The structure-function relationship of disulfide bonds in etanercept

Abstract

Etanercept is a TNFα receptor Fc fusion protein used for the treatment of rheumatic disease and psoriasis. Physicochemical and functional investigation of process fractions during development of the etanercept biosimilar GP2015 (Erelzi^®) revealed a correlation between reduced potency and incorrect disulfide bridging between specific cysteines in the receptor domain. This novel structure-function relationship was found to be the molecular basis for reduced potency in recent Enbrel^® batches, which exhibit higher levels of incorrect disulfide bridging. Interestingly, incorrect disulfide bridging was found to be reversible under serum-like redox conditions, restoring potency to normal levels. This redox dependent reversibility suggests that these variants are likely not relevant for clinical efficacy once the drug enters the bloodstream. Nonetheless, incorrect disulfide bridging in etanercept represents a new quality attribute that is critical for biopharmaceutical functionality and should thus be carefully monitored and controlled to guarantee patient safety.

Figure 1

Potency loss in late eluting large scale HIC fractions. Large scale HIC separation of GP2015 (Erelzi^®) resulted in the separation of a main peak from a late eluting shoulder peak. Assessment of potency in these two fractions revealed near 100% in the main peak while the late eluting fraction was found to be inactive. Although high and low molecular weight variants were slightly enriched in the late eluting fraction, this moderate decrease in purity could not explain the loss of potency.

Figure 2

Disulfide bridge structures in etanercept. X-ray crystallography of the GP2015 TNF-receptor domain (blue) in conjugation with TNFα (green) was achieved and provides important structural information for etanercept. The eleven disulfide bridge structures in the receptor domain are in line with those previously reported in the literature for crystal structures of etanercept. The LC-MS based triple digest peptide map (see methods for details) was able to identify each of the eleven disulfide bridge structures as well as four additional incorrect disulfide bonds between cysteines C₁₈-C₇₄, C₇₁-C₈₈, C₇₈-C₈₈ and C₇₁-C₇₄ in both GP2015 (Erelzi^®) and the reference product (Enbrel^®).

Figure 3

Approximation of overall incorrect disulfide bridging using analytical hydrophobic interaction chromatography. Samples with different potencies, such as GP2015 (Erelzi^®) drug substance, process intermediates and late eluting HIC fractions were analyzed using the analytical HIC and the non reducing peptide map methods. The abundance of incorrect disulfide bridging assessed by these methods was found to correlate, allowing approximation of how the relative amount of C₇₈-C₈₈ assessed by the non reducing peptide map corresponds with overall incorrect disulfide bridging in etanercept, as reflected by the HIC post peak.

Figure 4

Correlation of incorrect disulfide bridge C₇₈-C₈₈ and potency. The relative amount of C₇₈-C₈₈ was quantified by non reducing peptide map analysis in different samples including reference product (Enbrel^®), GP2015 (Erelzi^®), process intermediates and late eluting HIC fractions. Evaluation of incorrect disulfide bridging relative to potency in these samples revealed a strong inverse correlation between potency and the relative abundance of C₇₈-C₈₈.

Figure 5

Assessment of potency variability over time. Assessment of potency in reference product Enbrel^® batches sourced from both the EU and US since 2008 reveals a broad distribution with a tightening of values toward the lower end over time, likely due to one or more process changes. Assessment of potency for GP2015 (Erelzi^®) batches over time reveals low batch to batch variability well within the clinically verified range of the reference product. Manufacturing dates are given as month/year.