Background

Mature enamel is a thin outer protective layer and covers the crown of the tooth in the form of a shell [1]. Naturally, it is tough, hard, and highly mineralized translucent human tissue produced by ameloblasts and is epithelial in its origin [2]. The biochemical architecture of dental enamel is of crystals of substituted calcium hydroxyapatite (96%), and the 4% is of organic matter and water [3]. Amelogenesis is a highly intricate biomineralizing process controlled by the expression of several genes [2]. AI affects both the primary and permanent dentition with exceptionally variable severity of the disease conditions [4, 5].

Various accounts of both syndromic and non-syndromic/isolated cases of AI have been published in the literature. Depending upon the amount, structure, and composition of the dental enamel, the phenotypes of non-syndromic AI are highly variable and may be divided into hypoplastic, hypocalcified, and hypomaturation forms [3, 4].

To date, pathogenic variants causing non-syndromic AI have been identified in 20 genes at various chromosomal locations [3], including AMELX (OMIM 300391; Xp22.2), a candidate gene for X-linked dominant hypoplastic AI (OMIM: 301200) [6], encoding an enamel matrix protein (EMPs) called amelogenin and making up to 90% of the ameloblast secreted EMPs [7, 8]. ENAM (OMIM 606585; 4q13.3), encoding the largest EMP called enamelin, a tooth specific protein expressed by ameloblasts, causing an autosomal recessive (OMIM: 204650) and dominant forms of AI (OMIM 104500) [9, 10]. AMBN (OMIM 601259; 4q13.3) encodes a glycine, leucine, and proline-rich enamel matrix protein called ameloblastin, a second most abundant protein expressed during amelogenesis. AMBN associated AI segregates in an autosomal recessive fashion [11, 12]. LAMB3 (OMIM 150310; 1q32.2), LAMA3 (OMIM 600805; 18q11.2), COL17A1 (OMIM 113811; 10q25.1), ITGB6 (OMIM 147558; 2q24.2) and ACPT (OMIM: 606362; 19q13.33) are other genes that cause hypoplastic AI in their altered forms [13,14,15,16,17,18,19,20]. Mutations in FAM83H (OMIM 611927; 8q24.3) cause an autosomal dominant hypocalcified type of AI [6, 21]. However, SLC24A4 (OMIM 609840; 14q32.12), WDR72 (OMIM 613214; 15q21.3), MMP20 (OMIM 604629; 11q22.2), KLK4 (OMIM 603767; 19q13.41) and GPR68 (OMIM 601404; 14q32.11), cause autosomal recessive hypomaturation type of AI [6, 22,23,24,25,26]. MMP20 (OMIM 604629; 11q22.2) and KLK4 (OMIM 603767; 19q13.41) are the two proteinases secreted at the time of enamel formation [27]. Nevertheless, in the case of C4orf26 (OMIM 614829; 4q21.1), and AMTN (OMIM 610912; 4q13.3) mutations cause autosomal recessive and dominant hypo-mineralized amelogenesis imperfecta, respectively [28, 29]. Recently, RELT (OMIM 611211; 11q13.4) variants are identified, causing hypocalcified amelogenesis imperfecta type IIIC [30].

Occasionally, AI has been reported as a part of a syndrome. The most common of them include Tricho-Dento-Osseous (TDO; OMIM 190320) syndrome (DLX3, OMIM 600525), Laryngo-Onycho-Cutaneous (LOC; OMIM 245660) syndrome (LAMA3, OMIM 600805), Jalili syndrome (JS; OMIM 217080) (CNNM4, OMIM 607805), Amelogenesis Imperfecta and Nephrocalcinosis (OMIM 204690) (FAM20A, OMIM 611062), Kohlschutter-Tonz Syndrome (KTS; MIM 226750) (ROGDI, OMIM 614574), Amelo-Onycho-Hypohidrotic Syndrome (OMIM 104570), and Heimler Syndrome-1,2 (HMLR; OMIM 234580) (PEX1, PEX6, OMIM 602136, 601,498).

Here, we report a novel nonsense variant c.1192C > T (p.Gln398*) in exon-12 of SLC24A4 in non-syndromic AI patients in a family of Pakistani origin.

Methods

Patients recruitment, pedigree construction, and DNA extraction

The recommendations of the Declarations of Helsinki were strictly followed for the approval of the study from the Research and Ethical Committee of Kohat University of Science and Technology (KUST), Khyber Pakhtunkhwa, Pakistan. Informed written consent was obtained from the affected and unaffected participants. A five generational pedigree diagram was constructed after a thorough interview of the unaffected mother (III-4). The pedigree showed an autosomal recessive mode of inheritance (Fig. 1A). Venous blood samples were collected from seven members of the family, including two patients (IV-4, IV-5) and five phenotypically unaffected individuals (III-4, IV-1, IV-3, IV-7, V-1). The extraction of genomic DNA from whole peripheral blood was performed by using the GeneJET Genomic DNA extraction Kit (Thermo-scientific, Lithuania), strictly following the manufacturer’s protocol.

Fig. 1
figure 1

(A) Pedigree of the family where SLC24A4 variant c.1192C > T segregates in an autosomal recessive fashion. The asterisks show tested individuals. The red arrow indicates the index patient, who was subjected toexome sequencing. T shows the disease-allele while C is the wild-type presentation (B) The representation of amelogenesis imperfecta in the patients. (i), (ii), and (iii) are the clinical features of patient IV-4 showing yellow-brown discoloration, (iv) Orthopantomogram (OPG) of the patient IV-4 showing thin layer of enamel, high radio-density and distinction from the dentin, while (v) and (vi) are the clinical photographs of the patient IV-5 showing creamy type of discoloration, attrition and dental caries

Full size image

Exome sequencing, and validation of rare variants through DNA sequencing

DNA (70 ng/μl) of an affected member (IV-4) was subjected to exome sequencing. The enrichment of DNA for the intron-exon boundaries was carried out with the SeqCap EZ human exome library v2.0 kit. The Illumina HiSeq 4000 sequencing machine via a paired-end 100-bp protocol [31] was used to run the generated libraries. The Cologne Center for Genomics (CCG) Varbank pipeline v2.26 (https://varbank.ccg.uni-koeln.de/) was used for exome data analysis. The mean coverage of the data was 77%, while at 20X and 10X, the coverage of the targeted bases was 92.6 and 96.6%, respectively. Genome Aggregation Database (gnomAD; https://gnomad.broadinstitute.org/) was consulted to establish the minor allele frequency (MAF; value < 0.01) of the variants. As controls, an in-house database of 511 exomes, and a dataset of 65 exomes from the Pakistani population, including 44 exomes from Punjabi, Sindhi, and Balochi patients, and 21 exomes of ethnically matched Pakhtoon patients were also consulted. The rare variants in PSPH, CHCHD2, BNC2, and SLC24A4 were selected from the exome data and were considered for the co-segregation analysis. The online prediction tools like MutationTaster, PROVEAN, SIFT, and PolyPhen2.0 were used to predict the pathogenicity of the missense variants (Table 1). The reference sequences of PSPH, CHCHD2, BNC2, and SLC24A4 (NM_004577.3, NM_016139.2, NM_017637.5, NM_153646.3, respectively) were obtained from the University of California Santa Cruz (UCSC) genome database browser (http://genome.ucsc.edu/cgi-bin/hgGateway). Primer3Plus software (http://www.bioinformatics.nl/cgi-bin/primer3plus/primer3plus.cgi) was used for designing the primers for the amplification of the regions of interest. A nucleotide sequence of 600 bp up-and-downstream from the position of the rare variants was searched to find out a suitable primer pair (Table 1). PCR amplified the regions of interest and the Exo-Sap protocol (https://www.thermofisher.com) was used for purifying the PCR products. The DNA sequencing was performed on the ABI3730 genetic analyzer with BigDye chemistry v3.1. The sequence alignment against the reference sequence was carried out by a sequence alignment tool, BioEdit version 6.0.7 (http://www.mbio.ncsu.edu/BioEdit/bioedit.html).

Table 1 Rare variants extracted from the exome sequencing data of patient IV-4 and primer sequences for the respective variants
Full size table

Results

Clinical and radiological investigations

For clinical and radiological investigations, a 35-years old patient (IV-4) was referred to Khyber Medical College of Dentistry, Peshawar, Khyber Pakhtunkhwa, Pakistan. His major complaints were yellow-brown staining, eating, and chewing difficulties of all the teeth (Fig. 1B-i,ii,iii). The patient presented no complications of other body organs during the clinical evaluation. The Orthopantomogramm (OPG) of this patient showed a thin (hypoplastic) mandible with missing posterior teeth on the right side and carious molars with a periapical infection on the left side. The maxilla showed impacted canine in the right premolar region with a missing molar and spacing among the dentition on the right side of the arch.

Additionally, the teeth showed generalized horizontal bone loss, more prominent around the maxillary molars. OPG also showed the presence of a thin layer of enamel, especially in the region of molars of the upper jaw. Furthermore, enamel appeared to have higher radio-density compared to the dentin. Moreover, the dentin appeared normal and distinct from the enamel (Fig. 1B-iv).

Patient IV-5, the 27-year old brother of patient IV-4, presented with creamy discoloration and attrition of the frontal maxillary teeth while dental caries in the mandibular premolars and molars (Fig. 1B-v,vi).

Screening of pathogenic sequence variant

Exome sequencing revealed rare homozygous variants in four genes: PSPH (OMIM 172480; Exon-6, c.398A > G; p.Arg133Ser), CHCHD2 (OMIM 616244; Exon-3, c.418G > A; p.Val140Met), BNC2 (OMIM 608669; Exon-7, c.2860G > A; p.Ala954Thr), and SLC24A4 (c.1192C > T; p.Gln398*). These variants lie in three regions of homozygosity (ROH) on chromosome 7 (11.6 MB), 9 (3.8 MB), and 14 (4.7 MB). The variants in CHCHD2 and SLC24A4 are neither reported in gnomAD nor HGMD (Human Gene Mutation Database; http://www.hgmd.cf.ac.uk/ac/index.php). Both variants in PSPH and BNC2 are tremendously rare in gnomAD, where c.398A > G; p.Arg133Ser appears in 20 alleles out of 282,490 alleles (none homozygous) and c.2860G > A; p.A954T is found in 4 alleles (one is homozygous) out of 246,026 alleles. These variants are not identified in the in-house database of 511 exomes and 65 exomes of Pakistani patients with diverse phenotypes other than AI. The pathogenicity predictions of the variants in PSPH, CHCHD2, and BNC2 by four online prediction algorithms are described in Table 1.

Sanger sequencing was used to check the segregation of these variants with the disease. The homozygous missense variants in PSPH, CHCHD2, BNC2 did not segregate within the family while the homozygous nonsense variant (c.1192C > T; p.Gln398*) in SLC24A4 revealed its co-segregation in the family (Fig. 2A). The DNA sequencing results of this cohort showed three forms of genotypes for this variant, heterozygous (C/T) (III-4, IV-3, IV-7), homozygous (C/C) wild-type (IV-1, V-1) and homozygous (T/T) mutant (IV-4, IV-5) (Fig. 1A). A ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/variation/689492/) accession number (VCV000689492.1) for this variant has been allocated.

Fig. 2
figure 2

a Chromatograms of an unaffected individual (IV-1) in the upper panel, a carrier (III-4) member in the middle panel and an affected individual IV-4 in the lower panel. b Hypothetical structure of SLC24A4 containing all 17 exons, showing the positions of genetic alterations in the previous studies as well as in the present study (red)

Full size image

Exome data did not expose any rare variant in other genes (AMELX, ENAM, AMBN, LAMB3, LAMA3, COL17A1, ITGB6, ACPT, FAM83H, WDR72, MMP20, KLK4, GPR68, RELT, DLX3, CNNM4, ROGDI, PEX1, and PEX6) reported so far, to cause syndromic and non-syndromic AI.

Discussion

Five functionally different types of K+-dependent Na+/Ca+ 2 exchangers (NCKX1–5) have been characterized in humans [32, 33]. NCKXs are bidirectional membrane transporters; for example, NCKX4 transports an intracellular Ca+ 2 and a K+ ion in exchange for four extracellular Na+ ions [34]. Each NCKX protein has a unique role in various biochemical pathways governing the vision, olfaction, and skin pigmentation [35]. During the maturation stage of tooth development, SLC24A4 (NCKX4) is involved in the active transport of Ca+ 2 ions from ameloblasts into the enamel matrix. Genetic alterations in SLC24A4 in the human genome and its knockout mice Slc24a4−/− lead to the development of indisposed calcified enamel [36]. Clinical findings of Slc24a4−/− mice signify the essential role of this protein in enamel development [25].

SLC24A4 (OMIM 609840) encodes a protein of 622 amino acids, called solute carrier family 24 member 4 (SLC24A4), which is one of the members of K+-dependent Na+/Ca2+ exchanger family (SLC24A), comprising a total of five members. It has been mapped to the chromosome 14q32 [33, 36]. SLC24A4 has various transcripts (NM_153646, NM_153647, NM_153648) resulting from alternative splicing and the longest isoform (NM_153646) contains 17 coding exons. SLC24A4 is highly expressed in many types of tissues, such as aorta, brain, lungs, and thymus gland [34]. In the case of developing dentine, SLC24A4 is expressed in ameloblasts, and it borders to the membrane in contact with the developing enamel [37]. The predicted structure for full-length SLC24A4 protein consists of 11 transmembrane helices having two highly conserved transmembrane clusters (consisting of 5 transmembrane helices) linked together by an intracellular (cytoplasmic) loop. The Na+/Ca2+ exchanger domains are composed of these transmembrane pockets. Each domain contains a hydrophobic and highly conserved region of 30–40 residues called alpha-1 (139–179 amino acids), and alpha-2 (495–526 amino acids) repeats, respectively, which form ion-binding regions after undergoing highly intricate interactions with each other [25, 38].

We have identified a novel nonsense variant (c.1192C > T; p.Gln398*) in exon-12 of SLC24A4 by using exome sequencing. This unusual genetic alteration is expected to lead to the loss of function of SLC24A4 protein either by nonsense-mediated decay (NMD) or by the production of a truncated protein lacking the C-terminus. Since this nonsense variant introduces a premature stop codon at the position 398 in the cytoplasmic loop between the alpha-1 and alpha-2 repeats; hence the loss of remaining 225 amino acids (containing the alpha-2 repeat) is predicted. The two Na+/Ca2+ exchanger domains (alpha-1 and alpha-2 repeats) are crucial for the smooth transport of ions, which verifies the exceptional role of SLC24A4 during amelogenesis. The absence of one of the two Na+/Ca2+ exchanger domains, in this case, alpha-2-repeat only will ultimately render the protein nonfunctional and causes amelogenesis imperfecta, hypomaturation type AI2A5 (OMIM: 615887) phenotype [25].

To date, a total of five pathogenic variants causing AI have been identified in the SLC24A4, including three missense variants, one nonsense variant, and a gross deletion (Fig. 2B). Parry et al. in 2013 screened 15 Pakistani families and identified two homozygous variants in SLC24A4, including a missense c.1495A > T (p.Ser499Cys), and a nonsense variant c.1015C > T (p.Arg339*) in two consanguineous families. They performed Sanger sequencing of 37 AI patients of different ethnicities and suggested that pathogenic sequence variants in SLC24A4 are a rare cause of AI in general, but might be a frequent cause of AI in Pakistani population [25]. Researches on three consanguineous Turkish families have revealed two homozygous missense pathogenic variants c.437C > T; (p.Ala146Val), c.1317 T > G (p.Leu436Arg) and a 10 kb (10,042 bp) homozygous deletion, comprising of exons 15, 16 and most of the exon-17 (Chr14: 92,957,680-92,967,722del) [36, 39, 40]. During a comparison of AI phenotypes caused by SLC24A4 variants in patients reported so far in the literature, we have concluded that clinical manifestation of AI is moderately to severely variable among the cases (Table 2).

Table 2 Previously reported amelogenesis imperfecta patients carrying pathogenic variants in SLC24A4
Full size table

Conclusion

The present study aimed to perform a clinical and molecular evaluation of an autosomal recessive Pakistani family. We have identified the sixth disease-causing variant in SLC24A4 (Fig. 2B), which extends its mutation spectrum and confirms the role of this gene in the morphogenesis of human tooth enamel.