EA043811B1 - COMPOSITIONS CONTAINING CYCLODEXTRIN AND ANTIBODY-DRUG CONJUGATE - Google Patents

Description

This application claims priority to U.S. Provisional Patent Application No. 61/780,185, filed March 13, 2013, and U.S. Provisional Patent Application No. 61/782,231, filed March 14, 2013, the contents of which are incorporated herein by reference in their entirety. in its entirety for all purposes.

State of the art

Antibody-drug conjugates (ADCs) can be an effective means of delivering a drug to a target site in a tissue or body. Recognition of a target, such as a tumor, by an antibody minimizes exposure of non-target tissues to toxic chemotherapeutic agents and limits undesirable effects associated with free drug toxicity (ie, not associated with a carrier such as the antibody). ADCs can be obtained in several ways. Before administration to a human or other individual, the conjugate is purified to remove free drugs and other impurities.

Due to the very high activity of benzodiazepine-containing drugs, removing drug impurities from a mixture containing a benzodiazepine ADC and benzodiazepine drug impurities should be very effective. The present invention satisfies this need and other needs.

Brief description of drawings

In fig. 1 is a graph showing the percentage of high molecular weight (HMW) compounds present in various p2H12-1 preparations stored at 25°C. %HMW is determined at time points 0, 7 and 14 days.

In fig. 2 is a graph showing the percentage of high molecular weight (HMW) compounds present in various h1F6-1 preparations stored at 25°C. %HMW is determined at time points 0, 7 and 14 days.

In fig. 3 is a graph showing the percentage of high molecular weight (HMW) compounds present in various h2H12-3 preparations stored at 40°C. %HMW is determined at time points 0, 3 and 7 days.

In fig. 4 is a graph showing the percentage of high molecular weight (HMW) compounds present in various h1F6-3 preparations stored at 40°C. %HMW is determined at time points 0, 3 and 7 days.

In fig. Figure 5 is a graph showing the percentage of high molecular weight (HMW) compounds present in various h2H12-2 preparations stored at 40°C. %HMW is determined at time points 0, 3 and 7 days.

In fig. 6 is a graph showing the percentage of high molecular weight (HMW) compounds present in various h1F6-2 preparations stored at 40°C. %HMW is determined at time points 0, 3 and 7 days.

In fig. 7 is a graph showing the percentage of acidic compounds present in various h2H12-1 preparations stored at 25°C. The percentage of acidic compounds is determined at time points 0, 7 and 14 days.

In fig. 8 is a graph showing the percentage of acidic compounds present in various h21F6-1 preparations stored at 25°C. The percentage of acidic compounds is determined at time points 0, 7 and 14 days.

In fig. 9 is a graph showing the percentage of acidic compounds present in various h2H12-3 preparations stored at 40°C. The percentage of acidic compounds is determined at time points 0, 3 and 7 days.

In fig. 10 is a graph showing the percentage of acidic compounds present in various h1F6-3 preparations stored at 40°C. The percentage of acidic compounds is determined at time points 0, 3 and 7 days.

In fig. 11 is a graph showing the percentage of acidic compounds present in various h2H12-2 preparations stored at 40°C. The percentage of acidic compounds is determined at time points 0, 3 and 7 days.

In fig. 12 is a graph showing the percentage of acidic compounds present in various h21F6-2 preparations stored at 40°C. The percentage of acidic compounds is determined at time points 0, 3 and 7 days.

In fig. 13 is a graph showing the percentage of high molecular weight (HMW) compounds present in various h1F6-1 preparations stored at 25°C. % HMW compounds are determined at time points 0 and 7 days.

In fig. 14 is a graph showing the concentration of hydroxypropyl-betacyclodextrin during the diafiltration process. The diafiltration buffer contained 3% w/v. cyclodextrin. The data show that the membrane is permeable to cyclodextrin.

In fig. 15 is a graph showing the removal of quenched drug-linker molecules present in the reaction mixture used for conjugation after quenching while maintaining a cyclodextrin concentration of 10% w/v. (diamonds) or 3% w/v. (squares).

- 1 043811

In fig. 16 is a graph showing the removal of quenched drug-linker molecules present in the reaction mixture used for conjugation after quenching while maintaining a cyclodextrin concentration of 3% w/v.

In fig. 17 is a graph showing the removal of quenched drug-linker molecules present in the reaction mixture used for conjugation after quenching while maintaining a cyclodextrin concentration of 3% w/v.

In fig. 18 is a graph showing the removal of quenched drug-linker molecules present in the reaction mixture used for conjugation after quenching while maintaining a cyclodextrin concentration of 3% w/v.

General information

The present invention is based, in part, on the discovery that the removal of benzodiazepine drug impurities from a mixture containing a benzodiazepine ADC and benzodiazepine drug impurities (also referred to herein as an ADC mixture) is ineffective due to the nature of the benzodiazepines. agents, and the discovery that the addition of cyclodextrin to the mixture allows for effective clearance of impurities related to the benzodiazepine drug. The present inventors have found, in particular, that adding cyclodextrin to a mixture containing a benzodiazepine ADC and benzodiazepine drug impurities before performing tangential flow filtration allows for effective clearance of benzodiazepine drug impurities.

The present invention is also based, in part, on the discovery that cyclodextrin-containing benzodiazepine ADC compositions exhibit exceptional stability compared to formulations that do not contain cyclodextrin. Increased stability can be demonstrated, for example, by one or more of the following: (i) reduction in the rate and extent of aggregation, (ii) reduction in the growth of acidic species, and (iii) reduction in chemical degradation of the drug.

The essence of the invention

Provided herein are methods for removing benzodiazepine drug related impurities from mixtures containing a benzodiazepine ADC and benzodiazepine drug related impurities by tangential flow filtration using cyclodextrin. The methods involve exposing a mixture containing a benzodiazepine ADC and benzodiazepine drug-related impurities to a mixture using tangential flow filtration. The use of cyclodextrin during filtration facilitates the separation process.

Thus, the disclosure provides methods for removing impurities due to a benzodiazepine drug from a mixture containing a benzodiazepine ADC and benzodiazepine drug-related impurities, the methods comprising treating a mixture containing a benzodiazepine ADC and benzodiazepine drug-related impurities with using tangential flow filtration, where cyclodextrin is used to facilitate the purification process. In preferred aspects, the cyclodextrin is added in an amount sufficient to maintain significant solubility of the components in the ADC mixture and prevent aggregation. Cyclodextrin may, for example, be present in the mixture at the beginning of the tangential flow filtration process or, alternatively, may be added to the mixture only after the tangential flow filtration has begun (preferably before removing a significant portion of the impurities). In some aspects, the methods include subjecting a mixture containing a benzodiazepine ADC and benzodiazepine drug related impurities to tangential flow filtration while maintaining a concentration of at least about 1% w/v. cyclodextrin, at least about 2% w/v. cyclodextrin or at least about 3% w/v. cyclodextrin in the mixture. The present invention also provides methods including preparing a mixture containing a benzodiazepine ADC, benzodiazepine drug related impurities, and cyclodextrin, wherein the cyclodextrin is present in a concentration of at least about 1% w/v, and tangentially filtering the mixture at maintaining a concentration of at least about 1% w/v. cyclodextrin in the mixture; the methods include preparing a mixture containing a benzodiazepine ADC, benzodiazepine drug-related impurities, and cyclodextrin, wherein the cyclodextrin is present in a concentration of at least about 2% w/v, and tangentially filtering the mixture while maintaining the concentration at least about 2% w/v cyclodextrin in the mixture; and the methods include providing a mixture containing a benzodiazepine ADC, benzodiazepine drug-related impurities, and cyclodextrin, wherein the cyclodextrin is present in a concentration of at least about 3% w/v, and tangentially filtering the mixture while maintaining the concentration of at least approximately 3% w/v. cyclodextrin in the mixture.

- 2 043811

Tangential flow filtration can be, for example, constant volume diafiltration or batch diafiltration. The tangential flow filtration apparatus may include, for example, a pump, a filtration vessel having an inlet, a filtrate outlet, a retentate outlet, an ultrafiltration membrane having a pore size of approximately 50 kDa or less that divides the filtration vessel into an inlet compartment and an outlet compartment so that all filtrate must enter the inlet and pass through the ultrafiltration membrane before flowing out of the filtration vessel through the filtrate outlet, the sample vessel for storing the reaction mixture used for conjugation, and the buffer vessel , hydraulically connected to the sample container. In preferred aspects, the buffer container contains at least about 1% w/v. cyclodextrin, at least about 2% w/v. cyclodextrin or at least about 3% w/v. cyclodextrin. The ultrafiltration membrane may have a range of pore sizes including, for example, a pore size of approximately 30 kDa. The ultrafiltration membrane can be made from a variety of materials, including regenerated cellulose.

The mixture that is purified by tangential flow filtration may be a reaction mixture used for conjugation, including any of the reaction mixtures used for conjugation described herein. An impurity related to a benzodiazepine drug may be any of the impurities described herein. The benzodiazepine drug impurity may be extinguished or unextinguished. For example, the methods may include the steps of contacting an antibody or linker antibody with a linker-benzodiazepine drug compound under conditions sufficient to form a reaction mixture used for conjugation containing a benzodiazepine ADC. In some cases, the reaction mixture used for conjugation may be contacted with a quenching additive to produce a quenched reaction mixture used for conjugation. The quickliquid or quenched mixture used for conjugation is subjected to tangential flow filtration as described herein. Alternatively, the methods may include the steps of contacting the antibody or linker antibody with the free drug under conditions sufficient to produce a reaction mixture used for conjugation containing a benzodiazepine ADC. In some cases, the reaction mixture used for conjugation may be contacted with a quenching additive to produce a quenched reaction mixture used for conjugation. The quicklimited or quenched conjugation mixture is subjected to tangential flow filtration as described herein. An impurity related to a benzodiazepine drug may be, for example, a quenched or unquenched linker drug molecule or a quenched or unquenched drug molecule. The methods of the present invention are effective for removing impurities related to a benzodiazepine drug.

Preferably, the amount of impurities related to the benzodiazepine drug is reduced to a level of about 1 μM or less, 0.5 μM or less, 0.1 μM or less, or 0.05 μM or less.

Beta and gamma cyclodextrins, including chemically modified beta and gamma cyclodextrins, are particularly useful for use in the present invention. The cyclodextrin may be, for example, hydroxypropyl beta-cyclodextrin or beta-cyclodextrin sulfobutyl ether. In some aspects, if the cyclodextrin is a gammacyclodextrin, the cyclodextrin concentration is maintained at at least about 1% w/v. during tangential flow filtration, and if the cyclodextrin is a betacyclodextrin, maintain the cyclodextrin concentration at least about 2% w/v. or at least about 3% w/v. during tangential flow filtration.

Also provided are pharmaceutical compositions containing a benzodiazepine ADC and cyclodextrin at a concentration of from about to about 30% w/v, preferably at a concentration of from about 5% w/v. or 6% w/v. up to approximately 30% w/v. or from about 6% w/v. up to approximately 10% w/v. The compositions may be in aqueous or non-aqueous form. The compositions may contain additional excipients such as a lyoprotectant (preferably a sugar such as sucrose). The lyoprotectant can be at any concentration that is effective, for example, from about 4 to about 8% (wt./vol.), preferably about 6% (wt./vol.). The pH value of the composition is a physiologically appropriate pH value. Exemplary pH values are from about 6.0 to 8.0, or from about 6.5 to 7.5, or from about 7 to 7.5. The compositions typically contain a buffering agent. The buffering agent may be selected from a wide range of buffering agents, including tris, acetate, histidine, citrate, phosphate and succinate. Tris, for example, may be present at a concentration of approximately 20 mM. The concentration of the benzodiazepine ADC in the composition can vary widely. In preferred aspects, the ADC is present at a concentration of about

- 3 043811

0.5 to about 30 mg/mL, about 0.5 to about 10 mg/mL, about 1 to about 10 mg/mL, about 2 to about 10 mg/mL, about 2 to about 5 mg/mL ml, preferably about 3 mg/ml. Beta and gamma cyclodextrins are particularly effective for use in compositions, including chemically modified beta and gamma cyclodextrins. The cyclodextrin may be, for example, hydroxypropyl beta-cyclodextrin or beta-cyclodextrin sulfobutyl ether.

Provided are pharmaceutical compositions containing ADC PBD, wherein the concentration of ADC is from about 2 to about 5 mg/ml; hydroxypropyl-cyclodextrin in a concentration of approximately 5% w/v. up to about 10% w/v; sugar in a concentration of from about 4% to about 8%; and at least one buffering agent; wherein the composition is in aqueous solution, and wherein the concentration of at least one buffering agent is effective to maintain a physiologically appropriate pH value (e.g., from about 6 to about 8, more preferably from about 7 to about 8, or from about 7 to about 7 ,5).

Provided are pharmaceutical compositions containing ADC PBD, where the concentration of ADC is approximately 3 mg/ml; Hydroxypropyl-cyclodextrin is present at a concentration of approximately 6%; sucrose is present at a concentration of about 6%, tris is present at a concentration of about 20 mM, and wherein the pH of the composition is from about 7 to 7.5 (eg, about 7.3).

In the methods and compositions presented herein, the ADC may have any of the formulas described herein. The ADC may be, for example, an ADC PBD. For example, an ADC might have the formula

or be a pharmaceutically acceptable salt thereof; wherein Ab is a monoclonal antibody and p is the average number of drug-linker molecules per antibody and is approximately 2. In other aspects, the ADC may comprise a monoclonal antibody conjugated to an indolinobenzodiazepine dimer or a monoclonal antibody conjugated to an oxazolidinobenzodiazepine dimer. In the methods and composition described herein, the antibody may be any antibody, including any monoclonal antibody, including the humanized antibodies 2H12 or 1F6 described herein. Conjugation of an antibody to a drug-linker molecule can be accomplished by any of the methods known in the art, including conjugation using the sulfur atom of an introduced cysteine residue.

The description also provides methods for preparing a stabilized lyophilized antibody-drug conjugate composition. The methods may include preparing an aqueous composition as described herein; and lyophilizing the aqueous solution to form a lyophilized antibody-drug conjugate composition. Stabilized lyophilized antibody-drug conjugate compositions thus prepared are also provided.

The disclosure also provides methods for preventing chemical degradation and fragmentation of a benzodiazepine drug linker molecule attached to an antibody. The methods may include combining a benzodiazepine drug linker molecule attached to an antibody with at least about 6% w/v. gamma-cyclodextrin or chemically modified beta-cyclodextrin, as described in any of the embodiments presented in the description.

Definitions

Unless otherwise specified, all technical and scientific terms used in the description have the meaning commonly understood by one of ordinary skill in the art relating to the methods and compositions described. As used herein, the following terms and phrases have the meanings assigned to them unless otherwise noted.

The term heterocycle as used herein refers to a monocyclic, bicyclic or polycyclic ring system having from 3 to 14 ring atoms (also referred to as ring members), wherein at least one ring atom in at least one ring is a heteroatom selected from N, O, P or S (and to all combinations and subcombinations of ranges and specific values of carbon atoms and heteroatoms therein). The heterocycle may have from 1 to 4 ring heteroatoms independently selected from N, O, P or S. One or more N, C or S atoms in the heterocycle may be oxidized. The monocyclic heterocycle preferably has 3 to 7 ring members (e.g. 2 to 6 carbon atoms and 1 to 3 heteroatoms independently selected from N, O, P

- 4 043811 or S), and the bicyclic heterocycle preferably has from 5 to 10 ring members (eg 4 to 9 carbon atoms and 1 to 3 heteroatoms independently selected from N, O, P or S). The ring that includes the heteroatom may be aromatic or non-aromatic.

The term carbocycle in this context refers to a saturated or unsaturated aromatic non-monocyclic, bicyclic or polycyclic ring system having from 3 to 14 ring atoms (and all combinations and subcombinations of ranges and specific values of carbon atoms therein), where all ring atoms are carbon. Monocyclic carbocycles preferably have 3 to 6 ring atoms, more preferably 5 or 6 ring atoms. Carbocycles preferably have from 3 to 8 carbon atoms in the ring.

The phrase pharmaceutically acceptable salt as used herein refers to pharmaceutically acceptable organic or inorganic salts of a compound. The compound may contain at least one amino group, and accordingly, acid addition salts may be formed with the amino group. Typical salts include, but are not limited to, sulfate, trifluoroacetate, citrate, acetate, oxalate, chloride, bromide, iodide, nitrate, bisulfate, phosphate, acid phosphate, isonicotinate, lactate, salicylate, acid citrate, tartrate, oleate, tannate, pantothenate, bitartrate, ascorbate, succinate, maleate, gentisinate, fumarate, gluconate, glucuronate, saccharate, formate, benzoate, glutamate, methanesulfonate, ethanesulfonate, benzenesulfonate, p-toluenesulfonate and pamoate (i.e. 1,1'-methylene-bis -(2-hydroxy-3-naphthoate)).

The pharmaceutically acceptable salt may involve the attachment of another molecule, such as an acetate ion, a succinate ion, or another counterion. The counterion can be any organic or inorganic functional group that stabilizes the charge on the parent compound. In addition, a pharmaceutically acceptable salt may have more than one charged atom in its structure. In cases where multiple charged atoms are part of a pharmaceutically acceptable salt, multiple counterions may exist. Thus, a pharmaceutically acceptable salt may have one or more charged atoms and/or one or more counterions.

A polypeptide or polypeptide chain is a polymer consisting of amino acid residues linked by peptide bonds, either naturally produced or synthesized. Polypeptides consisting of less than about 10 amino acid residues are usually called peptides.

A protein is a macromolecule containing one or more polypeptide chains. The protein may also contain non-peptide components such as carbohydrate groups. Carbohydrates and other non-peptide substituents may be added to the protein by the cell in which the protein is produced, and these will vary depending on the type of cell. Proteins are defined herein in terms of the amino acid structures of their backbones; substituents such as carbohydrate groups are usually not specified, but may nevertheless be present.

The terms amino-terminal and carboxy-terminal are used throughout the description to refer to positions in polypeptides. When the context permits, the terms are used with reference to a specific sequence or portion of a polypeptide to indicate proximity or relative position. For example, a particular sequence located in the carboxy-terminal direction of a reference sequence within a polypeptide is located closer to the carboxyl terminus of the reference sequence, but not necessarily at the carboxyl terminus of the complete polypeptide.

The term antibody is used herein to refer to immunoglobulin proteins that are produced by the body in response to the presence of an antigen and bind to the antigen, as well as antigen-binding fragments and genetically engineered variants thereof. Thus, the term antibody includes, for example, intact monoclonal antibodies containing full-length immunoglobulin heavy and light chains (eg, antibodies produced using hybridoma technology) and antigen-binding antibody fragments, such as F(ab') ₂ and Fab fragments. Genetically engineered intact antibodies and fragments, such as chimeric antibodies, humanized antibodies, single-chain Fv fragments, single-chain antibodies, diabodies, mini-antibodies, linear antibodies, multivalent or multispecific (eg bispecific) hybrid antibodies, etc. also included. Thus, the term antibody is used broadly to include any protein that contains the antigen-binding portion of an antibody and is capable of specifically binding to an antigen.

The term genetically engineered antibodies refers to antibodies in which the amino acid sequence has been changed compared to the sequence of the native antibody. Due to the importance of recombinant DNA technologies in the creation of antibodies, there is no need to be limited by the amino acid sequences found in natural antibodies; antibodies can be engineered to produce desired characteristics. The possible options are numerous and range from substitution of only one or a few amino acids to complete reconstruction, for example, of the variable or constant region. Changes to the constant region are usually made to improve or change

- 5 043811 characteristics such as, for example, complement fixation, interaction with cells, and other effector functions. Typically, changes in the variable region are made to improve antigen-binding characteristics, increase stability of the variable region, or reduce the risk of immunogenicity.

The antigen-binding site of an antibody is that portion of the antibody that is sufficient to bind an antigen. The smallest such region is usually a variable domain or a genetically engineered variant thereof. Single domain binding sites can be derived from camel antibodies (see Muyldermans and Lauwereys, J. Mol. Recog. 12: 131-140, 1999; Nguyen et al., EMBO J. 19:921-930, 2000) or from VH- domains of antibodies of other species to create single-domain antibodies (dAbs; see Ward et al., Nature 341:544-546, 1989; US patent No. 6248516, Winter et al.). In specific embodiments, the antigen binding site is a polypeptide region having only 2 complementarity determining regions (CDRs) of a naturally occurring or non-naturally occurring (e.g., mutagenized) heavy chain variable domain or light chain variable domain, or a combination thereof (e.g. , Pessi et al., Nature 362:367-369, 1993; Qiu et al., Nature Biotechnol. 25:921-929, 2007). Typically, the antigen binding site of an antibody contains both a heavy chain variable domain (VH) and a light chain variable domain (VL) that bind to a common epitope. In the context of the present invention, an antibody may include one or more components in addition to an antigen binding site, such as, for example, a second antigen binding site of the antibody (which may bind the same or a different epitope or the same or a different antigen), a peptide linker , immunoglobulin constant region, immunoglobulin hinge region, amphipathic helix (see Pack and Pluckthun, Biochem. 31:1579-1584, 1992), non-peptide linker, oligonucleotide (see Chaudri et al., FEBS Letters 450:23-26, 1999), a cytostatic or cytotoxic drug, etc., and may be a monomeric or multimeric protein. Examples of molecules containing an antibody antigen binding site are known in the art and include, for example, Fv, single chain Fv (scFv), Fab, Fab', F(ab') ₂ , F(ab)c, diabodies, dAbs, miniantibodies, nanoantibodies , Fab-scFv fusions, bispecific (scFv) ₄ -IgG and bispecific (scFv) ₂ -Fab (see, for example, Hu et al, Cancer Res. 56: 3055-3061, 1996; Atwell et al, Molecular Immunology 33: 1301-1312, 1996; Carter and Merchant, Curr. Opin. Biotechnol. 8: 449454, 1997; Zuo et al., Protein Engineering 13: 361-367, 2000; and Lu et al., J. Immunol. Methods 267: 213-226, 2002).

As used herein, the term immunoglobulin refers to a protein consisting of one or more polypeptides substantially encoded by the immunoglobulin gene(s). One form of immunoglobulin constitutes the basic structural unit of native (i.e., natural) antibodies in vertebrates. This form is a tetramer and consists of two identical pairs of immunoglobulin chains, each pair having one light chain and one heavy chain. In each pair, the variable regions of the light and heavy chains (VL and VH) together are primarily responsible for binding to antigen, and the constant regions are primarily responsible for the effector functions of the antibody. Five classes of immunoglobulin protein (IgG, IgA, IgM, IgD, and IgE) have been identified in higher vertebrates. IgG constitutes the main class; it is typically the second most abundant protein found in blood plasma. In humans, IgG consists of four subclasses called IgG1, IgG2, IgG3 and IgG4. The heavy chain constant regions of the IgG class are designated by the Greek letter y. For example, immunoglobulins of the IgG1 subclass contain a γ1 heavy chain constant region. Each immunoglobulin heavy chain has a constant region, which consists of constant region protein domains (CH1, hinge, CH2 and CH3; IgG3 also contains a CH4 domain) that are essentially invariant for a given subclass in the species. DNA sequences encoding human and non-human immunoglobulin chains are known in the art (see, for example, Ellison et al., DNA 1:11-18, 1981; Ellison et al., Nucleic Acids Res. 10:4071-4079, 1982; Kenten et al., Proc. Natl. Acad. Sci. USA 79:66616665, 1982; Seno et al., Nuc. Acids Res. 11:719-726, 1983; Riechmann et al., Nature 332:323-327, 1988 Amster et al., Nuc. Acids Res. 8:2055-2065, 1980; Rusconi and Kohler, Nature 314:330-334, 1985; Boss et al., Nuc. Acids Res. 12:3791-3806, 1984; Bothwell et al., Nature 298:380-382, 1982; van der Loo et al., Immuno genetics 42:333-341, 1995; Karlin et al., J. Mol. Evol. 22: 195-208, 1985; Kindsvogel et al., DNA 1:335-343, 1982; Breiner et al., Gene 18: 165-174, 1982; Kondo et al., Eur. J. Immunol. 23:245-249, 1993; and from GenBank catalog No. J00228). For a review of the structure and function of immunoglobulins, see Putnam, The Plasma Proteins, Vol V, Academic Press, Inc., 49-140, 1987; and Padlan, Mol. Immunol. 31: 169-217, 1994. The term immunoglobulin is used in the description in its usual sense and refers to the intact antibody, its constituent chains or fragments of chains, depending on the context.

Full-length immunoglobulin light chains (approximately 25 kDa or 214 amino acids) are encoded at the amino terminus by a variable region gene (encoding approximately 110 amino acids) and a kappa or lambda constant region gene at the carboxyl terminus. Full-length immunoglobulin heavy chains (approximately 50 kDa or 446 amino acids) are encoded by the variable region gene (encoding approximately 116 amino acids) and the gamma, mu, alpha, delta or epsilon constant region gene (encoding approximately 330 amino acids), the latter op.

- 6 043811 defines the isotype of an antibody as IgG, IgM, IgA, IgD or IgE, respectively. In the light and heavy chains, the variable and constant regions are connected by a J' region of about 12 or more amino acids, with the heavy chain also including a D' region of about 10 or more amino acids (see generally Fundamental Immunology (Paul, ed. , Raven Press, N.Y., 2nd ed. 1989), Ch. 7).

The variable region of an immunoglobulin light or heavy chain (also referred to herein as the variable light chain domain (VL domain) or the variable heavy chain domain (VH domain), respectively) consists of a framework region interrupted by three hypervariable regions, also called complementarity determining regions or CDRs. . The framework regions serve to align the CDRs for specific binding to an antigen epitope. Thus, the term hypervariable region or CDR refers to the amino acid residues of an antibody that are primarily responsible for antigen binding. Starting from the amino terminus to the carboxy terminus, both VL and VH domains contain the following framework (FR) and CDR regions: FR1, CDR1, FR2, CDR2, FR3, CDR3, FR4. The assignment of amino acids to each domain is carried out in accordance with the definitions of Kabat, Sequences of Proteins of Immunological Interest (National Institutes of Health, Bethesda, MD, 1987 and 1991), or Chothia & Lesk., Mol. Biol. 196:901-917, 1987; Chothia et al., Nature 342:878-883, 1989. Kabat also provides a widely used numbering system (Kabat numbering) in which corresponding residues in different heavy chains or in different light chains are assigned the same number. CDRs 1, 2 and 3 of the VL domain are also referred to herein as CDR-L1, CDR-L2 and CDR-L3, respectively; CDRs 1, 2 and 3 of the VH domain are also referred to herein as CDR-H1, CDR-H2 and CDR-H3, respectively.

Unless the context indicates otherwise, the term monoclonal antibody as used herein is not limited to antibodies produced by hybridoma technology. The term monoclonal antibody refers to an antibody that is derived from a single clone, including any eukaryotic, prokaryotic or phage clone, and does not refer to the method by which it is produced.

The term chimeric antibody refers to an antibody having variable domains derived from a first species and constant regions derived from a second species. Chimeric immunoglobulins or antibodies can be constructed, for example, by genetic engineering from segments of immunoglobulin genes belonging to different species. The term humanized antibody, as defined below, does not include chimeric antibodies. Although humanized antibodies are chimeric in design (i.e., containing regions derived from more than one protein species), they contain additional features (i.e., variable regions containing CDR donor residues and framework acceptor residues) not found in chimeric immunoglobulins or antibodies, as defined below.

The term humanized VH domain or humanized VL domain refers to an immunoglobulin VH or VL domain containing some or all of the CDRs derived entirely or substantially from a non-human donor immunoglobulin (eg, mouse or rat) and variable region framework sequences derived wholly or mainly from human immunoglobulin sequences. The non-human immunoglobulin providing the CDR is called a donor and the human immunoglobulin providing the scaffold is called an acceptor. In some cases, humanized antibodies may retain non-human residues in the framework regions of the variable domains to improve the associated binding characteristics (for example, mutations in the framework regions may be required to maintain binding affinity when humanizing the antibody).

A humanized antibody is an antibody containing one or both of a humanized VH domain and a humanized VL domain. Immunoglobulin constant region(s) are not necessarily present, but if present, they are derived entirely or substantially from immunoglobulin constant regions.

A CDR in a humanized antibody is derived substantially from the corresponding CDR of a non-human antibody if at least 60%, at least 85%, at least 90%, at least 95, or 100% of the corresponding residues (as defined by Kabat) are identical in the relevant CDRs. In specific embodiments of the humanized VH or VL domain, in which the CDRs are substantially derived from a non-human immunoglobulin, the CDRs of the humanized VH or VL domain have no more than six (e.g., no more than five, no more than four, no more than three, no more than more than two or no more than one) amino acid substitutions among all three CDRs, compared with the corresponding non-human VH or VL CDRs. The antibody VH or VL variable domain framework sequences or, when present, immunoglobulin constant region sequences are derived substantially from a human VH or VL framework sequence or a human immunoglobulin constant region, respectively, if at least 85%, at least 90 %, at least 95 or 100% of the corresponding residues are identical as defined by Kabat. Thus, all parts of a humanized antibody, with the possible exception of the CDR, are derived entirely or substantially from the corresponding parts of natural sequences.

- 7 043811 importance of human immunoglobulin.

Specific binding of an antibody to its target antigen implies an affinity of at least 10 ⁶ , 10 ⁷ , 10 ⁸ , 10 ⁹ or 10 ¹⁰ M ^-1 . Specific binding is markedly higher in magnitude and differs from nonspecific binding, which occurs to at least one extraneous target. Specific binding may result from the formation of bonds between specific functional groups or may result from a specific spatial correspondence (eg, a key and a lock), whereas nonspecific binding is generally the result of van der Waals forces. However, specific binding does not necessarily imply that the monoclonal antibody binds to one and only one target.

With respect to the proteins that are described herein, reference to amino acid residues corresponding to those residues identified by SEQ ID NO includes post-translational modifications of such residues.

The term stabilized in the context of antibody-drug conjugate compositions as described herein refers to a composition in which the antibody-drug conjugate generally retains its physical and chemical identity and integrity upon storage. Various analytical methods for measuring protein stability are available in the art (see, for example, Peptide and Protein Drug Delivery, 247-301 (Vincent Lee Ed., Marcel Dekker, Inc., New York, N.Y., Pubs. 1991) and Jones, Adv Drug Delivery Rev 10:29-90, 1993). Typical techniques for measuring protein stability are also described herein (see examples below). Stability can be measured at a selected temperature for a selected period of time. For rapid testing, the drug can be stored at a higher or accelerated degradation temperature, such as 40°C for 1 week to 1 month or more, while stability is measured. In exemplary embodiments, the composition is resistant to the formation of antibody protein component by-products, such as high molecular weight aggregation products, low molecular weight degradation or fragmentation products, acidic compounds, chemical breakdown products, or mixtures thereof. The term stability refers to the length of time that a molecular species, such as an antibody, retains its original chemical identity, such as primary, secondary and/or tertiary structure.

The term by-product includes unwanted products that underestimate or reduce the proportion of therapeutic antibody-drug conjugate in a given composition. Typical by-products include antibody-drug conjugate aggregates, antibody-drug conjugate fragments (eg, those formed by degradation of the antibody protein by deamidation or hydrolysis or chemical degradation and fragmentation of the drug linker), acidic variants of the antibody-drug conjugate, or mixtures thereof.

An antibody-drug conjugate (ADC) is an antibody conjugated to a cytotoxic drug, typically through a linker. The linker may contain a cleavable portion or may be non-cleavable. Cleavable linkers include, for example, disulfide-containing linkers that are cleaved by disulfide exchange, acid-labile linkers that are cleaved at acidic pH, and linkers that are cleaved by hydrolases (e.g., glycosyl hydrolases such as glucuronidases), esterases and peptidases (eg, peptide linkers and glucuronide linkers). Non-cleavable linkers are believed to release drug through a mechanism of proteolytic degradation of the antibody.

The term high molecular weight aggregates includes antibody-drug conjugate (ADC) aggregates, as well as aggregates containing ADC fragments (eg, those formed by degradation of a polypeptide, such as by hydrolysis), and aggregates containing mixtures of ADCs and such fragments. The presence of high molecular weight aggregates can be determined, for example, using size exclusion chromatography (SEC). Typically, high molecular weight aggregates are complexes that have a molecular weight that is greater than the molecular weight of the therapeutic monomeric ADC. In the case of an ADC in which the antibody component is a tetramer consisting of two identical pairs of immunoglobulin chains, each pair having one light chain and one heavy chain (eg, IgG isotype), the molecular weight of such aggregates is greater than about 150 kDa. However, in the case of an ADC in which the antibody component has a molecular weight greater or less than the molecular weight of the protein of a typical monospecific, tetrameric antibody consisting of two immunoglobulin light chains and two immunoglobulin heavy chains (e.g., single chain antibodies or bispecific antibodies), the size of such aggregates may change accordingly.

The term low molecular weight degradation product includes, for example, antibody-drug conjugate (ADC) fragments, such as, for example, those generated by deamidation or hydrolysis. The presence of low molecular weight degradation products can be detected, for example, using gel chromatography (SEC). Typically, low molecular weight degradation products have a molecular weight that is less than the molecular weight of the therapeutic monomeric ADC. When

- 8 043811

An ADC in which the antibody component is a tetramer consisting of two identical pairs of immunoglobulin chains, each pair having one light chain and one heavy chain (eg, IgG isotype), the molecular weight of such aggregates being less than about 150 kDa. However, in the case of an ADC in which the antibody component has a molecular weight greater or less than the molecular weight of the protein of a typical monospecific, tetrameric antibody consisting of two immunoglobulin light chains and two immunoglobulin heavy chains (e.g., single chain antibodies or bispecific antibodies), the size of such products degradation can be changed accordingly.

An acidic variant of an investigational antibody-drug conjugate (ADC) is a variant of an ADC that is more acidic than the investigational PI ADC.

The presence of acidic variants can be detected, for example, using cation exchange chromatography or capillary isoelectric focusing with capillary detection (icIEF). An example of a sour variant is the deamidated variant. Deamidated variants of a protein molecule are variants in which one or more neutral side chain amides(s) have been converted to a residue with a general acidic character (eg, one or more asparagine residues of the parent polypeptide have been converted to aspartate).

The term diluent as used herein refers to a solution suitable to vary or achieve the typical or corresponding concentration or concentrations that are described herein.

The term container refers to anything in which an object or liquid can be placed or contained, such as for storage (eg, a reservoir, receptacle, vessel, or the like).

The term route of administration includes art-accepted routes of delivery of therapeutic protein, such as, for example, parenteral, intravenous, intramuscular, or subcutaneous routes of administration. When administering an ADC for the treatment of cancer, it may be desirable to administer it into the systemic circulation by intravenous or subcutaneous administration. For the treatment of cancer characterized by a solid tumor, administration may be limited to the tumor itself if desired.

The term treatment refers to the administration of a therapeutic agent to a patient who has a disease for the purpose of curing, healing, partially eliminating symptoms, slowing the progression of the disease, alleviating symptoms, modifying, eliminating the disease, improving the condition, curing or affecting the disease.

The term patient includes humans and other mammals who are receiving either prophylactic or therapeutic treatment.

The term effective amount, effective dose or effective dosage refers to an amount that is sufficient to achieve or at least partially achieve a desired effect, for example, sufficient to inhibit the occurrence or improve one or more symptoms of a disease or disorder. An effective amount of the pharmaceutical composition is administered in an effective manner. The term effective regimen refers to the combination of the amount of composition that is administered and the frequency of dosing to provide prophylactic or therapeutic treatment of a disease or disorder.

The term dosage form (or unit dosage form) in this context refers to a physically discrete unit suitable as unitary doses for a patient receiving treatment, each unit containing a predetermined amount of active compound (ADC according to the present invention) calculated to obtain the desired therapeutic effect in combination with the desired pharmaceutical carrier, diluent or excipient. The specification of the dosage forms of the invention is dictated by and directly depends on the unique characteristics of the active compound and the particular therapeutic effect to be achieved and the limitations existing in the art of preparing mixtures of such active compound for the treatment of patients.

The actual dosage levels of the ADCs in the composition of the present invention may be adjusted to provide an amount of ADCs that is effective to achieve the desired therapeutic response for a particular patient, composition, and route of administration, and is not toxic to the patient. The dose level selected will depend on various pharmacokinetic factors, including the potency of the specific compositions of the present invention used, route of administration, excretion rate of the particular compound used, duration of treatment, other drugs, compounds and/or materials used in combination with the particular compositions, age , gender, weight, disease, general health and medical history of the patient receiving treatment, and similar factors well known in the medical field.

The cytotoxic effect refers to the depletion, elimination and/or death of the target cell. A cytotoxic agent refers to an agent that has a cytotoxic effect on cells.

Cytostatic effect means inhibition of cell proliferation. Cytostatic agent means an agent that has a cytostatic effect on a cell, thereby inhibiting the growth and/or proliferation of a specific subpopulation of cells.

- 9 043811

Two amino acid sequences have 100% amino acid sequence identity if the amino acid residues of the two amino acid sequences are the same when they are aligned to determine the best match. Sequence comparisons can be made using standard software, such as the software included in the LASERGENE bioinformatics software package produced by DNASTAR (Madison, Wisconsin). Other methods for comparing two nucleotide or amino acid sequences by determining optimal alignment are well known to those skilled in the art (see, for example, Peruski and Peruski, The Internet and the New Biology: Tools for Genomic and Molecular Research (ASM Press, Inc. 1997); Wu et al. (eds.), Information Superhighway and Computer Databases of Nucleic Acids and Proteins, in Methods in Gene Biotechnology 123-151 (CRC Press, Inc. 1997); Bishop (ed.), Guide to Human Genome Computing (2nd ed., Academic Press, Inc. 1998)). Two amino acid sequences are considered to have significant sequence identity if the two sequences have at least 80%, at least 85%, at least 90%, or at least 95% sequence identity relative to each other.

Percentage sequence identity values are determined for antibody sequences maximally aligned to the Kabat numbering system. After alignment, if the antibody region of interest (e.g., the complete variable domain of the heavy or light chain) is compared with the same region of a reference antibody, the percentage of sequence identity between the antibody region of interest and the reference region is the number of positions occupied by the same amino acid in both the antibody region of interest and the reference region. reference antibody regions divided by the total number of aligned positions of the two regions, excluding gaps, multiplied by 100 to convert to a percentage.

The term pharmaceutical composition refers to a preparation that is in a form that provides the biological activity of the active ingredient that is intended to be effective (when administered to an individual) and which does not contain additional components that are unacceptably toxic to the individual to whom the composition is administered. Such compositions are sterile.

Compositions or methods containing one or more of the listed elements may include other elements not specifically listed.

A reference to a numeric range in a description (for example, X to Y or X to Y) includes the endpoints that define the range and all values that fall within the range.

Unless the context otherwise requires, when a value is expressed as approximately X or approximately X, the stated value of X should be understood to the nearest ±10%.

Description of Illustrative Embodiments Antibody-Benzodiazepine Drug Conjugates

An antibody-benzodiazepine drug conjugate refers to an antibody conjugated to a benzodiazepine dimer, usually, although not necessarily, via a linker. The term benzodiazepine drug-linker refers to a benzodiazepine dimer attached to a linker. The linker component of a benzodiazepine drug-linker compound typically has a functional group for attachment to an antibody. A benzodiazepine compound has a benzene ring in its central part fused to a diazepine ring. Typical ring structures of a benzene ring fused to a diazepine ring are the following structures:

3,4-dihydro-1H-3H-benzo[e][1,4]diazepinebenzo[e] [1,4]diazepin-5 (2H)-one 5(4H)-one

Benzodiazepine compounds differ in the number, type and position of substituents on both rings and in the degree of saturation of the diazepine ring. They also differ in the number of additional rings fused to the benzene and/or diazepine ring. The definition of a benzodiazepine compound includes compounds in which a benzene or diazepine ring is fused to one or more aromatic or non-aromatic carbocyclic or heterocyclic rings. A benzodiazepine dimer is a compound that is formed by joining two benzodiazepine units together using a linking group.

An antibody, as a component of an antibody-benzodiazepine drug conjugate, can be conjugated to one or more benzodiazepine drug linker molecules, such as 1 to 20 benzodiazepine drug linker molecules. In some aspects, the antibody, as a component of an antibody-benzodiazepine drug conjugate, will be conjugated to 1, 2, 3, or 4 drug-linker molecules. Conjugation can take place

- 10 043811 to flow through various positions in the antibody molecule. In some aspects, the conjugation will be accomplished using the sulfur atom of the cysteine residue. In some aspects, the cysteine residue is a cysteine residue of an antibody's interchain disulfides. In other aspects, a cysteine residue is incorporated into the antibody. In some aspects, a cysteine residue is inserted into the antibody at position 239 (human IgG), which is defined by the EU index (Kabat, Sequences of Proteins of Immunological Interest (National Institutes of Health, Bethesda, MD, 1987 and 1991)). In some aspects, there is an average of two drug-linker molecules per antibody in the benzodiazepine ADC mixture or composition, and the drug-linker molecules will be conjugated to a cysteine residue introduced into the antibody at position 239 (human IgG), which is defined according to the EU index (Kabat, Sequences of Proteins of Immunological Interest (National Institutes of Health, Bethesda, MD, 1987 and 1991)).

In one aspect, the benzodiazepine dimer is a pyrrolobenzodiazepine dimer (PBD). PBDs have a general structure:

PBDs vary in the number, type, and position of substituents on both the aromatic ring A and the pyrrole ring C, and in the degree of saturation of the C ring. Ring B contains either an imine (N=C), a carbinolamine (NH-CH(OH)), or carbinolamine ester (NH-CH(OR)) at the N10-C11 position, which is the electrophilic center responsible for DNA alkylation. All known natural products have a ^-configuration at the C11a chiral position, which produces a right-handed twist when viewed from the C-ring to the A-ring. This twisting gives them a corresponding three-dimensional shape, isospiral to the minor groove of the B-form of DNA, leading to a tight fit at the binding site.

The ability of PBDs to form an adduct in the minor groove ensures their ability to interfere with DNA processing, hence enabling their use as antitumor agents. The biological activity of these molecules can be enhanced, for example, by linking together two PBD units (eg, using C8/C'-hydroxyl groups through a flexible alkylene linker). PBD dimers are thought to create DNA sequence-selective lesions, such as palindromic 5'-Pu-GATC-Py-3' interstrand cross-links, which are believed to primarily account for their biological activity.

Typical PBD dimers that can be used as conjugates are the following dimers:

- 11 043811

or a salt thereof (eg, a pharmaceutically acceptable salt).

The PBD dimer can be linked to the antibody at any position suitable for conjugation to the linker. For example, in some embodiments, the PBD dimer will have a substituent at the C2 position (eg, a primary or secondary amine) that provides an anchor for binding of the compound to the antibody. The position of C2 is indicated by an arrow in the typical structures shown above. In alternative embodiments, the N10 position of the PBD dimer will provide an anchor for binding of the compound to the antibody.

- 12 043811

In another aspect, the benzodiazepine dimer is an indolinobenzodiazepine dimer or an oxazolidinobenzodiazepine dimer. Indolinobenzodiazepines (IBD) and oxazolidinobenzodiazepines (OBD) have a common structure:

Indolinobenzodiazepines and oxazolidinobenzodiazepines differ in the number, type and position of substituents on the rings. As with PBD, two indolinobenzodiazepine units or two oxazolidinobenzodiazepine units can be linked together to form dimers, for example, through the ester groups of the A-rings of the two monomer units. As with PBD, the indolinobenzodiazepine dimer or oxazolidinobenzodiazepine dimer can be linked to the antibody at any position suitable for conjugation to the linker.

A benzodiazepine ADC that contains a dimer as a drug moiety may also be designated as a PBD ADC. Likewise, a benzodiazepine ADC that contains an indolinobenzodiazepine dimer as a drug component may be referred to as an IBD ADC, and an ADC that contains an oxazolidinobenzodiazepine dimer as a drug component may be referred to as an OBD ADC. Typically, benzodiazepine ADCs, including PBD ADCs, IBD ADCs, and OBD ADCs, contain a linker between a benzodiazepine drug and an antibody. The linker may contain a cleavable portion (eg, an amino acid or a contiguous sequence of amino acids that is a target substrate for the enzyme) or may be a non-cleavable linker (eg, a linker released by degradation of an antibody). The linker may further contain a maleimide group for binding to the antibody, including, for example, maleimidocaproyl. The linker may contain an alternative group for binding to the antibody, including, for example, N-hydroxysuccinimidyl ester or a labile disulfide, including, for example, thiopyridyl disulfide. PBD dimers, IBD dimers, OBD dimers, linkers and conjugates thereof are known in the art (see, for example, WO 2010/091150, WO 2012/112708, WO 2012/128868, WO 2011/023883 and WO 2009/016516).

A typical linker for binding an antibody to a benzodiazepine dimer, including any of the dimers described herein, is shown below, in which a wavy line represents the site of drug attachment and the antibody is attached via a maleimide group.

Typical PBD antibody-drug conjugates include antibody-drug conjugates according to the formula below:

or a salt thereof (eg, a pharmaceutically acceptable salt), wherein Ab is an antibody (eg, a monoclonal antibody) and the drug content is represented by p, the number of linker drug molecules per antibody. Depending on the context, p may represent the average number of drug-linker molecules per antibody, also referred to as average drug content. p ranges from 1 to 20 and is preferably from 1 to 8. In some aspects, if p is the average drug content, p ranges from about 2 to about 5. In some aspects, p is about 2, about 3 , about 4 or about 5. In some aspects, the antibody is conjugated to a drug linker via the sulfur atom of a cysteine residue. In some aspects, the cysteine residue is the cysteine residue of the interchain disulfide bonds of the antibody. In other aspects, a cysteine residue is incorporated into the antibody. In some aspects, a cysteine residue is inserted into the antibody at position 239 (IgG1), which is identified by

- 13 043811 numbering according to the EU index (Kabat, Sequences of Proteins of Immunological Interest (National Institutes of Health, Bethesda, MD, 1987 and 1991)). In some such aspects, p is approximately 2. Methods for creating such ADCs are known in the art (see, for example, International Publication No. WO 2011/130613).

Conjugation reaction

The present invention relates, in particular, to methods for removing benzodiazepine drug related impurities from a mixture containing a benzodiazepine ADC and benzodiazepine drug related impurities. A benzodiazepine drug impurity is any drug impurity resulting from the conjugation reaction of an antibody with a benzodiazepine dimer or a benzodiazepine drug linker molecule. Impurities due to the benzodiazepine drug may contain, for example, free benzodiazepine dimer drugs, benzodiazepine drug linker molecules, quenched benzodiazepine drug linker molecules, or degradation products of benzodiazepine drug linker molecules. The impurities associated with the benzodiazepine drug are not benzodiazepine drugs conjugated to antibodies or drug-linker molecules conjugated to antibodies.

In some aspects of the present invention, the antibody is contacted with a benzodiazepine drug linker molecule under conditions sufficient to form a mixture containing the benzodiazepine ADC and benzodiazepine drug related impurities (also referred to herein as the conjugation reaction mixture). In other aspects, an antibody-linker molecule (an antibody conjugated to a linker) is contacted with a free benzodiazepine drug under conditions sufficient to create a mixture containing a benzodiazepine ADC and impurities related to the benzodiazepine drug (also referred to herein as the reaction mixture used for conjugation). In other aspects, an antibody linker molecule is contacted with a benzodiazepine drug linker molecule under conditions sufficient to form a mixture containing a benzodiazepine ADC and benzodiazepine drug related impurities (also referred to herein as a conjugation reaction mixture). General methods for conjugating antibodies to linkers or drug-linker molecules are known in the art and are not described in detail. In some aspects, conjugation will be to lysine residues of the antibody. In other aspects, the conjugation will be to a native or incorporated cysteine present in the antibody (eg, an interchain disulfide cysteine or a cysteine residue introduced into the heavy or light chain of the antibody). In some aspects, the conjugation will be to the embedded cysteine present in the antibody, the antibody will be reduced before contacting the benzodiazepine drug linker molecule, the antibody will be partially reoxidized (i.e., reoxidized to the interchain disulfides but not to the embedded cysteine ) and a benzodiazepine drug linker molecule will be conjugated to the embedded cysteine of the partially re-oxidized antibody. In some such aspects, the inserted cysteine will be at position 239 (IgG1, EU index numbering as reported by Kabat).

One skilled in the art will appreciate that the conditions used for conjugation of an antibody or antibody-linker molecule and a drug or drug-linker molecule will depend, in part, on the characteristics of the agent and linker. In general, conjugation reactions are carried out at a temperature of from about 0 to about 40°C. In some embodiments, the conjugation reaction is carried out at approximately 4°C. In some embodiments, the conjugation reaction is carried out at approximately 25°C. In some embodiments, the conjugation reaction is carried out at approximately 37°C. The conjugation reactions can be carried out for any suitable period of time. In general, mixtures for the conjugation reaction are incubated under suitable conditions from several minutes to several hours. Reactions can be carried out, for example, for about 1 minute or about 5 minutes, or about 30 minutes, or about 1 1/2 hours, or about 4 hours, or about 12 hours, or about 24 hours. In general, the mixtures used for the conjugation reaction, create a pH value ranging from about 6 to about 9 or from about 7 to about 8. Various buffering agents can be used to maintain a specific pH value. Examples of suitable buffering agents include, but are not limited to, 2-(N-morpholino)ethanesulfonic acid (MES), 2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid (HEPES), 3-morpholinopropane- 1sulfonic acid (MOPS), 2-amino-2-hydroxymethyl-propane-1,3-diol (TRIS), sodium citrate, sodium acetate and sodium borate. Co-solvents (eg, dimethylacetamide, propylene glycol, dimethyl sulfoxide, dimethylformamide, ethanol, methanol, tetrahydrofuran, acetone and acetic acid), salts (eg, NaCl, KCl, CaCl ₂ , and Mn ²⁺ and Mg ²⁺ salts), chelating agents (eg , ethylene glycol bis(2aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA), 2-({2-[bis(carboxymethyl)ami- 14 043811 but]ethyl}(carboxymethyl)amino)acetic acid (EDTA) and 1,2-bis(o-aminophenoxy)ethane-N,N,N,Ntetraacetic acid (BARTA)) may also be included if necessary. Buffers, cosolvents, salts and chelating agents can be used in any suitable concentration, which can be easily determined by one skilled in the art. In general, buffers, cosolvents, salts and chelating agents are included in the reaction mixtures at concentrations ranging from about 1 μM to about 1 M or even higher (eg, 0-50% v/v depending on the cosolvent). Any suitable amount of benzodiazepine drug or drug-linker compound can be used for conjugation.

In some aspects, after a conjugation reaction (eg, conjugation of an antibody to a linker drug molecule) and before performing tangential flow filtration, excess benzodiazepine drug-related impurities are rendered non-reactive using a quench additive. A quenching agent is a non-antibody reagent that is capable of eliminating the reactivity of a reactive group by covalently binding to the reactive group. One skilled in the art will appreciate that the quenching agent is selected based on the nature of the drug or linker. For example, a thiol-containing additive such as B-mercaptoethanol or N-acetylcysteine can be used to quench excess drug-linker compound containing a maleimide group or other thiol reactive group. An amine such as glycine can be used to quench excess drug-linker compound containing N-hydroxysuccinimide ester. Typically, the quenching agent is used in excess of the amount of antibody and drug-linker compound.

In some aspects, the benzodiazepine drug impurity will be a quenched benzodiazepine drug linker molecule. A quenched reaction mixture used for conjugation refers to the reaction mixture after a conjugation reaction (e.g., conjugation of an antibody to a drug-linker molecule, an antibody-linker molecule to a drug-linker molecule, or an antibody-linker molecule to a drug) has been carried out, administration of a quenching additives and extinguishing of impurities related to the benzodiazepine drug. Generally, the term quenched reaction mixture used for conjugation refers to the mixture prior to any purification steps.

The benzodiazepine dimers of the present invention are hydrophobic in nature. The inventors have surprisingly discovered that highly hydrophobic benzodiazepine drug impurities can be difficult to remove from ADC mixtures, even when the benzodiazepine drug impurity and the ADC are solubilized in the ADC mixture. The present inventors have discovered that benzodiazepine drug impurities having a SlogP value less than 7.50 are easier to remove from an ADC mixture using existing methods than benzodiazepine drug impurities having a SlogP value greater than 7.50. . Thus, in preferred embodiments, the benzodiazepine drug impurity will have a SlogP value of no more than 7.50, more preferably a SlogP value of no more than 7.0, even more preferably a SlogP value of no more than 6.5 or 6 .0, or even 5.8.

The inventors have also discovered that quenching agents can be used to reduce the hydrophobicity of impurities related to a benzodiazepine drug, thereby facilitating purification of the drug. By selecting quenching additives that help reduce the hydrophobicity of the compound to which they bind, the clearance of a benzodiazepine drug impurity from an ADC mixture can be improved. Accordingly, particularly preferred quenching additives are those which, when bound to the compound to be quenched, help reduce the hydrophobicity of the resulting compound. For example, a particularly preferred quenching agent will reduce the hydrophobicity of the benzodiazepine drug impurity (eg, drug or drug-linker molecule) to which it is attached. In some preferred aspects, the quenched benzodiazepine drug impurity will have a SlogP value of no more than 7.50, more preferably a SlogP value of no more than 7.0, even more preferably a SlogP value of no more than 6.5 or 6.0 , or even 5.8. In some such aspects, the unquenched benzodiazepine drug impurity has a higher SlogP value than the quenched benzodiazepine drug impurity.

SlogP is an indicator of hydrophobicity. SlogP is defined as the log value of the octanol/water partition coefficient (including implicit hydrogens) and can be calculated using the MOE program from the Chemical Computing group (SlogP values calculated using Wildman, S.A., Crippen, G.M.; Prediction of Physiochemical Parameters by Atomic Contributions; / Chem. Inf. Comput. Sci. 39 No. 5 (1999) 868-873).

Quenching agents that help reduce the hydrophobicity of the compound to which they bind include charged quenching agents as well as uncharged quenching agents which,

- 15 043811 are nevertheless hydrophilic. Hydrophilic quenching agents include thiol sugars such as thiol glucose and PEGylated quenching agents such as PEGylated thiols.

As an example of the synthesis of a drug-PBD-linker compound, WO 2011/130613 describes a method for synthesizing a drug-PBD-linker compound followed by conjugation of the drug-PBD-linker compound and antibodies. Briefly, antibodies in PBS containing 50 mM sodium borate at pH 7.4 were reduced with tris(carboxyethyl)phosphine hydrochloride (TCEP) at 37°C. Interchain antibody disulfides are converted by oxidation with dehydroascorbic acid, leaving the created cysteines in the thiol form available for alkylation by the drug linker. The reduced disulfide bonded antibody is then alkylated with approximately 1.5 maleimide equivalents of the drug-linker compound onto the antibody thiol, in the presence of sufficient cosolvent to solubilize the drug-linker compound. After approximately 90 minutes, the reaction is quenched by adding approximately 3 equivalents of N-acetylcysteine relative to the drug-linker compound.

Regardless of the conjugation methods used, impurities related to the benzodiazepine drug will be present in the mixture. Typically, but not in all cases, impurities related to the benzodiazepine drug will be present in the mixture in an amount of approximately 10 to 100 μM. In some aspects, the impurity related to the benzodiazepine drug will be a quenched or unquenched drug-linker compound, for example, a drug-linker compound quenched with N-acetylcysteine. In other aspects, the benzodiazepine drug impurity will be a quenched or unquenched free benzodiazepine drug (ie, a drug not attached to a linker or antibody). In other aspects, the benzodiazepine drug impurity will be a degradation product of the benzodiazepine drug-linker compound, such as, for example, an oxidized or hydrolyzed derivative of the drug-linker compound.

Tangential Flow Filtration (TFF)

Typically, after the conjugation reaction and optional quenching, the mixture containing the benzodiazepine ADC and benzodiazepine drug related impurities is subjected to tangential flow filtration (TFF). The inventors have found that to effectively remove drug impurities from a mixture, it is preferable to use cyclodextrin and maintain a minimum level of cyclodextrin during filtration. To optimize removal of benzodiazepine impurities, the cyclodextrin is maintained at a level that substantially solubilizes the components (eg, ADC) of the ADC mixture. Cyclodextrin is preferably added to the mixture before the filtration process begins (eg, after conjugation and optional quenching, but before tangential flow filtration begins) or at the beginning (or initiation) of the filtration process. In some aspects, the cyclodextrin will be added to the ADC mixture after the filtration process has begun, but before a significant portion of the impurities have been removed. A solubilizing component (eg cyclodextrin) is preferably added to the mixture prior to tangential flow filtration. The cyclodextrin concentration is preferably maintained at a minimum level throughout the filtration.

The inventors have found that cyclodextrin is a particularly preferred solubilizing component for use because its addition not only improves the removal of impurities related to the benzodiazepine drug, but its inclusion in the dosage form improves the stability of the dosage form. Cyclodextrin can be added to a mixture containing a benzodiazepine ADC and benzodiazepine drug-related impurities following a conjugation reaction and optional quenching. In some aspects, at least about 1% w/v. cyclodextrin (ie 10 g per liter) is added to the reaction mixture used for conjugation. In some aspects, at least about 2% w/v. cyclodextrin (ie 20 g per liter) is added to the reaction mixture used for conjugation. In some aspects, at least about 3% w/v. cyclodextrin (ie 30 g per liter) is added to the reaction mixture used for conjugation. In some aspects, from about 1% w/v. cyclodextrin to approximately 10% w/v. cyclodextrin, from about 2% w/v. cyclodextrin to approximately 10% w/v. cyclodextrin or from about 3% w/v. cyclodextrin to approximately 10% w/v. cyclodextrin, or from about 3% w/v. cyclodextrin to approximately 6% w/v. cyclodextrin, or from about 3% w/v. cyclodextrin to approximately 4% w/v. cyclodextrin is added to the cyclodextrin mixture.

Tangential flow filtration refers to a filtration process in which the sample to be purified circulates in a closed system tangentially along the surface of a semi-permeable membrane. Macromolecules that are too large to pass through the pores of the membrane are retained on the inlet side (the side of the membrane where the retentate is located) of the membrane, and molecules that are small enough to pass through the pores pass to the mem side

- 16 043811 branes where the filtrate is located. In general, determining which molecules are transferred to the filtrate and which molecules remain in the retentate depends primarily on the molecular weight, solubility, and pore size of the filter. Other factors, such as flow rate, transmembrane pressure, membrane type or composition, concentrations of components in the moving mixture, temperature and viscosity of the moving mixture, can affect the rate and extent of purification. General methods for using tangential flow filtration devices and performing tangential flow filtration to remove drug impurities from reaction mixtures used for conjugation are known in the art and can be optimized using the knowledge contained in the present invention in combination with knowledge known in the art to remove impurities related to a benzodiazepine drug.

In some aspects, the method of performing tangential flow filtration will be diafiltration. During diafiltration, a buffer is introduced while the filtrate is removed. Diafiltration may be, for example, constant volume diafiltration or batch diafiltration. In preferred embodiments, a constant level of cyclodextrin is maintained throughout the diafiltration. In preferred embodiments, the cyclodextrin is added to the mixture containing the benzodiazepine ADC and benzodiazepine drug related impurities before or at the start of the filtration process. In some aspects, the cyclodextrin concentration is maintained at at least about 1% w/v. during filtration. In some such aspects, the ADC mixture is supplemented with cyclodextrin before or at the beginning of the filtration process such that the mixture has a concentration of at least about 1% w/v. cyclodextrin and the diafiltration buffer contains at least about 1% w/v. cyclodextrin. In some aspects, a concentration of at least about 2% w/v. cyclodextrin is maintained throughout the filtration. In some such aspects, the ADC mixture is supplemented with cyclodextrin before or at the beginning of the filtration process such that the mixture has a concentration of at least about 2% w/v. cyclodextrin and the diafiltration buffer contains at least about 2% w/v. cyclodextrin. In some aspects, a concentration of at least about 3% w/v. cyclodextrin is maintained throughout the filtration. In some such aspects, the ADC mixture is supplemented with cyclodextrin before or at the beginning of the filtration process such that the mixture has a concentration of at least about 3% w/v. cyclodextrin and the diafiltration buffer contains at least about 3% w/v. cyclodextrin. In some aspects, the concentration is at least about 1%, at least about 2%, at least about 3% w/v. cyclodextrin is maintained throughout the filtration. In some such aspects, the ADC mixture is supplemented with cyclodextrin before or at the beginning of the filtration process such that the mixture has a concentration of at least about 1, 2, or 3% w/v. cyclodextrin and the diafiltration buffer contains at least about 1, 2 or 3% w/v. cyclodextrin. In other aspects, a level of from about 1% w/v. cyclodextrin to approximately 10% w/v. cyclodextrin, from about 2% w/v. cyclodextrin to approximately 10% w/v. cyclodextrin or from about 3% w/v. cyclodextrin to approximately 10% w/v. cyclodextrin, or from about 3% w/v. cyclodextrin to approximately 6% w/v. cyclodextrin, or from about 3% w/v. cyclodextrin to approximately 4% w/v. cyclodextrin is maintained during filtration. Preferably, the cyclodextrin is added to the ADC mixture before or at the beginning of the filtration process, although it is also contemplated that the cyclodextrin is first added to the ADC mixture after the filtration process has begun, but preferably before a significant portion of the impurities have been removed. The addition of cyclodextrin can be accomplished, for example, using a diafiltration buffer. The phrase maintained throughout filtration means that the specified concentration of cyclodextrin is maintained throughout the filtration process. In some aspects, such as those in which the cyclodextrin is added after the filtration process has begun (preferably before a significant portion of the impurities have been removed), the cyclodextrin will not be present at the start of the filtration process, but will be maintained at a specified concentration after the cyclodextrin is added.

As noted above, the diafiltration buffer may contain at least about 1%, at least about 2%, at least about 3% w/v. cyclodextrin or at least about 4% w/v. cyclodextrin. In some aspects, the diafiltration buffer will contain from about 1% w/v. cyclodextrin to approximately 10% w/v. cyclodextrin, from about 2% w/v. cyclodextrin to approximately 10% w/v. cyclodextrin or from about 3% w/v. cyclodextrin to approximately 10% w/v. cyclodextrin, or from about 3% w/v. cyclodextrin to approximately 6% w/v. cyclodextrin, or from about 3% w/v. cyclodextrin to approximately 4% w/v. cyclodextrin. Typically, the diafiltration buffer will additionally contain a buffering agent. Suitable buffers include, but are not limited to, acetate buffers, phosphate buffers, succinate buffers, histidine buffers, HEPES and MOPS. In one aspect, the buffering agent will be tris-(2-amino-2-(hydroxymethyl)-1,3-propanediol). Typical concentrations of boo

- 17 043811 fers in the case of a buffer used for diafiltration include concentrations from about 5 to about 100 mM, about 50 mM, about 10 mM to about 40 mM, or about 20 mM. In some embodiments, Tris buffer is used in an amount of about 5 mM to about 50 mM, about 10 mM to about 40 mM, or about 20 mM. A suitable pH for the diafiltration buffer is typically in the range of about 6.0 to about 8.0, but can be higher or lower. In some aspects, the pH ranges from about 6.5 to about 7.5, or from about 7 to 7.4. In some aspects, the pH is about 7.3 or about 7.2.

In some aspects, when beta-cyclodextrin is used, the concentration is maintained at at least about 2%, at least about 3%, or at least about 4% cyclodextrin during filtration. In some aspects, when gammacyclodextrin is used, the concentration is maintained at at least about 1% during filtration.

The present invention relates to methods for removing benzodiazepine drug impurities from an ADC mixture by tangential flow filtration. In some aspects, a tangential flow filtration apparatus comprises a filtration vessel having an inlet, a filtrate outlet (also called a permeate outlet), a retentate outlet, a semi-permeable ultrafiltration membrane, and a pump. The filtration membrane separates the filtration vessel into an inlet compartment and an outlet compartment so that all filtrate must enter the inlet and pass through the ultrafiltration membrane before flowing out of the filtration vessel through the filtrate outlet. The filtration membrane that is used can be made of any material or combination of materials suitable for use with ADC mixtures. Typical membranes include ultrafiltration membranes made from regenerated cellulose or polyethersulfone (eg, Millipore ULTRACEL® or BIOMAX® Membranes). The pore size of the membrane refers to the average pore size of the filter. For use in the present invention, the membrane typically has a pore size of less than 100 kDa, most often less than about 50 kDa or about 30 kDa. Thus, in some aspects, the filtration membrane will be a regenerated cellulose or polyethersulfone ultrafiltration membrane having a pore size of less than 100 kDa, most often less than about 50 kDa or about 30 kDa.

In some aspects, the tangential flow filtration system will further comprise a sample vessel for storing the reaction mixture used for conjugation and a buffer vessel fluidly coupled to the sample vessel. In some aspects, the buffer is designed to replace the volume of filtrate at the same rate as the flow rate of the filtrate so that the volume in the system remains constant. The mixture to be purified will be pressurized from the sample container, flow along the filtration membrane, and returned to the sample container. In some aspects, the line will contain a valve that can be used to reduce the flow of mixture into the sample container. Partially closing the valve increases the fraction of flow passing through the membrane to waste (filtrate), while the bulk of the flow is returned to the sample container (retentate). The filtrate entering the drain contains a buffer and those solute molecules that are small enough to pass through the pores of the membrane. Diafiltration buffer is added to the system to replace the loss of buffer as filtrate so that the total volume in the system remains constant. Since the diafiltration buffer does not contain any solute to be removed, the solute concentration gradually decreases.

As described, in some aspects of the present invention, constant volume diafiltration is a tangential flow filtration mode that is used to purify an ADC mixture and eliminate benzodiazepine drug related impurities. In constant volume diafiltration, the diafiltration volume refers to the total volume of solution at any given time, specifically the initial total volume of the mixture. In some aspects, samples will be recovered after one or more volumes of diafiltration to measure the concentration of impurities related to the benzodiazepine drug. Diafiltration is usually carried out using a number of diafiltration volumes, which are determined in a preliminary experiment to carry out purification to the desired target level. Typically, diafiltration is completed after the concentration of impurities related to the benzodiazepine drug reaches a predetermined level. In some aspects, the target level of purification is about 1 μM or less, preferably about 0.5 μM or less, 0.2 μM, or even 0.1 μM or less.

In some aspects, the retentate will contain from about 3 to about 10 mg/ml benzodiazepine ADC, from about 10 to about 30 mM Tris, from about 3% to about 10% w/v. cyclodextrin at a pH value of approximately 6.5 to approx.

- 18 043811 but 7.5. In some aspects, the retentate will contain about 3 mg/ml benzodiazepine ADC, about 20 mM Tris, about 3% cyclodextrin, or about 4% cyclodextrin, at a pH of about 7.3.

Cyclodextrin

Cyclodextrins are non-reducing, glucose-containing, cyclic oligosaccharides derived from starch. There are three main types of cyclodextrins, consisting of 6, 7 or 8 glucose residues (α-, β- and γ-cyclodextrins, respectively) linked by α-1,4 glycosidic bonds. Cyclodextrins can function as molecular containers by trapping a guest molecule into their internal cavity to form inclusion compounds, α-Cyclodextrins have a small cavity, whereas β- and γ-cyclodextrins have larger cavities.

Suitable cyclodextrins for use in the present invention include alpha, beta and gamma cyclodextrins, although beta and gamma cyclodextrins are preferred given that their internal cavities are larger. Chemical modifications of cyclodextrins, in particular β-cyclodextrins, were carried out to increase the solubility of the original cyclodextrin. Hydroxyethyl-β-cyclodextrin, hydroxypropyl-β-cyclodextrin (eg 2-hydroxypropyl-β-cyclodextrin), methylated β-cyclodextrin, glycosyl-β-cyclodextrin, and β-cyclodextrin sulfobutyl ether are examples of cyclodextrins that have been chemically modified to increase solubility. In some aspects of the present invention, β-cyclodextrin is used, including chemically modified β-cyclodextrins. In some aspects, the chemically modified β-cyclodextrin is hydroxypropyl-β-cyclodextrin or β-cyclodextrin sulfobutyl ether. In some aspects, the cyclodextrin is a gamma cyclodextrin, including chemically modified gamma cyclodextrins. In some aspects, the cyclodextrin will be a hydroxypropyl-cyclodextrin (eg, HP4.3-β-cyclodextrin, HP5.5-β-cyclodextrin, HP7.6-β-cyclodextrin, and HP4.5-γ-cyclodextrin). In other aspects, the cyclodextrin will be a sulfobutyl ester of β-cyclodextrins (e.g., SBE6.6-β-cyclodextrin, SBE6.7-β-cyclodextrin, SBE6.8-β-cyclodextrin, SBE4.1-β-cyclodextrin, and SBE4.6Et3.5 -β-cyclodextrin). In other aspects, the cyclodextrin will be a sulfobutyl ether of γ-cyclodextrins (eg, SBE4.3-γ-cyclodextrin, SBE4.6-Y-cyclodextrin, SBE5.2-γ-cyclodextrin, and SBE5.6Et6.3-γ-cyclodextrin). As used herein, a chemically modified beta-cyclodextrin is a beta-cyclodextrin that has been chemically modified such that it has at least increased solubility compared to the parent cyclodextrin (ie, unmodified cyclodextrin).

Compositions

The present invention relates to mixtures containing a benzodiazepine ADC and impurities related to the benzodiazepine drug (including conjugation reaction mixtures and quenched conjugation reaction mixtures), TFF retentate compositions, and pharmaceutical compositions containing Benzodiazepine ADC and cyclodextrin. TFF retentate formulation refers to a mixture that has been purified with TFF but not included in the final dosage form.

Cyclodextrins are typically added to ADC mixtures in concentrations of at least about 1% w/v, at least about 2% w/v. cyclodextrin, at least about 3% w/v. cyclodextrin or at least about 4% w/v. cyclodextrin (e.g., 1, 2, 3, or 4% w/v) In some aspects, the present invention provides a reaction mixture used for conjugation containing at least about 1%, at least about 2% w/v . cyclodextrin, at least about 3% w/v. cyclodextrin, at least about 3% w/v. cyclodextrin, to a quenched reaction mixture used for conjugation containing at least about 1%, at least about 2% wt./about. cyclodextrin or at least about 3% w/v. cyclodextrin, and/or to a TFF retentate composition containing at least about 1%, at least about 2% w/v. cyclodextrin or at least about 3% w/v. cyclodextrin. In some aspects, the cyclodextrins are provided in ADC mixtures (including conjugation reaction mixtures and quenched conjugation reaction mixtures) and TFF retentate formulations at a concentration of between about 2% and about 3% w/v. cyclodextrin to approximately 30% w/v. cyclodextrin, from about 2% or about 3% w/v. cyclodextrin to approximately 15% w/v. cyclodextrin, from about 2% or about 3% w/v. cyclodextrin to approximately 10% w/v. cyclodextrin. In some aspects, where the cyclodextrin is a betacyclodextrin, at least about 2% or about 3% cyclodextrin is present in the ADC mixture or TFF retentate composition. In some aspects, where the cyclodextrin is a gamma-cyclodextrin, at least about 1% or about 2% is present

- 19 043811 cyclodextrin or approximately 3% cyclodextrin in an ADC mixture or TFF retentate.

Cyclodextrins are typically present in pharmaceutical compositions at concentrations of approximately 3% w/v. cyclodextrin or more. In some aspects, the cyclodextrins are provided in pharmaceutical compositions at a concentration of approximately 5% w/v. or more, or approximately 6% w/v. or more. In some aspects, the cyclodextrins are provided in pharmaceutical compositions at concentrations ranging from 3% w/v. cyclodextrin, approximately 5% w/v. cyclodextrin or approximately 6% w/v. cyclodextrin to approximately 30% w/v. cyclodextrin, or from about 3% w/v. cyclodextrin, approximately 5% w/v. cyclodextrin or approximately 6% w/v. cyclodextrin to approximately 15% w/v. cyclodextrin, or from about 3% w/v. cyclodextrin, approximately 5% w/v. cyclodextrin or approximately 6% w/v. cyclodextrin to approximately 10% w/v. cyclodextrin, or from about 3% w/v. cyclodextrin, approximately 5% w/v. cyclodextrin or approximately 6% w/v. cyclodextrin to approximately 8% w/v. cyclodextrin. In some embodiments, the cyclodextrin is present at a concentration of approximately 3% w/v. In other embodiments, the cyclodextrin is present at a concentration of approximately 4% w/v. In other embodiments, the cyclodextrin is present at a concentration of about 5% w/v, about 6% w/v, about 7% w/v. or approximately 8% w/v. In some aspects, the pharmaceutical compositions described herein contain 1 μM or less, 0.5 μM or less, 0.1 μM or less, or 0.05 μM or less of benzodiazepine drug related impurities.

In some aspects, the present invention provides benzodiazepine ADC pharmaceutical compositions and TFF retentate compositions containing 1 μM or less, 0.5 μM or less, 0.1 μM or less, or 0.05 μM or less benzodiazepine drug impurities. Cyclodextrins may be present in such benzodiazepine ADC compositions at a concentration of about 1%, about 2%, or about 3% w/v. cyclodextrin to approximately 30% w/v. cyclodextrin, from about 1%, about 2% or about 3% w/v. cyclodextrin to approximately 15% w/v. cyclodextrin, from about 1%, about 2% or about 3% w/v. cyclodextrin to approximately 10% w/v. cyclodextrin. In some aspects, cyclodextrins are provided in benzodiazepine ADC pharmaceutical compositions and TFF retentate formulations containing 1 μM or less, 0.5 μM or less, 0.1 μM or less, or 0.05 μM or less of benzodiazepine drug-related impurities, at a concentration of about 5% or more, or about 6% or more. In some aspects, the cyclodextrins are present in such compositions at a concentration of about 1%, about 2%, about 3% w/v. cyclodextrin, approximately 5% w/v. cyclodextrin or approximately 6% w/v. cyclodextrin to approximately 30% w/v. cyclodextrin, or from about 3% w/v. cyclodextrin, approximately 5% w/v. cyclodextrin or approximately 6% w/v. cyclodextrin to approximately 15% w/v. cyclodextrin, or from about 3% w/v. cyclodextrin, approximately 5% w/v. cyclodextrin or approximately 6% w/v. cyclodextrin to approximately 10% w/v. cyclodextrin, or from about 3% w/v. cyclodextrin, approximately 5% w/v. cyclodextrin or approximately 6% w/v. cyclodextrin to approximately 8% w/v. cyclodextrin. In some embodiments, the cyclodextrin is present at a concentration of approximately 1% w/v. In some embodiments, the cyclodextrin is present at a concentration of approximately 2% w/v. In some embodiments, the cyclodextrin is present at a concentration of approximately 3% w/v. In other embodiments, the cyclodextrin is present at a concentration of approximately 4% w/v. In other embodiments, the cyclodextrin is present at a concentration of about 5% w/v, about 6% w/v, about 7% w/v. or approximately 8% w/v.

Pharmaceutical compositions

The present invention relates to pharmaceutical compositions containing a benzodiazepine ADC and a cyclodextrin. The inventors have found that cyclodextrin-containing compositions containing approximately 2% wt./vol. cyclodextrin or more, demonstrate a decrease in the rate and degree of aggregation, compared to compositions containing 0.5% wt./vol. cyclodextrin or less in the composition, and demonstrate a decrease in the growth of acidic compounds, compared to compositions not containing cyclodextrin. The inventors have also found that compositions containing approximately 6% or more wt./vol. cyclodextrin, demonstrate reduced chemical degradation of the benzodiazepine drug linker molecule compared to compositions containing 2% or less w/v. cyclodextrin in the composition. Thus, the present invention is based in particular on the discovery that benzodiazepine ADCs containing approximately 6% w/v. cyclodextrin exhibit increased stability compared to formulations containing cyclodextrin in smaller quantities or not containing cyclodextrin at all. Evidence of increased stability may include one or more of the following: (i) decreased speed

- 20 043811 growth and degree of aggregation, (ii) reduction in the growth of acidic compounds and (iii) reduction in chemical degradation of the drug. In certain aspects, the present invention relates to stabilized liquid or lyophilized benzodiazepine ADC compositions intended for therapeutic use. In particular, compositions are provided that are stabilized such that the therapeutic benzodiazepine ADC is stable over an extended period of time and can be administered via a variety of routes of administration. Such compositions are particularly effective, for example, as a benzodiazepine ADC for use in the treatment of a disease or disorder (eg, treatment of cancer) by exerting a cytotoxic or cytostatic effect on target cells expressing an antigen recognized by the antibody component of the ADC.

In one aspect, the present invention provides a benzodiazepine ADC composition comprising a benzodiazepine ADC, a cyclodextrin, and optionally at least one buffering agent, wherein the buffering agent is present in an amount sufficient to maintain a physiologically appropriate pH value. The cyclodextrin can be any of the cyclodextrins described herein, but is preferably a beta or gamma cyclodextrin, and more preferably a beta cyclodextrin that has been chemically modified to improve its solubility compared to the parent molecule. In some aspects, the cyclodextrin is a gamma-cyclodextrin. In some aspects, gamma cyclodextrin is chemically modified to improve its solubility compared to the parent molecule. Particularly preferred cyclodextrins are hydroxypropyl beta-cyclodextrin or betacyclodextrin sulfobutyl ether. Cyclodextrin may be formulated at any of the concentrations described herein. In some aspects, the cyclodextrin will be a gammacyclodextrin (unmodified or chemically modified molecule) or a chemically modified beta-cyclodextrin (e.g., methylated beta-cyclodextrin, glucosyl-betacyclodextrin, hydroxypropyl-beta-cyclodextrin, or sulfobutyl-beta-cyclodextrin) present in the composition at concentrations approximately 3% w/v, approximately 4% w/v, approximately 5% w/v. or approximately 6% w/v. or more. In some aspects, the cyclodextrin will be a gamma-cyclodextrin (unmodified or chemically modified molecule) or a chemically modified beta-cyclodextrin (e.g., methylated beta-cyclodextrin, glucosyl-beta-cyclodextrin, hydroxypropyl-beta-cyclodextrin, or sulfobutyl-beta-cyclodextrin) present in the composition in concentrations from approximately 3% w/v. cyclodextrin, approximately 5% w/v. cyclodextrin or approximately 6% w/v. cyclodextrin to approximately 30% w/v. cyclodextrin, or from about 3% w/v. cyclodextrin, approximately 5% w/v. cyclodextrin or approximately 6% w/v. cyclodextrin to approximately 15% w/v. cyclodextrin, or from about 3% w/v. cyclodextrin, approximately 5% w/v. cyclodextrin or from about 6% w/v. cyclodextrin to approximately 10% w/v. cyclodextrin, or from about 3% w/v. cyclodextrin, approximately 5% w/v. cyclodextrin or approximately 6% w/v. cyclodextrin to approximately 8% w/v. cyclodextrin.

As noted above, the aqueous benzodiazepine ADC composition of the present invention optionally contains a buffering agent to maintain a physiologically appropriate pH in the aqueous solution. Suitable pH values for the stabilized aqueous composition as described herein include, for example, a pH value in the range of about 6.0 to about 8.0 or about 6.5 to about 7.5. In certain embodiments, it may be desirable to include in the composition an ADC with a pH value of from about 6.8 to about 7.5; most commonly from about 7.0 to about 7.5, from about 7.1 to about 7.5, from about 7.2 to about 7.5, or from about 7.3 to about 7.5. In a specific embodiment, the pH value is approximately 7.3. In a specific embodiment, the pH value is approximately 7.2. Suitable buffers include tris-(2-amino-2-(hydroxymethyl)-1,3-propanediol) and citrate, as well as other physiologically acceptable buffers that are effective in the specified pH range and ensure that the solution reaches approximately physiological pH during preparation of the dosage form, including after dissolving the lyophilized composition to form an aqueous solution, as described herein. Examples of such buffers include phosphate, acetate, succinate and histidine buffers. Typical buffer concentrations for compositions of the present invention range from about 5 mM to about 100 mM, about 50 mM, about 10 mM to about 40 mM, or about 20 mM. In some embodiments, Tris is included in an amount of about 5 mM to about 50 mM, about 10 mM to about 40 mM, or about 20 mM. Other suitable concentrations of buffers for the compositions of the present invention can be readily determined by one skilled in the art.

In specific embodiments, compositions, such as compositions suitable for lyo

- 21 043811 filizations reconstituted from lyophilized form, or lyophilized compositions intended to be reconstituted into an aqueous composition as described herein, may contain one or more stabilizing agents to protect the antibody-drug conjugate. The stabilizing agent is also referred to in the description as a lyoprotector. Typically, a suitable stabilizing agent is a sugar or an amino acid. Typical stabilizing agents include sucrose, trehalose, mannitol, sorbitol, dextrose, maltose, dextran, arginine, glycine and histidine. Typically, the amount of stabilizing agent (eg, lyoprotectant) in the composition is such that, after reconstitution, the resulting composition is substantially isotonic. In addition, the amount of lyoprotectant should not be too low such that an unacceptable amount of ADC degradation/aggregation products occurs after lyophilization. If the lyoprotectant is a sugar (eg, sucrose or trehalose), typical concentrations of lyoprotectant in the aqueous composition may range from about 1% to about 10% (w/v), most often from about 2 to about 8% (w/v). vol.) and even more often from about 4 to about 8% (wt./vol.). In a specific embodiment, the lyoprotectant is sucrose at a concentration of approximately 6% (w/v).

In some aspects, arginine is present in the composition in an amount of 0.5 M or less. In some aspects, the arginine is present in the composition in an amount of 0.2 M or 0.1 M or less. In some aspects, arginine is absent from the composition. In some aspects, amino acids are not present in the composition.

Additional excipients for use in the compositions include, for example, salts (eg sodium chloride), surfactants (eg polysorbate 80, polysorbate 20, poloxamers), antioxidants (eg ascorbic acid, methionine, malic acid) and other excipients, such as polyethylene glycols, propylene glycols, carboxymethylcellulose and vinylpyrrolidone.

The concentration of ADC in the pharmaceutical composition of the present invention typically ranges from about 0.1 or 0.5 mg/ml to about 50 mg/ml or from about 1 to about 30 mg/ml. Most often, the ADC is presented at a concentration of from about 1 mg/ml to about 5 mg/ml, to about 10 mg/ml, or to about 15 mg/ml; or at a concentration of from about 2 mg/ml to about 5 mg/ml or up to about 10 mg/ml. In some aspects, the ADC is present at a concentration of from about 0.6 to about 3 mg/ml. In specific embodiments, the ADC is present at a concentration of about 1 mg/mL, about 2 mg/mL, about 3 mg/mL, about 4 mg/mL, about 5 mg/mL, about 6 mg/mL, about 7 mg/mL, about 8 mg/ml, about 9 or about 10 mg/ml.

In some embodiments, the antibody component of the ADC in ADC mixtures, conjugation reaction mixtures, quenched conjugation reaction mixtures, TFF retentate formulations, and in pharmaceutical compositions containing a benzodiazepine and cyclodextrin ADC described herein is a monoclonal antibody, selected from the following: a tetramer consisting of two identical pairs of immunoglobulin chains, each pair having one light chain and one heavy chain; Fv fragment of antibody; Fab fragment of antibody; Fab'(2)antibody fragment, Fd antibody fragment, single chain antibody (eg, scFv or scFvFc fusion); or a single domain antibody fragment (Dab). In a specific embodiment, the antibody comprises first and second polypeptide chains, wherein the first polypeptide chain comprises a light chain variable (VL) domain fused at the carboxyl terminus to a light chain constant region, and where the second polypeptide chain comprises a heavy chain variable (VH) domain fused at the carboxyl end with the heavy chain constant region. In such embodiments, the monoclonal antibody is typically a tetramer consisting of two identical pairs of immunoglobulin chains, each pair having one light chain and one heavy chain. The heavy chain constant region may be a natural constant region or a mutant form of a natural human immunoglobulin constant region having reduced binding to the Fcy receptor compared to a natural human immunoglobulin constant region. In some embodiments, the antibody is an isotype selected from IgG1, IgG2, IgG3, and IgG4. In a specific embodiment, the heavy chain constant region is an IgG1 isotype constant region.

Various methods for producing antibodies, including monoclonal antibodies, are well known in the art and are not described in detail herein. Antibodies for use in the present invention may be intact antibodies or antigen binding fragments thereof. Preferred antibodies are human or humanized antibodies, in particular human or humanized monoclonal antibodies. In some aspects, the antibody will specifically bind to a cancer cell antigen that is expressed on the surface of the cancer cell. In other aspects, the antibody will bind to activated lymphocytes that are caused by an autoimmune disease. In some aspects, the antibody will specifically bind to CD19, CD20, CD30, CD33, CD70, Glypican-3, Liv-1

- 22 043811 or with Lewis Y antigen.

In specific embodiments, the antibody is an aHTu-CD33 antibody that specifically binds to the extracellular domain of human CD33. The approximate human CD33 sequence has been assigned accession number P20138 in the Swiss Prot database. In certain embodiments, the anti-CD33 antibody comprises the complementarity determining regions of the light and/or heavy chain variable domains of a mouse anti-CD33 monoclonal antibody called 2H12 (amino acid sequences of the VL and VH domains are shown in SEQ ID NO: 1 and 2, respectively) . Accordingly, in some embodiments, the anti-CD33 antibody comprises a light chain variable domain (VL) comprising a CDR-L1 amino acid sequence as shown in SEQ ID NO: 5, a CDR-L2 amino acid sequence as shown in SEQ ID NO: 6, and the amino acid sequence of CDR-L3, which is shown in SEQ ID NO: 7; and/or a heavy chain variable domain (VH) comprising the amino acid sequence CDR-H1 which is shown in SEQ ID NO: 8, the amino acid sequence CDR-H2 which is shown in SEQ ID NO: 9, and the amino acid sequence CDR-H3 which shown in SEQ ID NO: 10. In some embodiments, the anti-CD33 antibody is a humanized antibody. For example, in specific embodiments of an anti-CD33 antibody containing VL and/or VH domains as described above, the VL domain is a humanized VL domain derived from the VL domain of the murine 2H12 antibody having the amino acid sequence as shown in SEQ ID NO: 1, and/or the VH domain is a humanized VH domain derived from the VH domain of the murine antibody 2H12 having the amino acid sequence as shown in SEQ ID NO: 2. Particularly suitable VL and VH domains have amino acid sequences that are at least 90%, at least 91%, at least 92%, at least 93%, at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, or 100% identical to the sequences of SEQ ID NO: 3 and SEQ ID NO: 4, respectively. In certain embodiments, an anti-CD33 antibody includes first and second polypeptide chains, wherein the first polypeptide chain comprises a VL domain fused at the carboxyl terminus to a light chain constant region (e.g., to a light chain constant region having the amino acid sequence SEQ ID NO : 21), and the second polypeptide chain contains a VH domain fused at the carboxyl terminus to a heavy chain constant region (eg, a heavy chain constant region having the amino acid sequence of SEQ ID NO: 22 or SEQ ID NO: 23). In such embodiments, the anti-CD33 antibody is typically a tetramer consisting of two identical pairs of immunoglobulin chains, each pair having one light chain and one heavy chain. Indicative anti-CD33 antibodies suitable for use in the present invention are also described in US Provisional Application No. 61/649,110, filed May 18, 2012, the disclosure of which is incorporated herein by reference in its entirety for all purposes.

In other embodiments, the antibody is an anti-CD70 antibody that specifically binds to the extracellular domain of human CD70. The approximate human CD70 sequence has been assigned accession number P32970.2 in the Swiss Prot database. In certain embodiments, the anti-CD70 antibody comprises the complementarity determining regions of the light and/or heavy chain variable domains of a murine anti-CD70 monoclonal antibody called 1F6 (amino acid sequences of the VL and VH domains are shown in SEQ ID NOs: 11 and 12, respectively) . In some such embodiments, the anti-CD70 antibody comprises a light chain variable domain (VL) comprising a CDR-L1 amino acid sequence as shown in SEQ ID NO: 15, a CDR-L2 amino acid sequence as shown in SEQ ID NO: 16, and an the CDR-L3 sequence, which is shown in SEQ ID NO: 17; and a heavy chain variable domain (VH) comprising the amino acid sequence of CDR-H1, which is shown in SEQ ID NO: 18, the amino acid sequence of CDR-H2, which is shown in SEQ ID NO: 19, and the amino acid sequence of CDR-H3, which is shown in SEQ ID NO: 20. In some embodiments, the anti-CD70 antibody is a humanized antibody. For example, in specific embodiments of an anti-CD70 antibody containing VL and VH domains as described above, the VL domain is a humanized VL domain derived from the VL domain of the mouse antibody 1F6 having the amino acid sequence that is shown in SEQ ID NO: 11, and/or the VH domain is a humanized VH domain derived from the VH domain of the mouse antibody 1F6 having the amino acid sequence that is shown in SEQ ID NO: 12. Particularly suitable VL and VH domains have amino acid sequences that at least 90%, at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least at least 97%, at least 98%, at least 99%, or 100% identical to the sequences of SEQ ID NO: 13 and SEQ ID NO: 14, respectively. In certain embodiments, an anti-CD70 antibody includes first and second polypeptide chains, wherein the first polypeptide chain comprises a VL domain fused at the carboxyl terminus to a light chain constant region (e.g., to a light chain constant region having the amino acid sequence SEQ ID NO : 21), and the second polypeptide chain contains VH

- 23 043811 domain fused at the carboxyl terminus to a heavy chain constant region (eg, to a heavy chain constant region having the amino acid sequence SEQ ID NO: 22). In such embodiments, the anti-CD70 antibody is typically a tetramer consisting of two identical pairs of immunoglobulin chains, each pair having one light chain and one heavy chain. Exemplary anti-CD70 antibodies suitable for use in the present invention are also described in US Pat. No. 8,067,546, the disclosure of which is incorporated by reference in its entirety for all purposes.

Representative sequences of variable domains and CDRs of anti-CD33 and anti-CD70, as well as representative sequences of constant regions of heavy and light chains of immunoglobulins, are presented in table. 1 below.

Table 1 SEQ Description Amino acid sequence ID NO: VL antibodies mice 2 × 12 DIKMTQSPSSMYASLGERVIINCKASQDINSYLSWFQQKPGK SPKTLIYRANRLVDGVPSRFSGSGSGQDYSLTISSLEYEDMG IYYCLQYDEFPLTFGAGTKLELK 1 VH antibodies mice 2 × 12 QVQLQQSGPELVRPGTFVKISCKASGYTFTNYDINWVNQRPG QGLEWIGWIYPGDGSTKYNEKFKAKATLTADKSSSTAYLQLN NLTSENSAVYFCASGYEDAMDYWGQGTSVTVSS 2 Humaniziro DIQMTQSPSSLSSASVGDRVTINCKASQDINSYLSWFQQKPGK bathroom VL APKTLIYRANRLVDGVPSRFSGSGSGQDYTLTISSLQPEDFA 3 2N12 TYYCLQYDEFPLTFGGGTKVEIK Humaniziro QVQLVQS GAEVKKPGASVKVS OKAS GYT FTNYDINWVRQAPG bathroom VH QGLEWIGWIYPGDGS TKYNEKFKAKATLTADT S T S TAYMELR 4 2N12 SLRSDDTAVYYCASGYEDAMDYWGQGTTVTVSS 2N12 CDR-L1 KASQDINSYLS 5 2N12 CDR-L2 RANRLLVD 6 2N12 CDR-L3 LQYDEFPLT 7 2N12 CDR-H1 NYDIN 8 2N12 CDR-H2 WIYPGDGSTKYNEKFKA 9 2N12 CDR-H3 GYEDAMDY 10 VL antibodies mice 1G 6 DIVLTQSPASLAVSLGQRATISCRASKSVSTSGYSFMHWYQQ KPGQPPKLLIYLASNLESGVPARFSGSGSGTDFTLNIHPVEE EDAATYYCQHSREVPWTFGGGTKLEIKR eleven VH antibodies mice 1F6 QIQLVQSGPEVKKPGETVKISCKASGYTFTNYGMNWVKQAPG KGLKWMGWINTYTGEPTYADAFKGRFAFSLETSASTAYLQIN NLKNEDTATYFCARDYGDYGMDYWGQGTSVTVSS 12 Humaniziro DIVMTQSPDSLAVSLGERATINCRASKSVSTSGYSFMHWYQQ bathroom VL KPGQPPKLLIYLASNLESGVPDRFSGSGSGTDFTLTISSLQA 13 1F6 EDVAVYYCQHSREVPWTFGQGTKVEIKR Humaniziro QVQLVQSGAEVKKPGASVKVSCKASGYTFTNYGMNWVRQAPG bathroom VH QGLKWMGWINTYTGEPTYADAFKGRVTMTRDTSISTAYMELS 14 1F6 RLRS DDTAVYYCARDYGDYGMDYWGQGT TVTVS S 1F6 CDR-L1 RASKSVSTSGYSFMH 15 1F6 CDR-L2 LASNLES 16 1F6 CDR-L3 QHSREVPWT 17

- 24 043811

1F6 CDR-H1 NYGMN 18 1F6 CDR-H2 WINTYTGEPTYADAFKG 19 1F6 CDR-H3 DYGDYGMDY 20 Human light chain constant region TVAAPSVFIFPPSDEQLKSGTASVVCLLNNFYPREAKVQWKV DNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVY ACEVTHQGLSSPVTKSFNRGEC 21 Human heavy chain constant region (without Terminal K) ASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWN SGALTSGVHTFPAVLQSSGLYSLSSWTVPSSSLGTQTYICN VNHKPSNTKVDKKVEPKSCDKTHTCPPCPAPELLGGPSVFLF PPKPKDTLMISRTPEVTCVVVDVSHEDPEVKFNWYVDGVEVH NAKTKPREEQYNSTYRWSVL TVLHQDWLNGKEYKCKVSNKA LPAPIEKTISKAKGQPREPQVYTLPPSRDELTKNQVSLTCLV KGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSKLT VDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPG 22 Human heavy chain constant region, S239C (no C-terminal K) ASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWN SGALTSGVHTFPAVLQSSGLYSLSSWTVPSSSLGTQTYICN VNHKPSNTKVDKKVEPKSCDKTHTCPPCPAPELLGGPCVFLF PPKPKDTLMISRTPEVTCVVVDVSHEDPEVKFNWYVDGVEVH NAKTKPREEQYNSTYRVVSV LTVLHQDWLNGKEYKCKVSNKA LPAPIEKTISKAKGQPREPQVYTLPPSRDELTKNQVSLTCLV KGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSKLT VDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPG 23

As noted above, typical pharmaceutical compositions of the present invention exhibit (i) a decrease in the rate and extent of aggregation, compared to compositions containing 0.5% w/v. cyclodextrin or less in the composition, (ii) reducing the increase in the amount of acidic compounds, compared to compositions not containing cyclodextrin, and/or (iii) reducing the chemical degradation of the drug-linker compound (for example, a benzodiazepine drug), compared to compositions , containing 2% or less w/v. cyclodextrin in the composition. Thus, the present invention relates to methods for preparing a composition that has at least one of the following: (i) increased resistance to chemical degradation of a benzodiazepine drug, compared to compositions containing 2% or less w/v. cyclodextrin in the composition, (ii) increased resistance to aggregation, compared to compositions containing 0.5% wt./vol. cyclodextrin or less in the composition, and (iii) increased resistance to the growth of acidic compounds, compared to compositions not containing cyclodextrin. Thus, in some aspects, the pharmaceutical compositions of the present invention containing about 5% or about 6% wt./vol. cyclodextrin or more, exhibit at least one of the following: (i) increased resistance to chemical degradation of the benzodiazepine drug, compared to compositions containing 2% cyclodextrin or less, (ii) increased resistance to aggregation, compared to compositions containing 0.5% cyclodextrin or less, and (iii) increased resistance to increased acidic compounds compared to compositions containing 0.5% cyclodextrin or less. Thus, in some aspects, the pharmaceutical compositions of the present invention containing about 5% or about 6% wt./vol. cyclodextrin or more exhibit (i) increased resistance to chemical degradation of the benzodiazepine drug compared to compositions containing 2% cyclodextrin or less, and (ii) increased resistance to aggregation compared to compositions containing 0.5% cyclodextrin or more. less. In other aspects, pharmaceutical compositions of the present invention containing about 5% or about 6% wt./vol. cyclodextrin or more, exhibit (i) increased resistance to chemical degradation of the benzodiazepine drug, compared to compositions containing 2% or less cyclodextrin, (ii) increased resistance to aggregation, compared to compositions containing 0.5% or less cyclodextrin , and (iii) increased resistance to increased acidic compounds compared to compositions containing 0.5% or less cyclodextrin.

In any of the mixtures or compositions described herein (including pharmaceutical compositions), benzodiazepine ADCs may contain a humanized 2H12 antibody or humani

- 25 043811 zited antibody 1F6 conjugated to PBD dimer 1, PBD dimer 2, PBD dimer 3, PBD dimer 4, PBD dimer 5, PBD dimer 6, PBD dimer 7, PBD dimer 8, PBD dimer 9, PBD dimer 10, dimer PBD 11, PBD 12 dimer, PBD 13 dimer, PBD 14 dimer, PBD 15 dimer, PBD 16 dimer, or with PBD 17 dimer, as described herein. In some aspects, conjugation of the PBD dimer to the antibody will be accomplished through a cysteine residue incorporated into the antibody. In some aspects, a cysteine residue is inserted into the antibody at position 239 (human IgG1), which is defined by the EU index (Kabat, Sequences of Proteins of Immunological Interest (National Institutes of Health, Bethesda, MD, 1987 and 1991)). In some aspects, the composition contains an average of 2 drug-linker molecules per antibody.

In some embodiments, the pharmaceutical composition of the present invention provides a concentrated drug ADC (eg, anti-CD33 or anti-CD70 ADC), often used as a bulk drug product. Additionally, in certain embodiments, the pharmaceutical composition of the present invention is resistant to freezing, lyophilization, and/or dissolution.

In some aspects, the pharmaceutical compositions described herein are stored at temperatures from about -80°C to about 8°C. In general, the pharmaceutical composition is stable and retains biological activity within the specified ranges. In certain aspects, the compositions described herein have increased stability compared to compositions not containing cyclodextrin when exposed to stressful conditions (for example, storage at 25 or 40°C for long periods of time, such as 7 and 14 days).

The pharmaceutical compositions of the present invention are suitable for delivery via various routes of administration. In certain embodiments, the pharmaceutical composition is administered parenterally, for example, intravenously, intramuscularly, or subcutaneously. When administering an ADC for the treatment of cancer, the pharmaceutical composition can be delivered into the systemic circulation by intravenous or subcutaneous administration. In a specific embodiment, the pharmaceutical composition is formulated for intravenous administration. Intravenous administration can be accomplished, for example, by infusion over a period of time, such as 30-90 minutes, or by a single bolus injection. In some aspects, administration will be by IV bolus injection (ie, over 30-60 seconds) into a peripherally inserted central catheter.

Effective doses of the pharmaceutical compositions of the present invention vary depending on many factors, including routes of administration, target location, physiological condition of the patient, whether the patient is human or animal, whether other drugs are administered, and whether the treatment is prophylactic or therapeutic. Typically the patient is human, but non-human mammals can also be treated. Treatment doses must be titrated to optimize safety and effectiveness.

Exemplary doses of the ADC compositions of the present invention include doses from about 1.0 μg/kg to about 5 mg/kg, from about 10 μg/kg to about 3 mg/kg, from about 10 μg/kg to about 2 mg/kg, from about 1.0 μg/kg to 1.0 mg/kg, or from about 1.0 to 500.0 μg/kg body weight of individuals.

The frequency of administration depends, among other factors, on the circulating half-life of the antibody-drug conjugate, the condition of the patient, and the route of administration. The frequency of administration may be daily, weekly, monthly, quarterly, or at irregular intervals in response to changes in the patient's condition or the progression of the disease (eg, cancer) being treated. The approximate frequency of intravenous administration is twice a week to once a quarter, during a continuous course of treatment, although more or less frequent dosing is also possible. Other exemplary intravenous dosing frequencies are once per week or once per month for a continuous course of treatment, although more or less frequent dosing is also possible. For subcutaneous administration, the approximate dosing frequency may range from daily to monthly, although more or less frequent dosing is also possible.

It is particularly preferred that the compositions of the invention are provided in unit dosage form for ease of administration and uniformity of dosage. The compositions of the invention may be presented in liquid or lyophilized form, for example, in capsules, glass vials, ampoules, multiple dose dosage packs or the like. The unit dosage form may contain any composition described herein. In some aspects, the ADC will be present as a lyophilized tablet or powder stored in a single-use orange glass vial designed to dissolve upon IV administration. Dissolution is carried out with a suitable solvent (for example, water for injection) to the desired concentration. Typically, the reconstituted drug product is diluted with sufficient solvent so that the reconstituted solution will have the same concentration of components as the composition before lyophilization. The reconstituted product can be further diluted depending on the dose to be administered to the patient. Subsequent breeding

- 26 043811 can be carried out, for example, with saline solution for injection. In some aspects of the invention, immediately after reconstitution or subsequent additional dilution, the ADC is administered IV (eg, by slow IV bolus) into the appropriate injection port of the central venous access device. The infusion system is usually flushed with saline.

In some aspects, the present invention provides a therapeutic product that includes a pharmaceutical unit dosage form containing a stabilized ADC composition of the present invention (e.g., a sealed container containing an ADC composition, either in liquid or lyophilized form, as described herein document). The therapeutic product may further include a directions for use. In some embodiments, the therapeutic product is provided as a kit, further including, for example, instructions for using the appropriate volume needed to achieve a therapeutic dose in a patient. The unit dosage form, container, or kit may be designed to provide sufficient amounts of the composition for multiple uses or for single use. In some embodiments, the kit further includes a diluent.

If different versions of the sequence refer to an account number with different dates, then the version referred to the account number for the valid filing date of the specified application is meant. The effective application filing date means the earlier of the existing application filing dates or the filing date of the priority application associated with the reference number, if applicable. Likewise, if different versions of a publication are published at different times, the version most recently published from the effective filing date of the application is meant, unless otherwise stated. Any feature, step, element, embodiment or aspect of the present invention may be used in combination with any others unless specifically stated otherwise. Although the present invention has been described in some detail by way of illustration and example for the purposes of clarity and understanding, it will be appreciated that certain changes and modifications may be made within the scope of the appended claims. All patent applications, websites, other publications, reference numbers, etc., mentioned above or below, are incorporated by reference in their entirety, for all purposes, to the same extent as if each individual subject matter had been specifically and separately identified for incorporation by reference.

Examples

PBD dimers 1-4 and their synthesis are described in WO 2011/130613. PBD dimers 5-10 and 16 can be synthesized using the methods described in WO 2011/130613 A1. Briefly, PBD dimers 9 and 16 can be prepared via the C3-linked bis-triflate of intermediate 8a described in WO 2011/130613 A1. The required C2-aryl groups in the form of boronic acids or pinacol boronates are introduced into sequential Suzuki coupling reactions, followed by reduction of SEM-dilactam to release the functional imino groups. PBD dimers 5-8 and 10 are prepared in a similar manner from the C5-linked bis-triflate of intermediate 8b described in WO 2011/130613 A1. PBD 11-15 dimers containing esters or carboxylic acids in C2 aryl groups can be prepared using the methods described in WO 2011/130613 A1 with minor modifications. PBD dimer 13 can be prepared from the C3-linked bis-triflate of intermediate 8a described in WO 2011/130613 A1. The bis-triflate is desymmetrized through a Suzuki coupling reaction with an appropriately functionalized boronic acid or pinacol borate to introduce a C2-aryl group bearing an amino function. The resulting monotriflate is then reduced with lithium triethylborohybrid to SEM-carbinol, which is then used in a second Suzuki coupling reaction to introduce a C2 aryl group containing the methyl ester. Finally, the SEM carbinols are converted to imines by stirring using silica gel for 3 days as described in WO 2011/130613 A1. PBD dimers 12 and 14 can be prepared in the same way, starting from the C5-linked bis-triflate 8b described in WO 2011/130613 A1. The conversion of PBD esters to the free carboxylic acid (11 and 15) can be accomplished by saponification. The preparation of cysteine mutant mAbs IgG1 is described generally in US 20100158909.

The PBD-linker drug compounds used in the following examples are presented below. Conjugation of drug-linker compounds with antibodies is carried out as described in WO 2011/130613 A1.

- 27 043811

The quenched drug-linker compound mentioned in the following examples has the following structure:

PBD-linker drug compounds were conjugated to h2H12ec or h1F6ec antibody via position 239 of the heavy chain.

(The designation ec after the antibody name indicates an antibody having an inserted cysteine at position 239.) Briefly, antibodies h2H12 and h1F6, having an inserted cysteine at position 239 (EU index numbering), were reduced, partially reoxidized (i.e., reoxidized interchain disulfides) and conjugated to the drug compound PBD linker according to the methods described in WO 2011/130613 to obtain ADC. The PBD linker drug was conjugated to the partially re-oxidized antibody via an inserted cysteine residue (an average of 2 drug-linker molecules per antibody). The term h2H12-1 means the humanized antibody 2H12ec conjugated to compound 1, while the term h1F6-1 means the humanized antibody 1F6ec conjugated to compound 1. The term h2H12-2 means the humanized antibody 2H12ec conjugated to compound 2, while the term h1F6-2 means the humanized antibody 1F6ec conjugated to compound 2. The term h2H12-3 means the humanized antibody 2H12ec conjugated to compound 3, while the term h1F6-3 means the humanized antibody 1F6ec conjugated to compound 3. Examples 4-7 were carried out using either antibody h2H12ec , or antibody h1F6ec conjugated to compound 1 through the introduced cysteine at position 239.

Example 1: Cyclodextrin-Based Formulations Reduced the Rate and Extent of Aggregation for All Six PBD ADCs

The compositions were prepared by exchanging the buffer in solutions of the starting substances with buffers of the test compositions using dialysis, a buffer exchange column, or a centrifugal filtration method. We used the compositions that are indicated in the comments to the figures. The concentration of arginine, if shown to be present, is 0.5 M. The final ADC concentration ranged from 1.3 to 3.1 mg/mL as measured by UV spectroscopy. Formulated ADCs were placed in pre-sterilized tubes and stored under the following ICH approved conditions: 2-8°C, 25°C/65% RH and/or 40°C/75% RH. %HMW was monitored using size exclusion chromatography. In fig. 1 composition contains h2H12-1 and storage is carried out at 25°C; in fig. 2 compositions contain h1F6-1 and storage is carried out at 25°C; in fig. 3 compositions contain h2H12-3 and storage is carried out at 40°C; in fig. 4 compositions contain h1F6-3 and storage is carried out at 40°C; in fig. 5 compositions contain h2H12-2 and storage is carried out at 40°C; in fig. 6 compositions contain h1F6-2 and storage is carried out at 40°C; in fig. 13 compositions contain h1F6-1 and storage is carried out at 25°C.

Example 2 HPBCD Formulations Reduced Acid Growth for Five of Six PBD ADCs

The compositions were prepared by exchanging the buffer in solutions of the starting substances with buffers of the test compositions using dialysis, a buffer exchange column, or the centrifugal filter method.

- 28 043811 vania. We used the compositions that are indicated in the comments to the figures. The concentration of arginine, if shown to be present, is 0.5 M. The final ADC concentration ranged from 1.3 to 3.1 mg/mL as measured by UV spectroscopy. The composition was placed in pre-sterilized tubes and stored under the following ICH approved conditions: 2-8°C, 25°C/65% RH and/or 40°C/75% RH. Samples were analyzed for the charge distribution of the compound molecules (acidic compounds, % basic compound, % basic compounds) using capillary isoelectric focusing with capillary detection. In fig. 7 compositions contain h2H12-1 and storage is carried out at 25°C; in fig. 8 compositions contain h1F6-1 and storage is carried out at 25°C; in fig. 9 compositions contain h2E12-3 and storage is carried out at 40°C; in fig. 10 compositions contain h1F6-3 and storage is carried out at 40°C; in fig. 11 compositions contain h2H12-2 and storage is carried out at 40°C; in fig. 12 compositions contain h1F6-2 and storage is carried out at 40°C.

Example 3: HPBCD-Based Compositions Reduce Chemical Degradation of the Drug-Linker Compound

The compositions were prepared by exchanging the buffer in solutions of the starting substances with buffers of the test compositions using dialysis, a buffer exchange column, or a centrifugal filtration method. In general, the compositions included an initial control (predominantly arginine), buffer, 3% cyclodextrin, and 6% cyclodextrin. The final ADC concentration ranged from 1.3 to 3.1 mg/mL as measured by UV spectroscopy. The composition was placed in pre-sterilized tubes and stored under the following ICH approved conditions: 2-8°C, 25°C/65% RH and/or 40°C/75% RH. Stability of the drug-linker compound was measured as % degradation products using the reductive PLRP/MS method or as % intact drug-linker compound using the pepsin cleavage map. The results are presented according to the compositions listed in the left column. The results presented in table. 2 show that the composition containing 6% cyclodextrin exhibited less degradation of the drug-linker junction than the composition containing 2% cyclodextrin after 1 week of storage at 40°C. The results presented in table. 3 show that h2E12-1 formulations containing 6% cyclodextrin exhibited less degradation of the drug-linker junction after 1 week of storage at 25°C, and the results are presented in table. 4 show that h1F6-1 formulations containing 6% hpb cyclodextrin retained higher levels of intact drug-linker compound after 1 week of incubation at 40°C.

table 2

Sample MF6-1 (after 1 week of storage at 40°C) % intact compound drug linker based on pepsin digestion test Initial substance (TO control) 82 20 mM Tris, 2% cyclodextrin 63 20 mM Tris, 6% cyclodextrin 70

Table 3

% compound degradant Sample h2H12-l (after 1 week medicine- storage at 25°C) linker (168 fragments Dalton) 0.5 M Arg (t=0, control) 0.2 0.5 M Arg (25°С, t=l week) 4.7 0.5 M Arg, 6% hppCD (25°С, t=l thirty a week) 6% hpPCD, 20 mM potassium phosphate 0.6 (t=0, control) 6% hpPCD, 20 mM potassium phosphate (25°С, t=l week) 1.4

- 29 043811

Table 4

Sample hlF6-l (after 1 week of storage at 40°C) % intact compound drug linker based on pepsin digestion test Initial substance (TO, control) 85 0.5 M Arg (pH 7.2, 40°C, t=l week) 63 20 mM Tris (pH 7.2, 40°C, t=l week) 66 20 mM 6% hppCD (pH 7.2, 40°C, t=l week) 69

Example 4 Cyclodextrin Concentration Profile

The experiment tested the usefulness of 3% w/v. hydroxypropyl-β-cyclodextrin (hpb-CD) as a component of the diafiltration buffer to prevent aggregation of the quenched conjugation reaction mixture (QCR) containing h2H12-1 during diafiltration. The main purpose of the experiment was to determine whether the membrane used for diafiltration was permeable to hpb-CD. If the membrane is poorly permeable to HPB-CD, its concentration in the batch will exceed the concentration in the diafiltration buffer, whereas if the membrane is easily permeable, the concentration of hpb-CD in the batch will approach the concentration in the diafiltration buffer and remain there.

The diafiltration (DF) process used an 88 cm ² Ultracel Pellicon 3 cassette from Millipore with a molecular weight pore size of 30 kDa. The batch volume for DF was 250 mL, the flow rate was 40 mL/min throughout the process, and the valve installed in the retentate withdrawal line was adjusted to maintain a transmembrane pressure of approximately 20 psi (138 kPa). The rate of addition of diafiltration buffer to maintain a constant batch volume was 9 ml/min at the beginning, but was maintained at 13-14 ml/min throughout most of the process. Samples were collected after completion of each volume of diafiltration. The concentration of hpb-CD was measured in each sample using the RP HPLC method with an evaporative light scattering detector.

The results showed that the hpb-CD concentration increased from 0 at the beginning of diafiltration to 2.8-2.9 wt.% over 5 volumes of diafiltration, then remained constant for an additional 5 volumes of diafiltration (Fig. 14). The concentration profile of hpb-CD corresponds to a reagent that readily penetrates the ultrafiltration membrane used.

Example 5. Tangential flow filtration with 3 and 10% w/v. cyclodextrin

The experiment used QCR containing h2H12-1, 50% v/v. propylene glycol, quenched drug-linker compound (compound 4) and NAC, in 50 mM Tris/5 mM EDTA, pH 8.0. This mixture was adjusted to 10% w/v. hpb-CD by adding 25% volume of 50% w/v. a solution of hpb-CD in 50 mM Tris/5 mM EDTA, pH 8.0. The mixture was then diafiltered under conditions similar to those of Example 4, but modified accordingly to suit the scale. Samples were collected after each volume of diafiltration and frozen until analysis. The concentration of quenched drug-linker compound was determined in each sample. The concentration of the quenched drug-linker compound decreased throughout diafiltration in a manner consistent with a reliable model of impurity removal during constant volume diafiltration, as shown by the linear relationship between In(C/Co) and diafiltration volume #: C/C ₀ = exp (-SN). C and _C0 are the concentrations of the analyte being measured, N is the number of diafiltration volumes, and S is the sieving factor, defined as the analyte concentration on the permeate side of the membrane divided by the analyte concentration on the retentate side of the membrane. The initial concentration was 16.3 μM, and the final concentration was 0.14 μM.

In the presence of 10% w/v. hpb-CD, the quenched compound drug PBD linker is effectively purified by the diafiltration process (Fig. 15). Unlike previous experiments in which purification stopped after a certain number of diafiltration volumes in the absence of hpb-CD, in this experiment purification was effective until very low impurity levels were reached. The experiment showed that effective and complete cleaning could be achieved under test conditions.

The experiment was repeated using a lower concentration of hpb-CD (Fig. 15). Essentially

- 30 043811 identical removal of impurities was observed when 3% w/v. hpb-CD was replaced with 10% w/v.

Example 6 Removal of Drug Impurities from a QCR Blend

The experiments described in Examples 4 and 5 showed that the quenched PBD linker drug compound could be removed from the mixture by a diafiltration process using cyclodextrin. The purpose of this experiment was to determine whether this method could successfully remove drug-linker quenched compound from a QCR having a higher initial concentration of drug-linker quenched compound PBD-linker.

The stored frozen reconstituted antibody (h2H12 antibody reduced at position 239 but not at interchain disulfides) was thawed. Propylene glycol (PG) was added, followed by drug-linker-PG/DMA (compound 1), to obtain the reaction mixture used for conjugation. The mixture contained 50% (v/v) PG, and excess drug-linker compound (compound 4). The reaction was carried out for approximately 90 minutes, after which 3.0 eq. NAC (relative to the drug-linker compound) and continued to react with the quenched reaction mixture for approximately 30 minutes.

A sufficient amount of 50% v/v was added. hpb-CD stock solution to bring the concentration of hpb-CD in the quenched conjugation reaction to 3%. The TFF process was started by turning on the TFF feed pump, and was carried out using 88 cm ² of regenerated cellulose membrane. As with previous TFF processes of solutions containing a high percentage of PG, the initial flow rate was limited due to the high transmembrane pressure, but during the first volume of diafiltration, the flow rate was gradually increased within the recommended operating range of the membrane. The valve installed in the retentate withdrawal line was then adjusted to maintain a TMP of 20 psi (138 kPa). Diafiltration buffer was added to maintain constant volume. The process was continued for 20 diafiltration volumes, and samples were obtained after each diafiltration volume and frozen for later analysis.

Measurement of the concentration of the PBD linker drug quenched compound (compound 4) in solution in samples obtained during the purification process indicated that the material was purified using the TFF process. The concentration was initially 27.47 μM and decreased to 0.03 μM after 20 volumes of diafiltration, indicating that the PBD-linker drug quenched compound could be removed by TFF from the QCR solution (Figure 16). Analysis of a sample of the final product showed that the process did not adversely affect the average drug content and distribution, disulfide bond integrity, charge distribution within the molecule, or the UV spectrum (data not shown).

Example 7 Removal of Drug Impurities from a QCR Blend Containing ADC 2H12 or ADC h1F6

The TFF process was performed on the quenched conjugation reaction mixture using a 0.1 m ² Ultracel Pellicon 3 membrane (Millipore) with a transmittance cutoff of 30 kDa. The TMP was maintained at 20 psi (138 kPa), first maintained at this level by the flow rate, then by a valve installed in the retentate withdrawal line. The TFF process was performed on 20 diafiltration volumes, with samples collected for every 2nd diafiltration volume. The concentration of the PBD linker drug quenched compound was determined in each sample.

In one experiment using a QCR mixture containing ADC 2H12, the concentration of the PBD-linker drug quenched compound (Compound 4) decreased from 29.56 to 0.02 mM during the TFF process. The impurity removal graph had a linear shape, with a screening factor of approximately 0.4. No recovery data was available during the TFF step, but the overall % ADC recovery was greater than 95, so the loss during the TFF process was less than 5 (FIG. 17).

The TFF method is as effective as activated carbon filtration to remove the quenched drug compound PBD-linker, in the presence of 3% hpb-CD. Based on testing of drug-related impurities during the TFF process, it was determined that 14 volumes of diafiltration were sufficient to ensure that impurity levels were sufficiently low.

In one experiment using a QCR mixture containing the ADC h1F6, the concentration of the PBD-linker drug quenched compound (Compound 4) decreased from 21.2 mM to 0.1 μM during the TFF process. The impurity removal graph had a linear shape, with the screening factor being approximately 0.6 (Fig. 18).

Example 8 Purification of Benzodiazepine ADC Using Tangential Flow Filtration Without Cyclodextrin

The quenched drug-linker compound (compound 4) was purified from QCR using constant volume diafiltration. The quenched reaction mixture used for conjugation was introduced into a tangential flow filtration device. The quenched reaction mixture used for conjugation contained Tris, NaCl and 50% propylene glycol. Buffer for tangential

-

Claims (71)

31 043811 in-line filtration also contained Tris, NaCl and 50% propylene glycol. After the ultrafiltration/diafiltration sequence, 1.1 μM of benzodiazepine drug impurity remained in the mixture, with purification ceased after four volumes of diafiltration (data not shown). Example 9 Cytotoxicity Cytotoxicity studies were performed to determine whether ADC compositions containing 6% cyclodextrin retained cytotoxic activity. The results showed that the ADC formulations retained cytotoxic activity (data not shown). Example 10 Removal of Drug Impurities from Quicklik Reaction Mixture Used for Conjugation A quicklink drug PBD linker with a SlogP value of 7.57 (compound 1) was purified from the quicklime conjugation reaction mixture using constant volume diafiltration. The cyclodextrin concentration was maintained at 3%. Surprisingly, it was found that removal of the PBD-linker quenched drug compound (data not shown) was not as effective as removal of the PBD-linker quenched drug compound as shown in the previous examples. The experiment showed that reducing the hydrophobicity of the PBD linker drug compound (compound 4 has a SlogP value of 5.76) improves the purification of the compound. CLAIM 1. A method for removing impurities related to a benzodiazepine drug from a mixture containing antibody-drug conjugates (ADC) of a benzodiazepine and impurities related to a benzodiazepine drug, including performing tangential flow filtration of the specified mixture while maintaining the concentration of cyclodextrin in the mixture from 3 to 30% w/v, wherein the process produces a stabilized mixture, wherein stabilization is defined as a reduction in low molecular weight degradation products, a reduction in the growth of acidic compounds, and/or a reduction in the rate and extent of aggregation compared to a mixture not containing cyclodextrin. 2. The method according to claim 1, wherein the cyclodextrin is present in the mixture at the beginning of the tangential flow filtration. 3. The method according to claim 1, wherein the cyclodextrin is added to the mixture before removing a significant portion of the impurities related to the benzodiazepine drug, and then the concentration of cyclodextrin in the mixture is maintained from 5 to 30% w/v. 4. The method according to any one of claims 1-3, where the concentration of cyclodextrin in the mixture is maintained from 6 to 10% w/v. 5. The method according to any one of claims 1-3, where the concentration of cyclodextrin in the mixture is maintained at 6% w/v. 6. A method as claimed in any one of the preceding claims, wherein the tangential flow filtration is constant volume diafiltration. 7. The method according to any one of claims 1 to 3, where the tangential flow filtration device is a device including a pump, a filtration container having an inlet pipe, a filtrate outlet pipe, a retentate outlet pipe, an ultrafiltration membrane with a pore size 50 kDa or less, which separates the filtration vessel into an inlet compartment and an outlet compartment so that all filtrate must enter the inlet and pass through the ultrafiltration membrane before flowing out of the filtration vessel through the filtrate outlet, sample container designed for storage a mixture containing a benzodiazepine ADC and impurities related to the benzodiazepine drug, and a buffer container fluidly connected to the sample container, and wherein the buffer contained in the buffer container contains 6% w/v. cyclodextrin. 8. The method of claim 7, wherein the buffer replaces the filtration volume at a rate equal to the filtration speed, such that the volume in the tangential flow filtration device remains constant. 9. Method according to any one of claims 1 to 5, where the filtration is batch diafiltration. 10. The method of any of the preceding claims, further comprising the steps of: (i) contacting the antibody with a benzodiazepine drug-linker compound under conditions sufficient to create a reaction mixture containing a benzodiazepine ADC and impurities related to the benzodiazepine drug, and (ii) contacting the reaction mixture with a quenching additive to obtain a quench mixture used for the conjugation reaction, wherein said mixture, subjected to tangential flow filtration, is a quenched reaction mixture used for the conjugation reaction. 11. The method according to any one of claims 1 to 10, where the impurities related to the benzodiazepine drug - 32 043811 drugs are quenched benzodiazepine drug-linker compounds. 12. The method of claim 11, wherein the SlogP value for the quenched benzodiazepine drug-linker compounds is no more than 7.50 or no more than 6.5. 13. The method as claimed in any one of the preceding claims, wherein the amount of impurities related to the benzodiazepine drug is reduced to a level of 1 μM or less, 0.5 μM or less, 0.1 μM or less, or 0.05 μM or less. 14. The method according to any of the preceding claims, wherein the cyclodextrin is hydroxypropyl beta-cyclodextrin or beta-cyclodextrin sulfobutyl ether. 15. The method according to claim 14, wherein the cyclodextrin is hydroxypropyl-beta-cyclodextrin. 16. The method of any one of the preceding claims, wherein the benzodiazepine ADC is a monoclonal antibody conjugated to a pyrrolobenzodiazepine dimer, an indolinobenzodiazepine dimer, or an oxazolidinobenzodiazepine dimer. 17. The method according to claim 16, wherein the pyrrolobenzodiazepine dimer is the compound shown below: ^Η γ ^Ν γ ^< γ ^ο ^/ ^ο '-<γ ^Ν; Λ, ^Η J ] O O Ύ 1 or its salt. 18. The method of claim 16, wherein the benzodiazepine ADC is a monoclonal antibody conjugated to a pyrrolobenzodiazepine dimer, wherein the benzodiazepine ADC has the following formula: or a salt thereof, where Ab is a monoclonal antibody and p is the average number of drug-linker molecules per antibody in the mixture and is approximately 2. 19. The method of claim 17 or 18, wherein the benzodiazepine drug-related impurities are a quenched drug-linker compound, wherein the drug-linker compound before quenching is a compound of the formula or its salt. 20. The method according to any one of claims 1 to 15, wherein the benzodiazepine ADC is a monoclonal antibody conjugated to an indolinobenzodiazepine dimer or an oxazolidinobenzodiazepine dimer. 21. Stabilized mixture containing A benzodiazepine ADC and impurities related to a benzodiazepine drug, and cyclodextrin at a concentration of from about 3 to about 30% w/v; wherein impurities related to the benzodiazepine drug are present in an amount of 0.1 μM or less, and wherein said stabilized mixture contains from 0.5 to 30 mg/ml benzodiazepine ADC, wherein said stabilized mixture is prepared by the method of any of the preceding claims , with stabilization defined as a reduction in low molecular weight degradation products, a reduction in the growth of acidic compounds, and/or a reduction in the rate and extent of aggregation compared to a mixture not containing cyclodextrin. 22. Liquid pharmaceutical composition containing from 0.5 to 10 mg/ml ADC benzodiazepine; 0.1 µM or less of an impurity related to a benzodiazepine drug; and β-cyclodextrin at a concentration of about 3% w/v. to approximately 30% w/v, wherein the liquid pharmaceutical composition is stabilized, with stabilization defined as a reduction in low molecular weight degradation products, a reduction in the growth of acidic compounds, and/or a reduction in the rate and extent of aggregation compared to a mixture not containing cyclodextrin. 23. The liquid pharmaceutical composition according to claim 22, where the composition further contains at least one buffering agent; where the composition is an aqueous solution and concentrate - 33 043811 tion wherein at least one buffering agent is effective to maintain a physiologically appropriate pH value. 24. Liquid pharmaceutical composition according to claim 22 or 23, where the composition is an aqueous solution, and the cyclodextrin is presented in a concentration of 5% w/v. or from 6% w/v. up to 30% w/v 25. Liquid pharmaceutical composition according to claim 24, where the cyclodextrin is presented in a concentration of 6% w/v. up to 10% w/v 26. The liquid pharmaceutical composition according to any one of claims 22 to 25, wherein the composition is an aqueous solution further comprising a lyoprotectant such that the aqueous solution is suitable for lyophilization. 27. Liquid pharmaceutical composition according to claim 26, where the lyoprotector is sugar. 28. The liquid pharmaceutical composition according to claim 27, wherein the sugar is sucrose. 29. Liquid pharmaceutical composition according to any one of claims 26-28, where the lyoprotector is presented in a concentration of 4 to 8% (wt./vol.). 30. Liquid pharmaceutical composition according to claim 29, where the lyoprotector is presented at a concentration of 6% (wt./vol.). 31. Liquid pharmaceutical composition according to any one of claims 22 to 30, where the composition is an aqueous solution having a pH value from 6.0 to 8.0. 32. Liquid pharmaceutical composition according to claim 31, wherein the aqueous solution has a pH value from 6.5 to 7.5. 33. Liquid pharmaceutical composition according to claim 32, wherein the aqueous solution has a pH value of 7.3. 34. The liquid pharmaceutical composition according to any one of claims 23 to 33, wherein at least one buffering agent is selected from the group consisting of tris, acetate, histidine, citrate, phosphate and succinate. 35. The liquid pharmaceutical composition according to claim 34, wherein at least one buffering agent is Tris at a concentration of 20 or 50 mM. 36. Liquid pharmaceutical composition according to any one of claims 22 to 35, wherein the benzodiazepine ADC is present at a concentration of 0.5, 1 or 2 to 10 mg/ml. 37. The liquid pharmaceutical composition according to claim 36, wherein the concentration of the benzodiazepine ADC is from 2 to 5 mg/ml. 38. The liquid pharmaceutical composition according to claim 37, wherein the concentration of the benzodiazepine ADC is 3 mg/ml. 39. Liquid pharmaceutical composition according to any one of claims 22 to 38, wherein the β-cyclodextrin is hydroxypropyl-beta-cyclodextrin or beta-cyclodextrin sulfobutyl ether. 40. Liquid pharmaceutical composition according to any one of claims 22 to 38, wherein the β-cyclodextrin is hydroxypropyl-beta-cyclodextrin. 41. The liquid pharmaceutical composition according to claim 22, wherein the benzodiazepine ADC is a monoclonal antibody conjugated to a pyrrolobenzodiazepine dimer, an indolinobenzodiazepine dimer, or an oxazolidinobenzodiazepine dimer. 42. The liquid pharmaceutical composition according to claim 41, wherein the pyrrolobenzodiazepine dimer is the following compound: or a pharmaceutically acceptable salt thereof. 43. The liquid pharmaceutical composition according to claim 41, wherein the benzodiazepine ADC is a monoclonal antibody conjugated to a pyrrolobenzodiazepine dimer and is the following compound: or a pharmaceutically acceptable salt thereof; where Ab is a monoclonal antibody and p is the average number of drug-linker molecules per antibody in the composition and is approximately 2. 44. The liquid pharmaceutical composition according to any one of claims 22 to 40, wherein the benzodiazepine ADC is a monoclonal antibody conjugated to an indolinobenzodiazepine dimer or an oxazolidinobenzodiazepine dimer. 45. The liquid pharmaceutical composition according to any one of claims 22 to 44, wherein the bezodiazepine ADC antibody component is an antibody that specifically binds to an antigen that - 34 043811 is expressed on the surface of a cancer cell, or with the extracellular domain of human CD19, human CD33 or human CD70. 46. The liquid pharmaceutical composition according to claim 45, wherein the antibody is an h2H12 antibody having a light chain variable sequence containing the amino acid sequence specified in SEQ ID NO: 3 and a heavy chain variable sequence containing the amino acid sequence specified in SEQ ID NO: : 4, or the antibody is an h1F6 antibody having a light chain variable sequence containing the amino acid sequence specified in SEQ ID NO: 13, and a heavy chain variable sequence containing the amino acid sequence specified in SEQ ID NO: 14. 47. Liquid pharmaceutical composition according to any one of claims 23-46, where the antibody, as a component of the ADC, contains a heavy chain constant region of the IgG1 isotype. 48. A liquid pharmaceutical composition containing pyrrolobenzodiazepine antibody-drug conjugates (PBD ADC), where the concentration of PBD ADC is from 2 to 5 mg/ml; hydroxypropyl beta-cyclodextrin at a concentration of 5 to 10% w/v; sugar in a concentration of 4 to 8% w/v. and at least one buffering agent; where the composition is an aqueous solution and the concentration of at least one buffering agent in it is sufficient to maintain a physiologically acceptable pH value. 49. Liquid pharmaceutical composition according to claim 48, where the concentration of PBD ADC is 3 mg/ml; the concentration of hydroxypropyl-β-cyclodextrin is 6% w/v; sugar is sucrose at a concentration of approximately 6% w/v; at least one buffering agent is Tris at a concentration of 20 mM and the pH value is from 7.0 to 7.5 or is 7.2 or 7.3. 50. The liquid pharmaceutical composition according to claim 48 or 49, where the PBD ADC is a compound of the formula: 'OME MeO' or a pharmaceutically acceptable salt thereof; where Ab is a monoclonal antibody and p is the average number of drug-linker molecules per antibody in the composition and is approximately 2. 51. The liquid pharmaceutical composition according to any one of claims 22 to 50, wherein the benzodiazepine drug, as a component of the ADC or PBD ADC, is conjugated to an antibody, where the antibody is a monoclonal antibody, using an integrated cysteine residue present in the monoclonal antibody , and the interchain disulfide bonds of the antibody are essentially intact. 52. The liquid pharmaceutical composition according to claim 51, wherein the cysteine residue is located at position 239 of the IgG1 heavy chain, where the numbering corresponds to the EU index as indicated by Kabat. 53. The liquid pharmaceutical composition according to any one of claims 48 to 52, wherein the monoclonal antibody is a humanized anti-CD33 or anti-CD70 monoclonal antibody. 54. The liquid pharmaceutical composition according to claim 53, wherein the humanized anti-CD33 antibody is an h2H12 antibody having a light chain variable sequence containing the amino acid sequence specified in SEQ ID NO: 3, and a heavy chain variable sequence containing the amino acid sequence specified in SEQ ID NO: 4, and the humanized anti-CD70 antibody is an h1F6 antibody having a light chain variable sequence comprising the amino acid sequence specified in SEQ ID NO: 13 and a heavy chain variable sequence containing the amino acid sequence specified in SEQ ID NO: 14. 55. A method for producing a stabilized, lyophilized benzodiazepine antibody-drug conjugate composition, comprising the steps of: (a) providing a liquid composition according to any one of claims 24-54 and a lyoprotector, wherein the liquid composition is an aqueous solution; and (b) lyophilizing the aqueous solution to create a lyophilized benzodiazepine antibody-drug conjugate composition, where stabilization is defined as a reduction in low molecular weight degradation products, a reduction in the growth of acidic species, and/or a reduction in the rate and extent of aggregation compared to a mixture not containing cyclodextrin. - 35 043811 56. Stabilized lyophilized benzodiazepine antibody-drug conjugate composition obtained by the method according to claim 55. 57. An aqueous pharmaceutical benzodiazepine antibody-drug conjugate composition prepared by reconstituting the stabilized, lyophilized benzodiazepine antibody-drug conjugate composition of claim 56 by adding a suitable diluent, wherein the diluent is water for injection. 58. A method for preventing chemical degradation and fragmentation of a benzodiazepine drug-linker compound attached to an antibody, comprising creating a composition of a benzodiazepine drug-linker compound attached to an antibody with 6% w/v. gamma-cyclodextrin or beta-cyclodextrin. 59. The liquid pharmaceutical composition according to claim 22, wherein the composition is an aqueous solution in which the benzodiazepine ADC is present at a concentration of 1 to 5 mg/ml. 60. The liquid pharmaceutical composition according to claim 60, wherein the benzodiazepine ADC in aqueous solution is presented at a concentration of 1, 2 or 3 mg/ml. 61. The liquid pharmaceutical composition according to claim 22, where the composition is an aqueous solution and the cyclodextrin in the aqueous solution is represented by hydroxypropyl-beta-cyclodextrin or beta-cyclodextrin sulfobutyl ether at a concentration of 6% w/v. 62. The liquid pharmaceutical composition of claim 22, wherein the composition is an aqueous solution, wherein the aqueous solution further comprises a buffering agent, wherein the buffering agent is Tris at a concentration effective to maintain a pH of from about 7 to 7.5. 63. The liquid pharmaceutical composition according to claim 22, wherein the composition is an aqueous solution, wherein the aqueous solution further contains a Tris-based buffer at a concentration of 5 to 50 mM. 64. The liquid pharmaceutical composition according to claim 22, wherein the composition is an aqueous solution, wherein the aqueous solution further contains a buffering agent, wherein the buffering agent is Tris at a concentration of 20 or 50 mM to maintain the pH at 7.3. 65. Liquid pharmaceutical composition according to claim 64, characterized in that the aqueous solution additionally contains a surfactant. 66. The liquid pharmaceutical composition according to claim 65, wherein the surfactant is polysorbate 20 or polysorbate 80. 67. A lyophilized benzodiazepine ADC composition obtained by lyophilizing the liquid composition according to claim 22, further containing a lyoprotector, wherein the liquid composition to be lyophilized is an aqueous solution. 68. The lyophilized composition of claim 67, wherein the amount of lyoprotectant in the aqueous composition to be lyophilized is such that when reconstituted into an aqueous composition, the reconstituted aqueous composition is substantially isotonic. 69. Lyophilized composition according to claim 68, in which the lyoprotectant is sucrose. 70. The lyophilized composition according to claim 68, characterized in that the aqueous composition to be lyophilized contains 1, 2 or 3 mg/ml benzodiazepine ADC, 20 or 50 mm Tris, 6% w/v. sucrose, 6% w/v. hydroxypropyl beta-cyclodextrin and optionally polysorbate 20. 71. The lyophilized composition according to claim 68, characterized in that the benzodiazepine ADC in the aqueous composition to be lyophilized is a monoclonal antibody conjugated to a pyrrolobenzodiazepine dimer, which is the following compound: 'OME MeO' or a pharmaceutically acceptable salt thereof; wherein Ab is a monoclonal antibody and p is the average number of drug-linker molecules per antibody in the composition and is approximately 2, wherein the monoclonal antibody specifically binds to a cancer cell antigen that is expressed on the surface of the cancer cell and to the extracellular domain Human CD19, human CD33, or human CD70. -