Achieve Baseline Separation of AAV Full and Empty Capsids for Multiple Serotypes

Separation of full and empty AAV capsids by anion exchange resin with dextran surface extenders Åsa Hagner-McWhirter, Brigitta Németh, Marcus Kjellander, Hans Blom, Ann-Christin Magnusson, and Jean-Luc Maloisel Cytiva, Björkgatan 30, 751 84 Uppsala, Sweden Introduction Adeno-associated virus (AAV) is the main vector for gene therapy, and there is need for scalable, cost-efficient, and robust filtration and chromatography-based purification processes. Key for a successful process is high overall yields of full capsids and efficient impurity removal. Our AAV process can be seen in Figure 1. The affinity step does not discriminate between full and empty capsids, and a polishing step is required to remove as much as possible of the empty capsids. Separation of full and empty capsids can be achieved with ion exchange by utilizing a small difference in pI (Fig 2). Here we show how full and empty separation of AAV2, AAV5, AAV8 and AAV9 can be improved using anion exchange, dextran surface extenders, and optimized conditions. Conclusions • A combination of dextran surface extenders, MgCl2 concentration, and step elution enables baseline separation of AAV full and empty capsids. • Capto Q resin resulted in > 80% VG recovery by qPCR and 75% to 88% full capsids by AUC analysis ( > 95% by qPCR: ELISA). • Reduction of partial capsids was observed for AAV5 and AAV9 by AUC analysis. • One single elution buffer (20 mM BTP, pH 9 with 2 mM MgCl2 and 250 mM sodium acetate) could be used for AAV2, AAV5, AAV8 and AAV9 (likely also other serotypes). • Prescreening is recommended to identify optimal two-step isocratic condition. Results and discussion We identified three critical parameters that enhance the full and empty capsid separation. Dextran surface extenders on the anion exchange resin (Fig 3), MgCl2 constant concentration in buffers, and step elution (Fig 4). We were able to develop an optimized step elution protocol that worked well for full and empty capsid separation of AAV2, AAV5, AAV8 and AAV9 using Capto Q with dextran surface extenders. We used 20 mM BTP, pH 9 containing 2 mM MgCl2 without any additives and replaced NaCl as an elution salt with sodium acetate, which is a softer (more kosmotropic) salt. The elution conductivity was identified in a prescreening procedure using incremental 5% B steps (Fig 5). The % B for empty capsid elution was selected based on the UV 260:280 ratio. The step before the ratio started to increase was selected, and the first elution step was prolonged to 20 CV to maximize the empty capsid removal (Fig 6). This was followed by a second shorter (5 CV) step for elution of full capsids using 100% B or lower (Fig 6). It is possible (and can be an advantage for some serotypes) to adjust the conductivity to maintain the empty capsids in the flowthrough only allowing binding of full capsids. cytiva.com Cytiva and the Drop logo are trademarks of Life Sciences IP Holdings Corporation or an affiliate doing business as Cytiva. ÄKTA, ÄKTA flux, ÄKTA pure, Capto, HyCell, ReadyToProcess, ReadyToProcess WAVE, Tricorn, ULTA, and Xcellerex are trademarks of Global Life Sciences Solutions USA LLC or an affiliate doing business as Cytiva. Denarase is a trademark of c-LEcta GmbH. Tween is a trademark of the Croda Group of Companies. Any other third-party trademarks are the property of their respective owners. © 2022-2023 Cytiva. For local office contact information, visit cytiva.com/contact. CY40940-17Nov23-PO The first peak contained mainly empty capsids while the second peak contained the full capsids as indicated by the UV260:280 ratio of ~ 0.6-0.7 for empty capsid and >1.2 for full capsids, respectively (1). The qPCR:ELISA ratio in the peaks and the calculated viral genome (VG) confirmed good separation, as shown in Table 1. The AAV2 start material contained a very low level of full capsids and the second peak, enriched with full capsids, is both smaller and had a lower UV260:280 ratio compared to the other serotypes. The results were further confirmed by analytical ultracentrifugation (AUC) with ~ 75% full capsids for AAV5 and AAV8 and ~ 88% full capsids for AAV9 in peak 2 (Fig 7). The partial capsids were reduced for AAV5 and AAV9 (Fig 7). The effect of the dextran surface extenders could also be confirmed by AUC. Capto Q resulted in ~ 75% AAV5 full capsids compared to ~ 35% using Capto Q ImpRes (Fig 8). The dynamic binding capacity was estimated to 1–3 x 1013 VP/mL Capto Q resin (results not shown). It is critical to dilute the eluate from affinity purification to reduce conductivity (~ 1–3 mS/cm depending on serotype), to ensure binding of full AAV capsids. We recommend bypassing the mixer on the ÄKTA pure 25 system to reduce dead volume and ensure sharp conductivity steps (important in small scale) and to use a 10 mm path length UV detector cell for increased sensitivity. Make sure to set the pH of the buffer before adding MgCl2 and perform a wash with deionized water before CIP with NaOH to avoid precipitation. Reference 1. Dickerson R, Argento C, Pieracci J, Bakhshayeshi M. Separating empty and full recombinant adeno-associated virus AAV particles using isocratic anion exchange chromatography. BioTechnol J. 2021; 16(1):e2000015. Fig 1. AAV process overview. Upstream production Xcellerex XDR-10 or ReadyToProcess WAVE 25 bioreactors HEK293T sus/ HyCell TransFx-H AAV -GFP triple plasmid transfection Clarification/NFF ÄKTA flux 6 system ULTA 5 + 2 µm GF + 0.6/0.2 HC Concentration and buffer exchange TFF with HF (300 000 NMWC) 10X UF/5X DF 20 mM Tris, 150 mM NaCl, pH 8 Affinity capture ÄKTA pure 150 or 25 system Capto AVB resin Anion exchange polishing ÄKTA pure 25 system Capto Q resin Harvest Cell lysis and DNA fragmentation 150mM NaCl, 0.5% Tween, Denarase 40U/mL, 1 mM MgCl2 Low pH treatment Fig 2. Principle of ion exchange separation in the polishing step. Fig 4. Separation of AAV full and empty capsids using Tricorn 5/100 columns, (2 mL) 1. Capto Q ImpRes or Capto Q with dextran surface extenders (AAV8) or 2. Capto Q with constant 2 mM or 18 mM MgCl2 (AAV8) or 3. Capto Q with 18 mM MgCl2 and linear or step NaCl gradient (AAV5). Buffers used were: 20 mM BTP pH 9.5, 2 or 18 mM MgCl2 + additives (1% sucrose and 0.1% poloxamer 188) (A buffer) and A buffer + 400 mM NaCl (B buffer). Cryo TEM imaging. Vironova AB, Stockholm Sweden. Serotype Start sample Peak 1 (empty capsids) Peak 2 (full capsids) qPCR:ELISA (% full capsids) UV 260:280 (peak area) VG recovery (%) qPCR:ELISA (% full capsids) UV 260:280 (peak area) VG recovery (%) qPCR:ELISA (% full capsids) AAV2 7-10% 0.75 NA NA 1.14 NA NA AAV5 47% 0.65 7 5 1.20 80 100 AAV8 11-35% 0.60 3 1 1.24 80 95 AAV9 40% 0.63 0.3 1 1.25 91 100 1. Dextran extenders 2. MgCl2 in buffers 3. Step elution mode Capto Q O OH HO O O OH O O OH O OH OH HO O O OH O OH O O OH O O OH n N+ OH ClO OH O N+ OH ClO N+ OH ClO HO HO OH O O OH O O HO O HO HO OH O O OH O HO O O OH O O HO n N+ OH ClO OH O N+ OH ClStrong quaternary ammonium (Q) anion exchange ligand Dextran surface extender Highly crosslinked agarose resin bead Fig 3. Capto Q extender and ligand structure. Prescreening Two-step elution Table 1. Results summary for two-step elution Serotype UV 260/280 AUC (% full capsids) Peak 1 E Peak 2 F Start Peak 1 E Peak 2 F AAV5 0.72 1.21 16.6 7.7 74.5 AAV8 0.61 1.26 1.4* 1.6 76.0 AAV9 0.59 1.25 35.7 2.3 87.7 AUC results for AAV5, AAV8, and AAV9 * No reliable data, % full expected to be > 15% from chromatogram peak areas. Sample related error (mix-up or sample prep) 0 20 40 60 80 100 120 AAV5 start AAV5 peak 1 AAV5 peak 2 AAV8 start AAV8 peak 1 AAV8 peak 2 AAV9 start AAV9 peak 1 AAV9 peak 2 % % full % empty % partials Fig 8. Comparison of Capto Q vs Capto Q ImpRes separation performance. Full AAV5 capsids purity obtained by analytical ultracentrifugation (AUC) of peak 2. Fig 7. UV 260:280 ratios and analytical ultracentrifugation (AUC) results of Capto Q peaks 1 and 2. Fig 5. Prescreening to identify optimal elution conductivity (%B) for Capto Q. Dotted line indicates selected %B buffer for elution of empty capsids (step 1). Fig 6. Final two-step elution protocol using optimal elution conductivity (%B, 20 CV) for empty capsids (step 1) and step 2 (5 CV) to elute the full capsids using Capto Q. Experimental Column:Capto Q resin packed in Tricorn 5/100, 2 mL Sample load:~ 1 × 1012VP/mL resin Flow rate:2 mL/min System:ÄKTA pure 25 Buffer A:20 mM BTP, pH 9.0, 2 mM MgCl2 Buffer B:20 mM BTP, pH 9.0, 2 mM MgCl2, 250 mM Na acetate Equilibration:Buffer A, 5 CV Wash:Buffer A, 5 CV Prescreening protocol Gradient:Step elution, 5% increments, 3 CV each Two-step protocol Gradient:Two-step elution for each serotype: rAAV2:Step 1 40% buffer B, 20 CV Step 2 100% buffer B, 5 CV rAAV5: Step 1 35% buffer B, 20 CV Step 2 100% buffer B, 5 CV rAAV8: Step 1 30% buffer B, 20 CV Step 2 100% buffer B, 5 CV rAAV9: Step 1 5% buffer B, 20 CV Step 2 30% buffer B, 5 CV We acknowledge Akash Bhattacharya and Shawn Sternisha at Beckman Coulter Life Sciences for collaboration and for kindly performing the analytical ultracentrifugation. Acknowledgements NA = not analyzed Learn more about our end-to-end AAV