AAV vectors

A step that makes the "seed stock" of cells that is the starting point of AAV manufacture, freezes it and keeps it ready. It is stored in two tiers, master (MCB) and working (WCB), and cells are taken from the WCB for each manufacture.

After dispensing uniformly into each vial, the cryoprotectant DMSO is added at a final 10% and the cells are frozen slowly at about 1 °C per minute. This rate, which does not damage the cells, is the key, and they are finally laid to rest for a long time at −130 to −150 °C in the liquid nitrogen vapor phase.

The aim is reproducibility. Only when vials are uniform, frozen without damage and kept stable at ultra-low temperature can lots years later be started up with the same quality.

In AAV manufacturing, cells are stored in 2 tiers, an MCB (master) and a WCB (working), and cells are taken from the WCB for each manufacture. Reproducible manufacturing requires 3 things: dispensing vials uniformly, freezing slowly so the cells are not damaged, and keeping them stable for a long time at ultra-low temperature. Cells are frozen with DMSO added at a final 10%, cooled slowly at about 1 °C per minute, and stored in the vapor phase of liquid nitrogen (roughly −130 to −150 °C).

Alternatives in practice

• Passive freezing · Thermo Scientific Mr. Frosty Nalgene 5100 (isopropanol type) or CoolCell (Corning) — Small scale, small MCB quantities. Simple control at about -1 °C/min. A controlled-rate freezer is recommended for GMP batches
• Mechanical ultra-low temperature storage · -86 °C ULT freezers (Thermo Scientific TSX / Eppendorf CryoCube / PHCbi VIP ECO) — Intermediate, short-term storage. Not below the glass transition point, so LN2 vapor phase is safer for long-term storage
• DMSO-free / low-toxicity freezing · CryoStor CS2/CS5 (BioLife Solutions) or trehalose / polysaccharide-based media (Sigma-Aldrich) — When DMSO toxicity is a concern. Protocol revalidation is needed

A testing step that supports whether the frozen seed stock (MCB/WCB) "can really be used for manufacturing." Without passing this, the production that follows cannot proceed.

There are two axes to confirm. One is cleanliness: whether sterility, mycoplasma and externally introduced viruses are absent. The other is identity: confirming in line with ICH Q5D that the cells are indeed the intended HEK293 line.

Genetic stability is also evaluated. This step serves to assure the identity and cleanliness of the cells before upstream is started.

To support that the cell banks (MCB/WCB) can be used for manufacturing, 2 things must be confirmed. The 1st is "cleanliness": whether anything other than the host has crept in, and whether any externally introduced virus is mixed in. The 2nd is "identity": confirming that the cells are indeed the intended cell line. This is the idea required by ICH Q5D. When making a gene therapy vector such as AAV, the vector is produced in cells called HEK293. For this cell bank, tests of sterility, mycoplasma, identity, adventitious viruses and genetic stability are imposed as regulatory requirements. The content of each test is determined in line with ICH Q5A(R2) and Q5B and the pharmacopoeias of each region (US, Europe, Japan).

Alternatives in practice

• Mycoplasma NAT · Roche MycoTOOL / Thermo Fisher MycoSEQ Real-Time PCR Detection Kit (QuantStudio) — Rapid compared with the 28-day culture method. Pharmacopoeial alternative-method validation is needed
• Adventitious viruses (NGS / in vitro) · Metagenomic NGS with Illumina NextSeq 2000, or in vitro indicator cell assay compliant with 9 CFR / Ph. Eur. — Complements TEM and indicator cells. NGS for broad-spectrum detection
• Cell identification · STR (AmpFLSTR) + isozyme analysis, or SNP array (Thermo Fisher Axiom) — Detects mix-ups of human lines. STR is the standard
• Genetic stability · Karyotyping (G-banding) + NGS (Illumina) for passage stability confirmation — Comparison of MCB/WCB/EOPC. Confirmation of copy number and insertion site

A step that receives the 3 plasmids (gene of interest, RepCap, AdHelper), the "blueprints" to be put into the cells, as raw material and judges their quality. It inspects the starting material of manufacturing itself.

In addition to purity, supercoiled fraction (usually 80% or more) and ITR integrity, residual impurities such as host DNA, RNA, endotoxin and antibiotics are tested to decide pass or fail.

If this is lax, it bears directly on the later full capsid ratio, titer and genotoxicity risk. That is why suitability as a starting material is confirmed strictly.

In transient triple transfection, the 3 plasmids define the AAV genome, capsid and production, so a low supercoiled fraction (usually 80% or more) or poor ITR integrity bears directly on the full capsid ratio, titer and genotoxicity risk. This is a standard step to confirm suitability as a starting material, based on USP <1040> and others.

Alternatives in practice

• Anion-exchange HPLC topology · Thermo Scientific DNAPac PA200 / Agilent 1290 Infinity II Bio LC + AEX column — Quantitative separation of supercoiled / open circular / linear. Complementary to CGE
• Quantitation of residual host DNA and kanamycin resistance gene · Bio-Rad QX600 Droplet Digital PCR System — Absolute quantitation of host DNA and antibiotic resistance genes by ddPCR
• NGS confirmation of ITR / full-length sequence · Confirmation of full length and ITR integrity with PacBio Revio (HiFi long reads) or Oxford Nanopore — Long reads for confirming the ITR hairpin structure. Short reads for mutation detection

A "run-up" step in which cells are revived from a frozen stock and expanded stepwise to a quantity that can be sent to the production vessel. They are supplied from the seed culture (seed train), through the N-1 vessel, to the N production bioreactor.

The cells used are HEK293 adapted to serum-free suspension culture, carrying the adenovirus-derived E1 region that supports virus production. They are bridged to the next vessel while being expanded.

The contest is in preparing the state just before transfection. Viability, the intended cell density and an appropriate cell age, these three, are aligned so that cells in good condition are delivered to the production vessel.

The producer cells used in AAV manufacturing are HEK293 cells adapted to grow in suspension. They carry the adenovirus-derived E1 region, which supports virus production. Immediately before the operation that introduces genes into these cells (transfection), 3 conditions must be aligned: the cells are well viable, they are at the intended concentration (density), and they are at an appropriate cell age, neither too young nor too old. To send cells to the N production vessel in this state, the seed cells are expanded step by step. Specifically, the reference approach is to start from shake flasks and enlarge the vessel in stages to a small bioreactor and then a single-use bioreactor (SUB). So that the culture environment does not shift as the vessel gets bigger, a practice has been established of scaling up while keeping the equipment geometry matched, such as holding the agitation power per unit volume (P/V) constant at each stage.

Alternatives in practice

• High-cell-density seed train by perfusion · Repligen XCell ATF 6 / KrosFlo TFDF + Sartorius Biostat STR — N-1 concentration allows the production culture to start at a high inoculation density. Shortens production time
• Orbital shake expansion · Kuhner SB10-X / Thermo Scientific shakers (large flasks, wave motion) — Simple but the upper scale limit is small. Suitable for the stage before the SUB
• Expansion on microcarriers (adherent system) · Cytiva Cytodex / Corning microcarriers + SUB — When adherent HEK293T is used. For lines that are difficult to adapt to suspension
• Benchtop rigid-vessel STR · Sartorius Biostat B / Eppendorf BioFlo 320 (glass vessel) — Process development and condition optimization. Obtains correlation with the SUB

A step that decides the foundation of how AAV is made. It selects the combination of cell substrate and genome delivery method, and cultures and expands suspension HEK293 in serum-free medium to build the basis for production.

The choice here sets the downstream steps all at once. It is linked from purification design (the ratio of empty to full capsids, the host-derived impurity profile) to the burden on later steps.

It also spills over into yield, COGS (manufacturing cost) and regulatory handling. It can be called the first and biggest fork in the road for upstream.

Which production platform to use is the first big fork in the road for upstream, because the choice made here decides the downstream steps all at once. Specifically, purification design (the ratio of empty to full capsids, the host-derived impurity profile), yield and COGS (cost of goods), and the explanation to regulators (how to control manufacturing-derived impurities and the transfection reagent) are all dragged along by this one move. The most widely used approach in both clinical and commercial settings is the following. First, suspension-adapted HEK293 cells are expanded from N-1 to N in single-use stirred-tank bioreactors (STR). Then 3 types of plasmids are transiently transfected to produce AAV. This approach has clear advantages over adherent systems. Combining serum-free suspension culture with transient expression using PEI-based reagents is independent of scale, easy to scale up, closed and GMP-compatible. That is exactly why the single-use STR is now the standard format. In practice, this approach, with its clinical and commercial track record, is placed first as the reference method. Alternatives are then considered according to yield, cost and the scale required.

Alternatives in practice

• Sf9/baculovirus system · Sf9 insect cells + Sartorius/Cytiva SUB, Thermo Gibco Sf-900 III medium — Suited to high titer and large scale. Post-translational modifications and titer ratios differ from HEK293
• Inducible / stable producer cell lines · HEK293-based inducible systems (e.g., Cevec / stable lines of various CDMOs) + SUB — No plasmids needed, advantageous in cost and scale. Line construction takes time
• Fixed-bed bioreactor · Univercells scale-X / Pall iCELLis nano/500 (structured fixed bed) — Adherent culture with a high surface area. Where volumetric productivity is higher than in suspension systems

A step that standardizes "how genes are put in" for transient AAV production. Plasmids with the information divided into 3 are introduced simultaneously into suspension HEK293.

The breakdown is 3 types: pITR, which carries the transgene flanked by ITRs, pRepCap, which supplies Rep and Cap, and pHelper, which supports production. These are brought together and added at once.

The 4 points to satisfy are efficiency, reproducibility, scalability and GMP compatibility. Delivery conditions that achieve all of them are fixed as the reference method, ensuring that the subsequent vector production is established.

When AAV is produced transiently, the required information is divided among 3 plasmids and put simultaneously into suspension-cultured HEK293 cells. The 1st is pITR (cis), which carries the transgene flanked by ITRs. The 2nd is pRepCap (trans), which supplies Rep and Cap. The 3rd is pHelper, which supplies the adenovirus-derived helper genes. Whether this delivery goes well shows up in 4 points: efficiency, reproducibility, ease of scale-up and GMP compatibility, and it directly governs vector yield and the ratio of particles with contents to empty particles (empty/full ratio). Deciding which transfection reagent to use and how to form the complex is therefore the center of this step. In clinical and commercial manufacturing, the method that uses PEI, a cationic polymer, to form precipitated complexes and bring the DNA into cells without relying on lipids has become established as the reference method, because it is inexpensive, can be used at any scale, and GMP-grade raw materials are readily available.

Alternatives in practice

• Large-scale transfection with FectoVIR-AAV · Polyplus FectoVIR-AAV / FectoVIR-AAV GMP + Sartorius Biostat STR — High complex stability, suited to large scale and high density. Higher productivity than PEIpro
• PEI MAX triple transfection · Polysciences PEI MAX 40000 (HCl salt, PD/research use) — Low cost, for PD and screening. GMP documentation is more complete with the Polyplus products
• Calcium phosphate coprecipitation · CaCl2/HEPES system (prepared in-house, Sigma-Aldrich reagents) — Classic, low cost. pH-sensitive, inferior in robustness and scalability
• Ratio optimization by DoE · Sartorius Ambr 250 HT + MODDE/DoE software — Comprehensive optimization of plasmid ratio, PEI:DNA ratio and seeding density

A "tightening" step that raises the ceiling of production. In batch / fed-batch, cell density plateaus at around several million cells per 1 mL of medium.

What is spreading to address this is perfusion operation. While the stirred tank runs, a cell retention device keeps the cells in the vessel and medium is exchanged continuously to maintain a high density.

The advantage is that more AAV can be taken from the same volume. As process intensification, it is an upstream enhancement measure that lifts productivity.

In batch / fed-batch culture, cell density plateaus at around several million cells per 1 mL of medium (up to several ×10^6 cells/mL). To make more from the same volume, high-cell-density (HCD) culture by perfusion is spreading. Medium is exchanged continuously while a cell retention device keeps the cells in the culture vessel. Being able to continue culture while maintaining a high density in this way is becoming the reference approach for today's "intensified upstream." This method can keep the cell density as high as 9 to 50 million cells per mL (9×10^6 to 50×10^6 cells/mL). Moreover, both the continuous harvest of secreted AAV and the harvest after lysing the cells can be handled by the same single device.

Alternatives in practice

• ATF perfusion · Repligen XCell ATF System (single-use/stainless steel, diaphragm type) — Backflow reduces clogging and maintains high cell density. Gentle shear
• Pseudo-perfusion (semi-perfusion) · Intermittent medium exchange by centrifugation (Thermo Sorvall) or depth filtration — Simple but less continuous. For small scale and PD
• Continuous harvest TFDF · Repligen KrosFlo TFDF — Continuous recovery of secreted AAV. A 3-fold yield versus batch has been reported for AAV8 and others

The first downstream step that "takes out" the finished AAV from the cells. In triple transfection, most of the virus made does not leave the cells and accumulates inside.

So the cell membrane is chemically dissolved with a nonionic surfactant, and the vector is released all at once as a lysate (lysis solution). It is the exit after the end of culture.

How well this lysis goes governs the later steps. Conditions change the load on clarification and capture, so the design looks at both recovery and the ease of passing the next step.

When rAAV is made by introducing 3 plasmids simultaneously into HEK293 cells (triple transfection), most of the virus produced does not leave the cells but accumulates inside. So after culture ends, the cells must first be broken to release the vector. This lysis is the prerequisite for the clarification and capture that follow. Moreover, depending on how lysis is done, the load on the subsequent nuclease treatment and depth filtration changes, and the final AAV recovery goes up or down. In other words, this step is the rate-limiting point (the bottleneck in throughput) that determines how well the whole process turns out. For that reason, chemical lysis of cells with a surfactant has been chosen as the reference on the shop floor.

Alternatives in practice

• Secreted harvest (lysis-free) · Direct supernatant recovery by TFF/depth filtration (Repligen/Merck) — For supernatant-secreted capsids (AAV9, etc.). The intracellular fraction is not recovered
• Mechanical lysis · Microfluidics Microfluidizer / high-pressure homogenizer (with freeze-thaw) — Surfactant-free. Watch for capsid damage from shear
• High-salt lysis · High-NaCl (~0.5 M) lysis buffer (prepared in-house) — Promotes dissociation of nucleic acids and host proteins. Consistent with the salt-tolerant nuclease downstream

A pretreatment step that "clarifies" the liquid scattered by lysis. Breaking cells releases a large amount of host genomic DNA, which clogs resin and filters through viscosity and aggregation.

First, endonuclease cuts that DNA into small pieces and lowers viscosity. Primary and secondary 2-stage depth filtration then removes cells, debris and colloids and makes the liquid clear.

The aim is to prevent clogging in the next step. Clarifying here lets the AAV be handed smoothly to the affinity capture that follows.

When cells are lysed, a large amount of host cell genomic DNA is released. This DNA raises the viscosity of the liquid and forms aggregates that clog filters and purification resin (chromatography resin). So the DNA is first cut into small pieces with an enzyme called a nuclease to lower viscosity. The material is then put through 2-stage depth filtration to remove cells, debris (cell fragments) and colloids. This series of pretreatment is an essential step before capture chromatography (AAVX affinity), which captures the target virus, and how well it is done governs recovery and ease of later processing.

Alternatives in practice

• High-salt digestion with a salt-tolerant endonuclease · c-LEcta Denarase (retains activity under high-salt conditions) — Consistent with the high-salt lysis process. More salt-tolerant than Benzonase
• Clarification by chromatography · Sartorius Sartoclear / flow-through charged depth membranes — Design that simplifies or omits the nuclease step
• TFDF clarification · Repligen KrosFlo TFDF — Integrates harvest and clarification. Reduces depth filtration load
• Centrifugation + depth filtration · Continuous centrifugation (Alfa Laval/GEA) + Merck Millistak+ — Upstream removal when solids are high. For large scale

The starting point of purification, "fishing out" only the target AAV capsid from a sea of impurities. AAV in the production bulk is present in very small amounts, with large quantities of HCP, DNA and medium components mixed around it.

A camelid single-domain antibody fragment is used. It recognizes the three-dimensional structure and conserved loops of the capsid with pinpoint accuracy and selectively captures AAV in a single pass.

This single stage removes most of the HCP, DNA and medium components and concentrates to over 90% purity with high recovery. It is the capture step that hands a high-quality starting point to the polishing steps that follow.

In an AAV production bulk, the target AAV capsid is present in only very small amounts. Around it are mixed large quantities of host cell proteins (HCP) and DNA derived from HEK293 cells, Pluronic and medium components, and precursors of empty capsids. Affinity capture recognizes the three-dimensional structure (conformation) and conserved loops of the AAV capsid. A single pass therefore reaches a purity above 90%, and recovery is kept high (in many cases above 70–80%). As a result, the load on the downstream anion exchange that separates empty from full capsids and on the TFF that handles concentration falls greatly. Because of this effectiveness, affinity capture has become established as the reference method for primary capture in clinical AAV processes.

Alternatives in practice

• AVB Sepharose-type VHH capture · Cytiva AVB Sepharose High Performance + ÄKTA process — The long-standing VHH resin that predates AAVX. Serotype coverage is more limited than AAVX
• Serotype-specific VHH / synthetic ligand resins · Thermo POROS CaptureSelect AAV8/AAV9 affinity resins — High DBC/selectivity for a single serotype. Less versatile than AAVX
• Non-affinity capture · Sartorius CIMmultus QA / Thermo POROS 50 HQ (AEX primary capture) — Avoids ligand dependence and low cost. Purity is inferior to affinity
• Synthetic peptide ligand capture · Emerging synthetic peptide affinity resins (research/development stage) — Avoids animal-derived ligands. Track record and GMP documentation are limited

The most important polishing step determining quality, which sorts "filled" from "empty" capsids. Only full capsids are concentrated, and empty and partially filled capsids are removed.

The only difference between the two is the very slight negative charge difference produced by the enclosed DNA. So anion exchange (AEX) at pH 8.5–9 gives the surface a negative charge, and the salt concentration is raised little by little to tell them apart.

This high-resolution gradient elution is the crux, extracting a slight difference as a difference in potency and titer. It is the key point that governs dose design and safety of the product.

The difference between full and empty capsids is nothing more than the very slight difference in negative charge brought by the enclosed genomic DNA. Anion exchange (AEX) chromatography can use this charge difference directly for separation. Combined with high-resolution gradient elution in which the salt concentration is raised little by little, it distinguishes the two most straightforwardly, so it has become established as the reference method on the manufacturing floor. In the procedure, the capsid surface is first negatively charged at pH 8.5–9 and bound to a Q-type ligand, and the salt concentration is then gradually raised. The weakly charged empty capsids come off first, followed by elution of the target full capsids.

Alternatives in practice

• AEX membrane flow-through · Sartorius Sartobind Q / Pall Mustang Q (membrane adsorbers) — High flow rate, low pressure. Removes impurities in flow-through; full/empty separation is inferior to monoliths
• Two-stage (2-pass) AEX strategy · CIMmultus QA HR (2 passes under different conditions) — Concentrates high-purity full capsids stepwise. Trade-off with yield
• Affinity-integrated full capsid enrichment · Optimizing pH/salt conditions in the affinity stage such as POROS AAVX — Moving the step earlier reduces the separation load. Shortens the process

A step that "concentrates and changes the clothes of" purified AAV. AAV, about 25 nm in diameter, is concentrated to the target concentration and exchanged into the formulation buffer.

What is used is the sieving effect of the membrane. AAV is larger than the membrane pores and cannot pass through, so it stays on the membrane while only water and low-molecular-weight species escape. This is the principle of ultrafiltration (UF/DF).

Using this property, concentration and buffer exchange can proceed together. Finally the material is passed through sterilizing-grade filtration and finished as the drug substance / bulk.

A step that concentrates the purified AAV (about 25 nm in diameter) to the target concentration and then exchanges it into the formulation buffer. Ultrafiltration (UF/DF, tangential flow filtration) is used here, and because AAV is sufficiently larger than the membrane pores, it does not pass through the membrane. Water and low-molecular-weight species can therefore escape through the membrane while only AAV stays on the membrane. Exploiting this property allows concentration and buffer exchange (diafiltration) to proceed together, which is why this is the reference method. Finally, the material is filtered through a 0.2/0.22 µm sterilizing-grade filter to remove bioburden such as bacteria. What has been through this final filtration is taken as the drug substance / bulk, and this is a standard step in line with the sterile product requirements set by each country's pharmacopoeia.

Alternatives in practice

• Hollow fiber TFF · Repligen KrosFlo hollow fiber module (mPES, 100 kDa) — Gentle shear, open channel and low clogging. Dead volume slightly large
• Flat-sheet cassette TFF · Cytiva ÄKTA flux + Hydrosart/Omega cassettes (Sartorius/Pall) — High area / small footprint. Shear is higher than hollow fiber
• Single-pass TFF (SPTFF) · Repligen SPTFF / Pall Cadence in-line — In-line concentration links steps. Buffer exchange is needed separately
• Choice of final sterilizing membrane · Pall Supor EKV (PES) / Merck Millipore Express SHC — Optimize capsid adsorption and yield through membrane material and pore size

A step that passes the final formulation through a 0.22 µm sterilizing-grade membrane to assure sterility. It is the gate to making a product that can be administered.

However, AAV drug substance has a low active ingredient concentration (0.01–0.1 mg/mL), and capsid adsorption to membrane and tubing surfaces alone loses 30–40% of titer.

That is why a low-protein-binding membrane is chosen to sterilize reliably while holding down adsorption loss. Success depends on whether sterility and prevention of vector loss can both be achieved.

The concentration of active ingredient in AAV drug substance is very low (mostly 0.01–0.1 mg/mL). Consequently, even capsids simply adsorbing to the surfaces of filter membranes, tubing and containers have been reported to cause loss of 30–40% of titer. Sterile filtration in the final step is indispensable for ensuring sterility. However, how this loss is controlled decides success or failure, so how the reference filtration method is chosen is important.

Alternatives in practice

• 0.2 µm PES double-layer sterile filtration · Sartorius Sartopore 2 (0.45 + 0.2 µm PES double layer) — Built-in prefilter gives clogging resistance. Where throughput is higher than Durapore
• Filter preconditioning / saturation filtration · Membrane saturation flush with poloxamer-containing buffer — Minimizes initial adsorption loss. Flush volume needs to be minimized
• Closed aseptic processing without filtration · Sterile connectors (Cytiva ReadyMate / Pall Kleenpak) + manual dispensing — Avoids capsid loss from filtration. Bioburden control is strict

A step that fills the sterile-filtered formulation into containers and seals them, finishing it in a "form that can be administered." Aseptic operation takes place inside an isolator / RABS in a Grade A environment.

The product is aseptically filled into glass or resin vials by peristaltic pump, followed by rubber stoppering. Aluminum caps are crimped in a separate chamber to complete the container and closure system.

What it takes on is preventing cross-contamination and assuring sterility. Only after passing through here does the AAV injectable become a final product that can be administered.

An AAV injectable becomes a final product that can be administered only after 0.22 µm sterile filtration, filling into glass/resin vials in a sterile environment, and closure with rubber stoppers and aluminum caps. Aseptic operation inside an isolator / RABS is a standard step essential for preventing cross-contamination and assuring sterility.

Alternatives in practice

• RABS filling · Filling line with a restricted access barrier system (Bosch/Syntegon, Optima) — More flexible equipment than an isolator. Decontamination level is inferior to an isolator
• Prefilled syringe filling · Bausch+Ströbel / Optima PFS filling lines — Improves ease of administration. Watch for AAV adsorption and silicone oil compatibility
• Single-use filling assembly · Cytiva/Sartorius single-use filling flow paths — No cleaning validation needed, reduces cross-contamination. Flexible for small batches

A design step that prevents AAV from "sticking to the wall and disappearing" through formulation. At low concentrations, up to 90% is lost by adsorption to container walls, tubing and filter surfaces, and titer also varies.

The countermeasure is to assemble the formulation. A combination of buffer system, salt, nonionic surfactant poloxamer 188, sugar, osmolality and pH suppresses both adsorption and aggregation.

The aim is to secure long-term stability. Depending on the serotype, particles aggregate with one another, so a formulation that suppresses loss and aggregation is built with reference to the design guidelines of approved products.

AAV capsids show a troublesome property at low concentrations. They adsorb to the walls of containers, to tubing and to filter surfaces, so that up to 90% of vector particles are lost and genome titer also varies. Depending on the serotype, particles also aggregate with one another. How to suppress such loss and aggregation is the crux of formulation design. Approved AAV products show the answer. Zolgensma contains 200 mM NaCl, 1 mM MgCl2 and 0.005% poloxamer 188 in 20 mM Tris (pH 8.0), and Luxturna contains 180 mM NaCl and 0.001% poloxamer 188 in 10 mM phosphate (pH 7.3). Both add the nonionic surfactant poloxamer 188 as standard, and combine high ionic strength (equivalent to 200 mM or higher) with sugars that protect the capsid to suppress aggregation and adsorption. This is the reference method in practice.

Alternatives in practice

• Changing the surfactant to PS80/PS20 · Polysorbate 80/20 (Merck / Croda Super Refined) formulation screening — Adsorption suppression and oxidative stability differ from poloxamer. Optimize for each product
• Stabilization based mainly on sugars/polyols · Sucrose/trehalose/sorbitol formulation (Pfanstiehl low-endotoxin grade) — Freeze and freeze-drying protection. Reduces dependence on surfactant
• Optimization of pH and ionic strength · DoE formulation screening + DLS/SEC-MALS evaluation — Aggregation control through buffer system optimization. Minimizes the amount of surfactant
• Freeze-dried formulation · Freeze dryer (SP Scientific LyoStar / Millrock) + bulking agent (mannitol) — Enables ambient/refrigerated distribution. Formulation design tolerant of freeze-drying stress is needed

A step that freezes and keeps AAV drug substance / drug product in an ultra-low-temperature freezer at −60 to −80 °C. It lets the product sleep stably for a long time at deep cold.

Capsids (the virus shell) are fragile and tear, aggregate or adsorb under mechanical stress from the ice-water boundary, the air-liquid boundary and solute recrystallization during freezing.

That is why it is kept deeply cooled together with a formulation containing surfactant and sugar. The aim is to maintain a stable state that can withstand freezing and thawing.

The AAV capsid (the virus shell) breaks easily when it touches the boundary between ice and water or the boundary between air and liquid. When components dissolved at freezing recrystallize, force is applied there as well. Under such mechanical stress the capsid can tear, clump together, or stick to container surfaces so that it can no longer be recovered. The conditions that prevent this are almost fixed: cooling deeply and keeping it at minus 60 °C or lower, combined with surfactant, sugar and containers to which the contents adhere poorly. This is established as the reference for long-term stability with activity retained, and the commercial AAV products Luxturna and Zolgensma adopt the same design.

Alternatives in practice

• LN2 vapor-phase cryopreservation · Chart MVE 1500 series vapor-phase LN2 tanks (−150 °C or below) — Best long-term stability (below the glass transition point). High equipment and LN2 supply cost
• Storage as a freeze-dried solid · SP Scientific LyoStar / Millrock freeze dryers — Refrigerated/ambient distribution possible. Freeze-drying process validation is needed
• Short-term refrigerated storage at 2–8 °C · GMP pharmaceutical refrigerators (Thermo Scientific/Helmer) — Only within the expiry period after thawing. Long-term stability is inferior to frozen storage

A step that carries frozen AAV "still cold" from shipment to administration. Maintaining temperature is itself the quality assurance.

If temperature control breaks down or freezing and thawing are repeated, AAV capsids rupture and aggregate and titer falls. That is why temperature must be assured over the entire route.

The means are dry ice (about −78.5 °C) packing or cryoshippers. −60 °C or below is maintained over the entire process as a validated cold chain, and the product is delivered frozen.

AAV is fragile if temperature control breaks down during transport. If it departs from the deep-cold temperature range or goes through repeated freezing and thawing, capsids rupture, aggregation occurs and titer falls. So temperature must be assured over the whole route from shipment to administration, and maintaining minus 60 °C or lower (or refrigeration) in a validated cold chain is essential. In fact, commercial AAV products are shipped frozen using dry ice or cryoshippers.

Alternatives in practice

• LN2 dry vapor-phase cryoshipper · Chart/MVE dewar-type dry shipper (−150 °C class) — For ultra-deep-cold products and long-distance/international transport. Container cost is high and handling is strict
• Validated transport at 2–8 °C under refrigeration · Pelican BioThermal/Sofrigam refrigerated passive shippers + loggers — Short distances for thawed or limited-quantity products. Not for frozen products

A step that accurately counts the amount of genome packaged in capsids (vg/mL) by digital PCR (ddPCR). It is the primary specification when the dose is set in vg/dose.

The value measured here is the "denominator" when the ratio to infectious titer (vg/IU) and the empty/full ratio are calculated. In other words, many quality indicators are built up from this number.

It is the basis that supports dose determination and dose uniformity between lots. It can be called the foundational quantitation that is the starting point of all critical quality attributes (CQAs).

This is the step that becomes the primary specification when the dose is set in vg/dose. The genome amount measured here is the denominator when the ratio to infectious titer (vg/IU) and the empty/full ratio are calculated, and it is the starting point of every CQA. It decides the dose, decides the numerator of the full:empty ratio (the amount of genome packaged), and also serves as the basis for whether the dose is uniform from lot to lot.

Alternatives in practice

• Quantitative PCR (qPCR) · Applied Biosystems QuantStudio 7 Pro + TaqMan Fast Advanced Master Mix — Depends on a standard curve, so CV is larger than ddPCR. Widely available
• Genome titer estimate from UV absorbance · Thermo NanoDrop One / spectrophotometer (A260/A280) — Simple, but cannot distinguish empty/full or free nucleic acids. For screening use
• CDMS (charge detection MS) · Thermo Scientific Orbitrap (CDMS mode) / Megadalton Solutions — Directly evaluates the encapsidated genome by mass. Specialized and expensive instrument

A step that measures the "total number" of particles administered and the proportion of them that contain contents. Total capsid titer (cp/mL) and full:empty ratio are quantitated by sedimentation velocity analytical ultracentrifugation (SV-AUC).

Combining the two gives the vg/cp ratio (genome packaging efficiency), a figure showing how efficiently the genome could be packed.

The empty capsid fraction bears on immunogenicity, so these become critical quality attributes (CQAs) representing product quality. It is a step that pins down both quality and quantity of particles at the same time.

A step to pin down the total number of particles administered and the proportion of empty capsids, which bears on immunogenicity. Expressed in addition as the vg/cp ratio (the efficiency with which genomes were packaged), it becomes a critical quality attribute (CQA) that governs product quality.

Alternatives in practice

• Total capsid ELISA · Progen AAV Titration ELISA (serotype-specific, sandwich format) + plate reader — Simple, high throughput. Possible underestimation due to dependence on conformational epitopes
• Mass photometry · Refeyn SamuxMP Mass Photometer — Small sample volume and rapid, obtains empty/full/partially filled as a distribution. Complementary to AUC
• AEX-HPLC · Thermo Scientific DNAPac/ProPac / Agilent 1290 Infinity II Bio LC + AEX column — Rapid quantitation of total capsid and full/empty ratio. Depends on resin conditions
• UV A260/A280 ratio · Spectrophotometer (Thermo NanoDrop / Agilent Cary) — Simple estimate of full:empty. Limited to high-purity samples

A step that divides capsids into 3 types, "empty, full and partial," and absolutely quantitates their ratio. It is measured by sedimentation velocity analytical ultracentrifugation and confirmed with a reference.

The proportion of full capsids that properly encapsulated the therapeutic genome directly governs quality. It decides titer, is the basis of dose design, and bears directly on safety.

That is why this ratio is a critical quality attribute (CQA) that regulators require to be shown in submissions without fail. It can be called the step that guarantees the contents of the product in numbers.

The proportion of full capsids that have properly encapsulated the therapeutic genome governs quality. It decides titer, is the basis of dose design, and bears directly on safety. That is why this proportion is a critical quality attribute (CQA) that regulators require to be shown in submissions without fail.

Alternatives in practice

• Mass photometry (MP) · Refeyn SamuxMP (calibrated with empty AAV9, 3.74 MDa) — Empty/full/partially filled ratio from the molecular weight distribution. Small sample volume, rapid, highly correlated with AUC
• AEX-HPLC/UPLC · Agilent 1290 Infinity II Bio LC + AEX column / Thermo DNAPac — Fast, highly reproducible full/empty separation. Calibration samples are needed
• CDMS / native MS · Thermo Orbitrap (CDMS) / Megadalton Solutions — Directly judges empty/full by single-particle mass. Specialized instrument
• UV spectroscopy A260/A280 · Spectrophotometer (Thermo NanoDrop) — Simple estimate of the full ratio. Accuracy is limited for mixtures

A step that confirms whether the encapsidated genome is "full length as designed." It examines whether it is continuous from ITR to ITR without breaks, and free of partial genomes, truncations and rearrangements.

Capillary gel electrophoresis under denaturing conditions and laser-induced fluorescence detection are used. Contamination with broken genomes is distinguished with high sensitivity.

Truncated species lead to poor transgene expression and the risk of immune response. That is why the full-length integrity of the genome is evaluated as a CQA that governs both titer and safety.

If the encapsidated genome is partial or broken partway, the transgene may not be expressed well. Such broken genomes (truncated species) also lead to the risk of triggering an immune response. That is why the full-length integrity of the genome is carefully evaluated as a critical quality attribute (CQA) that governs both titer and safety.

Alternatives in practice

• PacBio SMRT/HiFi long-read sequencing · PacBio Revio / Sequel IIe — Full length, ITR integrity and recombination at the single-molecule level. More detailed than CGE but higher cost
• Duplex/milepost ddPCR · Bio-Rad QX600 Droplet Digital PCR — Detects truncation by the genome 5'/3' ratio. Sequence detail is not obtained
• Alkaline denaturing agarose gel · Bio-Rad Sub-Cell + ChemiDoc — Simple confirmation of the encapsidated single-stranded genome. Resolution is inferior to CGE
• Short-read NGS (Illumina) · Illumina NextSeq 2000/MiSeq — Sequence identification and detection of mutations/contamination. Weak for long-range structure

A step that measures the blend ratio of the three proteins that assemble the capsid, VP1:VP2:VP3. It is quantitated by CE-SDS (UV detection).

This ratio bears on whether the virus shell assembles properly. It also affects the efficiency of infection and transduction (delivery of genes into cells).

That is why it is a critical quality attribute that serves as a yardstick for watching lot-to-lot variability and as an indicator of whether the manufacturing process is running stably.

The VP ratio is one of the critical quality attributes. It bears on whether the virus shell (capsid) assembles properly, and also affects the efficiency of infection and transduction. That is why it serves as a yardstick for watching lot-to-lot variability and as an indicator of whether the manufacturing process is running well.

Alternatives in practice

• Reversed-phase LC-UV / RP-UPLC · Agilent 1290 Infinity II Bio LC / Waters ACQUITY Premier + C4 column — Separation and ratio quantitation of intact VPs. Complementary to CE-SDS
• SDS-PAGE densitometry · Bio-Rad Criterion/Mini-PROTEAN + ChemiDoc (Coomassie/silver stain) — Simple, low cost. Quantitative reproducibility is inferior to CE/LC
• RP-LC-MS · Thermo Vanquish + Q Exactive/Orbitrap (intact VP mass + relative quantitation) — Obtains VP identification and ratio at the same time. Instruments and analysis are advanced
• Western blot · Bio-Rad transfer + anti-VP (B1) and serotype-specific antibodies (Progen) — For identity confirmation. Quantitation is limited

A step that confirms whether the capsid protein (VP) is "the serotype as designed" and also examines its chemical modifications. LC-MS/MS is used.

First, identity: confirming that the VP sequence is that of the intended serotype. Next, post-translational modifications (PTMs): profiling where deamidation, oxidation, acetylation, phosphorylation and so on are present.

PTMs can affect the vector's transduction efficiency and infectious titer, so the step goes beyond confirming the sequence and characterizes what modifications are present.

Identity confirmation and PTM characterization of the capsid protein (VP) are needed for 2 reasons. The 1st is to confirm that the VP sequence is that of the serotype as designed. The 2nd is that post-translational modifications (PTMs) such as deamidation, oxidation, acetylation and phosphorylation can affect the vector's transduction efficiency and infectious titer. So the identity of the sequence is confirmed first, and then which PTMs are present, and where, is profiled.

Alternatives in practice

• RP-LC-MS of intact VPs · Waters ACQUITY Premier + BioAccord/Xevo (intact mass) — Rapid serotype identification and grasp of proteoforms. Sequence detail is complemented by MS/MS
• Native mass spectrometry · Thermo Q Exactive UHMR / native MS configuration — Measures the mass of intact particles/VP complexes. Specialized instrument
• Terminus and sequence confirmation by MS/MS · Thermo Orbitrap (bottom-up/middle-down) — N-/C-terminus and sequence confirmation. Multiple enzymes to maximize coverage
• CGE-LIF / Western blot · SCIEX PA 800 Plus / Progen anti-VP and serotype-specific antibodies — Separation and simple identity confirmation of VP3-related species. Supplements mass spectrometry

A regulatory-required step that measures the biological activity (infectious titer, potency) of "whether it actually infects cells and works."

The measurement uses TCID50 or qPCR/ddPCR endpoint methods. By measuring genome copies after infection, the titer of the particles that truly work is determined.

It also looks at the ratio of the number of infectious particles (IU) to the number of genome-containing particles (vg). This vg/IU ratio is an indicator of the quality and consistency of the drug product.

This is the item that confirms "whether it actually infects cells and works." It is also required by regulation. In addition, the ratio of how many truly infectious particles (IU) there are relative to the number of genome-containing viral particles (vg) is examined to show the quality and consistency of the drug product.

Alternatives in practice

• ddPCR endpoint TCID50 · Bio-Rad QX600 ddPCR — Replaces qPCR with ddPCR for better sensitivity and precision. Makes the infectious titer absolute
• Relative infectivity assay · Cell culture + qPCR/ddPCR (versus reference standard) — Lot-to-lot relative comparison. Consistency control rather than absolute titer
• Cell-based expression potency · Reporter expression (luciferase/GFP) + plate reader/flow cytometry — Functional potency. Reflects the biological activity of the transgene
• vg/IU ratio monitoring · ddPCR (vg) ÷ infectious titer (IU) — An indicator of infection efficiency. Trend control of manufacturing consistency

A safety test that looks for "AAV that multiplies by itself" (rcAAV), which can be born inadvertently during manufacturing. It can arise at low frequency by recombination between RepCap and ITRs.

The detection method brings it out by amplification. Permissive human cells are infected with rAAV and helper adenovirus and passaged multiple times, and whatever grows is watched.

Each passage is screened by serotype-specific qPCR to confirm that the level is below the detection limit. It is a step required by the FDA and is the linchpin of safety.

Replication-competent AAV can arise at low frequency through recombination between RepCap and ITRs, so FDA makes rcAAV testing mandatory for rAAV products. Permissive human cells are infected with rAAV and helper adenovirus and amplified through multiple passages, and each passage is screened by serotype-specific qPCR to confirm that the level is below the detection limit.

Alternatives in practice

• Absolute rep/cap quantitation by ddPCR · Bio-Rad QX600 Droplet Digital PCR — Absolute quantitation of rep/cap sequences after amplification. Advantageous for detecting low-level rcAAV
• Infectious titer-based rcAAV functional assay · Permissive cells + serial passage (cytopathic effect / titer measurement) — Evaluates replication competence functionally. Complementary to PCR; sensitivity and duration are issues

A step that evaluates the amount of "aggregates", in which particles have clumped together, and the particle size distribution. Aggregation bears directly on all of AAV's efficacy, the risk of provoking immunity, and safety.

SEC-MALS is used. The sample is first separated by size with size exclusion chromatography and then measured by multi-angle light scattering. Separating first and measuring afterward is the beauty of this method.

This allows monomer, dimer and large aggregates to be separated and quantitated individually. It is a step to accurately capture aggregation that threatens product quality.

Aggregation bears directly on all of AAV's efficacy, the risk of provoking immunity, and safety. That is exactly why accurate measurement is wanted, and what is used here is the MALS method, which measures after first separating by size. With this method, monomer, dimer and aggregates can each be separated and quantified.

Alternatives in practice

• DLS / MADLS · Malvern Panalytical Zetasizer Ultra (MADLS) — Simple, obtains particle size distribution/aggregation from a small sample. No separation, so resolution of large aggregates is limited
• SV-AUC · Beckman Coulter Optima AUC — Separates and quantitates aggregates by sedimentation coefficient. Low throughput and high analytical expertise required
• Mass photometry · Refeyn SamuxMP — Directly observes dimers/aggregates by single-particle mass distribution. Small sample volume, rapid

A step that measures both the "number" and "size" of particles at the same time and also controls invisible particles. Physical titer (total capsid concentration) and particle size are measured by SEC-MALS.

Particle number (concentration) is the basis for deciding how much vector to administer. This measurement provides the yardstick for dose design.

In addition, for injectables, insoluble invisible particles (subvisible particles) must be kept within USP limits. It is a step that reconciles dose determination with particulate control.

Particle number (concentration) is the basis for deciding how much vector to administer. In addition, for injectables, the amount of invisible insoluble particles (subvisible particles) is required to be kept within USP limits. That is why the 2 are measured together.

Alternatives in practice

• NTA · Malvern Panalytical NanoSight Pro — Number and size distribution by nanoparticle tracking. Effective in the submicron range
• Flow imaging microscopy (FIM) · Yokogawa FlowCam / Halo Labs HORIZON — Morphological identification of subvisible particles. Advantageous for distinguishing air bubbles from protein particles
• Physical titer by DLS/MADLS · Malvern Zetasizer Ultra — Simple evaluation of particle size and concentration. SEC-MALS/particle counters are superior for absolute number

A step that measures "DNA taken up inside" the capsid. Encapsulated DNA cannot be degraded even by nuclease treatment from the outside.

The targets are host cell DNA and plasmid/helper-derived DNA. After DNase treatment, qPCR/ddPCR quantitates only the encapsulated DNA that survives.

Because this is administered to the patient as it is, it is a CQA governing safety. Free DNA below 10 ng per 1 dose and fragment length of roughly 200 bp or less are controlled as the guide.

DNA that has been taken up inside the capsid cannot be degraded from the outside even by nuclease treatment. It will be administered to the patient as it is, so it is treated as a critical quality attribute (CQA) that governs safety. That is why regulations also require its control. Specifically, free DNA not enclosed in the capsid is kept below 10 ng per 1 dose, and fragment length is also guided to be roughly 200 base pairs (bp) or less.

Alternatives in practice

• Absolute quantitation of encapsulated HCDNA by ddPCR · Bio-Rad QX600 ddPCR — No standard curve needed, quantitates low levels with high precision. Lower CV than qPCR
• Encapsulated contaminating DNA by NGS · Illumina NextSeq 2000 (short-read) / PacBio Revio (long-read) — Covers the types and distribution of encapsulated nucleic acids comprehensively. PCR is better for quantitation
• Commercial residual HCDNA quantitation kit · Thermo resDNASEQ HEK293 (qPCR) — Comes with standardization and regulatory documentation. Reduces design effort

A step that measures, all together, the "host-derived impurities left over" that purification could not fully remove, and shows conformance to specifications.

The targets are 3: residual host cell DNA, plasmid/helper sequences and host cell protein (HCP). All lead to risks of oncogenicity and immunogenicity.

That is why they are confirmed after being removed thoroughly. It is shown that FDA/WHO limits are met, such as free residual DNA below 10 ng per 1 dose and fragment length of roughly 200 bp or less.

Residual host-derived impurities are critical quality attributes that govern safety. Specifically, they are host cell DNA, plasmid- and helper-derived sequences, and host cell proteins (HCP). They lead to risks of oncogenicity and immunogenicity, so after careful removal, it is shown that the limits set by the FDA and WHO are met. For example, free residual DNA is kept below 10 nanograms per 1 dose, and fragment length at roughly 200 base pairs or less.

Alternatives in practice

• Fluorescent dye (PicoGreen-type) total DNA · Thermo Quant-iT PicoGreen + plate reader — Simple quantitation of total double-stranded DNA. Not sequence-specific, for screening
• Mass spectrometry proteomics · Thermo Orbitrap (HCP identification/quantitation) — Identifies individual HCP species. Complements hard-to-detect HCPs that ELISA misses
• Digital/qPCR for residual plasmid and adenovirus helper · Bio-Rad QX600 / QuantStudio — Individually quantitates plasmid/helper-derived sequences
• Evaluation of replication-competent AAV · rep/cap qPCR/ddPCR + functional assay — Monitors rcAAV as well. Complements the safety profile

A step that measures the "leftover plasmids" from triple transfection. A large amount of plasmid is put in with transient delivery, and its fragments can remain.

The targets are RepCap, AdHelper, kanamycin resistance gene and so on. Results with and without DNase treatment are compared to quantitate separately those encapsulated in capsids and those that are free.

These differ in origin from host cell DNA, so they are controlled as an independent specification. It is a quantitation targeting rep, cap, resistance genes and the like that is essential for safety and specifications.

In transient transfection a large amount of plasmid is put in, and its fragments can be encapsulated in the capsid or remain as free impurities, so quantitation of residual plasmid targeting rep/cap/antibiotic resistance genes and the like, independent of host cell DNA ([25]), is essential for safety and specifications.

Alternatives in practice

• Real-time qPCR · Applied Biosystems QuantStudio 7 Pro + TaqMan Master Mix — Quantitates rep/cap/antibiotic resistance genes. CV is larger than ddPCR but widely available
• Profiling of encapsulated nucleic acid impurities by NGS · Illumina NextSeq 2000 / PacBio Revio — Covers the sequence and distribution of encapsulated contaminating nucleic acids comprehensively. PCR is better for quantitation