Anion-exchange purification of recombinant factor IX from cell culture supernatant using different chromatography supports

• Anion-exchange chromatography of recombinant factor IX from cell culture supernatant was studied.
• Highest binding capacity was found for a monolith, when compared to resins and a membrane adsorber.
• A DOE-based study allowed selective pseudoaffinity elution of functional, gamma-carboxylated FIX molecules.
• Upon 8-fold scale-up, the monolith promoted 7182-fold and 1256-fold clearance of DNA and HCP, respectively.
• Activated FIX (FIXa) was monitored and the FIXa/FIX ratio in the eluate was in an acceptable range.

Both recombinant and plasma-derived factor IX concentrates are used in replacement therapies for the treatment of haemophilia B. In the present work, the capture step for a recombinant FIX (rFIX) purification process was investigated. Different strong anion-exchange chromatography media (the resins Q Sepharose® FF and Fractogel® TMAE, the monolith CIM® QA and the membrane adsorber Sartobind® Q) were tested for their rFIX binding capacity under dynamic conditions. In these experiments, crude supernatant from CHO cells was used, thus in the presence of supernatant contaminants and mimicking process conditions. The highest dynamic binding capacity was obtained for the monolith, which was then further investigated. To study pseudoaffinity elution of functional rFIX with Ca2+ ions, a design of experiments to evaluate the effects of pH, NaCl and CaCl2 on yield and purification factor was carried out. The effect of pH was not statistically significant, and a combination of no NaCl and 45 mM CaCl2 yielded a good purification factor combined with a high yield of active rFIX. Under these conditions, activity yield of rFIX was higher than the mass yield, confirming selective elution of functional, γ-carboxylated rFIX. Scaling-up of this process 8 fold resulted in very similar process performance. Monitoring of the undesired activated FIX (FIXa) revealed that the FIXa/FIX ratio (1.94%) was higher in the eluate than in the loaded sample, but was still within an acceptable range. HCP and DNA clearances were high (1256 and 7182 fold, respectively), indicating that the proposed process is adequate for the intended rFIX capture step.