Analysis of C3 Gene Variants in Patients With Idiopathic Recurrent Spontaneous Pregnancy Loss

Miscarriage is the most common complication of pregnancy. Approximately 1% of couples trying to conceive will experience recurrent miscarriages, defined as three or more consecutive pregnancy losses and many of these cases remain idiopathic. Complement is implicated both in the physiology and pathology of pregnancy. Therefore, we hypothesized that alterations in the C3 gene could potentially predispose to this disorder. We performed full Sanger sequencing of all exons of C3, in 192 childless women, with at least two miscarriages and without any known risk factors. All exons carrying non-synonymous alterations found in the patients were then sequenced in a control group of 192 women. None of the identified alterations were significantly associated with the disorder. Thirteen identified non-synonymous alterations (R102G, K155Q, L302P, P314L, Y325H, V326A, S327P, V330I, K633R, R735W, R1591G, G1606D, and S1619R) were expressed recombinantly, upon which C3 expression and secretion were determined. The L302P and S327P were not secreted from the cells, likely due to misfolding and intracellular degradation. Y325H, V326A, V3301I, R1591G, and G1606D yielded approximately half C3 concentration in the cell media compared with wild type (WT). We analyzed the hemolytic activity of the secreted C3 variants by reconstituting C3-depleted serum. In this assay, R1591G had impaired hemolytic activity while majority of remaining mutants instead had increased activity. R1591G also yielded more factor B activation in solution compared with WT. R1591G and G1606D showed impaired degradation by factor I, irrespectively if factor H, CD46, or C4b-binding protein were used as cofactors. These two C3 mutants showed impaired binding of the cofactors and/or factor I. Taken together, several alterations in C3 were identified and some of these affected the secretion and/or the function of the protein, which might contribute to the disorder but the degree of association must be evaluated in larger cohorts.