rAAV Manufacturing Process
The production and purification process of rAAV particles was adapted from methods outlined previously [37–40]. In brief, human embryonic kidney 293 (HEK293, #CRL-1573, ATCC) cells grown in FreeStyle™ F17 media (ThermoFisher, NY, USA) were cultured in suspension under controlled conditions (+ 37°C and 5% CO2). rAAV8 and rAAV9 vectors were produced by triple transfection of plasmids harboring a helper containing Adenovirus 5 genes, the packaging genes rep2/cap8 or rep2/cap9, and the added transgene of interest. rAAV vectors incorporate the transgene of a correct DNA sequence version of either 1) coagulation factor IX [41], which is deficient in hemophilia B disease, here termed as AAV8-FIX, or 2) the coagulation factor VIII, which is deficient in hemophilia A disease, here termed as AAV8-FVIII, or 3) the alpha-galactosidase (aGal) gene, which is deficient in Fabry disease, here termed as AAV9-aGAL [42], or 4) the alpha-glucosidase (GAA) gene, which is deficient in Pompe disease, termed here as rAAV9-GAA [43], or 5) the iduronate sulfatase (I2S) gene, which is deficient in Huntington’s disease, here termed as AAV9-I2S [44]. Transfection plasmids were complexed with polyethylenimine (PEI, Polysciences) according to the supplier's protocol. For rAAV8 serotypes, rAAVs released in the supernatant were harvested after five days post transfection. For rAAV9, three days post-transfection, cells were disrupted using an in-line disperser device to access the intracellular rAAVs. The suspension was diafiltrated and rAAVs were purified by ion exchange chromatography steps and ultracentrifugation as described elsewhere [37, 38].
Conditioning of rAAV Samples
The following sections describe the conditions that were applied to the rAAV8 and rAAV9 to verify their durability, stability and efficacy. For this purpose, a series of freeze/thaw cycles, pH exposure screening, and short-term storage and long-term storage were performed.
Freeze/Thaw Cycles
1 ml of rAAV9-aGAL vector was frozen at <-65°C and then thawed at room temperature (n = 2). Subsequently, a 50 µl aliquot stored in the refrigerator for a maximum of 8 hours at approx. +5°C pending subsequent use. The remaining vector was refrozen. This cycle was repeated up to ten times. Afterwards, the functionality of the treated vector aliquots was determined via the biopotency assay.
pH Exposure
rAAV9 samples were titrated with hydrochloric acid and adjusted to pH values of 8.5, 8.0, 6.0, 4.0 and 2.3. rAAV9-aGAL vectors derived from two different manufacturing processes underwent treatment. The vectors were then neutralized in cell culture medium and subsequently measured via biopotency assay.
Short-Term Storage of rAAV
Vector material was stored and analyzed at four distinct temperatures for a duration of up to 12 weeks. For this purpose, AAV9-aGAL vectors were filtered at 0.22 µm (Thermo Fisher Scientific, Austria) to prevent microbial growth at higher temperatures, divided into aliquots, and maintained at the specific temperatures for the duration of the respective study period. To assess the effect of incubation at ambient temperature, + 40°C, and + 5°C, an aliquot was taken for each point in time and tested for capsid titers in anti-AAV9 ELISA, vector titers in droplet digital PCR (ddPCR) analysis, and functionality in biopotency assay. For incubations at + 37°C, vector material was taken from the ultra-low temperature freezer in a more close-meshed manner over 72 hours, according to the same sampling scheme, and analyzed for its efficacy in biopotency.
Long-Term Storage of rAAV
For long term shelf life in the frozen liquid state, stability samples were stored in qualified, monitored storage areas at a temperature not exceeding − 60°C for a duration of up to 36 months. The initiation of the studies was defined by the set-down date of the samples in the stability chambers. The testing timepoints adhered to ICH guidance [33, 34], encompassing assessments at least every 3 months during the initial year, every 6 months in the subsequent year, and annually thereafter. At the designated testing timepoints, the samples were removed from the stability chambers, equilibrated at room temperature, aliquoted, and subjected to comprehensive analysis including physicochemical attributes and biopotency.
Lyophilized study samples were maintained at + 2 to + 8°C (monitored and controlled) for 10 months and were reconstituted at defined intervals. Lyophilization was performed as stabilizing process in which the substance in the liquid formulation is first frozen and then the quantity of the solvent is reduced first by sublimation (primary drying) and then by desorption (secondary drying) to values that will no longer support biological growth or chemical reactions [45]. First freezing was conducted at -60°C to solidify the water content, followed by primary drying where a vacuum is introduced to reduce the pressure within the drying chamber, leading to direct sublimation of ice crystals from solid to vapor, bypassing the liquid phase. In this experiment, the shelf temperature was maintained at -55°C under a pressure of 1.6 Pa. Once most of the water has been removed, the sample underwent secondary drying to further remove any remaining bound water molecules. This involved raising the shelf temperature to a target of + 25°C and reducing the chamber pressure to 1.1 Pa. Finally, reconstitution of the lyophilized samples was performed at 0 M (months), 1 M, 2 M, 3 M, 6M, and 10 M testing timepoints by adding WFI (water for injection) to the dried material, which represented the volume prior lyophilization. The sample was then rehydrated under gentle agitation until the drug product was fully dissolved and subsequently subjected to stability testing.
Analytical Methods
Droplet Digital PCR
For vector genome quantification, a Bio-Rad based ddPCR method was used, applying the fully automated QX One System or semi-automated QX 200 AutoDG system. The key steps of this analytical method involved partitioning of the sample into as many as 20,000 oil droplets, allowing each droplet to function as an independent compartment for PCR reaction. Following PCR amplification using fluorescent-probe-based method, the fluorescence was determined by a droplet reader. Droplets containing the target sequence were identified through fluorescence and categorized as positive, while droplets lacking fluorescence were categorized as negative. Poisson statistical analysis of the counts of positive and negative droplets enabled the absolute quantification of the target sequence.Sample preparation: To remove extraneous DNA, rAAV samples were treated with 4U DNase I (2000 units/mL, New England Biolabs, Ipswitch, MA, USA) for 60 min at + 37°C. This reaction was stopped with 0.5M EDTA pH 8.0 (Art. No. E177, VWR Life Science, Austria). In order to enhance the efficiency of DNase I activity, samples from intermediate production steps were pre-diluted at a ratio of 1:500 in a dilution buffer containing 0.1% of 10% Pluronic F-68 (Art. No. 24040032, Poloxamer 188 Non-ionic Surfactant (100X), Thermo Fisher Scientific, Austria), 2 ng/mL Salmon Sperm DNA, sheared 10 mg/mL (Art. No. AM9680, Thermo Fisher Scientific, Austria) and 1x GeneAmp PCR Buffer (Art. No. 4379876, 10X PCR Buffer, Thermo Fisher Scientific, Austria). The dilution buffer described was essential for ensuring an even distribution of rAAV capsids in droplets. It was also utilized for further diluting the samples following DNase I treatment. The dilution factor was determined based on the sample concentration to achieve droplets containing capsids, as well as droplets without. For an expected sample concentration of 1.00E + 13 vg/mL, the optimal total sample dilution, including all sample preparation steps prior to PCR, was determined to be 5.00E + 07 vg/ml.
Droplet generation, ddPCR cycling and readout: Mastermix for ddPCR was prepared by using 2X ddPCR Supermix for Probes (no dUTP) (Art. No. 1863025, Bio-Rad, Austria), 900 nM forward and reverse primer each (Suppl. Table 1), 200 nM probe (Microsynth AG, Austria) and 2 µL of sample preparation in 20 µL total volume for each replicate. Droplets were generated automatically in the Bio-Rad Droplet Generator.
A three-step PCR was carried out with a reduced ramp rate of 2°C per second, initiating with a single cycle at a temperature of + 95°C for 10 minutes to facilitate the degradation of the capsids and enable DNA amplification. Subsequent PCR denaturation steps were performed at + 95°C for 30 seconds and PCR extension steps at + 72°C for 15 minutes. The first 5 PCR cycles were performed with an annealing temperature of + 65°C, followed by 42 cycles of + 60°C for 60 seconds respectively. The PCR was completed with one step at + 98°C for 10 minutes. Plate reading was performed according to Bio-Rad instructions. Vector genome concentration was calculated by the appropriate Bio-Rad software and the vector genome titer (vg) per milliliter (mL) of the rAAV sample was determined.
ELISA of rAAV8 and rAAV9 Antigens
The quantification of rAAV8 capsids (cp) was conducted using an AAV-8 Titration ELISA Kit (Art. No. PRAAV8, Progen, Heidelberg, Germany) on a TECAN robotic system. Initially, microtiter strips were coated with a monoclonal antibody (ADK8), specifically targeting a conformational epitope on the assembled rAAV8 capsids. This coating facilitated the capture of rAAV8 particles. The detection of these captured rAAV8 particles was carried out in a two-step process. First, a biotin-conjugated monoclonal antibody, designed to specifically bind to the ADK8 antibody, was introduced to form an immune complex with the rAAV8 particles. Then, streptavidin peroxidase conjugates were added. These conjugates reacted with the biotin on the monoclonal antibody, forming a complex. For the detection an anti-AAV8 antibody (clone ADK8, Progen Germany) labelled with HRP (abcam HRP conjugation kit, ab102890 used as instructed by manufacturer) was added. To visualize the results, a peroxidase substrate solution was then applied. This addition initiated a chromogenic reaction, the intensity of which directly correlated with the quantity of rAAV8 particles present. Finally, The intensity was measured at a wavelength of 450, providing an estimate of the rAAV8 capsids concentration in the sample (cp/ml).
For quantification of rAAV9 capsids, a microtiter plate was coated with anti-AAV9 (clone ADK9, Progen Germany) overnight in PBS pH 7.4 at + 4°C. After four washing steps with PBS + 0.1% tween20, pH 7.4 (= PBST), samples and a rAAV9 standard of defined concentration were incubated in the plate for 1 hour at + 37°C. Again, after four washing steps, the detection antibody anti-AAV9 (clone ADK9, Progen Germany) labelled with HRP (abcam HRP conjugation kit, ab102890 used as instructed by manufacturer) was added, and incubated under the same conditions. Finally, the plate was washed 5x with PBST and TMB (3,3’m5,5’-tetramethylbenzidine, ThermoFisher Scientific, Austria) solution was added. Color development was stopped after approximately 10 min using 0.25 M sulfuric acid. The plate was measured at 450 nm and corrected for 620 nm absorbance. Samples were quantified relative to a 4-parametric fit of the rAAV9 standard curve.
In Vitro Biopotency Assay
The FIX in vitro biopotency assay was performed as described previously [46]. In brief, rAAV8-FIX vectors were quantified by ddPCR. Subsequently, the respective rAAV8 amount of each test item was used to infect the human liver cell line HepG2. During incubation, protein was expressed and released into the supernatant. In a second step, the activity of the FIX protein secreted into the supernatant was directly measured by a Rox Factor IX kit (Rossix, Moelndal, Sweden). The measurements of rAAV8-FIX samples are given as a percentage relative to a purified internal rAAV8-FIX vector standard material.
The FVIII in vitro biopotency assay was performed similarly as described above. Differing from the FIX assay, viral rAAV8-FVIII vector was used to infect HepG2 cells in the presence of 15 µg/mL VWF (von Willebrand factor), Takeda, Austria) and 7.5 µM EIPA (5-(N-Ethyl-N-isopropyl) amiloride, Sigma, Austria) supplemented in the cell supernatant to stabilize the expressed and secreted FVIII protein. Subsequently, the FVIII in the supernatant was used as cofactor for FX activation and its activity was measured in a Coatest® SP Factor VIII chromogenic assay (Chromogenix, Sweden).
Similarly, potency assays were used for rAAV9 therapeutic Fabry- and Pompe-vectors. These assays determine the metabolic activity of the transgenes aGAL, and GAA, respectively. Again, the potency is expressed relative to a purified internal vector standard material.
Size Exclusion Chromatography
rAAV9 samples were analyzed for aggregate formation by size exclusion chromatography (SEC). Analysis was performed on an Agilent 1260 HPLC system (Agilent, Waldbronn, Germany), consisting of a degasser, binary pump, autosampler, column oven coupled with a fluorescence detector, excitation at 280 nm and emission at 340 nm. For chromatographic separation an Agilent Bio SEC-5, 5 µm, 1000 Å, 7.8*300 mm (5190 − 2536) and a 1.47 mM KH2PO*2H2O, 2.68 mM KCl, 8.09 mM Na2HPO4, 350 mM NaCl, 0.02% NaN3, pH 7.4 running buffer was used. All chemicals were purchased from Sigma Aldrich (Saint Louis, MO, USA). Samples were transferred to an Agilent 300 µL high recovery, amber and after the column was equilibrated at 1 mL/min with run buffer, 4E + 11 rAAV9 capsids (based on ELISA) were injected. Aggregates were detected and integrated in a range of 5.5 to 8.5 min for signal intensity, and in the range of 8.5 to 10 min for monomer capsids. Aggregates % were calculated in relation to the total integrated area of aggregates plus monomer in each sample.
Differential Light Scattering (DLS)
Sample preparations were analyzed using the Wyatt PR III device (Wyatt Technology Corporation, Santa Barbara, CA, USA) and standard settings provided by the supplier. The rAAV particles were diafiltered and buffered in various buffers listed in Table 1. All samples were measured at a concentration of 1E + 13 cp/mL, which corresponds to reported dosages in pre-clinical and clinical studies ranging from 1E + 11 to 1E14 vg/ml [47–50]. With a theoretical molecular mass of 3729 kDa per empty capsid, the concentration of rAAV is approximately 0.064 mg/ml. This concentration was sufficient for DLS measurement due to the large size of rAAV capsids, compared to reference proteins such as BSA or lysozyme, which require higher minimum concentrations for detection. A data filter was employed to automatically exclude unsound measurements from the calculation of hydrodynamic radius (Rh).
Table 1
Hydrodynamic ratio and thermal stability of rAAV8 and rAAV9 in various buffers. This table illustrates the impact of different buffer compositions on the hydrodynamic diameter (measured in nanometers, nm) of recombinant adeno-associated virus serotypes 8 and 9 (AAV8-FIX and rAAV9-I2S). The data highlights variations in particle size due to changes in the buffer environment, reflecting potential influences on stability and bioactivity relevant to gene therapy applications. The temperatures reflect the onset points where significant changes in the structural stability of the viral capsids are observed, indicating their thermal tolerance in different buffer environments.
Code | Buffer Description | rAAV8 (nm) | rAAV9 (nm) | rAAV8 (°C) | rAAV9 (°C) |
B1 | 10mM Histidine, 50mM glycine, 5% trehalose,0.005% Tween80, pH 7.0 | 12.1 | 17.2 | 65.1 | 73.2 |
B2 | 10mM Histidine, 5mM sucrose, 0.005% Tween80, pH 7.0 | 13.6 | 15.7 | 65.8 | 68.2 |
B3 | 50mM Histidine, 0.005% Tween80. pH 7.0 | 13.5 | 15.5 | 65.8 | 71.7 |
B4 | 10mM Na3Citrate, 0.005% Tween80, pH 7.0 | 13.8 | 16.1 | 64.1 | 70.4 |
B5 | 80mM Glycine, 450mM NaCl, pH 5.5 | 14.5 | 13.6 | 63.3 | 69.8 |
B6 | 80mM Glycine, 150mM NaCl, pH 5.5 | 13.5 | 13.7 | 67.1 | 61.7 |
B7 | 80mM Glycine, 150mM NaCl, pH 2.5 | 13.8 | 33.6 | < 55.0 | < 55.0 |
B8 | 80mM Sodium Acetate, 150mM MgCl2, pH 5.5 | 13.3 | 13.5 | 64.8 | 69.6 |
B9 | 80mM Sodium Acetate, 450mM MgCl2, pH 5.5 | 13.1 | 13.6 | 64.8 | 68.0 |
B10 | 80mM Sodium Acetate, 1400mM MgCl2, pH 5.5 | 13.2 | 14.4 | 68.8 | 65.2 |
B11 | 80mM Glycine, 150mM, 150mM MgCl2, pH 3.5 | 13.3 | 13.8 | 58.9 | 61.4 |
B12 | 80mM NaAcetate buffer, 0.05% Tween80, pH 6 − 0 | 10.0 | 16.1 | 66.1 | < 55.0 |
B13 | 1mM HCl, pH 3.0 | 14.1 | 44.8 | 67.1 | n.a. |
B14 | 40mMTRIS, pH 8.0 | 14.2 | n.a. | 61.6 | n.a. |
B15 | 40mM TRIS, 100mM NaCl, pH 8.0 | 13.6 | n.a. | 64.8 | n.a. |
B16 | 40mM TRIS, 450mM NaCl, pH 8.0 | 14.1 | n.a. | 65.6 | n.a. |
B17 | 40mM TRIS, pH 10.0 | 14.3 | n.a. | 61.2 | n.a. |
B18 | 40mM TRIS, 100mM NaCl, pH 10.0 | 14.2 | n.a. | 60.9 | n.a. |
B19 | 40mM TRIS, 450mM NaCl, pH 10.0 | 14.2 | n.a. | 60.4 | n.a. |
The samples were stored in either Eppendorf tubes or Falcon tubes. Prior to measurement, liquid samples were briefly vortexed to ensure even distribution, but further vortexing was avoided to prevent aggregation. Instead, samples were centrifuged at 10,000 rpm for 5 minutes to remove airborne contaminants gently, and then pipetted into well plate.