Tatton‐Brown–Rahman syndrome: Novel pathogenic variants and new neuroimaging findings

Tatton‐Brown–Rahman syndrome (TBRS) or DNMT3A‐overgrowth syndrome is characterized by overgrowth and intellectual disability associated with minor dysmorphic features, obesity, and behavioral problems. It is caused by variants of the DNMT3A gene. We report four patients with this syndrome due to de novo DNMT3A pathogenic variants, contributing to a deeper understanding of the genetic basis and pathophysiology of this autosomal dominant syndrome. Clinical and magnetic resonance imaging assessments were also performed. All patients showed corpus callosum anomalies, small posterior fossa, and a deep left Sylvian fissure; as well as asymmetry of the uncinate and arcuate fascicles and marked increased cortical thickness. These results suggest that structural neuroimaging anomalies have been previously overlooked, where corpus callosum and brain tract alterations might be unrecognized neuroimaging traits of TBRS syndrome caused by DNMT3A variants.


| INTRODUCTION
Tatton-Brown-Rahman syndrome (TBRS) is a rare genetic disease caused by pathogenic variants in the DNMT3A gene.Its characteristics include a wide range of clinical features, such as intellectual disability, overgrowth, and facial dysmorphisms (Tatton-Brown et al., 2014, 2018;Tenorio et al., 2020).It overlaps with other overgrowth syndromes with intellectual disability (Tatton-Brown et al., 2018).Individuals with TBRS often exhibit excessive growth, including tall stature and macrocephaly.Most patients have developmental delay or some degree of intellectual disability, ranging from mild to severe (Lane et al., 2020); distinct facial features include a prominent forehead, horizontal and thick eyebrows, narrow palpebral fissures, and a round face.Epilepsy or heart defects have also been described in TBRS.
However, the clinical features of TBRS can vary widely, even among individuals with the same pathogenic variant in the DNMT3A gene (Tatton-Brown et al., 2014).
Limited studies have investigated the specific brain magnetic resonance imaging (MRI) findings in patients with TBRS.However, a few reports have suggested that individuals with DNMT3A pathogenic variants, particularly those with TBRS, may have subtle and unspecific structural brain anomalies like ventriculomegaly or Chiari malformations (Tatton-Brown et al., 2018;Yokoi et al., 2020).
Here, we describe four patients with de novo DNMT3A pathogenic variants.We report their clinical and cognitive features, as well as neuroimaging findings.Brain MRI was performed in all individuals, including DTI and 3D-SPGRT1 sequences, allowing for the analysis of structural anomalies, 3D-tractography, and cortical thickness measurements.

| Ethical considerations
The study was conducted in conformity with the World Medical Association's Declaration of Helsinki and approved by the Local Ethics Committees.From each family, written informed consent was acquired.

| Identification of cases with DNMT3A variants
In search of DNMT3A variants, we examined 910 cases of trio exome sequencing carried out in our department from 2014 onwards.All studies were performed on patients with neurodevelopmental disorders of likely genetic origin.Four patients were found to have pathogenic DNMT3A variants.

| Genetic analysis
Exome sequencing was performed using genomic DNA isolated from whole blood from proband and parents (MagnaPure, Roche) when arrays displayed normal results.Libraries were created using Life Technologies' Ion AmpliSeq™ Exome Kit and quantified by qPCR.The enriched libraries were prepared using Ion Chef™ and sequenced on PI™ Chip in the Ion Proton™ System (Life Technologies) to provide >90% of amplicons covered by at least 20X.Using the Torrent Suite™ Software, signal processing, base calling, alignment, and variant calling were carried out on a Proton™ Torrent Server.Ion Reporter™ Software was used to annotate the variations, and the Genetic Disease Screen (GDS) trio workflow was used to analyze the pedigrees.Using an inhouse software program and a local database, we carried out variant filtering and prioritization.Candidate variants were visualized through IGV (Integrative Genomics Viewer) and subjected to evaluation based on stringent assessments at both the gene and variant levels, taking into account the patient's phenotype and the inheritance pattern.Variants were categorized according to the standards established by the American College of Medical Genetics and Genomics (ACMG).A board of molecular clinical geneticists evaluated each variant classified as pathogenic, likely pathogenic, or variant of uncertain significance and decided which (if any) to report.Causal variants were reviewed with the referring doctor and/or clinical geneticist in every instance.Sanger sequencing was used to confirm the identified variants.

| Neuroimaging
The brain MRIs performed at our center always entail different sequences, including diffusion tensor imaging (DTI) and threedimensional, T1 spoiled gradient (3D-SPGRT1), regardless of the reason for the study.SS-SE echoplanar Diffusion Weighted Image (DWI) sequence (TR: 12000; FOV: 240 mm; sections thickness: 3 mm, 0 spacing; matrix 128 Â 128; bandwidth: 250; 1 nex; diffusion encoding in 45 directions) with a maximum b = 1000 sec/mm 2 was used to obtain DTI images with a 3T system (GE Medical System, Milwaukee, Wisconsin).3D-tractography was done in an off-line workstation using commercially available processing software from the manufacturer (Functool 3D Fiber Tracking, GE, France) based on fiber assignment by contiguous tracking (FACT) method, achieved by connecting voxel to voxel.The threshold values were 0.3 for FA and 45 for the trajectory angles between the regions of interest (ROIs).DTI tracts were also matched with the 3D-T1 weighted data.
Cortical thickness analysis was performed using FreeSurfer software.T1-weighted MRI images are preprocessed for skull stripping and intensity normalization using the recon-all pipeline in FreeSurfer.After preprocessing, the images are visually inspected for segmentation errors, and manual corrections are made if necessary.The corrected images are then run through the automated cortical thickness pipeline in FreeSurfer, which includes surface-based registration, intensity normalization, surface deformation, and cortical thickness estimation.MRI evaluation, tractography, and cortical thickness analysis are routinely performed in patients with intellectual disability and/or autism, as part of MRI interpretations at our institution.Of note, these analyses are always conducted by a radiologist specialized in these techniques who is unaware of the patient's genetic diagnosis.

| Patient 1
The patient is an 8-year-old female, the third child of healthy parents of Spanish origin.She showed an apparent developmental delay, walking at 21 months, and her first words were recorded at 2 years; she had febrile seizures in the first years of her life.There was no relevant family history.Both parents were 32.5 years old at her birth.
Physical examination revealed a weight of 40 kg (98th centile), a height of 141 cm (98th centile), and an OFC of 54.5 cm (97th centile).Some dysmorphic features were observed: round face, prominent forehead, downslanted palpebral fissures, and thick eyebrows.The neuropsychological assessment revealed a full-scale intellectual quotient (IQ) of 45 using the Wechsler Intelligence Scale for Children-Revised (WISC-R; verbal and performance IQ were 45 and 50, respectively).Despite necessary school support, school performance was below the expected level for this age group.Poor emotional regulation, social skills, and attentional functioning were also described.
Routine laboratory screening, including thyroid function and neurometabolic tests, were within the normal range.Results from the sleep video-EEG test and auditory evoked potentials were normal.No abnormalities were found by conventional genetic analyses (karyotype and array comparative genomic hybridization).

| Patient 2
The patient is an 8-year-old male, the only child of healthy parents of Spanish origin.There was no relevant family history.His mother and father were 32 and 34 years old, respectively, when he was born.The

| Patient 3
This patient is a 4-year-old male, the first monozygotic twin of healthy parents of Spanish origin with no relevant family history.Both parents were 37 years old when the children were born.He walked unsupported at 20 months; at the age of 2, he started using some bisyllabic

| Genetic results
Exome sequencing in trio revealed a de novo missense pathogenic variant in exon 19 of DNMT3A: hg19; Chr 2: 25240417; NM_022552.5;c.2207G > A, p.(Arg736His) in the first patient.This is a previously described variant (Tenorio et al., 2020) was present in Patients 3 and 4.This is a novel variant located in cysteine rich ADD domain; in silico predictors (BayesDel, EIGEN, SIFT, DEOGEN2, PROVEAN) describe it as pathogenic.It was classified as pathogenic, according to ACMG classification (PS1, PS2, PP3, PM1, PM2).Allele frequencies found in probands were 52%, 53%, and 44% respectively.All described variants are present in gnomAD (f < 0.0001).Whether these alleles were derived from clonal expansion of hematologic precursors, or from constitutional variants present in gnomAD participants with some TBRS features, is not clear at this time.

| Neuroimaging findings
In all individuals, brain MRI revealed a thin transition of the corpus and splenium of the callosum, small posterior fossa, and a deep left Sylvian fissure.DTI 3D-tractography showed an apparent asymmetry of the uncinate and arcuate fascicles in our four patients (see Figure 1).The analysis of the brain cortical thickness (see Figure 2) demonstrated a high cortical thickness in both cerebral hemispheres (>pc90) in the four patients studied for their age and sex, according to the parameters of our center and previously published data (Frangou et al., 2022).

| DISCUSSION
We report four patients with TBRS, three of them with previously unreported pathogenic variants of the DNMT3A gene.Dysmorphic features in all patients were consistent with those described in the literature (Tatton-Brown et al., 2018;Tenorio et al., 2020).Neurodevelopmental delay was observed in all four patients.Three of them showed mildmoderate psychomotor retardation (one with autistic traits), whereas the fourth had a moderate intellectual disability.Neurodevelopmental disorders are present in all patients with TBRS (Khazaei et al., 2023;Lane et al., 2020;Tatton-Brown et al., 2014, 2018;Tenorio et al., 2020;Yokoi et al., 2020).Patients with this syndrome tend to perform better on verbal tasks than nonverbal reasoning or spatial tasks, and autistic features appear in 45% of cases (Lane et al., 2020).TBRS is potentially an underdiagnosed condition, especially in patients without intellectual disability or autism.More than 30% of missense variants described as pathogenic and causative of this syndrome, particularly those located in the methyltransferase domain, are also present in individuals included in large-scale databases of genomic data (Karczewski et al., 2020;Tatton-Brown et al., 2018).On the one hand, as mentioned before the significance of described DNMT3A variants annotated in gnomAD is not fully clear.However, it is possible that some patients with this syndrome The DNMT3A gene encodes for the DNA methyltransferase alpha 3 enzyme, which plays an essential role in epigenetic regulation.This process is critical for gene expression and cellular differentiation during development.DNMT3A is essential in regulating DNA methylation patterns in stem cells and during early embryonic development (Okano et al., 1998;Yang et al., 2015).Studies have investigated the association of DNMT3A gene function with various diseases, including TBRS (Bozic et al., 2018;Khazaei et al., 2023;Tatton-Brown et al., 2014;Yang et al., 2015).DNMT3A is involved in several functions in the brain, including neurogenesis, synaptic plasticity, learning, and long-term memory (Morris et al., 2014;Sun et al., 2020).It is also necessary for de novo DNA methylation during development, which is crucial for proper neuronal differentiation and maturation (Wu et al., 2012;Zocher et al., 2021).
Given the functions of this gene in brain development, it would be expected to find structural brain abnormalities in patients with pathogenic variants of the gene.Therefore, the low frequency of structural brain abnormalities is interesting (Tatton-Brown et al., 2018;Tenorio et al., 2020;Yokoi et al., 2020).To our knowledge, previous studies have not directly explored tractography or cortical thickness in this context.Thus, the neuroimaging findings we highlight in the present study might be useful to better characterize the neuroanatomy of these patients.First, the omnipresent thin transition of the corpus and splenium of the callosum, small posterior fossa, amygdala herniation, and a deep left Sylvian fissure; these minor anomalies might be related to nonverbal reasoning and visospatial difficulties (Barnby et al., 2022;Chen et al., 2011;Frank et al., 2016;Hearne et al., 2019;Maxfield et al., 2023).Furthermore, 3D-tractography reconstruction revealed anomalies in arcuate and uncinate fasciculi; both tracts appear to play an essential role in social processing, emotional interaction, and cognitive control (Fletcher et al., 2010;Moseley et al., 2016;Olson et al., 2015).However, these alterations are unspecific and have been described in other genetic conditions causative of different neurodevelopmental disorders (Arlinghaus et al., 2011;de la Peña et al., 2021;DeLisi et al., 2005;Yoon et al., 2020).Lastly, the high cortical thickness recorded in all four patients is an intriguing finding.This feature can be related to an abnormal cortical development.Additionally, some studies have suggested that increased cortical thickness may be a compensatory response to reduced neuronal connectivity or function in other brain regions.In healthy individuals, during development, cortical thinning is directly associated with increased cortical connectivity (Ball et al., 2019;Storsve et al., 2016).Most studies on cortical thickness in genetic syndromes associated with intellectual disability have reported a normal hemispheric or global cortical thickness compared to the healthy population, except for significant variations in brain parcel areas, either because they are increased or decreased (Du et al., 2023;Fahim et al., 2012;Giedd et al., 2007; Kippenhan  et al., 2005;Levman et al., 2019).On the other hand, cortical thickness in overgrowth syndromes has been scarcely studied; most studies have shown a normal or even decreased cortical thickness (Zamary & Mamlouk, 2022).This high cortical thickness in patients with this syndrome could help understand the underlying neurobiology of this syndrome; the verification and confirmation of this finding in larger populations with TBRS might help in the diagnosis of patients with overgrowth, neurodevelopmental disorders, and variants of uncertain significance in the DNMT3A gene.
In conclusion, we have described two new pathogenic variants that cause TBRS.Additionally, we have reported newly discovered neuroimaging findings, which are likely due to microstructural cortical defects and/or primary anomalies in connection development (Poretti et al., 2013;Wahl et al., 2010).These alterations could also explain the patient walked unsupported at 24 months.Verbal development was severely delayed.Physical examination revealed a weight of 38 kg (98th centile), height of 143 cm (98th centile), and OFC of 54.5 cm (80th centile).Dysmorphic features were observed: round face, prominent forehead, deeply set and widely spaced eyes, short philtrum, and thick lips.During evaluation, he showed clear autistic features (restricted interests, little structured or symbolic play, stereotyped hand flapping in excitement, little eye contact).Verbal and nonverbal communication was severely impaired.A neurodevelopmental study conducted at age 3 demonstrated a development quotient of 70 with severe language impairment, and associated autistic features registered in formal assessment of autism spectrum disorder (ASD).No abnormalities were found by conventional genetic analyses (karyotype and array comparative genomic hybridization).
words.Physical examination showed a weight of 17 kg (55th centile), height of 105 cm (80th centile), and OFC of 53 cm (90th centile).Dysmorphic features were observed: long face, thick eyebrows, almondshaped and downslanted palpebral fissures, and anteverted nares.At the age of 3, a neuropsychological evaluation showed a global developmental quotient of 77.Results from the sleep video-EEG test and auditory evoked potentials were normal.Other genetic analyses, such as karyotype and array comparative genomic hybridization study, detected no anomalies.3.1.4| Patient 4The patient is a 4-year-old male twin brother of the previous patient.The patient walked unsupported at 22 months.Verbal development was delayed, with the first words at 2 years.Clinical examination revealed a weight of 16 kg (35th centile), height of 105 cm (80th centile), and OFC of 53 cm (90th centile).Facial dysmorphisms were identical to those observed in his brother.His neurodevelopmental assessment demonstrated a global developmental quotient of 75.Karyotype and auditory evoked potentials displayed normal results.

F
I G U R E 1 Brain MRI, tractography, and perfusion map results.(a) Patient 1. Upper line: sagittal reformations of the three-dimensional, T1-weighted, magnetization-prepared rapid gradient echo (3DT1 MPRAGE) sequence showing thin transition of the corpus and splenium of the callosum, small posterior fossa, amygdala herniation, and a deep left Sylvian fissure.Lower line: tractography reconstructions and perfusion map showing asymmetry of the left arcuate and uncinate tracts; and a normal perfusion map.(b) Patient 2. Upper line: sagittal reformations of the 3DT1 MPRAGE sequence showing the same MRI structural findings as Patient 1. Lower line: tractography reconstructions and perfusion map showing the same tractography findings as Patient 1.There is subtle hypoperfusion in the posterior right temporooccipital lobe.(c, d) Patients 3 and 4. Upper line: sagittal reformations of the 3DT1 MPRAGE sequence showing that both twin brothers have the same radiological structural findings, which are a thin transition of the corpus and splenium of the callosum, small posterior fossa, amygdala herniation, and a deep left Sylvian fissure.Lower line: tractography reconstructions and perfusion map showing that both twin brothers have the same radiological functional findings, which are asymmetry of the left arcuate and uncinate tracts and an evident right hemispheric hypoperfusion.MRI, magnetic resonance imaging.suffer from milder neurodevelopmental problems (being thus underdiagnosed) given the wide range of expressivity of the syndrome (Lemire et al., 2017), and that intellectual disability and autism are overestimated by the selection of the populations studied in previous cohorts.

F
I G U R E 2 Cortical thickness maps.(a) Patient 1. Global increase in cortical thickness, more pronounced in frontotemporal regions.(b) Patient 2. The same pattern of the volumetric cortical thickness map as Patient 1. (c, d) Patients 3 and 4. Both twin brothers have the same cortical thickness map, with high global brain thickness.
intellectual and behavioral features observed in patients with TBRS.The function of DNMT3A in brain development is still not fully understood.Our report highlights the need for further studies to determine whether abnormal cortical patterns and/or disturbances in connectivity are common neuroimaging features of this syndrome.Such research could aid in the diagnosis of patients with genetic variants of uncertain significance.AUTHOR CONTRIBUTIONS Mar Jiménez de la Peña: Formal analysis (brain MRI); writing-review and editing.Irene Rinc on-Pérez: Writing-original draft; writingreview and editing; visualization.Sara L opez-Martín: Writing-original draft; writing-review and editing; visualization.Jacobo Albert: Writing-original draft; writing-review and editing; visualization; funding acquisition.Daniel Martín Fernández-Mayoralas: Writing-review and editing.Ana Laura Fernández-Perrone: Writing-review and editing.Ana Jiménez de Domingo: Writing-review and editing.Pilar Tirado: Writing-review and editing.Beatriz Calleja-Pérez: Writingreview and editing.Javier Porta: Resources; formal analysis; writingreview and editing.Sara Álvarez: Resources; formal analysis; writingreview and editing.Alberto Fernández-Jaén: Conceptualization; methodology; formal analysis; investigation; resources; writing-original draft; writing-review and editing; visualization; project administration.