A Scalable Cell-Loading System for Non-Viral Gene Delivery and Other Applications

TECH REVIEW

By JOSEPH C. FRATANTONI, SERGEY DZEKUNOV, SARAH WANG, and LINDA N. LIU

cross many areas of bio- pharmaceutical development, the goal of consis- tently transfecting appropriate quantities of DNA Ainto cells has often been a significant bottleneck. Electroporation — a method of temporarily permeabilizing cell mem- branes by using a short electric pulse Figure 1. Schematic of the MaxCyte System. A cell suspension containing the biomolecule to be loaded (“payload”) is pumped through the processing chamber where an electric field — has gained ground in recent years as is applied synchronously with the flow rate. The loaded cells are collected in an appropriate an effective means of transfection. A container. Additional containers for priming fluid(s) and waste fluids can be incorporated into cell loading system based on electro- the sterile, closed, disposable processing unit. poration has been designed for ex vivo cell modification in a clinical setting and Introduction use of lipid vesicles that encapsulate or for incorporation into cGMP processing coat DNA and are taken up by the cell applications. In the MaxCyte system, All cells are bounded by an outer plas- or fuse with the plasma membrane; cells are suspended in buffer containing ma membrane that maintains a selective and electroporation. The latter was the biomolecule to be loaded are passed surface barrier to the entry of molecules developed based on the observation through the processing chamber, and that is essential to cellular integrity and that brief electric fields result in the then are available for further operations, viability. Many bioactive molecules that formation of transient permeability of cryopreservation, or administration. could potentially modify cellular activ- the plasma membrane through which The system is scalable, employs a sterile, ity and bring therapeutic benefit do not molecules, including DNA, can enter closed, disposable processing unit, and cross this barrier. Although molecules the cell by passive diffusion or electro- no chemical or biological substances are can be specially designed or selected phoretic transfer. If properly applied, added.1,2 that cross this barrier, many small mol- the permeability resolves and cells rap- The computer-controlled MaxCyte ecules, particularly highly charged ones, idly return to a basal state after applica- system (Fig. 1) consists of a power/con- and nearly all macromolecules, includ- tion of the electric field. The MaxCyte trol assembly approximately the size of a ing nucleic acids, do not efficiently enter System is based on electroporation and personal computer, and a disposable pro- cells under normal conditions.3 has the distinct advantage of not requir- cessing unit that attaches to the assembly. Several approaches have been devel- ing the inclusion of a secondary agent, The materials used in the components oped to allow the entry of molecules, (e.g. a lipid, viral package or carrier have been selected for biocompatibility particularly DNA into cells: packag- protein) which can be immunogenic or under the conditions of electroporation. ing a genetically engineered gene con- toxic. The only mechanism used is an struct into a virus or virus-like particle electric field.

Joseph C. Fratantoni, M.D. ([email protected]) is vice president of medical affairs and clinical development; Sergey Dzekunov, Ph.D. is director of research; Sarah Wang is a clinical instrumentation engineer; and Linda N. Liu, Ph.D. is director of biological product development; MaxCyte, Inc., Rockville, MD.

www.bioprocessingjournal.com • November/December 2004 49

JJ36NovDec.indd36NovDec.indd 4949 11/21/2005/21/2005 2:56:512:56:51 PMPM The Process of Electroporation bilayer of the cell membrane. This per- cell stream,cell stream, enabling enabling processing processing Although electroporation is a tech- meability develops in microseconds and volumes ranging from several nology that has been routinely used resolves in seconds to minutes. While microliters to several liters. in research laboratories throughout the a physical “pore” has been observed world for the past 30 years and has under some circumstances, in most situ- • The flow configuration provides been adequately reviewed, some general ations the permeability change is prob- the capability to design a closed comments on the process are appropri- ably related to transient reorientation system, which can be sterilized ate. The primary application has been of membrane phospholipids. During prior to use and can be incorpo- in delivering genetic constructs into the permeable period, both polar and rated into cGMP manufacturing eukaryotic and prokaryotic cells, result- nonpolar molecules of various sizes can processes. ing in non-viral transfection of such diffuse through the permeable areas cells. The process subjects cells to a according to concentration gradients. • The system does not require a con- pulsed electric field for a short duration, In addition, the electric field provides a trolled environment and can be resulting in permeabilization of the lipid force by which charged particles move operated in a routine laboratory, as into the cell (“electrophoretic” mecha- is the case with blood cell separa- nism). Various molecules can be insert- tors. This makes possible cell pro- ed into cells, including drugs, DNA, cessing at the point of care, i.e., at proteins, or other biomolecules. major medical centers that will be the primary sites for cell therapy. The MaxCyte System The cell loading system described in • The MaxCyte System is designed this report, while based on the physical for clinical use. All components process of electroporation, differs from that contact the cells undergoing instruments that have been available for processing are clinical grade. laboratory use in a number of ways: E.g., electrode materials are chosen so that electrolysis does not result • Existing electroporation technol- in toxic metalions in the processed ogy is limited by its ability to treat cell suspension. This requires Figure 2. Approximately 70% less hemoglo- only small suspensions of cells manufacturing electrodes with bin was detected in the Matrigel from mice (usually < mL). MaxCyte materials such as gold. injected with mIL-12 transfected cells (right utilizes a computer-controlled flow column) than the Matrigel from control mice system that synchronizes applica- • The process is rapid, permitting (left column). tion of electric field to the flowing processing of relatively labile cells and reagents.

Table 1. Efficiency and viability of processed cells • The final results of the MaxCyte Results with primary cells (blue shading) and cell lines (yellow shading). “Efficiency” cell loading process depend on is the percentage of observed cells displaying green fluorescence 24 to 48 hours more than the electroporation, and after processing with a plasmid coding for GFP. “Viability” is the percentage of cells extensive attention is given to pre- not displaying propidium iodide (PI) fluorescence (both assayed via flow cytometer). electroporation cell handling, processing buffers and post-electroporation cell management.

Capabilities

The MaxCyte system has been used to load a wide range of cell types using various biologically active molecules such as proteins, antibodies, plasmids, mRNA, and siRNA. A sampling of system performance using transfection efficiency as an indicator is presented in Table 1. Note that the “Efficiency” value is based on the expression of an enhanced green fluorescent protein (eGFP) plasmid, which is a function of cell loading and protein synthesis.

50 BioProcessing Journal • November/December 2004

JJ36NovDec.indd36NovDec.indd 5050 11/21/2005/21/2005 2:58:242:58:24 PMPM Some cells, such as T-lymphocytes, do protein, Ds-Red, highly expressed umes (e.g., 1 L) are processed, samples not readily synthesize protein and a bet- both proteins after the process. Green taken at intervals indicate similar values ter measure of the power of the system and red fluorescence was observed for efficiency and viability. Data are is seen by measuring incorporation of via flow cytometer and fluorescent displayed in the section on large-scale a fluorescence-labeled macromolecule. microscopy. This capability was used processing. With this procedure using, for instance, in developing viral vectors, which Cells are transfected with a high 500 kD FITC-dextran, 95 percent of is discussed in a subsequent section. level of efficiency. High efficiency was cells are loaded. The standard configu- When larger volumes are processed, obtained with the Jurkat cell line, pri- ration’s processing rate is approximately results remain consistent. When mod- mary human lymphocytes, and B-cells 600 million cells per minute.4 A modi- erate volumes (e.g., 50 ml) or large vol- from patients with chronic lymphocytic fied assembly (Table 1) achieves a processing rate that is ten-fold higher than the standard configuration rate. Therapeutic proteins produced by transfected cells function normally in biological systems with no observed toxicities. Following are the results of testing with erythropoietin and IL-12. The two test groups consist- ed of five mice each, and control mice received untreated cells. (a) Mouse fibroblasts were transfected with a DNA plasmid encoding human erythropoietin (a cytokine that stimu- lates production of red blood cells). The cells were subcutaneously injected into Figure 3. Expression of CD40L+IL-2 enhances stimulation of T cells. CLL cells were transfected BalbC/SCID mice (2 x 106 cells/mouse). using plasmids encoding hCD40L or GFP (control). The transformed cells were then incubated with allogeneic T cells and the production of interferon gamma used as an indicator of T The hematocrits (the percentage of red cell stimulation. blood cells in blood) were measured weekly. The erythropoietin functioned normally in vivo. EPO mice hemato- crit levels at 12 and 26 days were 49.9 +/- 3.2% and 48.5 +/-2.6%, respectively; control mice levels were 43.4 +/- 1.6% and 43.2 +/- 2.0%, respectively. By pro- tocol, the mice were sacrificed at day 25. (b) Mouse embryonic fibroblasts were transfected with a DNA plasmid encoding mouse interleukin 12 (IL-12), and subcutaneously injected into the right flank of C3H mice (2 x 106 cells/ Figure 4. Production of lentiviral vectors with the MaxCyte system. “Stock production cell” mouse). Matrigel (a biocompatible refers to a standard cell line (e.g., 293 cells). scaffold) mixed with a stimulator of blood vessel growth (basic fibroblast Figure 5. siRNA introduced by electropora- growth factor [bFGF]) was also injected tion significantly silenced target gene expres- into the right flank. bFGF initiates ves- sion. Human breast cancer MCF7 cells were sel growth and attracted red blood cells transfected with siRNA of GFP (green fluores- into the Matrigel, which was removed cence protein), a control, or with siRNA of an seven days post injection, homogenized, estrogen receptor (ER). The transfected cells and analyzed for hemoglobin concen- were lysed at day one or two post-treatment. Target gene (ER) expression was analyzed by tration. IL-12 inhibits blood vessel for- Western blot with anti-ER antibody. Tubulin mation resulting in fewer red blood cells expression revealed that an equal amount in the Matrigel of treated mice (Fig. 2). of protein was loaded in each lane. ER expression was reduced by more than 90 percent at Cells can be transfected with mul- both time points. Gene array studies showed that changes in the background gene activity tiple plasmids simultaneously. Cells were much lower than with other loading methods (e.g., cationic lipids). (These studies were processed with separate plasmids cod- conducted at the laboratory of Dr. Paul Meltzer, National Human Genome Research Institute, ing for GFP and for red fluorescent National Institutes of Health [NHGRI/NIH].)

www.bioprocessingjournal.com • November/December 2004 51

JJ36NovDec.indd36NovDec.indd 5151 11/21/2005/21/2005 2:58:292:58:29 PMPM leukemia (CLL). Data are presented in indicate cate that that a aretroviral retroviral vector vector selec- provide appropriate preclinical toxicol- the discussion of the CLL study. tively integrates into active gene ogy data on the transgene employed, areas, as compared to non-active but not on the vector per se, in accord Applications areas (introns) of the chromo- with the policy that derived from the some. 5 Non-Clinical Toxicology In Support of Cells that are reportedly difficult to Licensure of Gene Therapies Workshop.8 transfect have been processed success- • Nakai et al. describe studies that There is a Master File with data on the fully using the MaxCyte system. The indicate appreciable integration MaxCyte system at CBER/FDA. This system has been used in the prepa- with adeno-associated viral vec- file has been referenced for IND appli- ration of specific cell-based therapeu- tors. Such integration was previ- cations. tics, for incorporation into industrial ously thought to occur only rarely.6 bioprocessing, and for cell-based drug Example 1. Immunotherapy for Chronic discovery and development. Examples Clearly, alternative approaches are Lymphocytic Leukemia (CLL) of these applications are presented in needed.7 Electroporation provides an CLL, a B-cell malignancy, is the most Figure 3. alternative that simplifies preparation common leukemia in adults. Leukemic and testing logistics, and treads a less cells from patients do not stimulate a Cell-Based Gene Therapy onerous regulatory path. significant immune response, but if CLL Electroporation enables non-viral This has specifically been demon- cells express the protein CD40L, the gene delivery in an ex vivo mode while strated by the prompt clearance of appli- immune response is greatly enhanced avoiding persistent concerns about the cations through the Recombinant DNA and leads to diminished tumor mass. safety of viral vectors. Recent literature Advisory Committee and FDA, with Clinical studies using a mouse gene supports those concerns: regard to gene delivery aspects, for the to transfect autologous CLL cells with CLL trial described. In general, users the gene for CD40L showed promis- • Wu et al. present data which indi- of the MaxCyte system are required to ing results.9 Other studies have shown that the addition of IL-2 enhances the immune effects. The MaxCyte system delivers the genes for CD40L and IL-2 into CLL cells with very satisfactory efficiency and cell viability, and the cellular product is ready for release soon after processing. A Phase I/II clinical study has been developed in collaboration with the Center for Cell and Gene Therapy at Baylor University. In the study, blood is obtained from patients via standard techniques, and CLL cells are separated and processed using the MaxCyte system (to a sufficient volume to yield Figure 6. Screening candidate genes. B16 F10 melanoma cells were transiently transfected several doses, plus a sample for testing). 5 with various DNA plasmids. After 24 hours, mice were injected subcutaneously with 1 x 10 The blood is then frozen and stored. Six gene-modified or unmodified (no DNA) tumor cells. Tumor volumes were calculated bi-weekly doses are administered to patients every for three to four weeks (day 15 is shown). one to two weeks, while clinical and laboratory variables are monitored for Table 2. Data generated with the very large-scale configuration. safety. Efficacy will be demonstrated The model system chosen for the transfection of Jurkat cells was a standard by measuring the decreases in CLL cell plasmid carrying full-length cDNA encoding for eGFP driven by a cytomegalovi- counts and lymph node size. rus (CMV) promoter. Jurkat cells (2.5 x 1010 in a final total volume of one L) were washed and suspended in electroporation buffer and eGFP plasmid was added (100 µg/ml DNA). The flow system for this demonstration was equipped with a Example 2. A Cell-Based Therapeutic sample port and 1-ml samples were withdrawn at 100-ml intervals. The fluores- Approach for Pulmonary Arterial cence-activated cell sorting (FACS) results on these samples (direct gating and Hypertension. counting) are summarized and compared with four static experiments. MaxCyte is working with Northern Therapeutics to develop a cell-based gene therapy for pulmonary arterial hypertension (PAH), a progressive and often fatal disease. Using a rat model of PAH, both rat and human dermal fibroblasts

52 BioProcessing Journal • November/December 2004

JJ36NovDec.indd36NovDec.indd 5252 11/21/2005/21/2005 2:58:352:58:35 PMPM as well as erythroid progenitor cells were a large scale (discussed below) makes it single experiment is a demonstration of loaded with a plasmid encoding endo- suitable for all stages of clinical research the system’s volume capability and that thelial nitric oxide synthase (eNOS). The and for commercialization. Projects conditions were not optimal.) transfected cells can be cryopreserved with partners who are developing virus- with maintenance of eNOS transgene based therapeutics are underway. Summary expression upon thawing. To test in vivo Cell-based Drug Discovery and efficacy, Fisher 344 rats, which had been Development. The MaxCyte system has Electroporation, as used in the treated with monocrotaline to induce been used in a variety of applications: MaxCyte system, can effectively and PAH, received an injection through the reproducibly load cells with exogenous right jugular vein of either 1 x 106 • Loading cells with immunoglobu- DNA or other molecules. Because the syngeneic fibroblasts that had been lins to block selected functions. processing module is a closed, sterile transfected with an eNOS plasmid or IgG was loaded with high effi- unit, it can be used for point-of-care loaded with a null control plasmid.10 ciency, and specific functions were therapeutic protein production and can Right ventricular systolic pressure, observed to decrease. be incorporated into cGMP processing measured 25 days after treatment, was systems. The system’s use in FDA- reduced in animals receiving eNOS cell- • Rapid production of proteins from approved clinical studies should pave based gene therapy (38.5 ± 3.3 vs. 52 ± candidate genes of potential inter- the way for its large-volume capability to 3.9 mmHg [p=0.013, reversal]). A clini- est using transient transfection of be used in a wide range of bioprocessing cal study is in preparation. production cell lines. This tech- applications. Bioprocessing. The properties of nique can yield milligram quanti- the MaxCyte system that make it suit- ties of protein, allowing pilot test- able for production of selected bio- ing before embarking on viral REFERENCES pharmaceuticals (biocompatible mate- vector-based transfection for long- rials, scalability, closed and sterile term production of promising 1. Fratantoni JC, Dzekunov S, Singh V, Liu LN. A Non- Viral Gene Delivery System Designed for Clinical Use. processing unit) allow incorporation proteins. Cytotherapy 2003: 5; 208–210. of the system into cGMP manufactur- 2. Li L-H, Shivakumar R, Feller S, Allen C, Weiss JM, ing processes. The system’s capabil- • Loading cells with inhibitory RNA Dzekunov S, Singh V, Holaday J, Fratantoni JC, Liu ity to load two or more macromol- to provide model systems for test- L. Highly efficient, large volume flow electroporation. ecules simultaneously has been used ing potential therapeutics. An Technol Cancer Res Treatment 2002: 1; 341–350. 3. Stephens DJ and Pepperkok R. The many ways to to produce viral particles by loading example, using siRNA, is shown in cross the plasma membrane. Proc Natl Acad Sci 98: plasmids encoding the viral compo- Figure 5. 4295–4298, 2001. nents into routine production cell lines. 4. Richman E, Liu LN, Fratantoni JC. Loading bioac- Scaling up production of non-rep- • In vivo evaluation of candidate tive molecules into living cells. Gen Eng News 2002; licating viral vectors is a major hurdle genes by transfecting appropriate 22:50–1. 5. Wu X, Li Y, Crise B, Burgess SM. Transcription start for large gene therapy clinical trials. cell lines and observing the effects regions in the human genome are favored targets for 11 Transient, simultaneous transfection of in animals. An example of test- MLV integration. Science 2003: 300; 1749–51. cells with multiple plasmids results in ing potential antineo plastics is 6. Nakai H, Montini E, Fuess S, Storm TA, Grompe the production of vectors and decreases displayed in Figure 6. M, Kay MA. AAV serotype 2 vectors preferentially the possibility of viral-genome recom- integrate into active genes in mice. Nat Genet. 2003:4; 297–302. bination. Current transient transfection Large-Scale Processing 7. Nishikawa M and Huang L. Nonviral vectors in the methods (e.g., CaPO4) allow production new millenium: delivery barriers in gene transfer. Hum of small volumes of viral vectors, but the A configuration of the MaxCyte sys- Gene Ther 12: 861–870, 2001. process can introduce inconsistencies tem has been developed that can process 8. Report of the ASGT-FDA Non-Clinical Toxicology In from lot to lot, thus raising regula- cells swiftly for very large-scale applica- Support Of Licensure Of Gene Therapies Workshop, Arlington, VA, March 2003. tory issues. Furthermore, precipitation tions. Data have been generated for the 9. Weirda WG, Cantwell MJ, Woods SJ, RAssenti LZ, of CaPO4 interferes with downstream transfection of one liter of cell suspen- Prussak CE and Kipps TJ. CD40-ligand (CD154) gene purification and the concentration of sion, containing 2.5 x 1010 cells. The therapy for chronic lymphocytic leukemia. Blood 96: viral particles. The method presented configuration used to produce most of 2917–2924. 2000. here can co-transfect large volumes of the data displayed in Table 1 can process 10. Campbell AIM, Kuliszewski MA and Stewart DJ. Cell-based gene transfer to the pulmonary vasculature. cells, thereby scaling up production to 1 to 100 ml of cell suspension at a rate of Am J Respir Cell Mol Biol 21: 567–575, 1999. 8 research and clinical levels. The basic several ml/minute, or as high as 6 x 10 11. Weiss JM, Shivakumar R, Feller S, Li LH, Hanson elements of the process are shown sche- cells/minute. The very large-scale con- A, Fogler WE, Fratantoni JC, Liu LN. Rapid, in vivo, matically in Figure 4. figuration can process several liters of evaluation of anti-angiogenic and anti-neoplastic gene The system has demonstrated the cell suspension in a single operation at products by non-viral transfection of tumor cells. Cancer Gene Therapy 11: 346–353, 2004. capability to produce lentivirus- and a rate of 200 ml/minute. Summary data alphavirus-based therapeutics on a lab- with this configuration are displayed in oratory scale. Its ability to operate on Table 2. (It should be noted that this

www.bioprocessingjournal.com • November/December 2004 53

JJ36NovDec.indd36NovDec.indd 5353 11/21/2005/21/2005 2:58:392:58:39 PMPM