Multimodal metal affinity processed capsid

Abstract

A method for separating intact adeno-associated virus (AAV) from empty AAV capsids in a buffered mixture comprising intact AAV capsids and empty AAV capsids, comprising the steps of: - contacting the buffered mixture with a first matrix, said first matrix being provided with metal affinity ligands attached to the first matrix, said metal affinity ligands having the ability to complex metal ions via three or more nitrogen atoms, - separating the empty AAV capsids from the intact AAV capsids by elution with a pH gradient, a salt gradient, a metal ion gradient or a combination thereof in the presence of multivalent cations bound to the metal affinity ligands, to obtain a purified intact AAV capsid fraction. To remove contaminating DNA in the mixture or the purified AAV capsid fraction, the method of the invention can be combined with contacting the buffered mixture or the purified intact AAV capsid fraction with a second matrix, said second matrix being provided with metal affinity ligands attached to the second matrix, said metal affinity ligands comprising two or more negatively charged carboxylic acid residues, in the presence of multivalent cations bound to the metal affinity ligands.

Description

[0001] This invention relates to an improved method for separating intact AAV capsids from empty AAV capsids and contaminated DNA.

[0002] background

[0003] Adeno-associated virus (AAV) has become a popular candidate for DNA plasmid packaging and delivery in gene therapy. This has led to the evaluation or development of numerous materials and methods for their purification. Affinity chromatography is widely popular due to its simplicity. Most commonly, biological ligands (e.g., antibody-derived ligands) are immobilized on a solid-phase chromatographic surface and exposed to a raw sample containing the desired AAV and contaminants. In principle, AAV binds, and contaminants do not. Unbound contaminants are washed away. AAV can also be recovered by chemically interfering with the interaction between the ligand and AAV.

[0004] In many cases, immobilized metal affinity chromatography (IMAC) has been considered an effective alternative to bioaffinity chromatography [1,2]. A known practice is to genetically encode a multihistidine tail (His-tag) onto a protein so that it can be captured by nickel ions immobilized on an iminodiacetic acid (IDA) or nitrilotriacetic acid (NTA) ligand. This His-tag typically contains six or more linear histidine residues. IMAC purification of His-tagged AAVs is known [3,4]. IMAC techniques for capturing non-His-tagged AAVs are considered to have no practical application [3,4] because this eliminates the need to encode the His-tag onto the AAV. Purification of nucleic acids using IMAC is known [5,6].

[0005] Currently, no IMAC-purified His-tagged products have received approval from international regulatory agencies, which may reflect the fact that nickel ions, most commonly recommended for capturing His-tagged proteins, are carcinogenic. It is difficult to achieve, and even more difficult to prove, that nickel is completely removed from the final product because it binds very tightly to the tag and can bind to other sites on the protein. Toxic metals also increase the cost of hazardous waste disposal, a concern for those working with IMAC-nickel columns. Even at saturated levels of sodium chloride or guanidine hydrochloride, heavy metal ions like nickel can bind to proteins. Chelating agents may reduce their levels, but chelating agents can affect the stability of the AAV capsid, as the capsid requires magnesium and / or calcium ions to maintain its structural integrity.

[0006] Another limitation of affinity chromatography (including bioaffinity chromatography and IMAC) is that it cannot fully meet the purification requirements of AAV. It indiscriminately captures capsid fragments, defective capsids, empty capsids, and the desired intact capsid containing the therapeutic plasmid DNA intended for internal delivery. The term "empty capsid" refers to an AAV capsid that does not contain the gene therapy DNA plasmid intended for delivery to a patient. The terms "capsid fragments" and "defective capsid" refer to incomplete or nonfunctional capsids resulting from incorrect assembly or damage.

[0007] Another limitation of affinity chromatography is that it cannot remove contaminating DNA to sufficiently low levels to ensure patient safety and regulatory compliance. Contaminating DNA may be present in the form of host cell-derived DNA or plasmid DNA outside the capsid. AAVs have shown a tendency to bind such DNA to their outer capsid surface, leading to co-purification via affinity methods. This non-specifically bound DNA on the capsid surface can also interfere with subsequent methods for removing empty capsids.

[0008] The use of strong anion exchangers to reduce empty capsids and contaminated DNA is known [7-9]. Empty capsid removal is effective for some AAV serotypes, partially effective for many serotypes, and almost ineffective for others. Weak anion exchangers have been shown to be ineffective for separating empty AAV capsids from intact capsids.

[0009] The term "strong anion exchanger" refers to anion exchange ligands that maintain a consistent charge within a pH range of approximately 2 to 13. They are typically found in the form of quaternary ammonium compounds, commercially known as Q, QA, QAE, QAM, TEAE, TMAM, or TMAE, which respectively refer to quaternary ammonium, quaternary amino, quaternary aminoethyl, quaternary aminomethyl, triethylaminoethyl, trimethylaminomethyl, or trimethylaminomethyl.

[0010] The term "weak anion exchanger" refers to anion exchange ligands that lose their charge at high pH values. They are typically present in the form of tertiary amines, the most common of which is DEAE (diethylaminoethyl). DEAE has little effect on capsid separation because such separation requires alkaline pH values ​​close to pH 9 or higher, at which point DEAE has lost most of its charge. Even if it might be able to largely separate the desired capsid from the unwanted one, its charge loss within the necessary pH range reduces its capacity to the point that separation is not of practical value for manufacturing. This is especially true when strong anion exchangers perform superiorly.

[0011] It is known that cation exchange chromatography can achieve little or no actual separation of empty capsids from the desired intact capsid. In some respects, any separation it provides is less effective than that provided by strong anion exchangers.

[0012] Borates are sometimes referred to as synthetic lectins because they can form covalent complexes with certain hydroxyl groups on a variety of carbohydrates and glycosylated compounds. Borate complexes with cis-diols impart a negative charge to the complex.

[10] Borates can also alter the metal ion selectivity of metal chelating sites.

[11] Invention Summary

[0013] A method has been developed using immobilized metal affinity chromatography (IMAC) ligands with a remarkable ability to capture AAV, which lacks His-tags or other genetic modifications that promote binding to immobilized metal ions. It requires the use of cationic metal affinity ligands. It does not require the use of toxic heavy metals. It is effective against multivalent cationic species essential for human physiology.

[0014] The method offers a further surprising ability to separate empty capsids from intact capsids and to reduce DNA contamination more effectively than known methods. The capability of the basic method of this invention is compounded by combining it with methods for further reducing empty capsid and / or DNA content, such as by combining it with anion exchange chromatography, density gradient centrifugation, or other techniques.

[0015] In one aspect, the present invention relates to a method for separating an intact adeno-associated virus (AAV) capsid from an empty AAV capsid in a buffer mixture comprising an intact AAV capsid and an empty AAV capsid, comprising the following steps:

[0016] - The buffer mixture is contacted with a first substrate, the first substrate having metal affinity ligands attached to the first substrate, the metal affinity ligands having the ability to complex metal ions through three or more nitrogen atoms;

[0017] - In the presence of a multivalent cation bound to a metal affinity ligand, an empty AAV capsid is separated from a complete AAV capsid by elution using pH concentration, salt concentration, metal ion gradient, or a combination thereof, to obtain a purified complete AAV capsid fraction. For simplicity, the phrase "metal affinity ligand with the ability to complex metal ions through three or more nitrogen atoms" will also refer to "cationic ligand-metal solid phase".

[0018] In one embodiment of the method of the present invention, the first matrix is ​​loaded with polyvalent cations before, during, and / or during elution of the buffer mixture with the first matrix. In another embodiment of the method of the present invention, the solid matrix can be equilibrated with a buffer solution having an alkaline pH.

[0019] In another embodiment of the method described in this invention, the metal affinity ligand of the first matrix is ​​selected from the group consisting of: diethyltriamine, triethyltetraamine, tetraethylpentamine, polyamidoamine, polytriethylamine, deferriphosphine, N,N-bis(2-aminoethyl)-1,2-ethylenediamine, and tri(2-aminoethyl)amine (TREN).

[0020] In another embodiment of the method described in this invention, the polyvalent metal cation present during the elution process may be selected from the group consisting of iron (III), manganese (II), copper (II), zinc (II), cobalt (II), magnesium (II), calcium (II), barium (II), nickel (II), and combinations thereof. Specifically, copper (II), magnesium (II), calcium (II), or combinations thereof may be used. In one embodiment of the method described in this invention, the metal affinity ligand may be charged with a polyvalent cation before elution, and the AAV capsid can be eluted by increasing the concentration of a second polyvalent cation.

[0021] In another embodiment of the method described in this invention, an elution gradient is formed by applying an increased concentration of polyvalent metal cations. The concentration of the polyvalent cations at the end of the gradient can be in the range of 0.1 mM to 200 mM, or 1 mM to 100 mM, or 2 mM to 50 mM, or 5 mM to 25 mM. In some embodiments, the polyvalent ions can be magnesium, calcium, barium, copper, or mixtures thereof.

[0022] In another embodiment of the method described in this invention, a gradient buffer for separating intact and empty AAV capsids by elution has a salt concentration in the range of up to 500 mM. Typically, elution is performed by increasing the salt concentration in the ranges of 1 mM to 500 mM, 2 mM to 250 mM, 3 mM to 125 mM, 5 mM to 60 mM, or 7 mM to 30 mM.

[0023] The salt concentration range for any particular AAV serotype needs to be determined experimentally because each serotype has different surface chemistry characteristics. This phenomenon is well known in the field of ion exchange chromatography, where some serotypes bind more tightly than others and require more salt for removal. The amount of salt required for elution also depends on pH, but the pH-dependent pattern differs between cationic metal affinity matrices and strong anion exchangers. Binding on strong anion exchangers becomes stronger as pH increases. The strongest binding on TREN occurs at approximately neutral pH and weakens at higher and lower pH values. This means that more salt is required for elution at pH 7 than at pH 6 or pH 9. At pH 7, a gradient endpoint of at least 500 mM NaCl is suitable. At pH 9, a gradient endpoint of less than 250 mM NaCl is sufficient, as is the case at pH 6. The salt concentration also varies with the choice of metal ion bound to the cationic metal affinity matrix. When copper binds to a cationic metal affinity matrix, the salt concentration required to elute the capsid is typically about 3 times lower than for most other metals.

[0024] In another embodiment of the method described in this invention, the pH of the gradient buffer used to separate intact AAV capsids and empty AAV capsids can be within the ranges of pH 6.0 to pH 10, pH 7.0 to pH 9.75, or pH 8.0 to pH 9.5, or pH 8.5 to pH 9.5, or pH 9.0 to pH 9.5, or pH 8.75 to pH 9.25, or pH 8.9 to pH 9.1, or pH 8.95 to pH 9.05, or pH 8.75 to pH 9.25. During pH gradient elution, the elution pH will decrease due to the presence of salt. In principle, the higher the salt concentration, the greater the decrease in pH during capsid elution. Those skilled in the art will understand that exceptions to this rule are possible. Typically, if the salt concentration is 25 mM, the capsid will elute earlier than if the salt concentration is 10 mM, and so on. The salt concentration can remain constant during pH gradient elution, or it can vary independently.

[0025] In another embodiment of the method described in this invention, contaminating DNA present in a buffer mixture or purified intact AAV capsid fraction can be removed by binding DNA to a first and / or second matrix. Variations of the method described in this invention have been developed that utilize a surprising further discovery that, under the same conditions of capsid separation using a cationic ligand-metal solid phase as the first matrix, the AAV capsid cannot bind to the second matrix containing a metal affinity ligand attached to the second matrix, but the DNA binds, wherein the metal affinity ligand comprises two or more negatively charged carboxylic acid residues. For simplicity, the term "second matrix containing a metal affinity ligand attached to the second matrix, wherein the metal affinity ligand comprises two or more negatively charged carboxylic acid residues" is also referred to as "anionic ligand-metal solid phase".

[0026] Combining the steps involving cationic metal affinity ligands with those involving anionic metal affinity ligands can significantly reduce DNA contamination. If applied prior to treatment with the first matrix, the second matrix treatment can increase the capsid binding capacity of the first matrix.

[0027] In the presence of multivalent cations bound to metal-affinity ligands, a buffer mixture or a purified, intact AAV capping fraction is contacted with a second matrix, the second matrix carrying metal-affinity ligands connected to it, the metal-affinity ligands containing two or more negatively charged carboxylic acid residues. The buffer mixture can also be treated simultaneously with a first matrix (cationic ligand-metal solid phase) and a second matrix (anionic ligand-metal solid phase).

[0028] Therefore, in a second aspect, the present invention relates to the buffer mixture wherein the buffer mixture is pretreated by contacting the buffer mixture with a second matrix having an anionic metal affinity ligand loaded with a metal, so that the complete AAV capsid is not bound by the second matrix.

[0029] In some implementations, the conditions for the first matrix equilibration may differ from the conditions for the second matrix equilibration.

[0030] In another embodiment of the method described in this invention, contaminating DNA present in the buffer mixture can be removed by binding it to a second matrix.

[0031] In another embodiment of the method described in this invention, the anionic metal affinity ligand may be selected from the group consisting of aminodicarboxylic acids or aminotricarboxylic acids.

[0032] In another embodiment of the method described in this invention, the aminodicarboxylic acid may be iminodiacetic acid (IDA).

[0033] In another embodiment of the method described in this invention, the aminotricarboxylic acid is nitrotriacetic acid (NTA).

[0034] In another embodiment of the method described in this invention, the second matrix is ​​equilibrated with a buffer solution in the range of pH 6.0 to pH 10.0, or pH 7.0 to pH 9.5, or pH 8.0 to pH 9.25, or pH 8.5 to pH 9.0, or pH 8.75 to pH 9.25, or pH 8.9 to pH 9.1, or pH 8.95 to pH 9.05, particularly in the range of pH 8.75 to pH 9.25.

[0035] In another embodiment of the method described in this invention, the salt concentration of the buffer solution used to prevent the intact AAV capsid from binding to the second matrix is ​​in the range of up to 1M, or 0.1mM to 1.0M, or 1mM to 500mM, or 2mM to 250mM, or 5mM to 250mM, or 3mM to 125mM, or 10mM to 125mM, or 5mM to 60mM, or 20 to 62mM, or 7mM to 30mM.

[0036] In another embodiment of the method described in this invention, the buffer solution for equilibrating the first matrix and / or the buffer solution for separating the empty AAV from the intact AAV contains metal ions having at least two positive charges, preferably selected from the group consisting of calcium, magnesium, copper (divalent copper), iron (trivalent iron), manganese, zinc, barium, nickel, cobalt, and combinations thereof.

[0037] In another embodiment of the method described in this invention, the buffer solution used to balance the first matrix uses a different type of multivalent metal cation than the metal cation used to elute the capsid.

[0038] The overall effectiveness of the cationic metal affinity step and its combination with the anionic metal affinity step can be improved by placing one or two steps within the context of a complete multi-step purification process.

[0039] In a non-limiting example, the method of the present invention can be implemented as follows:

[0040] - The chromatographic solid phase with the cationic metal chelate ligand N,N-bis(2-aminoethyl)-1,2-ethylenediamine on its surface is exposed to a metal cation solution, so that the ligand carries the metal cation. Then, excess metal cations are washed away, leaving the solid phase surface with ligand:metal complex.

[0041] - The cationic metal ligand solid phase is equilibrated to a pH of approximately 9 using a buffer containing an excess of the same metal ion. This is unique in the field of IMAC for two reasons. Known methods specifically remove metal ions from the running buffer because the presence of metal ions is expected to compete with immobilized metal ions for binding sites on the target. This also results in elution products containing an excess of metal ions, which is toxic in many examples of this type of IMAC. The effective pH of 9.0 is also unique. Samples are typically loaded onto the IMAC column at near-neutral pH. This is the only known IMAC method that uses strongly alkaline pH conditions at every stage of processing.

[0042] - Equilibrate samples containing combinations of empty capsids, intact capsids, and contaminated DNA to conditions substantially similar to those of a cation-ligand solid phase.

[0043] - The sample is loaded onto a cationic metal-ligand solid phase. The AAV capping is retained, while unbound contaminants are not retained and are washed away from the cationic metal-ligand solid phase.

[0044] - By increasing the concentration of the non-chelating salt gradient while maintaining the pH at approximately 9, empty and intact AAV capsids are separated from each other. This is another unique feature of the IMAC field, as increasing the concentration of the non-chelating salt at a fixed pH does not elute proteins from the immobilized metal. In specific applications, it appears that the immobilized metal ions mediate a selective, weak attraction to empty capsids, thereby facilitating the separation of empty and intact capsids, while the immobilized metal also provides high-affinity binding sites for DNA.

[0045] This method is unique as the only known IMAC method capable of differentially separating any protein species from any immobilized metal species on any IMAC ligand. To date, the IMAC field has been characterized by proteins either binding or not binding. Bound proteins are eluted through a single step of terminating their interaction with the immobilized metal. No precedent has been found in the literature defining this field for achieving differential elution of closely related protein species.

[0046] In another non-limiting example, the method of the present invention can be implemented as follows:

[0047] - Expose the chromatographic solid phase with the anionic metal chelate ligand iminodiacetic acid on its surface to a metal cation solution, so that the ligand carries the metal cation, and then wash away the excess metal cation, leaving the solid phase surface with the ligand:metal complex.

[0048] - The solid phase of the anionic metal ligand is equilibrated to a pH of approximately 9 using a buffer solution containing an excess of the same metal ion.

[0049] - Equilibrate samples containing combinations of empty capsids, intact capsids, and contaminated DNA to conditions substantially similar to those of anion-ligand solid phases.

[0050] The sample is loaded onto an anionic metal-ligand solid phase. An AAV capsid flows through it. DNA is bound.

[0051] - The column is flushed to fully recover the unbound AAV capsid, which is then fractionated by the above-described cationic metal-ligand solid-phase method.

[0052] Brief description of the attached figures

[0053] Figure 1 Describe the different stationary configurations of TREN on a chromatographic solid phase. (A) shows a ligand bound to the solid via one amino group. Positive charges are shown as plus signs (+), only in (A). The derivatization state of each amino group is indicated by symbols: primary for 1°, secondary for 2°, and tertiary for 3°. (B) shows a ligand bound to the solid via two amino groups. (C) shows a ligand bound to the solid via three amino groups. (D) shows a single-amino-bound version of (A) that coordinates with a metal ion.

[0054] Figure 2 The method for identifying empty and intact AAV caps is described by the ratio of their respective UV absorbance at 260 nm and 280 nm. The symbol 4° refers to the quaternary ammonium (QA) anion exchanger. Gradient experiments were run using sodium chloride.

[0055] Figure 3 The method of detecting DNA using an online fluorescence monitor by detecting embedded DNA during the separation of empty and intact capsids is described. Figure 2 The ultraviolet chromatograms are shown in light gray for reference. It should be noted that most of the DNA is associated with the outer surface of the empty capsid.

[0056] Figure 4 The separation of empty AAV capsids and intact AAV capsids on immobilized TREN-magnesium at pH 7.0 is described.

[0057] Figure 5 The separation of empty and intact capsids on immobilized TREN-magnesium at pH 9.0 is described.

[0058] Figure 6 A comparison of contaminating DNA distribution when empty and intact capsids are separated using the method of this invention is described, and compared with the contaminating DNA distribution when empty and intact capsids are separated by a salt gradient on a strong anion exchanger. The results obtained were... Figure 4 and Figure 2The experimental data. In both cases, the intact capsid fraction is represented by "F". All other fractions represent empty capsids or fragments.

[0059] Figure 7 The analysis of the TREN-Mg fraction from Example 2 by anion exchange chromatography on a strong anion exchanger eluted with sodium chloride gradient is described.

[0060] Figure 8 The comparison results of separating empty and intact capsids using different elution methods and different chromatographic media are described. 4° amine refers to a quaternary amine anion exchanger eluted with salt. 1° amine refers to a primary amine anion exchanger eluted with an increasing pH gradient. TREN-Mg refers to the method of the present invention as described in Example 2.

[0061] Figure 9 A comparison of empty-intact capsid separation using metal-free TREN and magnesium-containing TREN is described.

[0062] Figure 10 The method described involves separating intact capsids from empty capsids and DNA using calcium at pH 7.0 via TREN-metal chromatography. All conditions were identical to those in Example 2, except that magnesium acetate was substituted for calcium acetate.

[0063] Figure 11 The method of separating intact capsids from empty capsids and DNA using TREN-metal chromatography with ferric iron at pH 7.0 is described. All conditions were identical to those in Example 2, except that ferric chloride was replaced with magnesium acetate.

[0064] Figure 12 Fractionation of the capsid purified by cation exchange on TREN-copper at pH 9.0 is described.

[0065] Figure 13 The elution curves of TREN-Cu at pH 9.0 and TREN-Mg at pH 9.0 are compared.

[0066] Figure 14 The application of AAV capping purified by cation exchange at pH 9 to magnesium-loaded anionic metal affinity ligands (IDA) is described.

[0067] Figure 15 The initial capture of the sample prepared on TREN-Mg is described. Complete chromatogram.

[0068] Figure 16 The initial capture of samples prepared on TREN-Mg is described. The areas for capsid elution and cleaning steps are as follows: Figure 15 As shown.

[0069] Figure 17 Described from Figure 16 Anion exchange polishing of the complete AAV capsid portion.

[0070] Figure 18 A comparison of anion exchange polishing curves after capture by two different capture columns is described. In both cases, the samples were prepared as described in Example 6. Left curve: Anion exchange polishing curve after capture by a known cation exchange chromatography method. Right curve: Anion exchange polishing curve after capture by the method described in this invention, from... Figure 17 A magnified image.

[0071] Figure 19 TREN-Mg capture (describes) Figure 15 After ,16), the reduction of empty coatings was compared between primary amine anion exchangers eluted with a pH gradient (left figure) and quaternary amine anion exchangers eluted with a salt gradient (right figure).

[0072] Figure 20 The study describes a comparison of AAV separation using TREN-Mg when extracting host cell DNA with and without immobilized iminodiacetic acid loaded with magnesium at pH 9.

[0073] Figure 21 The AAV capping was described as flowing through a magnesium-loaded monolithic IDA material at pH 9.

[0074] Figure 22 This paper describes the optimized separation of empty and intact AAV capsids using a magnesium-loaded TREN column after a prior DNA extraction step using a magnesium-loaded iminodiacetic acid monolithic material. After loading the capsid onto the TREN column, the empty capsid was replaced with a 10 mM magnesium chloride wash buffer. The intact capsid was then eluted using a sodium chloride gradient.

[0075] Figure 23 This paper describes the separation of empty AAV capsids and intact AAV capsids and the simultaneous extraction of DNA on a monolithic TREN material loaded with ferric iron and eluted with a magnesium gradient.

[0076] Figure 24 This describes the simultaneous extraction of DNA and separation of empty and intact AAV capsids on a monolithic TREN material loaded with ferric iron, followed by equilibration with a metal-free triborate buffer eluted with a magnesium gradient.

[0077] Figure 25 Described Figure 24 The magnified image shows the simultaneous extraction of DNA and separation of empty and intact AAV capsids on a monolithic TREN material loaded with ferric iron, equilibrated with a metal-free triborate buffer, and eluted with a magnesium gradient.

[0078] Figure 26 Described Figure 24 Cesium chloride density gradient fractionation of the contents of the NaCl peak.

[0079] Figure 27 The separation of the empty AAV cap and the intact AAV cap on the quaternary ammonium anion exchange monolithic material is described, equilibrated with a metal-free borate triester buffer and eluted with a magnesium gradient.

[0080] Figure 28 Described from Figure 27 A magnified image of the elution gradient.

[0081] Figure 29 Described Figure 24 The TREN-Fe3 monolithic material shown and Figure 27 The image shown is a superimposed image of the elution gradient of the quaternary ammonium feedstock. Invention Details

[0083] The cationic metal affinity ligand used to implement the basic form of the invention is from a family of amino derivatives, carrying at least one positive charge independent of any metal cation it may bind to or will bind to. One or more positively charged amino residues may consist of one or more primary, secondary, tertiary, quaternary, imide, or imine residues. The cationic amino residues may include any combination or conformation of the aforementioned residues. At least one cationic amino residue must maintain a positive charge at pH 9.0. The ligand may also include uncharged amide nitrogen residues, provided at least one positively charged amino derivative is present. Examples of cationic amino ligands that meet these requirements include diethyltriamine, triethyltetraamine, tetraethylpentamine, polytriethylamine, polyamide amine, deferoxamine (also known as deferox or desferal), and N,N-bis(2-aminoethyl)-1,2-ethylenediamine (also known as 2,2',2”-nitrotriethylamine, or 2,2',2”-triaminotriethylamine, or tri(2-aminoethyl)amine, or TAEA, or TREN.

[0084] Any of the aforementioned cationic metal affinity ligands can be directly covalently bonded to a solid-phase chromatography surface via one or more of their amino groups. They can also be indirectly attached to the solid-phase surface via linear or branched polymers that facilitate ligand access or generate multivalent linear tentacle-like, branched, or dendritic configurations. The term dendritic polymer refers to a tree-like branching pattern in which branches fork into a deep branching network, with each branch terminal bearing a ligand. Known examples include triethylamine-terminated dendritic polymers (polytriethylamine, pTEA) and polyamide amine (PAMAM) dendritic polymers, which are available from suppliers worldwide. The same approach can be applied to anionic metal affinity ligands.

[0085] In one embodiment, the cationic metal affinity ligand is TREN. The unfixed ligand comprises three primary amine residues branched from a central tertiary amine via ethyl groups. Immobilizing the ligand on a chromatographic solid phase can produce any one or any combination of the following configurations: a primary amine residue linked to the solid phase, converting the residue to a secondary amine, linked to an ethyl group of the tertiary amine residue, and linked to two other ethyl groups resulting in a terminal primary amine residue; two primary amine residues linked to the solid phase, converting these residues to secondary amines, each secondary amine linked to an ethyl group to produce a tertiary amine residue, and linked to another ethyl group to produce a primary amine; and three primary amine residues linked to the solid phase, each linked to a central tertiary amine residue via an ethyl group. Figure 1 ).

[0086] Those skilled in the art will recognize that primary, secondary, and tertiary amines are, to varying degrees, weak anion exchange groups based on their respective pKa values. In commercial literature, TRENs exhibit a positive charge up to approximately pH 9. The lack of practical methods for separating empty capsids from intact capsids is generally acknowledged in this field, highlighting the unexpected utility of TRENs, as the weak anion exchanger DEAE also maintains a positive charge up to approximately pH 9. It also emphasizes that the behavior of TRENs generally differs from that of commonly known weak anion exchangers. This suggests that, in configurations that bind polyamine ligands to weak anion exchange groups, the combination contributes to the distinctive characteristics of the methods of this invention in achieving their unique results.

[0087] Chromatographic solid phases containing TREN are available from commercial suppliers, namely...<https: / / www.bio-works.com / product / iex-resin / workbeads-tren> Many other global suppliers of chromatographic media manufacture and sell anion exchangers, demonstrating that the techniques for synthesizing chromatographic products suitable for implementing the methods of this invention are widely available.

[0088] The anionic metal affinity ligand in the compound form of the present invention is a ligand from a family of aminodicarboxylic acids lacking a positive charge, such as iminodiacetic acid (IDA) or an aminotricarboxylic acid lacking a positive charge, such as nitrilotriacetic acid (NTA).

[0089] Chromatographic solid phases with anionic metal affinity ligands, such as IDA and NTA, are widely available from all major commercial chromatography suppliers worldwide.

[0090] The methods described in this invention can be implemented with any one or any combination of a variety of polyvalent metal ions, including barium, calcium, iron, magnesium, manganese, copper, and / or zinc. This list favors non-toxic metals. Toxic heavy metals may provide positive results, but their use is discouraged because additional work is required to demonstrate the removal of toxic heavy metals from the final product. Generally, metals with known human nutritional, therapeutic, or general health value should be preferred, with the most preferred types being those that provide optimal empty and intact capsid separation and / or the greatest DNA reduction in any AAV serotype.

[0091] In some embodiments, the buffer used for column equilibration and the buffer used for elution may also contain metal ions. In some such embodiments, the type of metal ion will be the same as the type of charge carried by the solid-phase ligand. In some embodiments, such buffers may be used to metal-charge the solid-phase ligand instead of performing previous steps to metal-charge the ligand. In some embodiments, the types of metal ions in the equilibration buffer and the gradient buffer may be different from each other.

[0092] In some embodiments, the cationic metal affinity ligand is charged with multivalent metal cations before the sample is contacted with the cationic metal affinity matrix. In other embodiments, the cationic metal affinity ligand is charged with multivalent metal cations by an excess of multivalent metal cations in the sample. In other embodiments, the cationic metal affinity ligand is charged with multivalent metal cations after the sample is contacted with the cationic metal affinity matrix. In one such embodiment, the cationic metal affinity ligand is charged with multivalent metal cations using a wash buffer before elution begins. In a closely related embodiment, the cationic metal affinity matrix is ​​charged by including multivalent metal cations in a gradient initiation buffer. In another such embodiment, the cationic metal affinity matrix is ​​charged by including multivalent metal cations in a gradient endpoint buffer.

[0093] In various implementations, intact capsids can be eluted by increasing the pH gradient, or by increasing the salt gradient, or by increasing the gradient of polyvalent metal cations.

[0094] In some embodiments, the gradient initiation buffer may contain polyvalent metal cations at concentrations ranging from 0.0 mM to 10.0 mM, or 2.5 mM to 7.5 mM, or 4.0 mM to 6.0 mM, or lower, higher, or in between. Initial screening can be performed at approximately 5 mM. Based on the results at this concentration, the metal content can be adjusted upwards or downwards to obtain optimal results. The metal ion concentration in the gradient initiation buffer may be the same as the metal ion concentration in the gradient termination buffer.

[0095] In some embodiments, metal salts can be used as eluents and may be preferred because they alter the surface charge properties of the AAV capsid and improve the separation of empty capsids from intact capsids. Each event in which a multivalent metal binds to a metal binding site on the capsid surface neutralizes an equal number of negative charges at the dicarboxyl or tricarboxyl metal binding sites on the capsid. Since the negative charge facilitates interaction with the cationic metal-affinity ligands, this metal binding reduces the capsid's attractiveness to the ligands. This mechanism is fundamentally different from elution using nonmetallic salts, which work solely by increasing the overall conductivity of the mobile phase. The nonmetallic salt ions create a damping field that suppresses all types of ion interactions, regardless of where the charges are located, how close they are to each other, or whether they are positive or negative. Elution with nonmetallic salts is a non-specific field phenomenon. This is fundamentally different from the local and specific capabilities of metal salts, which neutralize the local charge of aminocarboxyl residues associated with metal binding sites on the capsid surface. This explains why metal salts can elute capsids at lower concentrations than nonmetallic salts.

[0096] In some embodiments where elution is performed by increasing the concentration of metal ions, the gradient can cover a range from 0 mM to 200 mM, or 0 mM to 100 mM, or 0 mM to 50 mM, or 0 mM to 25 mM, or 1 mM to 25 mM, or 2 mM to 20 mM, or 5 mM to 15 mM, or lower, higher, or intermediate ranges. An initial screening can be performed using a gradient range of 0 mM to 25 mM. Based on the results at this concentration, the metal content of the gradient initiation buffer and gradient termination buffer can be adjusted upwards or downwards to achieve optimal results.

[0097] In some embodiments, the metal affinity ligand can be passively charged with metal ions, meaning that the solid phase can be loaded with metal during buffer equilibration by including metal ions in the equilibration buffer, rather than having a discrete metal charging step, prior to sample introduction. Those skilled in the art will recognize that the term "equilibration buffer" refers to a buffer solution used in a chromatographic apparatus, the purpose of which is to create a specific chemical environment for the intended use of the chromatographic method.

[0098] In some embodiments, the metal ions contained in the sample can convert uncharged cationic metal affinity ligands into their metal-complexed forms for carrying out the methods of the present invention. This may occur intentionally by adding metal ions to the sample, or unintentionally due to metals remaining unintentionally in cell culture harvests, lysates, and partially purified samples.

[0099] It will be apparent to those skilled in developing purification methods that the choice of metal species may depend in part on where the method of the invention is desired to be placed within the overall multi-step purification process. Experimental data indicate that ferric iron and manganese have higher affinity for DNA and other phosphorylated contaminants, including endotoxins. This suggests that these metals may be advantageous when the method is used to polish already highly purified samples. When using iron or manganese to capture AAV from a roughage stream, increased affinity for DNA is unlikely to be advantageous, as high DNA content may consume excessive total column binding capacity. In such cases, metals with lower affinity for DNA, such as, but not limited to, calcium or magnesium, may prove more advantageous.

[0100] In some implementations, the efficiency of DNA removal can be improved by combining sample treatment with a cationic metal affinity matrix with treatment with an anionic metal affinity matrix.

[0101] In some implementations, it may be advantageous to use strong DNA-binding metals, such as iron and manganese, on anionic metal affinity ligands and AAV-stabilizing metals, such as calcium and / or magnesium, on cationic metal affinity ligands.

[0102] The methods of the present invention involving cationic metal affinity ligands can be carried out in a pH range of 6.0 to 10.0, or pH 7.0 to pH 9.75, or pH 8.0 to pH 9.5, or pH 8.5 to pH 9.5, or pH 9.0 to pH 9.5, or pH 8.75 to pH 9.25, or pH 8.9 to pH 9.1, or pH 8.95 to pH 9.05, or higher, lower, or intermediate ranges, preferably in the pH range of 8.75 to pH 9.25. For initial evaluation purposes, screening at a pH of approximately 9.0 is recommended, and further evaluation at lower and higher values ​​is recommended to optimize the results.

[0103] The methods of the present invention involving anionic metal affinity ligands can be implemented in pH 6.0 to 10.0, or pH 7.0 to pH 9.5, or pH 8.0 to pH 9.25, or pH 8.5 to pH 9.5, or pH 8.75 to pH 9.25, or a narrower, wider, or intermediate range.

[0104] In some implementations, the entire method is operated at a substantially constant pH. In one such implementation, the column and sample are equilibrated to approximately pH 9. The column is then loaded, washed, and eluted with a gradually increasing salt gradient, all at pH 9.0.

[0105] In some implementations, the column and sample used for implementing methods applicable to cationic metal affinity ligands can initially be equilibrated to a lower pH, such as in the pH range of 6.0 to 7.0, the sample is loaded onto the column and the column is washed to remove unbound contaminants, and then the loaded column is reequilibrated to a higher pH to prepare for the elution step. This process produces the unexpected benefit of increased capacity, as AAV binds more strongly to cationic ligand-metal complexes at pH 6.0 than at pH 7.0, or at pH 8.0, or at pH 9.0. The capacity increases as the pH decreases, presumably due to increased hydrogen bonding, reflecting an underlying increase in protonation. Once the sample is bound, the column can be reequilibrated to any desired conditions to maximize the separation of empty and intact capsids. Low-pH sample binding has additional utility for DNA removal, as it has been surprisingly found that solid-phase TREN-metal surfaces bind more strongly to DNA at acidic pH than at neutral or alkaline pH.

[0106] In some implementations, the cationic metal affinity column can be equilibrated at an alkaline pH (e.g., pH 9), but the sample to be loaded may be at a lower pH (e.g., pH 7). The lower pH of the sample causes a brief reequilibration of the column to a lower pH during sample loading. Experimental data suggest that AAV binds more tightly to TREN at lower pH values ​​(e.g., pH 7), which is generally considered to increase binding capacity. At the end of the sample loading phase, restoring the equilibration buffer flow rate will overcome the transient effect of sample loading and reequilibrate the column to pH 9. Those skilled in the art will recognize that this results in a more streamlined workflow, as it eliminates the need for buffer equilibration at pH 7 prior to sample loading.

[0107] In one embodiment, a cationic metal affinity solid phase loaded with polyvalent metal cations is equilibrated to pH 9.2 with a Tris-borate buffer, wherein the specific purpose of using the Tris-borate buffer is to provide strong buffering capacity at low ionic strengths. In one such embodiment, the buffer itself does not contain polyvalent metal cations. In one embodiment, a sample containing empty and intact AAV capsids is applied under substantially physiological conditions, wherein the term substantially physiological is understood to refer to a pH value in the range of about pH 6.5 to pH 7.5 and a conductivity in the range of about 50 mS / cm to about 200 mS / cm. The sample may contain, but does not need to contain, additional buffers. Applying the sample to the column will deequilibrate the pH and conductivity conditions of the column, after which restoring the equilibration buffer flow will restore the column to its original equilibration conditions, preparing it for initiating an elution gradient. In one such embodiment, the column elutes to polyvalent metal cations with a linear gradient. In one such embodiment, the metal cation is magnesium. In one such embodiment, a gradient or step with a non-metallic salt (such as sodium chloride) may be performed after the metal cation gradient. In another embodiment, a cleaning agent (such as NaOH) step may be performed after the metal cation gradient. In some such embodiments, the NaOH may be accompanied by other reagents, such as sodium chloride, designed to enhance its cleaning power. In some such embodiments, the cleaning buffer may contain 1M NaOH and 1-3M NaCl.

[0108] Those skilled in the art will recognize that the stronger binding at lower pH values ​​is the exact opposite of what is expected based on the known behavior of proteins on anion exchangers. Protein retention on anion exchangers is thought to weaken as pH decreases. The stronger binding at lower pH values ​​also differs from the behavior of known metal affinity chromatographic media, since they are typically eluted at low pH values. These findings underscore that, although chromatographic solid phases with cation ligand-metal complexes exhibit effects traceable to both IMAC and anion exchange, cation ligand-metal complexes produce a net selectivity distinct from both.

[0109] In some embodiments, AAV is captured at pH values ​​below 9, and the cationic metal affinity column can be eluted with an increasing pH gradient instead of a salt gradient. In one such example, the metal-loaded column is equilibrated to a pH of approximately 6 in a buffer containing approximately 15 mM sodium chloride. Sample conditions should be substantially the same. After loading the column with an AAV cap and washing to remove unbound sample components, the column is eluted with a gradually increasing pH gradient to a buffer containing approximately 100 mM sodium chloride at a pH of approximately 9; in other words, elution is performed with a gradually increasing pH gradient while maintaining a constant salt concentration. Co-pH elution also occurs when the salt concentration increases in parallel with the pH. It should be appreciated that many variations of this method are possible, including, but not limited to, increasing the salt concentration while increasing the pH, decreasing the salt concentration while increasing the pH, or eluting in a series of steps that independently change the pH and salt concentration. It will also be appreciated that this feature further distinguishes the method of the present invention from anion exchange chromatography on strong anion exchangers, where an increase in pH results in stronger AAV binding.

[0110] Within each of the above-mentioned ranges, buffering compounds and / or combinations of buffering compounds that provide good buffering capacity are known in the art. Compounds known to provide buffering capacity in the pH 9 region include glycine (pKa 9.6), arginine (pKa 9.1), tripropane (pKa 9.0), and boric acid (pKa 9.15–9.25). Such buffers are typically used in the art at concentrations of 20 mM to 50 mM, although concentrations can be reduced or increased to meet specific requirements. Generally, cation buffers are preferably used in conjunction with chromatography using a cationic surface, such as the chromatographic solid phase specified in the method described in this invention. This is advantageous for using the first three types listed above. Boric acid is an anion; negatively charged. One possible limitation of using anionic buffers with cationic chromatographic surfaces is that the interaction between the buffer and the surface can extract most of the buffer from the mobile phase, resulting in insufficient buffering capacity of the mobile phase. However, using a buffer with an opposite charge to the chromatographic surface can sometimes indeed support unique selectivity, thus benefiting a given separation. Combinations of anionic and cationic buffers are also possible, as are the use of zwitterionic buffers and combinations of zwitterionic buffers with cationic and / or anionic buffers. Zwitterionic buffers with a pKa close to 9 include 3-([1,1-dimethyl-2-hydroxyethyl]amino)-2-hydroxypropanesulfonic acid (AMPSO, pKa 9.36) and N-cyclohexyl-3-aminopropanesulfonic acid (CAPSO, pKa 9.43), etc.

[0111] In some implementations, the buffer solution can be selected specifically based on its ability to provide good buffering capacity at low conductivity. In one such implementation, boric acid is used as the buffer solution at a pH of approximately 9.0 to 9.2. In another such implementation, Tris is used as the counterion to raise the pH of the buffer solution to a selected target pH because Tris also has low molar conductivity and its buffering capacity is zero in the pH range of 9.0 to 9.2.

[0112] In some embodiments employing borate buffers, those skilled in the art of borate chemistry will recognize that borate anions are sometimes referred to as synthetic lectins because they are capable of forming covalent bonds with cis-diols, as is found on some carbohydrates. This bond formation gives the complex a negative borate charge where there were originally two uncharged hydroxyl groups. Because this bond is covalent, it survives exposure to high concentrations of salt. It should be understood that the electrostatic charge of the AAV capsid containing cis-diol carbohydrates will become more electronegative, with its net electronegativity increasing according to the number of cis-diol borate binding events on any given capsid surface. It should also be understood that, to some extent, imparting empty and intact capsids to cis-diol sugars to form covalent bonds with borate anions will alter their relative electronegativity, which may change their relative retention on cation chromatography supports. Regarding the relative enrichment of empty capsids in cis-diol-rich environments, increased retention is expected.

[0113] In other embodiments using borate buffers, those skilled in the art of borate chemistry will recognize that borate anions are known to alter the metal-binding selectivity in certain situations through chelating agents and compounds with chelating sites. By extension, this modification should be understood to potentially include sites on protein surfaces capable of chelating metals, including AAV capsid proteins. Further modification of the metal-binding selectivity of borates at their respective metal-binding sites, given the degree of difference between the metal-chelating sites of empty and intact capsids, could affect the chromatographic separation of empty and intact capsids.

[0114] In one implementation, the separation of empty and intact capsids is performed on a cationic metal affinity solid phase by eluting the chromatographic solid phase with an increasing salt gradient, wherein the salt is sodium chloride, or potassium chloride, or sodium acetate, or potassium acetate. Generally, sodium chloride is a suitable starting material, and evaluation of other materials is unnecessary. However, doing so may yield results of interest. It is known that AAV capsids from different serotypes have different affinities for anion exchangers, so it should be understood that a broad gradient (e.g., ending with 1M sodium chloride) as a starting point is preferable. The endpoint can then be adjusted to maximize the separation of empty and intact capsids. In one example of AAV serotype 2 / 8, an initial 50 column volume (CV) linear gradient was run to an endpoint of 1M NaCl. In subsequent experiments, the gradient length remained constant, but the endpoint was reduced to 500 mM NaCl, and in another subsequent experiment, it was reduced to 250 mM. Experiments demonstrating this characteristic are a routine part of normal process development. Generally, a gradient endpoint of the lowest salt concentration is preferred to separate the empty capsid from the intact capsid and achieve good recovery of the intact capsid, because such conditions should also allow the maximum proportion of DNA to remain bound to the solid phase.

[0115] In some implementations, arginine, histidine, or lysine may be used as eluents in place of inorganic salts.

[0116] Salts with strong metal cation binding capabilities are not recommended because they tend to remove metals bound to solid-phase ligands. They also tend to elute DNA prematurely and may prevent the method from reaching the level of DNA it should be able to remove. Salts of particular concern include citrate, phosphate, pyrophosphate, ethylenediaminetetraacetic acid (EDTA), ethylene glycol-bis(β-aminoethyl ether)-N,N,N',N'-tetraacetic acid (also known as ethyl acetate (EGTA)), aspartic acid, glutamic acid, glutamine, and other known chelating salts.

[0117] The buffer solution used to perform the methods of the present invention may contain compounds capable of stabilizing AAV capsids or inhibiting their nonspecific interactions with chromatographic surfaces. Such stabilizing compounds may include nonionic or zwitterionic surfactants, such as octagoglucoside, poloxamer 188, pranic acid F68, CHAPS, or CHAPSO. Such stabilizing compounds may alternatively or additionally include sugars, such as sucrose, sorbitol, xylose, mannitol, or trehalose. These stabilizing compounds may alternatively or additionally include amino acids, such as betaine, taurine-betaine, arginine, histidine, or lysine. They may also include glycine and alanine in a pH range of 4.0 to 7.5. All of these reagents are known in the biopharmaceutical field because they tend to improve solubility and / or stabilize product recovery. In some cases, they also improve the fractionation of desired products from unwanted substances. They are typically used at low concentrations unlikely to interfere with the implementation of the methods of the present invention.

[0118] In one embodiment, the AAV capsid may be eluted from a cationic ligand-metal solid phase by increasing a gradient of nonionic or zwitterionic strong hydrogen donor-acceptor (such as sugar or urea). This embodiment first requires loading the AAV capsid onto the column, then increasing the pH and salt concentration to a level at which the intact capsid will not be eluted. From this point, the intact capsid can be eluted by a one-step or gradient application of urea or sugar or other nonionic hydrogen donor-acceptor. In one such embodiment, the intact capsid is eluted with sorbitol at a gradient up to 300 mM at the same pH and salt concentration. In another such embodiment, the intact capsid is eluted with urea at a gradient up to 10 M at the same pH and salt concentration.

[0119] In some embodiments, the method of the present invention can simultaneously utilize the effects of two different metal species. In such embodiments, the cationic metal affinity matrix is ​​charged with ferric charge, while the elution of the AAV capsid is carried out by a gradient generated by increasing the magnesium concentration. In closely related embodiments, the metal used to charge the cationic metal affinity matrix can be manganese. In other closely related embodiments, the metal used to perform AAV capsid elution can be calcium, barium, or copper.

[0120] The validity of the invention can be demonstrated using known techniques. A common method is to measure the UV absorbance ratio of the chromatographic fraction at 260 nm and 280 nm. The AAV capsid is partially transparent to ultraviolet light. This allows for standard UV monitoring of the chromatogram to detect DNA within the capsid and DNA associated with the outer protein capsid wall. Empty capsids typically exhibit a 260 / 280 ratio below 1, usually in the range of 0.6 to 0.8. Highly purified intact capsids typically exhibit a 260 / 280 ratio greater than 1.3. Figure 2 The higher the ratio, the higher the proportion of intact capsids in the corresponding chromatographic fraction. For example, a fraction with a ratio of 1.35 is understood to contain a larger proportion of intact capsids than a fraction with a ratio of 1.33, which in turn is richer in intact capsids than a fraction with a ratio of 1.31, and so on. Secondary confirmation of intact and empty capsids may include analytical ultracentrifugation (AUC) and cryotransmission electron microscopy (cryoTEM). Wavelength ratios can also be used to indicate the elution position of contaminating DNA in the chromatogram, as DNA outside the AAV capsid also has a unique 260 / 280 ratio, ranging from approximately 1.5 to 2.2, with the proportion of pure DNA slightly above 2.0.

[0121] The amount of DNA associated with the capsid can be estimated by pre-incubating the sample with a dye called Picogreen before chromatography. Picogreen is a type of dye that can intercalate between DNA base pairs. This intercalation process is called intercalation. Picogreen itself has no important optical properties, but the Picogreen between DNA base pairs produces green fluorescence, which can be used to amplify the detection of DNA. Figure 3 Experiments have shown that Picogreen can only detect DNA outside the capsid. Therefore, estimating DNA content using both UV and Picogreen methods can provide more in-depth information than using either method alone.

[0122] The ratio of Picogreen fluorescence to UV absorbance can also be plotted to indicate the relative amount of DNA outside the capsid to the capsid itself, for example, a series of fractions obtained by practicing the method of the present invention.

[0123] In some embodiments, the method of the present invention can be implemented after the sample has been pretreated to reduce DNA contamination. In one such embodiment, the sample is pretreated with DNase enzyme. In another such embodiment, the sample is pretreated with positively charged solid particles to bind DNA, and then the DNA is removed by removing the particles bound to the DNA. In a related embodiment, the sample is treated with a positively charged polymer to form a large, insoluble complex with the DNA, which, after precipitation, can be removed by centrifugation, filtration, or a combination of both.

[0124] In one embodiment of the compound, a sample containing an AAV capsid and DNA is first treated with a positively charged solid phase for tangential flow filtration (TFF), then the sample is concentrated and percolated through the TFF into a buffer suitable for nuclease treatment. The sample is then treated with nucleases to break down the DNA. The sample is then further concentrated and percolated, while nucleotides and histones are removed through a membrane. In the final embodiment, the filter medium is preferably selected to have maximum porosity for retaining AAV, so as to eliminate maximum diversity of contaminants.

[0125] In another series of embodiments, following the cationic metal affinity step, the content of empty capsids and DNA can be reduced by additional methods. In one such embodiment, the additional method can be a chromatographic method, such as anion exchange chromatography on a strong anion exchanger. In another such embodiment, the additional method can be anion exchange chromatography on a primary amine exchanger, wherein empty and intact capsids are separated in an increasing pH gradient. In yet another such embodiment, the additional method can be density gradient centrifugation.

[0126] In another series of embodiments, the sample is first treated to reduce the amount of DNA before performing the method of the present invention, and then further methods are employed to reduce the empty capsid and DNA content.

[0127] In another series of embodiments, the method described in this invention is used as a polishing method to reduce the empty capsid and DNA content of samples that have already been treated to remove most contaminants. In one such embodiment, a first purification step is performed using a known cation exchange chromatography method. After recovering partially purified AAV from the elution fraction of the cation exchange step, these AAV fractions are treated with the method described in this invention. In another such embodiment, a first purification step is performed using a known bioaffinity chromatography method. After recovering partially purified AAV from the elution fraction of the bioaffinity step, these AAV fractions are treated with the method of this invention. In any of these embodiments, the sample may be pretreated as described above to reduce the DNA content prior to the first chromatography step.

[0128] In one embodiment, the method of the present invention can be performed in a configuration in which a TREN-metal solid phase and a quaternary ammonium anion exchange solid phase are connected in series. In one such embodiment, the quaternary ammonium solid phase may be placed first in sequence. Both solid phases are co-equilibrated with the same equilibration buffer, for example, at pH 9. The sample is loaded, the column is washed, and then eluted with a salt gradient. Because the salt concentration from which AAV is eluted from the strong anion exchanger is lower than that from the cation ligand-metal solid phase, the AAV capsid will be eluted from the strong anion exchanger and rebind to the cation ligand-metal solid phase. However, a large portion of the DNA will remain bound to the strong anion exchanger, including throughout the elution of the capsid from the cation ligand-metal solid phase. By preventing this DNA subset from residing on the cation ligand-metal solid phase, the full capacity of the cation ligand-metal solid phase can be used for capsid separation. Both columns may be washed in series with high salt and NaOH to release strongly bound contaminants. In another embodiment, the order of the solid phases is reversed.

[0129] In one relevant embodiment, the method of the present invention is combined with a step using a primary amine anion exchanger. In one such embodiment, the primary amine solid phase may be run first in this sequence. The primary amine anion exchanger is loaded at a pH range of pH 7 to pH 8, followed by elution with a pH gradient to a pH of approximately 10. This provides partial separation of empty and intact capsids and partial reduction of contaminating DNA. The eluted intact capsids can then be titrated to a pH of 9 or lower and treated with the method of the present invention to achieve further removal of empty capsids and contaminating DNA. In some embodiments, the process sequence may be reversed.

[0130] In another closely related embodiment, the method of the present invention is implemented in a configuration in which a cationic metal-ligand solid phase is connected in series with an anionic solid phase, wherein the anionic metal-ligand solid phase is the first in the sequence. In one such embodiment, the anionic metal-ligand solid phase is iminodiacetic acid (IDA). In another such embodiment, the anionic metal-ligand solid phase is hypozinotriacetic acid. In one such embodiment, the anionic solid phase carries the same metal species as the cationic metal-ligand solid phase. In one such embodiment, the anionic solid phase carries a different metal species than the cationic metal-ligand solid phase. In one such embodiment, the anionic metal-ligand solid phase carries an iron or manganese charge, while the cationic metal-ligand solid phase carries a calcium, magnesium, or another non-iron, non-manganese charge. In all these embodiments, the anionic metal-ligand solid phase is used to remove phosphorylated contaminants from the sample before the sample comes into contact with the cationic metal-ligand solid phase. In all such embodiments, such highly phosphorylated contaminants particularly include DNA, RNA, endotoxins, and cell membrane debris, including membranes from organelles and vesicles. In all these embodiments, pre-removal of these contaminants preserves the capacity of the cationic metal-ligand solid phase of the AAV capsid and allows for a higher purity capsid than when those contaminants are bound to the column along with the desired contaminant. In one embodiment, the tandem method is used to produce a purified, intact AAV capsid in a single step.

[0131] In another such embodiment, the anionic metal-ligand solid phase is not sequentially linked to the cationic metal-ligand solid phase. In one such embodiment, the sample is treated with the anionic metal-ligand solid phase before performing the cationic metal-ligand solid phase method. In one such method, the two steps are performed sequentially. In one such embodiment, cell culture harvests or lysates containing empty AAV capsids, intact AAV capsids, and DNA are treated with the anionic metal-ligand solid phase before performing the cationic metal-ligand solid phase step. In a closely related embodiment, one or more additional steps are performed after the anionic metal-ligand step and before the cationic metal-ligand step. In one such embodiment, cell culture harvests or lysates containing intact capsids, empty capsids, and DNA are treated with the anionic metal-ligand solid phase, and the DNA-deficient solution is separated by affinity chromatography before continuing to separate the empty and intact capsids by performing the cationic metal-ligand solid phase method. In another such implementation, the cell culture harvest or lysate treated with anionic metal ligands is separated by cation exchange chromatography before proceeding with the separation of empty and intact capsids by performing a cationic metal ligand solid-phase method.

[0132] In one embodiment, the DNA reduction step using a metal-loaded anionic metal affinity matrix is ​​combined with simultaneous DNA reduction via a DNase enzyme. In one such embodiment, the DNase enzyme is a salt-tolerant DNase, and the metal-loaded anionic metal affinity matrix exists in the form of a soluble polymer backbone with multiple metal-loaded anionic metal affinity ligands. In another such embodiment, the metal-loaded anionic metal affinity matrix is ​​in the form of solid, insoluble particles with multiple metal-loaded anionic metal affinity ligands. In both cases, a matrix with multiple metal-loaded anionic metal affinity ligands is present during the enzymatic cleavage of DNA. In some cases, the metal species loaded by the anionic metal affinity ligands is the same as the metal species cofactor required by the enzyme. In one case, the metal ion is magnesium; in other cases, the metal species loaded by the anionic metal affinity ligands is different from the metal species cofactor required by the enzyme. In one such case, the soluble enzyme cofactor is magnesium, but the anionic metal affinity ligands are preloaded with iron. In another case, the soluble enzyme cofactor is magnesium, but the anionic metal affinity ligand is preloaded with manganese.

[0133] In one embodiment, the solid phase with a cationic metal affinity ligand may be one or more porous particles, one or more porous membranes, one or more nanofibers, a monolithic material, a hydrogel, a depth filter, or another form of solid phase.

[0134] In a closely related embodiment, the solid phase with anionic metal affinity ligands can be one or more porous particles, or one or more porous membranes, or one or more nanofibers, or a monolithic material, or a hydrogel, a depth filter, or another form of solid phase.

[0135] In one embodiment, the solid phase used to carry out any aspect of the method described in this invention can be arranged as a flow-through device to facilitate the implementation of the chromatographic method. The flow-through device used to perform the chromatographic method is generally referred to as a chromatographic column, regardless of its form or the materials it contains.

[0136] In one embodiment, the materials and methods of the present invention, any prior processing steps, and any subsequent processing steps can be connected to and interfaced with a microprocessor to enable end-to-end automation of the multi-step process, which can be performed continuously if desired.

[0137] Those skilled in the art will recognize that while AAV capsids of different serotypes share many fundamental physical and chemical similarities, they also exhibit significant variability in surface chemistry and purification properties. Affinity chromatographic media tend to recognize one or more serotypes. Some broad-spectrum affinity ligands can recognize multiple AAV serotypes, but this does not mean they bind with the same affinity, provide the same capacity, or elute under the same conditions. Diversity is also observed in ion exchange chromatographic behavior. Different AAV serotypes bind to cation exchangers with different affinities and elute under different conditions. Different AAV serotypes also bind to anion exchangers with different affinities and elute under different conditions, with varying degrees of separation between empty and intact capsids. The optimal range for all these methods is known, thus excessive experimentation is unnecessary to optimize any given method for any given serotype. It is also expected that differences in binding, capacity, elution, and empty-intact capsid separation characteristics between serotypes can be observed using the methods described in this invention. Given prior knowledge of appropriate starting conditions, identification of key process variables, the range to which they may be evaluated, and the types of effects they produce, as provided in this specification, excessive experimentation is not required to optimize the method to best suit any given AAV serotype.

[0138] In one embodiment, the AAV serotype treated by the method of the present invention can be AAV1, or AAV2, or AAV3, or AAV4, or AAV5, or AAV6, or AAV7, or AAV8, or AAV9, or AAV10, or AAV11, or another serotype. In another embodiment, the AAV serotype treated by the method of the present invention can be a recombinant hybrid serotype, such as AAV2 / 8 or AAV2 / 9, or another hybrid serotype. In yet another embodiment, the AAV serotype treated by the method of the present invention can be a synthetic recombinant serotype.

[0139] Those skilled in the art will recognize that each serotype exhibits different retention characteristics on any type of adsorption chromatographic medium. This diversity across all chromatographic methods results in the need to optimize conditions to achieve the best overall purification for any particular serotype.

[0140] In one embodiment, the sample containing empty capsids, intact capsids, and DNA is a cell culture harvest. In a closely related embodiment, the sample containing empty capsids, intact capsids, and DNA is a cell lysate. In another closely related embodiment, the sample containing empty capsids, intact capsids, and DNA is a partially processed or partially purified formulation. In one such embodiment, the sample containing empty capsids, intact capsids, and DNA is a cell culture harvest or lysate that has been treated with enzymes to lyse the DNA. In another such embodiment, the sample containing empty capsids, intact capsids, and DNA is a cell culture harvest that has been treated with a cationic polymer or cationic solid phase to extract DNA. In another embodiment, the sample containing empty capsids, intact capsids, and DNA is a cell culture harvest or lysate that has been treated with a cationic solid phase, then concentrated and percolated by tangential flow filtration, and then treated with enzymes to lyse the DNA. In an extension of this embodiment, the sample is then concentrated and filtered by tangential flow filtration in preparation for chromatographic purification. In another embodiment, the cell harvest or lysate is treated to precipitate contaminants, leaving AAV in the supernatant. In another embodiment, the cell harvest or lysate is treated to precipitate AAV, leaving contaminants in the supernatant. The resuspended precipitate is separated from the supernatant and resuspended, this time without the initially unprecipitated contaminants. In another embodiment, a portion of the cell harvest or lysate is purified by chromatography. In one such embodiment, a portion of the harvest or lysate is purified by affinity chromatography. In another such embodiment, a portion of the harvest or lysate is purified by hydrophobic interaction chromatography. In another such embodiment, a portion of the harvest or lysate is purified by size exclusion chromatography. In another such embodiment, a portion of the harvest or lysate is partially purified by cation exchange chromatography. In another such embodiment, the harvest or lysate is purified by anion exchange chromatography.

[0141] The invention is further explained by the following non-limiting examples. Example

[0142] Example 1. Preparation of the reference baseline. A mixed sample of empty and intact AAV8 capsids, partially purified by cation exchange chromatography, was injected onto a strong (quaternary (4°)amine) monolithic column equilibrated with 50 mM Tris at pH 9. The anion exchanger was eluted with a sodium chloride gradient to 200 mM, followed by a one-step washing of the column with 500 mM sodium chloride. The elution profile is shown below. Figure 2As shown, it illustrates the partial separation of empty and intact capsids, with individual empty capsid peaks eluting before individual intact capsid peaks, but overlapping with them. The intact capsid peaks show a 260 / 280 wavelength ratio of 1.307. This can be interpreted as meaning that a certain proportion of empty capsids also elute in the intact capsid region, reducing the wavelength ratio. Note the relatively large size of the empty capsid peaks.

[0143] Example 2. Preliminary evaluation of the separation of empty and intact capsids by TREN-Mg at pH 7.0 and pH 9.0. The monolithic material with a TREN surface was loaded with divalent magnesium ions by exposing it to 100 mM magnesium acetate in water. Excess magnesium was washed away while the column was equilibrated to pH 7.0 with 50 mM HEPES. The column was then eluted with a linear gradient to 1 M sodium chloride. A one-step wash with 2 M sodium chloride, 50 mM HEPES, and then further wash with 1 M sodium hydroxide was performed. The intact capsid peak showed a 260 / 280 wavelength ratio of 1.240 (…). Figure 4 In a separate experiment, divalent magnesium ions were loaded onto the monolithic material with the TREN surface by exposing it to a 100 mM aqueous magnesium acetate solution. Excess magnesium was washed away with 50 mM tripropane and 2 mM magnesium chloride while the column equilibrated to pH 9.0. The same samples, both empty and intact AAV8 capsids purified by cation exchange chromatography as used in Example 1, were injected onto a TREN-Mg column. The column was eluted with a linear gradient of 500 mM sodium chloride and 2 mM magnesium chloride. It was then progressively cleaned with 2 M sodium chloride, 50 mM tripropane, and 2 mM magnesium chloride, followed by further washing with 1 M sodium hydroxide. The elution profiles are shown below. Figure 4 As shown, it illustrates the partial separation of empty and intact capsids, with two empty capsid peaks, one eluting before and one eluting after the intact capsid peak. The intact capsid peak shows a 260 / 280 wavelength ratio of 1.377. Figure 5 This can also be understood as the proportion eluted from the intact capsid region being less than the proportion when the sample was fractionated with the strong anion exchanger shown in Example 1. Overall results indicate that even at pH 7.0, some separation was achieved between empty and intact capsids, but at pH 9.0, the separation was substantially improved. Subsequent experiments at pH 9.25 showed essentially the same performance, but slightly worse results at pH 9.5.

[0144] Figure 6 This shows the difference in DNA distribution outside the capsid between the two methods. From Figure 2 and Figure 4In the chromatograms shown, the height of the bars is estimated by comparing the peak area of ​​Picogreen fluorescence at 260 nm UV absorbance. Fractions marked with F represent fractions with intact capsids. The two methods show a clear difference in DNA distribution, adding evidence that their selectivity differs from each other. This also highlights their complementarity and the potential value of the present invention's method using anion exchange chromatography, and vice versa.

[0145] Figure 7 The analytical anion exchange curves (quaternary ammonium) of the original cation exchange purified sample were compared with those of the original cation exchange purified sample. Figure 6 Fractions containing intact capsids, and Figure 6 The fraction contained a 2M sodium chloride wash. The intrinsic tryptophan fluorescence of the elution curve was monitored for two reasons: firstly, tryptophan fluorescence increases the sensitivity of protein detection by 15-20 times; secondly, tryptophan fluorescence is produced only by proteins. It does not detect DNA. This avoids confounding effects caused by the uncertain individual contributions of proteins and DNA to the 260nm and 280nm UV absorbance. As shown in the figure, the original sample was predominantly empty capsids. The intact capsid fraction was clearly dominated by fully intact capsids. The sodium chloride wash fraction consisted entirely of empty capsids. These findings indicate that the selectivity of the method described in this invention differs from any known method, especially from anion exchange chromatography using strong anion exchangers.

[0146] Example 3. Comparison of empty and intact capsid separation on a primary amine anion exchanger with an increasing pH gradient. The primary amine anion exchanger was equilibrated with 10 mM Tris, 10 mM ditripropane, and 2 mM magnesium chloride (pH 7.0). It was loaded with the same samples used to prepare Examples 1 and 2. Figure 8 This indicates the elution region containing the full peak in the rightmost center panel, labeled 1° amine. Compare this with... Figure 2 A comparison of the corresponding regions shows the separation of empty and intact capsids on a strong anion exchanger (left image). It is also compared with... Figure 5 The corresponding regions were compared, showing the separation of empty and intact capsids using the method described in this invention. As shown in the figure, the wavelength of the intact capsid fraction from the primary amine anion exchanger is 1.30, for the quaternary exchanger it is 1.307, and for the method described in this invention it is 1.377. Overall, Figure 8 The present invention emphasizes that the selectivity of the method differs from that of the two types of anion exchangers and points out the influence of immobilized metal ions.

[0147] Example 4. Comparison of capsid separation with and without metal added to chromatographic buffer. Figure 9Chromatograms from a control run in the absence of metal ions were compared (top plot). The TREN column was without multivalent metal cations, and the gradient buffer was also free of multivalent metal cations. The bottom plot illustrates the results when 2 mM magnesium was used in the gradient buffer. The results show a clear shift of the empty capsid to the later elution fraction, with a significant increase in the proportion of intact capsids in the main elution peak. This is evidenced by the higher 260 / 280 ratio. In another experiment, the TREN column was equilibrated with magnesium before sample loading. The results (not shown) were identical to those obtained when magnesium was provided only in the gradient buffer. These results, in addition to confirming the improved performance gained from the presence of magnesium, demonstrate that cationic metal affinity ligands do not need to be loaded with metal before sample application, representing a classic approach in the IMAC field. Metals can be loaded onto ligands at any time before capsid elution.

[0148] Example 5. Comparison of Chromatography with Different Metal Ions. The procedure in Example 2 was repeated, except that a different metal was used to charge the TREN solid phase in each case. Comparisons were made with calcium, iron (ferric), manganese, copper (divalent), zinc, and barium. The curves were consistent in the regions eluted with empty and intact capsids, but differed in the regions of the washing step. Iron and manganese, in particular, showed significantly larger peaks in the in-situ cleaning peak of sodium hydroxide. Figure 10 A chromatogram of calcium is shown. Figure 11 The chromatogram of iron is shown. These differences are attributed to the unusually strong binding of DNA to iron. Figure 12 The chromatogram of copper (divalent copper) at pH 9.0 is shown. The clean peak for copper is larger than that for calcium or magnesium, but smaller than that for iron. The use of manganese (not shown) produced a clean peak intermediate between that of iron and copper. Figure 13 The elution profiles of magnesium and copper in the capsid removal region were compared. Note the shift between the elution profiles; the capsid in copper eluted earlier. In the comparison, copper also produced the most favorable wavelength ratio, indicating that it produced the intact capsid fraction with the lowest proportion of empty capsids. The intact capsid peak containing copper was also the narrowest, and the variant subgroups in the empty fraction were resolved more clearly.

[0149] Example 6. Magnesium-loaded anionic metal affinity ligands for the assessment of pre-reduction of DNA. A monolithic material with iminodiacetic acid (IDA) chelating residues was loaded with magnesium and equilibrated to pH 9.0. A cation-exchange purified capsid samples were equilibrated to the same conditions and loaded onto a column. The AAV capsids passed through the column without binding and did not fractionate, highlighting the importance of cation-metal affinity columns for capturing and separating empty and intact capsids. However, DNA and contaminants in the sample remained bound together (…). Figure 14 ).

[0150] Example 7. Integrating the method of the present invention into a multi-step purification method. Filtered cell culture lysate containing DNA, empty and intact AAV capsids was treated with particles coated with ethylenediamine, giving them a positive charge. These were mixed with particles carrying negatively charged SO3 groups. The mixed particles were added to the sample at a volume ratio of 5% and incubated for 60 minutes. These particles bound a large amount of soluble host cell DNA. The particle-DNA complex was precipitated, and the supernatant was filtered through a membrane filter with a pore size cutoff of 0.45 μm. This treatment improved filterability, allowing the sample to be concentrated by tangential flow filtration. The sample was concentrated 10-fold using a membrane with a pore size cutoff of 300 kDa, which eliminated many small molecule contaminants, including proteins. The sample was also percolated by tangential flow filtration into a buffer suitable for DNase lysis with a salt-tolerant nuclease: 20 mM Tris, 500 mM NaCl, pH 8.0. Salt-tolerant nuclease was added together with 5 mM magnesium chloride, and the mixture was incubated at room temperature for 16 hours. Tangential flow filtration was repeated to remove histones released by digestion of host cell DNA, to eliminate DNA fragments and nucleotides, and the sample was equilibrated for capture on a column with the TREN-magnesium complex. The sample was loaded onto the column and processed as described in Example 2. A fraction of the TREN fraction containing the intact capsid was further fractionated on a primary amine anion exchanger eluted with a pH gradient. Another fraction of the TREN fraction was further fractionated on a strong anion exchanger eluted with a salt gradient. The results were compared with another experiment performed under the same conditions but with the initial chromatographic steps performed on a cation exchanger. Figure 15 The complete chromatogram of the initial capture step of TREN-Mg is shown. Figure 16 Emphasized Figure 15 The elution gradient and cleaning steps in the process. Figure 17 The anion-exchange polishing of the complete capsid portion captured by TREN-Mg is highlighted. Figure 18 The anion exchange polishing results of AAV purified using the method of this invention (right panel) were compared with those of AAV initially purified using cation exchange chromatography (left panel). Based on the relative peak height, the method of this invention eliminated 83% more empty capsids than cation exchange chromatography. This reduces the overlap between the two empty peaks and the full peak, resulting in a more effective reduction of empty capsids by polishing using anion exchange chromatography. Figure 19 A comparison was made between primary amine anion exchangers eluted with a pH gradient (left panel) and quaternary amine anion exchangers eluted with a salt gradient (right panel), showing a reduction in empty capsids after capturing TREN-Mg. Figure 15 Both results highlight the complementarity of the method of the present invention with anion exchange chromatography, and underscore the potential of using this combination to achieve an overall reduction in empty coating content more effectively than using either method alone.

[0151] Example 8. Integrating the method of the present invention into different multi-step purification processes. The method of Example 6 is repeated, except that a cation exchange chromatography step is used instead of the original capture step. The sample is then prepared for processing by the method of the present invention, and so on. In one variant of the method, the TREN column is loaded with copper. In another variant of the method, the TREN column is loaded with magnesium.

[0152] Example 9. DNA removal and separation of empty and intact capsids are enhanced by combining the method of the present invention with another method capable of separating them. After the sample is treated with the TREN step of Example 6, the intact capsid fraction is treated with a strong anion exchanger with a salt gradient using a known method for separating empty and intact capsids, as shown in Example 1.

[0153] Example 10. The separation of empty and intact capsids is enhanced by combining the method of the present invention with another method capable of separating them. The present invention employs strong anion exchange chromatography with salt gradient elution or primary amine anion exchange chromatography with pH gradient elution to separate the empty and intact capsids, followed by a further polishing step.

[0154] Example 11. The separation of empty and intact capsids is enhanced by combining the method of the present invention with another method capable of achieving their separation. A sample containing DNA, empty capsids, and intact capsids is first captured by affinity chromatography. The AA fraction eluted from the affinity column is treated by the method of the present invention to remove excess DNA and separate the empty capsid from the intact capsid. In one variant, the sample is treated prior to affinity chromatography to reduce the concentration of DNA, for example by any of the methods described in Example 7.

[0155] Example 12. Samples were treated with anionic metal affinity solid phases (AAVs) to enhance DNA reduction and separation of intact and empty capsids before separation of intact and empty capsids using cationic metal affinity solid phases. Samples containing DNA, intact capsids, and empty capsids were equilibrated to approximately pH 9 and then loaded onto an iminodiacetic acid (IDA) column packed with ferric iron. DNA was bound, but the AAV capsids were not. After treatment, the sample was loaded onto a magnesium-loaded TREN column without further preparation. Empty capsids were separated from intact capsids by increasing the sodium chloride gradient. In one variant of this method, the IDA solid phase was replaced by an NTA solid phase. In another variant, the anionic metal affinity solid phase was loaded with manganese. In yet another variant, the TREN solid phase was loaded with calcium. In yet another variation, both solid phases were loaded with the same metal. In one version of this method, the anionic metal affinity solid phase was present in the form of a flow chromatography device. In one variant, the anionic metal affinity chromatography device was connected in series with a TREN column. In another variant, an anionic metal affinity solid phase is added as particles to a sample containing intact capsids, empty capsids, and DNA, causing the DNA to bind, and then the particles carrying the DNA are removed. In yet another variant, the treated sample is subsequently treated with a TREN solid phase. In yet another variant, the treated sample is treated with a cation exchanger before being treated with a TREN solid phase. In yet another variant, the treated sample is treated by affinity chromatography before being treated with a TREN solid phase.

[0156] Example 13. Comparison of AAV separation at pH 9 using TREN-Mg with and without pre-extraction of host cell DNA via immobilized iminodiacetic acid loaded with magnesium. The monolithic material with the anionic metal affinity ligand iminodiacetic acid was loaded with magnesium ions. The monolithic material was then equilibrated with 50 mM bis-tri-propane at pH 9.0. A cation-exchange purified AAV8 sample was titrated to pH 9.0. The sample passed through the monolithic material. AAV did not bind and flowed through. Numerous DNA subsets bound and were removed from the sample. The sample was applied to magnesium-loaded cationic metal affinity (TREN) monolithic material and eluted with an increasing sodium chloride gradient. Another sample was prepared from cation-exchange purified AAV by titration to pH 9, but without treatment with the magnesium-IDA monolithic material. Figure 20 The TREN chromatograms were compared, and it can be seen that the pretreatment of the IDA column altered the elution behavior on the TREN column.

[0157] Example 14. Pre-extraction of host cell DNA. A monolithic material containing the anionic metal-affinity ligand iminodiacetic acid was loaded with magnesium ions. The monolithic material was then equilibrated with 50 mM bis-tri-propane at pH 9.0. A cation-exchange purified AAV8 sample was titrated to pH 9.0. The sample was passed through the monolithic material. AAV did not bind and flowed through. DNA bound and was thus removed from the AAV-containing sample. The DNA bound to the IDA column was subsequently removed using a cleaning step with 1 M NaOH. Figure 21 ).

[0158] Example 15. Optimization of empty and intact capsid separation from a TREN-magnesium column after pre-extraction of host cell DNA via an IDA-magnesium column. DNA was extracted from an AAV sample purified by cation exchange as described in Example 14. The sample was applied to a TREN-magnesium column at pH 9. Empty capsids were washed from the column with 10 mM magnesium chloride. Intact capsids were eluted using an increased salt gradient. Figure 22 ).

[0159] Example 16. Tandem extraction of host cell DNA and separation of empty and whole capsids, sequentially treated with a ferric iminodiacetic acid solid phase followed by gradient separation with a magnesium-loaded TREN solid phase. The IDA column was loaded with ferric iron and equilibrated to pH 9. The TREN column was loaded with magnesium and equilibrated to pH 9. The two columns were first tandemly connected to the IDA column and then washed with a pH 9 buffer. The sample, containing empty and whole capsids and host cell DNA, was equilibrated to pH 9. The sample was passed through both columns. DNA bound to the ferric IDA column. The AAV capsid passed through the IDA column and was captured by the magnesium-TREN column. The columns remained connected and were washed with equilibration buffer, followed by elution of the TREN column with an increasing salt gradient to bind the DNA to the IDA column. After collecting the desired whole AAV capsid fraction, the column was washed with 1M NaOH. In a variant of this method, the IDA column was disconnected after the washing step so that the TREN column could elute independently.

[0160] Example 17. Purification of AAV by anionic and cationic metal affinity and tangential flow filtration as an intermediate step. Cell lysate containing AAV8 and host cell DNA was equilibrated to pH 7.0 and 100 mM NaCl. Insoluble particles loaded with ferric iron and iminodiacetic acid were mixed with the cell lysate at a ratio of 5% particles. The sample was incubated for 60 minutes to allow the particles to settle to the bottom of the container. The supernatant was filtered through a membrane filter with an average porosity of approximately 0.45 micrometers. The sample was then percolated using a membrane with a molecular weight cutoff (MWCO) of 300 kDa via tangential flow filtration. The percolation buffer contained 2 mM magnesium chloride, 25 mM ditripropane, and pH 9.0. Six dialysis volumes were subjected to tangential flow filtration, during which proteins and lower molecular weight contaminants were removed through the membrane. The sample was then loaded onto a column containing a magnesium-loaded TREN and fractionated under the conditions described in Example 15.

[0161] Example 18. Simultaneous DNA extraction and empty / full capsid separation were performed using a cationic metal affinity matrix loaded with ferric iron and eluted with a linear gradient of magnesium chloride. TREN monolithic material was loaded with ferric iron and equilibrated to 25 mM tripropane, 1% sucrose, 0.1% poloxamer 188, pH 9.0. A cation exchange-purified AAV8 capsid, still containing host DNA, was equilibrated to the same conditions and loaded onto the TREN-Fe monolithic material. The column was washed with equilibration buffer to remove unbound material. The column was then eluted with a 50 CV linear gradient to 25 mM tripropane, 1% sucrose, 0.1% poloxamer 188, pH 9.0 to 25 mM magnesium chloride, followed by a 10 CV linear gradient to 50 mM magnesium chloride. A cleaning step was performed using 25 mM tripropane, 1% sucrose, 0.1% poloxamer 188, 2.0 M NaCl, and pH 9.0, followed by a stronger cleaning step using 1 M NaOH, 2 M NaCl, and pH 13. Results are shown below. Figure 23 As shown, a small subset of empty capsids eluted before the main intact capsid population. Another subset of empty capsids eluted long after the intact capsids. The initial cleaning step removed a small subset of capsid fragments, while the NaOH step removed a significant amount of host DNA. Notably, these results indicate that iron still binds to TREN in the presence of magnesium, which in turn suggests that TREN may have a stronger affinity for iron than for magnesium.

[0162] Example 19. IDA-Fe and NTA-Fe were used as precursors for comparison of the separation of empty and intact capsids by TREN-Fe elution with a magnesium gradient. IDA particles were loaded with ferric iron and equilibrated to 50 mM Hepes, pH 7. NTA particles were prepared using the same method. IDA-Fe particles were added to a sample of AAV8 particles purified by cation exchange chromatography to a final proportion of 5%. IDA-Fe particles were added to another sample of AAV8 particles purified by cation exchange chromatography to a final proportion of 5%. The samples were incubated for 120 min and then the particles were removed by centrifugation. The IDA-Fe-treated samples were applied to a TREN-Fe column at pH 7. The column was then reequilibrated to pH 9.0 (25 mM tripropane, 1% sucrose, 0.1% poloxamer 188) and eluted with a linear gradient to 50 mM magnesium chloride (in the same basal buffer). The column was then washed with 2 M NaCl and then with 1 M NaOH. The NTA-Fe-treated samples were treated in the same manner. After elution of TREN-Fe with magnesium, the 260 / 280 ratio of the sample treated with NTA-Fe was 3.02. After elution of TREN-Fe with magnesium, the 260 / 280 ratio of the sample treated with IDA-Fe was 3.05.

[0163] Example 20. Optimization of empty-intact capsid separation using borate buffer. A 1 mL monolithic TREN material containing ferric iron was equilibrated with 100 mM boric acid, 50 mM Tris, 1% sucrose, 0.01% poloxamer 188, pH 9.0, and a conductivity of 0.25 mS / cm. One part of cation-exchange purified AAV8 was diluted in 99 parts of 400 mM boric acid, 50 mM Tris, 1% sucrose, 0.01% poloxamer 188, pH 7.0, and a conductivity of 0.76 mS / cm, and loaded onto a column. Flow rate was restored with equilibration buffer until the buffer pH and conductivity stabilized, and then the column was eluted with a linear gradient to 100 mM boric acid, 50 mM Tris, 50 mM magnesium chloride, 1% sucrose, 0.01% poloxamer 188, pH 9.0, and a conductivity of 9.6 mS / cm. The empty capsid was gradually eluted with 2M NaCl, followed by column cleaning with 1M NaOH and 2M NaCl. The complete chromatogram is shown below. Figure 24 As shown. Figure 25 The area corresponding to the intact capsid after elution was magnified. The experiment was repeated, except that the sample dilution buffer was replaced with 25 mM Hepes, 1% sucrose, 0.01% poloxamer 188, pH 7.0, and conductivity 0.85 mS / cm. The results were essentially the same (not shown). Figure 24 and 25 The results shown present a significant and favorable contrast with all other materials and conditions. For example... Figure 2 and3 As shown, the curves produced by eluting the same sample with a salt gradient on a strong anion exchanger are characterized by a large empty capsid peak, followed by a large intact capsid peak. Figure 5 , 12 As shown in Figures 13 and 22, on a cationic metal affinity column containing multivalent metal cations, the separation of empty and intact capsids typically produces a curve indicating the onset of empty capsid elution, which partially overlaps with the tail boundary of the intact capsid peak. Figure 24 and 25 In the gradient, although intact capsids are eluted in the middle, only a small empty capsid peak precedes the intact capsid peak, and no obvious empty capsid peak is observed at the tail of the intact capsid peak. Conversely, except for the first small peak in the gradient, all empty capsids are eluted in the 2N salt step after the magnesium gradient. Figure 26 This indicates that their negative charge is enhanced. This suggests a hypothesis that borates enhance separation by modifying the negative charge on the empty capsid.

[0164] Example 21. Fractionation of AAV coating from quaternary ammonium (QA) anion exchange monolithic material. AAV8 coating was prepared as in Example 20. The quaternary ammonium anion exchanger was equilibrated and eluted as shown in Example 20. Elution curves are shown below. Figure 27 As shown. Under these conditions, the amount of DNA in the NaOH peak is higher than that in other peaks. Figure 27 The amount of chromatin is low. This document shows poor chromatin removal. Furthermore, the UV absorbance at 260 nm and 280 nm indicates that the QA column fails under these conditions and cannot effectively separate empty capsids from intact capsids. Figure 28 Results from TREN-Fe3 and QA are in Figure 29 A comparison was made. Besides the significant differences in relative chromatin extraction and separation of empty and intact capsids, this comparison highlights the substantial differences between the method of this invention and conventional anion exchange chromatography in terms of separation chemistry.

[0165] References

[0166] All references cited in this article are included by way of citation, provided that their inclusion does not contradict the explicit teachings herein.

[0167] [1] R Cheung, J Wong, T Ng, A review of the application of immobilized metal ion affinity chromatography, Applied Microbial Biotechnology, 96(2012)1411-1420.

[0168] [2] H Block, B Maertens, A Spriesterbach, N Brinker, J Kubicek, R Fabis, J Labahn, F Schafer, Immobilized metal affinity chromatography (IMAC): Review, Enzymatic Methods, 463(2009)439–473.

[0169] [3] J Koerber, JH Jang, J Yu, R Kane, D Schaffer, One-step purification of engineered adeno-associated virus by immobilized metal affinity chromatography, Human Gene Therapy, 18(2007)367–378.

[0170] https: / / www.liebertpub.com / doi / pdfplus / 10.1089 / hum.2006.139

[0171] [4] N Meyer, O Davulcu, Q Xie, M Silviera, G Zane, E Large, Expression and purification of adeno-associated particles in baculovirus systems and construction of the extracellular domain of Escherichia coli AAVR, *Biological Experimental Methods*, 2020, DOI: 10.21769 / BioProtoc.3513

[0172] https: / / bio-protocol.org / e3513

[0173] [5] L Tan, DS Kim, IK Yoo, WS Choe, Purification of plasmid DNA using metal ion affinity, Chemical Engineering Science, 62(2007)5809–5820.

[0174] [6] L Tan, WB Lai, CT Lee, DS Kim, WS Choe, Differential interactions of plasmid DNA, RNA and endotoxin with immobilized and free metal ions, Chromatography Journal, A, 1141 (2007) 226–234.

[0175] [7] M Lock, Luc Vandenberghe, J Wilson, Scalable Production Method for AAV, US Patent US9198984B2, effective August 31, 2028.

[0176] [8] M Lock, Luc Vandenberghe, J Wilson, Scalable Production Method for AAV, US Patent US20160040137A1, expected to expire on April 4, 2027.

[0177] [9] M Lock, M Alvira, Scalable purification method for AAV9, World Patent Application WO2017160360A9, priority date December 11, 2015.

[0178]

[10] B Pappin, M Kiefel, T Houston, Chapter 3. Boron-carbohydrate interactions, Integrative Research in Carbohydrate-Glycobiology and Glycotechnology, CF Chang (ed.), New Technology Open Ltd., London, 2012.

[0179]

[11] A Bousher, Review: Unidentified complexes involving borates, Journal of Coordination Chemistry, 34(1995)1-11.

Claims

1. A method for isolating intact adeno-associated virus (AAV) from empty AAV capsids in a mixture comprising a buffer, intact AAV capsids, and empty AAV capsids, comprising the steps of: - The mixture is contacted with a first matrix having metal affinity ligands attached to it, the metal affinity ligands having the ability to complex metal ions through three or more nitrogen atoms. - In the presence of multivalent cations bound to metal-affinity ligands, empty AAV capsids are separated from intact AAV capsids by elution using pH gradients, salt gradients, metal ion gradients, or combinations thereof, to obtain purified intact AAV capsid fractions. The metal affinity ligands are selected from the group consisting of: diethyltriamine, triethyltetraamine, tetraethylpentamine, polyamide amine, polytriethylamine, deferriphosphine, N,N-bis(2-aminoethyl)-1,2-ethylenediamine and tri(2-aminoethyl)amine (TREN); The polyvalent cations are selected from the group consisting of iron (III), manganese (II), copper (II), zinc (II), cobalt (II), magnesium (II), calcium (II), barium (II), nickel (II), and combinations thereof.

2. The method of claim 1, wherein the first matrix is ​​loaded with polyvalent cations before the mixture is contacted with the first matrix, during the contact of the mixture with the first matrix, and / or during elution.

3. The method of claim 1, wherein the polyvalent metal cation present during elution is selected from the group consisting of magnesium (II), calcium (II), barium (II), copper (II), and mixtures thereof.

4. The method of claim 1, wherein the concentration range of the polyvalent cation at the gradient endpoint is 0.1 mM to 200 mM, or 1 mM to 100 mM, or 2 mM to 50 mM, or 5 mM to 25 mM.

5. The method according to claim 1, wherein elution is performed at pH values ​​in the range of pH 6.0 to pH 10, pH 7.0 to pH 9.75, pH 8.0 to pH 9.5, pH 8.5 to pH 9.5, pH 9.0 to pH 9.5, pH 8.75 to pH 9.25, pH 8.9 to pH 9.1, or pH 8.95 to pH 9.

05.

6. The method of claim 5, wherein the pH value range is from pH 8.75 to pH 9.

25.

7. The method of claim 1, wherein elution is performed by increasing the salt concentration in the range of up to 1 M, 1 mM to 500 mM, 2 mM to 250 mM, 3 mM to 125 mM, 5 mM to 60 mM, or 7 mM to 30 mM.

8. The method of claim 1, wherein elution is performed by increasing the concentration of the metal salt, wherein the metal ion is a polyvalent metal cation.

9. The method of claim 8, wherein the multivalent metal cation is selected from the group consisting of copper(II), magnesium(II), calcium(II), or combinations thereof.

10. The method of claim 1, wherein a buffer containing borate is used for contacting and / or elution.

11. The method of claim 1, wherein the metal affinity ligand carries a multivalent cation prior to elution, and the AAV capsid is eluted by increasing the concentration of a second multivalent metal cation.

12. The method of claim 1, wherein the contact of the mixture or purified intact AAV capping fraction with a second matrix having a metal affinity ligand attached to the second matrix is ​​carried out in the presence of a multivalent cation bound to the metal affinity ligand, the metal affinity ligand comprising two or more negatively charged carboxylic acid residues.

13. The method of claim 12, wherein the mixture is treated simultaneously with a first matrix and a second matrix.

14. The method of claim 12, wherein the metal affinity ligand attached to the second matrix is ​​selected from the group consisting of aminodicarboxylic acids (such as iminodiacetic acid) and aminotricarboxylic acids (such as hypotriacetic acid).

15. The method of claim 12, wherein the anionic metal affinity ligand of the second matrix has a multivalent metal cation selected from the group consisting of iron (III), manganese (II), copper (II), zinc (II), cobalt (II), magnesium (II), calcium (II), barium (II), nickel (II), and combinations thereof.

16. The method of claim 15, wherein the multivalent metal cation is iron(III).

17. The method of claim 12, wherein the pH range in which the contact occurs between the mixture and the second matrix having a metal affinity ligand connected to the second matrix is ​​pH 6.0 to pH 10, pH 7.0 to pH 9.75, pH 8.0 to pH 9.5, pH 8.5 to pH 9.5, pH 9.0 to pH 9.5, pH 8.75 to pH 9.25, pH 8.9 to pH 9.1, or pH 8.95 to pH 9.

05.

18. The method of claim 17, wherein the pH range is from pH 8.75 to pH 9.

25.

19. The method of claim 12, wherein the salt concentration range of the contact between the mixture and the second matrix is ​​at most 1 M, or 0.1 to 1 mM to 500 mM, or 2 mM to 250 mM, or 5 mM to 250 mM, or 3 mM to 125 mM, or 10 mM to 125 mM, or 5 mM to 60 mM, or 20 to 62 mM, or 7 mM to 30 mM.

20. The method of any one of claims 1-19, wherein contaminating DNA present in the mixture or purified intact AAV capsid fraction is removed by binding DNA to the first and / or second matrix.

Images