Integrative analysis of Lunatic Fringe variants associated with spondylocostal dysostosis type-III

Lunatic Fringe (LFNG) is required for spinal development. Biallelic pathogenic variants cause spondylocostal dysostosis type-III (SCD3), a rare disease generally characterized by malformed, asymmetrical, and attenuated development of the vertebral column and ribs. However, a variety of SCD3 cases reported have presented with additional features such as auditory alterations and digit abnormalities. There has yet to be a single, comprehensive, functional evaluation of causative LFNG variants and such analyses could unveil molecular mechanisms for phenotypic variability in SCD3. Therefore, nine LFNG missense variants associated with SCD3, c.564C>A, c.583T>C, c.842C>A, c.467T>G, c.856C>T, c.601G>A, c.446C>T, c.521G>A, and c.766G>A, were assessed in vitro for subcellular localization and protein processing. Glycosyltransferase activity was quantified for the first time in the c.583T>C, c.842C>A, and c.446C>T variants. Primarily, our results are the first to satisfy American College of Medical Genetics and Genomics PS3 criteria (functional evidence via well-established assay) for the pathogenicity of c.583T>C, c.842C>A, and c.446C>T, and replicate this evidence for the remaining six variants. Secondly, this work indicates that all variants that prevent Golgi localization also lead to impaired protein processing. It appears that the FRINGE domain is responsible for this phenomenon. Thirdly, our data suggests that variant proximity to the catalytic residue may influence whether LFNG is improperly trafficked and/or enzymatically dysfunctional. Finally, the phenotype of the axial skeleton, but not elsewhere, may be modulated in a variant-specific fashion. More reports are needed to continue testing this hypothesis. We anticipate our data will be used as a basis for discussion of genotype-phenotype correlations in SCD3.