Chromatographic Methods for Purification of Insulin Analogs

Abstract

The present invention relates to a method for purifying non-glycosylated insulin analogs from a mixture containing glycosylated proteins and non-glycosylated proteins by removing glycosylated proteins. This removal is achieved by a combination of two RP-HPLC steps.

Description

Technical Field

[0001] The present invention relates generally to the field of protein purification methods. In particular, a chromatographic method for purifying non-glycosylated insulin analogs is provided. Background Art

[0002] Recombinant forms of insulin, insulin analogs and / or derivatives have been produced in various microbial expression systems. Organisms such as Escherichia coli and S. cerevisiae have been used in the commercial production of recombinant human insulin and its derivatives. Proteins or protein precursors may be derived from yeast expression systems, and the correlation between high expression levels and increased glycosylated protein precursors has been observed. Therefore, the demand for more effective methods for separating non-glycosylated proteins from glycosylated proteins becomes more obvious. Due to the shortcomings of these systems, such as low expression levels, downstream purification difficulties, etc., methylotrophic yeast Pichia pastoris is preferentially used as a protein expression system. In U.S. Patent No. 6,800,606, several advantages of Pichia pastoris expression systems, such as high expression, low production costs and high-density cultivation, are mentioned.

[0003] However, one of the major drawbacks of the Pichia pastoris expression system is the post-translational modification of recombinant proteins present as impurities in the final product, which is very challenging for purification. Although many post-translational modifications of proteins are known, the most common form of post-translational modification is glycosylation (Hart GW, Glycosylation, Curr. Opin. Cell. Biol 1992; 4: 1017). Depending on the expression system, glycosylation can be N-linked or O-linked (Gemmill TR et al., Overview of N-and O-linked oligosaccharide structures found in various yeast species, Biochemica et Biophysica Acta, 1999; 1426: 227). Glycosylation affects the stability of protein conformation, immunogenicity, clearance rate, prevention of proteolysis, and improves protein solubility (Walsh G, Biopharmaceutical benchmarks 2006, Nature Biotechnology, 2006; 24: 769).

[0004] U.S. Patent No. 6,180,757 relates to the use of calcium (Ca ++) A method for separating glycosylated proteins or protein precursors from non-glycosylated proteins or protein precursors by chromatography using an eluent. By using this method, a fraction containing non-glycosylated proteins substantially free of glycosylated proteins is obtained. However, the patent does not disclose a completely purified product.

[0005] US Patent No. 8,802,816 to Biocon Limited relates to a process for separating and / or purifying a product from impurities, which process yields purified heterologous recombinant insulin glargine with a purity of 96%.

[0006] US Patent Publication No. 20120178900 of Biocon Limited discloses a RP-HPLC method for purifying insulin Aspart, Atisoban, insulin Glargine and insulin Lispro. The method comprises using RP-HPLC and an ion pairing reagent in combination with an organic modifier.

[0007] Purpose of the Invention

[0008] The object of the present invention is to provide a method for obtaining at least 99.95% pure non-glycosylated insulin glargine and non-glycosylated insulin lispro that is substantially free of mono-glycosylated variants, di-glycosylated variants or tri-glycosylated variants. This is achieved by using a combination of two reverse phase HPLC steps, wherein the first step is carried out at a lower pH using octane sulfonic acid (OSA) as a preferred ion pairing agent, and the second step is carried out at a high pH using sodium perchlorate or tetrabutylammonium hydrogen sulfate (TBAB) as a preferred ion pairing agent. The first and second steps can be carried out in reverse order to achieve the removal of at least 99.95% of the glycosylated variants of insulin glargine or insulin lispro. Summary of the invention

[0009] The present invention relates to a method for purifying non-glycosylated insulin analogs such as insulin glargine and insulin lispro, which purifies the non-glycosylated insulin analogs from a complex mixture containing non-glycosylated, mono-glycosylated and poly-glycosylated variants of insulin glargine and insulin lispro by achieving at least 99.95% removal of the glycosylation variants.

[0010] The purification method includes two steps of reverse phase high performance liquid chromatography (RP-HPLC) performed in succession. In one possible method, the first step of RP-HPLC is performed with the complex mixture at an acidic pH in the presence of an ion pair reagent (preferably octane sulfonic acid (OSA)) in combination with an organic modifier (preferably acetonitrile) to obtain a first mixture containing partially purified non-glycosylated insulin analogs, wherein the level of glycosylated variants of the insulin analogs is significantly reduced. At an alkaline pH, the first mixture is subjected to a second step of RP-HPLC in combination with an ion pair reagent and an organic modifier (preferably acetonitrile) to obtain at least 99.95% pure non-glycosylated insulin analogs.

[0011] In a second possible method, the complex mixture is subjected to a first RP-HPLC step at alkaline pH in combination with an ion pair reagent and an organic modifier (preferably acetonitrile) to obtain a first mixture comprising partially purified non-glycosylated insulin analogs, wherein the level of glycosylated variants of the insulin analogs is significantly reduced. The first mixture is subjected to a second RP-HPLC step at acidic pH in the presence of a combination of an ion pair reagent (preferably octane sulfonic acid (OSA)) and an organic modifier (preferably acetonitrile) to obtain at least 99.95% pure non-glycosylated insulin analogs.

[0012] The insulin analogue can be insulin glargine or insulin lispro. When the insulin analogue is insulin glargine, the ion pair reagent used for RP-HPLC under alkaline pH conditions is preferably sodium perchlorate. When the insulin analogue is insulin lispro, the ion pair reagent used for RP-HPLC under alkaline pH conditions is preferably TBAB.

[0013] The removal of glycosylated forms of insulin analogs was achieved by a combination of low pH based RP-HPLC and high pH based RP-HPLC. As a result, at least 99.95% of the glycosylated variants of insulin glargine and insulin lispro were removed.

[0014] This Summary is not intended to identify essential features of the claimed subject matter, nor is it intended to be used to determine or limit the scope of the claimed subject matter. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The following drawings constitute part of this specification and are included to further illustrate various aspects of the present disclosure. The present disclosure may be better understood by referring to the drawings in conjunction with the detailed description of specific embodiments presented herein.

[0016] Figure 1 Flow chart showing the method of purification step-1.

[0017] Figure 2 Flow chart showing the process of purification step-2.

[0018] Figure 3A UV and TIC chromatograms of the reverse phase purified-1 final product representing the process of purifying insulin glargine using low pH chromatography followed by high pH chromatography.

[0019] Figure 3B UV and TIC chromatograms of the reverse phase purified-2 final product representing the method for purifying insulin glargine using low pH chromatography followed by high pH chromatography.

[0020] Figure 4A , 4B Mass spectral data showing the reverse phase purification-1 final product of the process for purification of insulin lispro using low pH chromatography followed by high pH chromatography - UV and TIC chromatograms.

[0021] Figure 5 is a UV and TIC chromatogram of the reverse phase purified-2 final product of the process for purifying insulin lispro using low pH chromatography followed by high pH chromatography.

[0022] Fig. 6A and 6B Mass spectral data showing the reverse phase purification-1 final product and the reverse phase purification-2 final product of the process for purifying insulin glargine using high pH chromatography followed by low pH chromatography.

[0023] Fig. 7A , 7B 7C are UV and TIC chromatograms of the reverse phase purification-1 final product of the method for purifying insulin lispro using high pH chromatography followed by low pH chromatography.

[0024] Fig. 8A , 8B 8C are UV and TIC chromatograms of the reverse phase purified-2 final product of the method for purifying insulin lispro using high pH chromatography followed by low pH chromatography. DETAILED DESCRIPTION

[0025] Those skilled in the art will appreciate that the invention described herein may be subject to changes and modifications other than those specifically described. It should be understood that the invention described herein includes all such changes and modifications. The present invention also includes any and all combinations of any two or more of all of the steps, features, compositions and methods mentioned or indicated in this specification, either individually or collectively, and any of the steps or features.

[0026] definition:

[0027] For convenience, before further describing the present invention, certain terms and examples used in the specification are collected here. Those skilled in the art should read and understand these definitions based on the remainder of this disclosure. Unless otherwise defined, all technical and scientific terms used in this article have the same meanings as those of ordinary skill in the art are generally understood. Unless otherwise limited in specific circumstances, the terms used in the entire specification are defined as follows.

[0028] The present disclosure is not to be limited in scope by the specific embodiments described herein, which are intended for purposes of illustrative purposes only.

[0029] As described herein, functionally equivalent processes and methods are apparent to be within the scope of the present disclosure.

[0030] The term "deglycosylated" refers specifically to a protein from which the sugar entities (glycans) have been removed by a purification process.

[0031] The term "glycosylated" specifically refers to proteins to which carbohydrate moieties (such as mannose, polymannose and ribose) are attached to amino acids of the A chain or B chain of the protein or to amino acids in both A and B chains. The term "glycosylated protein" includes monoglycosylated proteins as well as polyglycosylated proteins.

[0032] The term "non-glycosylated" refers specifically to proteins to which no sugar moieties (glycans) are attached to any chain of the protein.

[0033] The term "insulin glargine" refers in particular to rapid-acting and long-acting human insulin analogs which differ from human insulin in that the amino acid asparagine at position 21 of the insulin A chain is replaced by glycine and two arginines are added to the C-terminus of the B chain.

[0034] The term "insulin lispro" refers in particular to a rapid-acting human insulin analogue which is chemically different from human insulin. In insulin lispro, the amino acid proline at position B28 is substituted by lysine and the lysine at position B29 is substituted by proline.

[0035] The term "RP-HPLC" refers specifically to reverse phase high performance liquid chromatography, which involves the separation of molecules based on hydrophobicity.

[0036] The term "downstream purification" refers to the recovery and purification of the biosynthetic drug product from associated impurities and wastes generated during the production process.

[0037] The present invention relates to a method for purifying non-glycosylated insulin analogs by removing at least 99.95% of various forms of mono- and poly-glycosylated insulin analogs. The method is a selective purification of the product by downstream purification techniques optimized by a better understanding of the nature of the impurities present in the final product.

[0038] The purification method for purifying non-glycosylated insulin analogs is the result of the combined effects of two chromatographic techniques (referred to as purification step-1 and purification step-2) used in succession. Both chromatographic steps used to purify non-glycosylated insulin analogs are column chromatography, preferably RP-HPLC or hydrophobic interaction chromatography. The advantage of using the above-mentioned chromatography is that the purity of the fraction of non-glycosylated insulin analogs is improved. The purified final product is substantially (at least 99.95%) free of glycosylation variants. Preferably, in the purified fraction of non-glycosylated insulin, the concentration of the glycosylated form of insulin analogs is less than 0.05%; thereby, the glycosylation variants are removed to a level below quantification (BLOQ).

[0039] The present invention further relates to the identification of various glycoforms of insulin analogs, in particular insulin glargine and insulin lispro, by chemical methods combined with mass spectrometry techniques such as electrospray and matrix-assisted laser desorption ionization for identification.

[0040] In the present invention, the purification step performed at low pH uses octane sulfonic acid as the preferred ion pair reagent, preferably the sodium salt of octane sulfonic acid (OSA) in the reverse phase mode of chromatography. OSA is a cation pair reagent with a hydrophobic chain of 8 carbon atoms connected to a sulfonic acid group (SO3-). OSA interacts with the protein sample at acidic pH and changes the binding chemistry, thereby improving the separation between glycosylated and non-glycosylated proteins.

[0041] The present invention has proven to be particularly advantageous in separating glycosylated insulin from non-glycosylated insulin.Preferably, the concentration of glycosylated insulin in the purified non-glycosylated insulin fraction is less than 0.05%.

[0042] The present disclosure will now be illustrated by working examples which are intended to illustrate the working of the present disclosure and are not intended to restrictively imply any limitation on the scope of the present disclosure.

[0043] In an embodiment of the present invention, a method for purifying a non-glycosylated insulin analog from a mixture comprising non-glycosylated and glycosylated insulin analogs is provided.

[0044] In one embodiment, the insulin analog is insulin glargine. In another embodiment, the insulin analog is insulin lispro.

[0045] In one embodiment, the glycosylated insulin analog may be monoglycosylated, diglycosylated, triglycosylated, polyglycosylated, and mixtures thereof.

[0046] In one embodiment, the purification method of the invention results in the removal of at least 99.95% of the glycosylated forms of the insulin analog from a mixture comprising non-glycosylated and glycosylated insulin analogs. In preferred embodiments, the removal percentages are 99.96, 99.97, 99.98, 99.99 and 100.

[0047] In one embodiment, the purification method of the present invention results in the separation of glycosylated insulin glargine ≤ 0.05% of non-glycosylated insulin glargine. In a preferred embodiment, the purification method of the present invention results in the separation of glycosylated insulin glargine ≤ 0.05%, 0.04%, 0.03%, 0.02%, 0.01% or 0% of non-glycosylated insulin glargine.

[0048] In one embodiment, the purification method of the present invention results in the separation of glycosylated insulin lispro ≤ 0.05% of non-glycosylated insulin lispro. In a preferred embodiment, the purification method of the present invention results in the separation of glycosylated insulin lispro ≤ 0.05%, 0.04%, 0.03%, 0.02%, 0.01% or 0% of non-glycosylated insulin lispro.

[0049] In one embodiment, a method for purifying non-glycosylated insulin glargine from a mixture containing non-glycosylated and glycosylated insulin glargine is provided, the method consisting of a first step and a second step, wherein the first step comprises subjecting the mixture containing glycosylated and non-glycosylated insulin glargine to RP-HPLC at an acidic pH in the presence of an ion pair reagent and acetonitrile (organic modifier) ​​to obtain a first mixture containing partially purified non-glycosylated insulin glargine; wherein the second step comprises subjecting the first mixture from the first step to RP-HPLC at an alkaline pH in the presence of an ion pair reagent and acetonitrile (organic modifier) ​​to obtain at least 99.95% pure non-glycosylated insulin glargine. In a preferred embodiment, the ion pair reagent in the first step is octane sulfonic acid, more preferably the sodium salt of OSA, and the ion pair reagent in the second step is sodium perchlorate. In one embodiment, the pH range of the first step is 3.5-3.9, and the pH range of the second step is 8.3-8.7.

[0050] In one embodiment, a method for purifying non-glycosylated insulin lispro from a mixture containing non-glycosylated and glycosylated insulin lispro is provided, the method consisting of a first step and a second step, wherein the first step comprises subjecting the mixture containing glycosylated and non-glycosylated insulin lispro to RP-HPLC at an acidic pH in the presence of an ion pairing agent and acetonitrile (organic modifier) ​​to obtain a first mixture containing partially purified non-glycosylated insulin lispro; wherein the second step comprises subjecting the first mixture from the first step to RP-HPLC at an alkaline pH in the presence of an ion pairing agent and acetonitrile (organic modifier) ​​to obtain at least 99.95% pure non-glycosylated insulin lispro. In a preferred embodiment, the ion pairing agent in the first step is octane sulfonic acid, more preferably the sodium salt of OSA, and the ion pairing agent in the second step is TBAB. In one embodiment, the pH range of the first step is 3.5-3.9, and the pH range of the second step is 8.3-8.7.

[0051] In one embodiment, a method for purifying non-glycosylated insulin glargine from a mixture containing non-glycosylated and glycosylated insulin glargine is provided, the method consisting of a first step and a second step, wherein the first step comprises subjecting the mixture containing glycosylated and non-glycosylated insulin glargine to RP-HPLC at an alkaline pH in the presence of an ion pair reagent and acetonitrile (organic modifier) ​​to obtain a first mixture containing partially purified non-glycosylated insulin glargine; wherein the second step comprises subjecting the first mixture from the first step to RP-HPLC at an acidic pH in the presence of an ion pair reagent and acetonitrile (organic modifier) ​​to obtain at least 99.95% pure non-glycosylated insulin glargine. In a preferred embodiment, the ion pair reagent in the first step is sodium perchlorate, and the ion pair reagent in the second step is octane sulfonic acid, preferably the sodium salt of octane sulfonic acid. In one embodiment, the pH range of the first step is 8.3-8.7, and the pH range of the second step is 3.5-3.9.

[0052] In one embodiment, a method for purifying non-glycosylated insulin lispro from a mixture containing non-glycosylated and glycosylated insulin lispro is provided, the method consisting of a first step and a second step, wherein the first step comprises subjecting the mixture containing glycosylated and non-glycosylated insulin lispro to RP-HPLC at an alkaline pH in the presence of an ion pairing agent and acetonitrile (organic modifier) ​​to obtain a first mixture containing partially purified non-glycosylated insulin lispro; wherein the second step comprises subjecting the first mixture from the first step to RP-HPLC at an acidic pH in the presence of an ion pairing agent and acetonitrile (organic modifier) ​​to obtain at least 99.95% pure non-glycosylated insulin lispro. In a preferred embodiment, the ion pairing agent in the first step is TBAB, and the ion pairing agent in the second step is octane sulfonic acid, preferably the sodium salt of octane sulfonic acid. In one embodiment, the pH range of the first step is 8.3-8.7, and the pH range of the second step is 3.5-3.9.

[0053] In one embodiment, non-glycosylated insulin glargine that is at least 99.95% free of glycosylation variants of insulin glargine is provided. In one embodiment, non-glycosylated insulin glargine that is at least 99.96% free of glycosylation variants of insulin glargine is provided. In one embodiment, non-glycosylated insulin glargine that is at least 99.97% free of glycosylation variants of insulin glargine is provided. In one embodiment, non-glycosylated insulin glargine that is at least 99.98% free of glycosylation variants of insulin glargine is provided. In one embodiment, non-glycosylated insulin glargine that is at least 99.99% free of glycosylation variants of insulin glargine is provided. In one embodiment, non-glycosylated insulin glargine that is 100% free of glycosylation variants of insulin glargine is provided.

[0054] In one embodiment, non-glycosylated insulin lispro that is at least 99.95% free of glycosylation variants of insulin lispro is provided. In one embodiment, non-glycosylated insulin lispro that is at least 99.96% free of glycosylation variants of insulin lispro is provided. In one embodiment, non-glycosylated insulin lispro that is at least 99.97% free of glycosylation variants of insulin lispro is provided. In one embodiment, non-glycosylated insulin lispro that is at least 99.98% free of glycosylation variants of insulin lispro is provided. In one embodiment, non-glycosylated insulin lispro that is at least 99.99% free of glycosylation variants of insulin lispro is provided. In one embodiment, non-glycosylated insulin lispro that is 100% free of glycosylation variants of insulin lispro is provided.

[0055] Example

[0056] Purification of insulin glargine by low pH based RP-HPLC followed by high pH based RP-HPLC

[0057] Purification step at low pH - 1 Figure 1 The details of purification step-1 are as follows

[0058] Stationary Phase Details:

[0059] a. Medium: Kromasil

[0060] b. Height: 25.0±1.0cm

[0061] c. Linear velocity: All chromatographic steps ≤ 360 cm / h

[0062] Mobile phase details:

[0063] a. Mobile phase A: 100 mM sodium acetate + 0.05% (w / v) octane sulfonic acid (OSA), pH 3.7 ± 0.1 (pH adjusted with glacial acetic acid)

[0064] b. Mobile phase B: 100.0% acetonitrile

[0065] c. Temperature of mobile phase A and B during purification cycle: 20-30°C

[0066] Sample preparation:

[0067] a. Dilute the sample concentrate obtained in the previous step with a mixture of pure water and acetic acid at a ratio of ≥1.0:10.0.

[0068] b. After dilution, filter the sample through a 1.2 μm filter, a 0.45 μm filter, and then a 0.22 μm filter.

[0069] c. The filter for 1.2 μm filtration is made of polypropylene (PP), and the filter for 0.45 / 0.22 μm filtration is made of polyethersulfone (PES) material.

[0070] d. The product concentration in the above sample is expected to be in the range of 1.0 g / L to 2.0 g / L.

[0071] Process conditions used for purification:

[0072] a. Equilibrate the column with 4 to 5 CV (column volumes) using a mobile phase of 90.0% (A):10.0% (B) at a linear velocity of ≤360 cm / h.

[0073] b. Load the column at a linear velocity of ≤360 cm / h at 4-12 g of insulin glargine per liter of Kromasil.

[0074] c. After loading, wash the column with 4 to 5 CV (column volumes) at a linear velocity of ≤360 cm / h using a mobile phase of 85.0% (A):15.0% (B).

[0075] d. The product was eluted from the column using a linear gradient of 24-31% B over 25 CV at a linear velocity of ≤ 360 cm / h.

[0076] e. During the elution process, variable volume fractions were collected according to the increase in absorbance (UV) at 280 nm. The collection continued until UV 280 The volume required to elute the protein from the column is 2.0-3.0 CV.

[0077] f. Analyze and pool the eluted fractions according to their purity levels and store the prepared elution pool at 2-8°C.

[0078] g. After elution to remove residual protein (if any), the column was regenerated by passing 50.0% 1 M acetic acid and 50.0% acetonitrile in an upward flow direction for 2CV at a linear velocity of ≤360 cm / h, followed by 1V of 30.0% 1 M acetic acid and 70.0% acetonitrile.

[0079] The results of different relative retention times (RRT) are shown in the tabulated data in Table 1. According to the data, glycosylated proteins were reduced by about 4.5 times, from about 2.6% in the upper sample to about 0.6% in the elution pool (EP).

[0080] Table 1: Results of purification step-1 at low pH

[0081]

[0082] Note: The LOQ (limit of quantification) of the analytical method is 0.04%, the LOD (limit of detection) is 0.02%, and the BLOD - below the limit of detection

[0083] The elution pool from purification step-1 was subjected to purification step-2 to further purify non-glycosylated proteins from glycosylated proteins.

[0084] Purification step at high pH -2 Figure 2 The details of purification step-2 are as follows

[0085] Stationary Phase Details:

[0086] a. Medium: Kromasil

[0087] b. Height: 25.0±1.0cm

[0088] c. Linear velocity: All chromatographic steps ≤ 360 cm / h

[0089] Mobile phase details:

[0090] a. Mobile phase A: 100mM Tris + 50mM sodium perchlorate, pH 8.5±0.1 (pH adjusted with glacial acetic acid)

[0091] b. Mobile phase B: 100.0% acetonitrile

[0092] c. Temperature of mobile phase A and B during purification cycle: 20-30°C

[0093] Sample preparation:

[0094] a. Dilute the elution pool (EP) obtained in purification step-1 with pure water at a ratio of 1.0:2.0.

[0095] b. Adjust the pH of the diluted sample to 8.5 ± 0.1 using 2.5 M Tris.

[0096] c. After adjusting the pH, the sample was filtered through 1.2 μm, 0.45 μm, and then 0.22 μm.

[0097] d. The filter for 1.2 μm filtration is made of polypropylene (PP), and the filter for 0.45 / 0.22 μm filtration is made of polyethersulfone (PES) material.

[0098] e. The product concentration in the above sample is expected to be in the range of 1.0 g / L to 2.0 g / L.

[0099] Process conditions used for purification:

[0100] a. Equilibrate the column with 4 to 5 CV (column volumes) using a mobile phase of 90.0% (A):10.0% (B) at a linear velocity of ≤360 cm / h.

[0101] b. Load the column at a linear velocity of ≤360 cm / h at 4-12 g of insulin glargine per liter of Kromasil.

[0102] c. After loading, wash the column with 4 to 5 CV (column volumes) at a linear velocity of ≤360 cm / h using a mobile phase of 80.0% (A):20.0% (B).

[0103] d. The product was eluted from the column using a linear gradient of 25-29% B over 25 CV at a linear velocity of ≤ 360 cm / h.

[0104] e. During elution, variable volume fractions were collected based on the increase in absorbance (UV) at 280 nm. Collection continued until the UV dropped to the baseline. The volume required to elute the protein from the column was 2.0-3.0 CV.

[0105] f. After fraction collection, adjust the pH of the fractions to 4.0±0.1 using glacial acetic acid. The expected acetic acid consumption is 15.0%-25.0% (V / V).

[0106] g. Analyze and pool the eluted fractions according to their purity levels and store the prepared elution pool at 2-8°C.

[0107] h. After elution to remove residual protein (if any), the column was regenerated by passing 50.0% pure water and 50.0% acetonitrile in an upward flow direction for 3CV at a linear velocity of ≤360 cm / h.

[0108] The results are shown in the tabulated data in Table 2. According to the data, the glycosylated proteins decreased from about 0.5% in the upper sample to 0.00% in the elution pool.

[0109] Table 2: Results of purification step-2 at high pH

[0110]

[0111] Note: The LOQ of the analytical method is 0.04% and the LOD is 0.02%

[0112] A more complete understanding can be obtained by referring to the following examples, which are provided for purposes of illustration only and are not intended to limit the scope of the present disclosure.

[0113] Example 1 shows the purification step-1, which is the reverse phase purification-1 performed at low pH through experiments 1 to 10.

[0114] Experiment 1

[0115] The end product of the enzyme reaction containing human glargine insulin at a concentration of 40 mg / ml was adjusted to 5% acetonitrile, 0.5 M acetic acid, and diluted to 1.31 mg / ml with pure water for use as a loading sample. A pre-loaded reverse phase Kromasil 100-13-C8 column (1*25 cm) was used to purify glargine insulin. The column was initially equilibrated with 90% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 10% acetonitrile for about 5 column volumes. The sample was loaded onto the column at a binding amount of about 8.0 g of human glargine insulin per liter of resin. The loosely bound protein was washed with 85% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 15% acetonitrile for about 4 column volumes. The bound protein was eluted with a 25-32% acetonitrile gradient and 100mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1). According to the increase in absorbance at 280nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 2 column volumes of 50% 1M acetic acid and 50% acetonitrile, then 1 column volume of 30% 1M acetic acid and 70% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMC Co.Ltd was used to analyze the sample loading, elution fractions and elution pool by analytical methods. The results are listed in Table 3 below.

[0116] Table 3 Results of Experiment 1

[0117]

[0118] Experiment 2

[0119] The end product of the enzyme reaction containing human glargine insulin at a concentration of 40 mg / ml was adjusted to 5% acetonitrile, 0.5 M acetic acid, and diluted to 1.37 mg / ml with pure water for use as a loading sample. A pre-loaded reverse phase Kromasil 100-13-C8 column (1*25 cm) was used to purify glargine insulin. The column was initially equilibrated with 90% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 10% acetonitrile for about 5 column volumes. The sample was loaded onto the column at a binding amount of about 8.0 g of human glargine insulin per liter of resin. The loosely bound protein was washed with 85% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 15% acetonitrile for about 4 column volumes. The bound protein was eluted with a 25-32% acetonitrile gradient and 100mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1). Based on the increase in absorbance at 280nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 2 column volumes of 50% 1M acetic acid and 50% acetonitrile, then 1 column volume of 30% 1M acetic acid and 70% acetonitrile. The entire unit operation was performed at a linear flow rate of 360cm / h. The loading sample, elution fractions, and elution pool were analyzed by the YMC C18 method. The results are listed in Table 4 below.

[0120] Table 4 Results of Experiment 2

[0121]

[0122] Experiment 3

[0123] The end product of the enzyme reaction containing human glargine insulin at a concentration of 40 mg / ml was adjusted to 5% acetonitrile, 0.5 M acetic acid, and diluted to 1.42 mg / ml with pure water for use as a loading sample. A pre-loaded reverse phase Kromasil 100-13-C8 column (1*25 cm) was used to purify glargine insulin. Initially, the column was equilibrated with 90% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 10% acetonitrile for about 5 column volumes. The sample was loaded onto the column at a binding amount of about 7.6 g of human glargine insulin per liter of resin. The loosely bound protein was washed with 85% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 15% acetonitrile for about 4 column volumes. The bound protein was eluted with a 25-32% acetonitrile gradient and 100mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1). According to the increase in absorbance at 280nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 2 column volumes of 50% 1M acetic acid and 50% acetonitrile, then 1 column volume of 30% 1M acetic acid and 70% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMC Co.Ltd was used to analyze the sample loading, elution fractions and elution pool by analytical methods. The results are listed in Table 5 below.

[0124] Table 5 Results of Experiment 3

[0125]

[0126] Experiment 4

[0127] The end product of the enzyme reaction containing human glargine insulin at a concentration of 18 mg / ml was adjusted to 5% acetonitrile, 0.5 M acetic acid, and diluted to 1.5 mg / ml with pure water for use as a loading sample. A pre-loaded reverse phase Kromasil 100-13-C8 column (1*25 cm) was used to purify glargine insulin. The column was initially equilibrated with 90% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 10% acetonitrile for about 5 column volumes. The sample was loaded onto the column at a binding amount of about 10.0 g of human glargine insulin per liter of resin. The loosely bound protein was washed with 85% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 15% acetonitrile for about 4 column volumes. The bound protein was eluted with a 25-32% acetonitrile gradient and 100mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1). According to the increase in absorbance at 280nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 2 column volumes of 50% 1M acetic acid and 50% acetonitrile, then 1 column volume of 30% 1M acetic acid and 70% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMC Co.Ltd was used to analyze the sample loading, elution fractions and elution pool by analytical methods. The results are listed in Table 6 below.

[0128] Table 6 Results of Experiment 4

[0129]

[0130]

[0131] Experiment 5

[0132] The end product of the enzyme reaction containing human glargine insulin at a concentration of 18 mg / ml was adjusted to 5% acetonitrile, 0.5 M acetic acid, and diluted to 1.6 mg / ml with pure water for use as a loading sample. A pre-loaded reverse phase Kromasil 100-13-C8 column (2.1*25 cm) was used to purify glargine insulin. Initially, the column was equilibrated with 90% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 10% acetonitrile for about 5 column volumes. The sample was loaded onto the column at a binding amount of about 10 g of human glargine insulin per liter of resin. The loosely bound protein was washed with 85% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 15% acetonitrile for about 4 column volumes. The bound protein was eluted with a 25-32% acetonitrile gradient and 100mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1). According to the increase in absorbance at 280nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 2 column volumes of 50% 1M acetic acid and 50% acetonitrile, then 1 column volume of 30% 1M acetic acid and 70% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMC Co.Ltd was used to analyze the loading sample, elution fraction and elution pool by analytical methods. The results are listed in Table 7 below.

[0133] Table 7 Results of Experiment 5

[0134]

[0135] Experiment 6

[0136] The end product of the enzyme reaction containing human glargine insulin at a concentration of 32 mg / ml was adjusted to 5% acetonitrile, 0.5 M acetic acid, and diluted to 1.5 mg / ml with pure water for use as a loading sample. A pre-loaded reverse phase Kromasil 100-13-C8 column (2.1*25 cm) was used to purify glargine insulin. Initially, the column was equilibrated with 90% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 10% acetonitrile for about 5 column volumes. The sample was loaded onto the column at a binding amount of about 10 g of human glargine insulin per liter of resin. The loosely bound protein was washed with 85% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 15% acetonitrile for about 4 column volumes. The bound protein was eluted with a 25-32% acetonitrile gradient and 100mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1). According to the increase in absorbance at 280nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 2 column volumes of 50% 1M acetic acid and 50% acetonitrile, then 1 column volume of 30% 1M acetic acid and 70% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMC Co.Ltd was used to analyze the loading sample, elution fraction and elution pool by analytical methods. The results are listed in Table 8 below.

[0137] Table 8 Results of Experiment 6

[0138]

[0139] Experiment 7

[0140] The end product of the enzyme reaction containing human glargine insulin at a concentration of 16 mg / ml was adjusted to 5% acetonitrile, 0.5 M acetic acid, and diluted to 1.5 mg / ml with pure water for use as a loading sample. A pre-loaded reverse phase Kromasil 100-13-C8 column (2.1*25 cm) was used to purify glargine insulin. Initially, the column was equilibrated with 90% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 10% acetonitrile for about 5 column volumes. The sample was loaded onto the column at a binding amount of about 10 g of human glargine insulin per liter of resin. The loosely bound protein was washed with 85% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 15% acetonitrile for about 4 column volumes. The bound protein was eluted with a 25-32% acetonitrile gradient and 100mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1). According to the increase in absorbance at 280nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 2 column volumes of 50% 1M acetic acid and 50% acetonitrile, then 1 column volume of 30% 1M acetic acid and 70% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMC Co.Ltd was used to analyze the sample loading, elution fractions and elution pool by analytical methods. The results are listed in Table 9 below.

[0141] Table 9 Results of Experiment 7

[0142]

[0143] Experiment 8

[0144] The end product of the enzyme reaction containing human glargine insulin at a concentration of 36 mg / ml was adjusted to 5% acetonitrile, 0.5 M acetic acid, and diluted to 1.6 mg / ml with pure water for use as a loading sample. A Novosep stainless steel column (5*25 cm) packed with Kromasil 100-13-C8 resin was used to purify glargine insulin. Initially, the column was equilibrated with 90% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 10% acetonitrile for about 5 column volumes. The sample was loaded onto the column at a binding amount of about 9 g of human glargine insulin per liter of resin. The loosely bound protein was washed with 85% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 15% acetonitrile for about 4 column volumes. The bound protein was eluted with a 25-32% acetonitrile gradient and 100mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1). According to the increase in absorbance at 280nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 2 column volumes of 50% 1M acetic acid and 50% acetonitrile, then 1 column volume of 30% 1M acetic acid and 70% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMC Co.Ltd was used to analyze the sample loading, elution fractions and elution pool by analytical methods. The results are listed in Table 10 below.

[0145] Table 10 Results of Experiment 8

[0146]

[0147] Experiment 9

[0148] The end product of the enzyme reaction containing human glargine insulin at a concentration of 32 mg / ml was adjusted to 5% acetonitrile, 0.5 M acetic acid, and diluted to 1.7 mg / ml with pure water for use as a loading sample. A Novosep stainless steel column (5*25 cm) packed with Kromasil 100-13-C8 resin was used to purify glargine insulin. The column was initially equilibrated with 90% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 10% acetonitrile for about 5 column volumes. The sample was loaded onto the column at a binding amount of about 9 g of human glargine insulin per liter of resin. The loosely bound protein was washed with 85% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 15% acetonitrile for about 4 column volumes. The bound protein was eluted with a 25-32% acetonitrile gradient and 100mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1). According to the increase in absorbance at 280nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 2 column volumes of 50% 1M acetic acid and 50% acetonitrile, then 1 column volume of 30% 1M acetic acid and 70% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMC Co.Ltd was used to analyze the sample loading, elution fractions and elution pool by analytical methods. The results are listed in Table 11 below.

[0149] Table 11 Results of Experiment 9

[0150]

[0151] Experiment 10

[0152] The end product of the enzyme reaction containing human glargine insulin at a concentration of 32 mg / ml was adjusted to 5% acetonitrile, 0.5 M acetic acid, and diluted to 1.7 mg / ml with pure water for use as a loading sample. A Novosep stainless steel column (5*25 cm) packed with Kromasil 100-13-C8 resin was used to purify glargine insulin. Initially, the column was equilibrated with 90% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 10% acetonitrile for about 5 column volumes. The sample was loaded onto the column at a binding amount of about 11 g of human glargine insulin per liter of resin. The loosely bound protein was washed with 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1) and 15% acetonitrile for about 4 column volumes. The bound protein was eluted with a 25-32% acetonitrile gradient and 85% 100mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1). According to the increase in absorbance at 280nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 2 column volumes of 50% 1M acetic acid and 50% acetonitrile, then 1 column volume of 30% 1M acetic acid and 70% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMC Co.Ltd was used to analyze the sample loading, elution fractions and elution pool by analytical methods. The results are listed in Table 12 below.

[0153] Table 12 Results of Experiment 10

[0154]

[0155] A summary of experimental data for reverse phase purification of -1 at low pH is detailed in Table 13 below.

[0156] The first reverse phase purification step of human insulin glargine used a low pH buffer combined with the ion pairing agent octane sulfonic acid (OSA) to reduce glycosylation variants by 68-90%, thereby increasing purity from approximately 60% in the upper sample to >97% in the elution pool (EP).

[0157] Table 13 Summary of experimental data of reverse phase purification-1 at low pH

[0158]

[0159] Previous glycosylated impurities were eliminated using a second reverse phase purification step that employed high pH to increase the purity of insulin glargine to >99.5%.

[0160] Example 2 shows the purification step-2, which is a reverse phase purification method carried out at high pH by experiments 11 to 20. Purification-2.

[0161] Experiment 11

[0162] The final product of reverse phase purification-1 containing 2.8 mg / ml human glargine insulin and about 29% acetonitrile was adjusted to pH 8.5 with 2M Tris and diluted with pure water to obtain a final glargine insulin with a concentration of 1.3 mg / ml containing 10% acetonitrile, which was used as a loading sample. A pre-packed reverse phase Kromasil 100-13-C8 column (1*25 cm) was used to purify glargine insulin. Initially, the column was equilibrated with 90% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 10% acetonitrile for 5 column volumes. The sample was loaded onto the column at a binding amount of about 6.5 g human glargine insulin per liter of resin. The loosely bound protein was washed with 80% 100mMTris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 20% acetonitrile for about 4 column volumes. The bound protein was eluted with 25-29% acetonitrile gradient and 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) over 25 column volumes. According to the increase of absorbance at 280nm, multiple fractions of 0.25 column volumes were collected. Tightly bound protein was eluted with 4 column volumes of 50% pure water and 50% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMC Co.Ltd was used to analyze the loading sample, elution fraction and elution pool by analytical methods. The results are listed in Table 14 below.

[0163] Table 14 Results of Experiment 11

[0164]

[0165] Experiment 12

[0166] The final product of reverse phase purification-1 containing 2.8 mg / ml human glargine insulin and about 29% acetonitrile was adjusted to pH 8.5 with 2M Tris and diluted with pure water to obtain a final glargine insulin with a concentration of 1.3 mg / ml containing 10% acetonitrile, which was used as a loading sample. A pre-packed reverse phase Kromasil 100-13-C8 column (1*25 cm) was used to purify glargine insulin. Initially, the column was equilibrated with 90% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 10% acetonitrile for 5 column volumes. The sample was loaded onto the column at a binding amount of about 6.6 g human glargine insulin per liter of resin. The loosely bound protein was washed with 80% 100mMTris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 20% acetonitrile for about 4 column volumes. The bound protein was eluted with 25-29% acetonitrile gradient and 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) over 25 column volumes. According to the increase of absorbance at 280nm, multiple fractions of 0.25 column volumes were collected. Tightly bound protein was eluted with 4 column volumes of 50% pure water and 50% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMC Co.Ltd was used to analyze the loading sample, elution fraction and elution pool by analytical methods. The results are listed in Table 15 below.

[0167] Table 15 Results of Experiment 12

[0168]

[0169] Experiment 13

[0170] The final product of reverse phase purification-1 containing 2.8 mg / ml human glargine insulin and about 29% acetonitrile was adjusted to pH 8.5 with 2M Tris and diluted with pure water to obtain a final glargine insulin with a concentration of 1 mg / ml containing 10% acetonitrile, which was used as a loading sample. A pre-packed reverse phase Kromasil 100-13-C8 column (1*25 cm) was used to purify glargine insulin. Initially, the column was equilibrated with 90% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 10% acetonitrile for 5 column volumes. The sample was loaded onto the column at a binding amount of about 9 g human glargine insulin per liter of resin. The loosely bound protein was washed with 80% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 20% acetonitrile for about 4 column volumes. The bound protein was eluted with 25-29% acetonitrile gradient and 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) over 25 column volumes. According to the increase of absorbance at 280nm, multiple fractions of 0.25 column volumes were collected. Tightly bound protein was eluted with 4 column volumes of 50% pure water and 50% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMC Co. Ltd was used to analyze the sample loading, elution fractions and elution pool by analytical methods. The results are listed in Table 16 below.

[0171] Table 16 Results of Experiment 13

[0172]

[0173] Experiment 14

[0174] The final product of reverse phase purification-1 containing 2.9 mg / ml human glargine insulin and about 29% acetonitrile was adjusted to pH 8.5 with 2M Tris and diluted with pure water to obtain a final glargine insulin with a concentration of 1 mg / ml containing 10% acetonitrile, which was used as a loading sample. A pre-packed reverse phase Kromasil 100-13-C8 column (1*25 cm) was used to purify glargine insulin. Initially, the column was equilibrated with 90% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 10% acetonitrile for 5 column volumes. The sample was loaded onto the column at a binding amount of about 9 g human glargine insulin per liter of resin. The loosely bound protein was washed with 80% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 20% acetonitrile for about 4 column volumes. The bound protein was eluted with 25-29% acetonitrile gradient and 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) over 25 column volumes. According to the increase of absorbance at 280nm, multiple fractions of 0.25 column volumes were collected. Tightly bound protein was eluted with 4 column volumes of 50% pure water and 50% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMCCo.Ltd was used to analyze the loading sample, elution fraction and elution pool by analytical methods. The results are listed in Table 17 below.

[0175] Table 17 Results of Experiment 14

[0176]

[0177] Experiment 15

[0178] The final product of reverse phase purification-1 containing 3.8 mg / ml human glargine insulin and about 29% acetonitrile was adjusted to pH 8.5 with 2M Tris and diluted with pure water to obtain a final glargine insulin with a concentration of 0.5 mg / ml containing 10% acetonitrile, which was used as a loading sample. A pre-packed reverse phase Kromasil 100-13-C8 column (1*25 cm) was used to purify glargine insulin. Initially, the column was equilibrated with 90% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 10% acetonitrile for 5 column volumes. The sample was loaded onto the column at a binding amount of about 7 g human glargine insulin per liter of resin. The loosely bound protein was washed with 80% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 20% acetonitrile for about 4 column volumes. The bound protein was eluted with 25-29% acetonitrile gradient and 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) over 25 column volumes. According to the increase of absorbance at 280nm, multiple fractions of 0.25 column volumes were collected. Tightly bound protein was eluted with 4 column volumes of 50% pure water and 50% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMCCo.Ltd was used to analyze the loading sample, elution fraction and elution pool by analytical methods. The results are listed in Table 18 below.

[0179] Table 18 Results of Experiment 15

[0180]

[0181] Experiment 16

[0182] The final product of reverse phase purification-1 containing 3.8 mg / ml human glargine insulin and about 29% acetonitrile was adjusted to pH 8.5 with 2M Tris and diluted with pure water to obtain a final glargine insulin with a concentration of 1.2 mg / ml containing 10% acetonitrile, which was used as a loading sample. A pre-packed reverse phase Kromasil 100-13-C8 column (1*25 cm) was used to purify glargine insulin. Initially, the column was equilibrated with 90% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 10% acetonitrile for 5 column volumes. The sample was loaded onto the column at a binding amount of about 6 g human glargine insulin per liter of resin. The loosely bound protein was washed with 80% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 20% acetonitrile for about 4 column volumes. The bound protein was eluted with 25-29% acetonitrile gradient and 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) over 25 column volumes. According to the increase of absorbance at 280nm, multiple fractions of 0.25 column volumes were collected. Tightly bound protein was eluted with 4 column volumes of 50% pure water and 50% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMCCo.Ltd was used to analyze the loading sample, elution fraction and elution pool by analytical methods. The results are listed in Table 19 below.

[0183] Table 19 Results of Experiment 16

[0184]

[0185] Experiment 17

[0186] The final product of reverse phase purification-1 containing 3.4 mg / ml human glargine insulin and about 29% acetonitrile was adjusted to pH 8.5 with 2M Tris and diluted with pure water to obtain a final glargine insulin with a concentration of 1.2 mg / ml containing 10% acetonitrile, which was used as a loading sample. A pre-packed reverse phase Kromasil 100-13-C8 column (1*25 cm) was used to purify glargine insulin. Initially, the column was equilibrated with 90% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 10% acetonitrile for 5 column volumes. The sample was loaded onto the column at a binding amount of about 4 g human glargine insulin per liter of resin. The loosely bound protein was washed with 80% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 20% acetonitrile for about 4 column volumes. The bound protein was eluted with 25-29% acetonitrile gradient and 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) over 25 column volumes. According to the increase of absorbance at 280nm, multiple fractions of 0.25 column volumes were collected. Tightly bound protein was eluted with 4 column volumes of 50% pure water and 50% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMCCo.Ltd was used to analyze the loading sample, elution fraction and elution pool by analytical methods. The results are listed in Table 20 below.

[0187] Table 20 Results of Experiment 17

[0188]

[0189] Experiment 18

[0190] The reverse phase purification-1 final product containing 4.2 mg / ml human glargine insulin and about 29% acetonitrile was adjusted to pH 8.5 with 2M Tris and diluted with pure water to obtain a final glargine insulin with a concentration of 1.5 mg / ml containing 10% acetonitrile, which was used as a loading sample. A pre-loaded reverse phase Kromasil 100-13-C8 column (2.1*25 cm) was used to purify glargine insulin. Initially, the column was equilibrated with 90% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 10% acetonitrile for 5 column volumes. The sample was loaded onto the column at a binding amount of about 6 g human glargine insulin per liter of resin. The loosely bound protein was washed with 80% 100mMTris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 20% acetonitrile for about 4 column volumes. The bound protein was eluted with 25-29% acetonitrile gradient and 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1). According to the increase of absorbance at 280nm, multiple 0.25 column volume fractions were collected. Tightly bound protein was eluted with 4 column volumes of 50% pure water and 50% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMC Co.Ltd was used to analyze the sample loading, elution fractions and elution pool by analytical methods. The results are listed in Table 21 below.

[0191] Table 21 Results of Experiment 18

[0192]

[0193] Experiment 19

[0194] The reverse phase purification-1 final product containing 4.2 mg / ml human glargine insulin and 29% acetonitrile was adjusted to pH 8.5 with 2M Tris and diluted with pure water to obtain a final glargine insulin with a concentration of 1.5 mg / ml containing 10% acetonitrile, which was used as a loading sample. A pre-loaded reverse phase Kromasil 100-13-C8 column (2.1*25 cm) was used to purify glargine insulin. Initially, the column was balanced with 90% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 10% acetonitrile for 5 column volumes. The sample was loaded onto the column at a binding amount of about 9 g human glargine insulin per liter of resin. The loosely bound protein was washed with 80% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 20% acetonitrile for about 4 column volumes. The bound protein was eluted with 25-29% acetonitrile gradient and 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) over 25 column volumes. According to the increase of absorbance at 280nm, multiple fractions of 0.25 column volumes were collected. Tightly bound protein was eluted with 4 column volumes of 50% pure water and 50% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMCCo.Ltd was used to analyze the loading sample, elution fraction and elution pool by analytical methods. The results are listed in Table 22 below.

[0195] Table 22 Results of Experiment 19

[0196]

[0197] Experiment 20

[0198] The reverse phase purification-1 final product containing 4.5 mg / ml human glargine insulin and 29% acetonitrile was adjusted to pH 8.5 with 2M Tris and diluted with pure water to obtain a final glargine insulin with a concentration of 1.8 mg / ml containing 10% acetonitrile, which was used as a loading sample. A pre-loaded reverse phase Kromasil 100-13-C8 column (2.1*25 cm) was used to purify glargine insulin. Initially, the column was balanced with 90% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 10% acetonitrile for 5 column volumes. The sample was loaded onto the column at a binding amount of about 8 g human glargine insulin per liter of resin. The loosely bound protein was washed with 80% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 20% acetonitrile for about 4 column volumes. The bound protein was eluted with 25-29% acetonitrile gradient and 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) over 25 column volumes. According to the increase of absorbance at 280nm, multiple fractions of 0.25 column volumes were collected. Tightly bound protein was eluted with 4 column volumes of 50% pure water and 50% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMCCo.Ltd was used to analyze the loading sample, elution fraction and elution pool by analytical methods. The results are listed in Table 23 below.

[0199] Table 23 Results of Experiment 20

[0200]

[0201] A summary of experimental data for reverse phase purification of -2 at high pH is detailed in Table 24 below.

[0202] A second reverse phase purification step of insulin glargine using a high pH buffer combined with the ion pairing agent sodium perchlorate reduced previous glycosylation variants to below detection levels, thereby increasing purity from >96% in the upper sample to >99.5% in the elution pool.

[0203] Table 24 Summary of experimental data of reverse phase purification-2 at high pH

[0204]

[0205]

[0206] Note: The LOQ of the analytical method is 0.04% and the LOD is 0.02%

[0207] The mass characteristics and probable mass based IDs of the final products of the reverse phase purification steps 1 and 2 are detailed in Tables 25 and 26 below, respectively. The mass spectrometry data UV and TIC chromatograms of the final product of reverse phase purification-1 and reverse phase purification-2 are shown in Tables 25 and 26, respectively. Figure 3A and Figure 3B shown.

[0208] Table 25 Mass characteristics and possible mass-based IDs of glycoforms of the reverse phase purification step-1 final product.

[0209]

[0210] Table 26 Mass characteristics and possible IDs based on mass of the reverse phase purification step-2 final products.

[0211] RT RP Purification-2 Elution Pool Levels Observed in C18 HPLC Method Possible IDs based on quality 0.84 BLOD, no charge support, no isotopic distribution Monoglycosylated insulin glargine 0.92 BLOD 6094.8Da 0.99 BLOD, no charge support, no isotopic distribution Monoglycosylated insulin glargine 1.00 BLOD, no charge support, no isotopic distribution Insulin glargine

[0212] Purification of insulin lispro by low pH based RP-HPLC followed by high pH based RP-HPLC

[0213] Purification step at low pH - 1 Figure 1 The details of purification step-1 are as follows

[0214] Stationary Phase Details:

[0215] d. Medium: Daisopak

[0216] e. Height: 25.0±1.0cm

[0217] f. Linear velocity: All chromatographic steps ≤ 360 cm / h

[0218] Mobile phase details:

[0219] d. Mobile phase A: 400 mM trisodium citrate-citrate buffer + 0.1% (w / v) octane sulfonic acid (OSA) + 50 mM GuCl (guanidine chloride), pH 3.85±0.1.

[0220] e. Mobile phase B: 100.0% acetonitrile

[0221] f. Temperature of mobile phase A and B during purification cycle: 25-30°C

[0222] Sample preparation:

[0223] e. Dilute the sample concentrate obtained in the previous step with a mixture of pure water and acetic acid at a ratio of ≥1.0:10.0.

[0224] f. After dilution, filter the sample through a 1.2 μm filter, a 0.45 μm filter, and then a 0.22 μm filter.

[0225] g. The filter for 1.2 μm filtration is made of polypropylene (PP), and the filter for 0.45 / 0.22 μm filtration is made of polyethersulfone (PES) material.

[0226] h. The product concentration in the above samples ranged from 1.0 g / L to 2.0 g / L.

[0227] Process conditions used for purification:

[0228] h. Equilibrate the column with 5 to 10 CV (column volumes) using a mobile phase of 90.0% (A):10.0% (B) at a linear velocity of ≤360 cm / h.

[0229] i. Load the column at a linear velocity of ≤360 cm / h at a volume of 5-7.5 g of insulin lispro per liter of resin.

[0230] j. After loading, wash the column with 4 to 5 CV (column volumes) using a mobile phase of 78.0% (A):22.0% (B) at a linear velocity of ≤360 cm / h.

[0231] k. The product was eluted from the column using a linear gradient of 24-28% B over 20 CV at a linear velocity of ≤ 220 cm / h.

[0232] l. During the elution process, variable volume fractions were collected according to the increase in absorbance (UV) at 280 nm. The collection continued until UV 280 The volume required to elute the protein from the column is 2.0-3.0 CV.

[0233] m. Analyze and pool the eluted fractions according to their purity level and store the prepared elution pool at 2-8°C.

[0234] n. After elution to remove residual protein (if any), the column was regenerated by passing 2CV with 50.0% 1 M acetic acid and 50.0% acetonitrile in an upward flow direction at a linear velocity of ≤360 cm / h, followed by 1V with 30.0% 1 M acetic acid and 70.0% acetonitrile.

[0235] The results are shown in the tabulated data in Table 27. Based on the data, glycosylated proteins were reduced by about 3-fold, from about 4% in the upper sample to about 1.0% in the elution pool (EP).

[0236] Table 27: Results of purification step-1 at low pH

[0237]

[0238] Note: The LOQ of the analytical method is 0.05% and the LOD is 0.03%

[0239] The elution pool from purification step-1 was subjected to purification step-2 to further purify non-glycosylated proteins from glycosylated proteins.

[0240] Purification step at high pH -2 as Figure 2 The details of purification step-2 are as follows

[0241] Stationary Phase Details:

[0242] a. Medium: Phenomenex

[0243] b. Height: 25.0±1.0cm

[0244] c. Linear velocity: All chromatographic steps ≤ 360 cm / h

[0245] Mobile phase details:

[0246] a. Mobile phase A: 100 mM Tris + 50 mM imidazole + 0.2% (w / v) tetrabutylammonium hydrogen sulfate (TBAB), pH 7.5 ± 0.1 (pH adjusted with glacial acetic acid)

[0247] b. Mobile phase B: 100.0% acetonitrile

[0248] c. Temperature of mobile phase A and B during purification cycle: 26-30°C

[0249] Sample preparation:

[0250] a. Dilute the elution pool (EP) obtained in purification step-1 with pure water at a ratio of 1.0:2.0.

[0251] b. Adjust the pH of the diluted sample to 7.5 ± 0.1 using 2.0 M Tris.

[0252] c. After adjusting the pH, the sample was filtered through 1.2 μm, 0.45 μm, and then 0.22 μm.

[0253] d. The filter for 1.2 μm filtration is made of polypropylene (PP), and the filter for 0.45 / 0.22 μm filtration is made of polyethersulfone (PES) material.

[0254] e. The product concentration in the above sample is expected to be in the range of 1.0 g / L to 2.0 g / L.

[0255] Process conditions used for purification:

[0256] a. Equilibrate the column with 5 to 10 CV (column volumes) using a mobile phase of 90.0% (A):10.0% (B) at a linear velocity of ≤360 cm / h.

[0257] b. Load the column at a linear velocity of ≤360 cm / h at a volume of 5-8.5 g of lispro insulin per liter of resin.

[0258] c. After loading, wash the column with 4 to 5 CV (column volumes) at a linear velocity of ≤360 cm / h using a mobile phase of 80.0% (A):20.0% (B).

[0259] d. The product was eluted from the column using a linear gradient of 25-29% B over 13 CV at a linear velocity of ≤ 360 cm / h.

[0260] e. During elution, variable volume fractions were collected based on the increase in absorbance (UV) at 280 nm. Collection continued until the UV dropped to the baseline. The volume required to elute the protein from the column was 2.0-3.0 CV.

[0261] f. Analyze and pool the eluted fractions according to their purity levels and store the prepared elution pool at 2-8°C.

[0262] g. After elution to remove residual protein (if any), the column was regenerated by passing 3CV with 50.0% pure water and 50.0% acetonitrile in the upward flow direction at a linear velocity of ≤360 cm / h, followed by 30% 3M acetic acid and 70.0% acetonitrile.

[0263] The results are shown in the tabulated data in Table 28. Based on the data, glycosylated proteins were reduced from approximately 0.5% in the upper sample to the BLOQ in the elution pool.

[0264] Table 28: Results of purification step-2 at high pH

[0265]

[0266] Note: The LOQ of the analytical method is 0.05%, the LOD is 0.03%; BLOQ: below the level of quantification

[0267] Example 3 shows the purification step-1, which is a reverse phase purification method carried out at low pH by experiments 21 to 24. Purification-1.

[0268] A more complete understanding can be obtained by referring to the following examples, which are provided for illustration purposes only and are not intended to limit the scope of the present disclosure.

[0269] Experiment 21

[0270] The end product of the enzyme reaction containing human lispro insulin at a concentration of 12.7 mg / ml was adjusted to pH < 3.5 with glacial acetic acid and then with 10% acetonitrile, and diluted to 1.16 mg / ml with pure water for use as a loading sample. A pre-loaded reverse phase Daisopak 200-10-C18 column (1*25 cm) was used to purify lispro insulin. Initially, the column was equilibrated with 90% of 400 mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50 mM GuCl buffer (pH 3.85 ± 0.1) and 10% acetonitrile for about 10 column volumes. The sample was loaded onto the column at a binding amount of about 5.6 g of human lispro insulin per liter of resin. Loosely bound proteins were washed with 78% 400mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50mM GuCl buffer (pH 3.85 ± 0.1) and 22% acetonitrile with about 4 column volumes. The bound proteins were eluted with a 24-28% acetonitrile gradient and 400mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50mM GuCl buffer (pH 3.85 ± 0.1) at a flow rate of 220cm / h over 20 column volumes. Multiple 0.25 column volume fractions were collected based on the increase in 280nm absorbance. Tightly bound proteins were eluted with 4 column volumes of 70% 3M acetic acid and 30% acetonitrile. The entire unit operation was performed at a linear flow rate of 360cm / h, except for the elution phase. The load sample, elution fraction and elution pool were analyzed by an analytical method using a C18 column manufactured by Waters Ltd. The results are listed in Table 29 below.

[0271] Table 29 Results of Experiment 19

[0272]

[0273] Experiment 22

[0274] The end product of the enzyme reaction containing human lispro insulin at a concentration of 12.7 mg / ml was adjusted to pH < 3.5 with glacial acetic acid and then with 10% acetonitrile, and diluted to 1.16 mg / ml with pure water for use as a loading sample. A pre-loaded reverse phase Daisopak 200-10-C18 column (1*25 cm) was used to purify lispro insulin. Initially, the column was equilibrated with 90% 400 mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50 mM GuCl buffer (pH 3.85 ± 0.1) and 10% acetonitrile for about 10 column volumes. The sample was loaded onto the column at a binding amount of about 7.4 g human lispro insulin per liter of resin. Loosely bound proteins were washed with 78% 400mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50mM GuCl buffer (pH 3.85 ± 0.1) and 22% acetonitrile with about 4 column volumes. The bound proteins were eluted with a 24-28% acetonitrile gradient and 400mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50mM GuCl buffer (pH 3.85 ± 0.1) at a flow rate of 220cm / h over 20 column volumes. Multiple 0.25 column volume fractions were collected based on the increase in 280nm absorbance. Tightly bound proteins were eluted with 4 column volumes of 70% 3M acetic acid and 30% acetonitrile. The entire unit operation was performed at a linear flow rate of 360cm / h, except for the elution phase. The load sample, elution fraction and elution pool were analyzed by an analytical method using a C18 column manufactured by Waters Ltd. The results are listed in Table 30 below.

[0275] Table 30 Results of Experiment 22

[0276]

[0277] Experiment 23

[0278] The end product of the enzyme reaction containing human lispro insulin at a concentration of 10.8 mg / ml was adjusted to pH < 3.5 with glacial acetic acid and then with 10% acetonitrile, and diluted to 1.42 mg / ml with pure water for use as a loading sample. A pre-loaded reverse phase Daisopak 200-10-C18 column (1*25 cm) was used to purify lispro insulin. Initially, the column was equilibrated with 90% of 400 mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50 mM GuCl buffer (pH 3.85 ± 0.1) and 10% acetonitrile for about 10 column volumes. The sample was loaded onto the column at a binding amount of about 6.7 g of human lispro insulin per liter of resin. Loosely bound proteins were washed with 78% 400mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50mM GuCl buffer (pH 3.85 ± 0.1) and 22% acetonitrile with about 4 column volumes. The bound proteins were eluted with a 24-28% acetonitrile gradient and 400mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50mM GuCl buffer (pH 3.85 ± 0.1) at a flow rate of 220cm / h over 20 column volumes. Multiple 0.25 column volume fractions were collected based on the increase in 280nm absorbance. Tightly bound proteins were eluted with 4 column volumes of 70% 3M acetic acid and 30% acetonitrile. The entire unit operation was performed at a linear flow rate of 360cm / h, except for the elution phase. The load sample, elution fraction and elution pool were analyzed by an analytical method using a C18 column manufactured by Waters Ltd. The results are listed in Table 31 below.

[0279] Table 31 Results of Experiment 23

[0280]

[0281] Experiment 24

[0282] The end product of the enzyme reaction containing human lispro insulin at a concentration of 10.8 mg / ml was adjusted to pH < 3.5 with glacial acetic acid and then with 10% acetonitrile, and diluted to 1.06 mg / ml with pure water for use as a loading sample. A pre-loaded reverse phase Daisopak 200-10-C18 column (1*25 cm) was used to purify lispro insulin. Initially, the column was equilibrated with 90% of 400 mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50 mM GuCl buffer (pH 3.85 ± 0.1) and 10% acetonitrile for about 8 column volumes. The sample was loaded onto the column at a binding amount of about 6.7 g of human lispro insulin per liter of resin. Loosely bound proteins were washed with 78% 400mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50mM GuCl buffer (pH 3.85 ± 0.1) and 22% acetonitrile with about 4 column volumes. The bound proteins were eluted with a 24-28% acetonitrile gradient and 400mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50mM GuCl buffer (pH 3.85 ± 0.1) at a flow rate of 220cm / h over 20 column volumes. Multiple 0.25 column volume fractions were collected based on the increase in 280nm absorbance. Tightly bound proteins were eluted with 4 column volumes of 70% 3M acetic acid and 30% acetonitrile. The entire unit operation was performed at a linear flow rate of 360cm / h, except for the elution phase. The load sample, elution fraction and elution pool were analyzed by an analytical method using a C18 column manufactured by Waters Ltd. The results are listed in Table 32 below.

[0283] Table 32 Results of Experiment 24

[0284]

[0285] A summary of experimental data for reverse phase purification of -1 at low pH is detailed in Table 33 below.

[0286] Table 33 Summary of experimental data of reverse phase purification-1 at low pH

[0287]

[0288] Reverse phase purification step-1 of human insulin lispro uses a low pH buffer in combination with the ion pairing agent octane sulfonic acid (OSA) to reduce glycosylation variants by 70-95%, thereby increasing purity from approximately 58% in the upper sample to >83% in the elution pool.

[0289] Example 4 shows the reverse phase purification step-2 performed at high pH through experiments 25 to 26.

[0290] Experiment 25

[0291] The reverse phase purification-1 final product containing 2.4 mg / ml of human insulin lispro and about 26.4% acetonitrile was adjusted to pH 7.5 with 2M Tris and diluted with pure water to obtain a final insulin lispro of 0.82 mg / ml containing 10% acetonitrile for loading. A pre-loaded reverse phase Phenomenex 100-10-C8 column (1*25 cm) was used to purify insulin lispro. Initially, the column was balanced with 90% of 100 mM Tris+200 mM imidazole+0.2% (w / v) TBAB buffer (pH 7.5±0.1) and 10% acetonitrile for 10 column volumes. The sample was loaded onto the column at a binding amount of about 8 g of human insulin lispro per liter of resin. Loosely bound proteins were washed with 80% 100 mM Tris + 200 mM imidazole + 0.2% (w / v) TBAB buffer (pH 7.5 ± 0.1) and 20% acetonitrile for about 4 column volumes. Bound proteins were eluted with a 24-29% acetonitrile gradient with 100 mM Tris + 200 mM imidazole + 0.2% (w / v) TBAB buffer (pH 7.5 ± 0.1) over 13 column volumes. Based on the increase in absorbance at 280 nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 50% 100 mM Tris + 200 mM imidazole + 0.2% (w / v) TBAB buffer (pH 7.5 ± 0.1) and 50% acetonitrile for 2 column volumes, followed by 30% 3M acetic acid and 70% acetonitrile for 2 column volumes. The entire unit operation was carried out at a linear flow rate of 360 cm / h. The load sample, elution fraction and elution pool were analyzed by analytical methods using a C18 column manufactured by Waters. Ltd. The results are listed in Table 34 below.

[0292] Table 34 Results of Experiment 25

[0293]

[0294] Experiment 26

[0295] The reverse phase purification-1 final product containing 1.3 mg / ml human insulin lispro and about 26.4% acetonitrile was adjusted to pH 7.5 with 2M Tris and diluted with pure water to obtain a final insulin lispro of 0.5 mg / ml containing 10% acetonitrile for loading. A pre-loaded reverse phase Phenomenex 100-10-C8 column (1*25 cm) was used to purify insulin lispro. Initially, the column was balanced with 90% 100 mM Tris + 200 mM imidazole + 0.2% (w / v) TBAB buffer (pH 7.5 ± 0.1) and 10% acetonitrile for 5 column volumes. The sample was loaded onto the column at a binding amount of about 5.6 g human insulin lispro per liter of resin. Loosely bound proteins were washed with 80% 100 mM Tris + 200 mM imidazole + 0.2% (w / v) TBAB buffer (pH 7.5 ± 0.1) and 20% acetonitrile for about 4 column volumes. Bound proteins were eluted with a 24-29% acetonitrile gradient with 100 mM Tris + 200 mM imidazole + 0.2% (w / v) TBAB buffer (pH 7.5 ± 0.1) over 13 column volumes. Based on the increase in absorbance at 280 nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 50% 100 mM Tris + 200 mM imidazole + 0.2% (w / v) TBAB buffer (pH 7.5 ± 0.1) and 50% acetonitrile for 2 column volumes, followed by 30% 3M acetic acid and 70% acetonitrile for 2 column volumes. The entire unit operation was carried out at a linear flow rate of 360 cm / h. The load sample, elution fraction and elution pool were analyzed by analytical methods using a C18 column manufactured by Waters. Ltd. The results are listed in Table 35 below.

[0296] Table 35 Results of Experiment 26

[0297]

[0298] A summary of experimental data for reverse phase purification at high pH-2 (Experiments 25-26) is detailed in Table 36 below.

[0299] Table 36 Summary of experimental data for reverse phase purification-2 at high pH

[0300]

[0301] Note: The LOQ of the analytical method is 0.04% and the LOD is 0.02%

[0302] The second reverse phase purification step-2 of insulin lispro used a high pH buffer in combination with the ion pairing agent tetrabutylammonium hydrogen sulfate to reduce previous glycosylation variants to below detection levels, thereby increasing purity from >84% in the upper sample to >95% in the elution pool.

[0303] The mass characteristics and possible IDs based on mass of the final product of the reverse phase purification step-1 are detailed in Tables 37 (Experiment 21) and 38 (Experiment 22), and the mass characteristics and possible IDs based on mass of the final product of the reverse phase purification step-2 are detailed in Tables 39 (Experiment 25) and 40 (Experiment 26). The mass spectrometry data UV and TIC chromatograms of the final product of the reverse phase purification-1 are shown in Tables 37 (Experiment 21) and 38 (Experiment 22), respectively. Figure 4A (Experiment 21) and Figure 4B As shown in (Experiment 22), the mass spectrometry data UV and TIC chromatograms of the final product of reverse phase purification-2 are as follows Figure 5A (Experiment 25) and Figure 5B (Experiment 26) shown.

[0304] Table 37 Reverse phase purification step-1 final product (experiment 21, Figure 4A Mass characteristics and possible mass-based IDs of glycoforms of the same species (shown).

[0305]

[0306] Table 38 Reverse phase purification step-1 final product (experiment 22, Figure 4B Mass characteristics and possible mass-based IDs of glycoforms of the same species (shown).

[0307]

[0308] Table 39 Mass characteristics and possible mass-based IDs of glycoforms of the reverse phase purification step-2 final product (Experiment 25, shown in Figure 5).

[0309]

[0310] Table 40 Mass characteristics and possible mass-based IDs of glycoforms of the reverse phase purification step-2 final product (Experiment 26).

[0311]

[0312] Purification of insulin glargine by high pH based RP-HPLC followed by low pH based RP-HPLC

[0313] Purification step at high pH - 1 Figure 1 The details of purification step-1 are as follows

[0314] Stationary Phase Details:

[0315] a. Medium: Kromasil

[0316] b. Height: 25.0±1.0cm

[0317] c. Linear velocity: All chromatographic steps ≤ 360 cm / h

[0318] Mobile phase details:

[0319] a. Mobile phase A: 100mM Tris + 50mM sodium perchlorate, pH 8.5±0.1 (pH adjusted with glacial acetic acid)

[0320] b. Mobile phase B: 100.0% acetonitrile

[0321] c. Temperature of mobile phase A and B during purification cycle: 20-30°C

[0322] Sample preparation:

[0323] a. Dilute the loading concentrate obtained in the previous step with a mixture of 0.5 M Tris and 0.7 M arginine at a ratio of ≥1.0:10.0.

[0324] b. Adjust the pH of the diluted sample to 8.5 ± 0.1 using acetic acid.

[0325] c. After adjusting the pH, the sample was filtered through 1.2 μm, 0.45 μm, and then 0.22 μm.

[0326] d. The filter for 1.2 μm filtration is made of polypropylene (PP), and the filter for 0.45 / 0.22 μm filtration is made of polyethersulfone (PES) material.

[0327] e. The product concentration in the above sample is expected to be in the range of 1.0 g / L to 2.0 g / L.

[0328] Process conditions used for purification:

[0329] a. Equilibrate the column with 5 to 10 CV (column volumes) using a mobile phase of 90.0% (A):10.0% (B) at a linear velocity of ≤360 cm / h.

[0330] b. Load the column at a linear velocity of ≤360 cm / h at 9-10 g of insulin glargine per liter of resin.

[0331] c. After loading, wash the column with 4 to 5 CV (column volumes) at a linear velocity of ≤360 cm / h using a mobile phase of 80.0% (A):20.0% (B).

[0332] d. The product was eluted from the column using a linear gradient of 25-30% B over 25 CV at a linear velocity of ≤ 360 cm / h.

[0333] e. During elution, variable volume fractions were collected based on the increase in absorbance (UV) at 280 nm. Collection continued until the UV dropped to the baseline. The volume required to elute the protein from the column was 2.0-3.0 CV.

[0334] f. After fraction collection, adjust the pH of the fractions to 4.0±0.1 using glacial acetic acid. The expected acetic acid consumption is 15.0%-25.0% (V / V).

[0335] g. Analyze and pool the eluted fractions according to their purity levels and store the prepared elution pool at 2-8°C.

[0336] h. After elution to remove residual protein (if any), the column was regenerated by passing 50.0% pure water and 50.0% acetonitrile in an upward flow direction for 3CV at a linear velocity of ≤360 cm / h.

[0337] The results are shown in the tabulated data in Table 41. Based on the data, glycosylated proteins were reduced by about 9 times, from about 4.0% in the upper sample to about 0.25% in the elution pool.

[0338] Table 41: Results of purification step-1 at high pH

[0339]

[0340] Note: The LOQ of the analytical method is 0.04% and the LOD is 0.02%

[0341] The elution pool from purification step-1 was subjected to purification step-2 to further purify non-glycosylated proteins from glycosylated proteins.

[0342] Purification step at low pH -2 Figure 2 The details of purification step-2 are as follows

[0343] Stationary Phase Details:

[0344] a. Medium: Kromasil

[0345] b. Height: 25.0±1.0cm

[0346] c. Linear velocity: All chromatographic steps ≤ 360 cm / h

[0347] Mobile phase details:

[0348] a. Mobile phase A: 100 mM sodium acetate + 0.05% (w / v) octane sulfonic acid (OSA), pH 3.7 ± 0.1 (pH adjusted with glacial acetic acid)

[0349] b. Mobile phase B: 100.0% acetonitrile

[0350] c. Temperature of mobile phase A and B during purification cycle: 26-30°C

[0351] Sample preparation:

[0352] a. Adjust the pH of the elution pool obtained in the previous step to 3.7 ± 0.1 with glacial acetic acid and dilute with purified water at a ratio of ≥ 1.0:1.5.

[0353] b. After dilution, filter the sample through a 1.2 μm filter, a 0.45 μm filter, and then a 0.22 μm filter.

[0354] c. The filter for 1.2 μm filtration is made of polypropylene (PP), and the filter for 0.45 / 0.22 μm filtration is made of polyethersulfone (PES) material.

[0355] d. The product concentration in the above sample is expected to be in the range of 1.0 g / L to 2.0 g / L.

[0356] Process conditions used for purification:

[0357] a. Equilibrate the column with a mobile phase of 90.0% (A):10.0% (B) for 10 CV (column volume) at a linear velocity of ≤360 cm / h.

[0358] b. Load the column at a linear velocity of ≤360 cm / h at 6-7.5 g of insulin glargine per liter of resin.

[0359] c. After loading, wash the column with 4 to 5 CV (column volumes) at a linear velocity of ≤360 cm / h using a mobile phase of 85.0% (A):15.0% (B).

[0360] d. The product was eluted from the column using a linear gradient of 24-31% B over 25 CV at a linear velocity of ≤ 360 cm / h.

[0361] e. During the elution process, variable volume fractions were collected according to the increase in absorbance (UV) at 280 nm. The collection continued until UV 280 The volume required to elute the protein from the column is 2.0-3.0 CV.

[0362] f. Analyze and pool the eluted fractions according to their purity levels and store the prepared elution pool at 2-8°C.

[0363] g. After elution to remove residual protein (if any), the column was regenerated by passing 2CV with 50.0% 1 M acetic acid and 50.0% acetonitrile in an upward flow direction at a linear velocity of ≤360 cm / h, followed by 1V with 30.0% 1 M acetic acid and 70.0% acetonitrile.

[0364] The results are shown in the tabulated data in Table 42. According to the data, the glycosylated proteins were reduced from about 0.3% in the upper sample to the BLOQ level in the elution pool.

[0365] Table 42: Results of purification step-2 at low pH

[0366]

[0367] Note: The LOQ of the analytical method is 0.04% and the LOD is 0.02%

[0368] The elution pool from purification step-1 was subjected to purification step-2 to further purify non-glycosylated proteins from glycosylated proteins.

[0369] Example 5 shows the reverse phase purification of -1 performed at high pH by Experiments 27 to 29.

[0370] Experiment 27

[0371] The end product of the enzyme reaction containing human glargine insulin at a concentration of 10 mg / ml was adjusted to pH 8.5, adjusted to 10% acetonitrile, and diluted to 1.91 mg / ml with 0.5M Tris+0.7M arginine for use as a loading sample. A pre-packed reverse phase Kromasil 100-13-C8 column (1*25 cm) was used to purify glargine insulin. Initially, the column was equilibrated with 90% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 10% acetonitrile for 10 column volumes. The sample was loaded onto the column at a binding amount of about 9 g of human glargine insulin per liter of resin. The loosely bound protein was washed with 80% 100mM Tris+50mM sodium perchlorate buffer (pH8.5±0.1) and 20% acetonitrile for about 4 column volumes. The bound protein was eluted with a 25-30% acetonitrile gradient and 100mM sodium acetate + 50mM sodium perchlorate buffer (pH 8.5 ± 0.1). According to the increase in absorbance at 280nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 4 column volumes of 50% pure water and 50% acetonitrile. The entire unit operation was performed at a linear flow rate of 360cm / h. The C18 column manufactured by YMC Co.Ltd was used to analyze the sample loading, elution fractions and elution pool by analytical methods. The results are listed in Table 43:

[0372] *When purification-1 was performed at high pH, ​​L-arginine and Tris were used to prepare the top sample to ensure better solubility. The arginine used for the top sample preparation did not have any effect on the purification mechanism. All other aspects of the step remained unchanged, including the buffer composition and other input parameters.

[0373] Table 43 Results of Experiment 27

[0374]

[0375] Experiment 28

[0376] The end product of the enzyme reaction containing human glargine insulin at a concentration of 10 mg / ml was adjusted to pH 8.5, adjusted to 10% acetonitrile, and diluted to 1.91 mg / ml with 0.5M Tris+0.7M arginine* for use as a loading sample. A pre-packed reverse phase Kromasil 100-13-C8 column (1*25cm) was used to purify glargine insulin. The column was initially equilibrated with 90% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 10% acetonitrile for 5 column volumes. The sample was loaded onto the column at a binding amount of about 9.3g of human glargine insulin per liter of resin. The loosely bound protein was washed with 80% 100mM Tris+50mM sodium perchlorate buffer (pH8.5±0.1) and 20% acetonitrile for about 4 column volumes. The bound protein was eluted with a 25-30% acetonitrile gradient and 100mM sodium acetate + 50mM sodium perchlorate buffer (pH 8.5 ± 0.1). According to the increase in absorbance at 280nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 4 column volumes of 50% pure water and 50% acetonitrile. The entire unit operation was performed at a linear flow rate of 360cm / h. The C18 column manufactured by YMC Co.Ltd was used to analyze the sample loading, elution fractions and elution pool by analytical methods. The results are listed in Table 44:

[0377] *When purification-1 was performed at high pH, ​​L-arginine and Tris were used to prepare the top sample to ensure better solubility. The arginine used for the top sample preparation did not have any effect on the purification mechanism. All other aspects of the step remained unchanged, including the buffer composition and other input parameters.

[0378] Table 44 Results of Experiment 28

[0379]

[0380] Experiment 29

[0381] The end product of the enzyme reaction containing human glargine insulin at a concentration of 10 mg / ml was adjusted to pH 8.5, adjusted to 10% acetonitrile, and diluted to 1.91 mg / ml with 0.5M Tris+0.7M arginine* for use as a loading sample. A pre-packed reverse phase Kromasil 100-13-C8 column (2.1*25 cm) was used to purify glargine insulin. Initially, the column was equilibrated with 90% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 10% acetonitrile for 5 column volumes. The sample was loaded onto the column at a binding amount of about 9.75 g of human glargine insulin per liter of resin. The loosely bound protein was washed with 80% 100mM Tris+50mM sodium perchlorate buffer (pH 8.5±0.1) and 20% acetonitrile for about 4 column volumes. The bound protein was eluted with a 25-30% acetonitrile gradient and 100mM sodium acetate + 50mM sodium perchlorate buffer (pH 8.5 ± 0.1). According to the increase in absorbance at 280nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 4 column volumes of 50% pure water and 50% acetonitrile. The entire unit operation was performed at a linear flow rate of 360cm / h. The C18 column manufactured by YMC Co.Ltd was used to analyze the sample loading, elution fractions and elution pool by analytical methods. The results are listed in Table 45:

[0382] *When purification-1 was performed at high pH, ​​L-arginine and Tris were used to prepare the top sample to ensure better solubility. The arginine used for the top sample preparation did not have any effect on the purification mechanism. All other aspects of the step remained unchanged, including the buffer composition and other input parameters.

[0383] Table 45 Results of Experiment 29

[0384]

[0385] A summary of experimental data for reverse phase purification of -1 at high pH is detailed in Table 46 below.

[0386] Table 46 Summary of experimental data of reverse phase purification-1 at high pH

[0387]

[0388] The first reverse phase purification step of human insulin glargine used a high pH buffer combined with the ion pairing agent sodium perchlorate to reduce glycosylation variants by 90-93%, thereby increasing purity from approximately 56% in the upper sample to >85% in the elution pool.

[0389] Previous glycosylated impurities were eliminated using a second reverse phase purification step which employed low pH to increase the purity of insulin glargine to ≥99.5%.

[0390] Example 6 shows the reverse phase purification step-2 performed at low pH through experiments 30-32.

[0391] Experiment 30

[0392] The final product of reverse phase purification-1 containing about 5.06 mg / ml of human glargine insulin and about 28.8% acetonitrile was adjusted to pH 3.7 with acetic acid and diluted with pure water to obtain a final glargine insulin with a concentration of about 1.69 mg / ml containing 10% acetonitrile, which was used as a loading sample. A pre-packed reverse phase Kromasil 100-13-C8 column (1*25 cm) was used to purify glargine insulin. The column was initially equilibrated with 90% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7±0.1) and 10% acetonitrile for about 10 column volumes. The sample was loaded onto the column at a binding amount of about 6.9 g of human glargine insulin per liter of resin. The loosely bound protein was washed with 85% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7±0.1) and 15% acetonitrile for about 4 column volumes. The bound protein was eluted with a 24-31% acetonitrile gradient and 100mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1). Based on the increase in absorbance at 280nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 2 column volumes of 50% 1M acetic acid and 50% acetonitrile, then 1 column volume of 30% 1M acetic acid and 70% acetonitrile. The entire unit operation was performed at a linear flow rate of 360cm / h. The C18 column manufactured by YMC Co.Ltd was used to analyze the sample loading, elution fractions, and elution pool by analytical methods. The results are listed in Table 47 below.

[0393] Table 47 Results of Experiment 30

[0394]

[0395] Experiment 31

[0396] The final product of reverse phase purification-1 containing about 5.49 mg / ml of human glargine insulin and about 28.8% acetonitrile was adjusted to pH 3.7 with acetic acid and diluted with pure water to obtain a final glargine insulin with a concentration of about 1.96 mg / ml containing 10% acetonitrile, which was used as a loading sample. A pre-packed reverse phase Kromasil 100-13-C8 column (1*25 cm) was used to purify glargine insulin. The column was initially equilibrated with 90% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7±0.1) and 10% acetonitrile for about 10 column volumes. The sample was loaded onto the column at a binding amount of about 7.0 g of human glargine insulin per liter of resin. The loosely bound protein was washed with 85% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7±0.1) and 15% acetonitrile for about 4 column volumes. The bound protein was eluted with a 24-31% acetonitrile gradient and 100mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1). According to the increase in absorbance at 280nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 2 column volumes of 50% 1M acetic acid and 50% acetonitrile, then 1 column volume of 30% 1M acetic acid and 70% acetonitrile. The entire unit operation was carried out at a linear flow rate of 360cm / h. The C18 column manufactured by YMC Co.Ltd was used to analyze the sample loading, elution fractions and elution pool by analytical methods. The results are listed in Table 48 below.

[0397] Table 48 Results of Experiment 31

[0398]

[0399] Experiment 32

[0400] The reverse phase purification-1 final product containing about 4.3 mg / ml of human glargine insulin and about 28.6% acetonitrile was adjusted to pH 3.7 with acetic acid and diluted with pure water to obtain a final glargine insulin with a concentration of about 1.5 mg / ml containing 10% acetonitrile, which was used as a loading sample. A pre-packed reverse phase Kromasil 100-13-C8 column (1*25 cm) was used to purify glargine insulin. Initially, the column was equilibrated with 90% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7±0.1) and 10% acetonitrile for about 5 column volumes. The sample was loaded onto the column at a binding amount of about 7.5 g of human glargine insulin per liter of resin. The loosely bound protein was washed with 85% of 100 mM sodium acetate + 0.05% OSA buffer (pH 3.7±0.1) and 15% acetonitrile for about 4 column volumes. The bound protein was eluted with a 24-31% acetonitrile gradient and 100mM sodium acetate + 0.05% OSA buffer (pH 3.7 ± 0.1). Based on the increase in absorbance at 280nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 2 column volumes of 50% 1M acetic acid and 50% acetonitrile, then 1 column volume of 30% 1M acetic acid and 70% acetonitrile. The entire unit operation was performed at a linear flow rate of 360cm / h. The C18 column manufactured by YMC Co.Ltd was used to analyze the sample loading, elution fractions, and elution pool by analytical methods. The results are listed in Table 49 below.

[0401] Table 49 Results of Experiment 32

[0402]

[0403] A summary of experimental data for reverse phase purification of -1 at low pH is detailed in Table 50 below.

[0404] Table 50 Experimental data summary of reverse phase purification-2 at low pH

[0405]

[0406] Note: The LOQ of the analytical method is 0.04% and the LOD is 0.02%

[0407] The mass characteristics and possible mass-based IDs of the final products of the reverse phase purification steps 1 and 2 are detailed in Tables 51 and 52 below, respectively. The mass spectrometry data UV and TIC chromatograms of the final product of reverse phase purification-1 and reverse phase purification-2 are shown in Tables 51 and 52, respectively. Fig. 6A and Figure 6B shown.

[0408] Table 51 Mass characteristics and possible mass-based IDs of glycoforms of the reverse phase purification step-1 final product (Experiment 27).

[0409]

[0410] Table 52 Mass characteristics and possible mass-based IDs of glycoforms of the reverse phase purification step-2 final product (Experiment 30).

[0411]

[0412] Purification of insulin lispro by high pH based RP-HPLC followed by low pH based RP-HPLC

[0413] Purification step at high pH - 1 Figure 1 The details of purification step-1 are as follows

[0414] Stationary Phase Details:

[0415] a. Medium: Daisopak

[0416] b. Height: 25.0±1.0cm

[0417] c. Linear velocity: All chromatographic steps ≤ 360 cm / h

[0418] Mobile phase details:

[0419] a. Mobile phase A: 100 mM Tris + 200 mM imidazole + 0.2% tetrabutylammonium hydrogen sulfate (TBAB), pH 7.5 ± 0.1 (pH adjusted with glacial acetic acid)

[0420] b. Mobile phase B: 100.0% acetonitrile

[0421] c. Temperature of mobile phase A and B during purification cycle: 20-30°C

[0422] Sample preparation:

[0423] a. Adjust the pH of the sample concentrate obtained in the previous step to 7.5 ± 0.1 and dilute it with purified water at a ratio of ≥ 1.0:10.0.

[0424] b. Filter the sample through 1.2 μm, 0.45 μm, and then 0.22 μm.

[0425] c. The filter for 1.2 μm filtration is made of polypropylene (PP), and the filter for 0.45 / 0.22 μm filtration is made of polyethersulfone (PES) material.

[0426] d. The product concentration in the above sample is expected to be in the range of 1.0 g / L to 2.0 g / L.

[0427] Process conditions used for purification:

[0428] a. Equilibrate the column with 5 to 10 CV (column volumes) using a mobile phase of 90.0% (A):10.0% (B) at a linear velocity of ≤360 cm / h.

[0429] b. Load the column at a linear velocity of ≤360 cm / h at 9-10 g of insulin lispro per liter of resin.

[0430] c. After loading, wash the column with 4 to 5 CV (column volumes) at a linear velocity of ≤360 cm / h using a mobile phase of 80.0% (A):15.0% (B).

[0431] d. The product was eluted from the column using a linear gradient of 24-29% B over 13 CV at a linear velocity of ≤ 360 cm / h.

[0432] e. During elution, variable volume fractions were collected based on the increase in absorbance (UV) at 280 nm. Collection continued until the UV dropped to the baseline. The volume required to elute the protein from the column was 2.0-3.0 CV.

[0433] f. Analyze and pool the eluted fractions according to their purity levels and store the prepared elution pool at 2-8°C.

[0434] g. After elution to remove residual protein (if any), the column was regenerated using 50.0% mobile phase A and 50.0% acetonitrile in an upward flow direction for 2CV at a linear velocity of ≤360 cm / h, followed by 30% 3M acetic acid and 70.0% acetonitrile.

[0435] The results are shown in the tabulated data in Table 53. Based on the data, glycosylated proteins were reduced by about 3-fold, from about 4.0% in the upper sample to about 1.0% in the elution pool (EP).

[0436] Table 53: Results of purification step-1 at high pH

[0437]

[0438] Note: The LOQ of the analytical method is 0.05% and the LOD is 0.03%

[0439] The elution pool from purification step-1 was subjected to purification step-2 to further purify non-glycosylated proteins from glycosylated proteins.

[0440] Purification step at low pH -2 Figure 2 The details of purification step-2 are as follows

[0441] Stationary Phase Details:

[0442] a. Medium: Daisopak

[0443] b. Height: 25.0±1.0cm

[0444] c. Linear velocity: All chromatographic steps ≤ 360 cm / h

[0445] Mobile phase details:

[0446] a. Mobile phase A: 400 mM trisodium citrate-citrate buffer + 0.1% (w / v) octane sulfonic acid (OSA) + 50 mM GuCl, pH 3.85±0.1.

[0447] b. Mobile phase B: 100.0% acetonitrile

[0448] c. Temperature of mobile phase A and B during purification cycle: 26-30°C

[0449] Sample preparation:

[0450] a. Adjust the pH of the elution pool obtained in the previous step to 4.0 ± 0.1 with glacial acetic acid and dilute with purified water at a ratio of ≥ 1.0:1.5.

[0451] b. After dilution, the sample was filtered through a 1.2 μm filter, a 0.45 μm filter, and then a 0.22 μm filter.

[0452] c. The filter for 1.2 μm filtration is made of polypropylene (PP), and the filter for 0.45 / 0.22 μm filtration is made of polyethersulfone (PES) material.

[0453] d. The product concentration in the above sample is expected to be in the range of 0.8 g / L to 1.0 g / L.

[0454] Process conditions used for purification:

[0455] a. Equilibrate the column with 5 to 10 CV (column volumes) using a mobile phase of 90.0% (A):10.0% (B) at a linear velocity of ≤360 cm / h.

[0456] b. Load the column at a linear velocity of ≤360 cm / h at 7-8 g of insulin lispro per liter of resin.

[0457] c. After loading, wash the column with 4 to 5 CV (column volumes) using a mobile phase of 78.0% (A):22.0% (B) at a linear velocity of ≤360 cm / h.

[0458] d. The product was eluted from the column using a linear gradient of 24-28% B over 20 CV at a linear velocity of ≤ 220 cm / h.

[0459] e. During the elution process, variable volume fractions were collected according to the increase in absorbance (UV) at 280 nm. The collection continued until UV 280 The volume required to elute the protein from the column is 2.0-3.0 CV.

[0460] f. Analyze and pool the eluted fractions according to their purity levels and store the prepared elution pool at 2-8°C.

[0461] g. After elution to remove residual protein (if any), the column was regenerated by passing 70.0% 1 M acetic acid and 30.0% acetonitrile in an upward flow direction for 4CV at a linear velocity of ≤360 cm / h.

[0462] The results are shown in the tabulated data in Table 54. According to the data, the glycosylated proteins were reduced from about 0.3% in the upper sample to the BLOQ level in the elution pool.

[0463] Table 54: Results of purification step-2 at low pH

[0464]

[0465] Note: The LOQ of the analytical method is 0.05% and the LOD is 0.03%

[0466] Example 7 shows the reverse phase purification step-1 performed at high pH through experiments 33 to 35.

[0467] Experiment 33

[0468] The end product of the enzyme reaction containing human lispro insulin at a concentration of 13 mg / ml was adjusted to pH 7.5 and 10% acetonitrile, then diluted to 1.45 mg / ml with pure water and used as a loading sample. A pre-packed reverse phase Daisopak 200-10-C18 column (2*25 cm) was used to purify lispro insulin. The column was initially equilibrated with 90% 100 mM Tris + 200 mM imidazole + 0.2% (w / v) TBAB buffer (pH 7.5 ± 0.1) and 10% acetonitrile for 10 column volumes. The sample was loaded onto the column at a binding amount of about 7.3 g of human lispro insulin per liter of resin. The loosely bound protein was washed with 75% 100 mM Tris + 200 mM imidazole + 0.2% (w / v) TBAB buffer (pH 7.5 ± 0.1) and 15% acetonitrile for about 4 column volumes. The bound protein was eluted with a 24-29% acetonitrile gradient and 100mM Tris+200mM imidazole+0.2% (w / v) TBAB buffer (pH 7.5 ± 0.1) over 13 column volumes. Based on the increase in absorbance at 280nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 2 column volumes of 50% 100mM Tris+200mM imidazole+0.2% (w / v) TBAB buffer (pH 7.5 ± 0.1) and 50% acetonitrile, followed by 2 column volumes of 30% 3M acetic acid and 70% acetonitrile. The entire unit operation was performed at a linear flow rate of 360cm / h. The C18 column manufactured by Waters.Ltd was used to analyze the sample loading, elution fractions, and elution pool by analytical methods. The results are listed in Table 55:

[0469] Table 55 Results of Experiment 33

[0470]

[0471] Experiment 34

[0472] The end product of the enzyme reaction containing human lispro insulin at a concentration of 13 mg / ml was adjusted to pH 7.5 and 10% acetonitrile, then diluted to 1.45 mg / ml with pure water and used as a loading sample. A pre-packed reverse phase Daisopak 200-10-C18 column (2*25 cm) was used to purify lispro insulin. Initially, the column was equilibrated with 90% 100 mM Tris + 200 mM imidazole + 0.2% (w / v) TBAB buffer (pH 7.5 ± 0.1) and 10% acetonitrile for 5 column volumes. The sample was loaded onto the column at a binding amount of about 7.3 g of human lispro insulin per liter of resin. The loosely bound protein was washed with 75% 100 mM Tris + 200 mM imidazole + 0.2% (w / v) TBAB buffer (pH 7.5 ± 0.1) and 15% acetonitrile for about 4 column volumes. The bound protein was eluted with a 24-29% acetonitrile gradient and 100mMTris+200mM imidazole+0.2% (w / v) TBAB buffer (pH 7.5 ± 0.1) over 13 column volumes. Based on the increase in absorbance at 280nm, multiple fractions of 0.25 column volumes were collected. Tightly bound proteins were eluted with 2 column volumes of 50% 100mMTris+200mM imidazole+0.2% (w / v) TBAB buffer (pH 7.5 ± 0.1) and 50% acetonitrile, followed by 2 column volumes of 30% 3M acetic acid and 70% acetonitrile. The entire unit operation was performed at a linear flow rate of 360cm / h. The loaded sample, eluted fractions, and elution pool were analyzed by analytical methods using a C18 column manufactured by Waters.Ltd. The results are listed in Table 56:

[0473] Table 56 Results of Experiment 34

[0474]

[0475] Experiment 35

[0476] The end product of the enzyme reaction containing human lispro insulin at a concentration of 13 mg / ml was adjusted to pH 7.5 and 10% acetonitrile, then diluted to 1.45 mg / ml with pure water and used as a loading sample. A pre-packed reverse phase Daisopak 200-10-C18 column (2*25 cm) was used to purify lispro insulin. Initially, the column was equilibrated with 90% 100 mM Tris + 200 mM imidazole + 0.2% (w / v) TBAB buffer (pH 7.5 ± 0.1) and 10% acetonitrile for 5 column volumes. The sample was loaded onto the column at a binding amount of about 7.3 g of human lispro insulin per liter of resin. The loosely bound protein was washed with 75% 100 mM Tris + 200 mM imidazole + 0.2% (w / v) TBAB buffer (pH 7.5 ± 0.1) and 15% acetonitrile for about 4 column volumes. The bound protein was eluted with a 24-29% acetonitrile gradient and 100mMTris+200mM imidazole+0.2% (w / v) TBAB buffer (pH 7.5 ± 0.1) over 13 column volumes. Based on the increase in absorbance at 280nm, multiple 0.25 column volume fractions were collected. Tightly bound proteins were eluted with 2 column volumes of 50% 100mMTris+200mM imidazole+0.2% (w / v) TBAB buffer (pH 7.5 ± 0.1) and 50% acetonitrile, followed by 2 column volumes of 30% 3M acetic acid and 70% acetonitrile. The entire unit operation was performed at a linear flow rate of 360cm / h. The loaded sample, eluted fractions, and elution pool were analyzed by analytical methods using a C18 column manufactured by Waters.Ltd. The results are listed in Table 57:

[0477] Table 57 Results of Experiment 35

[0478]

[0479] A summary of experimental data for reverse phase purification of -1 at high pH is detailed in Table 58 below.

[0480] Table 58 Experimental data summary of reverse phase purification-1 at high pH

[0481]

[0482] The first reverse phase purification step of human insulin lispro used a high pH buffer in combination with the ion pairing agent tetrabutylammonium bisulfate to reduce glycosylation variants by 65-75%, thereby increasing purity from approximately 58% in the top sample to >82% in the elution pool.

[0483] Previous glycosylated impurities were eliminated using a second reverse phase purification step which employed low pH to increase the purity of insulin lispro to ≥95%.

[0484] Example 8 shows the reverse phase purification step-2 performed at low pH by Experiments 36-38.

[0485] Experiment 36

[0486] The reverse phase purification-1 final product containing about 3 mg / ml of human insulin glargine and about 28.3% of acetonitrile was adjusted to pH 4.0 with acetic acid and diluted with pure water to obtain a final insulin lispro containing 10% acetonitrile at a concentration of about 1.06 mg / ml, which was used as a loading sample. A pre-packed reverse phase Daisopak 200-10-C18 column (1*25 cm) was used to purify insulin lispro. Initially, the column was equilibrated with about 10 column volumes of 90% 400 mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50 mM GuCl (pH 3.85 ± 0.1) and 10% acetonitrile. The sample was loaded onto the column at a binding amount of about 7.5 g of human insulin lispro per liter of resin. Loosely bound proteins were washed with 78% of 400mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50mM GuCl (pH 3.85 ± 0.1) and 22% acetonitrile with about 4 column volumes. Bound proteins were eluted with a 24-28% acetonitrile gradient and 400mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50mM GuCl (pH 3.85 ± 0.1) over 20 column volumes at a linear flow rate of 220cm / h. Multiple 0.25 column volume fractions were collected based on the increase in absorbance at 280nm. Tightly bound proteins were eluted with 4 column volumes of 30% 1M acetic acid and 70% acetonitrile. Except for elution, the entire unit operation was performed at a linear flow rate of 360cm / h. The load sample, elution fraction and elution pool were analyzed by an analytical method using a C18 column manufactured by Waters Ltd. The results are listed in Table 59 below.

[0487] Table 59 Results of Experiment 36

[0488]

[0489] Experiment 37

[0490] The reverse phase purification-1 final product containing about 3 mg / ml of human insulin lispro and about 28.8% of acetonitrile was adjusted to pH 4.0 with acetic acid and diluted with pure water to obtain a final insulin lispro containing 10% acetonitrile at a concentration of about 1.06 mg / ml, which was used as a loading sample. A pre-loaded reverse phase Daisopak 200-10-C18 column (1*25 cm) was used to purify insulin lispro. Initially, the column was equilibrated with about 10 column volumes of 90% 400 mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50 mM GuCl (pH 3.85 ± 0.1) and 10% acetonitrile. The sample was loaded onto the column at a binding amount of about 7.2 g of human insulin lispro per liter of resin. Loosely bound proteins were washed with 78% of 400mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50mM GuCl (pH 3.85 ± 0.1) and 22% acetonitrile with about 4 column volumes. Bound proteins were eluted with a 24-28% acetonitrile gradient and 400mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50mM GuCl (pH 3.85 ± 0.1) over 20 column volumes at a linear flow rate of 220cm / h. Multiple 0.25 column volume fractions were collected based on the increase in absorbance at 280nm. Tightly bound proteins were eluted with 4 column volumes of 30% 1M acetic acid and 70% acetonitrile. Except for elution, the entire unit operation was performed at a linear flow rate of 360cm / h. The load sample, elution fraction and elution pool were analyzed by an analytical method using a C18 column manufactured by Waters Ltd. The results are listed in Table 60 below.

[0491] Table 60 Results of Experiment 37

[0492]

[0493] Experiment 38

[0494] The reverse phase purification-1 final product containing about 2.3 mg / ml of human insulin lispro and about 28.6% of acetonitrile was adjusted to pH 4.0 with acetic acid and diluted with pure water to obtain a final insulin lispro with a concentration of about 0.83 mg / ml containing 10% acetonitrile, which was used as a loading sample. A pre-packed reverse phase Daisopak 200-10-C18 column (1*25 cm) was used to purify insulin lispro. Initially, the column was equilibrated with about 5 column volumes of 90% 400 mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50 mM GuCl (pH 3.85 ± 0.1) and 10% acetonitrile. The sample was loaded onto the column at a binding amount of about 8.0 g of human insulin lispro per liter of resin. Loosely bound proteins were washed with 78% of 400mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50mM GuCl (pH 3.85 ± 0.1) and 22% acetonitrile with about 4 column volumes. Bound proteins were eluted with a 24-28% acetonitrile gradient and 400mM trisodium citrate-citric acid buffer + 0.1% (w / v) OSA + 50mM GuCl (pH 3.85 ± 0.1) over 20 column volumes at a linear flow rate of 220cm / h. Multiple 0.25 column volume fractions were collected based on the increase in absorbance at 280nm. Tightly bound proteins were eluted with 4 column volumes of 30% 1M acetic acid and 70% acetonitrile. Except for elution, the entire unit operation was performed at a linear flow rate of 360cm / h. The load sample, elution fraction and elution pool were analyzed by an analytical method using a C18 column manufactured by Waters Ltd. The results are listed in Table 61 below.

[0495] Table 61 Results of Experiment 38

[0496]

[0497] A summary of experimental data for reverse phase purification of -2 at low pH is detailed in Table 62 below.

[0498] Table 62 Summary of experimental data for reverse phase purification-2 at low pH

[0499]

[0500] Note: The LOQ of the analytical method is 0.05% and the LOD is 0.03%

[0501] The mass characteristics and possible IDs based on mass of the final product of the reverse phase purification step-1 are detailed in Tables 63 (Experiment 33), 64 (Experiment 34) and 65 (Experiment 35) below. The mass characteristics and possible IDs based on mass of the final product of the reverse phase purification step-2 are detailed in Tables 66 (Experiment 36), 67 (Experiment 37) and 68 (Experiment 38) below. The mass spectrometry data UV and TIC chromatograms of the final product of the reverse phase purification-1 are shown in Figures 1 and 2. Fig. 7A (Experiment 33), Figure 7B (Experiment 34) and Figure 7C (Experiment 35) shows. The mass spectrometry data UV and TIC chromatograms of the final product of reverse phase purification-2 are shown as follows: Fig. 8A (Experiment 36), Figure 8B (Experiment 37) and Figure 8C (Experiment 38) shown.

[0502] Table 63 Reverse phase purification step-1 final product (experiment 33, Fig. 7A Mass characteristics and possible mass-based IDs of glycoforms of the same species (shown).

[0503]

[0504] Table 64 Reverse phase purification step-1 final product (experiment 34, Figure 7B Mass characteristics and possible mass-based IDs of glycoforms of the same species (shown).

[0505]

[0506] Table 65 Reverse phase purification step-1 final product (experiment 35, Figure 7C Mass characteristics and possible mass-based IDs of glycoforms of the same species (shown).

[0507]

[0508] Table 66 Reverse Phase Purification Step-2 Final Product (Experiment 36, Fig. 8A Mass characteristics and possible mass-based IDs of glycoforms of the same species (shown).

[0509]

[0510] Table 67 Reverse Phase Purification Step-2 Final Product (Experiment 37, Figure 8B Mass characteristics and possible mass-based IDs of glycoforms of the same species (shown).

[0511]

[0512] Table 68 Reverse Phase Purification Step-2 Final Product (Experiment 38, Figure 8C Mass characteristics and possible mass-based IDs of glycoforms of the same species (shown).

[0513]

[0514] in conclusion

[0515] As described in detail in the examples, the present invention relates to a two-step purification method which achieves the removal of about 99.95% of the glycosylated forms of the insulin analogs by separating the non-glycosylated insulin analogs from the glycosylated insulin analogs using the concept of an ion pair reagent (preferably OSA) in one step and an alkaline pH-based eluent in another step.

Claims

1. A method for purifying non-glycosylated insulin glargine from a complex mixture containing both non-glycosylated insulin glargine and glycosylated insulin glargine, wherein: The purification method comprises the steps of: a1) performing RP-HPLC on the complex mixture in the presence of a first ion-pairing agent in combination with an organic modifier at acidic pH to obtain a first mixture comprising partially purified non-glycosylated insulin glargine; and b1) subjecting the first mixture to RP-HPLC in the presence of a second ion pairing agent and an organic modifier at a different alkaline pH from step a1 to obtain at least 99.95% pure non-glycosylated insulin glargine, or; a2) subjecting the complex mixture to RP-HPLC in the presence of a second ion-pairing agent in combination with an organic modifier at alkaline pH to obtain a first mixture comprising partially purified non-glycosylated insulin glargine; and b2) subjecting the first mixture to RP-HPLC in the presence of a first ion-pairing reagent and an organic modifier at an acidic pH different from that in step a2, to obtain at least 99.95% pure non-glycosylated insulin glargine, wherein performing RP-HPLC at an acidic pH comprises eluting with a first ion pair reagent at an acidic pH and an organic modifier, wherein the first ion pair reagent comprises octane sulfonic acid, and Performing RP-HPLC at alkaline pH includes eluting with a second ion-pairing reagent at alkaline pH and an organic modifier, the second ion-pairing reagent comprising sodium perchlorate, and The organic modifier is acetonitrile.

2. The method according to claim 1, wherein: Obtaining a first mixture containing partially purified non-glycosylated insulin glargine by low pH based RP-HPLC comprises the following steps: a) Packing RP-HPLC columns with silica-based resins; b) loading the complex mixture onto a RP-HPLC column at a volume of 4-12 g / L; c) washing the column with a first ion-pairing reagent in combination with an organic modifier at an acidic pH of 3.5 to 3.9; and d) performing a linear gradient from 24% to 31% to elute the first mixture of partially purified non-glycosylated insulin glargine.

3. The method according to claim 2, wherein the RP-HPLC column is a C8 column.

4. The method according to claim 1, wherein: Obtaining the at least 99.95% pure non-glycosylated insulin glargine by high pH based RP-HPLC followed by low pH based RP-HPLC steps comprises the following steps: a) Packing RP-HPLC columns with silica-based resins; b) loading the first mixture onto a RP-HPLC column at a volume of 4-12 g / L; c) washing the column with a second ion-pairing agent in combination with acetonitrile at a basic pH of 8.3 to 8.7; and d) Perform a linear gradient from 25% to 29% to elute at least 99.95% pure non-glycosylated insulin glargine.

5. The method according to claim 4, wherein the RP-HPLC column is a C8 column.

6. The method according to claim 1, wherein: A first mixture comprising partially purified non-glycosylated insulin glargine is obtained by high pH based RP-HPLC, wherein the high pH based RP-HPLC comprises the following steps: a) Packing RP-HPLC columns with silica-based resins; b) loading the complex mixture onto the RP-HPLC column at a concentration of 4-12 g / L; c) washing the column with a second ion-pairing agent in combination with acetonitrile at a basic pH of 8.3 to 8.7; and d) performing a 25% to 29% linear gradient to elute the first mixture of partially purified non-glycosylated insulin glargine.

7. The method according to claim 6, wherein the RP-HPLC column is a C8 column.

8. The method according to claim 1, wherein: Obtaining the at least 99.95% pure insulin glargine by low pH based RP-HPLC and then by high pH based RP-HPLC comprises the following steps: a) Packing RP-HPLC columns with silica-based resins; b) loading the first mixture onto a RP-HPLC column at a volume of 4-12 g / L; c) washing the column with a first ion-pairing reagent in combination with an organic modifier at an acidic pH of 3.5 to 3.9; and d) Perform a linear gradient from 24% to 31% to elute at least 99.95% pure non-glycosylated insulin glargine.

9. The method according to claim 8, wherein the RP-HPLC column is a C8 column.

10. The method according to claim 1, wherein: The insulin glargine is insulin glargine comprising less than 0.05% of insulin glargine glycosylation variants.

Images