WO2011035282A1 - Dual capture separation - Google Patents

Abstract

A process for a combined dual step separation of a mixture of two or more groups of molecular substances where the groups differ in their ability to be transported into and bound by two capture media, such that members of one group are largely excluded from the first media allowing them to flow through' to the second media is disclosed The capture media may both be based on similar base matnces and ligands but differ in other properties Examples of mixtures for processing in this manner include native and polysaccharide modified proteins, or native and synthetic polymer modified proteins The groups of molecules may themselves be heterogenous in regard to properties other than the above interactions This includes a group of monopolymer grafted proteins of protein to polymer molar ratio 1, whose members are heterogenous in regard to the position of polymer coupling to the protein

Description

Dual Capture Separation

Cross-Reference to Related Applications

This application claims priority to Swedish patent application number 0950688-2 filed September 21, 2009; the entire disclosure of which is incorporated herein.

Field of the Invention

The present invention relates to the field of biotechnology, and more specifically to the purification of biological compounds, such as proteins, antibodies, glycoproteins, nucleic acid polymers, other biomacromolecules and the like. More specifically, the present invention relates to a method of liquid chromatography using two columns that contain similar capture ligands, and are run using comparable mobile phase conditions.

This significantly, reduces the number and extent of mobile phase changes in a process.

The invention may be especially useful for separation of mixtures of proteins containing one or more larger molecular weight proteins, such as antibodies or polymer-modified proteins.

Background of the Invention

Separation of proteins and other target macromolecules by capture (e.g. ion exchange, metal ion affinity, hydrophobic interaction) chromatography is a well- established method. It is complementary to non-capture chromatography methods such as size exclusion chromatography. A particular advantage of capture chromatography is that two or more orthogonal methods which separate macromolecules on the basis of different physical properties (e.g. charge or affinity or hydrophobicity) can be employed in two or more different steps. This gives capture based processes great selectivity even though alteration from one method or unit operation to another (e.g. cation exchange to anion exchange, or metal ion affinity to hydrophobic affinity) typically requires costly alteration of the mobile liquid phase pH or salt composition and conductivity. Such "buffer changes" are dictated by two requirements. First because different separation approaches typically require different capture buffer (e.g. below target pi for cation exchange and above target pi for anion exchange). Secondly because adsorption and elution buffers often differ significantly [e.g. elution at high conductivity and above target isoelectric pH (pi) for cation exchange] so that the elution buffer from one separation may not be suitable as adsorption buffer for the next separation method. Mobile liquid phase buffer change is a particular problem in regard to large scale processing of proteins for use as biopharmaceuticals, or industrial catalysts, nutritional additives, etc. In such cases costs associated with buffer changes have a major impact on production costs.

Bioprocess engineers often try to develop processes using chromatographic steps arranged in order to reduce the number of buffer changes. One well-known example used in many bioseparation processes, is using the high conductivity elution buffer from an ion exchange chromatography (IEC) step as basis for formulating a high conductivity adsorption buffer in a following hydrophobic interaction chromatography (HIC) step. The latter typically involves adding even more salt to the IEC elution buffer to form a suitable HIC adsorption buffer.

Some mixtures of macromolecules offer particular challenges with the above approaches. For example proteins of similar charge or hydrophobic nature which differ in size, or proteins which are similar or identical except for being modified with

carbohydrate polymer groups. Such polymer modification may include proteins modified naturally with various polymers of glyco (sugar) groups, or proteins synthetically modified with polymers including naturally occurring polymers such as dextran polyglucose or starch polymers or synthetic polymers such as poly(ethylene glycol) [PEG]. Polymer modification of proteins or other bioactive pharmaceuticals is typically done to improve the pharmacological efficacy or other properties (e.g. protease resistance) of biopharmaceuticals, catalytic proteins, etc. There are, of course, other large scale commercially significant separation challenges which can be noted here. For example fractionation of plasma proteins involves separation of major protein components such as antibody and albumin proteins. Another is fractionation of major milk proteins such as lactoferrin, alpha-lactalbumin and beta-lactoglobulin. With the advent of production of recombinant proteins in plants, milk and eggs the ability to separate antibodies and other recombinant biopharmaceuticals from naturally occurring proteins, such as ovalbumin in recombinant egg protein processing, or proteins such as lactoferrin, lactalbumin and beta-lactoglobulin in recombinant milk processing, are becoming more important.

Modification of proteins by polymers, especially poly (ethylene glycol) [PEG] is gaining popularity. There are now several "PEGylated" proteins which are FDA approved and have yearly sales in excess of one billion dollars. Examples include PEGINTRON^® (PEGylated INTRON^®) from Schering, PEGASYS^® (PEGylated ROFERON^®) from Roche and NEULASTA^® (PEGylated NEUPOGEN^®) from Amgen. As indicated by these examples PEGylation is particularly suited to improving the serum half-life of small highly bioactive substances such as hormones including interferons. PEG-modified antibody and antibody fragments are also very promising. The polymer modification tends to reduce needed dosage and dosing frequency and development of drug-directed immune responses; thus improving both short and longer term patient tolerance to the treatments. In some cases (e. g. PEG-insulin) it may also allow for the drug to be taken orally, parenterally, or other routes than by injection. The result is PEGylated drug formulations which have greater chance of FDA approval, patient tolerance, ease of application, wider patient population (e. g. sales base) and longer term safety and efficacy.

In addition to PEG other hydrophilic polymers such as those containing mixtures of propylene glycol and ethylene glycol units, may also be useful for modification of biopharmaceuticals or synthetic pharmaceuticals. So too these neutral hydrophilic polymers appear to mimic hydrophilic carbohydrate polymers which might occur as a result of glycosylation and also increase the serum half life of biopharmaceuticals. There are also polymer conjugates of ethylene glycol units and carbohydrate units which are like PEG useful as nonimmunogenic polymers in formulations. One well known example is ethylhydroxyethylcellulose or EHEC. Another is use of dextran (polyglucose) polymer. In addition to modification of various drugs the above types of polymers, of which PEG is a prime example and in many ways representative of other such polymers, can be used to modify surfaces or other drug carriers including those used internally, parenterally, and externally (for recent review of such polymers, their size ranges and functionalities, see Poly(ethylene glycol) in Drug Delivery: Pros and Cons as Well as Potential Alternatives, K. Knop, R. Hoogenboom, D. Fischer, and U. S. Schubert, Angewandte Chem.

International Edn., vol. 49, pages 6288-6308, 2010). In some cases the active units are not proteins but other substances such as polynucleotides, or oligopeptides, or oligonucleotides, or other substances. One example is the drug MACUGEN^® (jointly developed by Pfizer and Gilead Sciences) and which is a PEGylated aptamer. One can refer to such a larger family as polymer modified pharmaceutical substances (PMPs).

What the above substances have in common is their active protein or other molecular regions are modified with neutral or other polymers which induce significant alteration in their hydrodynamic size, and charge density (net charge per unit area or volume) but not in their native number and type of charge groups. As a result of this such compounds are expected to exhibit a. Lower affinity to charged surfaces including ion exchange media, b. Greater hydrodynamic radius and slower diffusion rates than their native active counterparts, c. Other altered properties such as solubility, hydrophobicity (and interaction with hydrophobic interaction chromatography media) and hydrogen bonding (reviewed in C. Fee and J. Van Alstine, Bioconjugate Chemistry, Vol. 15, pp. 1304 - 1313, 2004).

Given the above it is not surprising that PEG-proteins typically show lower strength of binding on ion exchange media, or affinity media, and can be displaced by unmodified proteins. It is also not surprising that the much larger PEG-proteins typically exhibit much (often lOx) lower dynamic binding capacities (gram protein per volume of media) since their larger size means they occupy more media surface area per gram of protein bound, and they are more limited in the total area of media accessible to binding. The latter steric pore size distribution considerations are often very significant (C. Fee and J. Van Alstine, Chem. Engineering Sci. Vol. 61, pp. 924-939, 2006).

There are several distinct technical chemical separation challenges when processing PMPs - two of which relate to the present invention. They are noted below in relation to PEGylation but similar concerns hold when other types of polymers are used. In order to be useful PEG and related polymers must be chemically activated with various functional groups which may target the polymers to attach to specific sites on proteins or other substances of interest (e.g. PEG-n-hydroxysuccinimide reacting with protein amino- terminal or lysine peptide residue amine groups). Once the functionalized polymers are purified they can be reacted with target substances. A major bioprocessing challenge relates to fractionating the resulting reaction product mixture. The chemical reactions yield mixtures of unreacted PEGs, activated (modified) PEGs, polymer modified target substances, and side products (often lower MW charged substances). All of the above reaction products need to be fractionated and in most cases at least one group - the polymer modified targets of interest - need to be subfractionated on basis of their degree (polymer to target mole ratio) of polymer modification. The latter typically involves isolation of monopolymer modified targets (i.e. mole ratio = 1) from so called oligomodified target fraction (mole ratio > 1) which in the case of PEG-protein processing is often taken to include dimodified targets. The latter type of separation challenge is also expected to hold for proteins or other substances naturally or synthetically modified with other polymers including polysaccharides. This includes polymers which are considered as potential replacements for PEG in biopharmaceutical applications.

Separation of polymer modified substances is complicated and challenging and does not simply involve separating two components. That coupled to the value of the target substances, and low dynamic capacities necessitates a recognized need to optimize the purification process. As noted earlier optimization can have two primary goals (Fee and Van Alstine, 2006). First, isolation of target substances which are not modified by polymer (mole ratio 0) so that they will not take up column capacity and (due to their greater affinity) displace polymer modified targets, and also so that they might be recycled. Second, separation of mono- and oligo-polymer modified targets. The present invention is directed, in part, to these and other challenges related to mixtures of proteins differing in size and affinity for different solid phase separation related surfaces and formats.

WO 2006/011839 relates to a method for purification of at least one PEGylated compound, including providing a separation matrix comprised of a porous support, which presents an average pore radius of at least 40 nm and to the surfaces of which functional groups have been immobilised; and contacting the matrix with a liquid that comprises PEGylated compound(s) to allow interaction of one or more such compounds with the functional groups; and recovering PEGylated compound(s) as one or more fractions. This invention also describes a novel use of conventional size exclusion media base matrix, which has been provided with capture media functional groups such as ion exchange groups, hydrophobic groups, affinity groups, or the like and used to fractionate PEG- protein reaction product mixtures.

WO 2005/029065 relates to a separation matrix comprised of a support to the surfaces of which polymer chains have been coupled, wherein each polymer chain presents recurring proton-donating groups and at least the surface of the support is substantially hydrophilic. The primary function of the polymer coating is to bind PEG and related polymers via a large number of relatively weak hydrogen bonds. The polymers thus function as poly-ligands and are actively involved in the capture process. The polymer also fulfills another critical function which is to orient the topographical spacing of the hydrogen bond interacting groups so that they match the ethoxy groups on the PEG. For that reason, as shown in the patent filing, normal carboxy group containing IEC media will not perform sufficiently well to effectively bind suitable amounts of PEG polymers or related substances at pH below 5. Much better performance is achieved media presenting polycarboxylic acid polymer which affords the correct molecular topography.

US 6,428,707 describes a method for adsorption of a substance from a liquid sample on a fluidized bead or stirred suspension, in which the beads used comprise a base matrix and exhibit a structure having affinity to the substance, characterized in that the structure is covalently bound to the base matrix via an extender. Such extenders can be polymers modified with the ligands but differ from WO 2005/029065 where the hydrogen bond participating group is an inherent part of the polymer.

US 6,572,766 describes a matrix including a core showing a system of micropores and a surface in which the micropore system has openings. The characterizing feature is that the surface is coated with a polymer having such a large molecular weight that it cannot penetrate into the micropores. The presence of this polymer (which unlike in US 6,428,707 can be uncharged or otherwise contain no ligands or other groups which bind macromolecules) still has the ability to influence the transport of target or other macromolecules into the matrix, and therefore the capture and capacity properties of the matrix. Such selectivity is typically on the basis of size related to steric and perhaps diffusion or other hindrances noted below.

US 6,426,315 describes a process to introduce two or more functionalities in layers on a porous matrix. The characterizing feature is that the matrix is contacted with a functional deficiency of a reagent (which will react with functional groups on the matrix) under conditions where the reaction between reagent and the matrix is more rapid than diffusion of the reagent in the matrix. This can be a matrix with or without polymer coatings or tethers as described above.

US 2005-0267295 describes a method for purifying a desired nucleic acid related substance I from another nucleic acid related substance or group of substances II in a mixture, where both I and II have affinity for the same ligand structure, and wherein I is smaller than II. The method comprises the steps of (1) providing substances I and II in a liquid; (2) contacting the liquid with an adsorbent which selectively adsorbs substance I; (3) recovering the desired substance. The adsorbent has (a) an interior part which carries a ligand structure that is capable of binding to substances I and II, and is accessible to substance I, and (b) an outer surface layer that does not adsorb substance II, and is more easily penetrated by substance I than by substance II. US 5,886,155 describes a dual column system where Butyl SEPHAROSE™ hydrophobic interaction chromatography (HIC) media is used as an able-to-be- regenerated guard column upstream from a Phenyl SEPHAROSE™ HIC target capture column. Substances more hydrophobic than the target bind to the Butyl SEPHAROSE™ column under conditions where target does not bind and thus is available for binding to the second Phenyl SEPHAROSE™ column. Both types of HIC media are produced by GE Healthcare Bio-Sciences AB, Uppsala, Sweden.

Some properties of media whose mean pore size and pore size distribution can improve the ion exchange based purification of large target molecules, including PEGylated proteins from PEGylation reaction product mixtures, is given in WO 2006- 011839 which also notes how such media can be used in conjunction with various mobile phase buffers whose conductivity is varied in linear or step fashion.

It should be noted there that in some cases the purity of target obtained from one capture step may not be as high as desired and for that reason the same step may be repeated. US 2009-0118476, "Purification of Pegylated Polypeptides", refers to a method for the purification of monoPEGylated erythropoietin using two similar cation exchange chromatography steps which employ the same type of cation exchange material. In this case the examples only cover commercially available cation exchange media with fairly homogenious particle size distribution and ligand distribution, with target elution in continuous or step gradient modes. Typical target purification is reported to be 60% after the first cation exchange step and 90% or more after the second step.

Brief Description of the Invention

The present invention relates to the discovery that certain separation challenges can be solved by using two columns where selectivity differences do not come from the use of very different mobile phase buffers, or different (charge, hydrophobic, affinity, or other) ligands. Rather the selectivity derives from subtler chromatography media construction differences, or inherent nature (e.g. mass transfer into chromatography media) of various target and contaminant macromolecules as a function of their structure. This allows one molecule or group of molecules to be isolated from another molecule or group of molecules. Examples of such group separations include significantly larger and smaller substances, native and polymer modified proteins, native and glycosylated proteins, plus other protein separation challenges noted in the background section.

The present invention relates to the separation of complex mixtures of macromolecules using two or more columns in series. It involves separation matrices which exhibit differential interaction (uptake and binding) with two or more molecular substances, or groups of such substances. Although the illustrations here are in regard to packed particle bed chromatography formats they may also function with other formats such as filters, filtration beds, monolithic columns, expanded bed columns, radial flow columns, etc. The different formats such as two columns in series can employ similar base matrix supports and similar or even identical ligands to bind targets. These include sulfopropyl (SP) and sulfonate (S) cation exchange ligands or quaternary amine (Q) anion exchange ligands in the case of ion exchange. The added selectivity comes from modifying one matrix, such as that related to the first column, compared to the second. Suitable alterations include a. modifying the matrix with a neutral polymer coating which acts to sterically, diffusively, or otherwise limit adsorption binding of some target mixture components with the ligands on the matrix, b. tethering ligands to matrix using a polymer which acts similarly to alter ligand interaction with components in the mixture to be separated.

One or more such modified matrices may be used in series, either with unmodified matrix or with each other. In one aspect of the invention, the sequential two (or more) columns are run using the same mobile phase for adsorption with little significant alteration of the mobile phase. Thus the mobile phase buffer allows some components of the mixture to be separated to adsorb to the first column, but the target to flow through the first column. This same flow through then becomes, with minor or no alteration, the adsorption mobile phase for target on the second column. This allows significant saving of time and money during both process development and operation. Unlike in US 2009- 0118476 (above), and other typical methods of processing proteins via series linked ion exchange columns, or other solid phase based separation devices, the target does not need to bind to both columns, or be sequentially bound then eluted from one column before being adsorbed onto a following column, with each binding step requiring a different adsorption and elution phase.

An ideal situation is for the first column to bind none of the target material and exhibit significant affinity and capacity for unwanted contaminants. The second column may then bind other contaminants in addition to target allowing for a more effective, perhaps conductivity elution gradient mediated, separation of target from contaminants. However other possibilities can be considered as they may be commercially viable. To begin with, the first column may be chosen such that under relatively similar mobile phase conditions it affords significant capture of contaminant protein and a detectable but low degree of binding of target protein. The excess target flows through onto the second column. If the contaminant protein(s) exhibits greater molecular affinity for the initial capture surface it may displace target initially bound on the first column and in doing so promote its transfer to the second column. Such displacement effects mean that the efficiency of target capture on the second column, or target loss on the first column, are not drastically compromised as contaminant levels increase. Similar separation processes may be useful in separating serum antibody proteins from major contaminants of such serum albumin protein when processing blood proteins, or separating lactoferrin from antibodies in recombinant milk. It is also conceivable that a sample may contain not one but two target substances which need to be isolated from each other and from various contaminants. One example is isolating the valuable proteins alpha-lactalbumin and lactoferrin from milk samples. In this regard, one target binds preferentially on the first column and the second target bound preferentially on following column without significantly changing the mobile phase between columns. This should result in faster and more economically viable processes, especially when displacement effects can be used to enhance the efficiency of the process under increased load concentrations. Naturally the above might not be limited to two columns and one or two targets but with simple manifold fluid flow control could be expanded, in serial or parallel manner, to process several targets from one original sample, all using the same adsorption mobile phase.

More specifically, in one embodiment of the invention, it is provided a process for the separation of a mixture of macromolecules such as proteins or nucleic acid polymers, including polymer modified molecules comprising using two or more capture columns, which contain capture media with similar ligands in series such that adsorption mobile phase flow through from the first column is used as adsorption mobile phase for the second or follow-on column; wherein the capture columns separate the macromolecules based on similar interactions between the proteins and the capture media, further wherein said first column is not solely a size exclusion column and the capture interactions are not solely hydrophobic interactions.

In a related embodiment, it is provided a process for the separation of a mixture of macromolecules including polymer modified molecules comprising: (a) setting up, in series, two or more capture columns, which contain capture media which due to polymer or other modifications differ in their ability to capture selected macromolecules, (b) loading the first capture column and collect the adsorption mobile phase flow through; (c) performing, optionally, a simple modification of the flow through mobile phase; and (d) loading the second capture column with the flow through which contains sample molecules not bound on the first column; wherein the first column is not a size exclusion column and the interactions are not solely hydrophobic interactions. Preferably, the mobile phase modification is performed on-line. Preferably, the simple modification of the flow through is a pH adjustment, and both capture media are cationic exchangers or anionic exchangers.

The interactions between the macromolecules and the capture media are preferably ion exchange, immobilized metal affinity chromatography (IMAC), mixed mode, or protein A affinity interaction or similar protein affinity interactions, as well as other affinity interactions.

Also preferably, the two media differ in secondary construction such as an external coating or layer which effect the selective capture of the macromolecule targets. Alternatively, the two capture media are combinations of chromatographic media, filtration (scavenging) media, monolithic media, and other solid phase format separation media.

The two capture media can be both charged or both uncharged.

The proteins or modified proteins or other macromolecules to be separated may differ in their ability to be transported into the two capture media, such that one molecular type is largely excluded from the first capture media allowing it to flow through to the second capture media. For example, the different proteins to be separated including modified proteins differ in size such that the larger molecules are excluded from the first media. Alternatively, the proteins excluded from the first capture media are polymer modified proteins carrying a polymer which exhibits phase incompatibility with the external modification of the first capture media. For example, the proteins being separated can be (1) a mixture of mono-PEGylated, oligo-PEGylated and non-PEGylated proteins; (2) a mixture containing glycosylated and non-glycosylated proteins; (3) plasma proteins including antibodies and serum albumin; (4) antibody or other recombinant proteins being separated from milk, plant or other feed proteins. Furthermore, the macromolecules to be separated could include nucleic acid polymers and proteins, or nucleic acid polymers and polymer modified nucleic acid polymers.

When the macromolecules to be separated includes polymer modified proteins (e.g. PEGylated proteins) and non- modified proteins, or plasma proteins including antibodies and serum albumin, the first capture media can be a secondarily modified S or SP SEPHAROSE™ media and the second capture media can be a MACROCAP™ SP media. For example, the first capture media can be SEPHAROSE™ XL, FRACTOGEL^® or other media featuring polymer tether surface coatings. Furthermore, the second capture media can be SEPHAROSE™ Fast Flow, SEPHAROSE™ High Performance,

MACROCAP™ or analogous ion exchange media.

The two media used could have similar ligands and ligand density per unit area but different pore size distribution. These media facilitates the separation of, for example, a mixture of IgM and other antibodies. Under moderate loading conditions IgM flows through the first capture media but binds the second media, while IgG and other antibodies bind to the first column.

A mobile phase conduit can be optionally included such that the second column is able to be perfused and eluted independent of the first column for subfractionation.

In another embodiment, the invention provides a process for the separation of a mixture of macromolecules which include proteins and polymer modified proteins, comprising: using two or more capture media with similar ligands with the media in series such that adsorption mobile phase flow through from the first media is used as adsorption mobile phase for the second or follow-on media; wherein the capture media separate the proteins or modified proteins based on similar interactions between the proteins and the capture media, further wherein the interactions are not hydrophobic interactions and the media are in the form of filters, filtration beds, monolithic columns, expanded bed columns, radial flow columns, fluidized bed columns, or combinations thereof. Preferably, the first media is contained in a disposable buffer reservoir to adsorb contaminants from the mixture. The target macromolecule collected from the disposable buffer reservoir is loaded, without modification to the mobile phase, onto a column containing the second media to effect target capture.

In yet another embodiment, the invention relates to the use of a dual step approach for protein separation by chromatography, filtration, expanded bed, magnetic bead or bed, longitudinal or radial flow, simulated moving bed or other separation approaches.

In a variation of the embodiment, the invention relates to the use of a dual step process in a kit format in relation to sample preparation or assay for analytical, medical, diagnostic, process analytical or other uses based on either on-, off- or at-line monitoring by normal or high throughput method.

In another embodiment, the invention relates to the use of a dual bed column approach for protein separation where two or more distinct bed compartments are encased in a single physical column or column assembly.

Figure 1 shows an illustrative picture of how a neutral polymer coating, i.e. "lid", in the external regions of a porous particle chromatography support matrix, can be used to direct the transport and binding of larger macromolecules from the particles in the first column to particles in second column.

Figure 2 shows a confocal fluorescent microscopy showing aqueous polymer immiscibility with fluorescently labeled PEG of relatively small MW (hydrolyzed FITC PEG5000-NHS from Shearwater Polymers) is excluded from media with a neutral dextran polymer layer.

Figure 3 compares general trends in equilibrium (Qmax, 70 mM NaAcetate pH 5.3) and dynamic binding (QBIO , 25 mM NaAcetate pH 5.3) capacities for IgG on various dextran T fraction neutral lid modified SEPHAROSE™ Big Beads prototype gels from Table 1. For details see Example 2 and Table 1 notes.

Figure 4 shows cation exchange fractionation of a typical PEGylation reaction product mixture (mPEG20000 NHS modified cytochrome C protein) on commercial MACROCAP™ SP column under conditions of conductivity and pH where the oligoPEGylated protein flows through but the monoPEGylated protein and native protein fractions bind and are sequentially eluted respectively in two major peak regions when conductivity is increased.

Figure 5 shows the same PEG-protein reaction mixture as in Figure 4 run using same buffers but in dual column process where the first column, a prototype dextran lid modified SEPHAROSE™ 6 Fast Flow with S ligands, has a neutral lid which allows the unmodified protein to bind, yet directs flow through of the PEGylated protein

(oligoPEGylated and monoPEGylated) fractions to second commercial SP column (MACROCAP™ SP) where the oligoPEGylated fraction flows through but the monoPEGylated fraction binds.

Figure 6 shows the dynamic binding capacities at 10% breakthrough (QB10%, 50 mM NaAcetate pH 4.8) for various sample proteins on two commercial SP ligand modified ion exchange media SP SEPHAROSE™ Fast Flow (SPFF) and SP

SEPHAROSE™ XL (SPXL).

Figure 7 shows the dynamic binding capacities in SPFF and SPXL media from Figure 6 shown as ratios in order to illustrate significant selectivity differences between the two media in regard to certain protein pairs which often occur in biological samples (e.g. IgG and serum albumins such as BSA, or lactalbumin and lactoferrin).

Figure 8 shows the possible arrangement of dual columns for fractionating serum albumin and IgG or alpha-lactalbumin and lactoferrin under conditions noted in Figure 6.

Figure 9 shows the dual column series arrangement for fractionating of mixture of IgM and smaller MW proteins. The first column excludes the large (approx. 900 kDa) IgM molecules which flow through where they are captured on the second large pore size column (e.g. MACROCAP™ SP) and can be eluted by increasing mobile phase conductivity. In this figure the first column is dextran lid modified SP media however IgM exclusion might also occur on commercial media (Figure 10).

Figure 10 shows loading and flow through of IgM (1 mg/ml) sample on SP

SEPHAROSE™ High Performance commercial media. Note that most of the protein is in the flow through and not eluted by conductivity gradient.

Definitions

Capture based separations are those where molecules bind to solid phase surfaces, often chromatography media, monolithic media, or filtration media (see below). It is distinguishable from size exclusion chromatography where target or contaminant molecules are separated on the basis of differential migration rates, due to available volume or partition, rather than on being differentially adsorbed.

The term "ligand" is used herein in its conventional meaning for an entity comprising a functional group capable of interaction with a target compound. Examples of groups of ligands are positively charged groups (anion exchange ligands); negatively charged groups (cation exchange ligands); mixed mode (e.g. combined charge and hydrophobic group) ligands; groups with a biological affinity for a specific target compound, such as the affinity of an antigen for an antibody (affinity ligands); etc.

"Media" refers to solid (insoluble) phase material which in ligand modified form is used in capture based separations. This might refer to filter, bead, particle, porous particle, porous bead, fabric, membrane, monolithic or related supports used in filtration, packed bed chromatography, fluidized bed chromatography, expanded bed

chromatography, radial flow chromatography, monolithic column chromatography.

"Oligo" as in oligoPEGylated protein refers in the present document to having more than one and thus includes "di" as in diPEGylated protein.

"Similar capture ligands" are expected to show similar target binding. In the case of ion exchange ligands similar ligands can be molecularly identical anion or cation ligands, or molecularly similar ligands (e.g. S SP or CM ligands) or similarly charged anion or cation exchange ligands. In the case of affinity ligands similarity is defined as ligands which show comparable affinity to the same targets and interact with such targets at the same molecular sites.

"Similar mobile phases" induce analogous interactions (e.g. binding) between targets and similar ligands.

"Polymer coating" in the external regions of a porous chromatography support matrix is disclosed previously, see US 7,208,093, the disclosure of which is hereby incorporated by reference in its entirety.

The outer layer can be considered as a "lid or shell" that surrounds a

chromatographic particle of conventional composition, hence the denotation "lid beads" that is sometimes used for this kind of particles. Note that in reality the polymer coating may be attached to all surfaces, including possibly those of pores in the outer regions of the particle, and can be uncharged although the underlying particle surfaces may be charged or contain other ligands. Other polymer coatings noted below include dextran (SEPHAROSE™ XL media type) or polyacrylamide (FRACTOGEL^® media type) or analogous polymer tether coatings where the polymer itself may carry charged or other ligands, and the polymer coating may cover all surfaces.

Detailed Description of the Invention

The present invention relates to novel methods for the fractionation of a complex mixture of proteins or other macromolecules or compounds, using two or more capture chromatography columns, with similar types of ligands and base matrix. One compound or group of compounds binds to the first column and the other compound or group of compounds binds to the second column. This method uses the same mobile (adsorption buffer) phase for both columns, allowing the flow through from the first column to be used directly on the second column. This simplification can lead to considerable savings in process development and operation. As noted above the approach can also be extended to processes with more than one target with more than two columns used in serial or parallel manner via the obvious use of various valves and manifolds.

A first aspect of the invention is a separation method which uses two or more columns (or other solid phase ligand presenting supports) that contain similar capture ligands or groups under similar mobile phase conditions to effect separation of target and non-target proteins or other substances in a manner where flow through from the first column is used directly as adsorption step mobile phase on the second column.

This selectivity is made possible by heterogeneous chemical structuring of the chromatography media (sometimes referred to as resin), the particular properties of targets and contaminants to be fractionated, and the separation conditions. In this regard the proteins, other compounds or groups to be fractionated by adsorption (capture) are not expected to interact differently with ligand groups on the two capture columns run in series, than on analogous capture filter media run in series, or on capture filter (sometimes used as a scavenging filter) in series with a capture column. Similarly the adsorptive surfaces may be part of a radial column, or monolith, or cyrogel or other capture separation media. This latter is distinguished from the situation where one of the two first separation steps involves fractionation only on the basis of size, or where two or more orthogonal capture steps are used. In the present method two or more capture separation steps are involved, though one or more may have a size separation component which acts in addition to a capture component. In certain embodiments, both columns are chosen from commercially available media, and in others a mixture of commercial and other media, including media which has been modified, in part, to enhance the separation advantages of the present invention.

Some details related to the method in regard to media design and column arrangement are provided below. In considering certain phenomena which are used, in the present invention, to modify uptake and binding of proteins and other macromolecular substances, such as nucleic acids, glyco- or other polymers three general types of phenomena, which heretofore have not been considered in regard to processing, will become apparent. They may function together or independently, and can function for both neutral and charge group or other ligand modified polymers localized at separation media surfaces.

A. Steric inhibition as when a polymer layer sterically hinders transport of macromolecule to the surfaces of porous or nonporous chromatography particles. B. Diffusive inhibition as when a hydrophilic polymer layer provides a viscous microphase region osmotically or otherwise passively hinders transport of macromolecule to the pores and capture surfaces.

C. Aqueous polymer rich phase partition and miscibility as when a hydrophilic polymer layer inhibits transport of a certain macromolecule, or polymer modified substance due to phase partition related phenomena. Such phenomena are well appreciated in the literature and in terms of partition coefficients the efficient transport of a macromolecule into the polymer enriched phase can be controlled by operator influenced factors such as mobile phase temperature, pH, salt component type and salt concentrations.

What the above three types of phenomena have in common is that they can be controlled by addition of polymer coatings to chromatography media and can act to significantly differentiate the selectivity and dynamic capacity of separation media which are relatively similar except for the surface presence of the polymer layer. In addition this additional functionality behavior can be influenced to some extent by operator control of standard operating variables such as temperature, buffer salt compositions, and pH.

Polymer coating of capture media is not new. SEPHAROSE™ XL which comes in Q and SP or S forms is essentially SEPHAROSE™ Fast Flow media where the surfaces are coated with a dextran layer to which ion exchange groups are added at approximately the same concentration (micromole per gram of gel) as in Q and SP

SEPHAROSE™ Fast Flow media. CAPTO™ ion exchange media also features similar polymer ligand tethers. The above media are from GE Healthcare. Other polymer coated ion exchange media are noted including various FRACTOGEL^® or related polymer tentacle media from Merck, Darmstadt. However when selecting amongst such media to optimize one or more unit operations in a process, media choices would typically be made in regard to optimizing performance in each unit operation not with regard to the possibility to use the same mobile phase for two unit operations. Similarly while target flow through may sometimes be used at end of a process to remove a contaminant, normal procedure calls for first column in a process to effect target capture so as to achieve target concentration and process volume reduction effects. The approach here where contaminant not target capture may occur on the first column, with target capture on the second column using the same mobile phase is thus novel. So too is the approach where bound contaminant can later be eluted from the first column and then directed from the process flow path, prior to the same elution phase being used to elute target from the second column.

Figure 1 shows an illustrative picture of how a non-interacting layer in the external regions of a porous particle chromatography support matrix, can be used to direct the transport and binding of larger macromolecules from the particles in the first column to particles in second column. However media coated with a dextran or similar neutral polymer lid may also show some ability to hinder capture of PEGylated protein due to the two other phenomena noted above. Thus, for example Figure 2 illustrates what happens when a fluorescently labeled PEG 5000 molecule sample is incubated with a prototype dextran 75000 lid modified S SEPHAROSE™ Fast Flow prototype media under conditions where the relatively low MW PEG polymer might be expected on the basis of size exclusion to transport into the porous particle. Using confocal microscopy (An exclusion mechanism in ion exchange chromatography, Harinarayan, C., Mueller, J., Ljunglof, A., Fahrner, R., Van Alstine, J., Van Reis, R., Biotechnology and

Bioengineering 95 (5), pp. 775-787, 2006) it is observed that the PEG does not enter the media but is retained in the external liquid phase. Under these conditions it is believed that not only steric or viscosity barrier hindrances are functioning to keep the PEG in the mobile phase but that the relative incompatibility of phases rich in dextran and PEG (Chromatography-free recovery of biopharmaceuticals through aqueous two-phase processing, Ana M. Azevedo, Paula AJ. Rosa, I. Filipa Ferreira and M. Raquel Aires- Barros, Trends in Biotechnology, Vol 27, pp.240-247, 2009) are functioning significantly to promote the noted behavior. More details are given under Example 1.

Figure 3 based on Table 1 (see Example 2) provides more insight to the selectivity which a neutral layer can provide in terms of affecting both the dynamic and equilibrium binding capacities of various media. In this study SEPHAROSE™ Big Beads prototype (SPBB) media (GE Healthcare Bio-Sciences AB, Uppsala, Sweden) were modified with neutral lids generated using dextran T fractions of differing MW (Mr) including Tl (1000), T10 (10000), T70 (70000) and T500 (500000). Table 1 provides dynamic and equilibrium capacities for protein samples including lactoferrin, immunoglobulin, and lysozyme. In Figure 3 and Table 1 it can be seen that for the T modified SEPHAROSE™ Big Beads prototype media, including a TO control that the general trends for dynamic binding capacity (QB10 for 5 mg/ml protein, in 25 mM NaAcetate pH 5.3, 100 cm/h and 10% breakthrough) followed the same general trend as for equilibrium capacity (Qmaxl in 70 mM NaAcetate at pH 5.25) and that these trends held as the conductivity of the Qmax solution was increased by adding NaCl to 50 and 100 mM (Table 1). In general increasing the molecular weight of the polymer decreased the ability of the media to bind antibody under equilibrium or dynamic binding conditions. Smaller proteins such as lactoferrin and lysozyme showed similar effects although to less of a degree than for larger antibody protein. It thus appears possible to produce media which under normal conditions would largely exclude larger targets in a manner tunable by varying normal factors such as pH and conductivity. Table 1 also provides some insight to the ability of the neutral lid coatings to reduce fouling of media by milk which was not pasteurized or defatted. It is assumed similar results could be obtained with other crude mixtures such as stabilized blood plasma or various clarified fermentation feeds,

Figure 4 which is related to Example 3 illustrates cation exchange fractionation of a typical PEGylation reaction product mixture (PEG20000 NHS modified cytochrome C protein) on commercial MACROCAP™ SP column under conditions of conductivity and pH where the oligoPEGylated protein flows through but the monoPEGylated protein and native protein fractions bind and are sequentially eluted in two major peak regions when conductivity is increased. The newly launched commercial MACROCAP™ media (GE Healthcare), which does not have either a neutral or charged polymer tether coating, was designed to fractionate larger targets including PEGylation reaction product mixtures and the three peak region separation seen in Figure 4 represents the present state of the art. However obviously the column binds not only target monoPEGylated protein, which shows up in several peaks related to PEGylated proteins with mole ratio 1 but different sites of PEGylation (positional isomers or PEGamers) but also native protein (larger single peak).

In Figure 5 a similar PEG-protein reaction mixture to that used in Figure 4 was run using dual column process where the first column, a prototype dextran lid modified SEPHAROSE™ 6 Fast Flow with S ligands, has a neutral lid (as in Figure 1 and 2) which allows the unmodified protein to penetrate the lid and bind. It can later be eluted in a normal fashion simply by increasing the mobile phase conductivity. The lid directs flow of the PEGylated protein (oligoPEGylated and monoPEGylated) fractions to the second, commercial SP column (MACROCAP™ SP) where the oligoPEGylated fraction flows through (due to low surface charge per unit volume) but the monoPEGylated fraction binds. For experimental details see Example 3. Note that such a column arrangement might be used in manner such that unmodified protein could be recycled for PEGylation. Dual column arrangement where initial column has a neutral lid is not the only type of process which can be optimized according to the invention. Consider, for example, the situation where one wishes to purify antibody molecule from contaminant proteins in a sample. Table 2 below shows some standard proteins' molecular weights and isoelectric points. Figure 6 which is related to Example 4 and notes dynamic binding (QB10 breakthrough) at SPFF and the polymer coated SPXL media under one set of conditions (pH 4.75, 50 mM NaAcetate). It can be seen that in spite of the two media being based on SEPHAROSE™ Fast Flow (SPFF) base matrix, and having similar ligand and ligand concentrations, they exhibit quite varied dynamic binding for different proteins. The dynamic binding data is provided in Figure 6 and the ratio of dynamic binding in SPXL to SPFF is shown in Figure 7. Figure 7 suggests that polyclonal sample of human antibody molecules tested, under the conditions studied, had greater dynamic binding capacity on SPXL than on SPFF, whereas the opposite is true for BSA (bovine serum albumin) and some other proteins. Interestingly lactoferrin appeared to show greater differences between the SPXL and SPFF media (Figure 7) than between the control (TO) and neutral dextran (e.g. Tl to T500) modified SEPHAROSE™ Big Beads prototypes (Table 1). For experimental details see Example 4.

Dynamic binding capacities and equilibrium binding capacities are a function not only of target affinity for media but also various factors related to particle pore size distribution and target size and adsorption packing efficiency as well as transport kinetics into the particle, and the particular conditions used during their measurement. However in the present context the data in Figures 6 and 7 suggest possible paths to using the present invention. Thus to separate target antibody from serum albumin contaminant SP SEPHAROSE™ Fast Flow can be used in the first column (Figure 8 A), to allow the serum albumin to preferentially be captured and the antibody to flow through and concentrate on the second column e.g., SPXL. In the case of separating the milk proteins alpha-lactalbumin and lactoferrin the opposite column arrangement of dual columns (Figure 8B) might be used with lactoferrin showing less binding to the first SPXL column but more to the second SPFF column.

As noted above, and supported by experimental data in Tables 1 and Table 2 as well as Figure 7, the dual column approach is amenable to larger scale processing of native or recombinant proteins and other substances associated with various feeds including those related to milk, plants, and blood.

As noted above in regard to figures 1, 2 and 5, some polymers such as polyethers are not phase compatible (e.g. readily miscible) in solution with other polymers such as the polysaccharide dextran. Thus an aqueous mixture of dextran and poly(ethylene glycol) (PEG) polymers in solution will often rapidly separate into two immiscible phases both of which contain water (Azevedo, A.M. et al., 2009, see above).

Such phenomena also appear to affect the interaction of polymer modified proteins or other compounds with porous media which is also modified with polymers (see below). In this regard, commercial SP SEPHAROSE™ Fast Flow 6 (SPFF) media and SP SEPHAROSE™ 6 Fast Flow XL (SPXL) media (based on dextran polymer SEPHAROSE™ 6 Fast Flow matrix and having similar sulfopropyl (SP) ligands at similar grafting density per ml of gel) exhibit different chromatographic interactions and binding capacities with native and PEG modified proteins. The XL media has good capacity for many native proteins but not for PEG-modified proteins. So a column series of XL type media followed by SEPHAROSE™ Fast Flow, SEPHAROSE™ High Performance or MACROCAP™ should bind native protein on the first column (where it can subsequently be removed or recycled back to the PEGylation reaction) while PEG protein flows through and binds on the second column where it is concentrated and purified in manner analogous to Figure 5. In the case where PEGylation is extensive and the target PEG-protein is large SP SEPHAROSE™ 6 Fast Flow could be replaced by media with a larger mean pore size such as MACROCAP™ SP. The latter appears to offer lower ligand densities than the other media, however ligand density per unit area is similar as judged by elution conductivity for proteins which bind equally well to all the media. (Fee and Van Alstine, 2006).

Such polymer phase incompatibility may also be leveraged to promote capture selectivity in regard to the present invention and different polymer modified capture media. Especially as today not only PEG but many other polyethers and also

polysaccharides such as dextrans and starch polymers are being used to modify proteins and other biopharmaceuticals. Thus media modified with other polymers which show active polymer phase incompatibilities may also be of use in dual column approaches using separation media coated with polymers exhibiting such solution phase

incompatibility with the polymers used to modify the biopharmaceuticals.

Figures 9 and 10 relate to Example 5 and demonstrate the use of the selectivities noted above in purification of large macromolecules such as IgM (approx. MW 900 kDa) which are not polymer modified. Thus, in keeping with Figure 5 and results shown in Table 1 (see above) it is envisaged that a serum or other sample with IgM would be loaded onto a neutral lid ion exchange gel such as T10 or T70 (Table 1, Example 2) so that it passes through the first column and in the same mobile phase is captured by a second column. In the example given the second column is MACROCAP™ SP whose lack of polymer coating and large pore structure afford good capture of IgM (see MACROCAP™ SP Data File 28-40005-84 AA available via GE Healthcare's website). Choice of first column and operating conditions would have to take into account contaminant proteins to be removed. Thus when the sample also contains IgG and related antibodies of under 200 kDa then, per Table 1, the T10 or T70 lid modified type of capture media may be a better choice as under some easily identified operating conditions it might still afford reasonable uptake of contaminant IgG. Alternatively one might use a column which affords little capacity for IgM but is known from the literature to offer good capacity for IgG. An example of this is SP SEPHAROSE™ High Performance (SPHP). Figure 10 shows loading and salt gradient elution of IgM (1 mg/ml) sample on SPHP where it can clearly be seen that according to UV absorption most of the protein flows through the column.

It should be noted that the above examples typically feature dual column pairs of media which is either commercial media or lid modified media which can be produced commercially. It is important to understand that in addition to molecular weight the polymer lid structure is naturally affected by polymer grafting density. Thus the selectivity exhibited by a set of polymers of different MW, as in Table 1 , might be reproduced using same MW polymer grafted at different surface grafting densities; or via different grafting methods such as multi-focal versus single site grafting.

As noted above various types of separation matrix might be modified and used in the described fashion. In the case of solid or porous particles (beads) in a packed bed column, beads of any suitable size, such as in the interval 5-1000 microns in diameter, preferably 10-500 and more preferably 10-100 microns in diameter. An advantageous embodiment of the present invention, from the point of flow and column packing, is a relatively large sized bead, such as 30 - 200 microns in diameter. The support of the separation matrix according to the invention may be porous resin or other media such as filter media or monolith. In a specific embodiment, the supports are porous but made from cross-linked polymers, which improves its rigidity. The media should afford similar though not necessarily identical strengths of interaction. Neutral polymer lids can be produced from variety of polymers including dextran, agarose, or PEG. Their thickness will vary from 1 to 20 microns based on MW of the polymers employed (e.g. 500 to 1000000) and their grafting density (typically expected to be below 3 g/m² of porous surface area.

The support may be prepared from any commonly used material, such as synthetic polymers or native polymers, which polymers may be cross-linked to provide rigidity. Thus, in one embodiment, the support is prepared from a native polymer, such as a polysaccharide, e.g. agarose. In another embodiment, the support is prepared from a material selected from the group consisting of agarose, dextran, polymethacrylates, divinylbenzenes, polystyrenes, silica, glass, etc. One support may be modified by attachment of neutral or charge group modified polymers such as dextrans or other polymers such as polyethylene glycols, acrylamides.

In a second aspect, the present invention relates to a method of preparing a cassette or unit structure where instead of two columns operating in series there are two porous media fields operating in series.

In a specific embodiment the cassette or unit structure includes as a first field XL media or similar media (CAPTO™) and as a second field FF or similar media.

In a third aspect, the present invention relates to a process where more than one dual step is used to sequentially fractionate mixtures in a manner to save buffer and other costs.

Examples

The present examples are presented herein for illustrative purpose only, and should not be constructed to limit the invention as defined by the appended claims. Example 1:

Confocal Microscopy of PEG Polymer Diffusion into Dextran Lid Coated S

SEPHAROSE™ Fast Flow

Neutral dextran 75 coated S ligand modified SEPHAROSE™ 6 Fast Flow media was produced as described in the next examples and was examined for passive uptake of fluorescent isothionate labelled PEG 5000 (FITC-PEG5000) from Shearwater Polymers (now Nektar) Huntsville, Alabama. References to confocal microscopy and a description of the general system used can be found in Harinarayan, C. et al. Id The chromatography media was incubated in finite bath experiment with 8% FITC-PEG5000 (hydrolyzed form of FITC-PEG5000-NHS, Shearwater Polymers now Nektar) in 15mM NaAcetate, pH 5, with NaCl added to 10 mS/cm which allowed for significant passive uptake of the PEG into MACROCAP™ SP media under control studies (not shown). Results were analyzed optically (Figure 2) in regard to both detecting if fluorescent material had entered the media or remained in the external liquid phase, and in regard to comparing differences in light intensity at the media boundary. The results clearly indicate that under these conditions little if any PEG is diffusing into the media even though the relative size of the PEG 5000 is approximately comparable to a small 30 kDa protein (Fee and Van Alstine, 2004). Note that the lid media used here is similar to the lid media used in Example 3. Example 2:

Equilibrium and Dynamic Capacities of Proteins on Dextran Lid Coated SP

SEPHAROSE™ Big Beads Media

Experimental:

Commercial SEPHAROSE™ Big Beads (SPBB) base matrix media (catalog 17- 0657-03, GE Healthcare) was modified with a neutral dextran lid and SP ligands according to WO 1998/039364. See also next section. Different dextran samples (GE Healthcare) over a range of Mr molecular weights Tl (1,000), T10 (10,000), T70 (70,000) and T500 (500,000) were used with resulting control and media prototypes noted as shown in Table 1. Control media included commercial SEPHAROSE™ SP Big Beads (SPBB), plus the above prepared SPBB with slightly different ligand grafting than the commercial media (T) and media sample where polymer grafting reaction carried out without dextran (TO). All salts and reagents were pro analysis from Merck or Sigma. Methods and references to equilibrium and dynamic binding capacity determinations can be found in above reference to Harinarayan, C. et al. Instrument used was AKTAexplorer 100 using UNICORN™ software (GE Healthcare) and Millipore GS filters. Equilibrium binding capacities (Qmax in mg protein per ml wet gel) were measured in microtitre plates over 24 hour period while dynamic binding capacities (QBIO in mg protein per ml wet gel) were measured in XK16/20 columns (GE Healthcare) with 10 micron nets using 200 cm/h flow rate. Lactoferrin (Aria AB) equilibrium capacity measured in microtitre format in 30 mM NaPhosphate buffer at pH 6.8 with varied added NaCl at 0 mM

(Qmaxl), 50 mm (Qmax2) or 100 mM (Qmax3). IgG (Sigma G-5009) equilibrium capacity measured in microtitre format in 70 mM NaAcetate buffer at pH 5.25 with varied added NaCl at 0 mM (Qmaxl), 50 mm (Qmax2) or 100 mM (Qmax3).

IgG dynamic binding capacities all tended to decrease as flow increased from 100 to 800 cm/h (not shown) IgG (Sigma G-5009) dynamic capacity QBS_% measured using 5 mg/ml solution in 25 mM NaAcetate pH 5.3 solution. Column was XK 16/20 column with 10 cm bed height. Lysozyme (Sigma L-7001) dynamic capacity QBS_% measured using 5mg/ml solution in 25 mM NaAcetate pH 5.3 solution. In one study milk was passed through columns to reproduce extraction of protein from whole milk. In these studies the 10 micron nets supplied with the columns were replaced with larger 22 micron nets to reduce the effect of net fouling on backpressure measurement as an indication of chromatography media bed fouling. A Backpressure (bar) indication of fouling after 120 column volumes of non-Pasteurized, non defatted milk from cows run through column at room temperature

Results:

Neutral lid appears able to block uptake of proteins to some extent over fairly wide range of salt conditions. Similar effects were seen for both equilibrium binding and dynamic binding. In general media prototype capacities for larger proteins like IgG appear to be more influenced by the lids. Smaller proteins such as lactoferrin or lysozyme appear to be less influenced, especially by higher MW dextran lids. Protein size does not appear to be the only factor influencing lid effects with lysozyme appearing to be more excluded than lactoferrin in some experiments. Ability of the neutral lid to inhibit fouling by milk feed was indicated by decrease in backpressure seen after 120 column volumes of non-Pasteurized, nondefatted milk were pumped through the dextran 500000 modified lid column

Table 1

Equilibrium and Dynamic Binding Capacities for Proteins on Neutral Dextran Lid SPBB.

Notes

1. SPBB is commercial SP SEPHA OSE™ Big Beads (GE Healthcare), T is laboratory generated SPBB variant with somewhat lower SP ligand density, TO is T media treated with lid modification chemistry without dextran, Tl is T media lid modified with dextran of mean MW 1000, T10, T70 and T500 are T media lid modified with dextrans of mean MW 10000, 70000 and 500000, respectively.

2. Lactoferrin (supplied by Aria AB) equilibrium capacity Qmax measured in microtitre format in 30 mM NaPhosphate buffer at pH 6.8 with NaCl added to 0 mM (Qmaxl), 50 mm (Qmax2) or 100 mM (Qmax3).

3. IgG (Sigma G-5009) equilibrium capacity Qmax measured in microtitre format in 70 mM NaAcetate buffer at pH 5.25 with NaCl added to 0 mM (Qmaxl), 50 mm (Qmax2) or 100 mM (Qmax3).

4. IgG dynamic binding capacities at 100 cm/h. IgG dynamic capacities on all media tested decreased significantly (e.g. 80%) as flow increased from 100 to 800 cm/h but this could reflect much shorter residence times as much as flow effects.

5. IgG (Sigma G-5009) dynamic capacity Q_Bs measured using 5mg/ml solution in 25 mM NaAcetate pH 5.3 solution. Column was XK 16/20 column with 10 cm bed height.

6. Lysozyme (Sigma L-7001) dynamic capacity Q_Bs measured using 5mg/ml solution in 25 mM

NaAcetate pH 5.3 solution.

7. Backpressure (bar) indication of fouling after 120 column volumes of unpasteurized, non defatted milk from cows run through column at room temperature. Columns had 22 micron nets.

Example 3:

Lid Facilitated IEC for Polymer Modified Proteins

The single column (MACROCAP™ SP) purification of a PEG20000- Cytochrome

C reaction product mixture using commercial MACROCAP™ SP media (Figure 4) column is compared to dual column processing involving a neutral dextran lid coated S SEPHAROSE™ media column in series with a MACROCAP™ SP media column (Figure 5). Note that in Figure 5 the media is called Dx75 S SEPHAROSE™ media. In this case the number 75 does not (as in Table 1) refer to the MW of dextran used (which was dextran T 500 of 500000 MW) but to other aspects of the synthesis.

Example 3.1 :

Synthesis of Lid Dx S SEPHAROSE™ 6 Fast Flow

(A) Lid Dx AHP SEPHAROSE™ 6 Fast Flow: Layer functionalization of Allyl SEPHAROSE™ 6 Fast Flow with dextran 500

Chemicals and equipment:

2000 ml glass beaker; Three necked 500 ml reaction flask; Electric laboratory stirrer with Teflon stirrer/paddle on glass spindle; Magnet stirrer with heat control and a water bath; Glass filter funnel D3 ; allyl-SEPHAROSE™ 6 Fast Flow with an allyl content of 0,259 mmol/ml; Bromine Mw 159,82 d=3, 12 g/ml; Sodium sulfate water free, Mw 142,04; dextran 500 with a molecular weig ht of 5xl0⁶ Da.

Method:

100 gram vacuum drained AHP-SEPHAROSE™ 6 Fast Flow was brominated under heavy agitation in the presence of 1 ,26 g water free sodium sulfate dissolved in deionized water up to a total volume of 1 liter with 0,40 ml Bromine. Theoretical layer thickness 5 μιη corresponding to 29,8 % the volume of a 90 μιη spherical particle. The brominated gel was then washed with more than 5 bed volumes of deionized water and vacuum drained on a glass filter funnel. The weight of the vacuum drain gel was 102 g.

60 g dextran 500 was dissolved in 250 g water at 50°C. After 15- 20 min all of the dextran was dissolved and all of the vacuum drained gel was loaded. The slurry was equilibrated for 15 min before 20 g NaOH was loaded together with 0,5 g Sodium Borohydride. The reaction temperature was held at 50°C. After 21 h the reaction was stopped by slowly adding of 28,6 ml concentrated acetic acid and washing the modified gel media on a filter funnel with >3 L deionized water. Yield: 91 ,2 g vacuum drained Lid Dx AHP SEPHAROSE™ 6 Fast Flow.

(B) S functionalization of Lid Dx Allylhvdroxypropyl-SEPHAROSE™

6 Fast Flow Chemicals and equipment:

2000 ml plastic beaker; Three necked 500 ml reaction flask; Electric laboratory stirrer with Teflon stirrer/paddle on glass spindle; Magnet stirrer with heat control and a water bath; Glass filter funnel D3; Lid Dx75 Allyl-SEPHAROSE™ 6 Fast Flow 90μιη;

Bromine; Sodium sulfate; Sodium Sulfite; 0,1 M NaOH.

Method:

75 ml (75 g vacuum drained) Lid Dx Allyl-SEPHAROSE™ 6 Fast Flow was loaded together with 2,6 g water- free sodium sulfate in about 1000ml deionized water in a 2000ml beaker equipped with a stirrer.

Bromine was added drop vise under heavy stirring until a stable yellow coloring of the slurry was achieved. After the bromination the gel was washed with 1 L deionized water on a glass filter funnel and vacuum drained. (Yield: 75,2g). The vacuum drained gel was then loaded into a 3 -necked 250ml round flask already loaded with 36g Na₂S0₃ and 96g 0,1 M NaOH. Reaction under stirring at 50°C for 19 hours.

After the reaction the gel was washed with 4 L deionized water on a glass filter funnel. Yield: 76,0 g vacuum drained Lid Dx S SEPHAROSE™ 6 Fast Flow

Example 3.2:

Chromatography of PEGylated Cytochrome C Reaction Product Mixture

(A) PEGylation of Cytochrome C

Bovine cytochrome C (Sigma Aldrich, USA) was covalently modified with monomethoxy-PEG

20 000 Mr succinimidylproprionic acid (SPA) reagent (Nektar Therapeutics, USA) in manner similar to that described in Fee and Van Alstine 2004. In many cases a reagent to protein weight ratio of 1 or 2 is enough to assure a reasonable reaction product mixture containing significant oligo-, mono- and un-PEGylated protein. In other instances the PEG reagent to protein ratio may have to be increased to 5 or even 10. This can depend on reagent, temperature, protein and buffer. In the present case 100 mg of protein and 200 mg of PEG reagent were dissolved in 20 mL of 0.05M sodium borate pH 8 solution at RT. The reaction was allowed to proceed for 40 minutes and then stopped by adding approximately 200 uL of 1M acetic acid to acidify the solution. According to UV spectroscopy reaction product mixture contained protein more or less equally divided between oligoPEGylated protein, monoPEGylated protein and unmodified protein fractions.

The above reaction product mixture would be labeled according to its protein content so that following reaction it would be 100 mg protein in 20 ml or a 5 milligram per ml mixture. Such mixtures could be diluted with mobile phase adsorption buffer, e.g. say to 1 mg per ml.

A protein reaction product mixture was diluted in adsorption buffer (see below) and loaded onto a column. If protein is at 1 mg per ml and is loaded onto a 1 ml bed volume column at 5 mg per ml of media then 5 ml of load mix would be applied to the column. Thus sample load refers to the total protein loaded onto a column divided by the volume of the column.

(B) MACROCAP™ SP Single Column Chromatography of PEGylated Cytochrome C The protein reaction product mixture was chromatographed on MACROCAP™ SP media (GE Healthcare) in TRICORN™ 5/100 column (GE Healthcare) packed to 107 mm bed height (column volume CV 2.1 ml). Sample load was 1 ml of reaction product mixture. Adsorption buffer A was 0.05M NaPhosphate (Sigma) pH 6.8 to allow for oligoPEGylated protein flow through. Elution buffer B was buffer A plus 0.4M NaCl. Flow rate was 0.2 ml/min (61 cm/h, 10.5 min residence) with gradient 0 to 100% B over 20 CV. System was AKTAexplorer 10 with UNICORN™ software from GE Healthcare. Chromatogram shown in Figure 4. Note that the three different UV traces correspond to 254 (amide), 280 (aromatic) and 420 nm with the last being characteristic for red colored cytochrome C.

(C) Dual Column Run of PEGylated Cytochrome C reaction product mixture

Chromatography equipment and method:

Chromatography system used was an AKTAexplorer system from GE Healthcare Bio-sciences.

Sample: 1ml of PEG20 Cytochrome C reaction mixture in buffer A. Buffer A: 50 mM Na-Phosphate pH 6.8, Buffer B: A + 0.4 M NaCl. Buffer A could be adjusted with NaCl to allow for oligo flow through on the second column.

The chromatography media were packed in TRICORN™ 5/100 columns giving a column volume of 2 mL.

A first column was packed with Lid Dx S SEPHAROSE™ 6 Fast Flow (see above) and a second column was packed with MACROCAP™ SP, both cation exchangers. 1 mL of a reaction mixture from PEG 20 functionalization of cytochrome C was injected on the Lid Dx S SEPHAROSE™ 6 Fast Flow column already equilibrated with A-buffer.

Results and discussion

Figure 4 illustrates control cation exchange fractionation of a typical PEG protein (cytochrome C reacted with PEG20000 NHS reagent) reaction product mixture on a commercial MACROCAP™ SP column. At the conductivity and pH chosen for the mobile phase the oligoPEGylated (polyPEGylated) protein flows through the column while various monoPEGylated-cytochrome C proteins bind as does the native protein. Bound proteins are eluted by increasing the conductivity via a linear gradient.

Figure 5 shows proof of concept results for same PEG-protein mixture as Figure 4 run on dual column process where neutral dextran polymer modified S SEPHAROSE™ 6 Fast Flow media is run in series with MACROCAP™ SP. The conditions are similar as those for Figure 4. The first column Dex 75 (dextran T500 lid modified SEPHAROSE™ 6 Fast Flow with S ligands) has a neutral lid which allows the unmodified protein to be transported through the lid and bind. It can be eluted simply by increasing the mobile phase conductivity, yet directs flow of the PEGylated protein (oligoPEGylated and monoPEGylated) fractions through to second (MACROCAP™ SP) column where under the chosen conditions of mobile phase pH and conductivity the oligoPEGylated fraction flows through but the monoPEGylated fraction binds. The monoPEGylated target protein can then be eluted in normal manner, by increasing conductivity. The multiple peaks in the monoPEGamer region represent different positional isomers with mole ratio 1.

Unmodified protein adsorbed on the first column can be eluted and collected (or discarded) simply by increasing conductivity of mobile phase and then directing the flow via a T value and can be eluted simply by increasing the mobile phase conductivity

The PEG-cytochome C protein mixture is presented as an example and the concept applies to other similar protein mixtures where target and contaminant proteins differ in MW or size. In some cases, oligoPEGylated targets (e.g., PEG-hemoglobin) function better than monoPEGylated targets and there is a desire to separate the two so as to retain the oligoPEGylated fraction. This can also be accomplished using the above principle and process. Example 4:

Dynamic Binding Capacities of Specific Proteins on Certain Commercial Media and Implications for the Dual Column Approach

In this example the dynamic binding capacities of various commercial media for specific proteins are used to provide insight into how such media can be arranged to utilise the dual column approach of the invention. It should be noted that typically one of the dual columns involves media featuring a charged polymer layer. This is exemplified by SP SEPHAROSE™ XL (SPXL) media which has the same base matrix, ligand and ligand density (micromole per ml of wet gel) as SP SEPHAROSE™ 6 Fast Flow media. The SPXL media differs in possessing a dextran polymer coating to which some of the charged ligands may be attached. Several other commercial media

(CAPTO™ ion exchangers, FRACTOGEL^® ion exchangers) feature such neutral polymer ligand tethers. The polymer layers have been ostensibly added to enhance binding capacity by increasing ligand freedom of movement to interact with target (product literature, See also Harinarayan, C. et al. Id) however other interactions can be manipulated to enhance separations in dual column mode. Experimental

Dynamic binding capacities were determined in normal manner according to common methods (see above) and for some of the proteins such as antibodies are in agreement with some reported elsewhere in the literature (e.g. Harinarayan, C. et al. Id.). Basically 1 mg per ml solution of pure protein in buffer was loaded onto a column in 50 mM NaAcetate pH 4.8 and when the flow through A280 nm reached 10% of the load this was referred to the 10% breakthrough capacity. This is normally referred to in regard to the amount of protein adsorbed onto the column (mg per ml of wet gel) at the time the breakthrough reached 10%. Proteins were from Sigma except for polyclonal human IgG (GammaNorm, Octapharma AB).

Results and Discussion

Table 2

MW, pi and Dynamic Binding Capacities of Some Proteins.

^:Q_B10 is dynamic binding capacity at 10% breakthrough at pH 4.75 with 25 mM NaAcetate.

Figure 6 shows the dynamic binding capacities at 10% breakthrough (QB10%) for various sample proteins on two SP ligand modified ion exchange media SP

SEPHAROSE™ 6 Fast Flow (SPFF) and SP SEPHAROSE™ XL (SPXL). Figure 7 presents the dynamic binding capacities in SPFF and SPXL media from Figure 6 shown as ratios to illustrate significant selectivity differences between the two media in regard to certain protein pairs which often occur in biological samples (e.g. IgG and albumin, lactalbumin and lactoferrin).

As noted in Figure 3 dynamic binding capacities often correlate with equilibrium capacities and give some indication of the general usefulness of a media to be employed to effectively capture a desired target protein under a set of conditions. Figures 6 and 7 appear to indicate significant differences between the relative capture by SPXL and SPFF media of certain proteins which are often together in mixtures and required to be separated. These include antibodies and serum albumin in some antibody preparations, or alpha lactalbumin and lactoferrin in milk. Based on these results, possible dual column arrangements were set up for fractionating (A.) serum albumin such as BSA and IgG or (B.) alpha-lactalbumin and lactoferrin under conditions noted in Figure 5. Thus to purify IgG via dual column approach the first column might be SPFF and the second column SPXL. As can be seen in Figures 6 and 7 as well as noted in the literature (Harinarayan, C. et al. IcL) SPXL offers good capacity for antibodies. Some IgG would bind to the first column but in presence of albumin or another protein with higher affinity for the SPFF some displacement would occur. The opposite column arrangement would appear more useful for fractionating alpha-lactalbumin and lactoferrin. In this case lactoferrin interacts less with the first SPXL column (which the other protein has better capacity on) and will flow through, or possibly be displaced, to the second SPFF column where it should bind with good capacity.

Example 5:

Separation of IgM on Two SP Columns in Series

Figure 9 shows dual column series arrangement for fractionating mixture of IgM and smaller MW proteins. The IgM is excluded from the first column and passes to the larger pore MACROCAP™ SP media column where it is bound and can be eluted with a gradient. The first column could be a lid column or it might simply be a column whose pore size distribution does not allow for significant uptake of the very large IgM molecule. One example is SP SEPHAROSE™ High Performance (SPHP) as noted in Figure 10.

Experimental

TRICORN™ 5/100 (GE Healthcare) packed with SP SEPHAROSE™ High Performance (CV 2 ml). Sample IgM (human), 96% pure by HPLC (not shown), Sample Load: 0.5mg/ml medium. Buffer A 100 mM sodium acetate, pH 4.75, Buffer B is Buffer A + 0.5 M sodium chloride, Flow rate 0.3 ml/min (90 cm/h)

Gradient: 0 to 100% Buffer B in 20 CV, System: AKTAexplorer 10. Results

It can be seen in Figure 10 that most of the IgM UV signal passes in the flow through and that elution of any bound protein by increasing mobile phase conductivity results in small broad peak relative to the very sharp and significant flow through peak. All patents, patent publications, and other published references mentioned herein are hereby incorporated by reference in their entireties as if each had been individually and specifically incorporated by reference herein. While preferred illustrative embodiments of the present invention are described, one skilled in the art will appreciate that the present invention can be practiced by other than the described embodiments, which are presented for purposes of illustration only and not by way of limitation. The present invention is limited only by the claims that follow.

Claims

What is claimed is:

A process for the separation of a mixture of macromolecules such as proteins or nucleic acid polymers, including polymer modified molecules comprising:

using two or more capture columns, which contain capture media with similar ligands in series such that adsorption mobile phase flow through from the first column is used as adsorption mobile phase for the second or follow-on column;

wherein said capture columns separate said macromolecules based on similar interactions between the proteins and the capture media, further wherein said first column is not solely a size exclusion column and said capture interactions are not solely hydrophobic interactions.

A process for the separation of a mixture of macromolecules including polymer modified molecules comprising:

(a) setting up, in series, two or more capture columns, which contain capture media of claim 1 which differ in their ability to capture selected macromolecules;

(b) loading the first capture column and collect the adsorption mobile phase flow through;

(c) performing, optionally, a simple modification of the flow through mobile phase; and

(d) loading the second capture column with the flow through which contains sample molecules not bound on the first column;

wherein said first column is not a size exclusion column and said interactions are not solely hydrophobic interactions.

3. The process of claim 2, wherein said modification is a pH or conductivity

adjustment.

4. The process of claim 2, wherein said modification is performed on-line.

5. The process of claim 3, wherein both capture media are cationic exchangers or anionic exchangers.

6. The process of claim 1 or 2, further wherein said interactions are ion exchange, immobilized metal affinity chromatography (IMAC), mixed mode, or protein A affinity interaction or similar protein affinity interactions, as well as other affinity interactions.

7. The process of claim 1 or 2, wherein the two media differ in secondary

construction such as an external coating or layer which effect the above selective capture of said macromolecule targets. 8. The process of claim 1 or 2, wherein the two capture media are combinations of chromatographic media, filtration (scavenging) media, monolithic media, and other solid phase format separation media.

9. The process of claim 1 or 2, wherein the two capture media are both charged.

10. The process of claim 1 or 2, wherein the two capture media are both uncharged.

11. The process of claim 1 or 2, further wherein said proteins or modified proteins or other macromolecules to be separated differ in their ability to be transported into said two capture media, such that one molecular type is largely excluded from the first capture media allowing it to flow through to the second capture media.

12. The process of claim 11, wherein different proteins to be separated including modified proteins differ in size such that the larger molecules are excluded from the first media.

13. The process of claim 11, wherein proteins excluded from the first capture media are polymer modified proteins carrying a polymer which exhibits phase incompatibility with the external modification of the first capture media.

14. The process of claim 11, wherein the proteins being separated are a mixture of mono-PEGylated, poly-PEGylated and non-PEGylated proteins.

15. The process of claim 11, wherein the proteins being separated are a mixture

containing glycosylated and non-glycosylated proteins.

16. The process of claim 11, wherein the proteins being separated are plasma proteins including antibodies and serum albumin. 17. The process of claim 11, wherein the proteins being separated are antibody or other recombinant proteins being separated from milk, plant or other feed proteins.

18. The process of claim 11, wherein the macromolecules to be separated include nucleic acid polymers and proteins, or nucleic acid polymers and polymer modified nucleic acid polymers.

The process of claim 14 or 16, wherein said first capture media is a secondarily modified S or SP SEPHAROSE™ media and said second capture media is a

MACROCAP™ SP media.

20. The process of claim 14 or 16, wherein said first capture media is SEPHAROSE™ XL, FRACTOGEL^® or other media featuring polymer tether surface coatings. 21. The process of claim 20, wherein said second capture media is SEPHAROSE™ Fast Flow, SEPHAROSE™ High Performance, MACROCAP™ or analogous ion exchange media.

22. The process of claim 1 or 2, wherein the two media have similar ligands and ligand density per unit area but different pore size distribution.

23. The process of claim 22, wherein the proteins are a mixture of IgM and other antibodies. 24. The process of claim 22, wherein under moderate loading conditions IgM flows through the first capture media but binds the second media, while IgG and other antibodies bind to the first column.

The process of claim 1 or 2, wherein a mobile phase conduit is included such that the second column is able to be perfused and eluted independent of the first column for subfractionation.

Use of a dual step approach for protein separation by chromatography, filtration, expanded bed, magnetic bead or bed, longitudinal or radial flow, simulated moving bed or other separation approaches.

Use of a dual bed column approach for protein separation where two or more distinct bed compartments are encased in a single physical column or column assembly.

Use of a dual step process in a kit format in relation to sample preparation or assay for analytical, medical, diagnostic, process analytical or other uses based on either on-, off- or at-line monitoring by normal or high throughput method.

A process for the separation of a mixture of macromolecules which include proteins and polymer modified proteins, comprising:

using two or more capture media with similar ligands with the media in series such that adsorption mobile phase flow through from the first media is used as adsorption mobile phase for the second or follow-on media;

wherein said capture media separate said proteins or modified proteins based on similar interactions between the proteins and the capture media, further wherein said interactions are not hydrophobic interactions and said media are in the form of filters, filtration beds, monolithic columns, expanded bed columns, radial flow columns, fluidized bed columns, or combinations thereof.

The process of claim 29, wherein the first media is contained in a disposable buffer reservoir to adsorb contaminants from said mixture, the target

macromolecule is collected, and loaded, without modification to the mobile phase, onto a column containing the second media to effect target capture.