Method for separation of monomeric polypeptides from aggregated polypeptides

09783570 · 2017-10-10

Assignee

Inventors

Cpc classification

International classification

Abstract

Herein is reported a method for obtaining a polypeptide in monomeric form comprising the steps of a) providing a solution comprising the polypeptide m monomeric form and m aggregated form, wherein the ratio of monomeric to aggregated form is 4:1 or less as determined by size exclusion chromatography, b) performing a mixed-mode chromatography in bind-and-elute mode, or a hydrophobic interaction chromatography in flow-through mode, or a size-exclusion chromatography, and c) performing a weak cation exchange chromatography in bind-and-elute mode or flow-through mode, and thereby obtaining the polypeptide m monomeric form.

Claims

1. A method for purifying a monomeric Fc-region containing polypeptide comprising the steps: a) applying a solution comprising a Fc-region containing polypeptide in monomeric form and in aggregated form to a mixed-mode chromatography material, wherein the aggregated form is 20% or more as determined by size exclusion chromatography of the total Fc-region containing polypeptide in the solution, wherein said solution has a pH value of about 0.2 to about 2.5 pH units above the pl of the polypeptide, and recovering a solution comprising about 95% monomeric Fc-region containing polypeptide; and b) applying the solution recovered in step a) to a weak cation exchange chromatography material, wherein said solution has pH value of about 0.2 to about 2.5 pH units above the pl of the polypeptide, and recovering the monomeric form of the Fc-region containing polypeptide.

2. The method according to claim 1, wherein the monomeric form of the Fc-region containing polypeptide is recovered from the flow-through of the weak cation exchange chromatography material of step b).

3. The method according to claim 1, wherein the solution applied to the mixed-mode chromatography material of step a) and/or the weak cation exchange chromatography material of step b) is buffered with a citrate buffer, acetate buffer, and/or MES buffer.

Description

DESCRIPTION OF THE FIGURES

(1) FIG. 1 Exemplary elution profile of mutIL15-Fc from CM-Sepharose FF at pH 5.2, 200 mmol/l NaCl: 4 mg mutIL15-Fc in 10 mmol/l sodium citrate pH 5.2 were loaded on a 5 ml HiTrap™ CM-Sepharose FF column; elution was done by step elution at 200 mmol/l NaCl in the same buffer. The yield of monomeric mutIL15-Fc in the first peak (pool) was 77%; the amount of aggregate in the pooled fractions was 2.2%; fractions of the second peak contained 30-90% aggregated mutIL15-Fc; tightly bound aggregates were eluted at 1 mol/l NaCl (3rd peak).

(2) FIG. 2 Exemplary elution profile of mutIL15-Fc from CM-Sepharose FF in the MES buffer system (pH 5.0/pH 8.0): mutIL15-Fc previously purified to 95.5% monomer purity was loaded onto the CM-Sepharose FF column at pH 5.0; elution was initiated by increasing the pH (60% buffer B step, approx. pH 5.8); the yield of monomeric mutIL15-Fc in the first peak (pool) was 80%. The amount of aggregate in the pooled fractions was lowered to 1%; fractions of the second peak contained approx. 20% aggregated mutIL15-Fc.

(3) FIG. 3 Sequence of a) protein A affinity chromatography in bind-and-elute mode, b) mixed-mode chromatography in bind-and-elute mode, and c) weak cation exchange chromatography in bind-and-elute mode.

DEFINITIONS

(4) The term “recovery monomer” denotes the monomer found in the column efflux of the overall experiment and is calculated as the ratio of the monomer found in the total efflux vs. the monomer loaded onto the column.

(5) The term “yield monomer” denotes the purified monomer eluted from the column in a quality worth to be further processed and is calculated as the ratio of the eluted and purified monomer vs. the monomer loaded onto the column.

(6) The term “% monomer” denotes the partial fraction of mutIL15-Fc in monomeric form in the sample and is calculated as the ratio of monomeric mutIL15-Fc vs. the total amount of mutIL15-Fc, i.e. the sum of monomeric and aggregated form of the polypeptide, in the sample.

(7) The Term “HMW” (=high molecular weight) denotes the aggregated form of the polypeptide.

Example 1

(8) Production of mutIL15-Fc

(9) The interleukin-15/Fc fusion protein has been prepared in accordance with the data and methods reported in the international patent applications WO 1997/041232, WO 2005/100394 and WO 2005/100395 (incorporated herein by reference in their entirety).

Example 2

(10) Purification with a Protein a Affinity Chromatography

(11) The crude cultivation supernatant obtained in Example 1 was applied to the protein A affinity chromatography material MabSelectSuRe (GE Healthcare, Uppsala, Sweden). The column was loaded with up to 14 mg protein per ml column volume. The bound mutIL15-Fc fusion protein was eluted with 100 mM citrate buffer at pH 3.6. An additional wash step prior to the elution was introduced to remove host cell DNA and host cell protein. Of from 76.6% to 78.4% mutIL15-Fc has been recovered from the protein A affinity chromatography material.

(12) TABLE-US-00001 TABLE 1 Purification with a protein A affinity chromatography. load recovery yield [mg/ml] mode elution with monomer monomer HMWs up to 14 bind- 100 mM 76-78% 73% 35-40% and- citrate buffer elute pH 3.6

Example 3

(13) Purification with a Mixed-Mode Chromatography

(14) The protein A purified mutIL15-Fc was adjusted to the required pH value and salt concentration by dialysis and/or adding of appropriate stock solutions and afterwards applied to the mixed-mode chromatography column previously equilibrated with 25 mM acetate buffer pH 5.5. As mixed-mode chromatography material Capto MMC has been used (GE Healthcare, Uppsala, Sweden). The column was loaded with up to 20 mg protein per ml column volume. A solution comprising 25 mM sodium phosphate at pH 6.5 was used for recovery of the protein from the column.

(15) TABLE-US-00002 TABLE 2 Aggregate removal with a mixed-mode chromatography material. HMW [%] Total HMW Monomer 1 2 3 [%] [%] Load 5.2 12.0 14.7 31.9 68.1 Eluate 0.9 3.8 4.8 95.2 (Pooled) Regenerate 13.8 29.7 22.3 65.7 34.2

(16) TABLE-US-00003 TABLE 3 HCP removal with a mixed-mode chromatography material. HCP [ng/mg] Load 112 wash <1071 pool 23 Regenerate 202

(17) The mutIL15-Fc recovery was 58.3%.

(18) TABLE-US-00004 TABLE 4 Purification with a mixed-mode chromatography loading load elution yield buffer [mg/ml] mode with monomer monomer HMWs 25 mM up to 20 bind- 25 mM 58% 95% 5% acetate and- sodium buffer elute phosphate pH 5.5 pH 6.5

Example 4

(19) Purification with a Hydrophobic Interaction Chromatography

(20) a) Phenyl Sepharose

(21) Protein A purified mutIL15-Fc (approx. 12 mg) was adjusted to the required salt and 2-propanol concentrations by the adding of appropriate stock solutions and applied to a Phenyl Sepharose column HP (GE Healthcare, Uppsala, Sweden) previously equilibrated with 20 mM Tris-HCl buffer comprising 1.5 mol/l NaCl and 10% (v/v) 2-propanol, pH 7.0. The applied solution contained approximately 30% aggregated protein. Removal of tightly bound protein was achieved with 100% 20 mM Tris-HCl buffer comprising 50% (v/v) 2-propanol, pH 7.0. The column was operated at a flow rate of 5 ml/min.

(22) MutIL15-Fc was applied to the column in a 20 mM Tris-HCl buffer comprising 1.5 mol/l NaCl and 10% (v/v) 2-propanol, pH 7.0. Under these conditions mutIL15-Fc was found in the flow-through, whereas aggregates were bound to the chromatography material. In this non-binding method approx. 60% of the applied mutIL15-Fc monomer was recovered in the flow through at a purity of 94% with 6% aggregates.

(23) TABLE-US-00005 TABLE 5 Purification with Phenyl Sepharose HP. yield loading load mon- mon- buffer [mg/ml] mode elution with omer omer HMWs 20 mM 2.4 flow- loading 60% 94% 6% Tris-HCl through buffer buffer, (monomer), 1.5 20 mM Tris- mol/l HCl buffer, NaCl, 50% (v/v) 10% (v/v) 2-propanol, 2-propanol, pH 7.0 pH 7.0 (aggregates)
b) Toyopearl Butyl

(24) Protein A purified mutIL15-Fc (approx. 11 mg) was adjusted to the required salt and 2-propanol concentrations by the adding of appropriate stock solutions and applied to a Toyopearl Butyl-650M column (Tosoh Bioscience GmbH, Stuttgart, Germany) previously equilibrated with 20 mmol/l Tris/HCl pH 7.0, 0.5 mol/l NaCl, 15% (v/v) 2-propanol. The applied solution contained approximately 30% aggregated protein. Removal of tightly bound protein was achieved with 100% 20 mM Tris-HCl buffer comprising 50% (v/v) 2-propanol, pH 7.0. The column was operated at a flow rate of 5 ml/min.

(25) When protein A purified mutIL15-Fc was loaded onto the column in 20 mmol/1 Tris/HCl pH 7.0, 0.5 mol/l NaCl, 15% (v/v) 2-propanol 75% of monomeric mutIL15-Fc was recovered from the flow through at a purity of 95% with 5% of aggregates.

(26) TABLE-US-00006 TABLE 6 Purification with Toyopearl Butyl-650M. yield loading load mon- mon- buffer [mg/ml] mode elution with omer omer HMWs 20 2.2 flow- loading 75% 95% 5% mmol/l through buffer Tris/HCl (monomer), pH 7.0, 20 mM Tris- 0.5 HCl buffer, mol/l 50% (v/v) NaCl, 2-propanol, 15% (v/v) pH 7.0 2-propanol (aggregates)

Example 5

(27) Purification with a Weak Cation Exchange Chromatography

(28) a) Bind-and-Elute Mode

(29) The recovered solution of a protein A affinity chromatography comprising the mutIL15-Fc was equilibrated in buffer A. This was either performed using Amicon Ultra 15 devices (30 kDa MWCO, Millipore, order no. UFC803096) or by diafiltration. The sample was applied to the CM-Sepharose FF column (GE Healthcare, Uppsala, Sweden) previously equilibrated with buffer A. Elution was either achieved in a gradient or in a step elution profile using buffer B. Columns were operated at 2 to 2.5 ml/min.

(30) Using 10 mmol/l citrate buffer mutIL15-Fc bound completely to CM-Sepharose in the pH range from pH 3.0 to pH 5.4. At pH 5.5 about one third of the sample was recovered in the flow through. At pH 6.0 the main fraction of mutIL15-Fc did not bind to the CM-Sepharose. Therefore purification with CM-Sepharose in bind-and-elute mode was performed in the pH range between pH 5.0 and pH 5.5.

(31) When using a linear salt gradient monomeric mutIL15-Fc eluted prior to aggregated mutIL15-Fc. The higher the pH value was the better the separation of monomeric and aggregated mutIL15-Fc was resulting in aggregate-free fractions in the first half of the elution peak at pH 5.0 or at pH 5.2.

(32) When using step elution mutIL15-Fc was recovered with a 200 mmol/l sodium chloride solution. Using a 5 ml HiTrap column, mainly monomeric mutIL15-Fc (approx. 98% pure, yield 77%) eluted as a sharp peak followed by a second peak which contained aggregated mutIL15-Fc (approx. 90% aggregates) (see FIG. 1). Tightly bound protein was eluted with 1 mol/l NaCl. From the chromatogram it can be seen that during the elution of monomeric protein (first peak) the pH value decreased to approx. pH 4.0. During the elution of the second peak the pH partly recovered to pH 4.7. This shift in the pH value was also observed when the experiment was performed without loading of protein.

(33) The timely separation of the monomer and aggregate peak can be improved with increasing column length of the column. For example, at pH 5.2 (10 mmol/l citrate) monomeric mutIL15-Fc was eluted from a 15 ml column at same purity (98%) and better yield (90%) in comparison to the 5 ml HiTrap column using 200 mmol/l NaCl. The column load in this experiment was less than 1 mg/ml column bed volume.

(34) When the column load was stepwise increased the resolution of the CM-Sepharose FF column was gradually decreased. At low column load (0.36 mg per ml column bed volume, 10 mmol/l citrate pH 5.2) two totally separated peaks were eluted using a 200 mmol/l sodium chloride solution. At higher column load (5.3 mg/ml) the two peaks overlapped. The amount of aggregated mutIL15-Fc in the first (monomeric) peak was increased with increased amount of total protein. At low column load fractions of less than 3% aggregate could be recovered from the monomer peak. At higher load no fraction of the monomer peak contained less than 15% aggregates.

(35) Best results were obtained with 210 mmol/l NaCl at pH 5.2 (81% yield, 98.5% purity). Changing the salt concentration only +/−10 mmol/l had an impact on the separation, resulting in a lower purity and/or yield. When the pH value of the loading buffer was changed +/−0.2 pH units the elution of monomeric mutIL15-Fc at 210 mmol/l NaCl led to a strong decrease in purity (>20% aggregates). At pH 5.0 230 mmol/l NaCl were necessary to recover purified mutIL15-Fc (98.6% pure) with moderate yield (58%). At pH 5.4 elution with NaCl concentrations of more than 150 mmol/l results in the elution of aggregates.

(36) Alternatively to 10 mM citrate buffer pH 5.2 for column equilibration and protein loading 35 mmol/l acetate buffer at pH 5.0 or 10 mM MES buffer at pH 5.0 can be used, i.e. a buffer solution in the pH range of from pH 5.0 to pH 5.5 with a conductivity of about 1.9 mS/cm (same conductivity as the citrate buffer). For example with 35 mM acetate buffer, pH 5.0, mutIL15-Fc eluted in two separate peaks, where the first peak contained mainly monomeric protein (1%-8% aggregates) and the second peak contained mainly aggregates (60%-95%).

(37) Using KCl instead of NaCl for the step elution method at 180 mmol/l KCl resulted in a better separation than NaCl. With 180 mmol/l KCl a monomer shoulder in the rising edge of the elution peak was seen whereas no peak separation was seen with 180 mmol/l NaCl. The use of 220 mmol/l NaCl resulted in a good separation of monomer (100% purity, 70% yield).

(38) Using a pH gradient with acetate buffer and a gradient from pH 5.0 (buffer A) to pH 6.5 (buffer B), monomers eluted prior to aggregates. The shallower the gradient was the better the separation of monomer from aggregate was. The binding of mutIL15-Fc was complete at pH 5.0. Best purification results were obtained when equilibration and loading was performed at pH 5.0 and elution with 10% buffer B pH 6.0 (resulting in approx. pH 5.3 at the elution peak maximum). The main peak contained 68% of the loaded monomeric mutIL15-Fc at a purity of 99.5%. The column load was 0.3 mg/ml column bed volume. When the chromatography conditions were changed only slightly the purification of monomeric protein became worse again: when elution was performed at 11% buffer B, only 64% monomer at a purity of 97.5% could be recovered.

(39) Using MES buffer at pH 5.0 and pH 8.0 for buffer A and B, respectively, at 60% buffer B step (approx. pH 5.8) a monomer peak of 98.5% purity and a monomer yield of 96% was obtained. The influence of the column load on purification quality is summarized in the following table.

(40) TABLE-US-00007 TABLE 7 Influence of column load on yield and purity of mutIL15-Fc monomer from CM-Sepharose FF using a pH gradient with MES. column aggregates yield load input monomer HMWs run [mg/ml] [%] [%] [%] 1 0.3 40 96 1.5 2 3.1 40 75 2.0 3 10 38 76 4.5 4 7 4.5 80 1.0

(41) Comparative separation experiments performed with Toyopearl CM-650M (Tosoh Bioscience GmbH, Stuttgart, Germany, order no. 43203) provided for similar results as those obtained with the CM-Sepharose FF column.

(42) TABLE-US-00008 TABLE 8 Purification with CM-Sepharose FF in bind-and-elute mode. recov- load ery yield loading [mg/ elution mon- mon- buffer ml] mode with omer omer HMWs comment 40 mM 0.4 bind- 200 mM 73% 73% 25% sodium and- sodium acetate, elute chloride pH 5.3 25 mM 5 bind- 7.5 mM 84% 48% 0% 24% sodium and- sodium aggregate acetate, elute acetate, free, pH 5.2 200 mM monomer sodium in flow- chloride, through pH 5.5 10 mM 1 bind- 200 mM 72% 51% 0% sodium and- sodium acetate, elute chloride pH 5.5 15 mM 0.8 bind- 200 mM 77% 68% 3% sodium and- sodium acetate, elute chloride pH 5.5 20 mM 0.6 bind- 200 mM 96% 88% 14% sodium and- sodium acetate, elute chloride pH 5.5 25 mM 0.6 bind- 200 mM 90% 90% 22% sodium and- sodium acetate, elute chloride pH 5.5 30 mM 0.7 bind- 200 mM 69% 62% 27% 7% sodium and- sodium aggregate acetate, elute chloride free, pH 5.5 monomer in flow- through 60 mM 11 bind- 200 mM 95% 93% 27% sodium and- sodium acetate, elute chloride pH 5.1 90 mM 14 bind- 200 mM 91% 85% 21% sodium and- sodium acetate, elute chloride pH 5.1 10 5.6 bind- 220 mM n.d. >20% mmol/l and- sodium sodium elute chloride citrate, pH 5.0 10 3 bind- 230 mM 58% 1.4% mmol/l and- sodium sodium elute chloride citrate, pH 5.0 10 3 bind- 240 mM 51% 2% mmol/l and- sodium sodium elute chloride citrate, pH 5.0 10 4.2 bind- 200 mM n.d. >5% mmol/l and- sodium sodium elute chloride citrate, pH 5.2 10 3.5 bind- 210 mM 81% 1.5% mmol/l and- sodium sodium elute chloride citrate, pH 5.2 10 3.4 bind- 220 mM 70% 1.5% mmol/l and- sodium sodium elute chloride citrate, pH 5.2 10 2 bind- 130 mM n.d. >5% mmol/l and- sodium sodium elute chloride citrate, pH 5.4 10 1 bind- 150 mM n.d. 4% mmol/l and- sodium sodium elute chloride citrate, pH 5.4 10 2.5 bind- 200 mM n.d. >23% mmol/l and- sodium sodium elute chloride citrate, pH 5.4
b) Flow-Through Mode

(43) The solution was dialyzed against the respective loading buffer as listed in the following table. The monomeric mutIL15-Fc fusion protein was recovered in the flow-through in the loading buffer (isocratic elution). Aggregates and fragments were recovered with a single step elution with a solution comprising 200 mM sodium chloride.

(44) TABLE-US-00009 TABLE 9 Purification with CM-Sepharose FF in flow-through mode. loading load recovery yield buffer [mg/ml] mode monomer monomer HMWs comment 50 mM  0.6 flow- 88% 27% 0% sodium through acetate, pH 5.3 55 mM  0.4 flow- 71% 13% 0% sodium through acetate, pH 5.3 60 mM  0.4/ flow- 86/79/ 10/33/ 0/0/ sodium  0.55/ through 94/85% 29/53% 0/1% acetate,  1.4/ pH 5.3 11 70 mM  0.55/ flow- 83/80/ 72/58/ 0/0/ sodium  1.4/ through 86% 42% 3% acetate, 12 pH 5.3 85 mM  0.5/3/ flow- 75/95/ 64/80/ 0/5/ sodium 11 through 87% 76% 18%  acetate, pH 5.3 100 mM  0.3/ flow- 88/82/ 85/72/ 2/7/ sodium  0.55/ through 91% 81% 11%  acetate,  1.5 pH 5.3 40 mM  0.6 flow- 76% 57% 0% monomer sodium through also acetate, bound pH 5.5 (approx. 20%) 50 mM  0.6 flow- 77-84% 67-71% 3-5% monomer sodium through also acetate, bound pH 5.5 (<20%) 55 mM  0.5 flow- 74% 68% 4% sodium through acetate, pH 5.5 60 mM  0.4 flow- 68% 62% 6% sodium through acetate, pH 5.5 65 mM  0.6 flow- 101%  86% 10%  sodium through acetate, pH 5.5 70 mM  0.6 flow- 81% 76% 10%  sodium through acetate, pH 5.5 100 mM  1.6 flow- 117%  104%  23%  sodium through acetate, pH 5.5 40 mM  0.4 flow- 75% 71% 7% sodium through acetate, pH 5.7 50 mM  0.5 flow- 84-90% 78-87% 13-17% sodium through acetate, pH 5.7 55 mM  0.4 flow- 77% 73% 19%  sodium through acetate, pH 5.7 60 mM  0.4 flow- 69% 68% 12%  sodium through acetate, pH 5.7 10 mM  1 flow- 97% 55% 0% sodium through acetate, pH 6 15 mM  0.8 flow- 89% 79% 17%  sodium through acetate, pH 6 25 mM  0.8 flow- 78% 78% 21%  sodium through acetate, pH 6 50 mM  0.8 flow- 79% 79% 26%  sodium through acetate, pH 6 100 mM  1.6 flow- 123%  118%  35%  sodium through acetate, pH 6 10 mM  0.5 flow- 91% 91% 37%  potassium through phosphate, pH 7 100 mM  1.6 flow- 114%  114%  39%  monomer sodium through also acetate, 50 bound mM Tris, (approx. pH 7 20%) 100 mM 11 flow- 90% 43% 1% sodium through acetate, pH 5.1 110 mM 13 flow- 108%  73% 3% sodium through acetate, pH 5.1 120 mM 13 flow- 99% 70% 1% sodium through acetate, pH 5.1 60 mM 11 flow- 85% 53% 1% sodium through acetate, pH 5.3 85 mM 11 flow- 87% 76% 18%  sodium through acetate, pH 5.3

Example 6

(45) Purification with a Strong Cation Exchange Chromatography

(46) Prior to loading the column mutIL15-Fc was equilibrated in 10 mmol/l sodium citrate pH 5.0, 5.2 and 5.5, respectively, by dialysis. After the loading elution was either achieved in an increasing salt gradient or in a step elution profile using the equilibration buffer supplemented with 1 mol/l NaCl. Columns were operated at 2 ml/min.

(47) The mutIL15-Fc bound completely to SP-Sepharose FF (GE Healthcare, Uppsala, Sweden) at pH 5.0, 5.2 and 5.5 as well as to Toyopearl SP-650 M (Tosoh Bioscience GmbH, Stuttgart, Germany, order no. 43202) at pH 5.2. Elution of mutIL15-Fc started during re-equilibration of the column after end of load and before starting the salt gradient. In the main peak, monomeric mutIL15-Fc eluted earlier than aggregated protein, but with a total overlap not allowing significant separation of the monomeric form from the aggregates, independent of media, pH and elution conditions used.

(48) TABLE-US-00010 TABLE 10 Purification with sulfopropyl media. HMWs in the loading load main peak material buffer [mg/ml] mode elution with from . . . to Sepharose 10 mmol/l approx. bind- loading buffer, 2 to 46% sodium 0.4 and- linear gradient citrate elute from 0 . . . 1 pH 5.0 mol/l NaCl Sepharose 10 mmol/l approx. bind- loading buffer, 0 to 17% sodium 0.4 and- elution step citrate elute with 0.15 pH 5.0 mol/1 NaCl Sepharose 10 mmol/l approx. bind- loading buffer, 1 to 10% sodium 0.4 and- elution step citrate elute with 0.12 pH 5.0 mol/l NaCl Sepharose 10 mmol/l approx. bind- loading buffer, 4 to 11% sodium 0.4 and- elution step citrate elute with 0.15 pH 5.2 mol/l NaCl Toyopearl 10 mmol/l approx. bind- loading buffer, 0 to 14% sodium 0.4 and- linear gradient citrate elute from 0 . . . 1 pH 5.2 mol/l NaCl Toyopearl 10 mmol/l approx. bind- loading buffer, 0 to 12% sodium 0.4 and- elution step citrate elute with 0.12 pH 5.2 mol/l NaCl Sepharose 10 mmol/l approx. bind- loading buffer, 9 to 30% sodium 0.4 and- elution step citrate elute with 0.15 pH 5.5 mol/l NaCl

Example 7

(49) Purification with Protein a Chromatography, Size Exclusion Chromatography, and Carboxymethyl Sepharose

(50) The crude cultivation supernatant obtained in Example 1 was applied to the protein A affinity chromatography material MabSelectXtra (GE Healthcare, Uppsala, Sweden). The column was loaded with about 12 mg protein per ml column volume. The bound mutIL15-Fc fusion protein was eluted with 100 mM sodium acetate buffer at pH 3.2. An additional wash step prior to the elution was introduced to remove host cell DNA and host cell protein. About 80% mutIL15-Fc has been recovered.

(51) The protein A purified mutIL15-Fc was further purified using SEC. The recovered pooled monomer containing fractions contained approx. 95.5% purity (4.5% aggregates).

(52) For the carboxymethyl sepharose chromatography a MES buffer system with a pH gradient (pH 5.0/pH 8.0) as reported in Example 5 was used.

(53) As shown in FIG. 2 and summarized in Table 11 mutIL15-Fc was obtained with high purity when a moderately aggregated sample was loaded onto the column. Even at a relatively high column load of 7 mg/ml 99% pure monomer could be recovered at a yield of 80%.

(54) TABLE-US-00011 TABLE 11 Purification of pre-purified sample with CM-Sepharose FF loading load yield buffer [mg/ml] mode elution with monomer HMWs 180 mmol/l 7 bind- 60% step 80% 1% MES, and- gradient with pH 5.0 elute, buffer B = 30 (=buffer A) pH mmol/l MES, gradient pH 8.0

Example 8

(55) Purification with Protein a Chromatography, a Mixed-Mode Chromatography, and Carboxymethyl-Sepharose

(56) The crude cultivation supernatant obtained in Example 1 was applied to the protein A affinity chromatography material MabSelectSuRe (GE Healthcare, Uppsala, Sweden). The column was loaded with up to 14 mg protein per ml column volume. The bound mutIL15-Fc fusion protein was eluted with 100 mM citrate buffer at pH 3.6. An additional wash step prior to the elution was introduced to remove host cell DNA and host cell protein. Of from 76.6% to 78.4% mutIL15-Fc has been recovered (FIG. 3a) from the protein A affinity chromatography material.

(57) The protein A purified mutIL15-Fc was adjusted to the required pH value and salt concentration by dialysis and/or adding of appropriate stock solutions and afterwards applied to the mixed-mode chromatography column previously equilibrated with 25 mM acetate buffer pH 5.5. As mixed-mode chromatography material Capto MMC has been used (GE Healthcare, Uppsala, Sweden). The column was loaded with up to 20 mg protein per ml column volume. A solution comprising 25 mM sodium phosphate at pH 6.5 was used for recovery of the protein from the column (FIG. 3b).

(58) The Capto MMC eluate was dialyzed against 75 mM sodium acetate buffer pH 5.3. The CM-Sepharose column was loaded with up to 5 mg protein per ml column volume. The monomeric mutIL15-Fc fusion protein was recovered in the flow-through in the loading step. Aggregates and fragments were recovered with a single step elution with a solution comprising 200 mM sodium chloride (FIG. 3c). The aggregate removal capacity of the CM-Sepharose chromatography performed under the conditions described here is shown in Table 12. The yield of mutIL15-Fc is about 81% in this step.

(59) TABLE-US-00012 TABLE 12 Aggregate removal by CM Sepharose chromatography in flow-through mode as third chromatography step after affinity and mixed-mode chromatography HMW monomer LMW Load 6.70 91.67 1.63 Pool 1.80 97.31 0.89 Regenerate 20.45 67.39 12.15