Method for production of aav

Abstract

To provide a scalable system for the stable production of AAV vectors.SOLUTION: The present invention relates to methods for the production of Adeno-associated virus (AAV), comprising steps of providing a stable AAV producer cell line in which at least some or all genes encoding the components necessary for the production of AAV are stably integrated into the cell genome, and culturing said cells in perfusion culture during the AAV production step, wherein the perfusion culture encompasses continuous replacement of spent media with fresh media, and wherein the continuous replacement of spent media with fresh media continues after the induction of AAV production.SELECTED DRAWING: None

Description

[Technical field]

[0001] The present invention relates to a method for producing adeno-associated virus (AAV), comprising the steps of: providing a stable AAV producer host cell line, wherein at least some or all of the genes encoding components necessary for the production of AAV are stably integrated into the host cell genome; and culturing said cells in perfusion culture during the AAV production step (i.e., during the N step), wherein said perfusion culture involves continuous replacement of spent medium with fresh medium, and wherein said continuous replacement of spent medium with fresh medium follows induction of AAV production. [Background technology]

[0002] Recently, the number of gene therapy trials and products based on AAV-derived vectors has been rapidly increasing.The advantages of AAV vectors in gene therapy are their excellent safety profile, the fact that such vectors are not pathogenic, i.e., are not associated with any disease, the long-term expression of transgenes, and the possibility of transducing dividing and non-dividing cells.However, the main challenge to move this promising research into clinical development is the challenge of delivering a significant amount of AAV viral vectors with high quality.

[0003] The production of recombinant AAV requires in particular the expression of AAV Rep and Cap proteins, which are usually encoded in the AAV genome, for the production of recombinant virus supplied in trans. In addition, helper genes are used which can be derived from different helper viruses, the most common being the helper virus genes obtained from adenovirus (AV), such as E1A, E1B, E2A, E4orf6 or VA RNA. In addition, a transfer vector is required which contains the gene of interest (GOI) flanked by AAV inverted terminal repeats (ITRs).

[0004] Current production systems for AAV mostly rely on the following technologies, but have several drawbacks.

[0005] The most common system for producing recombinant AAV relies on the introduction of all genes required for AAV production into producer cells by transient transfection. Transient transfection usually requires a two or three plasmid system: a transfer vector containing the gene of interest; pHelper with adenovirus helper functions; and pAAV-Rep2CapX (CapX = capsid function of different AAV serotypes) that provides capsid and replicase functions. However, this approach has several drawbacks. In particular, due to the transient transfection step, the method requires expensive plasmid DNA and lacks sufficient scalability, robustness and reproducibility.

[0006] The step of transient transfection also hampers certain process optimization steps to increase product titer or product quality that are common in the production of other recombinant biological products, such as extensive medium and feed optimization, and variations during production of process parameters such as pH, temperature or cell density. In particular, the latter is a promising approach to increase overall product titer. Higher cell density generally leads to higher titer, unless cell-specific productivity is reduced ("cell density effect"). In particular, while using transient transfection, increasing cell density may be accompanied by a concomitant decrease in cell-specific productivity. This has also been described, in particular, for the production of viral vectors.

[0007] Production of AAV, independent of the specific production process, always results in a mixture of complete AAV particles, representing the packaged gene of interest (GOI) flanked by AAV ITRs, and empty particles, which do not contain a packaged GOI. Titers of AAV achieved by transient transfection of mammalian suspension cells, such as HEK293 cells, can range from 1 × 10 to 1 × 10, depending on the process. 13 vg / L (viral genome / L) ~ 1 × 10 15vg / L. With transient transfection, the percentage of complete particles has been reported up to about 30% complete particles, but typically ranges from 10% to 18%. This means that 82% to 90% are undesirable empty particles. Empty particles are therapeutically inactive and may enhance the immune response of the patient. Thus, in the post-production purification process, these empty particles must be removed as efficiently as possible. This can be achieved either via centrifugation, an effective but not scalable process, or by chromatographic methods. The latter is also available for large-scale production, but the fold enrichment of complete particles is usually only in the range of about 3-fold to 6-fold. Thus, for example, enrichment of a sample with approximately 10% complete particles results in a final percentage of only about 30% to 60% complete particles.

[0008] So-called producer cell lines, which already have the gene of interest stably integrated, are mostly based on HEK or HeLa cells. However, they require further infection with a helper virus, e.g. adenovirus. This addition of helper virus during AAV production requires, after the first production of the corresponding helper virus, extensive purification of the produced AAV vector to remove the helper virus from the final product, and to prove the absence of the helper virus at a cost. The same applies to herpes simplex virus (HSV)-based systems. Moreover, baculovirus-based systems lack sufficient scalability and robustness and add the risk of contamination with insect host cell proteins and immunogenic insect cell-specific glycosylation structures.

[0009] Thus, current AAV production systems are limited not only in terms of scalability, robustness, reproducibility, ease of use and cost-effectiveness, but also in terms of product quality. Summary of the Invention [Problem to be solved by the invention]

[0010] The technical problem underlying the present invention is therefore to provide a scalable system for stable production of AAV vectors that allows for robust, industrial and scalable production of AAV gene therapy vectors without the need for transient transfection or helper viruses, as well as optimization of large-scale upstream process development, in particular with the aim of increasing product quality in terms of product titer and / or cell-specific yield and even more importantly, in terms of the proportion of intact particles. [Means for solving the problem]

[0011] The solution to the above technical problem is achieved by the embodiments characterized in the claims.

[0012] In particular, in a first aspect, the present invention provides a method for producing a method for treating a cancer cell comprising the steps of: (a) providing a stable AAV producer host cell line, wherein at least some or all of the genes encoding components necessary for the production of AAV are stably integrated into the host cell genome; (b) culturing the cells in perfusion culture during the AAV production step (i.e., during the N step), wherein the perfusion culture comprises continuous replacement of spent medium with fresh medium, and wherein the continuous replacement of spent medium with fresh medium follows induction of AAV production; The present invention relates to a method for producing adeno-associated virus (AAV).

[0013] The AAV of the present invention is not limited to a specific AAV serotype.Accordingly, the AAV can be selected from the group consisting of AAV serotype 1 (AAV1), AAV2, AAV3, AAV4, AAV5, AAV6, AAV7, AAV8, AAV9, AAV10, AAV11, AAVDJ, AAVDJ8, AAVrhlO, a mixture of two or more different above serotypes, and the above serotypes with a mutation that changes the tropism of the AAV serotype.Preferably, the AAV is selected from the group consisting of AAV2, AAV5, AAV6, AAV8 and AAV9, where AAV8 is particularly preferred.

[0014] According to the invention, at least some or all of the genes encoding components necessary for AAV production are stably integrated into the host cell genome. Preferably, at least two, at least three, at least four, at least five, at least six or all of said genes are stably integrated into the host cell genome. More preferably, all of said genes are stably integrated into the host cell genome.

[0015] In this context, according to the present invention, the genes encoding the components necessary for AAV production are selected from the group consisting of the genes encoding the AAV Cap proteins VP1, VP2 and VP3, the genes encoding the AAV Rep proteins Rep78, Rep68, Rep52 and Rep40, the genes encoding the adenoviral helper functions E4orf6, E2A and VA-RNA, the genes encoding the Ad5 helper genes E1A and E1B, and the gene of interest (GOI) flanked by AAV ITRs. Preferably, the genes encoding the components necessary for AAV production include the genes encoding the AAV Cap proteins VP1, VP2 and VP3, the genes encoding the AAV Rep proteins Rep78 or Rep68, the genes encoding the AAV Rep proteins Rep52 or Rep40, the genes encoding the adenoviral helper function E4orf6, and the gene of interest (GOI) flanked by AAV ITRs. In certain embodiments, the genes encoding components necessary for AAV production further include at least one, at least two, at least three, at least four, at least five, or all of the genes encoding the AAV Rep proteins Rep78, Rep68, Rep52, and Rep40; and genes encoding the adenoviral helper functions E2A and VA-RNA; and genes encoding the Ad5 helper genes E1A and E1B.

[0016] Thus, in certain embodiments, the stable AAV producer host cell line provided in step (a) of the method of the invention comprises the following genes: A gene encoding the AAV Rep protein Rep78 or Rep68, A gene encoding the AAV Rep protein Rep52 or Rep40, and genes encoding the adenoviral helper functions E4orf6 and E2A, is stably integrated into the host cell genome.

[0017] In other particular embodiments, the stable AAV producer host cell line provided in step (a) of the method of the invention comprises the following genes: Genes encoding the AAV Cap proteins VP1, VP2, and VP3, A gene encoding the AAV Rep protein Rep78 or Rep68, A gene encoding the AAV Rep protein Rep52 or Rep40, Genes encoding the adenoviral helper functions E4orf6 and E2A, and A gene of interest (GOI) flanked by AAV ITRs; is stably integrated into the host cell genome.

[0018] In a further particular embodiment, the stable AAV producer host cell line provided in step (a) of the method of the invention comprises the gene(s) Genes encoding the AAV Cap proteins VP1, VP2, and VP3, A gene or genes encoding either or both of the AAV Rep proteins Rep78 or Rep68; A gene or genes encoding either or both of the AAV Rep proteins Rep52 or Rep40; A gene of interest (GOI) flanked by AAV ITRs, and Optionally, a gene or genes encoding any one, any two, any three or all of the adenoviral helper functions E1A, E1B, E4orf6 and E2A, is stably integrated into the host cell genome.

[0019] In a related specific embodiment, the stable AAV producer host cell line provided in step (a) of the method of the invention comprises the following genes: (i) a gene encoding the AAV Cap proteins VP1, VP2, and VP3; a gene encoding the AAV Rep proteins Rep78 or Rep68; a gene encoding the AAV Rep proteins Rep52 or Rep40; and a gene of interest (GOI) flanked by AAV ITRs, or (ii) a gene encoding the AAV Cap proteins VP1, VP2, and VP3; a gene encoding the AAV Rep proteins Rep78 or Rep68; a gene encoding the AAV Rep proteins Rep52 or Rep40; a gene encoding the adenoviral helper function E4orf6; and a gene of interest (GOI) flanked by the AAV ITRs, or (iii) a gene encoding the AAV Cap proteins VP1, VP2, and VP3; a gene encoding the AAV Rep proteins Rep78 or Rep68; a gene encoding the AAV Rep proteins Rep52 or Rep40; a gene encoding the adenoviral helper functions E4orf6 and E2A; and a gene of interest (GOI) flanked by the AAV ITRs, or (iv) a gene encoding the AAV Cap proteins VP1, VP2, and VP3; a gene encoding the AAV Rep proteins Rep78 or Rep68; a gene encoding the AAV Rep proteins Rep52 or Rep40; a gene encoding the adenoviral helper functions E1A, E1B, E4orf6, and E2A; and a gene of interest (GOI) flanked by the AAV ITRs, or (v) a gene encoding the AAV Cap proteins VP1, VP2, and VP3; a gene encoding the AAV Rep proteins Rep78, Rep68, Rep52, and Rep40; a gene encoding the adenoviral helper functions E4orf6 and E2A; and a gene of interest (GOI) flanked by the AAV ITRs, or (vi) genes encoding the AAV Cap proteins VP1, VP2, and VP3; genes encoding the AAV Rep proteins Rep78, Rep68, Rep52, and Rep40; genes encoding the adenoviral helper functions E1A, E1B, E4orf6, and E2A; and a gene of interest (GOI) flanked by the AAV ITRs, is stably integrated into the host cell genome.

[0020] In one particular embodiment, the method of producing AAV according to the invention comprises: (a) providing a stable AAV producer host cell line, wherein at least some or all of the genes encoding components necessary for the production of AAV are stably integrated into the host cell genome; (b) culturing the cells in perfusion culture during the AAV production step (i.e., during the N step), wherein the perfusion culture comprises continuous replacement of spent medium with fresh medium, and wherein the continuous replacement of spent medium with fresh medium follows induction of AAV production; where at least the following genes: Genes encoding the AAV Cap proteins VP1, VP2, and VP3, A gene encoding the AAV Rep protein Rep78 or Rep68, A gene encoding the AAV Rep protein Rep52 or Rep40, Genes encoding the adenoviral helper functions E4orf6 and E2A, and A gene of interest (GOI) flanked by AAV ITRs; is stably integrated into the host cell genome.

[0021] In this particular embodiment, the following additional genes: A gene encoding both AAV Rep proteins Rep78 and Rep68, A gene encoding both AAV Rep proteins Rep52 and Rep40, and A gene or genes encoding either or both of the adenoviral helper functions E1A and E1B, can be stably integrated into the host cell genome.

[0022] The GOI used in the context of the present invention is not particularly limited, and examples thereof include eye diseases, blindness diseases, muscle diseases, Duchenne muscular dystrophy, GM2 gangliosidosis and spinocerebellar ataxia, ALS, Huntington's disease, X-linked severe combined immunodeficiency (X-SCID), adenosine deaminase deficiency (ADA-SCID), central nervous system diseases, Parkinson's disease, Alzheimer's disease, liver diseases, liver enzyme ornithine transcarbamylase (OTC) deficiency, Leber's congenital amaurosis, hemophilia, β- and any gene whose transfer into a recipient of the AAV vector is of interest for the treatment of thalassemia, cancer disease, head and neck cancer, metastatic melanoma, cardiac disease, pulmonary disease or rare diseases including cystic fibrosis, Wiskott-Aldrich syndrome (WAS), metachromatic leukodystrophy (MLD) and severe lipoprotein lipase deficiency disorder (LPLD), infectious diseases, severe combined immunodeficiency syndrome, HIV infection, Niemann-Pick disease type C and ornithine transcarbamylase (OTC) deficiency.

[0023] The stable AAV producer host cell line used in the present invention may be derived from any suitable cell line known in the art. However, in a preferred embodiment, said cell line is derived from or is a cell line selected from the group consisting of CAP cells, AGE1.hn cells, HEK293 cells, PER.C6 cells, NSOL cells, COS-7 cells, BHK cells, CHO cells, CV1 cells, VERO cells, HeLa cells, MDCK cells, BRL3A cells, W138 cells and HepG2 cells, where CAP cells and HEK293 cells are particularly preferred. In a particular embodiment, said cell line is derived from CAP cells or is a CAP cell line. In another particular embodiment, said cell line is derived from HEK293 cells or is a HEK293 cell line. In this context, the term "stable AAV producer host cell line is derived from cell line X" indicates that said stable AAV producer host cell line is generated by stably integrating the respective genes as indicated above into cells of cell line X. Thus, the above terms expressly encompass that the stable AAV producer host cell line is cell line X, as discussed above. In this context, the term "the stable AAV producer host cell line is derived from cell line X" can be read as "the stable AAV producer host cell line is based on cell line X," and each term can be used interchangeably herein.

[0024] The scale of AAV production according to the present invention can vary from small to over moderate volumes in the milliliter range to large scale volumes using volumes of, for example, 2000 L or 25000 L per production run.

[0025] The perfusion culture of step (b) of the method of the present invention can be carried out in any suitable culture vessel or system known in the art. Different types of bioreactors that can be used in this regard include stirred tank bioreactors (STR), orbital shaker bioreactors (OSB), rocking bed bioreactors (e.g., WAVE, Rocking Motion), airlift bioreactors or tubular bioreactors. For any of these bioreactor types, glass, stainless steel or disposable bioreactor vessels can be used. Also, reasonable multi-parallel bioreactors can be used, including high throughput bioreactors coupled to perfusion devices. Perfusion bioreactors designed for adherent cells can also be used, such as hollow fiber bioreactors (HFBR), fixed bed bioreactors (e.g., packed bed bioreactors) and fluidized bed bioreactors.

[0026] The operation modes of perfusion culture include, but are not limited to, standard perfusion (continuous), concentrated perfusion (continuous), hybrid perfusion / fed-batch, volume-scaled perfusion and two-stage continuous processes, as known in the art.Perfusion devices include hollow fiber filters, floating membranes, spin filters and rotating tubular filters (also known as vortex filters or external spin filters) and rotating disk filters (also known as controlled-shear filters) operated under strategies such as TFF (tangential flow filtration) and ATF (alternating tangential flow), where ATF perfusion culture is particularly preferred.Other types of filtration devices, such as hollow-based or tubular membrane modules operated with peristaltic pumps, diaphragm pumps or centrifugal pumps based on magnetic levitation (e.g., Levitronix), can also be used. Still further perfusion devices include gravity settlers (lamella settlers, compact cell settlers), acoustic settlers (also known as acoustic filters), centrifuges and hydrocyclones, as known in the art.

[0027] According to the present invention, perfusion culture includes continuous replacement of spent medium with fresh medium (i.e., continuous medium exchange), where said continuous replacement of spent medium with fresh medium continues after induction of AAV production. In a preferred embodiment, said continuous replacement of spent medium with fresh medium continues for at least 24 hours after induction, at least 48 hours after induction, at least 72 hours after induction, and / or until harvest of AAV. As known in the art, the term "continuous replacement of spent medium with fresh medium" or "continuous medium exchange" does not necessarily require uninterrupted replacement of spent medium with fresh medium (i.e., uninterrupted medium exchange), and said replacement can be paused at intervals several times a day, as known in the art. By way of example, possible pauses in replacement or medium exchange are, for example, pauses of up to 1 hour or up to 2 hours once, twice, three times, four times, five times or six times a day, or pauses of up to 6 hours or up to 12 hours once or twice a day.

[0028] In a preferred embodiment, the cell specific perfusion rate (CSPR) during perfusion culture is about 0.01 nL / cell / day to about 0.20 nL / cell / day. Furthermore, the perfusion rate is preferably about 1 vvd (volume of fresh medium / working volume of reactor / day) to about 20 vvd.

[0029] According to the present invention, the method comprises: 6 cells / mL ~ approx. 2.0×10 6 Low viable cell density (VCD) in the range of cells / mL or approximately 2.0 x 10 6 cells / mL ~ approx. 20×10 6 One can start with a high inoculation cell density in the range of cells / mL of VCD. In a preferred embodiment, the inoculation cell density for perfusion culture is about 0.5×10 6 cells / mL ~ approx. 5×10 6 In cells / mL.

[0030] Furthermore, the cell density at the time of induction of AAV production was approximately 1.0 × 10 6 cells / mL ~ approx. 10×10 6Low VCD in the range of approximately 10 x 10 cells / mL 6 cells / mL ~ approx. 100×10 6 High VCD range of approximately 100 x 10 cells / mL 6 cells / mL ~ approx. 200×10 6 The VCD can be very high, in the range of cells / mL. In a preferred embodiment, the cell density at the time of induction of AAV production is about 20×10 6 cells / mL or more.

[0031] In certain embodiments, the methods of the invention can include a step of recovering AAV from the retentate of the perfusion culture.

[0032] In the methods of the present invention, the process time from post-inoculation to induction of AAV production can range from 0 to 21 days. Furthermore, the process time after induction can range from 2 to 10 days after induction.

[0033] To achieve maximum titer of AAV, process parameters such as seeding cell density, temperature, agitation rate, pH, DO (dissolved oxygen), osmolality and / or bioreactor work rate can be varied during the production process, as is known in the art. Also, as is known in the art, supplementation of the cell culture with missing components can be part of the production process to avoid deficiencies of certain amino acids, saccharides, organic acids, cofactors, vitamins, minerals and / or other elements.

[0034] In a second aspect, the present invention relates to an adeno-associated virus (AAV) obtainable by the method according to the invention as defined above.

[0035] As used herein, the term "comprising" / "comprises" expressly includes the terms "consisting essentially of" / "consists essentially of" and "consisting of", i.e., all of the above terms are interchangeable with each other.

[0036] As used herein, the term "about" refers to a variation of ±10%, preferably ±7%, ±5%, ±4%, ±3%, ±2%, ±1% or ±0.5% of a particular value. As an example, the term "about 10" as used herein can refer to a range of 9 to 11.

[0037] In the present invention, fully stable AAV producer cell lines represent an effective method for large-scale production of AAV vectors. These fully stable AAV producer cells consist of cells with stable integration of, for example, capsid genes (VP1, VP2, VP3), regulatory genes (Rep78, Rep68, Rep52, Rep40), adenoviral helper functions (E1A, E1B, E4orf6, E2A and VA-RNA) and ITR-flanked transgenes (gene of interest; GOI). In this system, the expression of the four rep proteins and the helper functions E2A and E4orf6 is under the control of an inducible system. Addition of an inducer results in the expression of the aforementioned proteins and the production of recombinant AAV (rAAV) vectors.

[0038] This stable system has several advantages since it does not require a transient transfection or helper infection step, thus reducing the cost of goods since the additional costs for plasmid cost or production of helper virus are eliminated, the system is scalable and highly reproducible with little or no batch-to-batch variation.

[0039] Most importantly, the stable AAV-producing cell lines used in the present invention, which do not require transient transfection or viral infection, also enable the development and optimization of large-scale upstream cell culture processes, e.g., to increase viral titer yield (volumetric productivity and / or cell-specific viral yield) and / or to increase product quality, especially in terms of the proportion of intact particles.

[0040] In general, upstream cell culture processes can be operated in batch, fed-batch, continuous, or continuous with cell recirculation (perfusion) mode, or in a hybrid mode (e.g., a hybrid of perfusion and fed-batch). The mode of operation determines the growth profile and production kinetics.

[0041] In perfusion mode, the supply of fresh medium (inlet) and removal of spent medium (outlet) are usually performed using the same flow rate at the inlet and outlet, so that the bioreactor volume remains constant. Cells are retained in the production system (bioreactor) using a cell retention (or perfusion) device. Thus, high cell densities can be obtained. Typically, perfusion allows higher volumetric productivity with smaller footprints compared to other operation modes.

[0042] Different perfusion devices can be used, using different mechanisms for cell separation: sedimentation (under gravity or centrifugal field), flocculation (followed by sedimentation) under an ultrasound field, and filtration. Thus, the separation efficiency depends on the choice of device. Examples are centrifuges, gravity settlers (lamella settlers, small cell settlers), acoustic settlers (also known as acoustic filters), hydrocyclones, and filters. Filtration is operated under different strategies: TFF (tangential flow filtration) and ATF (alternating tangential flow), floating membranes, spin filters, and rotating tubular filters (also known as vortex filters or external spin filters) and rotating disk filters (also known as controlled shear filters). Among the different types of filters, TFF and ATF are commonly used for perfusion. In both cases, hollow fiber modules with filter membranes are used. In the case of ATF, bidirectional flow is obtained via a diaphragm pump. At present, ATF is the cell retention device mostly used for recombinant protein production (biologics). Although perfusion is well established for biopharmaceutical manufacturing processes, it is not yet available for the scalable production of viral vectors.

[0043] In summary, the perfusion process of rAAV production according to the present invention enabled higher titers (vg / L) and higher cell-specific virus yields (vg / cell) compared to the conventional batch process using stable AAV producer cells. Unexpectedly, the application of perfusion in a stable AAV production system dramatically increased the ratio of complete virus particles to empty virus particles compared to the conventional batch process. Also, the recovery of AAV viral vectors could be performed from the retentate (the whole cell suspension collected from the bioreactor in bulk) instead of the permeate, which dramatically reduced the amount of subsequent AAV purification process. This makes the overall process more efficient.

[0044] The use of the fully stable AAV producer platform described in the present invention results in a similar percentage of complete particles compared to a completely transient approach. However, this only applies to the production of AAV particles in a batch or fed-batch approach. Unexpectedly, the production of AAV viral vectors using a fully stable producer cell system that utilizes perfusion advantageously dramatically increases the percentage of complete particles by up to about 10-fold compared to a reference batch process. Thus, in the subsequent complete particle concentration step, for example by utilizing AEX chromatography, a final percentage of complete particles of 75% or more can be achieved. [Brief description of the drawings]

[0045] [Figure 1] Graph showing growth of stable CAP AAV producers in shake flasks (SF) in simulated perfusion mode. A) Cell growth and viability of cultures in shake flasks (SF) 1 and 2 (biological replicates) up to high cell density. B) Cell-specific perfusion rate (CSPR) calculated using the average VCD of the biological replicates. [Diagram 2] Schematic of an ATF-based perfusion process for rAAV production. Continuous addition of fresh medium (feed) and removal of spent medium (permeate) occurs. ATF (alternating tangential flow) is shown as the perfusion device. [Diagram 3] Graph showing the ATF-based perfusion process for rAAV production using stable CAP AAV producer cells. Viable cell density (VCD, solid line) and viability (dashed line) of the CAP AAV producer cell line in the perfusion bioreactor. Three independent cultures: Runs 1 (◆), 2 (◇), and 3 (●). Perfusion mode starts on day 3 post-inoculation (gray vertical line). rAAV production starts on day 7 post-inoculation (vertical dashed line). The production phase runs until day 5 post-induction. [Figure 4]Graph showing a conventional batch process for rAAV production using stable CAP AAV producer cells. Viable cell density (VCD, solid line) and viability (dashed line) of the CAP AAV producer cell line in a bioreactor in batch mode. Two independent cultures: Runs 1 (closed squares) and 2 (open squares). rAAV production begins 3 days after inoculation (vertical dashed line). [Diagram 5] Graph showing rAAV production in perfusion mode compared to batch process using stable CAP AAV producer cells. Three runs in perfusion mode and two runs in batch mode were performed using the CAP AAV producer cell line. A) AAV8 genome titer determined by qPCR. B) Capsid concentration determined by ELISA. Data is normalized to the average of two batch processes at day 5 post-induction. [Figure 6] Graph showing higher percentage of total capsids in perfusion mode using stable CAP AAV producer cells. Comparison with batch reference process. Percentage was calculated as the ratio of genome titer (FIG. 5A) to capsid titer (FIG. 5B). Data is normalized to the batch run. A) Time course of the process in perfusion mode (n=3) and batch mode (n=2). n represents the number of biological replicates. B) Total capsid data is expressed as the mean ± standard deviation of biological replicates. [Figure 7] Figure 1 shows an increase in cell specific virus yield (CSVY) in perfusion mode using stable CAP AAV producer cells. Comparison to batch reference process. Yield was calculated as the ratio of genome titer at harvest to viable cell density at induction. Data are presented as mean ± standard deviation of biological replicates (perfusion: n=3, batch: n=2). Data are normalized to batch on day 5 post-induction. [Figure 8] Graph showing infectivity of CAP-derived AAV particles during perfusion and batch production. Data is normalized to day 3 post-induction for each production run. [Figure 9]Graph showing that interruption of the ATF-based perfusion process during the AAV production phase results in a decrease in viable cell density and cell viability after induction. Viable cell density (VCD, solid line) and viability (dashed line). Perfusion mode starts on day 3 after inoculation (gray vertical line). rAAV production starts on day 8 after inoculation (vertical dashed line). The production phase is performed in batch mode until day 4 after induction. [Figure 10] Graph showing that interruption of the ATF-based perfusion process during the AAV production phase results in a dramatic drop in AAV titer. AAV8 genomic titer determined by qPCR. Titer of high cell density (HCD) batch measured at harvest. Data are normalized to conventional low cell density (LCD) batch at day 5 post-induction. [Figure 11] Graph showing the ATF-based perfusion process for stable rAAV production using stable HEK293 AAV producer pool cells. Viable cell density (VCD, solid line) and viability (dashed line) of stable HEK293 AAV producer pool in perfusion bioreactor (◆). Perfusion begins 3 days after inoculation (gray vertical line). rAAV production begins 6 days after inoculation (vertical dashed line). The production phase continues until day 4 after induction. [Figure 12] Graph showing stable rAAV production using stable HEK293 AAV producer pool cells in batch mode. Viable cell density (VCD, solid line) and viability (dashed line) of stable HEK293 AAV producer pool in batch shake flasks (□). Production in batch mode begins 3 days post inoculation (vertical dashed line) and continues until day 5 post induction. [Figure 13] Graph showing rAAV production in perfusion mode using stable HEK293 AAV producer pool cells. One run in perfusion mode and one run in batch mode. A) AAV8 genome titer determined by qPCR. B) Capsid concentration determined by ELISA. Data is normalized to day 3 post-induction of the batch control. [Figure 14]Figure 1 shows an increase in cell-specific virus yield (CSVY) in perfusion mode using stable HEK293 AAV producer pool cells. Yields were calculated as the ratio of peak genome titer to viable cell density at induction. Results are normalized to batch controls. [Figure 15] 13A-B are graphs showing a higher percentage of total capsids in perfusion mode using stable HEK293 AAV producer pool cells. The percentage was calculated as the ratio of normalized genome titer (FIG. 13A) to normalized capsid titer (FIG. 13B). Data is normalized to batch control. [Figure 16] Photograph showing SyproRuby staining of AAV particles from various fractions of an AEX run reveals a VP1 to VP2 to VP3 ratio of approximately 1:1:10. Capsids from various fractions of two different AEX runs were visualized by SyproRuby staining. [Figure 17] Graph showing complete to empty ratios after AEX purification. After determining viral particle titers by ELISA and viral genome titers by qPCR, the complete to empty particle ratios of various fractions derived from two AEX runs from AAV production utilizing ATF perfusion were calculated. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0046] The present invention relates to the following nucleotide sequences: SEQ ID NO:1 SV40 polyA forward primer AGCAATAGCATCACAAATTTCACAA SEQ ID NO:2 SV40 polyA reverse primer CCAGACATGATAAGATACATTGATGAGTT SEQ ID NO:3 SV40 polyA probe AGCATTTTTTTCACTGCATTCTAGTTGTGGTTTGTC SEQ ID NO:4 GFP forward primer TTCTTCAAGTCCGCCATGCC SEQ ID NO:5 GFP reverse primer AAGTCGATGCCCTTCAGCTC SEQ ID NO:6 GFP probe CGCACCATCTTCTTCAAGGACGACGGCAACTACA

[0047] The present invention is further illustrated by the following non-limiting examples. EXAMPLES

[0048] Experimental procedure Producer cell line: Stable AAV producer cell lines that possess all the components necessary for AAV production were selected for the development of the perfusion process. Two cell lines were used: (1) a cell line based on CAP cells and (2) a cell line based on HEK293 cells. (1) A stable CAP AAV producer single cell clone containing the serotype-specific capsid (AAV8) inducers Tet3G, inducible replicase, inducible helper (E2A, E4orf6), VA-RNA, VP1, VP2 and VP3 and sequences for the ITR-flanked gene of interest (GOI) GFP. Also, the Ad5 helper genes E1A and E1B have already been stably integrated into the CAP cells, such that the CAP cells were already immortalized by them. (2) A stable HEK293 AAV producer pool containing the inducers Tet3G, inducible replicase, inducible helper (E2A, E4orf6), VA-RNA, VP1, VP2 and VP3 of a serotype-specific capsid (AAV8) and sequences for the ITR-flanked gene of interest (GOI) GFP-luciferase, and the Ad5 helper genes E1A and E1B already stably integrated into HEK cells, such that the HEK cells were already immortalized with these.

[0049] Cell culture: (1) For routine culture, CAP cells were cultured in Protein Expression Medium (PEM) (Life Technologies / Gibco) supplemented with 4 mM GlutaMAX (Gibco) in a non-baffled shake flask at 37 °C, 5% CO2, and 120 rpm to 185 rpm in a shaking incubator. Every 3 to 4 days, 0.5 × 10 6 cells / mL~1×10 6 Cells were routinely passaged with fresh medium to achieve a viable cell density of 1000 cells / mL. (2) For routine culture, HEK293 cells were cultured in Protein Expression Medium (PEM) (Life Technologies / Gibco) supplemented with 4 mM GlutaMAX (Gibco) in a non-baffled shake flask at 37 °C, 5% CO2, and 170 rpm to 200 rpm in a shaking incubator. Every 2 to 3 days, 0.3 × 10 6 cells / mL~0.5×10 6 Cells were routinely passaged with fresh medium to achieve a viable cell density of 1000 cells / mL.

[0050] Production of AAV using a stable AAV production platform: AAV production was induced by the addition of doxycycline to the cultures.

[0051] qPCR to determine viral titer: Viral titers were measured using the following primer / dual-labeled probe combinations (MWG, Eurofins; Table 1 ) against SV40 polyA or the GOI (GFP).

[0052] [Table 1]

[0053] As a standard, a linear transgene plasmid with a defined copy number was used. The qPCR reactions contained the following components: 2x Brilliant Multiplex qPCR Master Mix (Agilent), nuclease-free H2O (Thermo Fisher Scientific), primer / probe mix and samples / standards. qPCR was performed on an Agilent Mx3005P according to the manufacturer's instructions.

[0054] ELISA for assembled capsids: ELISA specific for assembled capsids was performed using a commercially available ELISA kit (Progen) according to the manufacturer's instructions.

[0055] SDS-PAGE and SyproRuby staining: To determine the ratio of capsid proteins VP1, VP2 and VP3, capsid proteins were separated by SDS PAGE and visualized by SyproRuby staining. Briefly, samples were mixed with 4x LDS buffer and then incubated at 70°C for 10 min in a dry heat block. The tubes were spun down and 10 μL of sample was loaded into the wells. 4%-12% Bis-Tris gels were run in MES SDS buffer at 200V for 40 min. SyproRuby staining was performed by overnight incubation in staining solution followed by a 30 min washing step in 10% methanol, 7% acetic acid and a final wash in ddH2O.

[0056] Separation of intact and empty AAV particles by anion exchange chromatography: AAV particles from whole cell suspension lysates were concentrated by chromatography. Total-empty separation was achieved by utilizing an anion exchange chromatography column.

[0057] Example 1: Simulated perfusion experiment (cell proliferation) With most cell culture media, higher cell densities require faster perfusion rates to maintain cell viability and sustain cell growth. As cell density changes during the production process, the cell specific perfusion rate (CSPR) is the determining factor for the perfusion rate during the entire process, based on cell density.

[0058] CSPR represents the nutrient supply per cell per day in a continuous process: CSPR=P / VCD where P = perfusion rate (1 / day) and VCD = viable cell density (1 x 10 6 cells / mL).

[0059] To determine the CSPR for the perfusion bioreactor process, stable CAP AAV producers (single cell clones) were cultured in shake flasks in PEM containing 4mM GlutaMAX in non-baffled shake flasks. The medium was changed daily from day 3 of culture. After centrifugation, the spent medium was discarded. The cell pellet was then resuspended in an equal volume of fresh medium, i.e., the perfusion rate was maintained at 1 vvd until a decrease in viability was observed.

[0060] Exponential cell growth and high viability (>90%) were observed by day 7, and these were observed in 20×10 6 This corresponded to a VCD of 100 cells / mL and a CSPR of approximately 0.05 nL / cell / day (Figure 1). These parameters were selected as the scale-up targets for the perfusion bioreactor. To sustain higher cell densities using supplemented PEM medium as the perfusion medium, a faster perfusion rate should be applied.

[0061] Example 2: AAV Production in Perfusion Mode Using Stable CAP AAV Producer Single Cells To evaluate the benefits of using ATF-based perfusion compared to a traditional batch process, AAV production was performed using stable CAP AAV producer cells either in a batch process or as an ATF-based perfusion process, both in stirred tank bioreactors under controlled conditions.

[0062] Traditional batch process: Cells were cultured in stirred tank bioreactors, e.g., disposable 10c vessels (Eppendorf), BioFlo320 systems (Eppendorf). Comparable low cell densities (0.5 × 10) in PEM medium supplemented with 4 mM GlutaMAX were used. 6 cells / mL~1×10 6 The bioreactor was inoculated with 1 μg / mL of doxycycline. The process was operated in batch mode. After the cells had grown to the target viable cell density for AAV production (typically 3 days after seeding), AAV production was initiated with 1 μg / mL doxycycline.

[0063] Perfusion configuration: The ATF-based perfusion setup consisted of a stirred tank bioreactor connected to an ATF unit. Cells were cultured in stirred tank bioreactors, e.g., disposable 3c vessels (Eppendorf), BioFlo320 systems (Eppendorf).

[0064] Perfusion-based AAV process: Comparable low cell densities (0.5 × 10) in PEM medium supplemented with 4 mM GlutaMAX 6 cells / mL~1×10 6 The bioreactor was inoculated with approximately 20 × 10 cells / mL. The process was operated in batch mode for 3 days. After that, perfusion was started to supply supplemented PEM medium according to a cell-specific perfusion rate (CSPR) of approximately 0.05 nL / cell / day to approximately 0.10 nL / cell / day. Approximately 20 × 10 cells were inoculated with the supplemented PEM medium. 6After expansion to viable cells / mL (day 7), AAV production was induced with 1 μg / mL doxycycline. During the production phase, the perfusion rate was kept fairly constant at approximately 1 vvd.

[0065] 20 x 10 in three perfusion runs 6 Robust cell growth to 100000 cells / mL was observed (Figure 3). Cell growth in the perfusion bioreactor was comparable to semi-perfusion experiments in shake flasks (Figure 1). An approximately 5-fold increase in VCD upon induction was observed compared to the batch process (Figure 4).

[0066] In perfusion mode, genome titers (vg / mL) were unusually high (Figure 5). Considering the maximum titer, a 42-fold increase was obtained with the CAP perfusion process (average of three independent cultures) compared to the reference batch process (average of two independent cultures). This is partly due to the higher VCD, but more importantly, due to a significant increase in cell-specific virus yields (vg / cell), which were on average 9-fold higher with perfusion (Figure 7).

[0067] Unexpectedly, the percentage of complete particles was also significantly increased in the perfusion process compared to the batch process (Figure 6). The maximum percentage of complete particles in the perfusion (approximately 4 d.pi) was about 10-fold higher compared to the value obtained in the reference batch process at the same time point. Comparing data from three perfusion runs with two batch runs, on average, the maximum percentage of complete AAV particles was 9.3-fold higher for the perfusion runs. Possible explanations could be that the perfusion environment contributes to better AAV vector DNA amplification or to increased packaging of the viral genome into AAV particles.

[0068] Infectious AAV particles were determined by transduction assay. The ratio of transducing units (TU) to viral genomes (VG) was calculated (TU:VG). For batch run 2, TU:VG was determined 3-6 days post-induction (dpi). The maximum value was obtained 3 dpi. For perfusion runs 1 and 2, the maximum value was also obtained 3 dpi. Therefore, for each run, the ratio of TU:VG for each sample was normalized to 3 dpi.

[0069] The results (Figure 8) suggest that higher infectivity was maintained over time in perfusion mode compared to batch: between days 3 and 5 post-induction, approximately 30% of infectious AAV particles lost infectivity in batch, whereas in perfusion mode (1vvd), less than 18% was lost over the same period.

[0070] Example 3: Utilizing ATF perfusion for the generation of high cell density CAP producer cell cultures To determine whether the significant increase in cell-specific productivity, volumetric titer, and percentage of complete particles was primarily due to the perfusion mode prior to induction of AAV production, the following experiment was performed. 6 Cells were cultured in N bioreactor in perfusion mode as described in Example 2 until the target viable cell density in cells / mL was reached. However, immediately after induction of AAV production by doxycycline induction, perfusion was stopped and the production phase was operated in batch mode, meaning that no further medium exchange was performed in the production phase.

[0071] Perfusion configuration: The ATF-based perfusion setup consisted of a stirred tank bioreactor connected to an ATF unit. Cells were cultured in stirred tank bioreactors, e.g., disposable 3c vessels (Eppendorf), BioFlo320 systems (Eppendorf).

[0072] Perfusion / batch AAV process: Comparable low cell densities (0.5 × 10) in PEM medium supplemented with 4 mM GlutaMAX 6 cells / mL~1×10 6 The bioreactor was inoculated with approximately 20 × 10 cells / mL. The process was operated in batch mode for 3 days. After that, perfusion was started to supply supplemented PEM medium according to a cell-specific perfusion rate (CSPR) of approximately 0.05 nL / cell / day to approximately 0.10 nL / cell / day. Approximately 20 × 10 cells were inoculated with the supplemented PEM medium. 6 After expansion to viable cells / mL, AAV production was induced with 1 μg / mL doxycycline. After induction, perfusion was stopped.

[0073] The perfusion culture followed the same growth profile as the previous run (Figure 3), with 2.0 × 10 6 After changing the operation mode to batch, the cell concentration was maintained for 24 hours after induction. Subsequently, there was a strong drop in cell concentration and viability. Interestingly, the volumetric productivity and cell-specific productivity were extremely low, and the genome titer at day 4 after induction was about 40-fold lower compared to the reference batch process (FIG. 10). This result indicates that it is not just the perfusion mode before induction, and therefore the increase in viable cell density at the time of induction, that is responsible for the surprising increase in volumetric productivity, cell-specific productivity, and the increase in the percentage of intact AAV particles mentioned above (Example 2).

[0074] The results of this example (Example 3) clearly show that only if the perfusion mode is continued further after induction and thus during the entire production phase can cell sustainability be ensured, resulting in efficient AAV production with high cell-specific virus yields and an increased percentage of intact AAV particles.

[0075] Example 4: Perfusion mode using stable HEK293 AAV producer cells for AAV production This experiment was performed to demonstrate that the beneficial effects of the perfusion process in terms of increasing cell-specific virus yield, volumetric productivity, and percentage of intact AAV particles are not limited to stable CAP AAV producer cells (Example 2) but can be transferred to other AAV production systems. Therefore, a similar perfusion process was developed for a stable HEK293 AAV producer pool. In addition, as an internal control, AAV was produced using HEK293 AAV producer pool cells cultured in shake flasks and operated in batch mode.

[0076] Batch process: Cells were cultured in 125 mL unbaffled shake flasks in a culture volume of 30 mL. A comparable low cell density (0.5 × 10) was cultured in PEM medium supplemented with 4 mM GlutaMAX. 6 cells / mL~1.0×10 6 Cultures were inoculated with 1 μg / mL of doxycycline. The process was operated in batch mode. After a 3-day cell growth period, AAV production was initiated with 1 μg / mL doxycycline.

[0077] Perfusion configuration: The ATF-based perfusion setup consisted of a stirred tank bioreactor connected to an ATF unit. Cells were cultured in stirred tank bioreactors, e.g., disposable 3c vessels (Eppendorf), BioFlo320 systems (Eppendorf).

[0078] Perfusion-based AAV process: Comparable low cell densities (0.5 × 10) in PEM medium supplemented with 4 mM GlutaMAX 6 cells / mL~1×10 6 The bioreactor was inoculated with approximately 20 × 10 cells / mL. The process was operated in batch mode for 3 days. After that, perfusion was started to supply supplemented PEM medium according to a cell-specific perfusion rate (CSPR) of approximately 0.05 nL / cell / day to approximately 0.10 nL / cell / day. Approximately 20 × 10 cells were inoculated with the supplemented PEM medium. 6After expansion to viable cells / mL (day 7), AAV production was induced with 1 μg / mL doxycycline. During the production phase, the perfusion rate was kept approximately constant at approximately 1.25 vvd.

[0079] 20×10 in perfusion with high survival rate 6 Cell proliferation to 100000000000000 cells / mL was observed before induction (Figure 11), which corresponded to approximately 6-fold higher viable cell density than at the time of induction of the batch (Figure 12).

[0080] As already seen for the stable CAP AAV producer single cells in Example 2, a huge increase in volumetric productivity of about 30-fold (Figure 13) could be achieved by culturing the stable HEK293 AAV producer pool cells in perfusion mode during the entire production phase (i.e., before and after induction of AAV production). Again, the increase in volumetric productivity was not simply due to an increase in viable cell number at the induction time point, as the increase in viable cell number at the induction time point was only approximately 6-fold. Thus, the overall titer increase was achieved by a 5-fold higher cell-specific virus yield in perfusion mode compared to batch mode (Figure 14).

[0081] As seen for the stable CAP AAV producer single cells (Example 2), the perfusion process using stable HEK293 AAV producer pool cells increases the percentage of intact AAV particles by approximately 4-fold compared to the batch process (Figure 15). The lower fold increase for the stable HEK293 perfusion process compared to the CAP perfusion process is most likely due to the fact that for the HEK293 process a polyclonal pool was utilized, whereas for the CAP perfusion process a monoclonal single cell clone was utilized.

[0082] Example 5: Enrichment of intact AAV particles from perfused material via AEX Unexpectedly, it was found that by applying ATF perfusion to the production of AAV particles using the sufficiently stable AAV producer CAP cell line, a significant approximately 9-fold increase in the ratio of intact particles could be achieved compared to the conventional batch process.

[0083] To test whether starting from an already higher percentage of intact particles, a further enrichment of intact particles could also be achieved using AEX chromatography, the following experiment was performed.

[0084] Cell suspensions from the retentate of perfusion runs with stable CAP AAV producer cells (Example 2) were lysed and AAV particles were enriched by chromatography. Pooled fractions containing AAV particles were then subjected to anion exchange chromatography. Intact AAV particles were enriched in the elution gradient in fraction 2 of perfusion AEX run 1 and in fractions 5 to 7 of perfusion AEX run 2 (Figure 16).

[0085] As shown in FIG. 16, separation of material from the perfusion process resulted in a percentage of intact particles up to 78%.

[0086] To determine the ratios of capsid proteins VP1, VP2, and VP3, various fractions from two separate AEX experiments were separated by SDS PAGE and visualized by SyproRuby staining. As shown in Figure 17, across all fractions from both experiments, the three capsid proteins showed the expected ratio of approximately 1:1:10, demonstrating that the perfusion process did not alter the composition of the AAV capsid.

[0087] In summary, this demonstrates that the process developed using ATF perfusion is an outstanding method to ensure a high percentage of intact AAV particles in the final product.

[0088] Consideration: The present invention clearly shows that the cell-specific virus yield was significantly increased in the perfusion process according to the invention compared to the batch process (Figure 7), demonstrating that the increased titer seen in the present invention is not only due to an increased viable cell density at the induction time point, but also due to an increased titer per cell.

[0089] Furthermore, Example 3 and Figures 9 and 10 show a perfusion process in which the viable cell density increased before induction of AAV production via doxycycline by the perfusion process, but the perfusion process was stopped immediately after induction. Thus, AAV production was in a high cell density batch process state, which proves that the increase in productivity was not only due to the high cell density at the time of induction, but also due to the subsequent perfusion process.

[0090] Furthermore, Example 4 and Figures 11-15 show the perfusion process according to the invention using stable HEK293 AAV producer cells, demonstrating that increased AAV productivity (volume and cell specific) and improved AAV quality through higher whole-to-empty ratios are not limited to CAP cells but are reproducible in another cell line, namely HEK293.

[0091] Thus, the present invention advantageously provides an increase in cell-specific productivity in the methods of the invention as well as an increase in whole versus empty AAV particles, both of which are unexpected and surprising. These advantageous properties are due to continued perfusion after induction and not simply due to an increase in total cell number, as shown in Figures 9 and 10.

Claims

1. (a) providing a stable AAV producer host cell line, wherein at least some or all of the genes encoding components necessary for the production of AAV are stably integrated into the host cell genome; (b) culturing the cells in perfusion culture during the AAV production step (i.e., during the N step), wherein the perfusion culture comprises continuous replacement of spent medium with fresh medium, and wherein the continuous replacement of spent medium with fresh medium continues for at least 24 hours after induction of AAV production; A method for producing an adeno-associated virus (AAV), comprising: The method, wherein the perfusion device used in the perfusion culture is a hollow fiber filter used in TFF (tangential flow filtration) mode or ATF (alternating tangential flow) mode; a floating membrane, a spin filter, a rotating tubular filter, a rotating disk filter, a centrifuge, a gravity settler, a lamellar settler, a small cell settler, an acoustic settler or a liquid cyclone.

2. At least the following genes: Genes encoding AAV Cap proteins VP1, VP2 and VP3, A gene encoding the AAV Rep protein Rep78 or Rep68, A gene encoding the AAV Rep protein Rep52 or Rep40, a gene encoding the adenoviral helper functions E4orf6 and E2A, and a gene of interest (GOI) flanked by AAV ITRs; The method of claim 1 , wherein said vector is stably integrated into said host cell genome.

3. 3. The method of claim 1 or 2, wherein the AAV is selected from the group consisting of AAV serotype 1 (AAV1), AAV2, AAV3, AAV4, AAV5, AAV6, AAV7, AAV8, AAV9, AAV10, AAV11, AAVDJ, AAVDJ8, AAVrhlO, a hybrid of two or more different said serotypes, and said serotypes having a mutation that alters the tropism of the AAV serotype.

4. 4. The method of any one of claims 1 to 3, wherein the genes encoding components necessary for AAV production are selected from the group consisting of genes encoding AAV Cap proteins VP1, VP2 and VP3; genes encoding AAV Rep proteins Rep78, Rep68, Rep52 and Rep40; genes encoding adenoviral helper functions E4orf6, E2A and VA-RNA; genes encoding Ad5 helper genes E1A and E1B; and a gene of interest (GOI) flanked by AAV inverted terminal repeats (ITRs).

5. Additional genes include: A gene encoding both AAV Rep proteins Rep78 and Rep68; A gene encoding both AAV Rep proteins Rep52 and Rep40, and one or more genes encoding either or both of the adenoviral helper functions E1A and E1B. The method according to any one of claims 1 to 4, wherein said vector is stably integrated into the host cell genome.

6. The method according to any one of claims 1 to 5, wherein the perfusion cultivation is carried out in a stirred tank bioreactor (STR), an orbital shaking bioreactor (OSB), a rocking bed bioreactor, an airlift bioreactor, a tubular bioreactor, a hollow fiber bioreactor (HFBR), a fixed bed bioreactor or a fluidized bed bioreactor.

7. The method according to any one of claims 1 to 6, wherein the perfusion device used for perfusion culture is a hollow fiber filter used in ATF mode.

8. 8. The method of any one of claims 1 to 7, wherein said continuous replacement of spent medium with fresh medium continues for at least 48 hours or at least 72 hours after induction of AAV production and / or until harvest of AAV.

9. 9. The method of any one of claims 1 to 8, wherein the cell specific perfusion rate (CSPR) is from about 0.01 nL / cell / day to about 0.20 nL / cell / day.

10. 10. The method according to any one of claims 1 to 9, wherein the perfusion rate is from about 1 vvd (volume of fresh medium / working volume of reactor / day) to about 20 vvd.

11. The cell density at the time of seeding in the perfusion culture was about 0.5×10 ６ cells / mL ~ approx. 5 x 10 ６ The method according to any one of claims 1 to 10, wherein the concentration is cells / mL.

12. The cell density at the time of induction of AAV production was approximately 20 × 10 ６ The method according to any one of claims 1 to 11, wherein the concentration of the serotonin in the blood is 100% or more than 100% cells / mL.

13. The method of any one of claims 1 to 12, further comprising a step of recovering AAV from the retentate of the perfusion culture.

Images