Functional characterization of 105 factor H variants associated with aHUS: lessons for variant classification

Abstract

Atypical hemolytic uremic syndrome (aHUS) is a life-threatening thrombotic microangiopathy that can progress, when untreated, to end-stage renal disease. Most frequently, aHUS is caused by complement dysregulation due to pathogenic variants in genes that encode complement components and regulators. Among these genes, the factor H (FH) gene, CFH, presents with the highest frequency (15% to 20%) of variants and is associated with the poorest prognosis. Correct classification of CFH variants as pathogenic or benign is essential to clinical care but remains challenging owing to the dearth of functional studies. As a result, significant numbers of variants are reported as variants of uncertain significance. To address this knowledge gap, we expressed and functionally characterized 105 aHUS-associated FH variants. All FH variants were categorized as pathogenic or benign and, for each, we fully documented the nature of the pathogenicity. Twenty-six previously characterized FH variants were used as controls to validate and confirm the robustness of the functional assays used. Of the remaining 79 uncharacterized variants, only 29 (36.7%) alter FH expression or function in vitro and, therefore, are proposed to be pathogenic. We show that rarity in control databases is not informative for variant classification, and we identify important limitations in applying prediction algorithms to FH variants. Based on structural and functional data, we suggest ways to circumvent these difficulties and, thereby, improve variant classification. Our work highlights the need for functional assays to interpret FH variants accurately if clinical care of patients with aHUS is to be individualized and optimized.

Graphical abstract

Figure 1.

Variants showing weak FH bands under nonreducing SDS-PAGE. Coomassie-stained gels showing each FH variant compared with equal amounts of wild-type (WT) FH protein (measured as optical density at absorbance at 280 nm). Asterisks indicate FH variants that were previously described to be associated with no-expression alleles or decreased FH plasma levels (see supplemental Table 3). Differences in expression levels were confirmed in all cases by enzyme-linked immunosorbent assay measurements.

Figure 2.

Mapping of the tested FH variants. FH N-terminal (A) and C-terminal (B) crystal structure. The surfaces that interact with C3b (N terminus) and C3d (C terminus) are depicted in yellow. The FI-interacting surface in the N terminus and the sialic acid binding site in the C terminus are depicted in purple. The position of all variants is indicated in orange. FH variant classification codes are based on functional assay results: benign (green), nonexpressed (black), and functionally impaired (red).

Figure 3.

Distribution of the tested FH variants in the FH sequence. (A) Schematic representation of the amino acid sequence of the 20 FH SCRs. Amino acids corresponding to the SCR consensus sequence are depicted in black, those functionally relevant are in yellow, and all other amino acids are in gray. FH variant classification codes are based on functional assay results: benign (green), nonexpressed (black), and functionally impaired (red). (B) Breakdown of CADD computational predictions using a default score value cutoff of 15: benign predictions (upper panel) and pathogenic predictions (lower panel). Residue types from left to right are “SCR consensus,” “functionally relevant,” and “other.” (C) Distribution of FH classified experimentally as pathogenic (red dots) or benign (green dots) according to CADD scores. Horizontal lines indicate the selected cutoff for each FH region. (D) Breakdown of CADD computational predictions using adjusted score value cutoff according to functional domain: benign predictions (upper panel) and pathogenic predictions (lower panel). Residue types from left to right are “SCR consensus,” “functionally relevant,” and “other.”