Genetic-environmental Risk Factors for Tooth Agenesis: A Familial Case-control Study in China

Abstract

Objective

Tooth agenesis (TA) is a common dental developmental anomaly influenced by genetic and environmental factors, though the specific contributions of these factors in large populations are not well understood. This study aimed to determine the genetic and environmental impact on TA and tooth development in a Chinese population.

Methods

We conducted a family-based case-control study involving 106 TA families (385 participants) and 197 controls (618 participants). Clinical examinations, evaluations of 25 potential risk factors, heritability and genetic analyses, including whole-exome sequencing and Sanger sequencing were performed.

Results

106 TA probands exhibited 60.0% hypodontia and 34.0% oligodontia/anodontia, with 17% presenting systemic changes. TA heritability was estimated at 82.1%, higher in females than in males. Genetic contributions(heritability) decreased with increased environmental exposure (from 95.0% to 86.5%, P < .05). Identified genetic patterns in TA families included monogenic (54.05%), oligogenic (27.03%), and other patterns (18.92%). Key TA genes in TA families were WNT10A (63.3%), EDA (23.3%) and PAX9 (20.0%). WNT10A appeared in both monogenic and oligogenic patterns, and was also found in some unaffected familial controls. A family history increased offspring TA risk by 20-fold. Environmental contribution of TA was 17.9%, with unprotected electromagnetic radiation exposure (≥4 hours/d) during pregnancy increasing 4-fold risk.

Conclusions

Genetic contribution is more statistically significant in TA patients with more severe phenotypes, and in females. Heritability of TA declined with increased environmental exposure. Multiple genetic modes exist in TA families. Additionally, environmental factors, like unprotected electromagnetic radiation during pregnancy, are likley to play a role in TA development.

Graphical Abstract

Introduction

Tooth agenesis (TA) is one of the most common developmental anomalies in children, affecting the number of teeth. The TA prevalence ranged from approximately 0.03% to 10.1% (excluded third molars).¹ Based on the number of missing teeth, it can be classified into 3 types: hypodontia (<6 missing teeth), oligodontia (≥6 missing teeth), and anodontia (all teeth are missing). TA can occur in isolation (non-syndromic type) or associated with other dental anomalies and/or the abnormalities in other organs or tissues (syndromic type). To date, over 90 different syndromes have been identified that are associated with TA.^2,3

The affected children may face significant functional, aesthetic, psychological challenges. The treatment of TA is also a complex and comprehensive process, entailing substantial costs.^4,5 Considering the detrimental impact and complexity of TA treatment, etiological prevention measures is paramount.

The aetiology of TA has been linked to genetic, epigenetic, environmental, evolutional or anatomical factors.6, 7, 8, 9, 10 Among these, genetic factors are widely recognised as primary contributors, with specific genes such as WNT10A, PAX9, MSX1, and EDA consistently identified in the affected individuals, some of them follow Mendelian inheritance.11, 12, 13 However, approximately half of cases with isolated TA lack detectable pathogenic variants in these well-characterised genes, indicating unresolved mechanisms in disease pathogenesis.¹⁴ Recent studies have further suggested oligogenic inheritance in some patients,¹⁵ highlighting that the genetic architecture of this condition is more complex than previously thought.

Additionally, variations in the severity and location of TA among monozygotic twins or triplets suggest the influence of environmental factors.¹⁶ Environmental factors such as maternal smoking, have also been linked to offspring TA.^17,18

Despite these insights, the genetic and environmental risk factors and their specific impacts on the occurrence of TA remain unclear. This study aimed to determine the genetic and environmental influences on the development of TA and its genetic architecture in the Chinese population.

Materials and methods

Study design and setting

The family-based study was conducted in affiliated hospital and local communities between May 2022 and June 2023. We analysed clinical profiles, heritability, risk factors and genotype-phenotype relationships in collected families (Figure 1).

Fig. 1.

Flow chart of this study design.

Study populations

To minimise investigation bias, our cases and controls were selected from both affiliated hospital and communities which largely ensured representativeness and comparability between 2 group. We recruited participants (the first participant in each family) with a 1:2 case-to-control ratio, matching for age and sex. Inclusion criteria included age ≥2 years, availability of a panoramic radiograph, and informed consent. TA cases were defined as individuals with missing permanent teeth, while controls had no missing teeth (excluding third molars). Individuals with missing teeth due to disease, trauma, or extractions were excluded from both groups. All control subjects were clinically and radiographically confirmed to be free of any form of TA. For eligible participants and all family members, this study covered the expenses associated with hospital visits, further diagnostic examinations, and genetic counselling.

Clinical analyses

Participants and their families underwent comprehensive medical assessments (Appendix 2), including, detailed dental examinations, standardised interviews and multidisciplinary evaluation. Oral and systemic characteristics were recorded, and X-ray imaging (panoramic radiographs and CBCT scans) were used to assess TA and other dental abnormalities. Two doctors confirmed the final diagnosis, which was then audited by a third doctor.

Hereditary mode and heritability

In family-based case-control study, clinical information from first-degree relatives in both groups were summarised, and prevalence rates were calculated. The Penrose method was used to estimate the hereditary mode, where different ratios of sibling prevalence (s) to population prevalence (q) suggest different inheritance patterns: s/q approaching 1/2q and 1/4q indicates monogenic dominant and recessive heredity respectively, and 1/√q suggests polygenic heredity.¹⁹ Heritability (h²) values in relatives were calculated using Falconer’s method: h² = 1/r *P_c (X_c-X_a) / a_c.²⁰

Risk factors analyses

Assessment protocol

Participants in the case and control groups, along with their parents, were invited to complete the Risk Factor Assessment Form (Appendix 3). The analysis of TA risk factors followed STROBE guidelines. We selected 25 potential risk factors based on previous literature and clinical expertise (detailed information see Appendix 3),^17,21 including socioeconomic, familial, and parental exposures. During the examination, well-trained doctors asked participants these questions respectfully, allowing them to skip uncomfortable ones.

Risk factor definitions

• 1.Socioeconomic status: the evaluation of household income, education and occupation.
• 2.Family history: the presence of TA in immediate or extended family members.
• 3.
  Maternal factors:

  • a.
    Abnormal fertility history: ie, spontaneous abortion, embryo demise or stillbirth.
  • b.
    Conceiving after the age of 35.
  • c.
    Presence of systemic diseases or medical conditions such as diabetes, hypertension, anaemia, or seizures during pregnancy.
  • d.
    Health status or lifestyle during pregnancy: anxiety history, cold, medication use, multivitamin intake. Anxiety was defined as a combination of anxious mood and a checklist of other related symptoms, persisting for a minimum of 4 weeks.
  • e.
    Smoking or exposure to second-hand smoking during pregnancy: Smoking status was defined as consuming at least one cigarette per day, while ‘second-hand smoking’ referred to involuntary inhalation of tobacco smoke for at least 15 minutes per day, at least once per week.
  • f.
    Exposure to chemicals: any type of agricultural or industrial chemical compound for any length of time during pregnancy.
• 4.Paternal factors: health conditions, medication use, smoking habit, alcohol consumption, lifestyle, exposure to chemicals and irradiation.

Family-based genetic analyses

To further evaluate genetic modes of TA, such as monogenic and oligogenic inheritance, we performed whole-exome sequencing (WES) in 65 individuals from 20 families, comprising 37 patients with TA and 28 unaffected individuals. Karyotype analysis and multiplex ligation-dependent probe amplification (MLPA) were complementarily conducted in 6 individuals with complex phenotypes from 5 families. Genomic DNA was extracted from peripheral blood samples of the patients using blood genomic DNA Mini Kit (CWBIO). WES involves exome capture, high-through put sequencing, and common filtering. Alignment of the sequence reads, indexing of the reference genome, variant calling, and annotation were carried out using the Agilent SureSelect Human All Exon V6 system (Agilent Technologies, Santa Clara). Sequence reads were aligned to the human reference genome (hg38) using the Burrows-Wheeler Aligner (BWA). 1000 Genomes Project was used to find rare variants. And 6 pathogenicity prediction software (SIFT, PolyPhen-2 HVAR, PolyPhen-2 HDIV, Mutation Assesor, Mutation Taster and CADD) were used for variant identification as likely damaging, deleterious, and disease causing. We prioritised the variants within genes associated with TA that scored as more deleterious. Variant classification followed ACMG 2015 criteria, including pathogenic/likely pathogenic, variants of uncertain significance, and benign/likely benign.²² We primarily focus on variants with minor allele frequencies (MAF) of less than 1% (rare variants). However, we also consider some variants that have been reported to exhibit high prevalence of incomplete penetrance or are classified as low-frequency (1%<MAF<5%).^23,24 These approaches were applied to the evaluation of monogenic inheritance patterns. For oligogenic inheritance patterns, we further referenced the OLIDA criteria.²⁵

Statistical analyses

In clinical analysis, Kruskal–Wallis test, chi square test, and modification were performed to compare the data. P values were calculated (*P < .05; ** P < .01; *** P < .001).

A sample size of 100 cases and 200 controls was deemed appropriate for risk factor evaluation, aiming to detect an odds ratio (OR) of at least 2.0. This sample size estimation considered the low prevalence of TA and challenges in recruiting family units, and was based on MAF of candidate genes like MSX1, PAX9, and EDA.¹⁷

Univariate and multivariate logistic regression procedures were conducted to estimate ORs and 95% CIs for the association between potential risk factors and TA. Socioeconomic status,²⁶ smoking habits (including second-hand smoking frequency and smoking severity), and alcohol consumption was treated as ordinal variables with 3 levels and entered into the model using 2 dummy variables. Other variables and the outcome variable, TA, were coded as binary variables. Student’s t-test was employed to compare age between groups.

Known risk factors for TA, such as smoking habits, were forced into the multivariable model and other possible predictor variables were investigated via Enter methods. Subgroup analyses were conducted to explore differences in risk factors between genders and different genetic backgrounds. Missing values were treated as being missing completely at random. Sensitivity analysis using different regression models was performed to test the robustness of our primary analysis. All data were analysed with SPSS software (IBM SPSS Statistics 26.0).

Results

Clinical features of study families

Family-based study included 106 case families and 197 control families. The general prevalence of TA among first-degree relatives was 21.40% in case families and 4.27% in control families. The sociodemographic characteristics of 2 study group are summarised in Table 1. Most of probands in the case families and the first participants in control families were female. The mean age of probands in case group was 16.7 years, compared to 17.6 years for the first participants in control group. A statistically significant difference in family socioeconomic status was observed between the groups.

Table 1.

General information in family-based study.

	Group		P value
	Case families (n = 106)	Control families (n = 197)
Total family members, n	385	618	-
Probands/First participants	106	197	-
First-degree relative	267	421	-
Extended family members	12	-	-
General TA prevalence of first-degree relative, n1/n2 (%)	57/267 (21.35%)	18/421(4.27%)	<.001
Father	26/106 (24.53%)	9/197(4.57%)	<.001
Mother	23/106 (21.70%)	9/197(4.57%)	<.001
Sibling	8/55 (14.54%)	*	-
Average reproductive age (SD)
Mother	27.65 (4.3)	27.13 (4.6)	.202
Father	28.26 (4.6)	27.40 (5.2)	.994
	Proband	First participants
Gender
Male	43 (40.6%)	81 (41.1%)	.926
Female	63 (59.4%)	116 (58.9%)
Age
Mean (SD)	16.7 (8.8)	17.6 (6.5)	.283
Median	18.0	19.0	-
Family socioeconomic status
High (≥11)	5 (5.4%)	24 (13.9%)	<.05
Medium (8-10)	45 (48.4%)	93 (53.8%)
Low (≤7)	43 (46.2%)	56 (32.3%)
Missing value	13	24

n_1, number of tooth agenesis patients; n₂, number of participants.

^⁎We do not find TA patients among 27 examined siblings in control group (we do not have information on the TA prevalence among the probands' children, as most of the probands from whom we collected data are under 18 years of age. This age selection was made to exclude the influence of tooth loss from periodontitis).

Of the 106 probands in case group, 83.0% (n = 88) had non-syndromic TA, while 17.0% (n = 18) exhibited syndromic TA. Syndromic TA (STA) patients sought medical advice at a younger age (10.8 years) compared to their non-syndromic TA (NSTA) counterparts (17.8 years), and they also had a higher average number of missing teeth (6.5 vs 4.9) (Table S1). Both NSTA and STA probands exhibited a higher prevalence of missing lateral incisors and second premolars, while the prevalence of the missing first molars was lower. Notably, STA patients had higher incidence of missing mandibular canines (Figure 2A).

Fig. 2.

Clinical analysis of probands with tooth agenesis and heritability evaluation. (A) Comparison of missing teeth in the maxillary and mandibular arches, combining the right and left sides, of patients with non-syndromic tooth agenesis (NSTA) and syndromic tooth agenesis (STA). Ca, canine; CI, central incisor; LI, lateral incisor; M, molar; PM, premolar. *P < .05. **P < .01. ***P < .001. (B) Comparison of tooth agenesis prevalence between male and female patients. (C) Comparison of tooth agenesis of patients in patients with and without a family history of tooth agenesis. (D) Proportion of tooth agenesis patients with oral malformations and the prevalence of these malformations across different type of tooth agenesis. (E) Proportion of tooth agenesis patients with systemic malformation. (F) TA prevalence rate among first-degree relative in case and control families. Differences in TA prevalence between male and female first-degree relatives were assessed across groups. Prevalence of NSTA, STA, hypodontia, and oligodontia/anodontia were also compared between groups. (G) Heritability analysis of tooth agenesis. h2, heritability; SE, standard error. #The specific heritability cannot be estimated for low incidence of affected individuals in control group.

Figure 2B and C illustrate the correlation between gender, family history, and TA. Males were found to have a higher number of missing teeth compared to females, particularly in the second premolars, maxillary premolars, lateral incisors, and mandibular incisors. A higher incidence of missing mandibular first molars was observed in affected patients without family history, compared to those with a family history.

Other oral malformations were observed in 32.0% of probands, with microdontia accounting for the highest proportion at 17.9% (Figure 2D, Table S2). Furthermore, the prevalence of oral malformations was statistically significantly higher in STA cases compared to NSTA, and was more frequently observed in individuals with oligodontia than in those with hypodontia.

For systemic changes, 6.60% (7/106) of probands were diagnosed with abnormal ectodermal diseases (eg, hypohidrotic ectodermal dysplasia and incontinentia pigmenti). Chromosomal disorders (eg, Down syndrome and Williams–Beuren syndrome) were present in 3.8% (4/106), and oral-facial-clefts and related syndromes in a further 3.8% (4/106). The aetiology remained unknown in 2.8% (3/106) (Table S2). Patients with syndromic TA also exhibited abnormities in more than 2 organs or tissues, central nervous system abnormalities were particularly prevalent (50%, 9/18), often manifesting as intellectual disability (Figure 2E, Tables S1 and S2).

A positive family history of TA was reported in 53.8% of probands (57/106). A further comparison of 40 probands and their affected parents revealed that 67.50% of offspring exhibited similar dental phenotypes, while 32.5% displayed more severe TA compared to their parents (Table S3). This suggests a strong familial link, with some offspring potentially experiencing more severe manifestations.

Familial aggregation

Maternal and paternal exposures to risk factors during pregnancy and preconception, respectively, are summarised in Table 2. Our analysis revealed that a positive family history of TA was associated with 20-fold higher risk of offspring TA (Table 3).

Table 2.

Univariate analysis of tooth agenesis risk factors.

	Case	Control	OR (95% CI)	P value
Family history of tooth agenesis			13.43 (7.09-25.46)	<.001
Yes	57/105 (54.3%)	16/197 (8.1%)	-	-
No	48/105 (45.7%)	181/197 (91.9%)	-	-
Missing value	1	-
Premature or post-term infant	5/106 (4.7%)	10/197 (5.1%)	0.93 (0.31-2.78)	.891
Maternal risk factors during pregnancy
Reproductive age ≥35	7/101 (6.9%)	13/190 (6.8%)	1.06 (0.57-21.97)	.860
Missing value	5	7	-	-
Parity ≥2	29/106 (27.4%)	32/175 (18.3%)	1.68 (0.95-2.99)	.075
Missing value	-	22	-	-
Abnormal fertility history	14/91 (15.4%)	10/148 (6.8%)	2.69 (1.14-6.34)	.024
Missing value	15	49	-	-
Systemic diseases	25/96 (26.0%)	23/182 (12.6%)	2.51 (1.33-4.72)	.004
Missing value	10	15	-	-
Catch a cold or fever	28/89 (31.5%)	31/163 (19.0%)	1.89 (1.05-3.42)	.035
Missing value	17	34	-	-
Anxiety or depression	24/91 (26.4%)	27/165 (16.4%)	1.83 (0.98-3.41)	.057
Missing value	15	32	-	-
Heavy pregnancy reaction	24/89 (27.0%)	44/143 (30.8%)	0.83 (0.46-1.50)	.536
Missing value	17	54	-	-
Medication use	21/92 (22.8%)	24/175 (13.7%)	1.86 (0.97-3.57)	.061
Missing value	14	22	-	-
Low frequency of protein intake	24/92 (26.1%)	32/180 (17.8%)	1.63 (0.89-2.98)	.111
Missing value	14	17	-	-
Supplement use	54/88 (61.4%)	93/174 (53.4%)	1.38 (0.82-2.33)	.223
Missing value	18	23	-	-
Smoking habits	1/97 (1.0%)	0/182	1.89 (0.04-96.05)	.751
Missing value	9	15	-	-
Second-hand smoking
<1 d/wk	55/94 (58.5%)	127/179 (70.9%)	1	-
1-2 d/wk	10/94 (10.6%)	17/179 (9.5%)	1.36 (0.59-3.16)	.476
≥3 d/wk	29/94 (30.9%)	35/179 (19.6%)	1.91 (1.07-3.44)	.030
Missing value	12	18	-	-
Alcohol consumption (>1 glass/wk)	3/97 (3.1%)	2/182 (1.1%)	2.87 (0.47-17.49)	.252
Missing value	9	15
Toxic chemicals or radiation exposure	10/95 (10.5%)	12/183 (7.3%)	1.68 (0.70-4.04)	.249
Missing value	11	14	-	-
Electromagnetic radiation exposure
<4 h/d	61/91 (67.0%)	145/175 (82.9%)	1
≥4 h/d+ protective clothing use	10/91 (11.0%)	16/175 (9.1%)	1.49 (0.64-3.46)	.358
≥4 h/d	19/91 (20.9%)	14/175 (8.0%)	3.17 (1.50-6.73)	.003
Missing value	15	22	-	-
Paternal exposure during preconception
Reproductive age ≥35	7/102 (6.7%)	13/173 (7.5%)	0.91 (0.35-2.35)	.841
Missing value	4	24	-	-
Systemic diseases	5/91 (5.5%)	3/183 (1.6%)	3.45 (0.82-14.93)	.092
Missing value	15	14	-	-
Smoking habits
Yes	42/83 (50.6%)	62/155 (40.0%)	1.54 (0.90-2.62)	.117
No	41/83 (49.4%)	93/155(60.0%)	-	-
Missing value	23	43	-	-
Number of cigarettes/d
0	35/83 (42.2%)	83/155 (53.5%)	1
1 to 9	17/83 (20.5%)	30/155 (19.4%)	1.34 (0.66-2.75)	.417
≥10	31/83 (37.3%)	42/155 (27.1%)	1.75 (0.95-3.22)	.072
Missing value	17	42	-	-
Alcohol consumption (>1 glass/wk)	21/90 (23.3%)	35/176 (19.9%)	1.23 (0.66-2.26)	.514
Missing value	16	21	-	-
Alcohol consumption
Former or never	59/80 (73.8%)	141/176 (80.1%)	1
Moderate (<4 d/wk)	19/80 (23.8%)	30/176 (17.0%)	1.51 (0.79-2.90)	.211
Heavy (≥4 d/wk)	2/80 (2.5%)	5/176 (2.8%)	0.96 (0.18-5.07)	.958
Missing value	26	21	-	-
Toxic chemicals or radiation exposure	13/98 (13.3%)	16/180 (8.9%)	1.57 (0.72-3.41)	.257
Missing value	8	17	-	-

Table 3.

Multivariable logistic regression models of tooth agenesis risk factors.

		OR (95% CI)	P value
Family-based study
	Socioeconomic status
	High (≥11)	1.00	-
	Medium (8-10)	6.96 (1.57-30.82)	.011
	Low (≤7)	6.50 (1.46-29.02)	.014
	Family history of tooth agenesis	20.48 (9.32-44.97)	<.001
Mother	Electromagnetic radiation exposure*
	<4 h/d	1.00	-
	≥4 h/d+ protective clothing use	1.10 (0.36-3.35)	.864
	≥4 h/d	4.00 (1.32-12.1)	.014
	Systemic diseases	2.88 (1.18-7.00)	.016
	Second-hand smoking
	<1 d/wk	1	-
	1-2 d/wk	0.45 (0.11-1.87)	.270
	≥3 d/wk	2.48 (1.07-5.75)	.034
	Abnormal fertility history	1.36 (0.44-4.21)	.598
	Catch a cold or fever	2.15 (0.85-5.44)	.103
	Anxiety	2.24 (0.96-5.22)	.061

^⁎No statistical significance was found between <4 h and <4 h + protective clothing use (P = .946).

Environmental risk factors

Electromagnetic radiation

Heavy maternal exposure to electromagnetic radiation (≥4 hours/d) during pregnancy was associated with a 4-fold higher risk of TA. However, this risk was significantly mitigated when protective clothing was used, as detailed in Table 3.

Maternal factors

Maternal systemic diseases during pregnancy were associated with an increased risk of TA in offspring. The adjusted odds of maternal systemic diseases were nearly 3 times higher in case group compared to controls.

Second-hand smoking

Maternal frequent passive smoking (≥3 days/wk) during pregnancy was associated with a 2.5-fold higher risk of TA in offspring compared to controls, as shown in Table 3.

Subgroup analysis

Environmental risks varied by sex and TA family history (Figure S1). Female patients were associated with frequent exposure to electronic radiation and passive smoking, while male patients exhibited links to maternal systemic diseases in addition to passive smoking. Notably, individuals without TA family history were associated with all 3 risk factors, whereas those with a family history were linked to only one.

Genetic factors

Hereditary mode and heritability

In family-based research, sibling prevalence (14.54%) and population prevalence (4.27%) supported a multifactorial inheritance, with the ratio of s/q approaching 1/√q (Table 1). The prevalence of TA among female first-degree relatives in the case group was 6 times that of the control group, while for males, it was 4 times higher. A significant difference in the prevalence of different types of TA was also observed between the 2 groups (Figure 2F).

The heritability of TA in the first-degree relative was 82.08%. The heritability of female individuals (87.15%) was higher than that of males (81.59%) (P < .05). Syndromic cases exhibited different heritability compared to non-syndromic cases, and oligodontia or anodontia also demonstrated distinct heritability patterns when compared to hypodontia, although the significance cannot be well estimated due to the low incidence of affected individuals in control group (Figure 2G).

Further analysis of heritability variation in TA patients revealed that heritability decreased from 95.0% in individuals without exposure to the 3 environmental risk factors to 86.5% in those with exposure (P < .05), suggesting that environmental factors may modulate the genetic contribution to TA.

Genetic inheritance of tooth agenesis

We investigated the genetic basis of TA in 65 individuals from 20 pedigrees using WES (Figure 3A). Of these individuals, 37 had TA phenotypes, while the remaining 28 individuals without TA did not and served as familial controls. Of 37 TA patients, 7 (18.92%) lacked the rare or low frequency variants in the reported TA genes, and 30 cases (81.08%) carried 21 different variants in the known TA genes. Among the 30 cases, 15 carry monoallelic variants (40.54%), 5 (13.51%) carry biallelic mutations in the same gene (homozygosity or compound heterozygosity) and 10 (27.03%) carry at least 2 mutant alleles from different genes (oligogenicity). While in the 28 familial controls, 18 (64.29%) did not carry any variants in the reported TA genes, and 10 (35.71%) had the variants in monoallelic (8 of 10) or oligogenic way (2 of 10). We further compared above ratio in the hypodontia group (13 cases) and the oligodontia/anodontia group (24 cases) and found that the latter exhibited an increasing proportion of both biallelic and oligogenic variants, along with a decreasing proportion of monoallelic variants compared to the former (Figure 3B).

Fig. 3.

Genetic analysis and tooth agenesis (TA) genetic architecture in affected families. (A) Pedigrees of 20 TA-affected families. ‘F’ indicates family number. The proband is indicated by an arrow. The available genotypes are indicated below each individual. Squares depict males, and circles depict females. (B) Number of alleles with variants in TA individuals and familial controls. (C) Percentages of affected individuals with monoallelic, biallelic and oligogenic variants in genes known to be mutated in TA. (For instance, among the 30 TA patients with identified variants, 63.3% (19/30) carry WNT10A variants, with different proportions of patients exhibiting monoallelic, biallelic, and oligogenic variants.)

Of the 21 identified variants across 6 genes, 8 were novel and 13 had been previously reported. Twelve of these 21 variants were deleterious, including 5 novel variants (Table 4). WNT10A variants were the most frequently detected, identified in 63.3% of TA patients (19 of 30), followed by EDA variants (23.3%,7 of 30) and PAX9 variants (20.0%,6 of 30) (Figure 3C). Specific WNT10A variants (c.637G>A, c.382C>T, c.511C>T) were observed in TA patients from over 2 affected families. Among the 8 familial controls carrying TA-associated variants, 7 carried WNT10A variants, including c.511C>T, c.637G>A, c.644G>T, c.1070C>T, and c.1248C>A, respectively (Table 4, Figure 3A).

Table 4.

Variants detected in tooth agenesis patients and their family controls.

Genes	Chr position	cDNA	Amino acid change	Type†	Highest MAF	SIFT	PolyPhen-2	VEST3	Mutation taster	ACMG criteria	Cases	Family controls	Inheritance	Known/Novel
WNT10A (NM_025216.3)	chr2: 218889989	c.382C>T	p. Arg128*	Nonsense	-	-	-	-	A	P (PVS1, PS1, PM2, PP3, PP4, PP5)	F59: Ⅰ-1, Ⅱ-1 F135: Ⅰ-1, Ⅱ-1	-	1) Monoallelic (mild cases) 2) Biallelic (severe cases)	K
	chr2: 218890118	c.511C>T	p. Arg171Cys	Missense	0.0186	0.0 (D)	0.93 (D)	0.451	D	VUS (PP3, PP5, BP6)	F262: Ⅰ-2, Ⅱ-1 F280: Ⅱ-1	F59: Ⅰ-1 F280: Ⅰ-2	1) Oligogenic 2) Biallelic	K
	chr2: 218890126	c.519G>T	p. Lys173Asn	Missense	0.0002	0.073 (T)	0.926 (D)	0.282	D	VUS (BP4)	F53: Ⅰ-1, Ⅱ-2	-	Oligogenic	K
	chr2: 218890176	c.569A>G	p. His190Arg	Missense	-	0.609 (T)	0.028 (B)	0.598	D	VUS (PM2)	F249: Ⅱ-2	-	Oligogenic	N
	chr2: 218890244	c.637G>A	p. Gly213Ser	Missense	0.0237	0.0 (D)	0.999 (D)	0.983	D	LP (PS1, PP3, PP4, PP5)	F60: 1-2, Ⅱ-1 F82: Ⅰ-1, Ⅱ-1, Ⅱ-2, Ⅱ-3 F262: Ⅰ-2 F281: Ⅱ-1	F82: Ⅱ-1 F281: 1-2	1) Monoallelic (mild, severe cases) 2) Biallelic (severe cases) 3) Oligogenic	K
	chr2: 218890251	c.644G>T	p. Ser215Ile	Missense	-	0.0 (D)	0.995 (D)	0.865	D	VUS (PM2, PP3, PP4)	F137: Ⅱ-1, Ⅱ-2	F137: Ⅰ-2	Biallelic	N
	chr2: 218893087	c.1070C>T	p. Thr357Ile	Missense	-	0.032 (D)	0.994 (D)	0.994	D	LP (PS1, PM2, PP3, PP4, PP5)	F135: Ⅱ-1	F135:1-2	Biallelic	K
	chr2: 218893265	c.1248C>A	p. Cys416*	Nonsense	-	-	-	-	D	P (PVS1, PM2, PP3, PP4)	F137: Ⅱ-1, Ⅱ-2	F137: Ⅰ-1, Ⅱ-3	Biallelic	N
EDA (NM_001399.5)	chrX:70033469	c.865C>T	p. Arg289Cys	Missense	-	0.0 (D)	0.987 (D)	0.973	D	P (PS1, PM2, PP2, PP3, PP4, PP5)	F276: Ⅱ-1	F276:1-2	Oligogenic (severe cases)	K
	chrX:70035434	c.1001G>A	p. Arg334His	Missense	0.0096	0.025 (D)	0.925 (P)	0.790	D	VUS (PP3)	F280: Ⅱ-1	F280: Ⅰ-2	Oligogenic	K
	chrX:70035446	c.1013C>T	-	Other type†	-	0.001 (D)	1.0 (D)	0.850	A	P (PVS1, PM2, PP3, PP4, PP5)	F247: Ⅰ-1, Ⅱ-2	-	1) Monoallelic 2) Oligogenic	K
	chrX:69255328	c.1045G>A	p. Ala349Thr	Missense	-	0.0 (D)	1.0 (D)	0.963	A	P (PS1, PM1, PM2, PP3, PP4)	F260: Ⅰ-1, Ⅱ-2	-	Monoallelic	K
	chrX:70035552	c.1119G>A	-	Other type†	-	0.0 (D)	0.86 (P)	0.952	D	P (PVS1, PM2, PP3, PP4)	F262: Ⅰ-2, Ⅱ-2	-	Oligogenic (severe cases)	N
PAX9 (NM_001372076.1)	chr14: 36663032	c.140G>C	p. Arg47Pro	Missense	-	0.0 (D)	0.998 (D)	0.988	D	LP (PS1, PM2, PP3, PP4, PP5)	F120: Ⅰ-1, Ⅱ-1	-	Monoallelic	K
	chr14:36663098	c.206G>A	p. Gly69Glu	Missense	-	0.0 (D)	1.0 (D)	0.957	D	LP (PM2, PP1, PP2, PP3, PP4)	F53: Ⅰ-1, Ⅱ-1, Ⅱ-2	-	1) Monoallelic 2) Oligogenic	N
	chr14: 36666470	c.640A>G	p. Ser214Gly	Missense	-	0.512 (T)	0.0 (B)	0.275	N	VUS (PM2, BP4)	F60: Ⅰ-1	-	Monoallelic	K
MSX1 (NM_002448.3)	chr4: 4860319	c.421delG	p. Glu141fs	Frameshift	-	-	-	-	-	P (PVS1, PM2, PM4, PP4)	F273: Ⅱ-1	-	Monoallelic	N
	chr4: 4860360	c.461C>T	p. Pro154Leu	Missense	0.001	0.038 (D)	0.003 (B)	0.324	D	VUS	F276: Ⅱ-1	F276: Ⅰ-2, Ⅱ-2	Oligogenic	K
KDF1 (NM_152365.3)	chr1: 26951536	c.845T>G	p. Ile282Ser	Missense	-	0.001 (D)	0.714 (P)	0.866	D	LP (PS2, PM2, PP3)	F68: Ⅱ-1	-	Monoallelic	N
LAMB3 (NM_000228.3)	chr1: 209617996	c.2962C>T	p. Arg988Trp	Missense	-	0.009 (D)	0.482 (P)	0.594	N	VUS (PM2, PP3)-	F249: Ⅱ-2	-	Oligogenic*	K
	chr1:209807877	c.479T>G	p. Val160Gly	Missense	-	D	D	NA	-	VUS (PM2, BP1)	F82: Ⅰ-1, Ⅱ-3	-	Oligogenic	N

NA, not available; "-", No relevant data available.

Highest MAF (Highest population minor allele frequency observed in any population, as reported in ExAC and 1000 Genomes Project databases).

SIFT (D, damaging; T, tolerated).

PolyPhen-2(D, probably damaging, P, possible damaging).

MutationTaster prediction (A, known to be deleterious; D, probably deleterious, N polymorphism).

VEST3 is a more powerful prediction tool that can make independent judgments when there are inconsistencies among multiple prediction software. When the VEST3 score meets the PP3 supporting evidence threshold (ie, ≥0.644), it may serve as standalone PP3 evidence – even if other computational tools predict benignity or yield uncertain results. For BP4 criteria, a VEST3 score <0.29 is applied (VEST3 exhibits inherent bias in assessing non-missense variants).⁴²

^⁎Require further validation.

^†Structural interaction variant (in this study, genome annotations and mutation types follow the sequence ontology [SO] standard. A ‘Structural interaction variant’ denotes a variant that disrupts internal interactions within the polypeptide structure, thereby affecting protein coding).

WNT10A was identified as a key gene in oligogenic inheritance, present in 8 of 10 individuals with oligogenic variation (Table 4, Figure 3C). Patients with compound heterozygous WNT10A variants had a higher incidence of missing teeth than those with simple heterozygous variants. Thus, WNT10A variants encompass both low-frequency variants (eg, c.511C>T) associated with oligogenic patterns and rare variants (eg, c.382C>T) that follow a monogenic inheritance pattern (Figure 3A and C).

Discussion

To the best of our knowledge, this study is the most extensive investigation to date of familial aggregation, high-risk factors, heritability and genetic architecture of TA. It involves the largest Chinese family-based cohort (n = 1003), with detailed information on TA in participants, their parents and siblings.

For the first time, this study quantified and compared the genetic and environmental contribution to TA across different subtypes and genders. Previous research has often focused on either genetic or environmental factors in isolation.^17,27 We found TA is a complex trait influenced by genetic (82.1%) and environmental (17.9%) factors. The influence of environmental factors is more prominent for hypodontia than oligodontia or anodontia. Moreover, environmental factors seem to have a stronger effect on males than on females, and on NSTA compared to STA. These results highlight the importance of considering these differences when formulating preventive interventions.

In this study, we found the TA heritability varies with familial aggregation and exposure to adverse environmental risks associated with TA. The heritability of TA patients was decreased when exposing to certain environmental factors, which suggest a gene-by-environment interaction for TA. This finding is consistent with the bioecological framework, which suggest that more favourable environments lead to higher heritability.²⁸

Apart from 3.8% of cases associated with chromosomal variations, TA exhibits more complex genetic patterns, involving monogenic, oligogenic inheritance, and other genetic modes. Yu et al¹⁴ reported that the molecular diagnostic rate among 131 unrelated Chinese patients with TA was 55.7%, which is comparable to our diagnostic rate for traditional Mendelian inheritance (54.05%). Notably, our expanded analysis identified an oligogenic inheritance pattern in 27.03% of cases, thereby expanding the potential etiological understanding of previously unresolved cases.

Several known TA variants, including WNT10A c.511C>T, c.637G>A, and c.1248C>A, have also been detected in control members, which highlights the role of incomplete penetrance and additional modifiers – such as epigenetic regulation and genotype-by-environment effects – in shaping phenotypic expression. For example, Song et al²⁹ analysed 451 normal Chinese control participants and observed that 6.2% of healthy individuals carried heterozygous TA genotypes but with lower penetrance, which was also found in individuals of Caucasian ethnicity.^30,31 Thus, these monoallelic variants may act as a susceptibility factor for TA. By contrast, previous research found biallelic genotypes detected in TA patients all had severe phenotypes in complete penetrance. Notably, no such biallelic genotypes were detected in healthy controls in these investigations,²⁹ consistent with our observations in Families 59, 135, 137, and 281. In our study, TA patients without TA family history were associated with 3 environmental risk factors (Figure S1), whereas those with a family history were linked to one, suggesting environmental effects on some potential healthy carriers. Additionally, the proband in family 280 had similar genetic variants but presented with both hypodontia and hearing disorders, unlike her mother, with the father's uranium radiation exposure potentially contributing to the proband's more severe phenotype. According to these findings, it should be cautious to assess the pathogenicity of some TA variations in sporadic TA cases or even in some ‘monogenic’ TA families.

Previous studies have shown that interactions between 2 distinct gene mutations can lead to a single disease. For instance, He et al³² found that digenic mutations in both WNT10A and EDA contribute to isolated and syndromic TA, which represents the most common combination and is consistent with our findings. Other previously reported gene pairs include WNT10A with PAX9, LRP6, EVC, LAMA3, PAG1 and TP63.^15,33, 34, 35 Additionally, our study identified a new combination involving WNT10A and LAMB3 in 2 independent families (F82 and F249). Beyond TA, amelogenesis imperfecta also exhibits oligogenic inheritance.³⁶ Given that over 300 genes have been confirmed to be involved in tooth development,³⁷ the phenomenon of oligogenic inheritance can not only explain certain cases of TA where no known pathogenic genes are detected but may also be extended to account for more other oral genetic diseases in the future.

In the context of non-shared environmental contributions to TA, previous research has identified relevant exposures. Maternal active smoking is a well-established risk factor for offspring dental developmental anomalies,¹⁷ while prolonged use of mobile phones and computers during pregnancy links to birth defects like Down syndrome, though evidence specific to TA remains limited.^9,38 Our findings align with this work: the significantly influence of heavy passive smoking (≥3 days/wk) extends understanding of tobacco’s deleterious effects on dental development. However, the associations of heavy electromagnetic radiation and maternal systemic diseases identified in this study require further validation.

Future preventive strategies may include genetic counselling (particularly for individuals with a family history of TA and healthy individuals carried WNT10A mutations), use of protective clothing, modification of maternal lifestyles, promotion of balanced diets, and enhanced awareness of social and occupational risk factors such as electromagnetic radiation and passive smoking.

Over 90 syndromes are linked to TA, this study identified at least 8 different syndromes in 18 of 106 (17%) TA patients, with ectodermal dysplasia and Down syndrome being prevalent, consistent with previous research.² The high ratio of central nervous system abnormalities in syndromic TA suggests shared developmental ways and molecular mechanisms between these tissues and teeth. One of our recent findings showed that Arghap29 deficiency results in cleft palate and tooth abnormalities in mice.³⁹ On the other hand, environmental factors may target different tissues including tooth at the same early embryonic stage.⁴⁰ Additionally, over half of the patients with non-syndromic cleft lip and palate had tooth-number anomalies, occurring on both the cleft and non-cleft sides, emphasising the impact of genetic and environmental factors on craniofacial development.⁴¹

Limitations: Severe cases of TA were more likely to seek medical advice, leading to selection bias. Additionally, self-reporting of risk factors may be influenced by recall bias. with cases recalling exposures more vividly than controls, and social desirability bias in parents might underreport exposures. Next, TA patients were predominantly from clinical settings, potentially inflating heritability estimates, which limited the study's generalisability. Last, as a first-line test, WES targets coding exons of established TA genes, limiting non-coding region analysis. Future studies could explore unknown pathogenic genes or non-coding variants via technologies like whole-genome sequencing in unresolved cases.

Conclusion

Our findings provide novel insights into the aetiology, clinical features, and molecular basis of TA. A deeper understanding of the impact of genetic and environmental factors on different subtypes of TA will facilitate the development of precision medicine. Future multicentre research is necessary to validate the risk factors and genetic architecture identified in this study. Whole-genome sequencing could be used in subsequent research to detect noncoding regulatory gene mutations. Additionally, methylation microarrays or ATAC-seq analyses of dental pulp stem cells or oral mucosal cells from patients with TA and controls may identify associations between differentially methylated regions (DMRs) or chromatin accessibility and the TA phenotype.

Conflict of interest

None disclosed.