The complement component C5 is not responsible for the alternative pathway activity in rabbit erythrocyte hemolytic assays during eculizumab treatment

Dear Editors,

Eculizumab is a therapeutic complement C5 inhibitor approved for the treatment of atypical hemolytic uremic syndrome (aHUS), paroxysmal nocturnal proteinuria (PNH), generalized myasthenia gravis and neuromyelitis optica spectrum disorder. Correct dosing of eculizumab is important to ensure complete complement blockade. Complement activity is often monitored by measuring either formation of the C5 activation product C5b-9 in vitro using an ELISA, independent of red blood cells, or by hemolytic assays measuring hemoglobin release from erythrocytes. Both approaches allow analysis of complement activation via the classical (CP) and alternative (AP) pathways.

The reported degree of complement inhibition varies in patients who attain target drug concentrations (>100 µg/mL), depending on the assays used. Studies with ELISA-based CP and AP assays^1–5 demonstrate full blockade in patient samples containing target drug concentrations. However, studies with hemolytic assays report various degrees of hemolysis in the presence of eculizumab, which could be interpreted as incomplete C5 blockade by the drug. When the CP hemolytic assay is used, complete eculizumab blockade is usually observed.^1,6–9 Nevertheless, considerable AP hemolytic activity is frequently reported as ongoing under eculizumab treatment even when eculizumab should be in excess.^1,6,10

The question of possibly incomplete blockade becomes especially relevant in cases where a patient does not respond well to eculizumab therapy. Then, it is important to know whether the lack of response may be related to residual C5 activity to determine further treatment strategies by increasing the dosage or choosing other treatment options. Thus, in this study, we investigated whether the residual hemolysis of erythrocytes detected by the complement AP assay in the presence of eculizumab may be caused by incomplete C5 blockade by the drug or could be explained by C5-independent mechanisms.

We analyzed samples from patients treated with eculizumab in erythrocyte-based hemolytic assays. Because the serum concentration used in hemolytic assays varies between the protocols of different laboratories, we tested samples in a broad range of serum concentrations.

Samples from five aHUS patients containing 207−367 µg/mL eculizumab were analyzed. In the AP erythrocyte assay, a high degree of hemolysis was observed, increasing to 73% at the highest serum concentration (50%) (Fig. 1a). In the CP hemolytic assay, all patient samples gave results comparable to those of heat-inactivated normal human serum (HI-NHS), used as a negative control (Fig. S1A), i.e., <5% hemolysis even at a 50% serum concentration. Furthermore, the normal human serum (NHS) control spiked with increasing amounts of eculizumab showed the same high lytic activity in AP with complete blockage of CP in hemolytic assays (Fig. S1B and S1C). Further, we tested whether hemolytic activity in eculizumab-containing samples can be caused by the residual formation of C5b-9. We chose commercial Wieslab CP and AP assays, as they are widely used. No C5b-9 formation was detected in the patient samples or NHS spiked with eculizumab at target values in the AP (Fig. 1b and Fig. S2A) or CP Wieslab assays (Fig. S2B and S2C). Moreover, no soluble C5b-9 (sC5b-9) was formed in NHS activated by zymosan A in the presence of eculizumab, while the formation of the C3 activation markers C3bBbP, C3bc and C3a was unaffected by the drug (Fig. 1c, d, Fig. S3A, S3B). Moreover, the serum of a genetically C5-deficient donor produced lysis of erythrocytes identical to that of NHS spiked with 500 µg/mL eculizumab, both reaching 85% lysis at the highest serum concentration in the AP hemolytic assay but not in the CP hemolytic assay or Wieslab tests (Fig. 1e, Fig. S4).

Fig. 1.

Samples of five aHUS patients containing eculizumab showed a high degree of hemolysis in the alternative pathway (AP) hemolytic assay (a). No C5b-9 was formed in these samples in the AP Wieslab test (b). Moreover, no soluble C5b-9 (sC5b-9, c) was formed during in vitro activation of normal human serum control (NHS) with zymosan A when eculizumab was added; however, generation of the C3 activation product C3bBbP remained unaffected (d). In the AP hemolytic assay, the serum of C5-deficient donor showed lysis similar to that of NHS spiked with 500 µg/mL eculizumab (e). This lysis was suppressed in the presence of 40 µg/mL compstatin Cp40 (f). Rabbit erythrocytes pre-exposed to NHS spiked with 500 µg/mL eculizumab (g) or C5-deficient serum (h) under AP-permissive conditions showed high osmotic fragility, which was restored by the addition of Cp40

As shown above, C3 activation was not affected by the presence of eculizumab. Thus, we investigated whether hemolysis in AP could be affected by C3 activity. We incubated NHS and C5-deficient serum with the potent C3 inhibitor compstatin Cp40. The presence of compstatin alone inhibited hemolysis in NHS and in C5-deficient serum (Fig. 1f).

Erythrocytes remaining after the AP assay (Fig. 1g, h) were more osmotically fragile, with >75% of the erythrocytes exposed to NHS with eculizumab or C5-deficient serum already lysed at 0.7% NaCl, whereas the addition of compstatin Cp40 substantially decreased lysis.

Overall, our results indicate that the erythrocyte lysis observed in the AP hemolytic assay during eculizumab treatment is not caused by residual C5 activity. Residual AP hemolysis most likely occurs due to changes in the fragility of the erythrocyte membrane, rendering it susceptible to mechanical lysis mediated by C3 activation and opsonization with C3 activation products. Erythrocytes are live cells with high sensitivity to storage time, buffer conditions and batch-to-batch variation, which is not the case for stable ELISA assays. This is a well-known phenomenon in complement laboratories over decades, including ours. Therefore, we propose that in general, erythrocyte-based assays, when possible, should be replaced by robust, reproducible ELISA methods in future complement research.

Competing interests

Prof. Dr. Jack F.M. Wetzels is a member of the international advisory board of Alexion and also received a grant from Alexion. Other authors have no conflict of interest.

Supplementary information

The online version of this article (10.1038/s41423-020-0406-y) contains supplementary material.