Preferential exclusion chromatography as a capture step for extracellular AAV harvest from adherent and suspension productions.

Preferential exclusion chromatography (PXC) sometimes described as hydrophobic interaction chromatography is a well-known, but not widely used technique for purification of Adeno-associated viruses. It employs high molarity of preferentially excluded cosolvent (salt in our case). The downside of this method is that high molarity of salt can lead to aggregation and precipitation of different compounds from the sample. In the case of viruses that are excreted to medium, the concentration of impurities is much lower compared to cell lysates, and PXC can be used as a first chromatographic, serotype independent step to concentrate and purify adeno-associated virus (AAV). Here, we explored PXC for adherent and suspension harvests using monolithic chromatographic columns (CIMmultus). Suspension extracellular adeno-associated virus, serotype 9 (AAV9) harvest had more impurities compared to adherent harvest, therefore it required higher input regarding method development. Final conditions for suspension harvest included higher molarity of binding salt and using more open channel format of chromatographic column (6 µm channel size). Vector genome analysis with droplet digital polymerase chain reaction (ddPCR) revealed 84% and 97% recovery for suspension and adherent AAV9 harvest, respectively. After PXC capture step, adherent AAV9 was purified by already described ion exchange techniques. Overall process vector genome recovery, from clarified harvest to anion exchange elution fraction, was 54% measured by ddPCR. Residual host cell DNA was measured at 40 ng per 1E13 vector genome, and empty AAV was below 5% in final anion exchange chromatography fraction.

Removal of empty capsids from adeno-associated virus preparations by multimodal metal affinity chromatography.

Separation of empty and full adeno-associated virus capsids by multimodal metal affinity chromatography was investigated using a positively charged metal affinity ligand. A subpopulation of empty capsids eluted first, followed by full capsids, and later by more empty capsids and debris. Empty and full capsid composition of chromatography fractions was evaluated by cesium chloride density gradient centrifugation followed by stratigraphic flow analysis of the centrifuge tube contents, monitored by intrinsic fluorescence. Columns charged with barium, calcium, magnesium, zinc, manganese, and ferric ions gave similar results with respect to capsid separation. Charging with cupric ions maintained resolution between early-eluting empty capsids and full capsids but caused them to elute at lower conductivity. Empty and full capsids were fractionated with Tris-borate gradients, sodium chloride gradients, and magnesium chloride gradients. Recovery of full serotype 9 capsids was 100% with complete elimination of empty capsids. All metal ions bound contaminant subsets that required sodium hydroxide for removal. Columns charged with ferric iron and manganese bound more contaminants than all other metals. Columns charged with calcium, magnesium, barium, and copper bound the least. Contaminant binding on zinc-charged columns was intermediate between the two groups.