By Ajay Lajmi, Senior Scientist, Pall Life Sciences[view feature in pdf format]
Antisense Oligonucleotides (AO) belong to a class of drugs that are currently being developed for the treatment of fatal diseases, and differ from small molecule therapeutics in their mechanism of action. Once these drugs become commercially available, they will expand the growing repertoire of biotherapeutics. These drugs are hailed for their specificity in targeting proteins that cause a wide range of ailments. AO work by hybridizing to messenger RNA to inhibit the production of proteins that cause any number of diseases, including cancer, viral infections, respiratory, inflammatory, metabolic, and cardiovascular.
With more than 25 different antisense molecules currently in various stages of clinical trials, the potential for this market is tremendous. However, current processing methods have limited capacity to support full-scale commercial production. Ion exchange membrane chromatography overcomes this challenge by providing high throughputs (i.e. fast flow rates and high dynamic binding capacities), reducing processing steps, while maintaining high purity levels to meet the full-scale production demands required to fully tap the commercial potential of antisense drugs.
Antisense TechnologyThe mechanism by which AO works relies on Watson-Crick base pairing of nucleotides. When a gene is transcribed, the resulting messenger RNA (mRNA) contains the “sense” sequence, and these are the genetic instructions that are translated into the manufacture of a protein. If the sequence of mRNA that is required to make the protein is known, a complementary, or “antisense” sequence of the mRNA that theoretically binds to the “sense” sequence, prevents it from functioning. Blocking the “sense” sequence, blocks the ability of mRNA to manufacture the protein being targeted. Therefore, once the gene involved in disease pathogenesis has been identified, short oligonucleotide molecules could be synthesized that would shut down the expression of those genes. These molecules are referred to as Antisense Oligonucleotides or AOs.
Antisense OligonucleotidesAO are short (17- to 27-nucleotides long) synthetic DNA molecules with phosphorothioate instead of the naturally occurring phosphodiester linkages, for in vivo nuclease resistance. AO including antisense deoxyoligonucleotides, siRNA, or CpG Immunostimulatory deoxyoligonucleotides, are synthesized on polymeric solid supports by covalently linking individual nucleic acids from the 3- to 5-end. These drugs differ from small-molecule pharmaceuticals in that instead of binding to a pocket in the protein to block certain symptoms, they prevent the protein from being expressed altogether. This minimizes side effects, and increases the effectiveness of the drug.
Ion exchange chromatography (IEX) takes advantage of several aspects unique to antisense drugs. Since the desired product in antisense molecules displays the highest negative charge, it binds most tightly to the pores that contain positively charged ion exchange groups. High concentration of sodium chloride is then used to elute the desired product from the ion-exchange bed. One factor contributing to the effectiveness of IEX in the production of antisense molecules is its ability to separate impurities that have a very close charge-to-size ratio. Here, the phenomenon of Sample Self Displacement further increases resolving power.
Displacement ChromatographyIn displacement chromatography, a sample is loaded on the ion-exchange media, followed by a solution of a molecule that binds to the media more tightly than any of the sample components. This displacer pushes the bound sample molecule through the media during which time a displacement train of adjacent bands in order of increasing binding strength is formed. Once all the sample molecules have been displaced off, the displacer is eluted and the media re-equilibrated with the loading buffer for the next purification cycle.In sample-self displacement chromatography of phosphorothioate oligonucleotides, the full-length, fully thioated oligonucleotide component in the sample that has the highest charge, hence the highest affinity for the stationary phase (the chromatography media), acts as the displacer. Sample loading is a critical parameter that affects the purity and recovery of the product. As the operating conditions vary for every process, this factor needs optimization prior to scale-up.
Ion Exchange Membrane Chromatography vs. Reversed Phase MethodsThe antisense molecules that are currently in pre-clinical and clinical trials are either purified by reversed phase chromatography (RPC) or a combination of RPC and IEX. The RPC method utilizes strong hydrophobic interactions to capture target molecules. However, RPC is inefficient for full-scale production, due to the presence of diffusive pores that result in low throughputs. Membrane chromatography media are pre-packed thereby eliminating costs related to packing and packing validation of RPC media. RPC also requires the use of organic solvents, which are both toxic and flammable, creating significant capital costs in storage, handling and disposal of these chemicals in bulk quantities.
Suitable for lab, pilot and full-scale production, ion exchange membrane chromatography reduces processing bottlenecks associated with traditional RPC conditions by virtue of its larger pore size, convective flow pattern, optimized module design, and capture protocol. As a result, membrane IEX can provide faster flow rates and higher dynamic binding capacities, that is, higher throughputs in the range of 10-100fold compared to RPC.
In studies published in the Journal of Chromatography1 Prof. Edwin Lightfoot, Ph.D. of the University of Wisconsin demonstrated that MUSTANG® chromatography membrane provides the same resolution as 15 µm beaded media. While the 15 µm beaded media provide high resolution, they lead to high pressures resulting in low operating flow rates that are unsuitable for large-scale production. Process-scale chromatography media typically range between 30 µm and 120 µm beads for that reason however, these larger beads do not provide the resolution needed to achieve the desired level of product purity for antisense drugs. Membrane IEX media with a 0.8 µm pore size achieve more than ten times the throughput of column chromatography methods, such as reversed phase techniques. In addition to not meeting the growing demand for higher purification throughputs at production scale, RPC columns are prone to channeling during or after packing, resulting in additional capital costs.
Diffusion vs. Convective FlowIn resin chromatography, the target molecules must diffuse into and out of the pores during the binding and elution stages. In membrane chromatography, diffusion limitations are so low that they have no discernable effect on the binding process. Figure 1 demonstrates the difference in pore structures for binding between packed columns and ion exchange membranes. In the packed column (left) the white areas represent resin beads, and blue areas represent flow paths. On the membrane (right) there is much more blue than white, illustrating the greater availability of surface areas for binding, which translates into a much higher flow rates and dynamic binding capacities.
Antisense Molecules and Membrane Properties: A Natural PairingMembrane chromatography uses a micro porous membrane with ion exchange groups in the membrane pores to capture target molecules by absorption. MUSTANG Q membrane using a proprietary polyethersulfone formulation provides low, non-specific protein absorption, which helps enhance performance and efficiency for specific chromatography processes, such as the purification of antisense oligonucleotides. This membrane uses Quaternary Ammonium (Q) chemistries to capture molecules by anion exchange, and features a large mass transfer surface, which is ideal for binding of biopolymers. The three-dimensional structure of membrane pores creates tortuous channels containing ion exchange groups on the surface of the membrane.
Module CharacteristicsThe design of the ion exchange chromatography module also plays a role in increasing flow rates. A lower plate height compared to a 15 µm beaded media column, uniform flow distribution, and an optimized header space in the module housing all contribute to higher separation efficiencies. Membrane modules that allow up to 80 membranes to be stacked together within stainless steel housings simulate a short, wide chromatography column. It also provides a sturdy design that can withstand pressures limited to a maximum of 100 psi, and about a hundred cleaning cycles with reproducible separation efficiencies.
Maximizing Sample Self DisplacementSample Self Displacement is a function of the chromatography process that serves to improve resolution. While this phenomenon is common to all chromatography techniques, characteristics of membrane ion exchange methods further speed processing with high resolution. Sample Self Displacement works on the premise that the fully-thioated oligonucleotide with the highest affinity for the stationary phase serves as the displacer. In other words, as the sample is transported through the pores in the ion exchange membranes, the molecule with the highest negative charge will bind to the stationary phase, displacing molecules that are lower in negative charge, thus forming adjacent zones of high concentration molecules of the same ionic charge. The molecules in each zone differ from those in the adjacent zones by the number of ionic charges. Since the target product in this case has the highest charge, these molecules are amassed together at the end of the product train. Impurities are separated from the product at the leading edge of the train, resulting in a highly purified fraction with improved separation of the target molecule from its closely related impurities. Profs. Steve Cramer, Casba Horvath and others2 have described this phenomenon in detail. The ability of ion exchange membranes to process at higher throughputs enables drug developers to take full advantage of Sample Self Displacement. However, the loading factor must be optimized in order to realize this effect.
Loading Optimization ExampleTo determine the loading capacity at which Sample Self Displacement occurs, initial fractions of the gradient elution peak of Isis Pharmaceuticals’ ISIS 2302 (Alicaforsen) were analyzed. This molecule is currently in Phase III trials for Crohn’s disease. Based on the analysis of these fractions, optimum loading capacity using the 10 ml MUSTANG® Q module was shown to be between 270 mg and 300 mg. In addition to using strong anion exchange chromatography (SAX) to analyze purification of the sample fractions, mass spectrometry was used to provide the most accurate and sensitive measure of purity.
Establishing a Primary Reference Standard for ISIS 2302In order to establish a benchmark for the highest level of purity, ISIS 2302 active pharmaceutical ingredient (API) was re-purified (polished) using MUSTANG Q membrane chromatography (Fig. 2).
This measure of purity was qualified as a primary reference standard (PRS). The purity of subsequent manufacturing batches is compared with the PRS. Following are the results on 10 ml and 1-liter MUSTANG? Q modules:
• 300 mg ISIS 2302 API of 87.7 % purity on a 10 ml MUSTANG? Q module produced a 95 percent pure oligonucleotide when analysis was performed by electrospray ionization mass spectrometry (ESI-MS).
• 30 g ISIS 2302 API of 88 % purity on 1 liter MUSTANG? Q module produced a 99 % pure oligonucleotide when analyzed by HPLC and 95 % purity.
• 177 g ISIS 2302 API was purified on the 1 liter strong anion exchange membrane chromatography unit in eight cycles including two recycles from side fractions in 82 % overall yield and 95 % product purity by ESI-MS3
Minimizing Oligonucleotide Purification StepsIn RPC, the DMT (dimethoxy trityl) protecting group not only functions to prevent the oligonucleotide from undergoing side reactions during synthesis, but it also gives it the necessary hydrophobic characteristic to bind with the reversed phase media. The molecule is cleaved from the solid support with ammonia, with the unreacted ammonia being evaporated under reduced pressure. Detritylation is then performed on either the RPC column or on the RPC purified product to remove the DMT group. The oligonucleotide molecule is then precipitated out by ethanol and lyophilized (freeze dried). After which, the oligonucleotide is dissolved in a mixture of water and acetonitrile or methanol (organic solvent) and loaded onto the reverse phase chromatography column. A gradient of increasing amount of organic solvent is then used to elute off the molecule, and recover the product.
Membrane IEX eliminates the need for ethanol precipitation, and lyophilization steps prior to chromatography. Time-consuming column packing is also omitted from the production process, as is the need for organic solvents, which are very expensive to re-purify, and must be used in an explosion proof facility. Because IEX uses a high salt elution buffer, it requires a final desalting step using ultrafiltration or diafiltration.
It may be argued that IEX introduces an additional unit operation that involves subsequent ultrafiltration and diafiltration (UF/DF), hence increases the purification cycle time. However, the total membrane-based purification cycle time (membrane IEX and UF/DF) can be maintained close to that of RPC by increasing the total filtration area in the UF/DF step. Thus, based on the amount of AO processed per hour (throughput), membrane-based AO purification proves to be more efficient than RPC. A proposed scheme for such a process is shown in Figure 3.
Linear Scale-UpLinear scalability is a critical consideration when selecting ion exchange membrane chromatography media because it will determine the ease with which a product is transitioned from early-stage clinical production to full-scale commercial manufacture. The process scale membrane IEX modules are linearly scalable in10 ml, 100 ml and 1 liter bed volumes (BV). These modules can be operated at flow rates between 1 BV/min and 5 BV/min at 15 to 75 psi with a total cycle time of 20-30 minutes compared to 0.1 CV/min with RPC media. These IEX membranes have a dynamic binding capacity of 20-30 mg of oligonucleotide per ml of membrane depending on the nature of oligonucleotide compared to 7 mg/ml with RPC media.
Summary of Benefits of Membrane IEX in Antisense Drug ProductionMembrane ion exchange chromatography provides high throughputs of antisense oligonucleotides due to its convective flow and high binding capacities compared to reversed phase chromatography. Cost savings related to column packing, packing validation as well as storage and disposal of organic solvents, give membrane ion exchange methods a significant economic and processing advantage compared to the disadvantages of reverse phase chromatography4. In addition, Membrane IEX can remove the crude DMT group from the feed without any additional steps and the superior resolving ability of membrane IEX ensures that multiple impurities are removed or minimized in the AO sample in one step.
Planning for Tapping a Successful Antisense Drug MarketImprovements in identification of more specific targets as well as in increasing bioavailability of antisense oligonucleotides is expected to result in significant advancements in the development of this class of biopharmaceuticals. As these advancements demonstrate the viability of these drugs, biopharmaceutical companies will need to rapidly plan for future demand. Ion exchange membrane chromatography media have proved to be valuable purification tools, and the new Membrane IEX technology can reduce bottlenecks, and costs, while improving throughputs in downstream processing of antisense oligonucleotides. This technology, as well as the near-future development of manufacturing scale Membrane IEX units, will also enable industry to tap the full potential of these drugs and to meet the projected ton scale per year demand of AO in the not too distant future.
1 Teeters, M.A.; Root, T.; Lightfoot, E.N. J. Chromatogr. A. 944 (2002) 129-139.
2 Cramer, S. M. and Horvath, Cs. Prep. Chromatogr. 1 (1988) 29-49.
3 Lajmi A.R., Schwartz L., and Sanghvi Y. Org. Process Res. Dev. 8 (2004) 651-657.
4 Warner, T. N. and Nochumson, S. BioPharm Int. 1 (2003) 58-60.
About the author: Dr. Lajmi (Ph.D.) is a Sr. Research Scientist at the Biopurification Laboratory of Pall Life Sciences, Pensacola, Florida responsible for applications research in protein, viral vector and oligonucleotide purification. In the past four years at Pall, he has focused on developing applications for oligonucleotide and gene therapy virus vector purification using membrane chromatography. He is also involved in chromatographic process development of proteins, and plasmid DNA at Pall. Prior to joining Pall Corp. Dr. Lajmi earned his Ph.D. from New York University in Bioorganic Chemistry. He later followed with postdoctoral research in DNA-binding protein expression, protein purification and protein refolding at the University of Pittsburgh. Dr. Lajmi has over 10 years experience in chromatography.