Functional characterization of 2,832 JAG1 variants supports reclassification for Alagille syndrome and improves guidance for clinical variant interpretation

Abstract

Pathogenic variants in the JAG1 gene are a primary cause of the multi-system disorder Alagille syndrome. Although variant detection rates are high for this disease, there is uncertainty associated with the classification of missense variants that leads to reduced diagnostic yield. Consequently, up to 85% of reported JAG1 missense variants have uncertain or conflicting classifications. We generated a library of 2,832 JAG1 nucleotide variants within exons 1-7, a region with a high number of reported missense variants, and designed a high-throughput assay to measure JAG1 membrane expression, a requirement for normal function. After calibration using a set of 175 known or predicted pathogenic and benign variants included within the variant library, 486 variants were characterized as functionally abnormal (n = 277 abnormal and n = 209 likely abnormal), of which 439 (90.3%) were missense. We identified divergent membrane expression occurring at specific residues, indicating that loss of the wild-type residue itself does not drive pathogenicity, a finding supported by structural modeling data and with broad implications for clinical variant classification both for Alagille syndrome and globally across other disease genes. Of 144 uncertain variants reported in patients undergoing clinical or research testing, 27 had functionally abnormal membrane expression, and inclusion of our data resulted in the reclassification of 26 to likely pathogenic. Functional evidence augments the classification of genomic variants, reducing uncertainty and improving diagnostics. Inclusion of this repository of functional evidence during JAG1 variant reclassification will significantly affect resolution of variant pathogenicity, making a critical impact on the molecular diagnosis of Alagille syndrome.

Keywords: ALGS; Alagille syndrome; JAG1; MAVEs; NOTCH2; VOUS; genomic diagnostics; multiplexed assay of variant effects; variant classification; variant of uncertain significance.

Graphical abstract

Figure 1

Development of an assay for JAG1 membrane expression (A–D) WT JAG1 overexpressed from a construct containing an IRES-eGFP cassette (A) shows membrane expression while the JAG1 pathogenic variants (B) Cys693Tyr (C693Y), (C) Cys911Tyr (C911Y), and (D) Leu37Ser (L37S), show intracellular expression of JAG1. (E) WT and mutant JAG1 cell lines were sorted by flow cytometry using an antibody against the extracellular region of JAG1. WT JAG1 has higher APC fluorescence, indicating membrane expression of JAG1, compared to all three mutants. A mixture of all four cell lines (Mix) shows a bimodal distribution of fluorescence intensity with peaks representing JAG1 non-membrane (APC low) and membrane (APC high) expression. Dotted line represents the lower threshold for the APC high population.

Figure 2

High-throughput functional characterization of a JAG1 variant library (A) Schematic of the JAG1 protein highlighting the SSVL targeted region. Blue lollipops indicate previously reported missense variants. Protein domains are color-coded: purple: C2-like; red: Delta/Serate/Lag2 (DSL); brown: NOTCH2-interacting epidermal growth factor-like (EGF-like) repeats; yellow: EGF-like repeats; orange: transmembrane. (B) Sorting strategy for membrane expression assay. Membrane (APC high; GFP+; blue) and non-membrane (APC low; GFP+; red) populations were gated using WT JAG1 and p.Leu16Arg (L16R) as controls. Red and blue coloration indicates the collected populations. Cells with intermediate APC fluorescence (light gray) were not collected. (C) Distribution of membrane expression scores across the entire SSVL. (D) Heatmap showing tolerance (yellow/green) or intolerance (blue/purple) of each JAG1 aa substitution. White indicates variants that were not tested in the assay. (E) Heatmap showing the tolerance of each aa residue to change. Up to seven alternate variants were assayed per residue, and the number of alternate residues with abnormal membrane expression is plotted. Dark blue indicates sites that are intolerant to variation, green indicates sites that are fully tolerant to variation, and intermediate shades indicate sites with some, but not all, changes disrupting membrane expression. NOTCH2-interacting residues are indicated by red (directly interacting) and orange (likely interacting) lollipops.

Figure 3

Specific aa residues show sensitivity to either gain or loss (A and B) The proportion of variants with low, medium, or high membrane expression scores was plotted for (A) secondary structure and (B) buried and interface residues. (C) Membrane expression scores for aa gain (blue) or loss (red) are plotted. Scores are averaged for each aa across the region. (D) Membrane expression scores plotted for cysteine gain versus cysteine loss across the entire region assayed and just the EGF-like domains. Data points for cysteine loss represent the average score for all variants at that residue. Green data points indicate four cysteines in the EGF domain that are tolerant to variation. Data were analyzed using an unpaired t test. (E) Membrane expression scores for aa that are most intolerant to gain are plotted across four regions. (F) Membrane expression scores for aa that are most intolerant to loss are plotted across four regions. All error bars represent SEM.

Figure 4

JAG1 structural modeling agrees with membrane expression scores (A) Structural modeling of Leu16–18 (L16–18) and Leu20–21 (L20–21) within the SP helix. (B and C) Phobius SP prediction of the (B) Leu20Met (L20M) and (C) Leu20Arg (L20R) variants show a decrease in the confidence of the SP signature for Leu20Arg (L20R) but not Leu20Met (L20M). Red box indicates location of variant. Black box indicates the last residue of the SP domain. (D) Structural modeling of the JAG1 region of study in contact with NOTCH2. Residues of interest are highlighted, including the four tolerant and 14 intolerant EGF-like cysteines. (E–G) Structural analysis of (E) Trp167 (W167) packed tightly between a glutamine and an arginine, (F) Pro269 (P269), and (G) hotspot residues Glu285 (E285), Thr286 (T286), Asn287 (N287), and Trp288 (W288) that are highly sensitive to variation.

Figure 5

JAG1 membrane expression correlates with pathogenic and benign variant classifications (A) Histogram of membrane expression scores for 106 benign and 69 pathogenic variants. Black dotted line indicates the threshold for the lower 5% of benign variants, below which variant function is classified as abnormal. Forty-five out of 69 (65.2%) pathogenic variants within the variant control set lie below this line. The distribution of membrane expression scores for the remainder of the library is shown below, with 486 variants scoring below the abnormal threshold. (B and C) Correlation of (B) AlphaMissense and (C) CADD scores for variants with abnormal membrane expression. Data were analyzed using an unpaired t test.

Figure 6

Functional evidence supports JAG1 VOUS reclassification (A) Sankey plot showing the percent of previously classified JAG1 variants from an internal ALGS research cohort with abnormal membrane expression data. (B) Sankey plot showing the distribution of previously classified variants from ClinVar (all classification categories) and HGMD (DM included with pathogenic classifications, DM? included with likely pathogenic classifications) with membrane expression data. (C) Analysis of VOUSs with abnormal membrane expression data resulted in reclassification of three VOUSs from the ALGS research cohort to likely pathogenic (upper) and evidence to support the reclassification of 24 VOUSs from ClinVar.