Patent Application
20220168667
The present invention relates to a chromatography system comprising at least two chromatography units connected in parallel.
Historically, conventional packed bed chromatography using porous beads has been an extremely powerful separation tool. In a porous bead-based system, the binding event between target molecule/impurity and the solid phase is dependent on diffusion into the porous bead. There is therefore a strong correlation between the residence time in a porous bead-based system and flowrate. Thus, binding capacity drops off with decreasing residence times. This type of chromatography can be called diffusion-based chromatography. A diffusion-based chromatography matrix includes any matrix which consists of particles and substantially exhibits a diffusion limitation of mass transfer, in that the rate of the adsorption and desorption processes is determined by the diffusion rate of the substance(s) into and out of the particles owing to the diffusion coefficients of the substance(s), which depend very heavily on the size, or the molecular weight, of the substances.
As alternatives to porous bead-based systems, monoliths or membranes may be used. The flow through such materials is convective rather than diffusional, and their binding capacity is therefore far less sensitive to flow than porous bead-based systems. These materials can be run at far higher flowrates than porous bead-based materials. In (membrane) adsorption chromatography, in contrast to gel-permeation chromatography, there is binding of components of a fluid, for example individual molecules, associates or particles, to the surface of a solid in contact with the fluid. The advantage of membrane adsorbers over packed chromatography columns is their suitability for being run with much higher flow rates. Conversely, however, membrane adsorbers have a lower dynamic capacity than, for example, gel-permeation chromatography columns. This is also called convection-based chromatography. A convection-based chromatography matrix includes any matrix in which application of a hydraulic pressure difference between the inflow and outflow of the matrix forces perfusion of the matrix, achieving substantially convective transport of the substance(s) into the matrix or out of the matrix, which is effected very rapidly at a high flow rate.
Convection-based chromatography and membrane adsorbers are described in for example US20140296464A1, US20160288089A1, WO2018011600A1 and WO2018037244A1, hereby incorporated by reference in their entireties.
The connection of chromatography units in parallel for increasing capacity is often problematic due to small differences in the different units which will lead to an uneven flow distribution through the different units. Hereby some of the units will not be optimally used. Furthermore a peak broadening of the eluted sample after chromatography is often the result when chromatography units are connected in parallel due to the non-uniform fluid flow through the units.
Hereby, either chromatography units which are as uniform as possible need to be provided for such systems and/or other control arrangements are needed for controlling the fluid flow between the units.
An object of the present invention is to provide an improved chromatography system with chromatography units connected in parallel.
A further object of the present invention is to provide a flexible and scalable chromatography system.
This is achieved by a chromatography system according to claim 1.
According to one aspect of the invention a chromatography system comprising at least two chromatography units connected in parallel is provided, wherein said at least two chromatography units each comprises a convection-based chromatography material, wherein an initial difference in back pressure provided from the different chromatography units is compensated dynamically during run of the system due to a change of chromatography unit properties provided during the chromatography process.
Hereby a flexible and scalable chromatography system is provided. Any number of chromatography units can be connected in parallel. Thanks to the dynamically changed delta column pressure the fluid flow through each of the chromatography units will also be dynamically changed during run of the system, i.e. the flow through the chromatography units which are connected in parallel will be regulated during run of the system automatically in dependence of the back pressure provided from each of the chromatography units. Surprisingly the system will hereby be self-regulated in a very effective way. The result will be an effective and uniform use of all of the chromatography units. This has not been the case in previous attempts to connect chromatography units in parallel. These types of chromatography units which are based on chromatography without diffusion into a chromatography medium, for example using cellulose fiber membranes, can be provided to allow for high flow rates which will allow for this effective self-regulation. Hereby all the units in the system will be effectively used.
In one embodiment of the invention each chromatography unit comprises a chromatography material having a chromatography material volume which is dimensioned to allow for a fluid flow rate through said chromatography unit of at least 2 chromatography material volumes per minute, wherein said system comprises a feed pump which is configured to flow a feed fluid through the chromatography units at a fluid flow rate of at least 2 chromatography material volumes per minute through each one of the chromatography units.
Because of these relatively high flow rates through the chromatography units a correspondingly high back pressure will be provided from each chromatography unit. This high back pressure can contribute to a fast and effective automatic regulation of the flow through the chromatography units.
Further embodiments are described in the dependent claims and in the following detailed description.
The invention relates to a chromatography system 1 comprising at least two chromatography units 3 connected in parallel. Such a parallel connection can be called a manifold and is schematically shown in
A convection-based chromatography material is a chromatography material which is not based on diffusion, such as diffusion into porous beads. A convection-based chromatography material can for example be monoliths or membranes, such as membrane adsorbers and can also be called adsorption chromatography. A cellulose fiber membrane is one example of a membrane adsorber. The convection-based chromatography material can be functionalized, for example protein A functionalized.
In one embodiment of the invention each chromatography unit 3 comprises a chromatography material having a chromatography material volume which is dimensioned to allow for a fluid flow rate through said chromatography unit of at least two or at least five chromatography material volumes per minute. The chromatography system 1 comprises hereby a feed pump 11 which is configured to flow a feed fluid through the chromatography units 3 at a fluid flow rate of at least 2 or at least 5 chromatography material volumes per minute through each one of the chromatography units.
The change of chromatography unit properties during the chromatography process can be due to different binding properties, for example binding of protein, in the chromatography material. Hereby the back pressure provided from the different chromatography units 3 will change in dependence of how much protein is bound in the chromatography material. Hereby the fluid flow through the chromatography units will be automatically adjusted during run of the system. Hereby all the units will be utilized optimally. The adjusting of fluid flow through the chromatography units is performed quickly thanks to the relatively high flow rate through the units.
This type of chromatography system 1 can be configured to recycle the fluid flow through the chromatography units 3 at least 10 or at least 20 times or at least 40 times.
Thanks to the quick adjusting of fluid flow in the system according to the invention the different chromatography units 3 of the chromatography system 1 does not need to be exactly identical. Chromatography units 3 which have some smaller differences, for example in initial delta column pressure or chromatography material volume can still be used and the utilization of these chromatography units 3 can still be effective in a chromatography system 1 according to the invention.
In some embodiments of the invention the feed pump 11 is configured to provide a fluid flow rate through each chromatography unit 3 of the system such that a back pressure provided from each of the chromatography units 3 is at least 5 bar.
The functionalised polymeric chromatography media of the present invention can suitably be formed from polymeric nanofibre substrates. Each substrate may be formed of one or more polymer nanofibres. The polymer nanofibres are typically electrospun polymer nanofibres. Such electrospun polymer nanofibres are well known to the person skilled in the art and optimised conditions for their production can be found in, for example, O. Hardick, et al., J. Mater. Sci. 46 (2011) 3890, the entirety of which is incorporated herein by reference. The processes of the present invention typically comprise an initial step of electrospinning a polymer to produce one or more polymer nanofibres. This may involve electrospinning a polymer to produce one or more non-woven sheets or layers, each comprising one or more polymer nanofibres. Suitably, the sheet(s) or layer(s) (10) each comprise a plurality of nanofibre-nanofibre fusion points (20) as illustrated in
Polymer nanofibres for use in the present invention typically have mean diameters from 10 nm to 1000 nm. For some applications, polymer nanofibres having mean diameters from 200 nm to 800 nm are appropriate. Polymer nanofibres having mean diameters from 200 nm to 400 nm may be appropriate for certain applications. The length of polymer nanofibres for use in the present invention is not particularly limited. Thus, conventional electrospinning processes can produce polymer nanofibres many hundreds of meters or even kilometers in length. Typically, though, the one or more polymer nanofibres have a length up to 10 km, preferably from 10 m to 10 km. The polymer nanofibres can suitably be monofilament nanofibres and may e.g. have a circular, ellipsoidal or essentially circular/ellipsoidal cross section.
Typically, the one or more polymer nanofibres are provided in the form of one or more non-woven sheets, each comprising one or more polymer nanofibres. Thus, the substrate is typically formed of one or more non-woven sheets, each comprising one or more polymer nanofibres. A non-woven sheet comprising one or more polymer nanofibres is a mat of said one or more polymer nanofibres with each nanofibre oriented essentially randomly, i.e. it has not been fabricated so that the nanofibre or nanofibres adopts a particular pattern. Non-woven sheets comprising polymer nanofibres are typically provided by known methods, such as that disclosed in O. Hardick, et al., J. Mater. Sci. 46 (2011) 3890. Non-woven sheets may, in certain circumstances, consist of a single polymer nanofibre. Alternatively, non-woven sheets may comprise two or more polymer nanofibres, for example 2, 3, 4, 5, 6, 7, 8, 9 or 10 polymer nanofibres.
Non-woven sheets typically have area densities from 1 to 40 g/m2, preferably from 5 to 25 g/m2, in some circumstances from 1 to 20 or 5 to 15 g/m2. Non-woven sheets can also typically have a thickness from 5 to 120 μm, preferably from 10 to 100 μm, in some circumstances from 50 to 90 μm, in other circumstances from 5 to 40, 10 to 30 or 15 to 25 μm.
The polymer used to produce the nanofibres used in the processes of the present invention is not particularly limited, provided the polymer is suitable for use in chromatography applications. Thus, typically, the polymer is a polymer suitable for use as a chromatography medium, i.e. an adsorbent, in a chromatography method. Suitable polymers include polyamides such as nylon, polyacrylic acid, polymethacrylic acid, polyacrylonitrile, polystyrene, polysulfones e.g. polyethersulfone (PES), polycaprolactone, collagen, chitosan, polyethylene oxide, agarose, agarose acetate, cellulose, cellulose acetate, and combinations thereof. Polyethersulfone (PES), cellulose, cellulose acetate, and combinations thereof are preferred. In some cases, cellulose, cellulose acetate, and combinations thereof are preferred.
In some embodiments, the substrate comprises one or more nanofibres formed from one or more polymer nanofibres formed from different polymers. Thus, in this embodiment, the substrate comprises one or more different polymers. Typical polymers are as defined above.
Typically, chromatography medium is a functionalised cellulose chromatography medium, prepared from a substrate formed of one or more cellulose acetate nanofibres. Preferably, a substrate is formed of one or more non-woven sheets or layers is provided, each comprising one or more cellulose acetate nanofibres. Cellulose acetate is readily electrospun, typically from a solution of cellulose acetate in one or more organic solvents, and can readily be transformed into cellulose after electrospinning. Thus, preferably the substrate is formed of one or more non-woven sheets/layers, each comprising one or more electrospun cellulose acetate nanofibres.
Physical Modification of Nanofibres
Provision of a substrate may involve physical modification of the polymer nanofibres optionally in the non-woven sheets/layers, prior to a grafting step. Specifically, physical modification may involve heating and/or pressing the polymer nanofibres/non-woven sheets/layers, preferably heating and pressing the polymer nanofibres/non-woven sheets/layers. These steps improve the structural stability of the material. The pressing and heating conditions may also be varied to alter the thickness and/or porosity of the resultant material. Use of multiple non-woven sheets of polymer nanofibres enables a thicker material to be prepared which has a greater capacity for adsorption (once grafted and functionalised). The provision of a substrate typically therefore may comprise providing two or more non-woven sheets/layers stacked one on top of the other, each said sheet comprising one or more polymer nanofibres, and simultaneously heating and pressing the stack of sheets/layers to fuse points of contact between the nanofibres of adjacent sheets/layers, creating interlayer fusion points. In the case of a cellulose chromatography medium, provision of a substrate typically comprises providing two or more non-woven sheets/layers stacked or folded over one on top of the other, each said sheet/layer comprising one or more cellulose acetate nanofibres, and simultaneously heating and pressing the stack of sheets/layers to fuse points of contact between the nanofibres of adjacent sheets. The functionalised chromatography medium can thus be a stack of a plurality of non-woven polymeric nanofibre layers, with a plurality of interlayer nanofibre-nanofibre fusion points, which bond at least two of the layers to each other. Preferred processing conditions for pressing and heating of polymer nanofibres/nonwoven sheets can be found in US 20160288089 and WO-A-2015/052465, the entireties of which are incorporated herein by reference.
Grafting Nanofibre Substrate
The preparation of the chromatography medium may involve a grafting step which typically involves grafting one or more neutral polymer chains from the substrate provided. Grafting one or more neutral polymer chains from a substrate typically comprises growing one or more polymer chains from one or more functional groups present on the substrate, optionally in the presence of one or more catalysts. Thus, typically, the substrate comprises one or more functional groups, preferably one or more functional groups from which a polymer chain may be grown. Growing a polymer chain from the one or more functional groups means building up a polymer at the one or more functional groups from individual monomer building blocks. The grafting step typically therefore involves growing polymer chains directly from the substrate, rather than bonding preformed polymer chains to the substrate. Thus, as the polymerisation progresses, individual monomers are added to the ends of growing polymer chains that are anchored distally to the substrate. The polymer coating will then comprise polymer molecules which are single-point covalently tethered to the substrate. The polymer molecules may be linear or branched and even hyperbranched, in which case >50% of the monomer residues are either branch points or terminal monomers. Growth of the polymer chains directly from the substrate enables control over the structure of the polymer coating, particularly using a polymerisation strategy whereby the polymers all grow simultaneously at a uniform rate. This enables formation of a dense and well-defined polymer coating layer. Thus, when the substrate is one or more layers of polymeric nanofibres fused together at their junctions, a well-defined thin coating will be formed—covering the core nanofibres and the fusion points between individual nanofibres. The coating may be a conformal coating, i.e. following the contours of the nanofibres and the fusion points. The thickness of the coating may e.g. be such that the coated nanofibres have an average diameter of 100-1000 nm, such as 100-700 nm or 200-700 nm. The average diameters of the non-coated nanofibres may then e.g. be 100-800 nm, such as 100-600 nm. The average pore size of the layer(s) may e.g. be 200-800 nm and the pore volume fraction may e.g. be 50-90%, such as 60-80%. The average pore size and the average fibre diameters can be calculated from SEM images of the layers and the pore volume fraction can be calculated from the total volume of a layer or substrate (the thickness multiplied by the cross-sectional area) and the specific gravity of the nanofibres. A specific example of a pore volume calculation of a representative multilayer cellulose acetate nanofibre disc is: Disc diameter 32 mm and disc thickness 0.44 mm gives a total volume of 0.354 cm3. The dry weight of the disc is 0.165 g, which with a cellulose acetate specific gravity of 1.31 g/cm3 gives a cellulose acetate volume of 0.126 cm3. The pore volume fraction is thus (0.354−0.126)/0.354=64%.
The pore structure of the layers is important for the performance in that a delicate balance between high dynamic binding capacity (large accessible surface and short diffusion paths) and low back pressures (large pores and high pore volume fractions) needs to be met. The grafted nanofibre layers are uniquely suited for meeting this balance with their small fibre diameters and high pore volume fractions. Particularly good combinations are achieved when the layers have the following pore sizes, as obtained from capillary flow porometry with a perfluoropolyether wetting liquid: bubblepoint pore size—0.9-1.2 μm, such as 1.0-1.2 μm; smallest pore size—0.2-0.4 μm and/or mean flow pore (MFP) pore size 0.3-0.5 μm. The measurements should be performed as detailed below in Example 4.
For good flow properties, it is advantageous if the open pore structure of the substrate is retained after the grafting process. In this way, the chromatography medium will have a first side and a second side, fluidically connected to each other through an open, and three-dimensionally connected, pore structure formed by the interstices (interstitial volume) between the grafted polymeric nanofibres. This pore structure is preferably free from, or substantially free from, graft polymer or any homopolymer accidentally formed during the grafting process. This can be achieved by limiting the monomer amounts added during grafting and the absence of any polymer obstructing the pore structure can easily be checked by measuring the flow rate, and/or by observing the pore structure by electron microscopy. The graft polymerisation process offers a unique possibility to introduce a functionalised polymer increasing the binding capacity, without obstructing the pore structure.
Study: Cycling mAb Perfusion Harvest on 4×4 mL PrismA Fiber Manifolds
The aim with this study was to show proof-of concept for manifolding of Fibro PrismA units, prepared as disclosed in WO2018011600 and WO2019137869, both of which are hereby incorporated by reference in their entireties. Four prototype units were selected for a set-up (array) of 4×4 mL manifold. The units within the set-up were selected to be similar in DBC and delta column pressure (dCP).
The lifetime study included CIP every 10th cycle using 1.0 M NaOH at a low flow rate in reverse, corresponding to a contact time of 30 minutes. 101 cycles were performed during two days where the first 51 cycles were performed overnight with a depth- and sterile-filtered mAb harvest material at ambient temperature. The last 50 cycles were performed on the next day with a harvest material of low temperature due to cold storage over-night. The low mAb harvest temperature resulted in a pressure increase (jump) from cycle 52 and onward, especially during harvest load.
The purification performance was good with recoveries ≥94%, elution pool volumes close to 4 MV and approximately log 3 reduction of HCP. The ligand leakage level was 26 ppm in the first cycle, however, the ligand leakage decreased over the 101 cycles and the in the bulk pool the concentration was <5 ppm.
The dCP and the DBC was measured on each single unit and the manifold pre- and post-cycling. After the cycling the average dCP for individual units, and the manifold had increased by approximately 30% whereas the DBC post cycling was 97% of the initial capacity.
This study provided additional evidence that manifolding of 4× PrismA fiber units in parallel is feasible. However, the flow rate had to be decreased due to relatively high initial back pressure for these chemistry prototypes.
Disposition of the Study
Cycling 4×4 mL PrismA Fiber Manifold, Day 1:
Cycling 4×4 mL PrismA Fiber Manifold, Day 2:
The dynamic binding capacity (DBC) and the dCP during equilibration of the manifold and the individual units was determined after cycling.
The units were connected in parallel according to the schematic picture in
mAb8 Perfusion Cell Culture Harvest
The cell culture was performed as a 25 L perfusion in a WAVE bioreactor (GE Healthcare). The CHO cells were grown in HyClone CDM4NS0 medium+CellBoost 1 and CellBoost 3 as perfusion medium. The cells were inoculated into the reactor at low volume and the volume was successively increased during three days to full volume. Two days later the perfusion started at a medium exchange rate of 12.5 L/day. Thereafter, as the cell density increased, the perfusion rate was increased to the final rate of 50 L/day. The perfusion was completed after approximately ten days at a cell density of 140×10e6 cells/mL. The mAb concentration in the perfusion increased to a final concentration of approximately 0.5 g/L.
A volume of 167 L from the last three days was obtained from this perfusion culture. The mAb concentration was determined to 0.4 (0.368) mg/mL by Biacore using a mAb standard.
Methods
Determination of DBC of the Individual 4 mL Units and the 4×4 mL Manifolds
A pre-purified mAb sample was diluted to the concentration of 2 mg/mL in wash 1 buffer (20 mM phosphate, 500 mM NaCl, pH 7). The sample was filtrated (0.22u) and the mAb concentration was determined after 4 times dilution in wash 1 buffer by measuring the UV absorbance at 280 nm and calculation of the concentration by Beer's law (Equation 1) using 1.48 ml/(mg×cm) as absorptivity coefficient (e).
A=εLc Equation 1 Beer's law
Where A is the UV absorption at 280 nm, E is the absorptivity coefficient, L is the path length of the cell holder and c is the concentration of the solution.
Before starting the BT analysis, the mAb sample was injected by-passing the column valve and the 100% UV absorbance at 280 nm was noted.
The fiber units were washed with elution buffer and equilibrated before start of the BT analysis. The Unicorn method for the BT analysis is described below.
Unicorn Method
Starting conditions: Alarm pre-column pressure 1.7 MPa and delta column pressure 1.0 or 1.5 MPa. UV averaging time (Noise reduction) 2.5 s.
Air trap and mixer valves by-pass.
A flow rate of 32 and 128 mL/min was used throughout the method for the single (4 mL) and manifolded units (16 mL) respectively.
The DBC at 10% BT of mAb (QB10%) was approximated to the amount of mAb in mg loaded per mL PrismA fiber at 10% BT by using following procedure:
The delay volume was determined to the volume where 50 V of conductivity shift occurred during the transition from equilibration buffer to the mAb sample diluted in 20 mM phosphate buffer containing 500 mM NaCl. The vertical marker reference point was set at 50% conductivity breakthrough and the delta volume was determined to the volume where the A280 nm curve reached 10% of the 100% absorbance of the mAb sample. The delta volume was multiplied by the concentration of the mAb sample and then divided by the volume of the unit (4 mL (single unit) or 16 mL (manifold)).
Cycling of Manifold Using Depth- and Sterile-Ter mAb Perfusion Harvest Including CIP Every 10th Cycle The Unicorn method in Table 5 was used with the addition of CIP after the loop phase.
CIP was done in reversed flow (Column 2 up-flow) where 6 MV of 1.0 M NaOH (A4) was applied at a flow rate of 100 mL/min. Then the flow rate was lowered to 6.6 mL/min during application of 12.5 MV (200 mL) 1.0 M NaOH resulting in a CIP contact time of 30 minutes. Re-equilibration was performed for 10 MV after the CIP.
The capacity of the manifold was 29.6 mg/mL which resulted in a load volume of 947.2 mL mAb harvest (titer of 0.4 mg/mL) per cycle for a load of 80% of QB10% (0.8*29.6*16/0.4).
100 L mAb harvest was filtered using two depth filters coupled in parallel (Zeta Plus Capsule filters, E102FSA90ZB05A, 1020 cm2, 3M) into a 200 L bag. Approximately 50 L of the depth filtered mAb harvest was further sterile filtered using a Ulta HC 20″ filter into a 50 L bag and used for the first 51 cycles. The temperature of this material was close to ambient when starting the cycling. The remaining 50 L depth filtered material was stored in the cold room until use following day for cycles 52-101.
The first 51 cycles were run over night using the scouting scheme according to Table 6.
On the second day the last 50 L of the mAb perfusion harvest were taken out from the cold room and sterile filtered into a 50 L bag using a new Ulta HC 20″ filter. This material was still cold when starting the cycles 52-101 according to Table 7.
Samples from all eluates (single cycles, pools of two cycles and bulk pools of 45 cycles) were taken out and prepared for HCP, leaked PrismA ligand, mAb concentration (Biacore and UV) and aggregate concentration by SEC analysis according to paragraphs.
Evaluation of dCP and UV Peak Area Over Cycles
The dCP and UV 300 nm curves were integrated and peaks corresponding to the delta column pressure during loading, elution and re-equilibration as well as the elution peak area were exported to excel using the “Multi result peak compare” function in UNICORN. The pressure peak height (MPa) and the normalized UV300 nm peak area was plotted for the 101 cycles on the manifold.
Analysis of Host Cell Protein, Leaked PrismA Ligand and mAb Concentration
450 μL of mAb harvest samples were taken out and mixed with 50 μL preservation solution and stored in the fridge until HCP analysis. 450 μL of eluate samples were taken out and mixed with 50 μL preservation solution and stored in the fridge until HCP and Protein A ligand (PrismA) analysis.
The mAb concentration of the eluates were determined using the UV absorbance at 280 nm on samples diluted five times. 40 μL sample was mixed with 160 μL elution buffer in a UV-readable plate (duplicated dilutions). The UV was measured using path check correction and the concentration was calculated using Beer's law (Equation 1) with 1.48 ml/(mg×cm) as absorptivity coefficient.
SEC Analysis for Aggregate Concentration Determination
200 μL of the elution samples were distributed in a 96-well plate and analyzed using SEC.
Buffer: PBS, 0.14 M NaCl, 0.0027 M KCl, 0.010 M phosphate, pH 7.4, was prepared by dissolving a buffer tablet (09-9400-100, Medicago) in MQ water and adjusting the final volume to 1 L in a volumetric flask.
Column: Superdex 200 Increase 10/300 GL, code 28-9909-44, lot 10263258 (Expiry 2022-06, ID 0094), CV 23.563 mL.
Unicorn Method:
Flow 1 mL/min
Wave lengths: 280, 260 and 215 nm
Alarm pressure 4.50 MPa
Column position 3
Equilibration 1 mL, auto zero UV
Sample injection Partial_905, 50 μL
Isocratic elution for 25 mL
SEC runs with blank (buffer) injections were initially done.
Criteria for Acceptance
Manifolds of 4×4 mL units with less pressure difference than 0.12 MPa between units initially should perform well in cycling studies, with purity and recovery comparable to MabSelect PrismA.
Results and Discussion
101 Cycles Including CIP Every 10th Cycle, on the Manifold Using Depth and Sterile Filtered mAb Perfusion Harvest
A manifold consisting of four 4 mL units made of PrismA fiber from there different chemistry batches was used.
The flow rate throughout the method was set to 100 mL/min. An example of a chromatogram is seen in
The 100 L perfusion mAb harvest material had been taken out of the cold room in the morning and kept at room temperature the whole day during the two filtration steps (depth and sterile filtration). Therefore, the harvest material was close to room temperature when the purification started in the afternoon. The first 51 cycles were performed over-night. The last 52-101 cycles were performed on the following day. At that time point, the 50 L mAb harvest depth filtered on previous day, had been kept in the cold room overnight and only kept at room temperature for less than an hour during the sterile filtration. Thus, the harvest material was still cold when starting the cycling.
The dCP during equilibration/re-equilibration in the first cycle was close to 0.5 MPa. The dCP increased approximately 0.1 MPa over the first 51 cycles where the largest pressure increase occurred during the first 10 cycles and then levelled off (
In cycle 52 to 101 the pressure during sample loading was related to the temperature of the mAb harvest material. In the overlay of every 10th cycle 61 to 101 the pressure during loading was highest in cycle 61 and then decreased over cycles as the temperature of the mAb harvest got closer to ambient (
The dCP for the loading, elution, and re-equilibration phases as well as the relative elution peak area for all cycles is presented in (
The recovery was ≥95% and the eluate pools was close to 4 MV for eluates from all cycles (single cycles, pools of eluates from two cycles and the bulk pools which are eluates from 45 cycles) (
Product Quality of the Eluates from Cycling of the Manifold
dCP and DBC of the Individual Units and Manifold Before and After Cycling
Before the cycling of the manifold the dCP of the four units ranged between 0.51 to 0.58 MPa as measured during the equilibration phase during BT analysis at a flow rate of 32 mL/min (Figure). After the cycling the dCP for the four units had increased to different extent with an average increase of 32% which also correlates to the dCP increase of the manifold. Largest difference in pressure between units before cycling was 0.07 MPa and after cycling it was 0.17 MPa. The manifold had a dCP of 0.60 MPa before cycling which had increased to 0.78 MPa after cycling. The DBC and dCP of individual units and the manifold before and after cycling is presented in Table. The post cycling manifold capacity was 28.6 mg/mL which corresponds to 97% of the initial capacity.
This study provided additional evidence that manifolding of 4× PrismA fiber units in parallel is feasible. However, the flow rate had to be decreased to 100 mL/min due to relatively high back pressure for these chemistry prototypes. 128 mL/min is aimed for in the project.
Pressure and capacity matching of the individual units in a manifold is believed to be crucial for good purification performance. However, it is also assumed that there is an intrinsic balance of pressure and flow within the manifold. If pressure is built up more on one of the units in the manifold compared to the other, the flow through that unit would decrease. This would lead to an increased flow rate during the remaining units with increased pressure build up on these. This self-regulating pressure build up and flow rate through the units would go on and level out any differences between the units at the end, considering that the initial difference between the units is not too large.
2 mL cellulose nanofiber membranes were grafted with glycidol and treated with divinylsulfone (DVS) as described in WO2018011600 and WO2019137869. They were then reacted with polyallylamine, resulting in an anion exchange functionality. Two of these 2 mL membranes were coupled in parallel and equilibrated in 20 mM Tris pH 8.0+100 mM NaCl. A BSA solution in the same buffer was loaded at 50 mL/min flow rate and the membranes were then washed with the equilibration buffer. The adsorbed BSA was then eluted with 20 mM Tris pH 8.0+2.0 M NaCl. As seen in
2 mL cellulose nanofiber membranes were grafted with glycidol and treated with divinylsulfone (DVS) as described in WO2018011600 and WO2019137869. They were then reacted with polyallylamine, resulting in an anion exchange functionality. Two of these 2 mL membranes were coupled in parallel and equilibrated in 20 mM Tris pH 8.0+100 mM NaCl. A suspension of adeno associated virus serotype 2 (AAV-2) in the same buffer was loaded at 50 mL flow rate and the membranes were then washed with the equilibration buffer. The adsorbed AAV-2 was eluted with 20 mM Tris pH 8.0+1.0 M NaCl. As seen in
The equipment used was a POROLUX™ 100 porometer (IB-FT GmbH, Berlin, Germany) and the methodology was as given in Table 9. Further details about the measurement principles can be found on the manufacturer's website http://www.ib-ft.com/measurement_principle.html
The three pore sizes obtained from the measurements are smallest pore size. Mean Flow Pore (MFP) size and bubblepoint pore size (the largest pore size).
The results from the capillary flow porometry, as well as estimates of the average nanofibre diameters from SEM images are shown in Table 10. These results are also applicable to the Fibro PrismA and polyallylamine membranes used in Examples 1-3.
| Number | Date | Country | Kind |
|---|---|---|---|
| 1904522.8 | Apr 2019 | GB | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/EP2020/059067 | 3/31/2020 | WO | 00 |
