PURIFICATION OF GLUCAGON-LIKE PEPTIDE 1 ANALOGS

Abstract

The present invention refers to a method of purifying a glucagon-like peptide 1 analogs, the method comprising a two dimensional reversed phase high performance liquid chromatography protocol, wherein the first step is carried out at a pH value between 7.0 to 7.8 using a mobile phase comprising a phosphate buffer and acetonitrile, and the second step is carried out at a pH value below 3.0 using a mobile phase comprising trifluoroacetic acid and acetonitrile.

Claims

1-15. (canceled)

16. A method for the purification of Liraglutide, comprising: a) providing a liquid composition C comprising Liraglutide and at least one unwanted component; b) subjecting the composition C to a first reversed phase high performance liquid chromatograph (RP-HPLC) purification at a pH between 7.0 and 7.8, wherein a hydrocarbon bonded silica is used as a stationary phase, a mobile phase comprising an aqueous phosphate buffer AB1 and acetonitrile is used, and elution is effected by gradually increasing the acetonitrile concentration within the mobile phase while collecting Liraglutide containing fractions; and c) subjecting the pooled Liraglutide containing fractions obtained in step b) to a second reversed phase HPLC purification at a pH below 3.0, wherein a hydrocarbon bonded silica is used as a stationary phase, a mobile phase comprising trifluoroacetic acid and acetonitrile is used, and elution is effected by gradually increasing the acetonitrile concentration within the mobile phase while collecting fractions containing purified Liraglutide.

17. The method according to claim 16, wherein the aqueous phosphate buffer AB1 in step b) is ammonium phosphate buffer, preferably at a concentration of 5 mM to 50 mM.

18. The method according to claim 16, wherein the gradient in step b) is selected from the range of 19 to 67% (v/v) acetonitrile and/or wherein the gradient in step c) is selected from the range of 31 to 100% (v/v) acetonitrile.

19. The method according to claim 16, wherein the trifluoroacetic acid concentration within the mobile phase used in step c) is selected from the range of 0.05 to 0.5% (v/v), preferably 0.05 to 0.1% (v/v).

20. The method according to claim 16, further comprising the step of: d) subjecting the Liraglutide obtained in step c) to a third reversed phase HPLC purification at a pH between 7.0 and 7.8, wherein a hydrocarbon bonded silica is used as a stationary phase, a mixture of an aqueous buffer AB2 with acetonitrile is used as a mobile phase, and elution is effected by gradually increasing the acetonitrile concentration within the mobile phase while collecting fractions containing purified Liraglutide.

21. The method according to claim 20, wherein said aqueous buffer AB2 is selected from the group consisting of: a mixture of sodium dihydrogen phosphate and disodium hydrogen phosphate, a mixture of potassium dihydrogen phosphate and dipotassium hydrogen phosphate, potassium acetate, and sodium acetate.

22. The method according to claim 16, wherein all or parts of step b) and/or step c) and/or step d), if present, is/are carried out at a temperature selected from the range of 4 C. to 25 C., preferably 4 C. to 10 C.

23. The method according to claim 16, wherein the stationary phase used in steps b) and c) and step d), if present, is C8 bonded silica or C18 bonded silica.

24. The method according to claim 16, further comprising a step e) of size exclusion chromatography.

25. The method according to claim 16, further comprising a step f) of desalting the peptide, preferably wherein desalting is performed by ion exchange chromatography, by size exclusion chromatography, or by ultrafiltration.

26. The method according to claim 24, wherein all or parts of the respective step is/are carried out at a temperature selected from the range of 4 C. to 20 C., preferably 4 C. to 10 C.

27. The method according to claim 16, wherein step a) comprises dissolving a dried crude Liraglutide peptide in an aqueous phosphate buffer AB0 at a pH selected from the range of 7.0 to 7.5.

28. The method according to claim 16, wherein the crude Liraglutide peptide is obtained by solid phase peptide synthesis, followed by trifluoroacetic acid mediated cleavage and peptide precipitation from the cleavage composition.

29. The method according to claim 16, wherein a purified Liraglutide is lyophilized, preferably at a pH selected from the range of 6.6 to 7.9, preferably 7.0 to 7.8, and most preferably 7.0 to 7.5.

30. A composition LC comprising Liraglutide obtainable from a method according to claim 16, characterized in that said composition contains Liraglutide at a purity above 99%, and does contain detectable levels, but not more than 0.5%, preferably not more than 0.3%, more preferably not more than 0.2%, and most preferably not more than 0.1% of each of i) any Liraglutide derivative, where the indole moiety in the side chain of Trp at position 25 is oxidized by incorporation of a single oxygen atom, and/or of ii) any Liraglutide derivative, where the indole moiety in the side chain of Trp at position 25 is oxidized by incorporation of two oxygen atoms, and/or of iii) a Liraglutide derivative comprising kynurenine instead of Trp at position 25, and/or of iv) a Liraglutide deletion variant lacking Gly31.

Description

BRIEF DESCRIPTION OF THE FIGURES

[0159] FIG. 1 shows the sequence alignment of selected glucagon-like peptides. Moieties sharing identity with the GLP-1 sequence are written in bold.

[0160] FIG. 2: Two dimensional Liraglutide purification of Example 2. Overlays of the analytical RP-UHPLC traces of the crude Liraglutide preparation used as starting material (crude, indicated as 1)) and of the pooled fractions obtained after the first purification dimension (1D, pool (NH.sub.4).sub.3PO.sub.4, pH 7.5, indicated as 2)) and second purification dimension (2D, pool TFA, indicated as 3)) are shown. Arrows highlight unwanted components, which were not removed in the first dimension, but in the second dimension.

[0161] FIG. 3: Plot of Conductivity of the retentate versus time during ultrafiltration.

EXAMPLES

Example 1: Determination of Purification Conditions

[0162] Small scale experiments were carried out to identify suitable purification conditions. The seven mobile phase buffers given in column 2 of Table 2 were tested each on four different stationary phases as indicated in line 1, columns 3-6 of Table 2. Each mobile phase buffer was used to prepare a Buffer A consisting of 3% (v/v) acetonitrile in an aqueous solution of said buffer, and a Buffer B consisting of 67 or 80% (v/v) acetonitrile in an aqueous solution of said buffer. The buffer concentrations in the aqueous solutions were between 20 and 400 mM, depending on the nature of the mobile phase buffer.

[0163] Each line of Table 2 represents four different one dimensional RP-HPLC runs, namely one run for each of the four stationary phases tested. For each of said runs, the Buffers A and B prepared with the mobile phase buffer indicated in column 2 the respective line were used. The following general protocol applied: Crude Liraglutide peptide produced by Fmoc SPPS (purity >=60%) was dissolved in Buffer A, and 18 mg each applied in parallel experiments to the four different stationary phases consisting of C8 or C18 bonded silica (column dimensions: 2504.6 mm). The protocol generally involved equilibration of the column for 15 min in Buffer A, sample loading, and elution for 1 min with Buffer A alone, followed by a gradient of 20-100% Buffer B. The flow rate was of 0.63 ml/min. Fractions of 0.37 ml were collected and analyzed by reversed phase UHPLC. Table 2 below indicates the purities determined in the purest fraction of each experiment in terms of relative peak area. Said relative peak area was calculated by dividing the Liraglutide peak area by the sum of all peak areas observed in analytical UHPLC, i.e. the area of the Liraglutide peak was expressed in percent of the total peak area.

TABLE-US-00003 TABLE 2 YMC Triart Luna PREP Kromasil Daiso ODS- Buffer C8-L C8 C18 Bio 1 NH.sub.4HCO.sub.3 95.40% 96.56% 96.86% 97.18% 2 (NH.sub.4).sub.3PO.sub.4 97.84% 98.02% 98.10% 97.76% 3 NH.sub.4OAc 95.84% 96.71% 94.79% 93.97% 4 TEAP 94.76% 95.58% 95.66% 95.52% 5 AcOH 94.18% 91.79% 94.90% 95.74% 6 H.sub.3PO.sub.4 95.55% 95.01% 95.37% 95.33% 7 TFA 96.97% 96.24% 97.00% 96.93%

Conclusions:

[0164] 1. Various C8 and C18 stationary phases give similar results. For each of the stationary phases, the following observations applied: [0165] 2. Under neutral to slightly basic (7.0pH<8.0) conditions, ammonium phosphate buffer is surprisingly superior to other buffers tested (cf. columns 3-6 of lines 1-4). [0166] 3. Under acidic conditions (pH<3.0), TFA buffer is surprisingly superior to other buffers tested (cf. columns 3-6 of lines 5-7).

Example 2: Two Dimensional RP-HPLC Purification

[0167] The purification involved a chromatographic purification at pH 7.5 in the first dimension, followed by a chromatographic purification under acidic conditions in the second dimension.

[0168] A 5 cm MODcol column (Grace) packed with C8-bonded silica (approx. bed-depth 32 cm) was used on a preparative HPLC system (Knauer HPLC pump 1800) with detection at 220 nm (Knauer smartline UV detector 2500) and an automated fraction collector (Bchi C-660). The same stationary phase was used in both dimensions of the purification protocol. Crude Liraglutide produced by Fmoc SPPS (purity >60%) was used as a starting material. The sample was loaded on the column at a flow rate of 90 ml/min. The detailed elution protocols for each step are given in Tables 3 and 4 below. The buffer concentration in the aqueous part of the mobile phase used in the first dimension was 20 mM.

TABLE-US-00004 TABLE 3 Parameters first purification dimension Sample loading buffer aqueous ammonium phosphate pH 7.5 Buffer A 3% (v/v) acetonitrile, aqueous ammonium phosphate, pH 7.5 Buffer B 67% (v/v) acetonitrile, aqueous ammonium phosphate, pH 7.5 Elution protocol Time Flow rate Buffer A Buffer B [min] [ml/min] [%] [%] Remarks 0 90 100 0 Flushing 20 90 100 0 post loading 21 90 76 24 Elution 103 90 0 100 linear gradient

[0169] The pooled fractions obtained from the first RP-HPLC step were further purified as set forth in Table 4.

TABLE-US-00005 TABLE 4 Parameters second purification dimension Sample loading buffer 3% (v/v) acetonitrile, aqueous ammonium phosphate pH 7.5 Buffer A 3% (v/v) acetonitrile, 0.1% (v/v) aqueous TFA Buffer B 0.1% (v/v) TFA in acetonitrile Elution protocol Time Flow rate Buffer A Buffer B [min] [ml/min] [%] [%] Remarks 0 90 100 0 Flushing 30 90 100 0 post loading 31 90 70 30 Elution: 146 90 0 100 linear gradient

[0170] The purity of the pooled fractions was after the second purification dimension was 98.8% as assessed by analytical RP-UHPLC, the overall yield after both steps was 35%. Comparison of the analytical RP-UHPLC traces of starting crude material and of the pooled fractions after the first and second HPLC pass demonstrated the surprising complementarity of both purification dimensions: Each purification dimension removed different unwanted components, such that the combination of both steps resulted in excellent product purity (cf. FIG. 2).

Example 3: Two Dimensional RP-HPLC Purification

[0171] The purification involved a chromatographic purification at pH 7.7 in the first dimension, followed by a chromatographic purification under acidic conditions in the second dimension.

[0172] A 5 cm MODcol column (Grace) packed with C8-bonded silica (approx. bed-depth 32 cm) was used on a preparative HPLC system (Knauer HPLC pump 1800) with detection at 220 nm (Knauer smartline UV detector 2500) and an automated fraction collector (Bchi C-660). The same stationary phase was used in both dimensions of the purification protocol. Crude Liraglutide produced by Fmoc SPPS was used as a starting material. The sample was loaded on the column at a flow rate of 43 ml/min (1.sup.st dimension) or 64 ml/min (2.sup.nd dimension). The detailed elution protocols for each step are given in Tables 5 and 6 below.

TABLE-US-00006 TABLE 5 Parameters first purification dimension Sample loading buffer aqueous ammonium phosphate pH 7.7 Buffer A 3% (m/m) acetonitrile, aqueous ammonium phosphate pH 7.7 Buffer B 61% (m/m) acetonitrile, aqueous ammonium phosphate pH 7.7 Elution protocol Time Flow rate Buffer A Buffer B [min] [ml/min] [%] [%] Remarks 0 90 100 0 Flushing 20 90 100 0 post loading 20.1 36.5 76 24 Elution: 102 36.5 0 100 linear gradient

[0173] The pooled main fraction obtained from the first RP-HPLC step was further purified as set forth in Table 6.

TABLE-US-00007 TABLE 6 Parameters second purification dimension Sample loading buffer 2% (m/m) acetonitrile, aqueous ammonium phosphate pH 7.7 Buffer A 2% (m/m) acetonitrile, 0.1% (v/v) aqueous TFA Buffer B 0.1% (v/v) TFA in 100% acetonitrile Elution protocol Time Flow rate Buffer A Buffer B [min] [ml/min] [%] [%] Remarks 0 90 100 0 Flushing 30 90 100 0 post-loading 30.1 36 70 30 Elution: 145 36 0 100 Linear gradient

[0174] The purity of the pooled main fraction was 99.38%, and the largest non-product peak was 0.18% as assessed by analytical RP-UHPLC. In other words, the preparation did not contain any unwanted component at a concentration above 0.3 as assessed by analytical RP-UHPLC.

[0175] Surprisingly, an attempt to swap the order of purification steps, i.e. to perform the run with the TFA-containing mobile phase first, failed due to column clogging.

Example 4: RP-HPLC Purification, Optional 3rd Dimension

[0176] A 5 cm MODcol column (Grace) packed with C8-bonded silica (approx. bed-depth 32 cm) was used on a preparative HPLC system (Knauer HPLC pump 1800) with detection at 220 nm (Knauer smartline UV detector 2500) and an automated fraction collector (Bchi C-660). Liraglutide purified by the two-dimensional approach given above was used as a starting material (purity: 99.2%). The column was equilibrated in Buffer A and the sample was loaded on the column at a flow rate of 43 ml/min. The detailed elution protocol is given in Table 7 below.

[0177] The purity of the pooled main fraction was 99.35% as assessed by analytical RP-UHPLC with UV detection at 220 nm. The preparation did not contain any peptidic impurity at a concentration above 0.3%.

TABLE-US-00008 TABLE 7 Parameters third purification dimension Sample loading buffer 3% (m/m) acetonitrile, aqueous sodium hydrogen phosphate, pH 7.7 Buffer A 3% (m/m) acetonitrile, aqueous sodium hydrogen phosphate, pH 7.7 Buffer B 61% (m/m) acetonitrile, aqueous sodium hydrogen phosphate, pH 7.7 Elution protocol Time Flow rate Buffer A Buffer B [min] [ml/min] [%] [%] Remarks 0 89 100 0 Flushing 20 89 100 0 post loading 20.1 36.5 76 24 Elution: 102 36.5 0 100 Linear gradient

Example 5: RP-HPLC Purification, Optional 3rd Dimension

[0178] A 5 cm MODcol column (Grace) packed with C8-bonded silica (approx. bed-depth 32 cm) was used on a preparative HPLC system (Knaur HPLC pump 1800) with detection at 220 nm (Knaur smartline UV detector 2500) and an automated fraction collector (Bchi C-660). Liraglutide purified by the two-dimensional approach given above was used as a starting material. The column was equilibrated in Buffer A and the sample was loaded on the column at a flow rate of 43 ml/min. The detailed elution protocol is given in Table 8 below.

TABLE-US-00009 TABLE 8 Parameters third purification dimension Sample loading buffer 3% (m/m) acetonitrile, aqueous disodium hydrogen phosphate, pH 7.5 Buffer A 3% (m/m) acetonitrile, aqueous disodium hydrogen phosphate, pH 7.5 Buffer B 61% (m/m) acetonitrile, aqueous disodium hydrogen phosphate, pH 7.5 Elution protocol Time Flow rate Buffer A Buffer B [min] [ml/min] [%] [%] Remarks 0 89 100 0 Flushing 20 89 100 0 post loading 20.1 36.5 76 24 Elution: 102 36.5 0 100 Linear gradient

[0179] The purity of the pooled main fraction was 99.36% as assessed by analytical RP-UHPLC. The preparation did not contain any peptidic impurity at a concentration above 0.3%.

Example 6: Desalting by Ultrafiltration

[0180] UHPLC purified Liraglutide (1.7 l, concentration approximately 35 g/1) was subjected to tangential flow filtration using standard equipment with a polyethersulfone (PES) membrane having a molecular weight cut-off of 1 kDa. A transmembrane pressure of 2.2 bar and a flow rate of 1 l/min were applied, and a permeate flow of 33 ml/min was observed. The volume loss in the retentate was compensated by constant addition of ultrapure water.

[0181] As shown in FIG. 3, the salt content as reflected by the retentate's conductivity decreased over time. When the volume of the filtrate reached 10 fold the volume of the retentate, the retentate contained only traces of residual salt. The peptide purity as detected by UHPLC analysis was 99.3, the overall net peptide yield was 91%.

Example 7: Removal of Unwanted Components During Purification

[0182] Liraglutide obtained from Fmoc-SPPS was subjected to the three-dimensional purification method of the present invention. C8-bonded silica was employed as stationary phase, and the aqueous component of the mobile phase contained phosphate buffer in the first dimension, TFA in the second dimension and acetate buffer in the third dimension. The pooled fractions obtained after each step were analyzed by LC-MS to evaluate the efficiency of the purification protocol. The findings relating to specific dominant unwanted components are summarized in Table 9 below. The concentrations are given in area percent of the Liraglutide main peak.

[0183] It can be seen that the Liraglutide truncation variant lacking Gly.sup.31 and Liraglutide with mono-oxygenated Trp.sup.25 are efficiently reduced by the first purification dimension, the second purification dimension achieves additional removal of Liraglutide with di-oxygenated Trp.sup.25 and the third purification dimension achieves control of Kyn.sup.25-Liraglutide. No other peptidic impurity was detectable at levels above 0.5% by analytic chromatography in the pooled fractions after the 3.sup.rd purification dimension.

TABLE-US-00010 TABLE 9 Sample Des-Gly.sup.31 Trp(O).sup.25- Trp(2O).sup.25- Kyn.sup.25- Liraglutide Liraglutide Liraglutide Liraglutide crude 0.47 2.17 0.20 0.85 Pool after 1.sup.st <0.01 0.33 0.05 0.33 dimension Pool after 2.sup.nd not 0.04 0.02 0.34 dimension detected Pool after 3.sup.rd not 0.02 0.01 0.01 dimension detected