Database of CLCN5 Pathogenic Variants Causing Dent Disease

Abstract

Introduction

Dent disease type 1 is an X-linked proximal tubulopathy caused by pathogenic variants in CLCN5, which encodes the chloride/proton exchanger, ClC-5. Loss of ClC-5 function disrupts receptor-mediated endocytosis, resulting in low–molecular-weight proteinuria, hypercalciuria, nephrocalcinosis, nephrolithiasis, and progressive kidney failure. Although many CLCN5 variants have been reported, existing resources remain incomplete, inconsistent, or based on different reference sequences (RefSeqs), thereby hindering variant comparison and clinical interpretation. A comprehensive and uniformly annotated catalog is needed to improve diagnostic accuracy and estimates of Dent disease type 1 prevalence.

Methods

Potentially pathogenic CLCN5 variants were compiled from firsthand Dent disease type 1 case reports, variant databases, Dent disease type 1–related online resources, and unpublished patients. Variants representing experimental constructs or benign or likely benign changes were excluded. All variants were standardized using The Human Genome Variation Society (HGVS) nomenclature and mapped to RefSeqs NM_001127898.4 and NM_000084.5 using Mutalyzer. Pathogenicity was assessed and cross-validated with the literature and public databases.

Results

We identified 524 unique pathogenic or likely pathogenic variants, exceeding all existing databases. Variant types included missense (31%), InDel insertions and deletions (37%), nonsense (14%), splicing defects (13%), and large deletions (5%). Most variants (74%) were reported only once. Exon 10 (NM_001127898.4) showed the highest density of pathogenic variants, driven by missense and in-frame insertions and deletions (InDel). Structural mapping revealed functional restrained regions in Helix H and O–Q helices. Source-variant counts indicated a minimum of 880 affected families, suggesting approximately 3520 globally.

Conclusion

This uniformly annotated catalog—the most comprehensive to date—enhances interpretation of CLCN5 variants, refines Dent disease type 1 burden estimates, and identifies key functional regions relevant for therapeutic development.

Graphical abstract

Dent disease type 1 is a proximal tubular disorder caused by mutations in the CLCN5 gene on the X chromosome encoding the chloride/proton exchanger, ClC-5.¹ ClC-5 plays an important role in receptor-mediated endocytosis in proximal tubule epithelial cells,2, 3, 4, 5, 6 and its deficiency impairs the functions of the proximal tubules to reabsorb small molecular proteins, water, and calcium filtered into the urine, causing proteinuria, hypercalciuria, nephrocalcinosis, and nephrolithiasis.⁷ Most patients with Dent disease type 1 develop life-threatening end-stage renal disease between the ages of 30 and 50 years.⁸ No cure is currently available for Dent disease type 1, except for supportive treatments that are associated with side effects and fail to halt disease progression.^8,9

Although it is a rare disease, Dent disease type 1 can be caused by hundreds of pathogenic variants. Mansour-Hendili et al.¹⁰ initially noted 234 unique Dent disease type 1–causing variants, including missense (∼ 33%), frameshift (∼ 29%), nonsense (∼ 18%), and splicing defects (∼ 12%). Following the 2015 compilation by Mansour-Hendili et al.,¹⁰ the 2020 review by Gianesello et al.¹¹ summarized 266 pathogenic CLCN5 variants. More pathogenic variants have been reported by many groups.12, 13, 14, 15 However, Dent disease type 1 case reports and variant reviewing papers often used different CLCN5 RefSeqs, including different versions of NM_001127898, NM_00084, and NM_00112789, describing the same variant differently, thus making it difficult to compare CLCN5 variants reported by different groups.

Several databases, including the National Center for Biotechnology Information clinical variant (ClinVar) database,¹⁶ the Human Gene Mutation Database (HGMD),¹⁷ and the Leiden Open Variation Database (LOVD)¹⁸ list CLCN5 variants. The ClinVar database includes 677 CLCN5 variants; only 230 are pathogenic or likely pathogenic (accessed in January, 2026). The current public domain of HGMD lists about 310 CLCN5 variants, and the information is not current. The Qiagen HGMD database, for which one must pay for access, was not comprehensive and current at the time of access (January 2026). The LOVD lists only 180 unique CLCN5 variants. Thus, there are currently no databases that are comprehensive and current on pathogenic CLCN5 variants. In addition, these databases used different CLCN5 RefSeqs for their listed variants, often giving the same variant a different description.

Accurate estimates of the number of families affected by Dent disease type 1 have been difficult. Recent papers,19, 20, 21, 22, 23, 24 the National Organization for Rare Disorders Rare Disease Database (https://rarediseases.org/rare-diseases/dent-disease/?utm_source=chatgpt.com#affected), the National Institutes of Health MedlinePlus Genetics (https://medlineplus.gov/genetics/condition/dent-disease/#frequency) and other Dent disease– related websites often cite an early estimate of about 250 families affected by Dent disease type 1 in the world.²⁵ With many more Dent disease cases reported, this number is no longer accurate. Thus, a better estimation of Dent disease prevalence is needed.

Obtaining a comprehensive spectrum of all pathogenic CLCN5 variants remains a significant challenge. Not only do databases use different RefSeqs to describe the same variant differently, but published papers use additional RefSeqs to describe CLCN5 variants. The ways of nucleotide numbering were inconsistent; in most papers the first nucleotide of the start codon was numbered as nucleotide 1, occasionally the first nucleotide of the 5’ UTR is numbered as nucleotide 1. Some variant descriptions were inconsistent with the evolving HGVS nomenclature guidance.^26,27 Typing errors were observed in variant descriptions. The lack of a comprehensive pathogenic CLCN5 variant database with a uniform description hinders diagnosis and examination of the roles of CLCN5 variation in Dent disease type 1. It is often difficult to tell whether newly observed CLCN5 variants from patients are novel or recurrent. Indeed, recurrent variants were often regarded as novel in the literature. Having an updated database for all potentially pathogenic CLCN5 variants with a uniform description will help us to get a full spectrum of pathogenic CLCN5 variants, appreciate the pathogenic role of newly diagnosed variants, and get a better idea of Dent disease type 1 prevalence.

To address this gap, we searched the literature and online databases and compiled a comprehensive database for potentially pathogenic CLCN5 variants. We noted 524 unique potentially disease-causing CLCN5 variants, the largest registration of pathogenic or likely pathogenic CLCN5 variants so far. We analyzed the mutation profiles, calculated the density of mutations related to exons and protein regions, and estimated the prevalence of Dent disease type 1 disease. Our analyses may help us to recognize ClC-5 regions with functional significance and find regions with high variant dense in CLCN5 cDNA.

Methods

Sources for Potentially Pathogenic CLCN5 Variants

The potentially pathogenic CLCN5 variants were from the following 4 sources: (i) first-hand Dent disease type 1 case reports and papers on Dent disease type 1 and CLCN5 mutation analysis and ClC-5 variant functional analysis indexed in PubMed; (ii) 4 online databases for human variants, including the ClinVar database (https://www.ncbi.nlm.nih.gov/clinvar/),¹⁶ the HGMD (https://www.hgmd.cf.ac.uk/ac/index.php),¹⁷ LOVD (https://www.lovd.nl/),¹⁸ and gnomAD v4.1.0²⁸; (iii) any website on Dent disease that lists CLCN5 variants, including OMIM²⁹; (iv) unpublished pathogenic variants observed in patients with Dent disease type 1 (Supplementary Table S1). All variants from these sources were recorded to minimize the possibility of missing validated pathogenic variants.

The following variants were then removed from the pathogenic variant list: (i) variants not observed in patients but created in the laboratory for functional studies; and (ii) variants observed in patients with Dent disease type 1 but are of known benign nature or unknown pathogenic significance (these are separately listed in Supplementary Table S2). Nonsense, canonical splicing and frameshifting variants from online databases were included in our list even if not associated with peer-reviewed papers, because they are certain to disrupt protein translation and likely pathogenic.

Four unpublished cases included in the database were confirmed through clinical genetic testing. Minimal anonymized clinical summaries are provided in Supplementary Table S1, consistent with institutional review board guidelines.

Compiling a List of Unique Potentially Pathogenic CLCN5 Variants

Because historical Dent disease type 1 literature primarily used reference NM_000084, and the exon numbering differs substantially from reference NM_001127898.4 (MANE Select) used by the most recent literature, all surviving variants from the above initial screens were described following the guidance of the HGVS nomenclature²⁶ based on both RefSeqs. To help researchers learn the relationship of the 2 references, we provided a mapping figure correlating cDNA and protein numbering and exon boundaries (Figure 1). All pathogenic variants in this study fall within coding regions shared by both transcripts; thus, protein-level interpretations and functional hotspot identification remain unchanged despite transcript choice. This effort provided 2 descriptions based on both references for all variants.

Figure 1.

Relationship between 2 major CLCN5 reference sequences. The upper panel compares the full-length cDNA RefSeqs NM_001127898.4 and NM_000084.5. Vertical dashed lines indicate exon boundaries; the number to the left of each dashed line denotes the last nucleotide of the preceding exon. The lower panel illustrates the corresponding protein RefSeqs NP_001121370.1 (816 amino acids) and NP_000075.1 (746 amino acids). Vertical solid lines mark the aligned amino acid positions between the 2 isoforms.

Some published variants matched RefSeqs other than the claimed one, judged by mismatch of both the cDNA and protein sequences with the claimed reference, but match of both with a different reference. Thus, these variants were described based on the matching reference, and new descriptions based on NM_001127898.4 and NM_000084.5 were generated.

Some original cDNA variant descriptions did not follow the 3’ rule of the HGVS nomenclature,²⁶ which recommends assigning the change to the most 3' position of the reference. The descriptions of these variants were corrected to follow the 3’ rule before generating new descriptions based on both references. Protein descriptions were predicted based on the nucleotide changes. The protein consequences of variants with splicing defects were uncertain for most variants and thus were indicated as “p.?”

Descriptions of all variants were verified using Mutalyzer³⁰ and the descriptions of > 30 variants were corrected following the HGVS nomenclature guidance.²⁶ Variants from different sources originally having different descriptions were grouped as the same one. Protein descriptions for > 80 variants from papers or websites were originally not standard or had errors, these descriptions were also corrected.

Obtaining a List of First-Hand Sources for the Variants

Only first-hand papers and databases were listed as sources of a specific variant. If a variant was found in online databases, websites, and CLCN5 variant reporting papers, only the PubMed Identifiers (PMIDs) of those firsthand variant reporting papers were included in the database as the source of that variant. Papers discussing previously published variants were not listed as sources for these variants. Multiple papers independently reporting the same variant observed in likely different patients with Dent disease type 1 were all listed as the sources for that variant. Online databases were listed as sources only if reporting truncating variants that are not reported in peer-reviewed papers.

Variant Analysis

CLCN5 variants with the following: (i) insertions or deletions (InDel-) that would disrupt the ClC-5 reading frame, (ii) nucleotide changes introducing premature stop codons, and (iii) canonical splicing defects, were defined as truncating variants. ClC-5 missense changes and short amino acid deletions or insertions were defined as nontruncating disease-causing ClC-5 variants if it was evaluated as pathogenic or likely pathogenic. Assignment of ClC-5 helices and domains was based on published models.31, 32, 33

Variant Pathogenicity Evaluation

Variants were categorized into a 5-tier classification according to American College of Medical Genetics and Genomics and Association for Molecular Pathology guidelines³⁴: pathogenic, likely pathogenic, uncertain significance, likely benign, and benign. Pathogenicity classification was done using Franklin (https://franklin.genoox.com).³⁵ When variant pathogenicity evaluation data were available in original sources, they were compared with our own evaluation. Discrepancies were carefully checked to justify our evaluation. Data on symptom, protein function, variant segregation and splicing effects, if available, were considered in pathogenicity classification. Variants with protein functional results^7,36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47 were indicated in the database. All variants classified by Franklin were further validated against the source literature, ClinVar, LOVD, and gnomAD to ensure concordance before final inclusion in the pathogenic or likely pathogenic dataset. Because the nonsense variant, p.Val798∗ was shown to be nonfunctional in electrophysiological experiments,⁴⁸ all variants causing truncations N-terminal to p.Val798, including nonsense variants, splicing defects and frameshifting InDel, were expected to produce a ClC-5 protein with impaired function. Several intronic variants outside ± 2 positions have been shown to impair normal splicing in minigene splicing assays49, 50, 51; these functional studies were considered in classification.

All variants classified as pathogenic or likely pathogenic were consolidated into a centralized database (Supplementary Table S3). Ten variants originally reported in patients with Dent disease type 1 or in individuals with kidney disease but lacking American College of Medical Genetics and Genomics interpretations in the original sources have been reassessed according to current American College of Medical Genetics and Genomics/Association for Molecular Pathology guidelines and population genetics data to be benign or of uncertain significance. Accordingly, they are excluded from the centralized pathogenic CLCN5 variant database but are listed in Supplementary Table S2 (With additional supporting information, some variants in Supplementary Table S2 could be reclassified as likely pathogenic or pathogenic).

Statistical Analysis

Student t tests were performed using GraphPad Prism 10 to examine whether the 2 groups were significantly different. P < 0.05 was regarded as significantly different. Outliers were identified using Robust regression and outlier removal method with a Q value of 1%.

Results

An Updated Database for Potentially Pathogenic CLNC5 Variants

We set out to compile a database with as many disease-causing CLCN5 variants as possible. For this purpose, we recorded all potentially pathogenic CLCN5 variants from papers on Dent disease type 1 case reports, Dent disease type 1 patient analysis and mutant functional analysis, removing all human-designed mutants not observed in Dent disease type 1 patients. In addition, we included 4 LOVD¹⁸-, 14 ClinVar¹⁶-, and 11 gnomAD²⁸-listed truncating variants not found in the literature, because these variations were expected to disrupt protein translation. Twelve truncating variants in the gnomAD database with stop codons before codon 70 of NP_001121370.1 (816AA) were excluded from the database because ClC-5 of 746AA (NP_000075.1) is fully functional (the relationship of the 2 ClC-5 variants is presented in Figure 1), and no patients with Dent disease type 1 with truncations before codon 70 of NP_001121370.1 have been observed (Ludwig et al.⁵² and our observations).

Because RefSeqs, NM_001127898.4 (MANE cDNA RefSeq) and NM_00084.5, have both been used in databases and publications for variant description, all variants were described based on both RefSeqs following the guidance of the HGVS Nomenclature.²⁶ Including NM_00084.5 as RefSeq makes it easy to connect variants to ClC-5 variant functional studies because most of these studies used NM_00084.5 and its protein RefSeq (NP_000075.1, 746AA) for description.

Altogether, we noted 524 unique CLCN5 variants in our database, all classified as pathogenic or likely pathogenic based on the American College of Medical Genetics and Genomics guidelines (Supplementary Table S3). Ten variants reported in papers or databases that were classified as of uncertain significance, likely benign or benign, were excluded from our comprehensive pathogenic database but are listed in Supplementary Table S2.

Previous compilations have listed < 300 pathogenic CLCN5 variants using different references.^10,11,13 Our present work expands this number to 524 and harmonizes all variants under MANE Select NM_001127898.4 and NM_000084.5. In addition, we associate all firsthand publications with the reporting variant and link all functional studies (if available) to each variant. We standardize cDNA and protein descriptions and eliminate redundant variants because of description inconsistency. We make this comprehensive and centralized database searchable to the public through website (https://www.DentDiseaseAtlas.com). To our knowledge, this is the largest pathogenic or likely pathogenic CLCN5 variant database, with >200 variants more than any other public database.

The variants included missense (31%), InDel (37%), nonsense (14%), splicing defect (13%), and large deletion (5%) (Figure 2). These numbers are similar to previously reported databased on smaller numbers of variants.10, 11, 12, 13, 14, 15 Except for 15 (3%) in-frame InDel variants causing the deletion or insertion of a few amino acid residues, all variants with InDel, nonsense changes, canonical splicing defects, and large deletions are expected to cause the lack of a full-length ClC-5 protein (63%).

Figure 2.

Percentage of different types of pathogenic CLCN5 variants.

Low Prevalence of CLCN5 Variants

These pathogenic or likely pathogenic CLCN5 variants were reported in 166 peer-reviewed firsthand CLCN5 variant reporting papers and websites. Three hundred eighty-six of the 524 variants (74%) in our database have only 1 reference. Only 10 variants were reported in > 10 peer-reviewed publications (Table 1). Attempts to use allele frequency data from the gnomAD4 database to estimate the Dent disease type 1 prevalence is not fruitful, because only 6 of the 524 pathogenic and likely pathogenic CLCN5 variants were observed in male genomes, each observed once in 400,897 male genomes. These observations suggest that pathogenic CLCN5 variants have low occurrence in human populations.

Table 1.

Recurrent CLCN5 variants reported in more than 10 papers

cDNA variation	Protein variation	Mutation type	Reporting publications (PMID)
2152C>T (CpG)	p.Arg718∗	Nonsense	16822791, 9853249, 10469281, 9259268, 38267993, 18976849, 36578800, 35100899, 9853249, 25907713, 8559248, 24081861, 32683654, 19806368, 12631345, 27889724, 33545640, 15895257, 20430975, 40420588 (20)
941C>T (CpG)	p.Ser314Leu	Missense	24081861, 8559248, 32683654, 9259268, 27889724, 9734595, 7915957, 9187673, 9452997, 20654030, 21955393, 32393202, 16807762, 34650943, 28580211, 14584882, 20804101, 40420588, 41361832 (19)
1249C>T (CpG)	p.Arg417∗	Nonsense	16822791, 9596078, 35100899, 11136179, 18976849, 9328929, 15814539, 25907713, 15086899, 9328941, 31674016, 26695661, 15895257, 36726441, 13679301, 16034421, 41361832 (17)
2119C>T (CpG)	p.Arg707∗	Nonsense	16822791,28815356, 11136179, 31674016, 18976849, 25907713, 15895257, 24912603, 18038239, 15719255, 19546586, 40133227, 29058463, 40420588, 41361832 (15)
2320C>T (CpG)	p.Arg774∗	Nonsense	10469281, 9893114, 8559248, 27625851, 9853249, 24810952, 32203225, 24081861, 27174143, 15086899, 32683654, 15895257, 20430975, 41361832 (14)
310C>T (CpG)	p.Arg104∗	Nonsense	24081861, 9734595, 32683654, 16807762, 10469281, 30581818, 31947599, 16822791, 28815356, 9734614, 27625851, 39327974 (12)
1609C>T (CpG)	p.Arg537∗	Nonsense	16822791, 19076289, 25907713, 15895257, 31674016, 18184518, 19546586, 24081861, 32683654, 29084614, 31597132, 38726999 (12)
1745-2A>C	p. ?	Splicing defect	24912603, 32495484, 11136179, 15895257, 18038239, 19579036, 24081861, 24912603, 27889724, 19076289, 17982249 (11)
c.1756C>T	p.Arg586Trp	Missense	9328929, 15086899, 19546591, 24081861, 24912603, 32683654, 33532864, 38726999, 13679301, 40420588, 41291521 (11)
c.1776_1778del	p.Val593del	INDEL	19546591, 33987651, 27625851, 31674016, 24081861, 25907713, 40133227, 24912603, 37284679, 32683654, 41361832 (11)

PMID, PubMed identifiers.

∗ indicates a stop codon.

We reviewed published data on the geographic distribution of patients carrying the c.2152C>T variant, which is the most frequently reported mutation across 20 publications (listed as the first variant in Table 1). These most likely nonoverlapping patients were from 12 different countries, and most likely do not have relationships. We reason that this CLCN5 variant most likely arose independently in some of these affected families.

Seven of the 10 most frequently reported variants involve CpG to TpG transition. This is not surprising because the cytosine in the CpG context can be methylated to form 5-methylcytosine, which undergoes spontaneous deamination and converts into thymine.⁵³ Seven of the 10 variants cause nonsense mutations or splicing defects that disrupt protein translation, and only 2 are missense mutations, consistent with the fact that truncating variants are more likely to disrupt protein function.

Estimating the Number of Families Diagnosed With Pathogenic CLCN5 Variants

Only 6 of the 524 unique pathogenic and likely pathogenic CLCN5 variants have male allele frequency > 0 in gnomAD v4.1.0. Therefore, it is difficult to estimate the number of patients based on allele frequencies. We explored a different strategy to get a better idea of the number of families diagnosed with pathogenic CLCN5 variants. We recorded 880 different source∗variants, where sources could be variant-reporting papers or databases, or our unpublished observations. Because it is very unlikely that a patient would perform genetic tests multiple times in different clinical centers, this number of 880 most likely reflects the minimum number of families diagnosed with pathogenic CLCN5 variants. We reason that this is possibly an underestimate of the number of families diagnosed, because not all pathogenic variants have been published following diagnosis. For example, the c.603G>A variant (Supplementary Table S1) from a family approaching us for consultation was not published following diagnosis.

The vast majority of Dent disease type 1 case reports were from the United States, Europe, Japan, Israel, Korea, and China, which account for about one-fourth of the world population. If pathogenic CLCN5 variants are present in people from different countries with similar frequency, then about 3520 (880 × 4) families can be affected by Dent disease type 1 worldwide. The affected individuals could be significantly larger than the affected family numbers, depending on the number of individuals affected by Dent disease type 1 in each family.

CLCN5 Exon 10 Has the Highest Density of Pathogenic CLCN5 Variants

We observed pathogenic CLCN5 variants in most protein-coding exons (exons 5–15) except for exons 3 and 4 (NM_001127898.4) (Figure 3a). We reason that this is most likely because exons 3 and 4 of NM_001127898.4 fall in the intron of alternatively spliced CLCN5 transcript, NM_000084.5, which encodes a fully functional ClC-5 of 746AA (NP_000075.1; full correspondence is presented in Figure 1).⁵²

Figure 3.

Analysis of pathogenic CLCN5 variants in exons. (a) Number of pathogenic variants in coding exons. Large deletions involving multiple exons are not counted. (b) Pathogenic variant densities of coding exons normalized by exon size. (c) Densities of truncating variants of coding exons. (d) Densities of nontruncating disease-causing ClC-5 variants for coding exons. # indicates outlier in Robust regression and Outlier removal analysis.

Exon 10 has the highest variant density of 50.6 variants/100 nt (Figure 3b), almost 3 times more than the average of all protein-coding exons (17.8 variants/100 nt). To determine whether the high pathogenic CLCN5 variant density in exon 10 is caused by features of the DNA sequence in the region, we compared the distribution of truncating variants (nonsense mutations and frame shifting mutations) that disrupt protein expression regardless of the function of the encoding protein domain. The density of truncating variants was not increased in exon 10 (Figure 3c), suggesting that the high variant density in exon 10 is not caused by the intrinsic feature of the DNA sequence. However, the high pathogenic variant density of exon 10 is caused by the high frequency of missense variants and in-frame insertions/deletions (Figure 3d), which may impair ClC-5 functions. Exon 10 has 4 times more nontruncating disease-causing ClC-5 variant density (35.8 nontruncating disease-causing ClC-5 variants/100 nt) than the average of all other exons (7.2 nontruncating disease-causing ClC-5 variants/100 nt). Robust regression and outlier removal analysis identified exon 10 as an outlier of all exons. Therefore, high pathogenic CLCN5 variant density in exon 10 is most likely caused by ClC-5 functional restraint but not by unique DNA sequence features.

Helix H has the Highest Density for Pathogenic CLCN5 Variants

ClC-5 has 18 helices³¹ and 2 intracellular cystathionine β-synthetase (CBS) domains³² (Figure 4a). To determine whether high nontruncating disease-causing ClC-5 variant density in exon 10 reflects functional constraints of the protein regions (helix H and the majority of helix G) encoded by exon 10, we compared nontruncating disease-causing ClC-5 variants in each known ClC-5 domain. Helix H has the highest nontruncating disease-causing ClC-5 variant density of 200 variants/100 residues, followed by helices O to Q (75.9), helices F and G (56.8), helix D (55.6) and helix N (45.5), all greatly higher than the average of all domains (22.0 variants/100 residues) (Figure 4b). Robust regression and outlier removal analysis identified helix H as an outlier. Plotting different types of pathogenic variants along ClC-5 protein, it is clear that whereas nontruncating disease-causing ClC-5 variants show a clear high density in these regions, other types of variants such as nonsense variants, InDel, and splicing defects do not show such trend (Figure 5a). Missense mutations on 9 of 11 positions of helix H are pathogenic, except for I328 and F334 (Figures 4a and 5b). Seven positions in helix H have >1 pathogenic variant (Figure 5b).

Figure 4.

Distribution of nontruncating disease-causing CLCN5 variants in different domains of ClC-5. (a) Diagram showing the ClC-5 domains and positions with pathogenic variants. The 18 helices and 2 CBS domains are shown. Positions with nontruncating disease-causing ClC-5 variants are indicated. (b) Nontruncating disease-causing ClC-5 variant densities for ClC-5 Helices and domains. Single letters indicate helices. Two letters indicate the sequences between the 2 helices. Outliers identified by Robust regression and Outlier removal analysis was marked by #. ∗∗ indicates P < 0.05 comparing the indicated region with all regions of the protein.

Figure 5.

Distribution of pathogenic CLCN5 variants along ClC-5 protein. (a) Distribution of different types of pathogenic CLCN5 variants along ClC-5 protein. The helices with high nontruncating disease-causing CLCN5 variant densities are indicated. (b) Nontruncating disease-causing ClC-5 variants in helix H. The highlighted variant has 1 amino acid deletion and 2 amino acid insertions. A dash indicates a single amino acid deletion. (c) Nontruncating disease-causing ClC-5 variants in helices O–Q. For (b and c), the wildtype sequence is in green. Numbering is based on NP_001121371.1 (816AA).

Helices O to Q have the second highest density of nontruncating disease-causing ClC-5 variants (75.9 variants/100 residues), which is significantly higher than the average of all regions (22 variants/100 residues; P < 0.01, t test). Changes in half of the amino acid residues (28 in 54) in this region cause ClC-5 functional defects (Figure 5c). The nontruncating disease-causing ClC-5 variant density decreased abruptly in sequences immediately adjacent to helices O to Q. Studying a smaller cohort of nontruncating pathogenic variants, Lourdel et al.⁵⁴ observed that many pathogenic missense variants involve amino acids in helices O to Q forming the dimer interface of ClC-5 homodimers. Our analysis based on a much larger database confirmed the original findings. In addition, helices F and G (56.8 variants/100 residues), helix D, and helix N have relatively high nontruncating disease-causing ClC-5 variant density (Figure 4b). Possible explanations will be discussed later in this article.

A Website for Pathogenic or Likely Pathogenic CLCN5 Variants

We developed a dedicated website (https://www.DentDiseaseAtlas.com) to facilitate access to pathogenic and likely pathogenic CLCN5 variants for Dent disease type 1 researchers and patients. The database is searchable by cDNA notation, protein consequence (3 letter code), PMID, or relevant keywords (Figure 6a). Direct links to PMIDs and external resources are provided to help users easily access original publications and related websites. Functional studies of variants are also included, with PMIDs linked to their sources.

Figure 6.

A database for variants causing Dent disease type 1. (a) Pathogenic or likely pathogenic CLCN5 variants search results. (b) Newly diagnosed CLCN5 variant submission form.

In addition, the site includes a feedback feature that allows users to submit newly observed CLCN5 variants—whether novel or recurrent (Figure 6b). Although recurrent variants does not expand the variant list, their submission helps us better estimate the number of affected families. After expert review, variants reported from new patients with Dent disease type 1 will be incorporated into the database, ensuring it remains a dynamic and continuously updated resource.

Discussion

Here, we compiled an updated database of 524 unique likely pathogenic or pathogenic CLCN5 variants. This new database is currently the most comprehensive one, with >200 more likely pathogenic or pathogenic CLCN5 variants than any other database, including the ClinVar database, the HGMD, the LOVD, and any published pathogenic CLCN5 variants.10, 11, 12, 13, 14, 15 Our database has more pathogenic CLCN5 variants than the variant aggregator, Varsome database, which lists 190 pathogenic variants (131 pathogenic and 59 likely pathogenic), 197 variants with uncertain significance, 65 likely benign variants, and 10 benign variants.⁵⁵ The availability of a comprehensive list of potentially pathogenic variants described according to the MANE RefSeq (NM_001127898.4) and the RefSeq often used in early publications and functional studies (NM_00084.5) makes it possible to compare newly diagnosed CLCN5 variants against all known pathogenic ones, facilitating the evaluation of the possible roles of the variants in Dent disease type 1 pathogenesis. For a disease having >500 potentially pathogenic variants with no dominant ones, such a database is especially useful. With the PMIDs or the websites reporting the variants listed, it is convenient for the users to read the original sources for more information. In addition, the PMIDs of functional studies, if available, are listed with the variants, helping researchers and doctors to evaluate newly diagnosed variants.

We reasoned that the 880 CLCN5 variants reflect the lower limit of the number of families diagnosed with potentially pathogenic CLCN5 variants. This number is >3 times the 250 families often cited in recent papers and websites. We estimate that the families affected by Dent disease type I in the world could reach about 3520 (880 × 4), and this estimate could be the lower limit of families affected by Dent disease type 1 worldwide. Nationwide registries from Japan,⁵⁶ France,¹⁰ and Great Britain,⁵⁷ suggest a Dent disease type 1 prevalence of 1 in 400,000 to 1 in 1 million.⁵⁸ Based on this prevalence, Dent disease type 1 is estimated to affect 8000 to 20,000 people in the world’s population of 8.1 billion. Our estimate of 3520 Dent disease type 1 affected families is consistent with this estimate from a different method.

Our estimate assumes that our variant counts approximate distinct families. However, recurrent variants may arise through independent mutational events or founder effects. Therefore, though approximately 880 families represent a conservative lower limit, the true number may vary, depending on the relative contribution of these mechanisms.

The 524 unique potentially pathogenic CLCN5 variants are distributed in all protein-coding exons except for exons 3 and 4. Most variants were only diagnosed in a single family, and even the most frequently reported variant was reported in ≤20 families. The large number of pathogenic variants, the scattered distribution of the variants, and the small number of patients for each variant make developing Dent disease type 1 gene editing therapies, including base editing^59,60 and prime editing,⁶¹ impractical because these therapeutic strategies are sequence-specific. Techniques capable of targeted integration of large DNA, such as PASTE⁶² and PASSIGE,⁶³ could be used to insert the full-length wild-type CLCN5 cDNA into the locus. However, the efficiency of targeted integration of large DNA fragments has yet to be improved for in vivo applications. CLCN5 Exon 13 has the most pathogenic variants of 88 with 168 source∗variants (corresponding to 168 families) and is short enough (399 bp) to be relatively efficiently replaced by CRISPR/Cas9–mediated targeted integration. Patients with variants in this exon may benefit from exon replacement therapy by targeted integration.

Developing gene supplementation therapy for Dent disease type 1 is currently the most practical strategy, because one therapy may benefit the majority of patients with Dent disease type 1. Supporting this view, our recent work in Dent disease type 1 mouse models showed that gene supplementation therapy is effective in CLCN5 null variants.^64,65 It remains to be determined whether Dent disease type 1 caused by missense variants can also benefit from gene supplementation therapy.

Nontruncating disease-causing ClC-5 variant density in ClC-5 domains followed the order of helix H > helices O to Q > helices F and G > helices D and N (the underlying protein regions exhibiting high functional constraint—namely helices H, O–Q, F, G, D, and N—are unequivocally defined and their biological significance is independent of transcript choice). This observation can be explained by structural and functional constraints. Structure studies found that helices H, I, P, and Q are the 4 major helices that are involved in the formation of the interface between the ClC-5 homodimers.^31,33,66,67 These observations explain our observation of high pathogenic density in helix H and helices O–Q. Helix H not only plays an important role in homodimer interface formation, but also hosts the proton glutamate (E268, numbering based on ClC-5 746AA) important for Cl(−) and H(+) transport,⁶⁸ organizing a water proton pathway⁶⁹ and regulating transport probability.⁷⁰ In addition, G261 in helix H is important to form a normal kink in helix H.⁶⁹ Further, it was found that mutations in helix H affected the expression or transport function of ClC-5.³⁶ The functional restraints, in addition to structural roles, are consistent with helix H’s sensitivity to amino acid changes. To the best of our knowledge, this is the first observation that helix H is the functional hotspot of variation in ClC-5 because of functional restraints.

Helices F and G are part of the ion transport pathway.³³ E211 of ClC-5 helix F is the gating glutamate coupling movement of protons to Cl⁻ flux.^4,5 G212 of ClC-5 helix F, which is equivalent to G149 of ecCLC helix F, is part of the anion transport pathway.³³ Helix D is part of the anion transport pathway,³³ and the S238 immediately N-terminal to helix D is important for determining anion specificity.⁷¹ Helix N forms the anion-selectivity filter together with helix F.³³ Thus, functional restraints may also explain the high nontruncating pathogenic variant density in helices F and G, D, and N.

In summary, we compiled an updated database containing 524 unique potentially pathogenic CLCN5 variants, >200 more than current databases or publications on the same topic. The database presents the full spectrum of all potentially pathogenic CLCN5 variants for the first time. The comprehensive pathogenic CLCN5 variant database with uniform descriptions will be very useful for researchers and clinicians on Dent disease type 1 to evaluate the pathogenicity of variants diagnosed in patients with Dent disease type 1. Our counting of source∗variants from firsthand papers and databases helps us to estimate the number of families affected by Dent disease type 1, an estimate close to estimates using a different method.^10,56, 57, 58 The ClC-5 helices we identified with high pathogenic variant densities (helices H, O–Q, F & G, D, and N) coincide with their functional importance in homodimer interface formation^31,33,66,67 and ion transport.^4,5,33,68, 69, 70, 71

Disclosure

All the authors declared no competing interests.

Supplementary Material

Supplementary File (PDF)

Table S1. Unpublished pathogenic variants observed in patients with Dent disease type 1. Table S2. Variants reported from patients classified as VUS or benign. Table S3. All pathogenic and likely pathogenic CLCN5 variants.