Genotype-­p henotype correlations

J Med Genet: first published as 10.1136/jmg-2023-109510 on 18 November 2023. Downloaded from http://jmg.bmj.com/ on September 29, 2024 by guest. Protected by copyright.
Original research

Heterozygous COL17A1 variants are a frequent cause

of amelogenesis imperfecta
Ummey Hany ‍ ‍,1 Christopher M Watson ‍ ‍,1,2 Lu Liu,1,3 Claire E L Smith,1
Asmaa Harfoush,3 James A Poulter ‍ ‍,1 Georgios Nikolopoulos,4 Richard Balmer,3
Catriona J Brown,5 Anesha Patel,6 Jenny Simmonds,2 Ruth Charlton,2
María Gabriela Acosta de Camargo,7 Helen D Rodd,8 Hussain Jafri,9
Agne Antanaviciute,10 Michelle Moffat,11 Maisoon Al-­Jawad,3 Chris F Inglehearn,1
Alan J Mighell3

► Additional supplemental ABSTRACT

material is published online Background Collagen XVII is most typically associated WHAT IS ALREADY KNOWN ON THIS TOPIC
only. To view, please visit ⇒ There is an established understanding that
the journal online (http://​dx.​ with human disease when biallelic COL17A1 variants
doi.​org/​10.​1136/​jmg-​2023-​ (>230) cause junctional epidermolysis bullosa (JEB), biallelic COL17A1 variants are a cause of
109510). a rare, genetically heterogeneous, mucocutaneous junctional epidermolysis bullosa (JEB).
blistering disease with amelogenesis imperfecta (AI), WHAT THIS STUDY ADDS
For numbered affiliations see
a developmental enamel defect. Despite recognition
end of article. ⇒ Heterozygous COL17A1 variants are a much
that heterozygous carriers in JEB families can have AI,
and that heterozygous COL17A1 variants also cause more common cause than previously recognised
Correspondence to of isolated autosomal dominant amelogenesis
Alan J Mighell, School of dominant corneal epithelial recurrent erosion dystrophy
Dentistry, Clarendon Way, (ERED), the importance of heterozygous COL17A1 imperfecta (AI) (developmental enamel defects)
University of Leeds, Leeds LS2 variants causing dominant non-­syndromic AI is not in the absence of mucocutaneous disease.
9NL, UK;
widely recognised. HOW THIS STUDY MIGHT AFFECT RESEARCH,
a​ .​j.m
​ ighell@​leeds.​ac.​uk
Methods Probands from an AI cohort were screened by PRACTICE OR POLICY
CFI and AJM are joint senior single molecule molecular inversion probes or targeted
⇒ As genetic testing availability increases,
authors. hybridisation capture (both a custom panel and whole
including as part of AI and corneal dystrophy
exome sequencing) for COL17A1 variants. Patient
Received 13 July 2023 care, more individuals will receive reports of
Accepted 17 October 2023 phenotypes were assessed by clinical examination and
heterozygous COL17A1 variants.
analyses of affected teeth.
⇒ This study provides a reference point to inform
Results Nineteen unrelated probands with isolated
how genetic counselling and clinical care are
AI (no co-­segregating features) had 17 heterozygous,
advanced.
potentially pathogenic COL17A1 variants, including
⇒ Furthermore, there is a need to consider
missense, premature termination codons, frameshift
specialist dental and ophthalmic evaluation of
and splice site variants in both the endo-­domains and
carriers in JEB families to ensure that their care
the ecto-­domains of the protein. The AI phenotype was
needs are also being met.
consistent with enamel of near normal thickness and
variable focal hypoplasia with surface irregularities
including pitting.
Conclusion These results indicate that COL17A1 in humans. It has diverse biological functions in
variants are a frequent cause of dominantly inherited cell adhesion, morphogenesis, neuromuscular
non-­syndromic AI. Comparison of variants implicated in signalling and host defence.1 As a hemidesmosome
AI and JEB identifies similarities in type and distribution, component, it is present in the cutaneous basement
with five identified in both conditions, one of which membrane zone, which connects the skin epidermis
may also cause ERED. Increased availability of genetic and dermis.2 3 It is also expressed during amelogen-
testing means that more individuals will receive reports esis, the process by which dental enamel is formed,
of heterozygous COL17A1 variants. We propose that and contributes to the differentiation of amelo-
patients with isolated AI or ERED, due to COL17A1 blasts.4 5 Col17 knockout (Col17−/−) mice exhibit
variants, should be considered as potential carriers for distorted Tomes’ processes, a reduced volume
© Author(s) (or their JEB and counselled accordingly, reflecting the importance of enamel matrix during the secretory stage and
employer(s)) 2023. Re-­use of multidisciplinary care.
permitted under CC BY. prolonged calcification in the maturation stage of
Published by BMJ. amelogenesis.4
The 56 exons of the COL17A1 gene encode a
To cite: Hany U, Watson CM,
1497-­ amino acid protein which acts as a homo-
Liu L, et al. J Med Genet Epub
ahead of print: [please INTRODUCTION trimer composed of three alpha (α1) chains,
include Day Month Year]. Collagen type XVII alpha 1 chain (COL17A1), each with a molecular mass of 180 kDa. These
doi:10.1136/jmg-2023- hereafter referred to as collagen XVII, is a hemides- consist of a globular amino-­terminal intracellular
109510 mosomal transmembrane protein widely expressed endodomain, a short transmembrane domain and
Hany U, et al. J Med Genet 2023;0:1–9. doi:10.1136/jmg-2023-109510 1
 Genotype-­p henotype correlations

J Med Genet: first published as 10.1136/jmg-2023-109510 on 18 November 2023. Downloaded from http://jmg.bmj.com/ on September 29, 2024 by guest. Protected by copyright.
Ectodomain (exon 18-56, 490–1497 aa)

Non-collagenous domains

Collagenous domains

Endodomain (exon 2-17, 1-466 aa)

Transmembrane domain (exon 17, 466-489 aa) c.2947+2T>C c.3277+1G>A

p.(Pro971Glnfs*95) p.(Tyr1099*)
c.2435-1G>A p.(Arg1133Cys)
Basal keratinocyte p.(Gly677Asp) p.(Pro1154Leufs*97)
plasma membrane
p.(Gly1155Leufs*7)
p.(Gly671Ser)
p.(Glu1199Gln)
p.(Asn181Profs*13) p.(Gly621Ser) p.(Ser1202Leu)
p.(Arg154*) p.(Ser1383Hisfs*71)
p.(Ser114Valfs*60) NC16A
528–547 aa

NH2 NC16C COL15 COL1 COOH

p.(Trp549*) p.(Arg625*) p.(Pro1110Argfs*21)

c.1141+1G>A p.(1745-2A>C) p.(Gly803*)

Figure 1 Schematic representation of the domain organisation of the collagen XVII protein. The extracellular domain or ectodomain is comprised of 15
collagenous (COL1–COL15, yellow vertical boxes) flanked by stretches of non-­collagen sequence (NC1–NC16a, green horizontal lines). The non-­collagen
domain, NC16, spans from the extracellular matrix to cytoplasm and comprises a transmembrane domain NC16b adjoined by NC16a and NC16c to the
C-­terminal and N-­terminal ends respectively. COL17A1 variants identified in this study are denoted in blue text above the protein domains and variants
published by others as causes of amelogenesis imperfecta are displayed in orange text below the protein domains.25 50 The circled variants have been
previously published in association with junctional epidermolysis bullosa.

a flexible rod-­like carboxy-­terminal extracellular ectodomain.6 defects, typically using the non-­specific descriptive term enamel
The ectodomain consists of 15 collagenous (COL1–COL15) hypoplasia.20 21
sequences containing repeating Gly-­ X-­
Y tripeptides which, in Monoallelic COL17A1 variants can also cause dominantly
the homotrimer, form the characteristic collagen triple helices. inherited epithelial recurrent erosion dystrophy (ERED, OMIM
These are flanked by 16 non-­ collagenous sequences (NC1– 122400), a corneal disease with the potential for lifelong progres-
NC16) (figure 1).7 A notable characteristic of collagen XVII is sion and vision loss (three variants reported).22 23 Furthermore,
the shedding of the ectodomain after cleavage at the cell surface there are documented cases of monoallelic COL17A1 variants
by the sheddases ADAM 9, 10 and 17, to yield its soluble intra- causing dominantly inherited amelogenesis imperfecta (AI) in
cellular form; the biological significance of this remains to be the absence of other co-­segregating features or any family history
determined.8 of JEB. AI is a developmental failure of normal dental enamel
Biallelic variants in COL17A1 (OMIM 113811) are a well-­ formation affecting all teeth, which can be inherited as a domi-
documented cause of the recessively inherited, genetically nant, recessive or X-­linked trait, either in isolation or as a compo-
heterogeneous mucocutaneous blistering condition junctional nent of syndromic conditions.24 Dominant isolated AI caused by
epidermolysis bullosa (JEB).9 JEB is genetically heterogeneous heterozygous COL17A1 variants has only been reported in two
and characterised by erosions and blistering of the skin and cohort studies, where COL17A1 was only one of several genes
mucous membranes, with cleavage at the basement membrane implicated, and in one report of a genetically complex AI family
zone. There are a range of clinical presentations, but it is gener- (total six variants),25–27 and COL17A1 variants were not listed
ally classified into one of two major subtypes: intermediate or as a cause of AI in OMIM at the time of submission (13 July
severe.10 JEB prevalence is estimated to be approximately 2 per 2023). The distinction between AI and descriptive terms such as
million live births in the USA and 1 per million in England and enamel hypoplasia is important. The latter term does not link to
Wales.11–13 aetiology or inheritance, unlike AI.
Corneal epithelial erosions and enamel hypoplasia are distinct Here, we describe 19 unrelated families with isolated AI in
features in JEB families.13 Corneal erosions are found in only a which probands are heterozygous for 17 different monoallelic
proportion of JEB cases,14 while it has been reported that enamel COL17A1 variants, consistent with this being a frequent cause
hypoplasia is always associated with JEB.15 Hintner and Wolff16 of autosomal dominant AI presenting in the absence of other
first reported defective enamel in their patients with JEB, and clinical features.
since then, enamel hypoplasia in association with JEB has been
further corroborated.17 18 In comparison to healthy enamel,
the enamel of patients with JEB has increased tissue porosity, MATERIALS AND METHODS
reduced mineral content and contains serum albumin, with Patient recruitment
enamel hypoplasia.19 These studies identified JEB on clinical Patients were recruited though UK dental clinics, with informed
features, without knowing which of the genes known to cause written consent and local ethical approval (REC 13/YH/0028),
JEB was responsible. In papers primarily about JEB due to bial- in accordance with the principles of the Declaration of Helsinki.
lelic COL17A1 changes, carrier parents or siblings of patients Genomic DNA was obtained from venous blood using conven-
with JEB have been described as having developmental enamel tional extraction techniques, or from saliva using Oragene DNA
2 Hany U, et al. J Med Genet 2023;0:1–9. doi:10.1136/jmg-2023-109510
 Genotype-­p henotype correlations

J Med Genet: first published as 10.1136/jmg-2023-109510 on 18 November 2023. Downloaded from http://jmg.bmj.com/ on September 29, 2024 by guest. Protected by copyright.
Sample Collection kits (DNA Genotek). Screening of families using standard tools BWA (V.0.7.12), Picard (V.1.102.0) and the
F2-­F19 was carried out by the University of Leeds Amelogen- GATK HaplotypeCaller (V.3.2–2).
esis Research Group, while family F1 was screened via the NHS
testing service for AI (https://www.england.nhs.uk/publication/​
national-genomic-test-directories/).
Variant classification
The pathogenicity of the variants was assessed according to the
American College of Medical Genetics and Genomics (ACMG)
Reference genome and transcript criteria using Franklin by Genoox (https://franklin.genoox.com/​
The human reference genome used for this study was clinical-db/home).33 Allele frequencies were obtained from the
GRCh37/hg19, the transcript sequence of COL17A1 used was Genome Aggregation Database V.2.1.1 (https://gnomad.broa-
NM_000494.4 and the collagen XVII protein sequence used was dinstitute.org/).34 Splicing predictions were generated using
NP_000485. SpliceAI (https://spliceailookup.broadinstitute.org).35

Whole exome sequencing (WES) Sanger sequencing verification

Two different hybridisation capture reagents were used for WES Primers were designed using AutoPrimer3 (https://github.com/​
library preparation: the SureSelect Human All Exon V6 kit david-a-parry/autoprimer3) and synthesised by IDT. About
(Agilent Technologies) and the Human Comprehensive Exome 25 ng genomic DNA was amplified using Q5 High-­Fidelity 2X
kit (10–50 Mb) (Twist Bioscience). For SureSelect, 3 µg genomic Master Mix (NEB) according to manufacturer’s instructions.
DNA was used to make libraries, following the manufacturer’s PCR products were purified using ExoSAP-­IT (Applied Biosys-
protocol. These were sequenced on a HiSeq 3000 which gener- tems) then sequenced using BigDye Terminator V.3.1 chemistry
ated paired-­end 150 bp reads (Illumina). For the Twist kit, 50 ng on an ABI3130xl Genetic Analyser (Applied Biosystems). Elec-
genomic DNA was used to make libraries following the manufac- tropherograms were analysed using SeqScape software V.2.5
turer’s protocol. Sequencing was carried out on a NextSeq 2000 (Applied Biosystems).
using a P3 kit to generate paired-­end 150 bp reads (Illumina).
The quality of the raw sequence reads was reviewed using
FastQC (V.0.11.3) (https://www.bioinformatics.babraham.ac.uk/​ Micro-computed tomography (µCT)
projects/fastqc/). These were then aligned to an indexed human Intact teeth were analysed using a high resolution µ-CT SkyScan
reference genome using BWA (V.0.7.12) (https://bio-bwa.source- 1172 (Bruker, Belgium) scanner to quantify mineral density.
forge.net/).28 PCR duplicates were removed using Picard (V.2.5.0) Mineral density values were calculated relative to three hydroxy-
(https://broadinstitute.github.io/picard/). Non-­reference bases apatite standards of 0.25 and 0.75 g/cm3 (Bruker, Belgium) and
were identified and recorded in variant call format using the 2.9 g/cm3 (Himed, USA). Fiji/ImageJ was used to analyse enamel
Genome Analysis Tool Kit (GATK) HaplotypeCaller (V.3.5) density, with a pixel threshold above 2.0 g/cm3.
(https://gatk.broadinstitute.org).29 Prior to filtering, identified
variants were annotated with functionally relevant biological Scanning electron microscopy (SEM)
information and observed population allele frequencies using Longitudinal mid buccal slices of teeth were obtained using
the Variant Effect Predictor (VEP) (V.83).30 an Accutom 10 cutting machine and diamond cutting wheel
(Struers, Germany). After removing surface debris, slices were
Targeted sequencing of known AI genes (NHS) gold coated (Agar Scientific, Elektron Technology, UK). Imaging
Genes in the R340 (amelogenesis imperfecta) panel of the UK was performed by S-­3400N (Hitachi, Japan) SEM.
NHS National Genomic Test Directory were subject to hybri-
disation capture using a custom SureSelect reagent. Libraries
RESULTS
were generated from 3 µg genomic DNA according to the
Cohort screening
manufacturer’s instructions (Agilent Technologies). Libraries
Genomic DNA from probands in a large cohort of appar-
were sequenced using a NovaSeq 6000 (Illumina). The resulting
ently unrelated families with non-­ syndromic AI were inves-
FASTQ files were processed and aligned as described above.
tigated either by targeted smMIP screening, NHS diagnostic
AI screening or WES. Variants identified were excluded if the
Single molecule molecular inversion probes (smMIP) CADD score was <15 or minor allele frequency was >0.001.
sequencing The variant list for each case was then filtered further to include
smMIPs targeting the coding sequences of 19 genes (online only candidate pathogenic variants in known or potential
supplemental table S4) implicated in non-­syndromic AI were candidate AI genes. Where possible, additional filtering was
designed using MIPGEN (https://github.com/shendurelab/​ performed based on family history. Probands from 19 families
MIPGEN)31 and synthesised by Integrated DNA Technologies were identified as carrying potentially pathogenic variants in
(IDT, Leuven, Belgium) at 100 nmol scale. Then, 100 ng genomic the COL17A1 gene. In each case, this was the only variant in
DNA was subjected to targeted capture and ligation using the the known non-­syndromic AI genes that met these criteria. Of
smMIPs probe pool diluted to reach a ratio of 800 smMIPs these, inheritance in 12 families appeared dominant, while the
copies for every one DNA molecule in the final capture reac- mode of inheritance could not be determined in the remaining
tion.32 Sequencing was carried out on a NextSeq 500 (Illumina) families due to incomplete clinical information about additional
which generated paired-­end 150 bp reads. family members. Variant pathogenicity was assigned according
Data processing was performed using the MIPVAR pipeline to the ACMG criteria, with the outcome being pathogenic, likely
(https://sourceforge.net/projects/mipvar/) which was modified pathogenic or a variant of unknown significance (VUS). All the
for compatibility with local computing hardware. This enabled suspected variants were re-­sequenced by Sanger sequencing and
sample demultiplexing and the removal of unique molecular their segregation with disease was checked in available family
identifiers prior to ligation-­arm and extension-­arm processing members (figure 2 and online supplemental figure S2).
Hany U, et al. J Med Genet 2023;0:1–9. doi:10.1136/jmg-2023-109510 3
 Genotype-­p henotype correlations

J Med Genet: first published as 10.1136/jmg-2023-109510 on 18 November 2023. Downloaded from http://jmg.bmj.com/ on September 29, 2024 by guest. Protected by copyright.
Figure 2 Pedigrees of 18 of the 19 families recruited for this study. The Sanger sequencing chromatogram from the proband from each family is displayed
beneath each pedigree. Details of the family pedigree and variant identified in family F1 are displayed in online supplemental figure S2. A ‘?’ mark in the
pedigree means ‘individuals with possible AI not clinically assessed’.

COL17A1 variants occurring in only one family. None of the 17 variants described
Probands in the 19 non-­ syndromic AI families displayed in in this study were previously associated with AI. Only seven
figure 2 and online supplemental figure S2 were found to carry are present in the gnomAD database (table 1). Of the 17 vari-
heterozygous, potentially pathogenic variants in COL17A1. Two ants, 6 are missense: c.1861G>A; p.(Gly621Ser), c.2011G>A;
variants were present in two families, with the remainder each p.(Gly671Ser), c.2030G>A; p.(Gly677Asp), c.3397C>T;

Table 1 Details of COL17A1 variants reported in this study

Variants
gnomAD
Family ID ACMG criteria Genomic nomenclature Transcript nomenclature Predicted protein nomenclature CADD score frequency
F1 P (PVS1, PM2, PP5) g.105833981del c.340del p.(Ser114Valfs*60) 32.0 0.00001591
F2 P (PP1, PP4, PVS1, PM2, PP5) g.105831793G>A c.460C>T p.(Arg154*) 36.0 0.000003977
F3 P (PP1, PP4, PVS1, PM2) g.105830245_105830254del c.541_550del p.(Asn181Profs*13) 32.0 Absent
F4 LP (PP1, PP3, PP4, PM2) g.105812867C>T c.1861G>A p.(Gly621Ser) 23.9 Absent
F5 LP (PP1, PP3, PP4, PM2) g.105811266C>T c.2011G>A p.(Gly671Ser) 26.3 0.0001135
F6 LP (PP3, PP4, PM2) g.105811247C>T c.2030G>A p.(Gly677Asp) 26.1 Absent
F7 LP (PP1, PP4, PVS1, PM2) g.105803340C>T c.2435–1G>A p.? 34.0 Absent
F8 P (PP4, PVS1, PM2) g.105798865del c.2912del p.(Pro971Glnfs*95) 33.0 Absent
F9 LP (PP1, PP4, PVS1, PM2) g.105798827A>G c.2947+2T>C p.? 30.0 Absent
F10 P (PP1, PP4, PVS1, PM2, PP5) g.105796802C>T c.3277+1G>A p.? 27.7 0.00006312
F11 P (PP1, PP4, PVS1, PM2) g.105796371G>T c.3297C>A p.(Tyr1099*) 36.0 Absent
F12 VUS (PP4, PM2) g.105796271G>A c.3397C>T p.(Arg1133Cys) 33.0 Absent
F13 P (PP4, PVS1, PM2) g.105795287del c.3456del p.(Pro1154Leufs*97) 20.6 0.000008021
F14 P (PP1, PP4, PVS1, PM2) g.105795287del c.3456del p.(Pro1154Leufs*97) 20.6 0.000008021
F15 P (PP1, PP4, PVS1, PM2) g.105795277_105795278del c.3462_3463del p.(Gly1155Leufs*7) 33.0 Absent
F16 LP (PP1, PP4, PS4, PM2) g.105795045C>G c.3595G>C p.(Glu1199Gln) 25.1 Absent
F17 LP (PP1, PP4, PS4, PM2) g.105795045C>G c.3595G>C p.(Glu1199Gln) 25.1 Absent
F18 VUS (PP4, PM2) g.105795035G>A c.3605C>T p.(Ser1202Leu) 27.0 0.00002800
F19 P (PP1, PP4, PVS1, PM2) g.105793715_105793716del c.4147_4148del p.(Ser1383Hisfs*71) 34.0 Absent
ACMG scoring criteria: PP1: segregation data pathogenic supporting; PP4: phenotype pathogenic supporting; PS4: case control studies pathogenic strong; PVS1: effect on protein
pathogenic very strong; PM2: population data pathogenic moderate; PP5: reputable source data pathogenic supporting.
Nomenclature is reported according to COL17A1 transcript NM_000494.4 and chromosome 10 of human reference genome build hg19.
CADD V.1.3, combined annotation dependent depletion; gnomAD V.2.1.1, genome aggregation database; LP, likely pathogenic; P, pathogenic; VUS, variant of unknown
significance.

4 Hany U, et al. J Med Genet 2023;0:1–9. doi:10.1136/jmg-2023-109510

 Genotype-­p henotype correlations

J Med Genet: first published as 10.1136/jmg-2023-109510 on 18 November 2023. Downloaded from http://jmg.bmj.com/ on September 29, 2024 by guest. Protected by copyright.
quantitative changes (figure 3 and online supplemental figure
S4).
Clinical enamel changes in the primary dentition were minimal
and could be easily overlooked. Hypomaturation changes were
more evident where there had been some post-­eruptive enamel
loss. Focal surface pitting was subtle.
In the secondary dentition, there was generalised, but clin-
ically variable enamel hypomaturation characterised by white
to yellow/brown colouration and greater enamel opacity
than expected. Surface irregularities were also variable, with
distinct, deep pits that in some instances were obvious due to
extrinsic staining or formed linear, vertical defects in the most
pronounced cases. Shallow surface irregularities were also
present in some teeth. Regional enamel hypoplasia involving the
middle third of the labial aspect of anterior teeth was observed in
some cases. Dental radiographs confirmed that enamel thickness
Figure 3 Intraoral images and dental radiographs illustrating the was for the most part within expected normal limits. A clear
variation in enamel phenotypes associated with heterozygous COL17A1 distinction between the radiodensity of enamel compared with
variants in primary and secondary teeth. (i) Primary tooth enamel changes the supporting dentine confirmed that any reduction in enamel
can be minimal and easily missed and are primarily characterised by mineralisation was at the mild end of the spectrum, consistent
hypomaturation changes with subtle surface focal pitting (F10). (ii) with the clinical hypomaturation phenotype. Post-­ eruption
A predominantly hypomaturation AI phenotype with some surface enamel loss was not obviously exaggerated.
irregularities (F9). (iii) Surface pits and other irregularities are the Tooth root morphology including pulp spaces was within
clinically dominant feature, on a background of hypomaturation (F4). (iv) expected normal limits with no taurodontism. No oral mucosal
Hypomaturation enamel is combined with more exaggerated surface pits or other oral cavity changes were evident.
merging into grooves with mid-­third crown regional hypoplasia (arrow)
(F3). (v) Section of an orthopantomogram of a mixed dentition illustrating Laboratory analysis of teeth
near normal enamel thickness with a normal difference in radiodensity Upper primary molar teeth from affected members of families
between the enamel and the supporting dentine (F15). (vi) Intraoral F9 and F14 were analysed by three-­dimensional µCT and SEM
radiograph illustrating near normal enamel thickness, but with enamel and compared with the relevant control teeth (figure 4, online
irregularities and a lesser difference in radiodensity between enamel and supplemental figure S3). µCT revealed near normal enamel
dentine than would be expected (F5). Further clinical images and dental volume in the affected teeth, but they lacked a hard outer enamel
radiographs are included in online supplemental figure S4. layer and mineral density gradation from higher to lower moving
from the outer enamel towards the dental enamel junction
(DEJ), by comparison with the control teeth. µCT also revealed
p.(Arg1133Cys), c.3595G>C; p.(Glu1199Gln), c.3605C>T; a pitted and uneven enamel surface in the probands’ teeth, in
p.(Ser1202Leu); six result in frameshifts: c.340del; p.(Ser- both primary and permanent dentitions, confirmed by SEM
114Valfs*60), c.541_550del; p.(Asn181Profs*13), c.2912del; analysis, which showed the presence of pitting, and disruption of
p.(Pro971Glnfs*95), c.3456del; p.(Pro1154Leufs*97), the enamel layers appearing as a stack of lamellae, with patches
c.3462_3463del; p.(Gly1155Leufs*7) and c.4147_4148del; of fused rod-­interrod regions hard to distinguish between them
p.(Ser1383Hisfs*71); 3 are predicted to affect splice sites: (figure 4 and online supplemental figure S3).
c.2435–1G>A (acceptor loss score 0.99, acceptor gain
score 0.92), c.2947+2T>C (donor loss score 0.98) and Wider clinical phenotype
c.3277+1G>A (donor loss score 0.97); and 2 create prema- None of the affected individuals described here were noted to
ture termination codons (PTC): c.460C>T; p.(Arg154*) and have skin or mucosal abnormalities, corneal problems or any
c.3297C>A; p.(Tyr1099*). other associated conditions. However, all were recruited in
All the stop and frameshift variants identified are classified as dental clinics as cases of non-­syndromic AI and have not been
pathogenic, while the three splice site variants are classified as examined by other clinical specialists for subtle skin or corneal
pathogenic or likely pathogenic. Among the six missense vari- presentations.
ants three are glycine substitutions, and these are classified as
likely pathogenic. Of the three remaining missense variants, one,
DISCUSSION
p.(Glu1199Gln), was initially classed as a VUS, but is absent
Here, we report 15 pathogenic/likely pathogenic heterozygous
from gnomAD and was observed to co-­segregate in two families
COL17A1 variants as the likely cause of non-­syndromic AI in
reported here, leading to reclassification as likely pathogenic.
17 probands, as well as 2 further cases with VUSs that may also
The remaining two non-­glycine missense variants are currently
be causative. This greatly increases the previous tally of six,
classified as variants of unknown significance (VUS). All missense
within an increasingly clear context that COL17A1 variants are
variants identified are in the extracellular domain of the protein
a frequent and under recognised cause of dominantly inherited
(figure 1).
AI. The AI phenotype observed is consistent with the limited
clinical images, radiographs and other data in the peer-­reviewed
Oral clinical phenotype literature from carriers in JEB families who are heterozygous for
All families presented as isolated AI with no history of co-­seg- COL17A1 variants.
regating health issues. Variability in the clinical AI phenotype AI due to heterozygous COL17A1 variants has been linked
was evident, with features that reflected enamel qualitative and to the Witkop classification type 1a pitted hypoplastic AI.26 36
Hany U, et al. J Med Genet 2023;0:1–9. doi:10.1136/jmg-2023-109510 5
 Genotype-­p henotype correlations

J Med Genet: first published as 10.1136/jmg-2023-109510 on 18 November 2023. Downloaded from http://jmg.bmj.com/ on September 29, 2024 by guest. Protected by copyright.
in rows and columns, but with comment that some teeth may
appear normal in both dentitions. The Witkop classification
evolved over time but remained primarily clinically descriptive,
with patterns of inheritance included in some instances. Data
presented here highlight the advantages of switching to classi-
fication where genetic diagnosis has primacy and is correlated
to the clinical enamel phenotype that can vary within certain
parameters. This is also with recognition that enamel does not
have cellular capacity for repair and that the enamel pheno-
type is altered by post-­eruption changes. The enamel present
is generally well mineralised but shows disrupted enamel rod
morphology, which can be expected to adversely impact enamel
functional longevity. Teeth from individuals with JEB due to
biallelic COL17A1 variants were not available for comparative
analysis. In summary, in this series affected enamel has hypomat-
uration characteristics with variable focal hypoplasia (pits and
indentations) and in some instances, partial regional hypoplasia
of the middle third of the tooth enamel.
The profound adverse impact of JEB on the affected indi-
viduals and their families has driven our understanding of
how the condition is caused by pathological biallelic variants
in COL17A1, LAMA3, LAMB3, LAMC2, ITGA6, ITGB4 and
ITGA3. According to the England and Wales EB database,
JEB prevalence is around 1 per million, with most pathogenic
variants detected in LAMB3 (40%–50%), followed by LAMC2
(15%–20%) and LAMA3 (10%–15%), with only a small propor-
tion (5%–10%) in COL17A1 (John McGrath, personal commu-
nication). By contrast, the association of AI with heterozygous
variants in these genes in families with dominant inheritance and
no history of JEB, are less obviously presented in the published
literature, which also fails to clarify whether affected individuals
are carriers for JEB. While all individuals with JEB have AI (or
enamel hypoplasia), there are very few reports of AI in carriers
of JEB due to COL17A1 variants, and it remains unclear what
proportion of carriers will manifest enamel or corneal abnor-
malities.21 37 Assuming 1 in 10 million people have JEB due to
COL17A1 variants, Hardy-­Weinberg equilibrium would predict
a carrier frequency of approximately 1 in 1600, not inconsistent
with published estimates of the frequency of AI,38 39 especially
Figure 4 Laboratory analysis of teeth. (i) Micro-­computed tomography given that many such individuals may have been considered to
(µ-CT) imaging of a permanent upper first premolar tooth from the have enamel hypoplasia or enamel opacities rather than inher-
proband of F14 and (iii) a primary upper molar tooth from the affected ited AI. This highlights two important related points where a
individual from F9. Panels (ii) and (iv) have images from corresponding molecular diagnosis can inform clinical decision-­making. First,
control teeth. No significant differences were observed in average enamel distinguishing between more subtle forms of AI and other
mineral density (EMD) between affected and control samples. F14 and enamel development defects. Second, that dental changes offer
its corresponding control had EMD values of 2.58 and 2.60 g/cm3, an opportunity to identify carrier status for JEB in families with
respectively, while F9 and its corresponding control were 2.40 and 2.52 no history of this condition.
g/cm3, respectively. (v–vi) Line graphs showing the distribution of mineral If it is assumed that all JEB carriers have AI, then one might
density from the enamel surface to the dental enamel junction (DEJ), as expect cases of AI due to variants in LAMB3, LAMC2 and LAMA3
shown by the arrows in (i–iv). Affected samples (red) lack high mineral to be more common than those with COL17A1 variants, given
density at the surface, as opposed to the control teeth (black). Scanning the frequency of the different forms of JEB. Variants in all three
electron microscopy (SEM) images of the enamel in F14 (vii–viii) and F9 genes have been reported in patients with AI in the literature but
show clear pitting extending towards DEJ.51 SEM of enamel from F14 (x– only in a handful of cases for each,40 41 while our findings show
xii) shows generally disrupted and poorly formed prismatic microstructure that variants in COL17A1 are a relatively common cause of AI.
compared with the corresponding control teeth in the images (xiii–xv), It is therefore evident that further research is needed into the
respectively. SEM images of the enamel prism in F9 (xvi–xviii) appears as link between dominant AI and recessive JEB due to COL17A1
a stack of lamellae, with patches of fused rod-­interrod regions hard to variants.
distinguish between them. However, enamel from corresponding controls We identified missense, PTC, frameshift and splice site variants
show distinguishable rod interrod regions (xix–xxi). in both the endo-­domains and the ecto-­domains of the protein.
Fifteen of the variants described are novel, while two have been
previously reported as pathogenic in JEB but not in isolated AI.
Witkop described hypoplastic, pitted autosomal dominant type Patients in this study were not reported to have any associated
enamel with pits from pinpoint to pinhead size primarily on skin or corneal problems but have not been examined by derma-
labial or buccal surfaces of permanent teeth, often arranged tologists or cornea specialists, meaning that subtle versions of
6 Hany U, et al. J Med Genet 2023;0:1–9. doi:10.1136/jmg-2023-109510
 Genotype-­p henotype correlations

J Med Genet: first published as 10.1136/jmg-2023-109510 on 18 November 2023. Downloaded from http://jmg.bmj.com/ on September 29, 2024 by guest. Protected by copyright.
either condition could potentially have been overlooked. We dentition, with complete loss of all teeth by age 14. The proband’s
wanted to understand whether the COL17A1 variants associ- daughter, who is a heterozygous carrier of the p.Gly627Val
ated with AI differ from those that cause JEB. By combining a variant, showed no skin abnormalities but had extensive enamel
literature search on the NCBI database (https://pubmed.ncbi.​ hypoplasia and pitting. The proband’s granddaughter, who was
nlm.nih.gov/) with data from the HGMD professional database also a carrier of the p.Gly627Val variant, manifested dental
(accessed 30 March 2023),42 we identified 232 COL17A1 vari- abnormalities and trauma-­ induced skin blistering, especially
ants reported to cause JEB (online supplemental table S1). The around the knees. The authors concluded that p.Gly627Val has
distribution of mutations and mutation types in JEB and AI are a dominant negative effect on the collagen XVII protein, causing
similar (online supplemental figure S1). Variants c.460C>T; autosomal dominant JEB in the granddaughter.20 48 As well as
p.(Arg154*) and c.1861G>A; p.(Gly621Ser), reported here as providing further evidence of a dominant negative disease mech-
causing AI, and variant c.1745–2A>C, c.2407G>T; p.(Gly803*) anism and of overlap between the COL17A1 variants causing
and c.3327del; p.(Pro1110Argfs*21) reported to cause AI by JEB and AI, this case illustrates the importance of a multidisci-
Prasad and colleagues,25 have also been identified as pathogenic plinary approach to the clinical care of such patients.
in JEB,5 43 44 showing there is overlap in the underlying genetic These findings have significant implications for future care of
basis of these conditions. individuals and their families with diagnoses of JEB, AI or ERED
Only three COL17A1 variants have been reported in the liter- due to pathogenic variants in COL17A1. Further studies are
ature as causing the corneal disease ERED (online supplemental needed to better understand links between these conditions, but
table S3). None of these have been implicated in AI or JEB. The it seems likely that there is overlap between carrier status for JEB
nonsense variant, p.(Arg154*), detected in an AI proband in this and a diagnosis of AI or ERED when they result from hetero-
study and in JEB, was reported in ClinVar (VCV000931124.3) zygous COL17A1 pathogenic variants. This may not have been
as causing autosomal dominant ERED. However, no further fully appreciated by disparate groups of clinicians treating each
evidence was provided, and this result remains unpublished at condition in isolation. The mucocutaneous lesions of JEB are
the time of writing, meaning this should be considered uncon- generally so severe that corneal or dental problems may not have
firmed at this stage. It is unknown if individuals with ERED due been prioritised in patients and could have been overlooked in
to dominant COL17A1 variants have AI, but the expectation their carrier parents or siblings. ERED manifests at around 5
until demonstrated otherwise is that they will, although many years, but may resolve by the early 20s, meaning many adults
of these patients will not have been thoroughly examined by with the condition are without symptoms. AI may be dismissed
dentists. by non-­experts as resulting from poor dental hygiene, especially
Most variants associated with JEB, AI and ERED are frame- in children with EB, who have considerable difficulty in main-
shift, splice or PTC. The consequences of these variants have taining oral hygiene for multiple reasons.49
not been determined experimentally, but it seems likely that To summarise, we identified 17 families with AI due to patho-
they will be subject to nonsense mediated decay,45 meaning no genic/likely pathogenic heterozygous variants in the COL17A1
functional protein is produced from those alleles. Many of the gene, and a further 2 families with variants of unknown signifi-
individuals with JEB due to COL17A1 variants are homozygous cance in COL17A1 that may also be pathogenic. These findings
for such variants, meaning their phenotype is in effect the result suggest that the significance of COL17A1 variants as a cause of
of complete knockout of COL17A1. It therefore seems likely AI has not been fully appreciated and this may in fact be a rela-
that many with JEB suffer from near-­complete loss of Collagen tively common form of dominantly inherited AI. We detail the
XVII protein. Since two of the PTCs implicated in JEB were also spectrum of the enamel phenotype observed and review all the
found in AI, and given the unconfirmed report of one of the pathogenic COL17A1 variants known to cause AI. A compar-
same variants in an ERED case, it therefore seems likely that AI ison with those causing the recessive skin disorder JEB suggests
and ERED due to heterozygous COL17A1 pathogenic variants they are similar in mutation type and distribution, and there
is caused, at least in some cases, by haploinsufficiency. Further is also direct overlap between the variants implicated in both
evidence for this disease mechanism comes from the work of conditions, and possibly in a third, the dominantly inherited
Yuen and colleagues,37 who used immunofluorescence staining corneal disorder ERED. People with AI or ERED due to hetero-
with antibodies targeted to mouse Col17 to show reduced base- zygous COL17A1 variants should be considered carriers for JEB.
ment membrane zone and apical–lateral staining in skin from Furthermore, these results highlight the need for a multidisci-
both JEB patients and carriers compared with control skin. plinary approach to the care of families and individuals with
A proportion of the COL17A1 variants observed in individ- JEB, including carriers, and those with dominant AI or ERED
uals with JEB, AI and ERED are amino acid substitutions. These due to COL17A1 variants.
may also be functional knockouts, but an alternative disease
mechanism has been proposed in some of these cases. Missense Author affiliations
1
variants, and most commonly glycine substitutions, have been Leeds Institute of Medical Research, University of Leeds, St. James’s University
Hospital, Leeds, UK
reported to be associated with milder JEB phenotypes.17 Substi- 2
North East and Yorkshire Genomic Laboratory Hub, Central Lab, St. James’s
tution of glycine residues within the ectodomain, and partic- University Hospital, Leeds, UK
ularly within the COL15 collagenous sequence (figure 1), is 3
School of Dentistry, Clarendon Way, University of Leeds, Leeds, UK
4
thought to destabilise the collagenous triple helix, making the Institute for Fundamental Biomedical Research, B.S.R.C. ’Alexander Fleming’, Vari,
protein unstable, with the mutated protein predicted to exert a Attica, Greece
5
Birmingham Dental Hospital, Mill Pool Way, Edgbaston, Birmingham, UK
dominant negative effect on the wild-­type protein.46 47 Interest- 6
LCRN West Midlands Core Team, NIHR Clinical Research Network (CRN),
ingly, of the six missense variants reported here in patients with Birmingham Research Park (West Wing), Vincent Drive, Edgbaston, Birmingham, UK
7
AI, three were glycine substitutions in the COL15 region. Department of Dentistry of the Child and Adolescent, Universidad de Carabobo,
An interesting case describes a patient with JEB who is a Carabodo, Venezuela
8
Academic Unit of Oral Health Dentistry and Society, School of Clinical Dentistry,
compound heterozygote for glycine substitution p.Gly627Val University of Sheffield, Sheffield, UK
within the COL15 domain and frameshift insertion c.3514ins25 9
Fatima Jinnah Medical University, Punjab Thalassaemia and Other Genetic Disorders
within the COL6 domain.20 The proband had an abnormal Prevention and Research Institute, Lahore, Pakistan

Hany U, et al. J Med Genet 2023;0:1–9. doi:10.1136/jmg-2023-109510 7

 Genotype-­p henotype correlations

J Med Genet: first published as 10.1136/jmg-2023-109510 on 18 November 2023. Downloaded from http://jmg.bmj.com/ on September 29, 2024 by guest. Protected by copyright.
10
MRC Human Immunology Unit, University of Oxford, Oxford, UK 12 Petrof G, Papanikolaou M, Martinez AE, et al. The epidemiology of epidermolysis
11
Paediatric Dentistry, The Newcastle Upon Tyne Hospitals NHS Foundation Trust, bullosa in England and Wales: data from the National Epidermolysis Bullosa database.
Newcastle upon Tyne, UK Br J Dermatol 2022;186:843–8.
13 Nishie W. Collagen XVII processing and blistering skin diseases. Acta Derm Venereol
2020;100:5662.
Twitter Christopher M Watson @ChrisM_Watson, James A Poulter @jamesapoulter
14 Mellado F, Fuentes I, Palisson F, et al. Ophthalmologic approach in epidermolysis
and María Gabriela Acosta de Camargo @gaviota113
bullosa: a cross-­sectional study with phenotype-­genotype correlations. Cornea
Acknowledgements The authors thank the families involved for their support for 2018;37:442–7.
this study. 15 Wright JT, Johnson LB, Fine JD. Development defects of enamel in humans with
Contributors UH, CMW, CELS, JAP, CI and AM contributed to conception, design, hereditary epidermolysis bullosa. Arch Oral Biol 1993;38:945–55.
data acquisition and interpretation, drafting and critical review of the manuscript. LL, 16 Hintner H, Wolff K. Generalized atrophic benign epidermolysis bullosa. Arch Dermatol
AH, GN, RB, CJB, AP, JS, RC, MGAdC, HDR, HJ, AA, MM and MA-­J contributed to data 1982;118:375–84.
acquisition, interpretation and critical review of the manuscript. All authors approved 17 Pasmooij AMG, Pas HH, Jansen GHL, et al. Localized and generalized forms of
the final version of the manuscript. Guarantor: CFI. blistering in junctional epidermolysis bullosa due to COL17A1 mutations in the
Netherlands. Br J Dermatol 2007;156:861–70.
Funding This work was supported by Rosetrees Trust Grant PGS19-­2/10111, 18 Gedde-­Dahl T. Phenotype-­genotype correlations in epidermolysis bullosa. Birth Defects
Wellcome Trust grant number WT093113MA and a Leeds Doctoral Scholarship Orig Artic Ser 1971;7:107–17.
awarded to UH. 19 Kirkham J, Robinson C, Strafford SM, et al. The chemical composition of tooth enamel
Competing interests None declared. in junctional epidermolysis bullosa. Arch Oral Biol 2000;45:377–86.
20 McGrath JA, Gatalica B, Li K, et al. Compound heterozygosity for a dominant glycine
Patient consent for publication Not applicable.
substitution and a recessive internal duplication mutation in the type XVII collagen
Ethics approval This study involves human participants and was approved by REC gene results in junctional epidermolysis bullosa and abnormal dentition. Am J Pathol
13/YH/0028. Participants gave informed consent to participate in the study before 1996;148:1787–96.
taking part. 21 Floeth M, Fiedorowicz J, Schäcke H, et al. Novel homozygous and compound
Provenance and peer review Not commissioned; externally peer reviewed. heterozygous COL17A1 mutations associated with junctional epidermolysis bullosa. J
Invest Dermatol 1998;111:528–33.
Data availability statement All data relevant to the study are included in the 22 Jonsson F, Byström B, Davidson AE, et al. Mutations in collagen, type XVII, alpha
article or uploaded as supplementary information. 1 (COL17A1) cause epithelial recurrent erosion dystrophy (ERED). Hum Mutat
Supplemental material This content has been supplied by the author(s). It 2015;36:463–73.
has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have 23 Oliver VF, van Bysterveldt KA, Cadzow M, et al. A COL17A1 splice-­altering mutation is
been peer-­reviewed. Any opinions or recommendations discussed are solely those prevalent in inherited recurrent corneal erosions. Ophthalmology 2016;123:709–22.
of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and 24 Smith CEL, Poulter JA, Antanaviciute A, et al. Amelogenesis imperfecta; genes,
responsibility arising from any reliance placed on the content. Where the content proteins, and pathways. Front Physiol 2017;8:435.
includes any translated material, BMJ does not warrant the accuracy and reliability 25 Prasad MK, Geoffroy V, Vicaire S, et al. A targeted next-­generation sequencing assay
of the translations (including but not limited to local regulations, clinical guidelines, for the molecular diagnosis of genetic disorders with Orodental involvement. J Med
terminology, drug names and drug dosages), and is not responsible for any error Genet 2016;53:98–110.
and/or omissions arising from translation and adaptation or otherwise. 26 Bloch-­Zupan A, Rey T, Jimenez-­Armijo A, et al. Amelogenesis imperfecta: next-­
generation sequencing sheds light on witkop’s classification. Front Physiol
Open access This is an open access article distributed in accordance with the
2023;14:1130175.
Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits
27 Prasad MK, Laouina S, El Alloussi M, et al. Amelogenesis imperfecta:1 family, 2
others to copy, redistribute, remix, transform and build upon this work for any
phenotypes, and 2 mutated genes. J Dent Res 2016;95:1457–63.
purpose, provided the original work is properly cited, a link to the licence is given,
28 Li H, Durbin R. Fast and accurate short read alignment with burrows-­wheeler
and indication of whether changes were made. See: https://creativecommons.org/​
transform. Bioinformatics 2009;25:1754–60.
licenses/by/4.0/.
29 DePristo MA, Banks E, Poplin R, et al. A framework for variation discovery and
ORCID iDs genotyping using next-­generation DNA sequencing data. Nat Genet 2011;43:491–8.
Ummey Hany http://orcid.org/0000-0002-4486-1625 30 McLaren W, Gil L, Hunt SE, et al. The Ensembl variant effect Predictor. Genome Biol
Christopher M Watson http://orcid.org/0000-0003-2371-1844 2016;17:122.
James A Poulter http://orcid.org/0000-0003-2048-5693 31 Boyle EA, O’Roak BJ, Martin BK, et al. Mipgen: optimized modeling and
design of molecular inversion probes for targeted resequencing. Bioinformatics
2014;30:2670–2.
32 Hiatt JB, Pritchard CC, Salipante SJ, et al. Single molecule molecular inversion probes
REFERENCES
1 Franzke CW, Bruckner P, Bruckner-­Tuderman L. Collagenous transmembrane proteins: for targeted, high-­accuracy detection of low-­frequency variation. Genome Res
recent insights into biology and pathology. J Biol Chem 2005;280:4005–8. 2013;23:843–54.
2 Eady RAJ, McGrath JA, McMillan JR. Ultrastructural clues to genetic disorders of skin: 33 Richards S, Aziz N, Bale S, et al. Standards and guidelines for the interpretation of
the dermal-­epidermal junction. J Invest Dermatol 1994;103:S13–8. sequence variants: a joint consensus recommendation of the American College of
3 Pulkkinen L, Uitto J. Mutation analysis and molecular genetics of epidermolysis Medical Genetics and Genomics and the Association for Molecular Pathology. Genet
bullosa. Matrix Biol 1999;18:29–42. Med 2015;17:405–24.
4 Asaka T, Akiyama M, Domon T, et al. Type XVII collagen is a key player in tooth enamel 34 Karczewski KJ, Francioli LC, Tiao G, et al. The mutational constraint spectrum
formation. Am J Pathol 2009;174:91–100. quantified from variation in 141,456 humans. Nature 2020;581:434–43.
5 Kiritsi D, Kern JS, Schumann H, et al. Molecular mechanisms of phenotypic variability 35 Jaganathan K, Kyriazopoulou Panagiotopoulou S, McRae JF, et al. Predicting splicing
in junctional epidermolysis Bullosa. J Med Genet 2011;48:450–7. from primary sequence with deep learning. Cell 2019;176:535–48.
6 Diaz LA, Ratrie H 3rd, Saunders WS, et al. Isolation of a human epidermal cDNA 36 Witkop CJ. Amelogenesis imperfecta, dentinogenesis imperfecta and dentin dysplasia
corresponding to the 180-­kD autoantigen recognized by Bullous pemphigoid and Revisited: problems in classification. J Oral Pathol 1988;17:547–53.
herpes gestationis sera. Immunolocalization of this protein to the hemidesmosome. J 37 Yuen WY, Di Zenzo G, Jonkman MF, et al. New versatile monoclonal antibodies
Clin Invest 1990;86:1088–94. against type XVII collagen endodomain for diagnosis and Subtyping COL17A1-­
7 Giudice GJ, Emery DJ, Diaz LA. Cloning and primary structural analysis of the bullous associated junctional epidermolysis bullosa. J Eur Acad Dermatol Venereol
pemphigoid autoantigen Bp180. J Invest Dermatol 1992;99:243–50. 2016;30:1426–7.
8 Franzke C-­W, Bruckner-­Tuderman L, Blobel CP. Shedding of collagen XVII/Bp180 in 38 Altug-­Atac AT, Erdem D. Prevalence and distribution of dental anomalies in
skin depends on both ADAM10 and ADAM9. J Biol Chem 2009;284:23386–96. orthodontic patients. Am J Orthod Dentofacial Orthop 2007;131:510–4.
9 Has C, Bauer JW, Bodemer C, et al. Consensus reclassification of inherited 39 Bäckman B, Holm AK. Amelogenesis imperfecta: prevalence and incidence in a
epidermolysis Bullosa and other disorders with skin fragility. Br J Dermatol northern Swedish County. Community Dent Oral Epidemiol 1986;14:43–7.
2020;183:614–27. 40 Poulter JA, El-­Sayed W, Shore RC, et al. Whole-­exome sequencing, without prior
10 Pfendner E, Uitto J, Fine JD. Epidermolysis bullosa carrier frequencies in the US linkage, identifies a mutation in LAMB3 as a cause of dominant hypoplastic
population. J Invest Dermatol 2001;116:483–4. amelogenesis imperfecta. Eur J Hum Genet 2014;22:132–5.
11 Fine J-­D. Epidemiology of inherited epidermolysis Bullosa based on incidence and 41 Gostyńska KB, Yan Yuen W, Pasmooij AMG, et al. Carriers with functional null
prevalence estimates from the National Epidermolysis Bullosa Registry. JAMA mutations in LAMA3 have localized enamel abnormalities due to haploinsufficiency.
Dermatol 2016;152:1231–8. Eur J Hum Genet 2016;25:94–9.

8 Hany U, et al. J Med Genet 2023;0:1–9. doi:10.1136/jmg-2023-109510

 Genotype-­p henotype correlations

J Med Genet: first published as 10.1136/jmg-2023-109510 on 18 November 2023. Downloaded from http://jmg.bmj.com/ on September 29, 2024 by guest. Protected by copyright.
42 Stenson PD, Mort M, Ball EV, et al. The human gene mutation database: towards a 47 Tasanen K, Floeth M, Schumann H, et al. Hemizygosity for a glycine substitution
comprehensive repository of inherited mutation data for medical research, genetic in collagen XVII: unfolding and degradation of the ectodomain. J Invest Dermatol
diagnosis and next-­generation sequencing studies. Hum Genet 2017;136:665–77. 2000;115:207–12.
43 Condrat I, He Y, Cosgarea R, et al. Junctional epidermolysis bullosa: allelic 48 Almaani N, Liu L, Dopping-­Hepenstal PJC, et al. Autosomal dominant junctional
heterogeneity and mutation stratification for precision medicine. Front Med epidermolysis bullosa. Br J Dermatol 2009;160:1094–7.
(Lausanne) 2018;5:363. 49 Marty M, Chiaverini C, Milon C, et al. Perception of oral health-­related quality of life
44 Pasmooij AMG, van Zalen S, Nijenhuis AM, et al. A very mild form of non-­Herlitz in children with epidermolysis bullosa: a quantitative and qualitative study. JDR Clin
junctional epidermolysis bullosa: BP180 rescue by outsplicing of mutated exon 30 Trans Res 2023;8:349–55.
coding for the COL15 domain. Exp Dermatol 2004;13:125–8. 50 Bloch-­Zupan A, Rey T, Jimenez-­Armijo A, et al. Amelogenesis imperfecta: next-­
45 Chang YF, Imam JS, Wilkinson MF. The nonsense-­mediated decay RNA surveillance generation sequencing sheds light on Witkop’s classification. Front Physiol
pathway. Annu Rev Biochem 2007;76:51–74. 2023;14:1130175.
46 Väisänen L, Has C, Franzke C, et al. Molecular mechanisms of junctional epidermolysis 51 Brookes SJ, Barron MJ, Smith CEL, et al. Amelogenesis imperfecta caused by
bullosa: col 15 domain mutations decrease the thermal stability of collagen XVII. J N-­terminal enamelin point mutations in mice and men is driven by endoplasmic
Invest Dermatol 2005;125:1112–8. reticulum stress. Hum Mol Genet 2017;26:1863–76.

Hany U, et al. J Med Genet 2023;0:1–9. doi:10.1136/jmg-2023-109510 9