A Rare Genetic Polymorphism In C5 Confers Poor Response To The Anti-C5 Monoclonal Antibody Eculizumab In 11 Japanese Patients With PNH

Eculizumab is a humanized monoclonal antibody targeting the terminal complement protein C5 and inhibiting terminal complement-mediated hemolysis associated with paroxysmal nocturnal hemoglobinuria (PNH). In the Japanese AEGIS PNH-eculizumab study, 2 poor-responders were identified out of 29 cases. Currently, more than 300 patients have been treated with eculizumab, and a total of 11 poor-responders were identified all of whom are Japanese.

To clarify the mechanism of difference in the responsiveness of eculizumab, blood samples from poor and good responders were analyzed after obtaining informed consent. Approval for these studies was obtained from the institutional review boards at each study site taking care of patients as well as from Osaka University. The levels of lactate dehydrogenase in these two patients were markedly elevated before eculizumab treatment, and were not decreased during the 12 weeks AEGIS study. From the pharmacokinetic analysis, peak and trough levels of eculizumab during the study were well above the minimal level required to completely inhibit complement-mediated hemolysis in PNH patients. The pharmacodynamics of eculizumab were determined by measuring the capacity of the patients’ serum to lyse chicken erythrocytes in a standard hemolytic assay. Serum samples analysed from these two patients failed over the entire treatment period, to show a suppression of hemolysis, prompting further study of the effect of exogenous eculizumab on the hemolytic activity of patient pre-drug sera. Eculizumab up to 2000μg/mL did not block hemolytic activity in the sera of either poor-responder. However, hemolytic activity both in the two poor-responders and in control patient was blocked completely using a different anti-C5 antibody (N19/8) at 50μg/mL. Therefore, the DNA of C5 from Japanese PNH patients with a good or poor response to eculizumab was sequenced, and a single missense C5 heterozygous mutation at exon 21, c.2654G>A, which predicts p.Arg885His, was found in all of the 11 poor responders identified to date, but not in any of the responders. Among about 300 Japanese patients treated with eculizumab, 11 patients (about 3.7%) have been identified as poor responders. A similar prevalence (3.5%) was seen in healthy volunteers, since we determined that 10 out of 288 Japanese healthy volunteers have the same mutation. This polymorphism was also identified in 1 out of 120 China Han healthy volunteers, but not in 100 persons of British ancestry living in England and Scotland, and not in 90 persons of Mexican ancestry in Los Angels.

To close the genotype-phenotype loop, electrophoretically pure recombinant C5 (rC5) and rC5 mutant (rC5m) containing c.2654G>A were generated and functionally compared in various in vitro experiments. As a preliminary experiment, we confirmed that natural C5, rC5, and rC5m restored classical pathway lysis equivalently when added to C5-depleted serum. Eculizumab did not block classical pathway lysis reconstituted with rC5m but did block rC5 and nC5-dependent lysis. By contrast, as observed with patient sera, N19/8 inhibited lysis reconstituted with nC5, rC5, and rC5m. Finally, while eculizumab bound nanomolar concentrations of rC5 using surface plasmon resonance, with clear association and dissociation phases, there was no detectable binding with rC5m in the same assay up to the highest concentration (1 µM) of eculizumab examined.

A single missense C5 heterozygous mutation, c.2654G>A, which predicts p.Arg885His, was commonly identified in poor-responders, but not in responders. This polymorphism had at least spread to other East Asian countries. After determining that the poor responders likely express both wild-type C5 and a structural variant C5, we then showed that the hemolytic activity supported by this structural variant in vitro, like the effects on patient sera, was not blocked by eculizumab but was fully blocked by N19/8, and that the variant was incapable of binding eculizumab. Collectively, these data are consistent with the hypothesis that the functional capacity of the mutant C5 together with its inability to bind to and undergo blockade by eculizumab fully account for the poor response in patients carrying this mutation.

(JN and MY contributed equally to this work)