Original ArticleUniversal Method for the Purificationof Recombinant AAV Vectorsof Differing SerotypesShelley A. Nass,1Maryellen A. Mattingly,1Denise A. Woodcock,1Brenda L. Burnham,1Jeffrey A. Ardinger,1Shayla E. Osmond,1Amy M. Frederick,1Abraham Scaria,1Seng H. Cheng,1and Catherine R. O’Riordan11Gene Therapy, Sanofi, 49 New York Avenue, Framingham, MA 01701, USAThe generation of clinical good manufacturing practices(GMP)-grade adeno-associated virus (AAV) vectors requirespurification strategies that support the generation of vectorsof high purity, and that exhibit a good safety and efficacy pro-file. To date, most reported purification schemas are serotypedependent, requiring method development for each AAVgene therapy product. Here, we describe a platform purifica-tion process that is compatible with the purification of multipleAAV serotypes. The method generates vector preparations ofhigh purity that are enriched for capsids with full vectorgenomes, and that minimizes the fractional content of emptycapsids. The two-column purification method, a combinationof affinity and ion exchange chromatographies, is compatiblewith a range of AAV serotypes generated by either the transienttriple transfection method or the more scalable producer cellline platform. In summary, the adaptable purification methoddescribed can be used for the production of a variety of high-quality AAV vectors suitable for preclinical testing in animalmodels of diseases.INTRODUCTIONAdeno-associated virus (AAV) vector-based gene therapy has nowreported many successes. In 2012, Europe approved thefirst genetherapy product, Glybera, an AAV1-LDL vector for the treatmentof lipoprotein lipase deficiency.1Additionally, positive results havebeen reported from AAV vector-based clinical trials for the treatmentof an early childhood blindness, Leber’s congenital amaurosis(LCA2),2–5and hemophilia B.6AAV-based gene delivery vectorscomprise an AAV capsid harboring the therapeutic transgene, withcapsid selection based on tropism for the target tissue. For example,the LCA2 trial evaluated AAV2 because of its predilection for retinalpigmented epithelial (RPE) cells. AAV8 transduces human hepato-cytes and was the choice for the hemophilia B trial, while Glybera isan AAV1-based therapeutic that targets muscle;1–6consequently,large-scale good manufacturing practices (GMPs) that can supportthe production and purification of a range of AAV serotypes areessential, especially as the repertoire of new and more diverse AAVserotypes expands.7,8In support of this concept, scalable AAV pro-duction methods, compatible with a range of AAV serotypes, havebeen described;9however, purification methods are less generic andare traditionally based on the unique properties of each AAV capsid,necessitating the optimization of capsid-specific purificationmethods.10–12AAV purification methods based on affinity chromatography and,more specifically, AVB Sepharose, have increased in popularity.13,14AVB Sepharose is an affinity resin based on single-domain antibodyfragments from the family Camelidae. These antibodies have highphysical and thermal stability and excellent binding characteristics.14The ligand used in AVB Sepharose was isolated from llamas naturallyexposed to wild-type AAV; consequently, many AAV serotypes bindto this resin.15Recently, an AVB-binding epitope, residing in a sur-face-exposed region of the AAV capsid, was identified. AAV capsidsharboring this canonical epitope bound with high affinity to the AVBresin, while the substitution of this epitope into AAV capsids withpoor affinity for AVB converted the capsids into AVB“binders.”15A major disadvantage of affinity chromatography is the indiscrimi-nate purification of both empty and vector-containing particles, anexpected result based on their identical amino acid composition.Empty particles are considered a product-related impurity and areproduced at a significant level during the biosynthesis of AAVvectors. Their presence in preclinical and clinical AAV vector stocksis problematic, because they contribute to an increased level ofAAV antigen and unnecessary immune responses in animals andhumans.16Chromatographic separation methods based on chargedifferences can be useful in the separation of empty from vectorgenome-containing AAV capsids, suggesting that a subtle differencein charge between the two populations exists, facilitating their separa-tion by traditional ion exchange chromatography (IEX).17Herein, a universal purification method that combines affinity andIEX is described. The method is compatible with a range of AAV se-rotypes and production platforms, including the triple transfectionReceived 25 September 2017; accepted 19 December 2017;https://doi.org/10.1016/j.omtm.2017.12.004.Correspondence:Catherine R. O’Riordan, PhD, Gene Therapy, Sanofi, 49 NewYork Avenue, Framingham, MA 01701, USA.E-mail:catherine.o'riordan@sanofi.comMolecular Therapy: Methods & Clinical Development Vol. 9 June 2018ª2017 The Authors. 33This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
method and the more scalable producer cell line method. The utilityof the AVB Sepharose affinity resin for the purification of multipleAAV serotypes is confirmed; however, these studies suggest thatvector recovery is improved when serotype-specificaffinity resinsare employed. Moreover, serotype-specific resins are useful for thepurification of related AAV serotypes with conserved epitopes. Addi-tionally, AAV capsids that fail to bind to AVB or serotype-specificresins, such as AAVDJ, can be converted to AVB binders by epitopeswapping. Finally, the chromatographic-based purification strategydescribed here generates AAV vectors of high quality with a reducedfractional content of empty capsids and robustin vivopotency, asdemonstrated by widespread retinal transduction in the context ofan AAV5EGFPvector.RESULTSPurification of Various AAV Serotypes over AVB ResinThe AVB affinity resin was evaluated as the initial resin in the pu-rification process, and a range of AAV serotypes was assessed fortheir binding kinetics under similarflow rates, ionic conditions,and elution parameters. The AVB resin was compatible with thepurification of multiple AAV serotypes, including AAV1, AAV2,AAV5, AAV6, and AAVrh10.Tables 1,2,and3include represen-tative recoveries following AVB chromatography for theseserotypes, with AAV5 showing the best performance in terms ofrecovery (>90%). Notably, the recoveries of serotypes, such asAAV8 and AAVrh8R, trended lower with AVB chromatography,while others, such as AAV9 or AAVDJ, failed to bind to theAVB resin at any appreciable level (data not shown). The AVBligand was selected from a library created from llamas naturallyexposed to AAV; consequently, it is not specificforanyoneserotype.13Improved recoveries were achieved using affinity resinsharboring an affinity ligand that was generated specifically againstthat serotype, e.g., POROS AAV9 and AAV8 CaptureSelect Affinitymatrices. The yield of AAV9 following chromatography withPOROS AAV9 CaptureSelect Affinity matrices was approximately73%; similarly, the purification of an AAV8 vector using a POROSAAV8 CaptureSelect Affinity matrix yielded improved recoveriesof approximately 80%, as shown inTables 1,2,and3. Interestingly,AAVrh8R, which shares sequence homology with AAV9, failed tobind to the POROS AAV9 CaptureSelect yet was successfullypurified to high yield (56% recovery) with a POROS AAV8CaptureSelect affinity matrix (Tables 1,2,and3). Despite havingsignificant homology to AAV8, no improvement in vector yieldwas achieved by purifying AAVDJ using the POROS AAV8CaptureSelect affinity matrix, giving only a 1% yield (Tables 1,2,and3).8The purification of the various AAV serotypes using theaffinity resins generated vectors of high purity for all serotypesevaluated.Figure 1A represents the SDS-PAGE analysis of affin-ity-purified vectors, and the predominant bands were VP1, VP2,and VP3, with few other contaminants.Table 1. AVB Sepharose High PerformanceSerotypeVolume (mL)vg/mLTotal vg% RecoveryAAV1Load1,6251.43E112.32E14–FT1,6351.01E91.65E12<1Wash107.55.9E86.34E10<1Elution8.371.4E131.17E1450AAV2Load5004.09E112.04E14–FT5003.90E71.95E10<1Wash203.83E77.70E8<1Elution151.08E131.62E1479AAV5Load5133.93E112.0E14–FT5139.39E94.81E122.4Wash1441.38E91.99E11<1Elution247.64E121.83E1491.5AAV6Load4331.81E117.84E13–FT4331.35E105.85E127.5Wash755.51E94.13E11<1Elution6.69.76E126.44E1382rh10Load1381.15E111.59E13–FT1381.5E102.07E1213Wash226.12E91.35E11<1Elution6.61.33E128.78E1255FT,flow-through; vg, vector genomes.Table 2. POROS CaptureSelect AAV8 Affinity MatrixSerotypeVolume (mL)vg/mLTotal vg% RecoveryAAV8Load1501.15e121.73e14–FT1504.07e107.05e124.1Wash312.16e106.7e11<1Elution6.62.11e131.39e1480AAVrh8RLoad2.61.43E123.7E12–FT2.6LOD––Wash52.19E91E10–Elution1.51.38E122.1E1256AAVDJLoad433.85E111.65E13–FT433.15E111.35E1282Wash234.46E101E126Elution43.98E101.59E111FT,flow-through; vg, vector genomes.Molecular Therapy: Methods & Clinical Development34 Molecular Therapy: Methods & Clinical Development Vol. 9 June 2018