Methods for the Automated Manufacturing of an Autologous Dendritic-Cell Immunotherapy

The dendritic cell (DC), the most powerful antigen-presenting cell in the immune system, has been a popular choice as a basis for personalized or autologous cellular immunotherapies. For these autologous DC immunotherapies, a batch of drug product is generated for each individual patient using his or her own cells. Scientists use three main approaches to obtain or generate DCs for a patient.

One approach is to isolate circulating DCs directly from the blood or from white blood cells collected through leukapheresis. The number of circulating DCs in the peripheral blood is extremely low, which limits the potential yields and doses that can be obtained using this approach. The advantage of this approach is that the DCs require minimal manipulation once isolated. Scientists essentially load the DCs with the antigen or antigens of interest against which an immune response is desired.

The two other approaches involve isolating precursor cells and culturing them to generate DCs. These precursor cells are either hematopoietic stem cells (HSCs) or monocytes. HSCs require longer culturing times to generate DCs compared with monocytes and require the patient to be mobilized (e.g., pretreatment with granulocyte colony-stimulating factor) before the leukapheresis. The advantage of the HSC approach is the ability to proliferate the cells before differentiation. Monocytes reduce culture time and manipulations compared with HSCs with the limitation that the cells have with no proliferation capacity. Therefore, the key to autologous DC processing from monocytes is managing losses and optimizing recoveries at each step to ensure an efficient process.

Generating a dendritic-cell immunotherapy

Using monocytes, Argos Therapeutics developed a robust method for generating its autologous DC therapy from leukapheresis for clinical trials in renal-cell carcinoma (RCC) and human immunodeficiency virus (HIV) indications (see Figure 1). The monocytes are isolated from the leukapheresis using the Elutra Cell Separation System (Caridian BCT). This instrument uses elutriation, also known as counterflow centrifugation, to fractionate the cells in the leukapheresis based primarily on cell volume, thus providing a final fraction of enriched monocytes. These monocytes can be cultured immediately to generate DCs, or can be frozen to be thawed and cultured when the antigen is available. Monocytes are cultured with granulocyte-macrophage colony-stimulating factor (GM-CSF) and interleukin 4 (IL-4) for 5 days to generate immature DCs in culture bags. Maturation media containing tumor necrosis factor α, interferon γ, and prostaglandin E₂ is added, and the DCs are cultured overnight to mature before the antigen is introduced.

Figure 1: Overview of an autologous dendritic cell-manufacturing process for oncology and infectious-disease indications. (ALL IMAGES ARE COURTESY OF THE AUTHORS)

Antigen is added to the DCs in the form of ribonucleic acid (RNA) using electroporation. The antigen RNAs are amplified for each patient from samples of his or her tumor or virus (1, 2). This method provides antigen that is unique to that individual and tailors the DC immunotherapy for that individual. This technique, however, makes the manufacturing more complex than alternative approaches in which the antigen is universal for all patients and can be manufactured in bulk quantities. The RNA is manufactured for each patient by isolating the total tumor or viral RNA from the patient's sample. That isolated RNA is converted to first-strand cDNA through reverse transcription and amplified by polymerase chain reaction (PCR) using nonsequence-specific primers and methods, thus making them universal for all samples. RNA is generated using the resulting amplified cDNA through an in vitro transcription (IVT) reaction and is post-transcriptionally capped to ensure a high capping efficiency. Using this process, milligrams of amplified RCC messenger RNA (mRNA) are generated consistently from micrograms of total RNA. For HIV, RNAs for the quasispecies of gag, vpr, rev, and nef proteins in the viral sample are amplified (2). From isolated viral RNA, the concentration of which cannot be measured by standard spectrophotometric methods, milligrams of RNA for each antigen are amplified for each patient.

In addition to the amplified RNA from the patient's disease, an RNA-encoding cluster of differentiation 40 ligand (CD40L) is added to the RNA payload. The purpose of adding this CD40L RNA is to provide the CD40–CD40L ligation signal required by the DC to induce IL-12 secretion (3–5). IL-12 is linked to the functionality of the DC because IL-12 secretion is one of the three signals required for a typical adaptive immune response (6). A technique for quantifying the release of IL-12 from the DC immunotherapy is in development as the drug product's potency assay (7).

This CD40L RNA is generated in bulk from a plasmid with one batch of CD40L RNA used for several batches of the DC immunotherapeutic drug product (8). For each batch of CD40L RNA, the plasmid is linearized, and uncapped CD40L RNA is generated using IVT methods. The uncapped CD40L RNA is capped and polyadenylated to generate the final CD40L RNA that is added to the RNA payload during electroporation. This process of maturing the DCs before electroporating them and adding CD40L RNA to the RNA payload has been called the postmaturation electroporation CD40L or the PME CD40L process (9). Dendritric cells resulting from this maturation process expand the central and effector memory T cells (CD8⁺CD28⁺) associated with favorable clinical outcomes (10).

Following electroporation with the amplified RNA from the tumor or viral sample and the CD40L RNA, the DCs are cultured for 4 h with GM-CSF and IL-4 to recover, translate the RNAs, and process and present the resulting tumor or viral peptides. After culture, the DCs are harvested, formulated in autologous plasma collected during leukapheresis and cryoprotectants (i.e., dimethyl sulfoxide and dextrose), and frozen in multiple vials. Each vial is a single dose of drug product. These vials are stored cryogenically and shipped individually to the clinical site for administration to each subject. Implementing this cellular process based on elutriation, culture bags, and the PME-CD40L maturation method, yields a mean number of doses produced per batch greater than 20 for the RCC and HIV indications (see Table I). This method provides multiple years of dosing for a patient from a single leukapheresis.

Table I: Results of the cellular process based on elutriation, culture bags, and PME-CD40L maturation methods for manufacturing clinical-scale batches of RCC and HIV.

The drug-product release testing includes post-thaw total viable cell count and viability to verify the dose strength and immunophenotyping for identity (see Table I and Figure 2). Cell-surface markers CD80, CD86, CD83, and CD209 identify the cells as mature DCs with the appropriate co-stimulatory molecules to generate an immunostimulatory T-cell response. CD14 is a monocyte marker; therefore, the low percentage confirms that the monocytes were converted to DCs. Human leukocyte antigen-DR indicates the presence of major histocompatibility complex Class II receptor for peptide antigen presentation. These data demonstrate the consistency in the formulation and quality of the DCs despite the significant biological variability in the starting materials. The consistency in results between batches for each disease indication and for the two different disease indications establishes that the clinical manufacturing methods are robust.

Figure 2: Post-thaw immunophenotyping results confirmed the identity and quality of the dendritic-cell drug products generated for renal-cell carcinoma (RCC) and human immunodeficiency virus (HIV) after implementing the cellular process based on elutriation, culture bags, and PME-CD40L maturation methods. CD is cluster of differentiation, and HLA is human leukocyte antigen.

Automating the process for commercialization

The process for generating this autologous DC immunotherapy has been consistent and robust for the RCC and HIV indications during Phase II clinical trials. Elutriation and culture bags have moved key cellular processes to functionally closed, single-use disposable units. However, some manipulations and processing are still performed using manual methods, including open manipulations in biological safety cabinets. Although this approach may be feasible for the number of batches required for clinical trials, it is not practical for commercial manufacturing because every patient requires a new batch of material to be produced.

Advanced RCC has a relatively low incidence of an estimated 9400 new cases in the US in 2010 (11–13). This number alone would mean a considerable number of batches to manufacture per year. An HIV product would likely require manufacturing for more than three times as many patients per year in the US, compared with an advanced RCC product. Even more staggering is the number of patients for Provenge (sipuleucel-T), the first approved autologous cellular therapy produced by Dendreon. Provenge was approved for metastatic castrate-resistant (i.e., hormone refractory) prostate cancer, an indication for which more than 100,000 patients are treated per year (11, 14). These numbers of batches per year require feasible processing methods to meet the demands of commercial manufacturing for autologous cellular therapies.

If a company used the method described above to manufacture an autologous DC immunotherapy, which could generate years of drug product with one leukapheresis, the process scale would remain the same for commercialization. The company would not need to scale up for autologous therapies. The question would be how to scale out or address the throughput needed for commercialization. Answering this question was the goal of an automation project that developed novel manufacturing equipment. The cornerstone of the developed approach is the incorporation of single-use, functionally closed disposables throughout the process. This method was recognized as crucial for autologous cellular-therapy manufacturing to eliminate cross-contamination concerns and minimize turnaround time between processes, given the throughput needs.

Automated cellular equipment. Personnel began the process by adapting the cellular processing methods to incorporate elutriation and culture bags. Equipment to perform the remaining cellular-processing steps at the scale required for autologous cellular therapies, however, did not exist. Therefore, two similar instruments were designed with the functionality to perform each of these process stages, including processing the autologous plasma collected during the leukapheresis procedure for use in drug-product formulation, and designing custom disposable sets to handle the nuances of each process. The reason for designing two instruments was to create one for plasma and monocyte processing, which are less complex. The second instrument addressed the unique needs and volumes required for mature DCs, such as the electroporation step, including handling the addition of precise volumes of RNA normalized to the DC concentrations, and formulation.

The programming and disposables were designed so that reagents and the cellular inputs for each process were connected to the appropriate disposable set using standard functionally closed disposable manipulation methods (e.g., tube welding). The design incorporated the removal of culture bags or other process outputs using tube-sealing methods. Therefore, the cellular equipment would never be exposed to patient material, and all products and processing will be closed, thus eliminating potential contamination events and product losses resulting from open manipulations. Although it is always crucial to minimize batch failures, the need is greater for autologous cellular therapies, given that each batch is specifically manufactured for a patient, and the patient undergoes leukapheresis to provide the starting cellular material for manufacturing. Because the high variability of the biological starting material influences manufacturing success, personnel need to incorporate any possible additional controls during manufacturing to ensure the generation of product. Processing with single-use, functionally closed disposables and automated methods is a way to significantly improve process control.

The development of the automated cellular equipment and associated disposable sets began with identifying the key processing steps and needs. Moving from elutriation to culture (or freezing to thawing and then culture) required media-exchange steps and the ability to resuspend and distribute cells in a functionally closed disposable component. Similarly, harvesting the DCs, electroporation, culturing following electroporation, and final formulation all required cell-concentration and media-exchange steps. Methods for reliably and accurately addressing this manipulation of cells were developed for even the small-volume manipulations required for electroporation or formulation (i.e., 5–30 mL) while not compromising processing time when large-volume manipulations (large volume for an autologous cellular therapy means as much as 5 L) are required. Overall, the time required for cellular processing using automated equipment is similar to that for the manual methods. The majority of the time required for cellular processing, however, is dedicated to incubation to generate DCs from monocytes, followed by culturing for maturation and culturing for recovery after electroporation.

The biggest challenges to resolve were the management of losses and the maximization of recovery at each step. Minimizing the manipulations for harvesting procedures and media exchanges was critical to ensure that process efficiencies (which were measured by the percentage of monocytes in the initial leukapheresis resulting in RNA-electroporated, mature DCs vialed as drug product) for the automated process were similar to those of manual processing.

Table II: Results for the four initial feasibility runs using the developed automation equipment and functionally closed disposables.

Four initial cellular-feasibility runs performed on prototype automated equipment using the developed disposable sets demonstrated the cellular equipment's ability to perform the processing efficiently with well-controlled, small volumes (see Table II). The formulation of the drug product was on target; it had high DC viability post-thaw. The number of doses produced met expectations; the variability in dose numbers related to the variability in the number of starting monocytes present in the leukapheresis for each run. The drug-product immunophenotyping results for these automated runs confirmed identity and consistency in quality (see Figure 3). Results were similar to those generated in clinical manufacturing (see Figure 2).

Figure 3: Post-thaw immunophenotyping results confirmed the identity and quality of the dendritic-cell products generated in the four initial feasibility runs using the developed automation equipment and functionally closed disposables. CD is cluster of differentiation, and HLA is human leukocyte antigen.

Automated RNA equipment. As the cellular automated equipment and disposable sets were developed, equipment also was developed to amplify autologous RNA from a tumor sample. Though various platforms for automated equipment for nucleic acid manipulations exist, they are generally based on high-throughput methods and open manipulations of plates. For isolating and amplifying nucleic acids for an autologous therapy, this type of equipment could potentially be placed in a barrier isolator to achieve the isolation required for manufacturing. The cleaning requirements between patient samples to prevent cross-contamination, however, would be time consuming. Ensuring that existing automation platforms could withstand vaporous hydrogen-peroxide decontamination between processes would have required additional instrument development. Also, the cleaning validation would have been extremely challenging because the products generated were nucleic acids. The concept, therefore, was to use a functionally closed disposable container for processing, and to design that disposable so that patient material was never in direct contact with the equipment (see Figure 4).

Figure 4: Prototype equipment for automated autologous RNA processing.

The developed RNA disposable container had two main components. The first component was a rigid tray that incorporated all that was necessary to isolate and process the nucleic acid (e.g., pipette tips, reagents, mechanism for nucleic acid isolations and purifications, and PCR plate for all incubation steps). The disposable container also includes spectrophotometer cuvettes and specially designed volumetric cuvettes required to determine the concentration and volume of the isolated and amplified nucleic acids. These cuvettes ensured that the equipment can calculate yields and the volumes required for concentration normalization. They also enabled the equipment to perform the entire process without interruption or data from an outside source.

The second component was a flexible barrier with an incorporated pipette head. This flexible barrier was sealed onto the rigid tray to generate the closed RNA disposable. When closed, the flexible barrier enabled the six-axis robot arm incorporated in the equipment to access all areas in the rigid tray to perform the liquid transfers and other manipulations required for processing. Along with the robotic arm, the automated RNA equipment incorporated the thermal cycler needed for PCR and all incubation steps, as well as the spectrophotometer needed to determine concentration. In initial feasibility runs, the prototype automated equipment and functionally closed disposable container generated amplified RNA comparable to the amplified RNA generated using current manual clinical processing methods, thus demonstrating that this automated concept is appropriate for manufacturing drugs for oncology indications. This concept can be readily adapted to infectious-disease indications to amplify RNA from a viral sample.

Conclusion

These prototypes demonstrated the feasibility of this automated approach. The automated equipment and the associated disposable sets or components provides a platform technology to enable commercial manufacturing of complex autologous DC immunotherapies.

Tamara T. Monesmith is a director of manufacturing and process development at Argos Therapeutics, 4233 Technology Dr., Durham, NC 27704, tel. 919.287.6300, fax 919.287.6301, tmonesmith@argostherapeutics.com.

References

1. J. Harris et al., Biochim. Biophys. Acta 1724 (1-2), 127–136 (2005).

2. I. Tcherepanova et al., PLoS One 1 (e1489), 1–10 (2008).

3. S.R. Bennett et al., Nature 393 (6684), 478–480 (1998).

4. J.P. Ridge, R.F. Di, and P. Matzinger, Nature 393 (6684), 474–478 (1998).

5. S.P. Schoenberger et al., Nature 393 (6684), 480–483 (1998).

6. P.T. Coates et al., Expert Rev. Mol. Med. 4 (3),1–21 (2002).

7. M.A. DeBenedette et al., J. Immunother. 34 (1), 45–57 (2011).

8. I.Y. Tcherepanova et al., BMC Mol. Biol. 9 (90), 1–13 (2008).

9. D.M. Calderhead et al., J. Immunother. 31 (8), 731-741 (2008).

10. M.A. DeBenedette et al., J. Immunol. 181 (8), 5296–5305 (2008).

11. A. Jemal et al., CA Cancer J. Clin. 60 (5), 277–300 (2010).

12. S.A. Karumanchi, J. Merchan, and V.P. Sukhatme, Curr. Opin. Nephrol. Hypertens. 11 (1), 37–42 (2002).

13. National Cancer Institute, SEER Cancer Statistics Review, 1975–2007, S.F. Altekruse et al., Eds. (NCI, Bethesda, MD, 2010).

14. K.J. Pienta and D.C. Smith, CA Cancer J. Clin. 55 (5), 300–318 (2005).