Genetics of Moebius syndrome and other congenital facial weakness disorders

National Human Genome Research Institute

Investigators

Linked publications & trials

Abstract

In 2014, we established an NHGRI protocol (14-HG-0055, ClinicalTrials.gov ID: NCT02055248; PI: Eirini Manoli) dedicated to defining the phenotypes and the genetic factors associated with MBS and other congenital facial weakness (CFW) disorders. This protocol was partially funded by a 2014 U01 grant 1U01HD079068-03 (coPIs: Jabs, Engle, Manoli, Brooks, Pierpaoli) and a 2-year research grant from the Moebius Syndrome Foundation, awarded in January 2017. We have enrolled 80 probands and 96 family members as well as 31 healthy controls for brain imaging data analysis, for a total of 207 subjects. Standardized multisystem phenotyping of 143/207 participants has been completed at the NIH CRC. Since June 2019, the NIH protocol is open only for data analysis. Our multisite collaboration enrolled 584 individuals, 211 affecteds from 203 families, including 17 multiplex families, and established a patient registry/clinical database in REDCap. We have sequenced the exomes of 148 individuals, including 37 proband-parent trios and the genomes of 18 individuals (including 5 trios). The Gabriella Miller Kids First Pediatric Research Program awarded resources (Dr. Engle, BCH) for whole genome sequencing at the Baylor Sequencing Core of 142 individuals and their family members with MBS or hereditary congenital facial paresis (HCFP). By jointly analyzing 148 exomes sequenced at the NIH and 142 whole genomes sequenced by our collaborator (Engle, BCH), we identified 445,980 variants (SNPs/Indels) in the coding region of several candidate genes. We prioritized candidate variants for follow-up based on allele frequencies and segregation with affection/disease status. We identified autosomal recessive mutations in myomaker (MYMK) as the cause for Carey-Fineman-Ziter syndrome (CFZ) in two multiplex families enrolled through John Carey (University of Utah) and Elizabeth Engle (BCH). In collaboration with Stephen Robertson (University of Otago, New Zealand) and Eric Olson (University of Texas Southwestern) we confirmed the functional relevance of these mutations in human myoblasts and the zebrafish model system. These results were published at Nature Communications. Genome analyses in families with HCFP1 identified six heterozygous non-coding single nucleotide variants in a conserved cis-regulatory element (cRE2), located 3 of DNAJB8 and GATA2, in seven families and two overlapping tandem duplications in two families, for a total of 34 mutation-positive individuals. ENCODE chromatin state data for a neuroblastoma cell line supports a potential regulatory role for this region and follow-up functional studies suggests the presence of a cis-acting silencer operating in conjunction with nearby enhancers to regulate cell-type specific expression of GATA2. The mutated cis-regulatory element no longer binds NR2F1 (COUP-TF1) or functions as a silencer, resulting in a disruption of the developmental switch of neuronal progenitors from the inner ear efferent neurons (IEEs) to facial branchial motor neurons in rhombomere 4 early in embryonic development. Two mouse models related to this work were developed and characterized by the Boston and Mt Sinai groups, with a humanized cRE1 duplication mouse manifesting absent whisker movement consistent with the human phenotype of HCFP1, while additional affected families were identified in collaboration with Hans Van Bokhoven (Nijmegen, Netherlands). This work has recently been published at Nature Genetics. From the remaining 158 sporadic cases, we diagnosed one subject with a de novo pathogenic variant in TPM2, causing an autosomal dominant myopathy, and identified three subjects with variants of unknown significance in genes associated with different diagnoses (PIEZO2, MEPE and CHRNE). Seven genes with de novo variants (USP15, POU2F1, CNTRL, MINDY1, KPNA3, STMN3, PALM) and 5 genes with biallelic variants (ZNF407, KCNAB2, ZRANB1, MORC2, and PBXIP1) were submitted to GeneMatcher. Three of these genes encode deubiquitinating enzymes USP15, MINDY1, and ZRANB1. A fourth gene, KPNA3, facilitates nuclear import of a deubiquitinating enzyme, ATXN3. Putative matches were made for several genes, but none of the individuals with mutations in these genes had MBS or CFW phenotypes. None of the genes were mutated in more than one proband and there were no variants in PLXND1 and REV3L, two genes previously reported in association with MBS, in our cohort. We suggest that MBS is not likely caused by highly penetrant germline mutations but perhaps to somatic, non-coding, oligogenic or other complex modes of inheritance that alone or in combination with environmental factors alter brainstem and other organ embryogenesis. This is supported by our observation that sporadic MBS is not transmitted to offspring as the family histories of our cohort show no vertical (9 probands with 17 unaffected offspring) or horizontal transmission (178 unaffected siblings). We also performed genome sequencing of brainstem-derived DNA from two unrelated probands and blood-derived DNA from proband/parent trios to investigate the role of somatic mutations. Two full-body research autopsies were performed at the NIH CRC. Small areas of calcification (not observed in brain MRI or CT) were identified at the predicted location of Cn VI, with gliosis in the region of Cn VII. Aberrant dorsoventral tracts were documented near the expected Cn VI tract. The case of complex Moebius was a younger child with Moebius syndrome associated with severe intellectual disability, short stature, club feet, history of ileal atresia and severe gastrointestinal dysmotility/recurrent obstructions requiring multiple surgeries. We used tensor-based morphometry (TBM), developed by co-PI Carlo Pierpaoli (NIBIB) to image the brain and identified a small region of highly significant volumetric reduction in brainstem structures important for the initiation and coordination of conjugate horizontal gaze. This indicates the presence of atrophy or hypoplasia of the brainstem at this level in MBS subjects. This study was published in Brain Communications. We pursued functional MRI studies with collaborators at NIMH (Japee Shruti and Leslie Ungerleider) to understand differences in emotion processing in MBS subjects compared to controls. We identified deficits in fearful emotion processing in subjects with MBS that are accompanied by reduced engagement of the amygdala. Our manuscript describing this finding was accepted in Cortex. In collaboration with colleagues at NINDS, NIMH and NIBIB (Zhen Ni, Mark Hallett, Shruti Japee, Carlo Pierpaoli), we performed fMRI, DTI, and electrophysiology studies for seven MBS subjects with mirror movements and compared with 14 age/gender-matched control subjects. We identified three patient groups with bilateral, unilateral and no mirror movements. Notably, ipsilateral motor evoked and somatosensory evoked potentials, and decreased interhemispheric inhibition were observed on both sides in patients with bilateral mirror movement. DTI findings showed decreased or absent pyramidal decussation at the level of the medulla. We conclude that a certain portion of the motor and sensory pathways do not decussate in Moebius syndrome subjects with mirror movement, suggesting abnormalities in neuronal circuits outside the cranial nerve VI and VII region in the brainstem. Lastly, we performed multimodality analysis of craniofacial and dental phenotypes in 60 subjects evaluated at the NIH CC with collaborators at NIDCR (Janice Lee and Konstantinia Almpani). Distinct craniofacial shape and moderate-severe intraoral anomalies were identified between each of the syndromes causing CFW, illustrating the impact of facial nerve or muscle pathology on craniofacial skeletal development.