Interface-guided phenotyping of coding variants in the transcription factor RUNX1

Area of Science:

• Genomics
• Molecular Biology
• Cancer Research

Background:

• Interpreting single-gene missense mutations, particularly in cancer-related genes like RUNX1, presents a significant challenge.
• RUNX1 mutations are implicated in various hematological malignancies, but their precise functional consequences are often unclear.

Purpose of the Study:

• To functionally characterize RUNX1 missense mutations using a scalable screening approach.
• To categorize mutations based on their impact on cellular programs and downstream gene expression.
• To improve the classification of variants of uncertain significance (VUS) in RUNX1.

Main Methods:

• Deployment of scalable functional screening by sequencing (SEUSS), a Perturb-seq method, to introduce and assess 115 RUNX1 mutations.
• Analysis of single-cell RNA sequencing profiles to categorize mutations into wild-type (WT)-like, loss-of-function (LoF)-like, and hypomorphic groups.
• Orthogonal assays for validation of functional categories and computational classifier training for VUS prediction.

Main Results:

• Identification of three distinct functional categories for RUNX1 mutations: WT-like, LoF-like, and hypomorphic.
• LoF-like variants were enriched at the DNA-binding site and frequently recurrent in cancer, though recurrence did not solely predict function.
• Hypomorphic variants influenced protein interactions, affecting gene expression related to nerve growth factor (NGF) response and neutrophil cytokine recruitment.
• RUNX1-binding motifs were found in accessible DNA near differentially expressed genes.
• Reclassification of 16 VUS and development of a classifier to predict the function of 103 additional variants.

Conclusions:

• Scalable functional screening is effective for interpreting missense mutations in genes like RUNX1.
• Mutation recurrence alone is insufficient for predicting functional impact; functional categorization is crucial.
• Understanding the distinct functional impacts of LoF-like and hypomorphic variants provides insights into RUNX1-driven phenotypes.
• Targeting protein interactions offers a promising avenue for defining the phenotypic landscape of missense mutations.