Clinicopathologic Implications of Complement Genetic Variants in Kidney Transplantation

Abstract

Genetic testing has uncovered rare variants in complement proteins associated with thrombotic microangiopathy (TMA) and C3 glomerulopathy (C3G). Approximately 50% are classified as variants of uncertain significance (VUS). Clinical risk assessment of patients carrying a VUS remains challenging primarily due to a lack of functional information, especially in the context of multiple confounding factors in the setting of kidney transplantation. Our objective was to evaluate the clinicopathologic significance of genetic variants in TMA and C3G in a kidney transplant cohort. We used whole exome next-generation sequencing to analyze complement genes in 76 patients, comprising 60 patients with a TMA and 16 with C3G. Ten variants in complement factor H (CFH) were identified; of these, four were known to be pathogenic, one was likely benign and five were classified as a VUS (I372V, I453L, G918E, T956M, L1207I). Each VUS was subjected to a structural analysis and was recombinantly produced; if expressed, its function was then characterized relative to the wild-type (WT) protein. Our data indicate that I372V, I453L, and G918E were deleterious while T956M and L1207I demonstrated normal functional activity. Four common polymorphisms in CFH (E936D, N1050Y, I1059T, Q1143E) were also characterized. We also assessed a family with a pathogenic variant in membrane cofactor protein (MCP) in addition to CFH with a unique clinical presentation featuring valvular dysfunction. Our analyses helped to determine disease etiology and defined the recurrence risk after kidney transplant, thereby facilitating clinical decision making for our patients. This work further illustrates the limitations of the prediction models and highlights the importance of conducting functional analysis of genetic variants particularly in a complex clinicopathologic scenario such as kidney transplantation.

Keywords: C3 glomerulopathy; atypical hemolytic uremic syndrome; complement; complement regulators; kidney transplantation; thrombotic microangiopathy; variants of uncertain significance.

Figure 1

Schematic diagram of Factor H (FH) and its functional domains. (A) Factor H is a 155-kDa plasma complement regulatory protein composed of 20 repeating homologous domains of approximately 60 amino acids each, known as complement control protein (CCP) repeats, short consensus repeats (SCRs), or sushi domains (12). Factor H prevents the formation and accelerates the decay of the alternative pathway C3 convertase (C3bBb; decay accelerating activity) and is a cofactor for Factor I-mediated cleavage and inactivation of C3b (cofactor activity). Factor H has fluid-phase and cell-surface regulatory capabilities (13). The N-terminal portion contains the regulatory domain that mediates the cofactor activity and decay accelerating activity (repeats 1–4). The surface binding recognition motifs are located in repeats 6–8 and 18–20. (B–D) Factor H constructs used to prepare the variants (as indicated). Residue numbering includes the 18 amino acids of the signal peptide (SP). CA, cofactor activity; DAA, decay accelerating activity; C3b, C3b binding site; C3d, C3d binding site; GAG, glycosaminoglycan binding site.

Figure 2

Recombinant expression of CFH variants. Western blots of supernatants from transfected wild-type and variant constructs. (A) Recombinant expression of variants I372V (lane 2) and I453L (lane 3) shows secretion comparable to wild type CCP 1–8 (lane 1). (B) Wild type CCP 15–20 as control (lane 1). Secretion of variant protein G918E is undetectable (lane 2). The SNP N1050Y displays higher secretion than WT (lane 6). The combined variant G918E/N1050Y has markedly reduced secretion (lane 7). Variant T956M has normal secretion (lane 5). Combined variant T956M/E936D and the SNP E936D also demonstrate normal secretion (lanes 3 and 4). (C) Wild type CCP 18–20 as control (lane 1). The SNPs N1050Y, I1059T, Q1143E, and variant L1207I show normal secretion (lanes 2–6).

Figure 3

Amino acid sequence alignment of complement control protein (CCP) repeats in FH. Sequence alignments are generated with NCBI protein BLAST software based on position-specific scoring matrix (PSSM). Red to blue color scale represents the degree of conservation; red represents highly conserved residues; blue represents less conserved residues. Upper-case residues are aligned, lower-case residues are unaligned. Mutated residues of FH in the cohort are indicated with arrows. Two disulfide bonds form in each CCP, one between cysteine I and III and the other between cysteine II and IV (red and highlighted).

Figure 4

Structural analysis of I372V and I453L (A) The overall structure of FH CCP 6–8 is shown in a cartoon representation (PDBID:2UWN) (20). CCP 6 is in red, CCP 7 in yellow, and CCP 8 in blue. H402 and the two rare variants, I372V and I453L, are in cyan. Crystallographic structural data indicates that CCP 6–8 contain four glycosaminoglycan (GAG) binding sites (20). The primary binding site is H402 (20); the other three putative binding sites are H360 (not shown) in CCP 6, R341 in CCP 6 (in B), and R444 in CCP 7 (colored green at the linker between CCP 7 and CCP 8). (B) Interface structure of FH CCP 6–7 in complex with FH binding protein (fHbp) derived from Neisseria meningitidis (PDBID: 2W80) (21). Side chains from both proteins involved in forming salt bridges across the interaction surface are shown. The putative heparin binding site is composed of H337 (not shown), R341 and H371 in CCP 6 (21). E304 and R341 form a salt bridge over a distance of approximately 2.6Å. E304 and E283 in cyan from fHbP; R341 and K351 from CCP 6 in red; the variant I372V in orange is flanked by H371 and H373 in blue. Structures are prepared using PyMol. Functional analysis of I372V and I453L (C) (i, iv) Absorbance is plotted against logarithmic protein concentrations. (ii, v) Box-and-whisker plots demonstrating C3b and heparin binding (by ELISA) of I372V and I453L compared to WT. Relative absorbance (RA) is computed as absorbance of the variant divided by absorbance of the WT. Bottom of the box shows the 25^th percentile, the line within the box indicates the median, and the top of the box shows the 75^th percentile. (iii, vi) Representation of C3b and heparin binding of I372V and I453L compared to WT using bar graphs. For C3b binding the P-value for the percentage differences of I372V and I453L compared to WT were 0.69 and 0.65, respectively. For heparin binding, the P-value for the percentage differences of I372V and I453L compared to WT were both <0.001. Data represent three separate experiments with bars corresponding to SEM (standard error of mean). ns, no significant difference; ***P < 0.001. (D) (i, iii) Fluid-phase C3b cofactor activity (CA) of I372V and I453L assessed by the cleavage of purified C3b to iC3b compared to WT. The percentage of alpha chain remaining and the generation of α1 indicates the cleavage of C3b to iC3b. A kinetic analysis of CA was conducted at 0, 5, 10, and 20 min. Cleavage rate is measured by densitometric analysis of the generation of α1 relative to the β chain. Representative Western blot is shown. (ii, iv) Densitometric quantification of the Western blot. Lane 9 represents a negative control employing concentrated supernatant in the presence of FI but absence of FH. Lane 10 is a positive control using full-length FH (FH-FL). Upon comparison to WT FH, there is no difference in CA of I372V and I453L.

Figure 5

Structural analysis of G918E, E936D, and T956M. Nuclear magnetic resonance spectroscopy solution structure of CCP 15–16 (PDB ID: 1HFH) (32). (A) The highest frequency rotamer of the G918E side chain and its clashing residue I868 is highlighted in the orange circle. (B) The highest frequency rotamer of the T956M and E936D side chains is highlighted in cyan. C958 and C990 form a disulfide bond next to the Y957 residue to stabilize the CCP 16 tertiary structure. E936D and T956M are located on the protein surface, spatially distant from the disulfide bond. CCP 15 (orange); CCP 16 (green); linker between CCP 15 and 16 (yellow); cysteine side chains involved in disulfide bond formation (red); G918E, E936D, and T956M variants (cyan sticks, zoomed and reoriented in inset box). Functional analysis of T956M and E936D. (C) (i, iv) Absorbance is plotted against logarithmic protein concentrations. (ii, v) Box-and-whisker plots demonstrating C3b and heparin binding (by ELISA) of E936D, T956M, and combined variant E936D/T956M compared to WT. (iii, vi) Representation of C3b and heparin binding of E936D, T956M, and E936D/T956M compared to WT using bar graphs. Compared to WT there was no difference in heparin binding for E936D, T956M, or E936D/T956M. T956M showed ~35% decrease in C3b binding compared to WT (P < 0.001). Data represent three separate experiments with bars corresponding to SEM. ***P < 0.001.

Figure 6

Structural analysis of L1207I. (A) Structure of the CCP 19–20:C3d complex containing one molecule of CCP 19–20 and two molecules of C3d (PDB ID 2XQW, green and purple) (40). Residues located at the CCP 20 interface are annotated and shown as cyan sticks. Residues located at the C3d interface are shown as magenta sticks. Close-up view of the CCP 20-C3d interface is shown and reoriented in the inset box. R1182, W1183, and T1184 mutations shown as cyan sticks were reported in aHUS cases (40). Functional study of W1183L demonstrated impaired interaction with surface-bound C3b (14). E160 and R1203, D163 and K1230, D292 and R1231 form salt bridges with distances of approximately 2.5, 2.5, and 3.3 Å, respectively. L1207 (in cyan) is located away from the C3d-CCP 20 interaction surface. (B) Sialic acid binding site on complement CCP 20 (PDB ID 4ONT) (41)—C3d (magenta); CCP 19 (orange); CCP 20 (green); α-N-acetylneuraminic acid (Neu5AC, orange stick). L1207 (cyan) is spatially distant from the binding surface between Neu5AC and CCP 20. R1210 mutations (green sticks) have been reported in aHUS. (42). Functional analysis of L1207I (C) (i, iv) Absorbance is plotted against logarithmic protein concentrations. (ii, v) Box-and-whisker plots demonstrating C3b and heparin binding (by ELISA) of L1207I compared to WT. (iii, vi) Representation of C3b and heparin binding of L1207I compared to WT using bar graphs. There is no difference in C3b or heparin binding for L1207I compared to WT. Data represent three separate experiments, bars correspond to SEM.

Figure 7

Algorithm with recommended testing and steps to define clinical significance of genetic variants in patients with TMA or C3G.