Structural Modeling Analysis to Prioritize ZNF408 Variants as Candidates Linked to Familial Exudative Vitreo-Retinopathy (FEVR)

Document Type

Conference Proceeding

Publication Date

6-2025

Publication Title

Investigative Ophthalmology and Visual Science

Abstract

Purpose : ZNF408 (PRDM17) is a zinc-finger transcription factor critical for normal retinal vascular development, and some variants have been implicated in FEVR, such as the H455Y variant, known to impair DNA binding. This study examines seven ZNF408 variants, found in a cohort of 33 patients, using AI predictive structural modelling combined with allele population frequency data to prioritize candidates for potential experimental validation in FEVR pathogenesis.

Methods : DNA sequencing used an ERI Ampliseq targeted gene panel, or PCR-Sanger sequencing. Subjects were consented for participation in the ARC Eye-Bank and DNA-sequencing under Oakland University IRB approval. Target genes included: NDP, CTNNB1, TSPAN12, KIF11, FZD4, LRP5, ZNF408, and RS1. AlphaFold3 was used to model secondary and tertiary structure of ZNF408 and variants to predict structural impact. Models were then visualized and annotated in PyMOL. Variant impact and allele frequency (AF) were determined using the Variant Effect Predictor, ClinVar, and GnomAD databases as well as any published.

Results : Two variants, Glu230Gly (AF 0.000371) and Ser225Phe (AF 0.0000006912), were identified as top candidates for experimental validation based on rarity, clinical association with FEVR, and in silico models. Both were found in FEVR patients without known concurrent mutations in other FEVR-linked genes. Alphafold3 predicted that Ser225Phe forms a new alpha helix with the loss of a beta sheet. Glu230Gly resulted in a loop distortion and less pronounced secondary structural elements near the PR-SET domain, losing a predicted beta sheet.

Conclusions : Based on predicted structural changes, allele frequency, and absence of other FEVR-linked variants in the affected persons, we prioritized two variants (Glu230Gly, Ser225Phe) as top candidates for further investigation. The structural changes suggest structural destabilization, potentially impairing scaffolding activity, DNA or protein interactions, or chromatin remodeling activity. Assays to assess DNA binding, chromatin remodeling, and transcriptional activity are justified. The findings show the utility of integrating in silico AI modeling for prioritizing variants for which no empirical functional data is available, and provide a foundation for further investigation into the mechanisms underlying ZNF408-mediated FEVR pathogenesis.

Volume

66

Issue

8

First Page

1832

Comments

Association for Research in Vision and Ophthalmology ARVO Annual Meeting, May 4-8, 2025, Salt Lake City, UT

Last Page

1832

Link to Full Text

Share

COinS