A Scalable Adenovirus Production Process from Cell Culture to Purified Bulk Product

Adenovirus (AdV) vectors are commonly used in cancer gene therapy trials, evaluated in gene therapy, and used in vaccines for various diseases. In this study, we have combined technical evaluation of process steps and process economy calculations, from AdV production in cell culture to purified bulk product in up to 10 L scale^1-8. Human AdV5, expressing the green fluorescent protein (GFP), was used for process development. The cells were lysed using Tween™ 20 as an alternative to Triton™ X-100 that is now on the authorization list (Annex XIV) of Registration, Evaluation, Authorization, and Restriction of Chemicals (REACH).

Analytical methods for determination of virus titer are challenging and depend on purity and quality of the sample. We used traditional analytics but also developed new sensitive and reproducible assays for virus titer. Based on analytical data, the novel downstream process was compared with a reference process regarding virus load capacity, recovery, and purity (Figure 1).

Materials and methods

Adenovirus production

HEK293 suspension cells (HEK-293.2sus, ATCC, grown in HyClone CDM4HEK293 using the Xcellerex™ XDR-10 or ReadyToProcess WAVE™ 25 bioreactor systems, were infected at a cell density of about 1 × 10⁶ cells/mL  and a multiplicity of infection (MOI) of 10 with E1/E3-deleted recombinant AdV5 coding for the GFP reporter protein.^1–3

Harvest and filtration

Forty-two hours post infection, the cells were lysed with 0.5% Tween 20 and treated with 20 U/mL Benzonase™ + 1 mM MgCl2 for 4 hr in the bioreactor. The harvest was thereafter clarified by normal flow filtration (NFF) using a combination of 2 μm and 0.6 μm ULTA GF filters. Concentration and buffer exchange were performed by tangential flow filtration (TFF) on a ReadyToProcess™ hollow-fiber filter with a nominal molecular weight cutoff (NMWC) of M_r 300,000 (10 × ultrafiltration [UF]/5 × diafiltration [DF] into 20 mM Tris, pH 8.0 + 300 mM NaCl, 2 mM MgCl₂).⁴

Chromatography steps

Columns were operated on an ÄKTA™ pure 150 system.^5–7

Novel process

For capture, Capto™ Q ImpRes resin (HiScale™ 26 column, 88 mL [3 L scale] or HiScale 50 column, 294 mL [10 L scale]) was used. Elution was conducted with 20 mM Tris, pH 8.0 + 2 mM MgCl₂ using a linear gradient of 480–570 mM NaCl. Polishing was conducted on Capto Core 700 resin (HiScale 16, 10 mL [3 L scale] and 29 mL [10 L scale]).

Reference process

For capture, Q Sepharose™ XL resin (HiScale 50 column, 249 mL [3 L scale]) was used. Elution was performed in two steps using 20 mM sodium phosphate, pH 7.3 + 2 mM MgCl₂ + 2% sucrose with 500 mM + 750 mM NaCl. Polishing was performed using Sepharose 4 Fast Flow resin (HiScale 50 column, 382 mL [3 L scale]).

Formulation and sterile filtration

Sample was concentrated and buffer exchanged (5 × UF/ 5 × DF) into 20 mM Tris, pH 8 + 25 mM NaCl, 2 mM MgCl₂, 2.5% glycerol. Process aliquots of the final purified bulk were sterile filtered using a syringe filter (polyethersulfone, 0.2 μm).

Analytics⁶

Infectious virus titer was determined using the 50% tissue culture infective dose (TCID₅₀) assay and by automated fluorescence microscopy (AFM) using the IN Cell Analyzer. Total virus titer was determined by hexon DNA qPCR and size exclusion chromatography (SEC)-HPLC using a Superose™ 6 Increase 10/300 GL column. The Biacore™ T200 system was used for determination of virus particles through binding of virus fiber or hexon protein to CAR or FX protein, respectively, immobilized on a Biacore Sensor Chip CM5. Host cell protein (HCP) was determined using an ELISA assay; total protein, using a BCA assay kit; host cell DNA (hcDNA), by qPCR and total DNA using Quant-iT™ PicoGreen™ dsDNA Reagent (Invitrogen). Analysis of AdV5 samples by transmission electron microscopy (TEM) was conducted in collaboration with Vironova AB using the MiniTEM™ system.

Results

Cell growth, cell viability, and Adv5 productivity were similar between the two bioreactor cultures (Figure 2). To minimize co-elution with virus, Capto Q ImpRes with gradient elution was applied in the novel process (Figure 3A). Polishing with Capto Core 700 allowed binding of impurities inside the bead, while the virus eluted in the flowthrough (Figure 3C). The reference capture step, using Q Sepharose XL, was performed with step elution (Figure 3B) followed by polishing with SEC (Figure 3D).

Cell growth, cell viability, and Adv5 productivity were similar between the two bioreactor cultures (Figure 2). To minimize co-elution with virus, Capto Q ImpRes with gradient elution was applied in the novel process (Figure 3A). Polishing with Capto Core 700 allowed binding of impurities inside the bead, while the virus eluted in the flowthrough (Figure 3C). The reference capture step, using Q Sepharose XL, was performed with step elution (Figure 3B) followed by polishing with SEC (Figure 3D).

The final purified bulk meets the regulatory requirements for virus purity (Table).

Capto Core 700 allows up to 300-fold higher sample loads compared with SEC, but it is not designed to remove full-length DNA, giving the SEC approach a certain advantage. Comparable yields were achieved in both processes, but the novel process showed lower impurity levels. This was confirmed by TEM, showing significantly lower levels of debris with the novel process.

As seen by SEC-HPLC analysis of clarified feed and the final purified bulk, impurities were efficiently removed in the novel process. The Biacore assays for total virus titer were shown to be sensitive, reproducible, and robust, with the benefit of reduced assay and hands-on time while showing similar results to qPCR.  AFM using the IN Cell analyzer was shown to give similar results as TCID₅₀

Process economic evaluations indicate that the novel process is favorable over the reference process at increasing titers and scales, mainly due to the choice of polishing resin. Additionally, single-use process setups proved more cost-efficient than their stainless-steel counterparts in all investigated scenarios.⁸

Conclusions

We propose a novel process for Adv5 production with technical and process economical advantages compared with an existing reference process. The novel process, using a combination of Capto Q ImpRes and Capto Core 700 in the downstream purification steps, produced pure final bulk product that meets regulatory requirements. New convenient and sensitive virus titer assays using Biacore T200 and IN Cell Analyzer were developed with comparable results to traditional methods.

References

1. Application note: Evaluation of HEK293 cell growth and adenovirus productivity in HyClone CDM4HEK293 medium. GE Healthcare, 29264715, Edition AA (2017).
2. Application note: Adenovirus production in single-use Xcellerex XDR-10 bioreactor system. GE Healthcare, KA874021017AN (2017).
3. Application note: Adenovirus production in single-use ReadyToProcess WAVE 25 bioreactor system. GE Healthcare, KA879160418AN (2018).
4. Application note: Optimization of midstream cell lysis and virus filtration steps in an adenovirus purification process.
   GE Healthcare, KA875220218AN (2018).
5. Application note: Downstream process development for efficient purification of adenovirus. GE Healthcare, KA876080618AN (2018).
6. Application note: Determination of adenovirus concentration using Biacore T200. GE Healthcare, KA878080618AN (2018).
7. Application note: Scalable process for adenovirus production. GE Healthcare, KA877080618AN (2018).
8. Application note: Process economic simulation for scalable production of adenovirus. GE Healthcare, KA3941080618AN (2018).

This bioprocess tutorial is a modified version of a GE Healthcare poster presentation that can be found in its entirely at www.GENengnews.com.

Åsa Hagner-McWhirter, Gustaf Ahlén, Magnus Bergman, Eva Blanck, Sara Häggblad-Sahlberg, Pelle Sjöholm, Maria Soultsioti, Sravani Musunuri, Elisabeth Wallby, Anna Åkerblom, Åsa Lagerlöf, and Mats Lundgren are affliated with GE Healthcare Bio-Sciences, Sweden.