Elevated Plasma Complement Factors in CRB1-Associated Inherited Retinal Dystrophies

Abstract

Purpose: To determine the profile of inflammation-related proteins and complement system factors in the plasma of CRB1-associated inherited retinal dystrophies (CRB1-IRDs).

Methods: We used the Olink Explore 384 Inflammation II panel for targeted proteomics in 30 cases and 29 controls (cohort I) to identify immune pathways involved in CRB1-IRDs. Genotyping was performed in cohort I and a second cohort of 123 patients from 14 countries and 1292 controls (cohort II).

Results: A significant shift in complement cascade factors was observed in plasma proteomes of CRB1-IRD patients (enrichment for complement cascade, Padj = 3.03 × 10-15). We detected higher plasma levels of complement factor I and complement factor H [CFH] (q = 0.008 and q = 0.046, respectively, adjusted for age and sex), inhibitors of complement component 3 (C3), which correlated significantly (Pearson's coefficient >0.6) with elevated levels of C3 (q = 0.064). The CRB1 missense variants frequently found in patients showed a strong linkage disequilibrium with the common CFH variant rs7535263 (D' = 0.97 for p.(Cys948Tyr); D' = 1.0 for p.(Arg764Cys)), known to be linked with altered plasma CFH-related protein levels. Correction for the CFH genotype revealed significantly elevated plasma levels of CFH-related 2 (CFHR2) in CRB1-IRD patients (q = 0.041).

Conclusions: CRB1-IRDs are characterized by changes in plasma levels of complement factors and proteins of the innate immune system, and linkage between CRB1 and CFH genes implicates functional variants of the CFH-CFHR locus with specific pathogenic variants of CRB1.

Figure 1.

Quality control of plasma proteomic profiling by proximity extension assays combined with NGS. (A) The median and the interquartile range (IQR) of the distribution of NPX values for each of 59 plasma samples (blue) measured in this study and internal plate and array controls (gray). The horizontal and vertical dashed lines indicate ±2.5 standard deviations of all sample medians (x-axis) and IQRs (y-axis). (B) The sample-wise distributions of NPX values for the Inflammation II panel. Boxplots of NPX values for each plasma sample are along the x-axis and NPX values along the y-axis. The center line represents the median. Plasma samples are colored by “QC_Warning” (Warning = >0.3 NPX from the median value of all samples on the plate). (C, D) The first two principal components of 50 plasma samples and 359 proteins before and after batch correction with Combat.

Figure 2.

Plasma complement and coagulation proteins are associated with CRB1-IRDs. (A) Volcano plot showing the –log₁₀(P values) versus the log₂(mean NPX difference between groups) in plasma protein concentrations. Red indicates proteins with a significant increase in protein concentration, and blue indicates proteins with a significant decrease in protein concentration in patients compared to controls. The dotted line indicates P < 0.05. (B) Scatterplots of the levels of CFI, ADAMTS1, PROS1, and SERPIND1 in the plasma of patients (orange) and controls (blue). The median and the interquartile range (IQR) of the distribution of NPX values for each plasma sample are shown. (C) Correlation plot of the 62 proteins most associated with CRB1-IRDs (P < 0.05 from differential expression analysis). Red indicates a positive correlation, white indicates no correlation, and blue indicates a negative correlation between plasma protein analytes. (D) Dot plot showing the top 10 results from enrichment analysis using the plasma proteins associated with CRB1-IRDs, as shown in C.

Figure 3.

Skewed allelic distribution for a common CFH variant in patients with CRB1-IRDs. (A, B) Linkage disequilibrium (r² metric, left; D′ metric, right) for rs7535263 in the CEU superpopulation of the 1000 Genomes Project. Genes in the CFH extended locus and CRB1 are highlighted. (C) Allele frequency distribution for rs7535263 in cohort 1 in cases and controls. (D) The odds ratio, 95% confidence interval, and P value from Fisher's exact test for the A allele of rs7535263 in case-control analysis of cohort 1. (E) The frequency of each CRB1 pathogenic variant in cohort 1. (F) Heatmap of the frequency of patients with rs7535263 genotype and CRB1 Met1041Thr variant dosage in cohort 1.

Figure 4.

Elevated CFI and CFH levels correlate with plasma C3 in CRB1-IRDs. (A) Results from differential expression analysis before (x-axis) and after (y-axis) corrected for the genotype of rs7535263. The –log₁₀(P value) for each protein analyte is shown. Red indicates proteins that became statistically significantly different between cases and controls after adjusting for the genotype of rs7535263, and black indicates the 10 significant plasma proteins (q < 0.05) in our initial group comparison. Purple highlights plasma proteins involved in complement pathways. The dotted line indicates P < 0.05. (B) Boxplots show the levels for CFH and CFHR2 in cases and controls in bulk and stratified by the genotype of CFH variant rs7535263. (C) Correlation plot of the 18 proteins involved in complement pathways. Red indicates a positive correlation, white indicates no correlation, and blue indicates a negative correlation between plasma protein analytes. (D) Boxplot and correlation plots of the levels of CFI, CFH, CFHR2, and C3 in plasma in patients (orange) and controls (blue).