Clinical Phenotypes of a Pediatric Cohort with GDF2-Related Hereditary Hemorrhagic Telangiectasia

Abstract

Background/Objectives: Pathogenic variants in the growth differentiation factor 2 (GDF2) gene have been linked to a hereditary hemorrhagic telangiectasia (HHT)-like syndrome, yet their clinical significance remains under investigation. This study reports seven pediatric patients with GDF2 variants from a single center. Methods: We identified children with GDF2 pathogenic variants and variants of uncertain significance (VUS) from the Children's Hospital of Philadelphia Comprehensive HHT Program and cross-referenced the list with a full-text query by GDF2 gene name on >53,000,000 visits to ensure complete ascertainment. Medical records were reviewed retrospectively, and variables of interest were abstracted. Results: The median age at genetic testing was 12 years (range 1.75-16). Reasons for genetic testing included telangiectasias, pulmonary hypertension, familial testing, respiratory symptoms, seizures, developmental disabilities, and lung arteriovenous malformations (AVMs). Four patients had missense VUS, including two novel VUS (c.34C>G; p.Leu12Val, c.41C>T; p.Ser14Phe), while three had pathogenic deletions. All patients experienced epistaxis, starting at a median age of 6 years (range 2-12). Three had telangiectasias. One patient had both a GDF2 VUS and a de novo partial endoglin (ENG) gene deletion. While this patient's symptoms of HHT are likely related to her ENG variant, synergy cannot be excluded, and two first-degree family members with clinically significant epistaxis also have the same GDF2 VUS. Notably, two patients had visceral AVMs-one with a lung AVM and another with a vein of Galen malformation. Conclusions: Interpretation of GDF2 VUS and their relationship to clinical symptoms is challenging given the rarity of these genetic variants and the inadequate diagnostic utility of the current clinical criteria for HHT in the pediatric population. Further research with larger cohorts is necessary to improve the genotype-phenotype correlation in GDF2-related HHT. Carefully collected clinical information with longitudinal follow-up may also assist in refining classification of GDF2 VUS as benign or pathogenic in the future.

Keywords: GDF2; arteriovenous malformation; epistaxis; hereditary hemorrhagic telangiectasia; nosebleeds; pulmonary hypertension.

Figure 1

GDF2 gene map. Exons 1 and 2 indicated by blue boxes. Important domains indicated by the light blue boxes. Variants in CHOP patients are indicated by gray dots on the top. Variants found in the literature indicated by gray and orange dots on the bottom. Gray indicates missense, orange indicates nonsense variants. Of note, three CHOP patients with full gene deletions and two literature patients with full gene deletions are not included in this figure.

Figure 2

Pedigrees of 5/7 patients with GDF2 variants. (a) Pedigree of patients 2 and 3 showing maternally inherited GDF2 VUS. (b) Pedigree of patient 4 showing paternally inherited GDF2 VUS. Father (II, 2) and sister (III, 2) have epistaxis. (c) Pedigree of patient 6 showing proband with de novo ENG and paternally inherited GDF2 VUS. Epistaxis in father (II, 3) and paternal half-brother (III, 4) tracks with the GDF2 VUS. (d) Pedigree of patient 7 showing the possible cerebral vascular malformation in maternal grandmother (I, 4) and cerebral aneurysm at age 28 in maternal first cousin once removed (II, 7). No family member cascade testing has been completed, but the maternal grandmother with a “knot” in the cerebral veins could be relevant if the variant is found to be inherited from the maternal side.

Figure 2

Pedigrees of 5/7 patients with GDF2 variants. (a) Pedigree of patients 2 and 3 showing maternally inherited GDF2 VUS. (b) Pedigree of patient 4 showing paternally inherited GDF2 VUS. Father (II, 2) and sister (III, 2) have epistaxis. (c) Pedigree of patient 6 showing proband with de novo ENG and paternally inherited GDF2 VUS. Epistaxis in father (II, 3) and paternal half-brother (III, 4) tracks with the GDF2 VUS. (d) Pedigree of patient 7 showing the possible cerebral vascular malformation in maternal grandmother (I, 4) and cerebral aneurysm at age 28 in maternal first cousin once removed (II, 7). No family member cascade testing has been completed, but the maternal grandmother with a “knot” in the cerebral veins could be relevant if the variant is found to be inherited from the maternal side.