Disposable Process for cGMP Manufacture of Plasmid DNA

Abstract

Disposables are increasingly being used in the manufacture of biopharmaceuticals. This article describes the design of a fully disposable process for the cGMP manufacture of clinical trial grade plasmid DNA. It addresses the rationale for implementing such a process with respect to the manufacture of patient-specific plasmid DNA vaccines for the treatment of leukemia. The process incorporates a number of disposable technologies, which are simple to use and thus reduce the need for investment in expensive equipment and cleaning validation.

Disposable technologies are being used either as replacements for conventional technologies or for the development of new processes for the manufacture of current good manufacturing practices (cGMP) grade biopharmaceuticals. These technologies are particularly applicable to manufacturers of investigational medical products (IMPs) where the range of products required to be manufactured is often diverse and at a small scale. In this article, we discuss the design of a disposable platform for the manufacture of patient-specific plasmid DNA (pDNA) vaccines for the treatment of leukemia.

pDNA Vaccine Therapy

The potential for treating a range of diseases by the delivery of DNA in vivo is being realized in the clinic in several different areas such as cancer, infectious diseases, and autoimmune diseases. The advantages of using naked pDNA over viral delivery routes include low immunogenicity (allowing a course of multiple treatments) and a simpler manufacturing system.¹ The pDNA vector cassette is highly suitable for the rapid development of various vaccine targets, such as personalized vaccines. One limitation for pDNA vaccination is the relative efficiency of the targeting and delivery of pDNA, which is being currently addressed by improving delivery modes by either chemical or physical means (e.g., adjuvants, liposomes, gene gun, or electroporation).

The Clinical Biotechnology Centre (a part of the UK's National Blood Service) has manufactured a range of pDNA vaccines for academic collaborators, and those vaccines are currently undergoing Phase 1 and 2 trials. Since 2005, we have been involved in the manufacture of anti-idiotypic bespoke pDNA vaccines for patients suffering from multiple myeloma. This B-cell malignancy is characterized by the secretion of a serum immunoglobulin paraprotein, which can be identified molecularly and re-engineered into a single-chain Fv construct. The construct is cloned into the pDNA vector cassette containing an immunostimulatory tetanus toxoid fragment.² Vaccination with the pDNA encoding the patient's paraprotein should induce the patient's immune system to mount anti-idiotypic responses against the tumor. The pDNA construct from one patient to the next may only differ by less than 2% of the backbone, making it suitable for designing a manufacturing platform. The vaccination regimen requires less than 50 mg of pDNA to be manufactured.

The Rationale

The promise of being able to deliver personally tailored vaccines for patients is a big challenge for the biopharmaceutical industry. Often, investments in such "orphan technologies" are not necessarily favored over more conventional generic approaches, where cGMP issues are less complex and the expected returns are greater. An ideal platform process would address the need to reduce manufacturing costs and allow rapid turnaround of batches of product.

Processing costs can be significantly reduced by eliminating the capital outlay in expensive, dedicated equipment (such as fermenters or chromatography systems) and opt for a process that is both disposable and simple to perform by an operator. A single-use process also eliminates the need for equipment decontamination procedures and cleaning validation studies, which are cumbersome and costly.

Process Design

Some prerequisites are necessary to underpin the platform process to ensure compliance with current regulatory requirements. The appropriate sourcing of key materials is critical, such as the exclusion of animal-derived ingredients and the use of pharmaceutical-grade (USP, class VI) disposable plastics throughout. The prudent selection of the appropriate E. coli host cell and pDNA vector will also maximize plasmid quality and yields.³

The use of validated and pharmacopoeia-compliant quality control assays is essential to demonstrate the effectiveness of the process to purify pDNA away from process-related impurities.⁴ The platform should be validated according to the cGMP guidelines and achieve a high quality of product to satisfy the current regulatory requirements.^4,5

An outline of the design of the cGMP process for manufacturing pDNA vaccines is shown in Figure 1. Each step should be attributed a specific function.

Figure 1

Fermentation

Plasmid DNA is transfected into E.coli by heat-shock, and clones containing the pDNA are identified under a selective pressure present on the vector (such as an antibiotic marker or by regulation of an essential E. coli protein). In an industrial setting, a master cell bank (MCB) is laid down from a clone, which is then cultured in large-scale fermenters by fed-batch, high-cell density culture.⁶ We have demonstrated that it is possible to omit the clone selection or MCB steps and culture a randomly selected clone in low-tech disposable shaker flasks without compromising high pDNA yields (>3 mg pDNA per gram of wet weight E. coli). We are able to routinely produce up to 200 mg of pDNA in a culture volume of less than 4 L in 24 hours. High specific yields are favored over high volumetric yields (mg pDNA per liter of ferment) because they reduce the relative starting load of E. coli-sourced contaminants. The ferment is harvested by centrifugation into single-use plastic centrifuge bottles at late log for maximal pDNA yields. Because of cost, we decided not to use disposable fermenter systems or crossflow applications for this step.

Lysate

Standard alkaline lysis methods must be optimized to ensure the efficient lysis of cells and subsequent removal of major contaminants such as genomic DNA and proteins.⁷ The lysis is performed in a plastic container containing a low-level outlet attached to prefiltration and 0.22 μM filtration devices. The whole assembly and container are autoclaved together before use. Addition of a secondary salt such as ammonium acetate or calcium chloride precipitates and removes RNA (up to 45% reduction in contaminating nucleic acid load observed).⁸ The secondary salt partitions the flocculent (precipitated material) in an upper layer away from the lysate, which enables the subsequent clarification of the lysate by dead-end filtration (Figure 2). The clarified lysate can be stored in a bioprocess container or can be pumped directly into the crossflow filtration container for the start of the purification step.

Figure 2

Figure 3 shows the complex nature of the lysate, which is known to contain large amounts of RNA, endotoxin, and trace amounts of contaminating genomic DNA, protein, and other non-supercoiled pDNA related isoforms. The pDNA constitutes less than 5% of the total mixture, which presents a challenge to purify the desired monomeric supercoiled pDNA to homogeneity from contaminants that have similar physiochemical properties. Supercoiled pDNA is tightly coiled and is widely accepted as being the therapeutically efficacious form, mainly due to its relatively small size compared to other isoforms.¹

Figure 3

Purification

The purification step initially involves the diafiltration and concentration of the lysate with a hollow-fiber crossflow filtration cartridge. This device is assembled with a vacuum-resistant plastic container fitted with a series of three-way valves, vent filters, and Pharmed tubing in an enclosed system. The buffers, which are stored in bioprocess containers (BPCs), are attached to a three-way valve on the diafiltration line (Figure 4). An appropriate molecular weight cut-off membrane is chosen (300 kDa or smaller) to allow the retention and high recovery of supercoiled pDNA from the lysate. The SC nature of plasmid requires the selection of a membrane with a MW cut-off approximately 10 times smaller than expected to prevent losses across the membrane. Carefully defined operating conditions are also necessary to reduce the effect of shear, because plasmid is prone to affects that are exacerbated by the prevailing high salt conditions. The lysate is diafiltered, concentrated, and filtered into a BPC before anion exchange chromatography. Crossflow filtration of the lysate has been previously reported to be an excellent RNA removal step with more than 90% clearance demonstrated in practice (Figure 3).⁹

Figure 4

Anion exchange chromatography media in packed columns is the conventional approach for the initial capture of pDNA. The negatively charged pDNA molecules bind electrostatically to the media, but limitations in the binding capacity is a problem that remains unresolved, mainly because of the large size of the plasmid molecule. This has recently been addressed by a shift in emphasis from beaded media to a membrane approach. Membrane devices offer many benefits, including disposability (no need for column packing and qualification), an increased dynamic binding capacity for larger molecules, and high process flow rates.¹⁰

In our platform, a Mustang Q (Pall Corporation, East Hills, NY) capsule is used downstream from the crossflowed lysate (in contrast to other organizations that have used this step as a direct capture from E. coli lysate).¹¹ Pretreatment by crossflow ultrafiltration is advantageous for increasing the Mustang Q's binding capacity for pDNA (since the RNA no longer competes once removed), improving recovery of eluted pDNA, and achieving higher product purities. The membrane device is attached to Pharmed tubing and three-way valves for processing during purification (Figure 5). The apparatus is autoclaved, operated in a Grade A laminar flow hood by a pump, and all fluid movements are contained in BPCs. If the step has been validated in terms of the wash conditions, elution volumes, and flow rates, the cartridge does not need to be attached to expensive instrumentation. The loading and washing of the sample is rapid, followed by a longer and staggered multiple high salt elution step. This ensures that the recoveries of pDNA achieve greater than 70% of total plasmid load, because membranes are characterized by slower off rates for pDNA.¹² Mass balance analysis across this step efficiently removes 2 log₁₀ orders of endotoxin and more than 50% of the remaining residual protein.

Figure 5

The eluate is collected into a BPC and ammonium sulphate is added to the BPC to achieve a final concentration of 1–1.5 M for the negative capture hydrophobic interaction step. At a set concentration of salt, impurities such as endotoxin (>90% removal) and RNA (>95% RNA removal) are selectively bound onto the HIC media while pDNA is recovered. The media is preconditioned with ammonium sulphate, autoclaved, and pumped into the BPC to achieve a specific nucleic acid ratio load to resin. The batch-mode HIC is incubated on a rotating mixer before 0.22 μM filtration into another BPC to remove the resin and recover the product.

Formulation and Quality Control

Purified pDNA is formulated into the buffer of choice by a disposable scaled-down version of the previous hollow fiber membrane assembly (Figure 6). The pDNA is formulated at the desired concentration in the appropriate solution (normally isotonic) requested by the sponsor. The choice of filter membrane is also critical to avoid unnecessary losses of plasmid by nonspecific binding.

Figure 6

The final product is fully tested, and satisfies current recommended guidelines (Table 1). Limited in-process QC testing is performed because of the continuous nature of the process.

Table 1. Final QC release testing and specifications for clinical grade pDNA

Conclusion

The continuous process is suitable for the rapid turnaround required for the preparation of these vaccines at a low cost. It is aseptic, contained, and completely disposable to avoid product cross contamination. It has been estimated that on current demand, with two operators in two cleanrooms, up to 40 individual vaccines could be manufactured annually.

By careful design, it is possible to eliminate some of the main obstacles and limitations normally associated with conventional processes (e.g., to reduce the reliance on high-tech equipment and to eliminate the use of column chromatography). The process has been shown to achieve a recovery of pDNA of greater than 45% from bacterial lysate with the quality of the final product satisfying current regulatory requirements.

Paul Lloyd-Evans, PhD is an operations manager, Denise A. Phillips is a production supervisor, Antony C.C. Wright is a production scientist, and R. Keith Williams is a business development manager, all at the Clinical Biotechnology Centre, Bristol, UK, 011.792.89388, paul.lloyd-evans@nbs.nhs.uk

References

1. Schleef M, et al. DNA Pharmaceuticals. Wiley-VCH Verlag GmBH & Co KGaA, Weinheim;2005.

2. Stevenson FK, et al. DNA vaccines to attack cancer. PNAS. 2004;10(2):14646–14652.

3. Werner RG, et al. pDNA—from process science to commercial manufacture. Pharm Technol Eur. 2002;March;34–40.

4. European Pharmacopoeia Supplement 5.6. Gene transfer of medicinal products;2007.

5. Rules and guidance for pharmaceutical manufacturers and distributors. Pharmaceutical Press, London;2007.

6. Lee SY. High cell density culture of Escherichia coli. TIBTECH. 1996;14:94–105.

7. Birnboim HC, Doly JA. A rapid alkaline extraction procedure for screening recombinant plasmid DNA. 1979;7:1513–1523.

8. Eon-Duval A, Gumbs K, Ellet C. Precipitation of RNA impurities with high salt in a plasmid DNA purification process: use of experimental design to determine reaction conditions. Biotech Bioeng. 2003;83:544–553.

9. Eon-Duval A, et al. Anal Biochem. 2003;316:66–73.

10. Shukla AA, et al. Process scale bioseparations for the biopharmaceutical industry. CRC Press Florida. 2007.

11. Shiying Z, Krivosheyeva A, Nochumson S. Large-scale capture and purification of plasmid DNA using anion-exchange membrane capsules. Biotechnol Appl Biochem. 2003;37:245–249.

12. Endres HN, et al. Evaluation of an ion-exchange membrane for the purification of plasmid DNA. Biotechnol. Appl. Biochem. 2003;37:259–266.