002/mgg3.1475 2 of 10 |    GARCÍA-CASTAÑO et al. 1 | IN TRO D U C T ION regions. Only one large deletion has been identified in the CLDN16 gene of a patient with FHHNC (Yamaguti et al., Loss-of-function mutations in two genes involved in renal 2015). paracellular reabsorption of calcium and magnesium, In this study, we aimed to describe the clinical and ge- CLDN16 (OMIM *603959) and CLDN19 (OMIM *610036), netic features of a patient with early onset of the first FHHNC cause two types of a rare autosomal recessive disease known symptoms. Since direct DNA sequencing of coding exons as Familial Hypomagnesemia with Hypercalciuria and and flanking intronic sequences only detected a missense het- Nephrocalcinosis, respectively (FHHNC; type 1, OMIM erozygous mutation of CLDN16, we designed a Quantitative #248250 and type 2, #248190) (Konrad et al., 2006; Simon Multiplex PCR of Short Fluorescent Fragments (QMPSF) et al., 1999). This tubulopathy is characterized by excessive assay to identify a novel large deletion in the other allele. urinary losses of renal magnesium and calcium, bilateral nephrocalcinosis, and progressive chronic renal failure early in life (Claverie-Martin, 2015; Praga et al., 1995). Patients 2 | M ATERIAL S AND M ETHOD S usually present with recurrent urinary tract infections, poly- uria, polydipsia, and nephrolithiasis. The symptoms usually 2.1 | Ethical compliance appear in early childhood or adolescence. Common bio- chemical characteristics include low levels of magnesium The Ethics Committee of Hospital Universitario Nuestra in serum, high urine levels of calcium and magnesium, high Señora de Candelaria (Santa Cruz de Tenerife, Spain) ap- serum levels of parathyroid hormone, and reduced glomer- proved the protocols of this study, which was conducted ular filtration rate (Godron et al., 2012; Konrad et al., 2008; according to the Declaration of Helsinki. Written informed Weber et al., 2001; Claverie-Martin et al., 2013). Defects in consent was obtained from the patient's parents. enamel formation have been reported in some FHHNC pa- tients (Bardet et al., 2016; Yamaguti et al., 2017). In addition, patients with mutations in CLDN19 display severe ocular ab- 2.2 | Genomic DNA purification and direct normalities such as myopia, macular colobamata, and nys- sequence analysis tagmus (Konrad et al., 2006; Claverie-Martin et al., 2013; Godron et al., 2012). After obtaining written informed consent, peripheral blood CLDN16 and CLDN19 encode tight junction proteins samples of patients and relatives were collected for genetic claudin-16 and claudin-19, respectively, which are members analysis. Genomic DNA was extracted using the GenElute of a family of membrane proteins that contain four transmem- Blood Genomic DNA kit (Sigma-Aldrich, St. Louis, MO, brane domains, two extracellular loops, a cytoplasmic loop, USA) following the manufacturer's instructions. The cod- and cytoplasmic amino- and carboxy ends (Meoli & Günzel, ing exons and flanking intronic sequences of CLDN16 and 2020; Suzuki et al., 2014). Both claudins interact with each CLDN19 were amplified by polymerase chain reaction (PCR) other on adjacent epithelial cells of the thick ascending loop using primers and conditions previously described (Claverie- of Henle to form a cation paracellular selective barrier that Martin et al., 2013; Perdomo-Ramirez et al., 2019; Simon regulates the paracellular reabsorption of magnesium and et al., 1999). PCR products were purified with the QIAquick calcium in the kidney (Hou et al., 2008). Claudin-16 and PCR purification kit (Qiagen, Hilden, Germany) and se- claudin-19 are also expressed in the tight junction of amelo- quenced with the BigDye Terminator v3.1 Cycle Sequencing blasts, which might explain their association with defective Kit (Applied Biosystems, Foster City, CA, USA). Sequence enamel formation in FHHNC patients (Bardet et al., 2016; reactions were purified with Performa®DTR Gel Filtration Yamaguti et al., 2017). Moreover, severe ocular involvement Cartridges (EdgeBio BioSystems, Gaithersburg, Maryland, in FHHNC patients with mutation in claudin-19 might be USA), and analyzed on a 3500 Series Genetic Analyzer attributed to the expression of this claudin in fetal retinal (Applied Biosystems, Foster City, CA, USA). Mutations pigment epithelium (Peng, Rao, Adelman, & Rizzolo, 2011; were identified by comparison to the respective reference Wang et al., 2019). sequences (GenBank accession numbers NG_008993.1 and The clinical diagnosis of FHHNC needs to be confirmed NG_008149.1, for CLDN19 and CLDN16, respectively), and by the detection of CLDN16 or CLDN19 pathogenic variants confirmed by sequencing additional independent amplifi- in both alleles. Most CLDN16 and CLDN19 variants identi- cation products. We inspected several databases, including fied in FHHNC patients are missense mutations found in ho- ClinVar (https://www.ncbi.nlm.nih.gov/clinv​ ar/), Human mozygous or compound heterozygous state (Claverie-Martin Gene Mutation database (HGMD, http://www.hgmd.cf.ac. et al., 2015; Prot-Bertoye & Houillier, 2020). These muta- uk/ac/index.php), 1000 Genomes Project (http://www.1000g​ tions are mainly located in the two extracellular loops but a enomes.org/), and gnomAD database (https://gnomad.broad​ few affect the transmembrane domains and the cytoplasmic insti​tute.org/), to verify that the variant detected in our patient GARCÍA-CASTAÑO et al.    | 3 of 10 was not a common polymorphism and to confirm that it was the control-amplified product. Finally, the results obtained novel. There are two potential start codons (methionine 1 and from the test samples were compared with the results from a methionine 71) in the CLDN16 gene that would produce a healthy individual and positive control samples. short (235 amino acids) or a long (305 amino acids) clau- din-16 isoform (Hou, Paul, & Goodenough, 2005; Weber et al., 2001). It is currently unknown whether both iso- 2.4 | Bioinformatics analysis forms are functional. For all CLDN16 mutations described in the literature, nucleotide numbering starts with the A of The potential detrimental effect of the amino acid substitu- the first ATG translation initiation site as nucleotide 1 (see, tion on the structure and function of claudin-16 was evalu- table 4 of Prot-Bertoye & Houillier, 2020, and HGMD). ated using the following bioinformatics tools, which are We used this same numbering for the two mutations de- based on different principles: PolyPhen-2 (http://genet​ics. scribed here (Reference NCBI sequence used for numbering: bwh.harva​ rd.edu/pph2/) (Adzhubei et al., 2010), Panther NM_006580.3). We followed the recommendations of the v.15.0 (http://www.panth​erdb.org/tools/​csnpS​coreF​orm. Human Genome Variation Society (http://varno​ men.hgvs. jsp) (Tang & Thomas, 2016), MutPred2 (http://mutpr​ org/) for the description of sequence variants. ed.mutdb.org/) (Pejaver, Mooney, & Radivojac, 2017), MutationTaster2 (http://www.mutat​ionta​ster.org/) (Schwarz, Cooper, Schuelke, & Seelow, 2014), and Fathmm-XF (http:// 2.3 | Quantitative Multiplex PCR of Short fathmm.bioco​ mpute.org.uk/fathm​ m-xf/) (Rogers et al., Fluorescent Fragments analysis 2018). The protein sequence of human claudin-16 was ob- tained from the Uniprot database (https://www.unipr​ot.org/, In order to detect the potential gross deletions or duplications entry identifier Q9Y5I7). Protein stability changes resulting in the CLDN16 region, we used a QMPSF assay. Primer pairs from the missense variant were estimated using web-based for amplification of short exonic fragments corresponding to programs MUpro (http://mupro.prote​ omics.ics.uci.edu/) exons 1–5 of the CLDN16 gene and a sequence control of the (Cheng, Randall, & Baldi, 2006) and I-Mutant 3.0 (http:// HNF1B gene were designed (Table 1). Simultaneous PCR- gpcr2.bioco​mp.unibo.it/cgi/predi​ctors/​I-Mutan​t3.0/I-Mutan​ amplification was performed in a 20µL reaction mixture con- t3.0.cgi) (Capriotti, Fariselli, & Casadio, 2005). These tools taining dNTPs (10 mM), MgCl2, 10% DMSO, KAPA Taq provide the predicted free energy change value (DDG) and DNA Polymerase (Kapabiosystems, Boston, Massachusetts), the sign of the prediction. The DDG value is calculated from dye-labeled primers (0.3 µM of each primer), and 40 ng of the unfolding Gibbs free energy value of the mutant protein genomic DNA. The amplification conditions consisted of minus the unfolding Gibbs free energy value of the wild type an initial step of denaturation at 94°C for 5 min, followed (kcal/mol). A DDG value below 0 means that the stability by 24 cycles of denaturation at 94°C for 20 s, annealing at of the protein has decreased, whereas a DDG superior to 0 60°C for 20 s, and extension at 72°C for 20 s. A final ex- means it has increased. tension was performed at 72°C for 10 min. Then, the sam- ples were denatured and loaded onto an ABI3130xl Genetic Analyzer (Life Technologies). We used the GeneMapper® 2.5 | Protein modeling Software v. 4.0 (Applied Biosystems, California) for data analysis. Normalization was performed by dividing the peak There is no experimentally determined 3D structure avail- height of each amplification product by the peak height of able for claudin-16 protein in the Protein Data Bank (PDB). TABLE 1 Primer used for QMPSF analysis of the human CLDN16 gene Amplicon Exona Forward primer (5′–3′) Reverse primer (5′–3′) size (bp) 1 GACCACCACTAGCCCACAGT 56-FAM-TGGTACCTGGCAATGTGAAA 243 2 AATGCTTTTGATGGGATTCG 56-FAM-CCATTACAAACTGGACCGAAC 185 3 TGGTAACTCGAGCGTTGATG 56-FAM-GTTGCTAGTCCAGCCAGACC 181 4 CCAGGAATCATTGGCTCTGT 56-FAM-GAACAGCTCCAGCCAAAAAG 160 5 TTGGACCTGAGAGAAACTATCCTT 56-FAM-AGCATACATTTTGGCCGTCT 113 b Control 56-FAM-TGTCTAGTGAGGACCCTTGG GGATCTCTCGTTGCTTTCTG 199 a Numbering is according to DNA sequence (Ensembl: ENST00000264734.2). b HNF1B gene. 4 of 10 |    GARCÍA-CASTAÑO et al. Therefore, wild-type and mutant claudin-16 were modeled 3.2 | Detection of a novel CLDN16missense using the SWISS-MODEL server (Waterhouse et al., 2018) mutation by direct DNA sequencing (https://swiss​ model.expasy.org/). Protein sequences in FASTA format were uploaded and modeled with the crystal DNA sequence analysis showed that the CLDN19 coding structure of human claudin-4 in complex with the C-terminal exons and their flanking intronic regions had the normal se- fragment of Clostridium perfringens enterotoxin (PDBe: quence. Analysis of the CLDN16 gene revealed a heterozy- 5b2g.2) (Shinoda et al., 2016). gous variant c.277G>A, in exon 1 (Figure 1a). This variant was also detected in heterozygosis in the patient's mother, who was unaffected. Sequence analysis of the patient's fa- 3 | R ES U LTS ther and brother, both healthy, showed the CLDN16 normal sequence (Figure 1a). The pedigree of the family is shown 3.1 | Clinical description in Figure 1b.The G to A substitution leads to the replace- ment of alanine for threonine in residue 93 of the claudin-16 An 11 month old previously healthy boy consulted for protein, p.(Ala93Thr). Alanine 93 is located next to the last fever and respiratory symptoms, and leukocyturia was amino acid of the first transmembrane domain (Figure 1c), identified incidentally. Urinary tract infection was sus- and is highly conserved throughout the evolution (Figure pected but the finding by renal ultrasound of severe and 1d). We searched several databases to verify that the vari- bilateral nephrocalcinosis led to additional diagnostic ant detected in our patient, c.277G>A; p.(Ala93Thr), was work-up. His physical exam, height, and weight were not a common SNP and to confirm that it was a novel muta- normal by age (percentile >99), as well as his blood pres- tion not previously reported. This variant was not found in sure. Previously unnoticed polyuria and hyposthenuria ClinVar, HGMDor 1000 Genomes Project database. In the were also identified. Biochemistry serum tests revealed gnomAD database, it was found in only one allele from a increased creatinine (0.61 mg/dl) [estimated Glomerular genome sample from African population with a frequency Filtration Rate (eGFR): 56 ml/min/1.73 m2)], without asso- of 0.00003 (variant ID: 3-190106185-G-A). In silico predic- ciated electrolytes anomalies (potassium 5.3 mEq/L, chlo- tion tools PolyPhen-2, Panther, MutPred2, MutationTaster, ride 106 mEq/L, total calcium 10.2 mg/dl, and phosphorus and Fathmm-XF indicated that this rare variant was probably 4.9 mg/dl) except mild hypomagnesemia (magnesium pathogenic (Table 2). Furthermore, this variant caused a de- 1.54 mg/dl) and elevated uric acid (7.4 mg/dl), and remark- crease in claudin-16 protein stability according to programs ably significant high intact parathyroid hormone (175 pg/ that predict the protein stability of mutated proteins including ml), for the degree of renal function impairment, with I-Mutant 3.0 (DDG in kcal/mol: −1.51).We submitted this adequate 25-hydroxy-vitamin D levels (25 ng/dl). Acid- new missense CLDN16 mutation to ClinVar and it was in- base balance was preserved. Urine exam demonstrated in- cluded with accession number VCV000930213.1. creased urinary calcium and magnesium excretion: urinary calcium/creatinine (UCa/Cr) 0.39 mg/mg, urinary magne- sium/creatinine (UMg/Cr) 0.21 mg/mg (at the high-normal 3.3 | Identification by QMPSF of a new level but inappropriately increased for the level of serum CLDN16 deletion magnesium), and elevated fractional excretion of magne- sium (FEMg, 8.2%). At the age of 2 years, under the clinical Direct sequencing of CLDN16 revealed only a mutated allele, suspect of FHHNC, genetic analysis was performed (see, therefore, since FHHNC is an autosomal recessive disease, below). The patient's follow-up 7 years later was consist- we suspected that the mutation in the other allele was a large ent with FHHNC with moderate chronic kidney disease heterozygous deletion in this gene. In order to detect a po- (serum creatinine 1.16 mg/dl (eGFR: 40 ml/min/1.73 m2), tential heterozygous deletion of CLDN16 that could explain persistent hypomagnesemia despite oral supplements the clinical phenotype and to map precisely the deletion, we (magnesium 1.3 mg/dL) due to maintained hypermagne- designed a QMPSF assay, a method based on simultaneous siuria (UMg/Cr 0.27 mg/mg, FEMg 24%), associated with sig- amplification of multiple short sequences under quantitative nificant hypercalciuria (UCa/Cr 0.42 mg/mg; VCa 5.4 mg/ conditions. The results of this assay revealed a novel large de- Kg/day). His PTH levels were normalized under medical letion in the CLDN16 gene of the patient that, at least, lacked treatment. Dentistry careful evaluation did not demonstrate exons 4 and 5, c.(840+25_?)del. As illustrated by Figure 2, any enamel defects. Furthermore, the patient had no ocu- exonic deletions were easily detected by a 50%decrease of lar abnormalities. His parents are not consanguineous, and the corresponding peaks; the probes designed for exons 4 and there is no family history of nephropathy except a maternal 5 presented half the expected signal in the patient, brother great grandmother with a single kidney and nephrolithiasis. and father. In contrast, exons 1, 2, and 3 were not deleted. GARCÍA-CASTAÑO et al.    | 5 of 10 F I G U R E 1 Identification of novel CLDN16 heterozygous missense mutation c.277G>A; p.(Ala93Thr) in patient with FHHNC. (a) Electropherograms showing the heterozygous substitution in exon 1 of the patient and his mother. Arrows indicate the position affected by the mutation in the patient and his mother. The brother and father revealed the normal sequence. (b) Pedigree of the family showing segregation of both CLDN16 variants (missense mutation and deletion in blue and black, respectively). Circle, female individual; squares, male individual; P, proband; del, c.(840+25_?)del detected in the other allele. The patient inherited the missense mutation and the deletion from his mother and father, respectively. (c) Schematic representation of the CLDN16 gene, cDNA, and claudin-16 protein. Arrows indicate the position affected by missense mutation c.277G>A; p.(Ala93Thr). Colored boxes represent the five coding exons and black lines indicate intron sequences; exons missing in the deletion mutant are in red. Exons and introns sizes are not at scale. The positions of the ATG start codon and the TAA stop codon are also shown. Exons are associated by dotted lines to the schematic representation of the claudin-16 protein where transmembrane domains (TM1–TM4), extracellular segments (EC1 and EC2) and cytoplasmic regions (white boxes) are indicated. (d) Multiple sequence alignment of claudin-16 or related proteins from different species showing evolutionary conservation of alanine 93. 1T. rubripes does not contain the ortholog for mammalian claudin-16. The sequence shown is that of claudin-11B 6 of 10 |    GARCÍA-CASTAÑO et al. TABLE 2 Pathogenicity prediction for claudin-16 mutation p.(Ala93Thr) using bioinformatics tools Fathmm- Tool PolyPhen-2a Pantherb MutPred2c MutationTasterd XFe Score 0.85 362 0.747 58 0.615 Significance Possibly damaging Possibly damaging Probably pathogenic Disease causing Pathogenic a The PolyPhen-2 score varies from 0 to 1 Variants with scores in the range 0.85 to 1.0 are more confidently predicted to be damaging. b Panther measures the length of time (in millions of years, my) an amino acid position in the protein has been preserved. The longer a position has been preserved, the more likely that amino acid change will have a deleterious effect. The thresholds are: probably damaging (preservation time is greater than 450 my), possibly damaging (preservation time is between 200 my and 450 my), and probably benign (preservation time is less than 200 my). c The MutPred2 prediction ranges from 0.0 and 1.0, in which a higher score indicates a greater propensity to be pathogenic. d The MutationTaster score ranges from 0.0 to 215. It is derived from the Grantham Matrix for amino acid substitutions and shows the physicochemical difference between the original and the mutated amino acid. e Fathmm-XF predictions are given as p-values in the range 0–1: values above 0.5 are predicted to be deleterious, while those below 0.5 are predicted to be neutral or benign. F I G U R E 2 Detection of the CLDN16 deletion using a QMPSF assay. Panels show the QMPSF electropherograms of: (a) normal control, (b) proband, (c) brother, and (d) father. Blue peaks and red peaks correspond to CLDN16 exons 1–5 fragments and molecular weight markers, respectively. The Y-axis indicates fluorescence in arbitrary units, and the X-axis displays the size in base pairs. The heterozygous deletion is easily detected by a reduction of the peaks corresponding to exons 4 and 5 (arrows) in the proband, his brother and father compared to a normal control The patient's unaffected brother and father were found to be protein with the missense mutation (Figure 3). The predicted carriers of this deletion (Figure 2). This novel CLDN16 dele- 3D structure of the claudin-16 protein encoded by the allele tion was submitted to ClinVar and it was included with ac- with the deletion of exons 4 and 5 resulted in the removal cession number SCV001335520. of nine amino acids of the third transmembrane domain, the second extracellular loop, the fourth transmembrane domain (encoded by exon 4) and the carboxy-terminal cytoplasmic 3.4 | Modeling of claudin-16 region containing a PDZ-binding motif and potential phos- protein structure phorylation sites (encoded by exon 5) (Figure 3). To analyze the structural impact of the two CLDN16 muta- tions detected in the patient, the 3D structure of the claudin-16 4 | DISCUSSION protein was modeled based on the solved protein structure of claudin-4 using SWISS-MODEL. The results did not reveal FHHNC is a rare hereditary disease characterized by any derangement in the structure of the mutant claudin-16 excessive renal losses of magnesium and calcium, GARCÍA-CASTAÑO et al.    | 7 of 10 F I G U R E 3 Predicted 3D structures of wild-type and mutant claudin-16 proteins generated using homology modeling with SWISS-MODEL. TM1 to 4, transmembrane domains 1 to 4; EC1 and EC2, extracellular loops 1 and 2; CS, cytoplasmic segment. (a) Wild type claudin-16. (b) Mutant p.(Ala93Thr). Blue arrows indicate the location of the original alanine 93 residue and missense mutation p.(Ala93Thr). The NH2 and COOH termini correspond to threonine 70 and cysteine 255, respectively. (c) Deletion mutant lacking part of TM3, EC2, TM4, and COOH terminus. The COOH terminus corresponds now to alanine 197. The models are based on template 5b2 g.2 of claudin-14 bilateral nephrocalcinosis, and progressive chronic renal fail- others remain in the endoplasmic reticulum, Golgi apparatus, ure (Claverie-Martin, 2015; Praga et al., 1995). Clinical ob- or lysosomes. Even mutant proteins that are correctly local- servations and clearance studies in patients suggested that the ized to the tight junction frequently show defective magne- primary defect was associated with reduced paracellular re- sium paracellular transport (Hou et al., 2005; Konrad et al., absorption of magnesium and calcium in the thick ascending 2008). Hou and colleagues have shown that claudin-16 inter- loop of Henle (Rodriguez-Soriano, Vallo, & Garcia-Fuentes, acts with claudin-19 conferring cation selectivity to the tight 1987). The disease is caused by loss-of-function mutations junction in a synergistic manner (Hou et al., 2008). Mutations ofCLDN16 (previously known as PCLN1) (type 1), (Simon that disturb this interaction cause loss of cation selectivity. et al., 1999) or CLDN19 (type 2).The majority of Spanish Direct sequencing of CLDN16 exons identified heterozy- patients with FHHNC are homozygous for CLDN19 founder gous missense variant c.277G>A; p.(Ala93Thr) in our pa- mutation p.(Gly20Asp), and they also have severe ocular tient, which was inherited from his mother. This very rare defects (Konrad et al., 2006, Claverie-Martin et al., 2013; variant, which results in the replacement of an alanine residue Martin-Nuñez et al., 2015). We examined the phenotype for threonine, has not been described previously in FHHNC. and genotype of a Spanish child with the typical character- The substitution of a hydrophobic amino acid for a polar istics of FHHNC and no ocular abnormalities. He suffered amino acid in the first transmembrane domain of claudin-16 from polyuria and hyposthenuria. Bilateral nephrocalcinosis, could lead to misfolding or defects in membrane insertion. mild hypomagnesemia, hypercalciuria, elevated fractional Therefore, p.(Ala93Thr) could affect the formation of clau- excretion of magnesium, increased parathyroid hormone, din-16/claudin-19 tight junctions. We consider this variant is and decreased glomerular filtration rate were found. Genetic a pathogenic mutation based on the following: (1) it affects a analysis of the proband and family members revealed two highly conserved amino acid residue, (2) bioinformatics anal- novel mutations of CLDN16 in a compound heterozygous ysis predicts deleterious consequences, (3) it's extremely low state, which confirmed the clinical diagnosis of FHHNC type frequency, and (4) it results in a decrease of protein stability. 1. However, this would have to be confirmed experimentally by The CLDN16 gene is located on chromosome 3q27, and functional analysis. Most CLDN16 mutations affect the two to date, 69 pathogenic mutations have been identified: 43 extracellular loops of the protein, and only two previously re- missense, 10 nonsense, 5 splicing, 5 small deletions, 2 small ported missense mutations are located in the first transmem- insertion, 2 small indels, 1 complex, and 1 gross deletion brane domain, p.(Cys80Tyr) and p.(Gly88Glu) (Prot-Bertoye (Prot-Bertoye & Houillier, 2020). Because of the autosomal & Houillier, 2020). Functional analysis of p.(Gly88Glu) has recessive pattern of inheritance in FHHNC type 1, patients revealed that this mutation results in a complete loss of func- are homozygous or compound heterozygous for the particular tion (Konrad et al., 2008). mutations. Functional studies have shown that CLDN16 mu- Since only one heterozygous mutation of CLDN16 could tations may cause partial or complete loss of claudin-16 func- not explain the FHHNC phenotype of our patient, we used tion (Hou et al., 2005; Kausalya et al., 2006; Konrad et al., a QMPSF assay designed by us to detect the mutation in 2008; Müller et al., 2003). The majority of mutant claudin-16 the other allele. The results showed a heterozygous dele- proteins display normal trafficking to the cell membrane, but tion of exons 4 and 5 that was inherited from his father. The 8 of 10 |    GARCÍA-CASTAÑO et al. claudin-16 protein encoded by this mutant would lack part residual function, while the deletion involves exons 4 and 5 of the third transmembrane domain, the second extracellular and most likely leads to a complete loss of function. These loop, the fourth transmembrane domain, and the cytoplasmic mutations were inherited from his mother and father, respec- carboxy terminus, which plays an important role in protein tively. Our study provides further insights into the molecu- stability and trafficking to the tight junction (Itoh et al., 1999; lar basis of FHHNC. Moreover, the QMPSF assay described Müller et al., 2003). Therefore, we assumed that this deletion here should facilitate the genetic diagnosis of FHHNC. induces a complete loss of function of the mutant claudin-16 protein. Only one large homozygous deletion in the CLDN16 ACKNOWLEDGMENTS gene of a patient with FHHNC has been reported in the litera- This study was supported by grant PI17/00153, integrated in ture (Yamaguti et al., 2015). This deletion was detected using the Plan Nacional de I+D+I 2013-2016 and co-financed by a multiplex ligation-dependent probe amplification assay, the ISCIII-Subdirección General de Evaluación y Fomento and included exons 2 to 5. CLDN16 large deletions may be de la Investigación and the European Regional Development underestimated, and we suggest that in cases where only one Fund “Another way to build Europe,” and by the Hipofam mutated allele is detected, the genetic analysis should include Association. We thank the patient and family members for a method to detect large deletions. their participation in this study The pathogenesis of chronic renal disease in patients with FHHNC remains unclear. The hypercalciuria and CONFLICT OF INTEREST nephrocalcinosis present in these patients may contribute to The authors declare that they have no competing interests. progression to end-stage renal disease, however, a suitable correlation has not been found (Praga et al., 1995). Moreover, AUTHOR CONTRIBUTION not all hereditary tubulopathies characterized by nephrocal- AG-C designed and implemented the QMPSF assay. AP-R cinosis lead to end-stage renal disease. Other factors related and M-VP performed PCR amplifications and analyzed the to the progression of chronic kidney disease could be the DNA sequences. AP-R, ER-T, and FC-M contributed to the activation of the inflammasome by crystal nephropathy or bioinformatics analysis. GA and LM collected the patient's the anomalies caused by the defective claudin-16 function data and followed up the case. FC-M, AG-C, and GA wrote early in the development of tubular tight junctions (Godron the paper. FC-M and GA acquired funding. All authors read et al., 2012; Claverie-Martin et al., 2015). Konrad and col- and approved the final version of the manuscript. leagues have suggested a genotype-phenotype correlation re- lated to the progression of renal failure in FHHNC patients ORCID with CLDN16 mutations (Konrad et al., 2008). They found Alejandro García-Castaño https://orcid. that the progression of renal failure is significantly faster in org/0000-0002-9734-6600 patients with complete loss-of-function mutations in both Ana Perdomo-Ramirez https://orcid. alleles as compared with patients with partial loss-of-func- org/0000-0002-2562-2119 tion mutations in one or both alleles. As most patients with Mònica Vall-Palomar https://orcid. FHHNC, our patient presented the typical characteristics of org/0000-0002-4133-5814 FHHNC, including moderate chronic kidney disease early in Elena Ramos-Trujillo https://orcid. infanthood. However, during the last 7 years of follow-up, he org/0000-0002-7080-7649 showed a slow deterioration in renal function (eGFR from Leire Madariaga https://orcid.org/0000-0002-4032-9842 56–40 ml/min/1.73 m2) in comparison with other patients Gema Ariceta https://orcid.org/0000-0003-1763-1098 (Weber et al., 2001). This suggests that at least one of the two Felix Claverie-Martin https://orcid. mutations detected in the patient has a partial loss of func- org/0000-0003-2468-3267 tion. Since the large deletion obviously has a complete loss of function, we speculate that p.(Ala93Thr) has a residual clau- R E F E R E NC E S din-16 function that delays progression of renal failure. Adzhubei, I. A., Schmidt, S., Peshkin, L., Ramensky, V. E., Gerasimova, In conclusion, we report the clinical and genetic features A., Bork, P., … Sunyaev, S. R. (2010). A method and server for of a Spanish boy with early onset of FHHNC symptoms but predicting damaging missense mutations. Nature Methods, 7(4), 248–249. https://doi.org/10.1038/nmeth​0410-248 slow deterioration in renal function. By applying direct DNA Bardet, C., Courson, F., Wu, Y., Khaddam, M., Salmon, B., Ribes, S., sequencing and a QMPSF analysis designed by us, two novel … Chaussain, C. (2016). Claudin-16 deficiency impairs tight junc- compound heterozygous pathogenic variants of CLDN16 tion function in ameloblasts, leading to abnormal enamel formation. were identified; a missense mutation,c.277G>A; p.(Ala- Journal of Bone and Mineral Research, 31(3), 498–513. https://doi. 93Thr), in one allele and a large deletion, c.(840+25_?)del, in org/10.1002/jbmr.2726 the other allele. The missense mutation affects the first trans- Capriotti, E., Fariselli, P., & Casadio, R. (2005). I-Mutant2.0: predict- membrane domain of claudin-16 and probably conserves a ing stability changes upon mutation from the protein sequence or GARCÍA-CASTAÑO et al.    | 9 of 10 structure. Nucleic Acids Research, 33(Web Server), W306–W310. nephrocalcinosis. World Journal of Pediatrics, 11(3), 272–275. https://doi.org/10.1093/nar/gki375 https://doi.org/10.1007/s1251​9-014-0528-3 Cheng, J., Randall, A., & Baldi, P. (2006). Prediction of protein stabil- Meoli, L., & Günzel, D. (2020). Channel functions of claudins in the ity changes for single-site mutations using support vector machines. organization of biological systems. Biochimica Et Biophysica Acta Proteins: Structure, Function, and Bioinformatics, 62(4), 1125– (BBA) - Biomembranes, 1862(9), 183344. https://doi.org/10.1016/j. 1132. https://doi.org/10.1002/prot.20810 bbamem.2020.183344 Claverie-Martin, F. (2015). Familial hypomagnesaemia with hypercal- Müller, D., Kausalya, P. J., Claverie-Martin, F., Meij, I. C., Eggert, ciuria and nephrocalcinosis: Clinical and molecular characteristics. P., Garcia-Nieto, V., & Hunziker, W. (2003). A novel claudin 16 Clinical Kidney Journal, 8(6), 656–664. https://doi.org/10.1093/ckj/ mutation associated with childhood hypercalciuria abolishes bind- sfv081 ing to ZO-1 and results in lysosomal mistargeting. The American Claverie-Martín, F., García-Nieto, V., Loris, C., Ariceta, G., Nadal, I., Journal of Human Genetics, 73(6), 1293–1301. https://doi. Espinosa, L., … Muley, R. (2013). Claudin-19 mutations and clin- org/10.1086/380418 ical phenotype in Spanish patients with familial hypomagnesemia Pejaver, V., Mooney, S. D., & Radivojac, P. (2017). Missense variant with hypercalciuria and nephrocalcinosis. PLoS One, 8(1), e53151. pathogenicity predictors generalize well across a range of func- https://doi.org/10.1371/journ​al.pone.0053151 tion-specific prediction challenges. Human Mutation, 38(9), 1092– Claverie-Martín, F., Vargas-Poussou, R., Müller, D., & García-Nieto, 1108. https://doi.org/10.1002/humu.23258 V. (2015). Clinical utility gene card for: Familial hypomagnesemia Peng, S., Rao, V. S., Adelman, R. A., & Rizzolo, L. J. (2011). Claudin-19 with hypercalciuria and nephrocalcinosis with/without severe ocu- and the barrier properties of the human retinal pigment epithelium. lar involvement. European Journal of Human Genetics, 23(6), 889. Investigative Opthalmology & Visual Science, 52(3), 1392. https:// https://doi.org/10.1038/ejhg.2014.176 doi.org/10.1167/iovs.10-5984 Godron, A., Harambat, J., Boccio, V., Mensire, A., May, A., Rigothier, Perdomo-Ramirez, A., Aguirre, M., Davitaia, T., Ariceta, G., Ramos- C., … Vargas-Poussou, R. (2012). Familial hypomagnesemia with Trujillo, E., & Claverie-Martin, F. (2019). Characterization of two hypercalciuria and nephrocalcinosis: Phenotype-genotype correla- novel mutations in the claudin-16 and claudin-19 genes that cause tion and outcome in 32 patients with CLDN16 or CLDN19 muta- familial hypomagnesemia with hypercalciuria and nephrocalcinosis. tions. Clinical Journal of the American Society of Nephrology, 7(5), Gene, 689, 227–234. https://doi.org/10.1016/j.gene.2018.12.024 801–809. https://doi.org/10.2215/cjn.12841211 Praga, M., Vara, J., González-Parra, E., Andrés, A., Alamo, C., Araque, Hou, J., Paul, D., & Goodenough, D. (2005). Paracellin-1 and the mod- A., … Rodicio, J. L. (1995). Familial hypomagnesemia with hyper- ulation of ion selectivity of tight junctions. Journal of Cell Science, calciuria and nephrocalcinosis. Kidney International, 47(5), 1419– 118(21), 5109–5118. https://doi.org/10.1242/jcs.02631 1425. https://doi.org/10.1038/ki.1995.199 Hou, J., Renigunta, A., Konrad, M., Gomes, A. S., Schneeberger, E. E., Prot-Bertoye, C., & Houillier, P. (2020). Claudins in renal physiology Paul, D. L., … Goodenough, D. A. (2008). Claudin-16 and clau- and pathology. Genes, 11(3), 290. https://doi.org/10.3390/genes​ din-19 interact and form a cation-selective tight junction complex. 11030290 Journal of Clinical Investigation, 118(2), 619–628. https://doi. Rodriguez-Soriano, J., Vallo, A., & Garcia-Fuentes, M. (1987). org/10.1172/jci33970 Hypomagnesaemia of hereditary renal origin. Pediatric Nephrology, Itoh, M., Furuse, M., Morita, K., Kubota, K., Saitou, M., & Tsukita, S. 1(3), 465–472. https://doi.org/10.1007/bf008​49255 (1999). Direct binding of three tight junction-associated MAGUKs, Rogers, M. F., Shihab, H. A., Mort, M., Cooper, D. N., Gaunt, T. R., & ZO-1, ZO-2, and ZO-3, with the COOH termini of claudins. The Campbell, C. (2018). FATHMM-XF: Accurate prediction of patho- Journal of Cell Biology, 147(6), 1351–1363. https://doi.org/10.1083/ genic point mutations via extended features. Bioinformatics, 34(3), jcb.147.6.1351 511–513. https://doi.org/10.1093/bioin​forma​tics/btx536 Kausalya, P. J., Amasheh, S., Günzel, D., Wurps, H., Müller, D., Schwarz, J. M., Cooper, D. N., Schuelke, M., & Seelow, D. (2014). Fromm, M., & Hunziker, W. (2006). Disease-associated mutations MutationTaster2: Mutation prediction for the deep-sequencing affect intracellular traffic and paracellular Mg2+ transport function age. Nature Methods, 11(4), 361–362. https://doi.org/10.1038/ of Claudin-16. Journal of Clinical Investigation, 116(4), 878–891. nmeth.2890 https://doi.org/10.1172/jci26323 Shinoda, T., Shinya, N., Ito, K., Ohsawa, N., Terada, T., Hirata, K., … Konrad, M., Hou, J., Weber, S., Dötsch, J., Kari, J. A., Seeman, T., … Shirouzu, M. (2016). Structural basis for disruption of claudin as- Schaller, A. (2008). CLDN16 genotype predicts renal decline in fa- sembly in tight junctions by an enterotoxin. Scientific Reports, 6(1), milial hypomagnesemia with hypercalciuria and nephrocalcinosis. https://doi.org/10.1038/srep3​3632 Journal of the American Society of Nephrology, 19(1), 171–181. Simon, D. B., Lu, Y., Choate, K., Velazquez, H., Al-Sabban, E., Praga, https://doi.org/10.1681/asn.20070​60709 M., … Lifton, R. (1999). Paracellin-1, a renal tight junction pro- Konrad, M., Schaller, A., Seelow, D., Pandey, A. V., Waldegger, S., tein required for paracellular Mg2+ resorption. Science, 285(5424), Lesslauer, A., … Weber, S. (2006). Mutations in the tight-junc- 103–106. https://doi.org/10.1126/scien​ce.285.5424.103 tion gene claudin 19 (CLDN19) are associated with renal mag- Suzuki, H., Nishizawa, T., Tani, K., Yamazaki, Y., Tamura, A., Ishitani, nesium wasting, renal failure, and severe ocular involvement. The R., … Fujiyoshi, Y. (2014). Crystal structure of a claudin provides American Journal of Human Genetics, 79(5), 949–957. https://doi. insight into the architecture of tight junctions. Science, 344(6181), org/10.1086/508617 304–307. https://doi.org/10.1126/scien​ce.1248571 Martin-Nuñez, E., Cordoba-Lanus, E., Gonzalez-Acosta, H., Oliet, Tang, H., & Thomas, P. D. (2016). PANTHER-PSEP: Predicting dis- A., Izquierdo, E., & Claverie-Martin, F. (2015). Haplotype anal- ease-causing genetic variants using position-specific evolution- ysis of CLDN19 single nucleotide polymorphisms in Spanish ary preservation. Bioinformatics, 32(14), 2230–2232. https://doi. patients with familial hypomagnesemia with hypercalciuria and org/10.1093/bioin​forma​tics/btw222 10 of 10 |    GARCÍA-CASTAÑO et al. Wang, S.-B., Xu, T., Peng, S., Singh, D., Ghiassi-Nejad, M., Adelman, nephrocalcinosis caused byCLDN19gene mutations. Journal of R. A., & Rizzolo, L. J. (2019). Disease-associated mutations Medical Genetics, 54(1), 26–37. https://doi.org/10.1136/jmedg​ of claudin-19 disrupt retinal neurogenesis and visual function. enet-2016-103956 Communications Biology, 2(113). https://doi.org/10.1038/s4200​ Yamaguti, P. M., dos Santos, P. A. C., Leal, B. S., Santana, V. B. B. D. 3-019-0355-0 M., Mazzeu, J. F., Acevedo, A. C., & Neves, F. D. A. R. (2015). Waterhouse, A., Bertoni, M., Bienert, S., Studer, G., Tauriello, G., Identification of the first large deletion in the CLDN16 gene in a Gumienny, R., … Schwede, T. (2018). SWISS-MODEL: Homology patient with FHHNC and late-onset of chronic kidney disease: case modelling of protein structures and complexes. Nucleic Acids report. BMC Nephrology, 16(92). https://doi.org/10.1186/s1288​ Research, 46(W1), W296–W303. https://doi.org/10.1093/nar/ 2-015-0079-4 gky427 Weber, S., Schneider, L., Peters, M., Misselwitz, J., Rönnefarth, G., Böswald, M., … Konrad, M. (2001). Novel paracellin-1 mutations in How to cite this article: García-Castaño A, Perdomo- 25 families with familial hypomagnesemia with hypercalciuria and Ramirez A, Vall-Palomar M, et al. Novel compound nephrocalcinosis. Journal of the American Society of Nephrology, heterozygous mutations of CLDN16 in a patient with 12(9), 1872–1881. Retrieved from https://jasn.asnjo​urnals.org/conte​ familial hypomagnesemia with hypercalciuria and nt/12/9/1872.long Yamaguti, P. M., de Neves, F. A. R., Hotton, D., Bardet, C., de La Dure- nephrocalcinosis. Mol Genet Genomic Med. Molla, M., Castro, L. C., …Acevedo, A. C. (2017). Amelogenesis 2020;8:e1475. https://doi.org/10.1002/mgg3.1475 imperfecta in familial hypomagnesaemia and hypercalciuria with