Thesis (M.S.) -- Worcester Polytechnic Institute. / Keywords: Protein separation. Includes bibliographical references (p. 76-79).

Membrane technology has been actively used as a separation tool in the chemical, environmental, and biopharmaceutical industries for several decades. As membrane quality requirement in the industry has increased, efforts have been directed towards enhancement in mechanical strength, chemical durability and functionality of membranes. One of the approaches for membrane quality enhancement is based on the combination of hydrogel technology with membrane technology. This thesis focused on the application and development of hydrogel based membranes, notably hydrophilized PVDF (polyvinylidene fluoride) membrane for hydrophobic interaction membrane chromatography; the fabrication of paper-hydrogel composite membranes for membrane chromatography; development of a technique for coating alginate (a natural hydrogel) on the outer surface of a hollow fiber membrane for potential application in bioreactors and the use of hollow fiber membranes as mold for fabrication calcium alginate fibers for biomedical and tissue engineering applications. A membrane chromatography-based polishing technique was developed for removing leached protein-A and aggregates from monoclonal antibody (mAb). A commercial synthetic membrane that is known to be hydrophilized by hydrogel grafting was employed to develop this polishing process that resulted in highly pure mAb, free from aggregates and protein-A. This mAb polishing technique could easily be integrated with a hydrophobic interaction membrane chromatography based mAb purification process. A paper-hydrogel composite membrane was developed as an inexpensive alternative to commercial synthetic membranes used for carrying hydrophobic interaction membrane chromatography. Poly(N-vinylcaprolactam) or PVCL hydrogel was coated on Whatman filter paper to prepare these membranes. These environment responsive membrane which responded to changes in salt concentration, gave excellent fractionation of multi-component protein mixtures. As case study, a mixture of immunoglobulin G, human serum albumin and insulin was fractionated. A technique for modifying the surface of synthetic hollow fiber membranes with alginate (a natural hydrogel) was developed. This manner of surface modification led to the improvement in membrane mass transport. The alginate was cross-linked on the outer surface of the membrane by diffusion of the cross-linker (calcium ions) through the membrane pores. The calcium alginate coating layer was characterized by optical and transmission electron microscopy, contact angle measurement, hydraulic permeability measurement and by examining solute transport. Hollow and solid calcium alginate fibers were fabricated using a novel hollow fiber membrane based moulding technique. The pore present on the hollow fiber membrane served as the reservoir for the calcium chloride solution with cross-linked the alginate within the lumen. The calcium alginate fibers produced were characterized by optical, transmission electron, and scanning electron microscopy. Cell immobilization experiments were carried out to demonstrate biocompatibility and potential for tissue engineering applications. / Thesis / Doctor of Philosophy (PhD)

Membrane chromatography is a promising process for the isolation, purification, and recovery of proteins, enzymes, and nuclear acids. Comparing with traditional beads column chromatography, membrane chromatography can faster, easier and cheaper to mass-produce. And also, it is easy to set up and scale up. In this thesis, we are trying to study the performance of membrane chromatography, and the mixture of HSA and chicken egg white is used as an example. We are investigating the purification of Human serum albumin (HSA) from chicken egg white in terms of precondition, dilution, purification method, product recovery, product purity and product cost. HSA, is a very important clinical protein. In order to obtain low cost, high efficiency and less risk HSA, recombinant DNA technology is used. Many kinds of host organism have been used to produce recombinant HSA (rHSA). In this thesis, a kind of ion-exchange membrane (Mustang Q membrane capsule) chromatography was used. The membrane capsule is disposable because it is designed for use in pharmaceutical production. For this project, a cleaning method was used which made the membrane capsule reusable. Washing with 4 mL 1 M NaCl and 4 mL NaOH was sufficient for this purpose. Since the egg white protein solution was very viscous, it needs to be diluted before loaded on FPLC. Dilute experiment was done to find the best dilution level. In this thesis, we found that 5 times dilution was best not only for high efficiency but also for FPLC operation. After getting the basic conditions, some purification experiments were done to find the optimal operation condition to purify HSA form chicken egg white protein solution by changing buffer pH, salt concentration in elution buffer and gradient used to elute proteins. The best purification condition for loading buffer is Tris-HCl buffer A (4.75g/L, pH 9.5) and the elution buffer is Tris-HCl buffer A + 0.2M NaCl. The purity of HSA recovered was 93% on the Mustang Q membrane capsule at 1 ml/min when the mixture of HSA and chicken egg white was diluted 10 times. And the yield was 85%. The impurity is probably ovoglobulin as suggested by the result of SDS-PAGE, whose molecular weight is close to 40kd. To characterize the separation capability of the Mustang Q membrane capsules, equilibrium adsorption and breakthrough curve studies were made using bovine serum albumin (BSA). 1mg/mL BSA solution was used to get the breakthrough curve with different flow rate ranging from 1 to 4 ml/min. With a flow rate is 1 ml/min, breakthrough curve were obtained with different concentrations of BSA ranging from 1 to 16 mg/mL. The dynamic binding capacity was found to be from 9.1 to 119.1 mg/mL. The equilibrium adsorption isotherm showed Langmuir isotherm behavior with dissociation constant and a maximum adsorption capability. According to the result of isotherm adsorption, a multi-plate mathematical model was used to get the theoretical breakthrough curve. By fitting the theoretical breakthrough curve to the experimental breakthrough curve, constants in the multi-plate model were obtained and were used to estimate the axial dispersion coefficient of the membrane capsule. The estimated axial dispersion coefficient of 2.45*10-6 cm2/s is very small which means that the axial ispersion is not significant. The adsorption process is therefore controlled by radial radius dispersion or film dispersion.

Whey protein fractionation is an important industrial process that requires effective large-scale processes. Although packed bed chromatography has been used extensively, it suffers from low processing rates due to high back-pressures generated at high flow rates. Batch chromatography has been applied but generally has a low efficiency. More recently, adsorptive membranes have shown great promise for large-scale protein purification, particularly from large-volume dilute feedstocks. A new method for producing versatile adsorptive membranes by combining membrane and chromatographic resin matrices has been developed but not previously applied to whey protein fractionation. In this work, a series of mixed matrix membranes (MMMs) were developed for membrane chromatography using ethylene vinyl alcohol (EVAL) based membranes and various types of adsorbent resin. The feasibility of MMM was tested in bovine whey protein fractionation processes. Flat sheet anion exchange MMMs were cast using EVAL and crushed Lewatit® MP500 (Lanxess, Leverkusen, Germany) anion resin, expected to bind the acidic whey proteins β-lactoglobulin (β-Lac), α-lactalbumin (α-Lac) and bovine serum albumin (BSA). The MMM showed a static binding capacity of 120 mg β-Lac g⁻¹ membrane (36 mg β-Lac mL⁻¹ membrane) and 90 mg α-Lac g⁻¹ membrane (27 mg α-Lac mL⁻¹ membrane). It had a selective binding towards β-Lac in whey with a binding preference order of β-Lac > BSA > α-Lac. In batch whey fractionation, average binding capacities of 75.6 mg β-Lac g⁻¹ membrane, 3.5 mg α-Lac g⁻¹ membrane and 0.5 mg BSA g⁻¹ membrane were achieved with a β-Lac elution recovery of around 80%. Crushed SP Sepharose™ Fast Flow (GE Healthcare Technologies, Uppsala, Sweden) resin was used as an adsorbent particle in preparing cation exchange MMMs for lactoferrin (LF) recovery from whey. The static binding capacity of the cationic MMM was 384 mg LF g⁻¹membrane or 155 mg LF mL⁻¹ membrane, exceeding the capacity of several commercial adsorptive membranes. Adsorption of lysozyme onto the embedded ion exchange resin was visualized by confocal laser scanning microscopy. In LF isolation from whey, cross-flow operation was used to minimize membrane fouling and to enhance the protein binding capacity. LF recovery as high as of 91% with a high purity (as judged by the presence of a single band in gel electrophoresis) was achieved from 150 mL feed whey. The MMM preparation concept was extended, for the first time, to produce a hydrophobic interaction membrane using crushed Phenyl Sepharose™ (GE Healthcare Technologies, Uppsala, Sweden) resin and tested for the feasibility in whey protein fractionation. Phenyl Sepharose MMM showed binding capacities of 20.54 mg mL⁻¹ of β-Lac, 45.58 mg mL⁻¹ of α-Lac, 38.65 mg mL⁻¹ of BSA and 42.05 mg mL⁻¹ of LF for a pure protein solution (binding capacity values given on a membrane volume basis). In flow through whey fractionation, the adsorption performance of the Phenyl Sepharose MMM was similar to the HiTrap™ Phenyl hydrophobic interaction chromatography column. However, in terms of processing speed and low pressure drop across the column, the benefits of using MMM over a packed bed column were clear. A novel mixed mode interaction membrane was synthesized in a single membrane by incorporating a certain ratio of SP Sepharose cation resin and Lewatit MP500 anion resin into an EVAL base polymer solution. The mixed mode cation and anion membrane chromatography developed was able to bind basic and acidic proteins simultaneously from a solution. Furthermore, the ratio of the different types of adsorptive resin incorporated into the membrane matrix could be customised for protein recovery from a specific feedstream. The customized mixed mode MMM consisting of 42.5 wt% of MP500 anionic resin and 7.5 wt% SP Sepharose cationic resin showed a binding capacity of 7.16 mg α-Lac g⁻¹ membrane, 11.40 mg LF g⁻¹ membrane, 59.21 mg β-Lac g⁻¹ membrane and 6.79 mg IgG g⁻¹ membrane from batch fractionation of 1 mL LF-spiked whey. A tangential flow process using this membrane was predicted to be able to produce 125 g total whey protein per L membrane per h.

Viruses have emerged as a new class of biotherapeutics used as vectors in gene and cell therapies, vaccines, and as oncolytic agents in novel cancer immunotherapies. While these new and potentially curative new therapies bring great promise for patients, the large-scale purification of viruses is hampered by complicated unit operations, poor overall yields, and high costs. Membrane chromatography (MC) is one of the most ideal options for the removal of host-cell impurities in virus manufacturing. Centred on developing and improving MC processes for virus purification, this thesis focuses on different aspects of downstream processes that are directly related to MC. It describes the development of the first hydrophobic interaction MC process for the purification of vesicular stomatitis virus as a scalable method for the removal of host-cell protein and DNA. It also describes the development of MC for adenovirus purification, and how device design and membrane type impact the resolution; here, the novel laterally-fed membrane chromatography (LFMC) was proven to provide higher resolution than conventional MC devices, and allowed for the first direct comparison between the most popularly used membranes in virus manufacturing – Sartobind Q and Mustang Q. Beyond MC, this thesis also addresses how other downstream unit operations contribute to the final purity. Through an integrated study optimizing clarification, DNA digestion, and MC simultaneously, significant improvement in adenovirus purity was obtained. Finally, the collection of experimental results was used to model complete adenovirus production processes using BioSolve Process and determine the cost-of-goods (COG) of manufacturing for clinical applications. Through simulations of multiple scenarios, critical process parameters were identified and can serve as a guide for future process development decisions. It is anticipated that the contributions herein described will help address critically outstanding questions related to virus purification and thus enable the development of the economical processes for various manufacturing scales. / Thesis / Doctor of Philosophy (PhD) / Certain viruses can be used for human benefit and there are now more than a dozen approved therapies worldwide that use a virus as the main therapeutic agent or as the vector to instruct the patient’s cells to fight cancer and other diseases. The area keeps growing as thousands of other clinical trials continue to be conducted. One of the main challenges that can inhibit patient access to these ground-breaking new options is related to difficulties in producing and purifying enough virus. This study tackles the virus purification challenge by applying and improving membrane chromatography (MC), a promising and scalable technique where virus and impurities are separated based on how differently they interact with a membrane. Different experimental and modelling and simulation tools were applied to optimize MC and other directly-related steps of the production process. The findings in this study can contribute to the development of new virus-based therapeutics so they can reach patients in safe, effective, and affordable ways.

Downstream purification often represents the most cost-intensive step in the manufacturing of recombinant proteins. Conventional purification processes are lengthy, technically complicated, product specific and time-consuming. To address this issue, herein we develop a one step purification system that due to the nature of the non-selective secretion system and the versatility of ion-exchange membrane chromatography can be widely applied to the production of many recombinant proteins. This was achieved through the integration of the intrinsically coupled upstream, midstream and downstream processes, a connection that is rarely exploited. A bioprocess for effective production and purification of penicillin G acylase (PAC) was developed. PAC was overexpressed in a genetically engineered Escherichia coli strain, secreted into the cultivation medium, harvested, and purified in a single step by anion-exchange chromatography. The cultivation medium developed had a sufficiently low conductivity to allow direct application of the extracellular fraction to the anion-exchange chromatography medium while providing all of the required nutrients for sustaining cell growth and PAC overexpression. It was contrived with the purposes of (i) providing sufficient osmolarity and buffering capacity, (ii) minimizing ionic species to facilitate the binding of extracellular proteins to anion-exchange medium, and (iii) enhancing PAC expression level and secretion efficiency. Employing this medium recipe the specific PAC activity reached a high level of 487 U/L/OD600, with more than 90% was localized in the extracellular medium. Both, the osmotic pressure and induction conditions were found to be critical for optimal culture performance. Furthermore, formation of inclusion bodies associated with PAC overexpression tended to arrest cell growth, leading to potential cell lysis. iv At harvest, the whole non-clarified culture broth was applied directly to a tangential flow filtration anion-exchange membrane chromatography system. One-step purification of recombinant PAC was achieved based on the dual nature of membrane chromatography (i.e. microfiltration-sized pores and anion-exchange chemistry). Due to their size, cells remained in the retentate while the extracellular medium penetrated the membrane. Most contaminate proteins were captured by the anion-exchange membrane, whereas the purified PAC was collected in the filtrate. The batch time for both cultivation and purification was less than 24 h and recombinant PAC with high purity (19 U/mg), process yield (74%), and productivity (41 mg/L) was obtained.

Downstream purification often represents the most cost-intensive step in the manufacturing of recombinant proteins. Conventional purification processes are lengthy, technically complicated, product specific and time-consuming. To address this issue, herein we develop a one step purification system that due to the nature of the non-selective secretion system and the versatility of ion-exchange membrane chromatography can be widely applied to the production of many recombinant proteins. This was achieved through the integration of the intrinsically coupled upstream, midstream and downstream processes, a connection that is rarely exploited. A bioprocess for effective production and purification of penicillin G acylase (PAC) was developed. PAC was overexpressed in a genetically engineered Escherichia coli strain, secreted into the cultivation medium, harvested, and purified in a single step by anion-exchange chromatography. The cultivation medium developed had a sufficiently low conductivity to allow direct application of the extracellular fraction to the anion-exchange chromatography medium while providing all of the required nutrients for sustaining cell growth and PAC overexpression. It was contrived with the purposes of (i) providing sufficient osmolarity and buffering capacity, (ii) minimizing ionic species to facilitate the binding of extracellular proteins to anion-exchange medium, and (iii) enhancing PAC expression level and secretion efficiency. Employing this medium recipe the specific PAC activity reached a high level of 487 U/L/OD600, with more than 90% was localized in the extracellular medium. Both, the osmotic pressure and induction conditions were found to be critical for optimal culture performance. Furthermore, formation of inclusion bodies associated with PAC overexpression tended to arrest cell growth, leading to potential cell lysis. iv At harvest, the whole non-clarified culture broth was applied directly to a tangential flow filtration anion-exchange membrane chromatography system. One-step purification of recombinant PAC was achieved based on the dual nature of membrane chromatography (i.e. microfiltration-sized pores and anion-exchange chemistry). Due to their size, cells remained in the retentate while the extracellular medium penetrated the membrane. Most contaminate proteins were captured by the anion-exchange membrane, whereas the purified PAC was collected in the filtrate. The batch time for both cultivation and purification was less than 24 h and recombinant PAC with high purity (19 U/mg), process yield (74%), and productivity (41 mg/L) was obtained.

The need of purified biomolecules, such as proteins or antibodies, has required the biopharmaceutical industries to look for new recovering solutions to reduce time and costs of bioseparations. In the last decade, the emergent field of membrane chromatography has gained attention as possible substituent of the common used protein A affinity chromatography for bioseparations. In this scenario, gold nanoparticles can be used as means for offering affinity, mainly because of their biocompatible and reversible binding behavior, together with their high surface area-to-volume ratio, which offers a large number of binding sites. This work introduces a new procedure for purification of monoclonal antibodies based on polymeric membranes functionalized with gold nanoparticles. This novel approach shortens the process of purification by promoting selective binding of antibodies, while separating a mixture of biomolecules during a filtration process. The effects of gold nanoparticles and the surrounding ligand on the proteins adsorption and filtration are investigated. The results confirm that the functionalization helps in inducing a selective binding, preventing the non-selective one, and it also improves the selectivity of the separation process.

Current high-throughput screening (HTS) tools (i.e., single-use 96-well filter plate) are limited to the few membrane types that are sold commercially, restricting the ability to screen membrane materials for targeted applications. In this thesis, a multi-well device capable of screening any flat-sheet membrane was designed, where multiple devices can be stacked for extensive HTS (>32 experiments). Confocal imaging of a Natrix Q cross-section – a membrane type not sold in a commercial filter plate – was carried out after 24 h in contact with green fluorescent protein to visually confirm protein-membrane interactions. The static binding capacity (SBC) of bovine serum albumin (BSA) and Herring testes DNA was found for specific parameters: membrane type (Mustang Q, Sartobind Q, Natrix Q, Durapore), salt concentration (0, 50, 100 mM NaCl), and contact time (1 min, 4 h, 8 h, 24 h). Considering solution conditions, the highest BSA SBC was observed with Natrix Q at 0 M NaCl with a contact time of 24 h. The DNA and BSA SBC values for Natrix Q were the highest among the membrane types evaluated, demonstrating consistency with literature trends. These findings suggest that SBC experiments can predict promising membrane materials for scaled-up applications. Finally, the chromatography process was replicated in this multi-well device (Natrix Q), showing 50% BSA elution from the membrane. The results of this thesis confirmed this ability to accommodate any membrane adsorber, simultaneously compare different membrane materials, and extract the membrane for post-experimental analysis. This work’s significance was emphasized in its future potential to aid with membrane material selection, particularly with exploring the properties of next-generation membrane materials (e.g., 3D-printed membranes). Three future areas for optimization with this multi-well device were highlighted: biotherapeutic purification, sequencing of membrane materials within a process, and applying it as a tool to understand ion selectivity. / Thesis / Master of Applied Science (MASc) / Membranes are used in many industries, such as water treatment, environmental remediation, and biopharmaceuticals. In the biopharmaceutical industry, high-throughput screening (HTS) tools (e.g., filter plates), which allow for miniaturized experiments, are used to perform extensive experimental analysis to determine optimal solution conditions (e.g., pH) for biomolecule binding. Unfortunately, commercial filter plates are limited in customizability for HTS of membrane materials. To address these limitations, this thesis focuses on designing and validating a multi-well device capable of incorporating any membrane adsorber. Different biomolecules (proteins, DNA), solution conditions, and membrane materials were evaluated. The results of this thesis confirmed this ability to accommodate any membrane adsorber, simultaneously compare different membrane materials, and extract the membrane for post-experimental analysis. This work also discussed using this device for future rapid membrane material selection in multiple industries (e.g., biotherapeutics, ion extraction).

The biopharmaceutical industry has experienced remarkable progress in the upstream production capacity of life-saving proteins. This is while the downstream processing has failed to keep pace, including unit operations which are working close to their physical limit with no economy of scale. Column chromatography which is an integral unit in different stages of downstream purification is considered as the major bottleneck in this section. The packed-bed resin media is costly and the processes are labor-intensive and extremely time consuming. Membrane chromatography which uses a stack of adsorptive membranes as the chromatographic media is one of the most promising alternatives for conventional chromatography techniques. The performance of membrane adsorbers is consistent over a wide range of flow rates which is owing to the dominance of convective solute transport as opposed to the diffusion-based nature of mass transfer within the pores of the resin beads. This translates to much higher productivity and considerably lower buffer consumption (even as high as 95%), leading to much lower overall processing costs. The other advantages are significantly lower footprints and decreased pressure drops, both contributing to diminished capital costs. Membrane adsorbers are greatly scalable and used in a single-use manner. The latter eliminates the cleaning and validation steps and brings about much shorter processing times and higher flexibility in process development. Due to the performance advantages of membrane chromatography, this technique is now widely used in purification of high volumes of samples in late-stage polishing. Currently available membrane adsorbers have radial-flow spiral-wound configuration with high frontal surface area to bed height ratio according to which dilute impurities are removed in a flow-through format at very high flow rates and low pressure drops. Nevertheless, they fail to give high-resolution for bind-and-elute separations which makes them unsuitable for many unit operations, highly restricting their application. Severe design deficiencies such as large dead volumes and varying membrane area over the bed height result in broad and poorly resolved peaks. Herein, a novel device design was successfully developed which addresses the abovementioned shortcomings. The laterally-fed membrane chromatography (LFMC) devices house a stack of rectangular membrane sheets with two rectangular lateral channels on both sides of the stack as the feed and permeate channels. The design offers balanced pressure over the sides of the stack as well as even solute flow path lengths due to which the solute residence time is very uniform. Also, the small dead volumes minimize the dispersion effects. These features make the LFMC technology highly suitable for bind-and-elute applications, the improvement which is brought about by a simple design. The devices are easy to fabricate and highly scalable. The LFMC devices containing cation-exchange (CEX) membranes with 7 mL bed volume were examined for bind-and-elute separation where they outperformed the equivalent commercially available radial-flow devices. The design was further modified to give even lower dead volumes and more cost-effective fabrication. The latest embodiment of the device gave resolutions which were comparable with the ones obtained with the commercially packed resin columns in 1 mL and 5 mL scale with consistency over wide range of flow rates. The results were all acquired using a three component model protein system. Upon the approval of suitability of the device for bind-and-elute separation, the CEX-LFMC was used for purification of monoclonal antibodies (mAbs), the largest class of biopharmaceuticals. The device showed great performance in separation of mAb charge variants when extensively shallow gradients (60 membrane bed volumes) were required. The devices offered very stable conductivity gradients at high flow rates. LFMC devices in three different preparative scales gave great performance in separation of mAb aggregates which was approved for different mAb samples. The other application studied with the CEX-LFMC devices was the single-step preparative purification of mono-PEGylated proteins which is as well very challenging due to the physicochemical similarities between the target molecules and the impurities. Collectively, the LFMC devices combine the high-resolution with high-productivity which is highly desirable in downstream purification of biological molecules with great potential to expand the application of membrane chromatography. Finally, the LFMC devices were modified to adapt the analytical scale where they were integrated with a stack of hydrophilized PVDF membranes. The device successfully delivered ultra-fast separation of mAb aggregates in less than 1.5 minutes based on hydrophobic interaction membrane chromatography (HIMC). The assay times achieved with the HI-LFMC technique outclassed the currently available ultra-high performance chromatography (UPLC) methods at the same time with being extremely cost-effective. The application of the LFMC technology in analytical scale has great potential to offer cheap and rapid analysis in process development and quality control section of biopharmaceutical manufacturing. / Thesis / Doctor of Philosophy (PhD)