Amelogenesis imperfecta: Next-generation sequencing sheds light on Witkop's classification

Abstract

Amelogenesis imperfecta (AI) is a heterogeneous group of genetic rare diseases disrupting enamel development (Smith et al., Front Physiol, 2017a, 8, 333). The clinical enamel phenotypes can be described as hypoplastic, hypomineralized or hypomature and serve as a basis, together with the mode of inheritance, to Witkop's classification (Witkop, J Oral Pathol, 1988, 17, 547-553). AI can be described in isolation or associated with others symptoms in syndromes. Its occurrence was estimated to range from 1/700 to 1/14,000. More than 70 genes have currently been identified as causative. Objectives: We analyzed using next-generation sequencing (NGS) a heterogeneous cohort of AI patients in order to determine the molecular etiology of AI and to improve diagnosis and disease management. Methods: Individuals presenting with so called "isolated" or syndromic AI were enrolled and examined at the Reference Centre for Rare Oral and Dental Diseases (O-Rares) using D4/phenodent protocol (www.phenodent.org). Families gave written informed consents for both phenotyping and molecular analysis and diagnosis using a dedicated NGS panel named GenoDENT. This panel explores currently simultaneously 567 genes. The study is registered under NCT01746121 and NCT02397824 (https://clinicaltrials.gov/). Results: GenoDENT obtained a 60% diagnostic rate. We reported genetics results for 221 persons divided between 115 AI index cases and their 106 associated relatives from a total of 111 families. From this index cohort, 73% were diagnosed with non-syndromic amelogenesis imperfecta and 27% with syndromic amelogenesis imperfecta. Each individual was classified according to the AI phenotype. Type I hypoplastic AI represented 61 individuals (53%), Type II hypomature AI affected 31 individuals (27%), Type III hypomineralized AI was diagnosed in 18 individuals (16%) and Type IV hypoplastic-hypomature AI with taurodontism concerned 5 individuals (4%). We validated the genetic diagnosis, with class 4 (likely pathogenic) or class 5 (pathogenic) variants, for 81% of the cohort, and identified candidate variants (variant of uncertain significance or VUS) for 19% of index cases. Among the 151 sequenced variants, 47 are newly reported and classified as class 4 or 5. The most frequently discovered genotypes were associated with MMP20 and FAM83H for isolated AI. FAM20A and LTBP3 genes were the most frequent genes identified for syndromic AI. Patients negative to the panel were resolved with exome sequencing elucidating for example the gene involved ie ACP4 or digenic inheritance. Conclusion: NGS GenoDENT panel is a validated and cost-efficient technique offering new perspectives to understand underlying molecular mechanisms of AI. Discovering variants in genes involved in syndromic AI (CNNM4, WDR72, FAM20A … ) transformed patient overall care. Unravelling the genetic basis of AI sheds light on Witkop's AI classification.

Keywords: NGS; amelogenesis imperfecta; enamel; genetics; next-generation sequencing; rare diseases.

FIGURE 1

Identified variants within the cohort grouped by pathogenic class and novelty. Variant were classified following ACMG recommendations. (A) Variant’s repartition by pathogenic class and novelty for individual affected by isolated AI. Forty variants (37%) are newly identified variants (orange gradient), 69 (63%) variants are already reported in literature (blue gradient). Among variants reported, 50 (46%) are pathogenic (class 5), 38 (35%) are likely pathogenic (class 4) and 21 (19%) are of uncertain significance. (B). Variant’s repartition by pathogenic class and novelty for individual affected by syndromic AI. Seven variants (24%) are newly identified variants (orange gradient), 32 variants (76%) are already reported in literature (blue gradient). Among variants reported, 5 (12%) are pathogenic (class 5), 32 (77%) are likely pathogenic (class 4), 4 (9%) are of uncertain significance (class 3) and 1 (2%) is probably begin (class 2).

FIGURE 2

Phenotypic and genetic diagnosis in 86 individuals with isolated AI. Number of patients per main type of AI and per gene. Type I hypoplastic AI represents 39 individuals (45.3%) in blue associated with 9 different genes (AMELX, ENAM, COL17A1, LAMB3, ACP4, AMBN, COL7A1, LAMA3, LAMC2). Type II hypomature AI represents 28 individuals (32.6%) in orange associated with 6 different genes (MMP20, WDR72, AMELX, SLC24A4, KLK4). Type III hypomineralized AI represents 16 individuals (18.6%) in grey associated with 2 different genes (FAM83H, WDR72). Type IV hypoplastic-hypomature with taurodontism AI represents 3 individuals (3.5%) in yellow in 1 gene (DLX3).

FIGURE 3

Phenotypic and genetic diagnosis in 31 individuals with syndromic AI. Number of patients per main type of AI and per gene. Type I hypoplastic AI represents 22 individuals (71%) in blue associated with 7 different genes (FAM20A, LTBP3, GALNS, ARHGAP6, RAI1, SLC13A5, TGFBR2). Type II hypomature AI represents 3 individuals (9.7%) in orange associated with 1 gene (ROGDI). Type III hypomineralized AI represents 4 individuals (12.9%) in grey associated with 2 different genes (CNNM4, SLC10A7). Type IV hypoplastic-hypomature with taurodontism AI represents 2 individuals (6.5%) in yellow in 1 gene (DLX3).

FIGURE 4

Phenotype/Genotype correlation for isolated AI. Typical phenotype/genotype correlation observed in patients presenting with isolated Amelogenesis imperfecta (intraoral pictures and radiographs). Patients who are carrier of AMELX mutations can present with different types of AI. Indeed, the phenotype can be either hypoplastic (severe with almost no enamel in male and with a lyonization banding pattern in female) or hypomature depending on the mutation’s localisation. When the mutations occur in a MMP20 cleavage site the AMELX related AI observed is a X-linked hypomature AI.

FIGURE 5

Phenotype/Genotype correlation for syndromic AI.