Beyond CHD7 gene: unveiling genetic diversity in clinically suspected CHARGE syndrome

Introduction

CHARGE syndrome (OMIM #214800), formerly known as CHARGE association, represents multiple congenital anomalies such as coloboma, heart anomalies, atresia of choanae, restricted growth and development, genital anomalies, and ear anomalies. CHD7 was discovered as the causative gene in 2004 [1, 2]. The estimated birth incidence of CHARGE syndrome is 1 in 10,000 [3]. CHARGE syndrome is primarily diagnosed clinically according to the diagnostic criteria established by Blake et al. in 1998, Verloes in 2005, and Hale et al. in 2015 [4,5,6].

Approximately 70%–90% of patients clinically diagnosed with CHARGE syndrome harbor heterozygous pathogenic variants in CHD7 [6, 7]. This gene is located on chromosome 8q12.2 and is responsible for encoding the chromodomain-helicase-DNA-binding protein 7 [8]. This protein plays a crucial role in regulating gene expression during embryonic development by participating in chromatin remodeling and transcriptional reprogramming processes associated with neural crest migration [9]. Haploinsufficiency of CHD7 disrupts the normal transcription of the various genes it modulates, resulting in the developmental anomaly characteristic of CHARGE syndrome [10].

A proportion of patients clinically diagnosed with CHARGE syndrome do not harbor pathogenic variants of CHD7. This raises considerations for variants in the intronic regions, 5ʹ untranslated regions, or 3ʹ regions of CHD7, as well as abnormalities in regulatory elements [8]. It also suggests the existence of other candidate genes and genetic heterogeneity that should be explored. SEMA3E has been identified as a candidate gene implicated in CHARGE syndrome [7, 11]. SEMA3E is crucial for regulating angiogenesis and development in the central nervous system, heart, and bone [12].

Various genetic disorders also share overlapping features with CHARGE syndrome, including 22q11.2 deletion syndrome, Kabuki syndrome, PAX2-related disorder, EFTUD2-related mandibulofacial dysostosis, and SOX2-related disorder [13,14,15,16,17,18].

This study was conducted to verify the detection rate of pathogenic variants in CHD7 among patients with clinically suspected CHARGE syndrome and to examine patients who did not harbor pathogenic variants in CHD7 but were found to harbor other genetic mutations. The results could help expand the insights on CHARGE syndrome and conditions that have similar symptoms, thus improving our understanding of the genetic diversity associated with these disorders.

Subjects and methods

Subjects and clinical evaluations

This study included subjects who were clinically suspected with CHARGE syndrome and referred for CHD7 sequencing analysis from November 2004 to December 2023 at Asan Medical Center Children’s Hospital, Seoul, Korea. Subjects who did not fully meet the clinical diagnostic criteria for CHAGE syndrome but had some features were also included.

Medical records were retrospectively reviewed, which included age at the time of referral, sex, family history of CHARGE syndrome, and related phenotypes according to previous known features [1, 2]. Phenotypes were evaluated based on comprehensive physical examination, ophthalmological examination, audiometry, echocardiogram, magnetic resonance imaging or computerized tomography scans encompassing the semicircular canals and olfactory bulbs, and renal ultrasound.

Molecular analysis

For single-gene sequencing, genomic DNA was extracted from peripheral blood leukocytes using the Puregene DNA isolation kit (Qiagen). Polymerase chain reaction (PCR) was used to amplify all coding exons and exon–intron boundaries of CHD7 using specific oligonucleotide primers. The PCR products were then sequenced using an ABI3130xl Genetic Analyzer (Applied Biosystems, Foster City, CA).

Multiplex ligation-dependent probe amplification (MLPA) analysis was conducted using the SALSA MLPA kit P201 CHARGE probemix (MRC Holland, Amsterdam, Netherlands) to detect deletion or duplication in CHD7. Genomic DNA was subjected to denaturation, hybridization, ligation reaction, and PCR amplification in order. Then, the PCR amplicons were run on an ABI3130xl Genetic Analyzer (Applied Biosystems, Foster City, CA), and the gene dosage was determined using the GeneMarker software version1.7 (SoftGenetics, State College, PA).

For exome sequencing (ES), genomic DNA was extracted from either peripheral blood or buccal swab samples. Exons of all human genes were captured using one of the following three kits: Agilent Sure Select kit (version C2, December 2019), Twist capture kit (Twist Bioscience HQ, San Francisco, CA, USA), or IDTxGen Exome Research Panel v2 (Integrated DNA Technologies, Coralville, Iowa, USA). Genomic DNA was extracted from peripheral blood using the Allprep DNA/RNA kit (Qiagen, Venlo, Netherlands) for genome sequencing (GS). Sequencing was performed on the NovaSeq platform (Illumina, San Diego, CA, USA), resulting in a mean coverage depth of 100× for ES and 30× for GS. The sequencing reads were aligned to the human reference genome using the BWA-MEM algorithm, with GRCh37 used for ES and GRCh38 for GS. To call single nucleotide variant and small insertions/deletions (INDEL), HaplotypeCaller2 was used in ES and HaplotypeCaller2 and Strelka2 were used in GS. Furthermore, copy number variants were called from ES based on the depth-of-coverage information using CoNIFER [19] and 3bCNV, an internally developed tool. Structural variant from GS was done using Delly. Variants were filtered and categorized according to the guidelines of both the American College of Medical Genetics and Genomics (ACMG) [20] and ClinGen (https://clinicalgenome.org/working-groups/sequence-variant-interpretation/), using the RareVision system (Inocras, San Diego, CA, USA). Medical geneticists manually reviewed variant pathogenicity evaluations, considering the patient’s clinical features and family history. Candidate variant sequences were validated by Sanger sequencing of the patients’ and/or their parents’ DNA.

Ethical statement

This study was approved by the Institutional Review Board (IRB) of Asan Medical Center (IRB No. 2023-1298) and was conducted in accordance with the principles of the Declaration of Helsinki. Blood samples for molecular analyses were collected after obtaining written informed consent from the patients.

Results

Classification of cohort and clinical characteristics

This study included 59 patients from 59 families (36 boys and 23 girls) suspected with CHARGE syndrome, with an average age at evaluation of 3 years and 2 months (range 0 months to 19 years). Patients were classified according to the Verloes and Hale criteria [5, 6], as illustrated in Fig. 1. Of the 59 patients, 44 (74.6%) were diagnosed with CHARGE syndrome according to the Verloes criteria, and 40 patients (67.8%) met the Hale criteria. A total of 38 patients met both criteria, whereas 13 patients suspected with CHARGE syndrome did not meet either the Verloes or Hale criteria.

Patients were categorized as follows: those meeting both diagnostic criteria [Group A, 38 patients (64.4%)], those meeting only the Verloes criteria [Group B, 6 patients (10.2%)], those meeting only the Hale criteria [Group C, 2 patients (3.4%)], and those meeting neither criteria [Group D, 13 patients (22.0%)]. The phenotypes of the patients are detailed in Table 1 and Supplementary Table 1. The frequency of phenotypes for each group was analyzed, i.e., those meeting at least one diagnostic criterion (Groups A–C) and the entire cohort (Groups A–D), and the obtained frequency was consistent with that reported in previous studies [3, 6, 21,22,23,24,25,26,27] (Table 1).

Table 1 Frequency of clinical manifestations in patients suspected with CHARGE syndrome: A comparative analysis with previous literature

Full size table

Major criteria including coloboma, choanal atresia, and ear anomalies were highly frequent (> 70%), with ear anomalies present in up to 81.4% of the entire cohort (Groups A–D) and in 91.3% of those fulfilling either the Verloes or Hale criteria (Groups A–C). In the minor criteria, hearing impairment, short stature, cardiac malformation, and structural brain anomalies were also highly frequent ( > 70%). However, choanal atresia in the major criteria, anosmia, growth hormone deficiency, tracheoesophageal fistula, cleft lip or palate, renal anomalies, and skeletal or limb anomalies in the minor criteria were observed at a lower frequency ( < 30%), differing from previous reports.

Detection of CHD7 pathogenic variants

The results of genetic testing are shown in Fig. 2. Among the 59 patients, 36 (60%) harbored pathogenic variants in CHD7 as evaluated by single-gene sequencing or MLPA analysis (listed in Supplementary Table 1 and Supplementary Table 2). The detection of CHD7 pathogenic variants varied by group classification as follows: Group A (33/38, 86.8%), Group B (1/6, 1.7%), Group C (2/2, 100%), and Group D (0/13, 0%). Of those who met either the Verloes or Hale criteria (Groups A–C), 78.3% (36/46) harbored pathogenic variants in CHD7.

The CHD7 pathogenic variants were categorized by genotype as follows: nonsense variants (n = 17) were the most frequent, followed by frameshift (n = 10), missense (n = 4), and splicing variants (n = 3). Two CNVs of CHD7 were identified through MLPA analysis, viz., an exon 3-38 deletion and a whole gene deletion. Four variants in CHD7, viz., c.1451del, c.2998dup, c.3424 G > T, and c.3522 + 1 G > T, have not been reported previously.

Exploring alternative genetic explanations for clinically suspected CHARGE syndrome in patients lacking CHD7 pathogenic variants

For 23 patients who did not harbor pathogenic variants, further investigations were conducted, including karyotyping (n = 18), chromosome microarray (CMA, n = 9), ES (n = 7), and GS (n = 7).

Among 10 patients diagnosed with CHARGE syndrome according to either the Verloes or Hale criteria (Groups A–C), two different genetic diseases were identified in 2 patients (Supplementary Table 1).

Patient S40, a 1-year-old boy, was diagnosed with atypical CHARGE syndrome according to the Verloes criteria, but not the Hale criteria. He had intrauterine growth retardation (IUGR), macrocephaly, frontal bossing, short palpebral fissures, facial asymmetry, cleft palate, hypotonia, club feet, hypospadias, bilateral cryptorchidism, ventriculomegaly, duplex kidney, hearing loss, developmental delay, and congenital heart anomaly. Additional findings encompassed a choroid fissure cyst, hydrocephalus requiring endoscopic third ventriculostomy, hypothyroidism, and pigmented retinopathy. Single-gene testing for CHD7 showed normal findings, and his karyotype was 46,XY. At 2.7 years, CMA revealed a pathogenic 2.5-Mb deletion at 19q13.32-q13.33, comprising 99 genes, of which 63 are remarkable OMIM-listed genes.

Patient S18, a female born at full term with IUGR, was diagnosed with typical CHARGE syndrome according to both Verloes and Hale criteria. She had microcephaly, unilateral cataract, coloboma, anotia, ventriculomegaly, thinning of the corpus callosum, and congenital heart anomaly. GS revealed a previously unreported, de novo pathogenic variant in OTX2, c.728dup (p.Ala244Serfs*16), and her diagnosis was changed to microphthalmia, syndromic 5 (MCOPS5; OMIM#610125).

Identification of genetic background in patients sharing phenotype with CHARGE syndrome

In Group D, different genetic causes were identified in five patients (Supplementary Table 1).

Patient S51 had congenital heart anomaly, ear anomaly, facial palsy, right ptosis, frontal bossing, and hypotonia. She died due to progressive heart failure at the age of 4 months. Chromosome analysis revealed 46,XX,del(4)(p15.3) (Wolf–Hirschhorn syndrome; OMIM#194190).

Patient S52 had microcephaly, finger clinodactyly, hearing loss, agenesis of the corpus callosum, epilepsy, and developmental delay. ES revealed a 3-Mb deletion at 4p16.3 (Wolf–Hirschhorn syndrome; OMIM#194190).

Patient S53 showed frontal bossing, facial asymmetry, microphthalmia, low-set ears, webbed neck, hypertrophic cardiomyopathy, developmental delay, and short stature. ES revealed a de novo pathogenic variant, c.1924G>A (p.Asp642 Asn), in PLCB4 (auriculocondylar syndrome 2A; OMIM#614669).

Patient S56 had low-set ears, high-arched palate, short digital phalanges, congenital heart anomaly, multicystic dysplastic kidney, hearing loss, cryptorchidism, and hypoplastic inferior vermis of the cerebellum. GS revealed a de novo pathogenic variant, c.3316 G > A (p.Glu1106Lys), in TRRAP (developmental delay with or without dysmorphic facial features and autism) (DEDDFA; OMIM#618454).

Patient S59 presented hypertelorism, flat nasal root, micrognathia, high-arched palate, facial asymmetry, polydactyly, congenital heart anomaly, laryngomalacia, and feeding difficulty. She had only two minor features of CHARGE syndrome with a normal karyotype and no pathogenic variant in CHD7. CMA revealed a pathogenic 2.1-Mb deletion at 1q21.1-q21.2.

To summarize, this comprehensive genetic evaluation identified seven different genetic causes in 43 of 59 patients (72.9%), including CHD7 (36 patients, 61.0%) variants, Wolf–Hirschhorn syndrome (2 patients, 3.4%), 1q21.1 deletion (1 patient, 1.7%), 19q13 deletion (1 patient, 1.7%), OTX2 (1 patient, 1.7%), PLCB4 (1 patient, 1.7%), and TRRAP (1 patient, 1.7%) variant.

Discussion

In this study involving 59 patients suspected with CHARGE syndrome, the results emphasize the complexity of diagnosing this disease, which often requires a multifaceted genetic evaluation. This study categorized patients based on two diagnostic criteria. The Verloes criteria rely solely on clinical features, whereas the Hale criteria include CHD7 pathogenic variants as major criterion, allowing for the classification of patients with minor or atypical features of CHARGE syndrome [5, 6]. By presenting the proportion of patients meeting each criterion and the detection rate of CHD7 variants, this study highlights the utility and complementary roles of these diagnostic frameworks in identifying clinically suspected CHARGE syndrome.

Pathogenic variants in CHD7 were detected in 61% of the cohort. For patients who met at least one of the clinical diagnostic criteria, the variant detection rate increased to 78.3%, and for those who met both criteria, the variant detection rate increased to 86.8%, corroborating with the variant detection rate of CHD7 reported by previous studies [6, 7]. Moreover, the discovery of four novel pathogenic variants in CHD7 contributes to the expanding genetic landscape associated with CHARGE syndrome.

In patients in Groups A–C who met at least one of the Verloes or Hale criteria, the frequency of most of the major or minor clinical manifestations was similar to that reported in previous studies; however, the frequency of some findings such as coloboma, choanal atresia, anosmia, dysphagia, and skeletal or limb anomalies was lower than the estimated prevalence in previous studies on CHARGE syndrome [3, 6, 21,22,23, 28,29,30,31]. This discrepancy could be attributed to age at evaluation and missing information in the physical examination and history-taking. Short stature was frequently observed, but growth hormone provocation test was conducted only in patients who showed stunted growth velocity or low IGF-1 levels. Therefore, growth hormone deficiency was confirmed in fewer subjects than in the estimated. The overall cohort analysis indicated that genetic testing for CHARGE syndrome was often considered for patients with ear anomalies, hearing impairment, facial palsy, short stature, laryngomalacia, and multiple structural anomalies of the heart, brain, and kidneys.

The results of our study indicated alternative genetic explanations for patients lacking CHD7 pathogenic variants but presenting with CHARGE-like phenotypes. In fact, the comprehensive genetic testing revealed the six different genetic etiologies of complex phenotypes in seven patients (11.9%).

The 19q13.32-q13.33 microdeletion is a rare genetic disorder with fewer than 30 cases reported since its first description in 1998 [24]. Depending on various genes included in the deletion interval, it can exhibit variable clinical manifestations. Genes such as NAPA, NPAS1, DHX34, SLC8A2, MEIS3, ZNF541, PPP5C, CALM3, CCDC8, PTGIR, DACT3, STRN4, FKRP, SLC1A5, and DMPK on 19q13.32-q13.33 are associated with the development of the brain parenchyma, heart, gastrointestinal tract, and musculoskeletal system [24, 25]. A deletion in this region can cause growth retardation, congenital hearing loss, feeding difficulty, and, in some cases, coloboma and cryptorchidism, sharing some features of CHARGE syndrome. However, the 19q13 microdeletion syndrome can be distinguished from CHARGE syndrome by its characteristic ectodermal dysplasia, such as sparse fine hair, sparse eyebrows, nail dysplasia, and localized skin aplasia. The phenotype of Patient S40 was similar to that of a previously reported case with an overlapping deletion [24, 26].

The gene OTX2, located at 14q22.3, is a member of the subfamily of homeodomain-containing transcription factors and plays a vital role in the development of ocular and craniofacial structures and the pituitary gland [27]. Loss of function due to OTX2 pathogenic variants can result in phenotypes ranging from isolated ophthalmological manifestations such as microphthalmia, anophthalmia, and coloboma to pituitary hormone deficiency and developmental delay. Some patients may also show cleft palate, agenesis of the corpus callosum, hypotonia, and seizure [32, 33]. As OTX is not expressed in the heart, the congenital heart anomaly (atrial septal defect) found in Patient S18 remains unknown to be associated with OTX2.

Variants in the regulatory elements of homeobox genes, including OTX2 and GBX2, which are binding targets of CHD7, may disrupt co-activating and repressing factors, leading to altered Fgf8 expression [34]. Additionally, Chd7 haploinsufficiency has been associated with reduced expression of Otx2 and Fgf10 [35]. The potential mechanistic link between CHD7 and OTX2 may provide an explanation for the overlapping phenotypes.

Pathogenic variants in key transcription factors such as OTX2, SOX2, PAX2, and MITF cause syndromic microphthalmia, anophthalmia, and coloboma (MAC spectrum), related to early ocular development [36,37,38]. Beyond ocular manifestations, these genetic alterations can result in a broad spectrum of symptoms, including structural brain anomalies, pituitary dysfunction, and hearing loss. The presence of ocular symptoms together with anomalies in other systems requires differential diagnosis from CHARGE syndrome, with prominent ocular symptoms serving as a distinguishing factor.

Furthermore, the discovery of other genetic disorders such as 1q21.1 deletion syndrome, Wolf–Hirschhorn syndrome, auriculocondylar syndrome 2 A, and DEDDFA sharing dysmorphic facial features, microcephaly, developmental delay, congenital hearing loss, and various degrees of anomalies in other organs with CHARGE syndrome emphasized the need to broaden the differential diagnosis for CHARGE syndrome-like disorders.

Since its first description in 1961, there has been an accumulation of data on the phenotype associated with Wolf–Hirschhorn syndrome, a disorder resulting from the deletion of the distal short arm of chromosome 4, with an estimated birth prevalence of 1 in 20,000 to 50,000 [39, 40]. This disease is characterized by typical craniofacial features, prenatal and postnatal growth impairment, intellectual disability, seizures, and hypotonia [41, 42]. It may exhibit ophthalmological and auditory anomalies, as well as congenital heart anomaly and phenotypes such as cleft palate, hypospadias, and cryptorchidism [41]. Although Wolf–Hirschhorn syndrome shares overlapping features with CHARGE syndrome, such as developmental delay, cleft palate, and renal and cardiac anomalies, differences in typical facial dysmorphism between each disorder and the prominent prenatal growth retardation observed in Wolf–Hirschhorn syndrome can serve as distinguishing factors.

TRRAP is a key component of the coactivator complex with histone acetyltransferase activity, essential for transcriptional activation, DNA damage repair and chromatin remodeling [43]. Pathogenic variants in TRRAP exhibit a highly variable phenotype, ranging from isolated neuropsychiatric disorders such as autism spectrum disorder and schizophrenia to varying degrees of impaired intellectual development starting in infancy [44, 45]. Moreover, these variants can be accompanied by facial dysmorphism, short stature, feeding difficulties, and multisystem malformations affecting the brain, heart, kidneys, genitalia, and hearing [46]. There exists a genotype–phenotype correlation associated with missense variants in TRRAP [46]. It shares characteristics with CHARGE syndrome, such as hearing loss, short stature, heart defects, cryptorchidism, and kidney anomalies. Nonetheless, developmental delay and hearing loss are the most distinguishing features of TRRAP-related DEDDFA.

Although this study provides significant insights into the genetic underpinnings of CHARGE syndrome and related disorders, there are several limitations. Despite the fact that the cohort of 59 patients presents a rare genetic disorder, this sample size does not completely represent the broader population of all individuals suspected with CHARGE syndrome. Furthermore, as 13 (22.0%) of our patients manifested the partial overlapping phenotypes of CHARGE syndrome (Group D), the genetic spectrum of our cohort does not represent patients with clinically confirmative CHARGE syndrome but those in whom CHARGE syndrome was considered in the differential diagnosis. Another limitation is that this is a single tertiary-center study, raising a potential bias due to regional or specific diagnostic practices. Despite comprehensive genetic testing, because this study collected the data in a retrospective manner and the genetic diagnostic process was individualized and not generalized, there exists a possibility of missing genetic defects in certain patients. For instance, 11 patients underwent sequencing and MLPA analysis of CHD7 and CMA analysis; however, ES or GS was not performed because these patients were lost to follow-up.

In conclusion, this study emphasizes the complex genetic landscape of patients suspected with CHARGE syndrome. The discoveries of rare genetic conditions overlapping with CHARGE syndrome phenotypes not only enrich our understanding of these ultra-rare disorders but also broaden the list of genetic diseases in the differential diagnosis of CHARGE-like features. Identification of the underlying genetic cause is essential for appropriate genetic counseling and clinical management. This study has also expanded the phenotypic spectrum of patients with CHARGE syndrome. Moreover, considering that the genetic defect was not revealed in even 10.5% of the patients who met both the criteria, further genetic work-up such as long-read genome or transcriptome sequencing is necessary to understand the genetic spectrum of CHARGE or CHARGE-like syndrome.