Application Note USD Capture of Human Serum Albumin from Plasma Using HyperCel STAR AX Salt Tolerant Anion Exchange Chromatography Sorbent

Transcription

1 Application Note USD 2871 Captre of Hman Serm Albmin from Plasma Using HyperCel STAR AX Salt Tolerant Anion Exchange Chromatography Sorbent Impact on Process Economics in a Two-Step Streamlined Process

2 1. Smmary This stdy describes the development of a model two-step prification process for serm albmin from hman plasma sing HyperCel STAR AX salt tolerant anion-exchange sorbent as captre step. Comparative assessment of HyperCel STAR AX sorbent with a conventional rigid DEAE agarose sorbent, typically sed in plasma fractionation, was performed. Design of Experiment (DoE) / high throghpt screening optimization of prification conditions was performed for both sorbents. A process economics analysis was performed to evalate the impact of the elimination of feedstock diltion on process cost. 2. Introdction Captre of netral to acidic proteins generally ses anion exchange (AEX) sorbents. Conventional anion exchangers reqires feed diltion to decrease ionic strength, or needs diafiltration to achieve sfficient protein binding capacity. These operations, however, increase bffer consmption and processing time, and increase overall process cost. Using a salt tolerant anion exchanger sch as HyperCel STAR AX sorbent wold allow direct captre from ndilted feedstocks and reslt in significant process economics benefits at prodction scale. HyperCel STAR AX sorbent was evalated in a two-step prification process withot diltion, taking the prification of Hman Serm Albmin (HSA) from cryo-poor plasma as a model (Figre 1). Table 1 Properties of HyperCel STAR AX Sorbent Average particle size 8 μm Ion exchange ligand Primary amine Dynamic binding capacity 1 at condctivity 15 ms/cm >1 mg BSA/mL within ph range Recommended operating range of feedstock condctivity 2 15 ms/cm Recommended cleaning conditions 2 1 M NaOH 1 Determined sing 5 mg/ml BSA in 25 mm Tris-HCl,.14 M NaCl at 2 minte residence time 2 Injection of 5 colmn volmes (CV) of.5 1 M NaOH, 1 hor contact time Figre 1 Overview of Process Development for Prification of HSA from Cryo-Poor Plasma Clarified Cryo-Poor Plasma Captre Step Bind/ Elte on Anion Exchange HyperCel STAR AX Sorbent Rigid DEAE Agarose Sorbent Albmin in Eltion Immnogloblins, Transferrin in Flowthrogh Second Step Bind/ Elte on Cation Exchange S HyperCel Sorbent Albmin in Eltion 2

3 3. Materials and Methods 3.1. Plasma Preparation and Analytical Methods Frozen hman plasma was thawed at 4 C for one day and centrifged at 45 x g for 1 mintes at 4 C to remove small aggregates, and then filtered throgh a Pall.2 μm Spor 2 membrane. SDS-PAGE in non-redcing conditions Immnogloblin (Ig) qantification Transferrin (Trf) qantification Albmin qantification Npage 4-12% Bis-Tris precast gels, staining with Coomassie SimplyBle SafeStain (Life Technologies) Protein A HPLC colmn (Applied Biosystems) ELISA assay kit (Cygns Technologies) Bromocresol green colorimetric assay (Fisher Diagnostics) Recovery and prity were calclated as shown below. Prity was measred as the ratio between the amont of albmin and the total amont of major plasma proteins (albmin, immnogloblins and transferrin) measred in the samples. Albmin recovery (% of load) = Albmin in eltion (mg) x 1 Albmin in load (mg) Albmin prity (%) = Albmin (mg) x 1 Ig (mg) + Trf (mg)+ Albmin (mg) 3.2. Chromatography Rns DBC vs. Plasma Diltion on HyperCel STAR AX Sorbent and Rigid DEAE Agarose Sorbent DBC was determined in.5 cm I.D. colmns packed with 1 ml of HyperCel STAR AX sorbent or rigid DEAE agarose sorbent. DBC at 1% breakthrogh (DBC 1%BT ) was evalated by qantification of albmin in the flowthrogh fractions after loading ndilted and 1.6- and 3-fold dilted plasma (respectively 11, 7 and 4 ms/cm). Flow rate Eqilibration bffer Sample load Wash Strip.5 ml/min (2 minte residence time) 5 mm Tris-HCl, ph 7.6 at 11, 7 or 4 ms/cm (1 CV) Plasma (ndilted or 1.6-fold or 3-fold dilted) Eqilibration Bffer (1 CV) 5 mm Na acetate, ph M NaCl (1 CV) HSA Captre Rns on Colmn The optimal conditions determined by 96-well plate / high-throghpt screening for HyperCel STAR AX sorbent and rigid DEAE agarose sorbent (see 3.3) were transferred to chromatography rns on colmns with load at 6% of DBC 1%BT of ndilted plasma on HyperCel STAR AX sorbent, and 3-fold dilted plasma on rigid DEAE agarose sorbent. Flow rate Eqilibration bffer Sample load Wash Eltion Strip CIP.5 ml/min 5 mm Tris-HCl, ph 7.6, adjsted to optimal condctivity for each sorbent Plasma at optimal diltion Determined by 96-well plates optimization (1 CV) 5 mm Na acetate, ph M NaCl (5 CV) 1 M NaOH (2 CV) 3

4 3.3. DoE Optimization of Wash and Eltion Conditions on 96-Well Plates High throghpt screening optimization of ph and condctivity conditions based on DoE was performed. An extended central composite design was established sing the Minitab software, as described in Table 2, for the for parameters Wash ph, Wash condctivity, Eltion ph, Eltion Condctivity. Table 2 Test Conditions for High Throghpt Screening Optimization of Wash and Eltion Conditions of HSA Captre Step on AEX Sorbents Wash Eltion Condctivity Eltion Condctivity Sorbent Load Wash ph (ms/cm) ph (ms/cm) HyperCel STAR AX Cryo-poor plasma 5.5 to to to to 5 Rigid DEAE Agarose 3-fold dilted cryo-poor plasma Based on this DoE, 27 combinations were tested for their impact on albmin yield and prity. The screening of ph / condctivity combinations was done sing AcroPrep Advance 96-well filter plate containing 1 μl of sorbent following this seqence: Eqilibration Load Wash Eltion Volme (µl) Incbation time (min) HSA yield and prity obtained for each condition were recorded, data modelled and models validity checked. Contor plots were generated in order to analyze the impact of each factor on albmin recovery yield and prity, and optimal conditions were determined Process Economics Analysis The process economics analysis was condcted sing the BioSolve Cost of Goods (CoG) analysis software from Biopharm Services Ltd. A concentration of 5 g/l HSA in plasma was fixed. CoG analysis was performed considering 2, L batches, with an average nmber of 2 batches per year, corresponding to the processing of 4, L of plasma per year, meeting the average capacity of a fractionation plant. The HSA recovery for diltion steps was considered to be 1%. The flow rate, colmn bed height, and nmber of cycles were kept constant, respectively 2 cm/hr, 2 cm and 2 cycles. 4. Reslts and Discssion 4.1. Optimization of a Two-Step Process for HSA Prification Effect of Feed Condctivity on Dynamic Binding Capacity (DBC) of Salt-Tolerant vs. Conventional AEX Sorbent In order to address the effect of feed condctivity on DBC, evalation of DBC on HyperCel STAR AX sorbent and rigid DEAE agarose sorbent was performed sing plasma at different diltions (Figre 2). DBC vales obtained on HyperCel STAR AX sorbent were maintained arond 3 mg/ml with load at condctivities ranging from 4 ms/cm (3-fold dilted) to 11 ms/cm (ndilted). In contrast, an increase of load condctivity reslted in a significant decrease of DBC of the rigid DEAE agarose sorbent. 4

5 This data confirmed the salt tolerant behavior of HyperCel STAR AX sorbent and indicated that this sorbent cold be sed for high capacity captre of HSA directly from non dilted plasma. In this condition, DBC was more than 2-fold higher on HyperCel STAR AX sorbent than on the standard AEX sorbent. Additionally, the DBC of HyperCel STAR AX sorbent was maintained in a broad range of condctivities, bringing more process robstness and flexibility, by accommodating potential feedstock ionic strengths variations dring the process. Figre 2 Dynamic Binding Capacity of HyperCel STAR AX Sorbent and Rigid DEAE Agarose Sorbent for Hman Serm Albmin Using Undilted (11 ms/cm), 1.6-fold Dilted (7 ms/cm) and 3-fold Dilted (4 ms/cm) Plasma 4 DBC for HSA (mg/ml) ms/cm 7 ms/cm 11 ms/cm HyperCel STAR AX Sorbent DEAE Rigid Agarose Sorbent Optimization of Process Conditions for HSA Captre Optimization of wash and eltion conditions was performed sing DoE / high-throghpt screening on AcroPrep Advance 96-well filter plates (Figre 3 and Table 2). Optimal wash and eltion conditions allowing best eltion yield combined to lowest contaminants content (Immnogloblins + Transferrin) in eltion were determined by response srface modelling analysis. 5

6 Figre 3 96-Well Plate / High Throghpt Screening for Optimization of HSA Captre 1. Design of Experiment (DoE) Critical parameters (wash and eltion ph and condctivity) Qality attribtes (eltion yield, prity) Bffer Sorbent bed Filter plate 2. Screening on 96-Well Plates AcroPrep Advance Filter Plates 3. Analytical Testing HSA qantification (Bromocresol Green) Qantification of Transferrin (ELISA) Qantification of IgG (Protein A-HPLC) AU Min 4. Reslt Analysis Design space for optimm performances Eltion ph HCP (ppm) Load ph Transfer to Colmn Response srface modelling of HyperCel STAR AX sorbent data revealed that HSA prity was significantly impacted by wash conditions (Figre 4A). Increase in wash condctivity led to higher HSA prity, with an optimal zone for wash condctivity >15 ms/cm. Optimal eltion conditions zone was determined as ph 3.5 to 4.2 at condctivities of 2 to 27 ms/cm (Figre 4B). Finally, optimal combinations of wash and eltion conditions providing estimated yield and prity 9% were chosen as follows according to the model predictions: Wash: ph 7.5, 2 ms/cm, allowing maximal prity. Eltion: ph 4., 2 ms/cm, providing high eltion yield and the possibility to directly load eltion from HyperCel STAR AX sorbent on an orthogonal cation exchange sorbent (see process in Figre 1). 6

7 Figre 4 Optimization of Wash and Eltion Conditions for HSA Prification on HyperCel STAR AX Sorbent. (A) Effect of Wash Conditions on Prity. (B) Effect of Eltion Conditions on Eltion Yield Wash Condctivity (ms/cm) (A) Optimal wash condition Prity (%) < > 98 Hold vales E ph 3.5 E cond. 2. Eltion Condctivity (ms/cm) (B) Optimal eltion condition Yield (%) < > 95 Hold vales W ph 7.5 W cond Wash ph Eltion ph Fixed conditions: (A) Eltion at ph 4, condctivity 2 ms/cm; (B) Wash ph 7.5, condctivity 2 ms/cm. A similar optimization carried ot with the rigid DEAE agarose sorbent yielded the following optimal combinations of wash and eltion conditions providing yield and prity 9%: Wash: ph 7.5, 9 ms/cm Eltion: ph 4., 25 ms/cm Transfer of Conditions to Colmn Chromatography Conditions optimized on 96-well plates for both sorbents were next applied to colmn rns (Figre 5). For HyperCel STAR AX sorbent, elimination of contaminants with high condc tivity wash as predicted by DoE optimization was efficient, yielding 99% pre HSA. Eltion driven by ph drop also reslted in a satisfying eltion yield (9%). Application of optimized conditions inclding a high condctivity wash on rigid DEAE agarose sorbent also allowed to obtain pre HSA (98%), whereas a combination of low ph and high condctivity was reqired to obtain satisfying recovery at eltion (yield 9%). On rigid DEAE agarose sorbent, sing alternative eltion condition (low ph and condctivity, as for HyperCel STAR AX sorbent) preparing a direct load to the following cation exchanger reslted in good prity bt only partial eltion (73% yield, see Table 3). 7

8 Figre 5 HSA Captre on HyperCel STAR AX Sorbent and Rigid DEAE Agarose Sorbent mau 25 2 UV ph Condctivity FT from neat plasma load Wash 1 Wash 2, ph 7.5, 2 ms/cm FT W1 W2 Eltion, ph 4., 2 ms/cm E Albmin IgG FT W1 W2 E P 15 Transferrin Albmin ml (A) HyperCel STAR AX Sorbent Eltion, ph ms/cm mau E 25 Albmin FT W1 W2 E P 2 15 FT from 3x dilted plasma load Wash 1 Wash 2 ph 7.5, 9 ms/cm FT W1 W2 ph drop delayed IgG Transferrin Albmin ml (B) Rigid DEAE Agarose Sorbent Rns were performed in optimal conditions depicted in Table 3. Chromatograms (left) and SDS- PAGE gels (right). FT: Flowthrogh; W1: Wash 1; W2: Wash 2; E: Eltion, P: Plasma Load, non-dilted for (A) and 3-fold dilted for (B). Table 3 Performance of HyperCel STAR AX Sorbent and Rigid DEAE Agarose Sorbent for Captre Prification of HSA on Colmn in Optimal or Alternative Conditions (Load = 6% of DBC 1%BT ) Recovery HSA Sorbent / Conditions Load Wash Conditions Eltion Conditions Yield (%) Prity (%) HyperCel STAR AX / Cryo-poor plasma ph 7.5, 2 ms/cm ph 4., 2 ms/cm Optimal Rigid DEAE Agarose / 3-fold dilted ph 7.5, 9 ms/cm ph 4., 25 ms/cm Optimal cryo-poor plasma Rigid DEAE Agarose / 3-fold dilted ph 7.5, 9 ms/cm ph 4., 2 ms/cm Alternative cryo-poor plasma 8

9 Second HSA Prification Step on Cation Exchange Sorbent HSA prified on HyperCel STAR AX sorbent and rigid DEAE agarose sorbent was next sed as loading feed to optimize second prification step of HSA on S HyperCel cation exchanger. To obtain condctivity compatible with CEX rn, the eltion from rigid DEAE agarose sorbent in optimal conditions had to be dilted 5 times (condctivity 6 ms/cm). HSA fractions elted at low condctivity were loaded withot modification for optimization of the second step. After optimization of S HyperCel chromatography on 96-well plates (not shown), HSA was elted from this second colmn at ph 7, 15 ms/cm with 95% recovery and more than 99% prity Process Economics Analysis Process Scenarios Different scenarios derived from the present stdy were compared in a process economics analysis (Figre 6). The process sing HyperCel STAR AX sorbent (Scenario 1) is the easiest since optimal performance is obtained withot plasma diltion. In contrast, different options exist when sing rigid DEAE Agarose sorbent depending on diltion. Using different diltion steps before captre and second chromatography step leads to different performance as smmarized in Figre 6. Figre 6 Process Scenarios Used for Process Economics Analysis Cryo Poor Plasma ph 7.6, Condctivity 11 ms/cm Scenario 3 Plasma Diltion 3X 1% Yield, ph 7.6, Condctivity 4 ms/cm Scenario 1 Scenario 2 Scenario 3a Scenario 3b HyperCel STAR AX 89% Yield, DBC 3 mg/ml, Eltion in 5 mm Na Acetate, ph 4. Rigid DEAE Agarose 73% Yield, DBC 11 mg/ml, Eltion in 5 mm Na Acetate, ph 4. Rigid DEAE Agarose 73% Yield, DBC 4 mg/ml, Eltion in 5 mm Na Acetate, ph 4. Rigid DEAE Agarose 89% Yield, DBC 4 mg/ml, Eltion in 5 mm Na Acetate,.3 M NaCl, ph 4. Diltion 5X post DEAE 1% Yield S HyperCel 95% Yield, DBC 6 mg/ml, Eltion in 5 mm Na Phosphate,.15 M NaCl, ph 7. S HyperCel 95% Yield, DBC 6 mg/ml, Eltion in 5 mm Na Phosphate,.15 M NaCl, ph 7. S HyperCel 95% Yield, DBC 6 mg/ml, Eltion in 5 mm Na Phosphate,.15 M NaCl, ph 7. S HyperCel 95% Yield, DBC 6 mg/ml, Eltion in 5 mm Na Phosphate,.15 M NaCl, ph Cost of Goods (CoG) Comparison Process 1: CoG corresponding to the HyperCel STAR AX sorbent process (Scenario 1) was compared to each of the scenarios sing rigid DEAE agarose sorbent (Figre 7). In all cases, the total cost of goods in US$/g of HSA was lower with HyperCel STAR AX sorbent, allowing 14 to 67% savings compared to processes inclding rigid DEAE agarose sorbent. 9

10 Process 2 corresponds to the least favorable conditions, with an increase in costs of 67% compared to Scenario 1. For one part, this is de to the lower DBC obtained at captre step with ndilted plasma, which leads to a higher nmber of rns to process the same amont of HSA. This reslts in a higher bffer consmption, water sage and material / tankage costs. In addition, lower throghpt in kg/year (-22%) was obtained de to the poorer eltion yield from rigid DEAE agarose sorbent with low salt ph. Process 3a showed less extra-costs compared to process 1, althogh the savings sing HyperCel STAR AX sorbent were close to 2%. In this case, the higher cost/g of HSA was mainly driven by the decreased throghpt, de to the low eltion yield from rigid agarose DEAE as in process 2. Process 3b was the most cost-effective process based on the se of rigid DEAE agarose sorbent. In this case, cost increase compared to scenario 1 is de to the two diltion steps applied before first and second prification rns. Since these diltions allow optimal conditions for all chromatography steps, the throghpt is preserved and a lowest cost increase is anticipated. However, there is still a 14% cost difference, and several nit operation steps are added de to diltion reqirements. This leads still to a significant advantage of the HyperCel STAR AX sorbent based process. To estimate the impact of the 5-fold diltion between first and second step on the cost advantage of process 1 vs. process 3b, simlation of a process employing a low (2-fold) diltion factor was done. This scenario cold be applied if a salt-tolerant cation exchange sorbent was sed for the second step. In this case, the process cost sing HyperCel STAR AX sorbent was still redced vs. a process based on rigid DEAE agarose sorbent (4%). Figre 7 Cost Comparison Between Scenario 1 (HyperCel STAR AX sorbent) with Scenarios 2, 3a and 3b (Rigid DEAE Agarose Sorbent) Rigid DEAE agarose: No diltion, Scenario 2 1 diltion, Scenario 3a 2 diltions, Scenario 3b Variaton vs. HyperCel STAR AX (%) 2% 1% % -1% -2% -3% -4% -5% CoG (US$/g) CoG (US$/g) Annal CoG (US$) Annal CoG (US$) Capital Capital Materials Materials Consmables Consmables Labor Labor Water sage (m 3 per batch) Water sage (m3 per batch) Throghpt (kg (kg of doses or doses per year) per year) -6% -7% Percentage variations from scenarios sing rigid DEAE agarose sorbent to scenario 1 are presented for global cost of goods in US$/g and US$/year. Savings on capital, materials, consmables and labor are expressed as the percentage of cost variation; variation in water sage and throghpt are expressed as percentage variation in volme (m 3 ) sed per year and kg of HSA prodced per year respectively. 1

11 5. Conclsion HyperCel STAR AX sorbent can efficiently captre HSA from plasma withot diltion, allowing good elimination of major plasma contaminants sch as immnogloblins and transferrin. DBC of HyperCel STAR AX sorbent is maintained in a wide range of condctivities, bringing process flexibility and robstness. DBC of HyperCel STAR AX sorbent with crde plasma is more than 2-fold higher than that of a conventional rigid DEAE agarose sorbent. The se of HyperCel STAR AX sorbent allows developing a streamlined two-step process for HSA prification, avoiding any diltion reqirement. Process economics analysis indicates that elimination of diltion steps thanks to the se of HyperCel STAR AX sorbent provides strong economical benefits (redced capital and consmable cost, redced water sage) compared to conventional anion exchange sorbents. Corporate Headqarters Port Washington, NY, USA toll free (USA) phone biopharm@pall.com Eropean Headqarters Friborg, Switzerland +41 () phone LifeSciences.EU@pall.com Asia-Pacific Headqarters Singapore phone sgcstomerservice@pall.com Visit s on the Web at s at biopharm@pall.com International Offices Pall Corporation has offices and plants throghot the world in locations sch as: Argentina, Astralia, Astria, Belgim, Brazil, Canada, China, France, Germany, India, Indonesia, Ireland, Italy, Japan, Korea, Malaysia, Mexico, the Netherlands, New Zealand, Norway, Poland, Perto Rico, Rssia, Singapore, Soth Africa, Spain, Sweden, Switzerland, Taiwan, Thailand, the United Kingdom, the United States, and Venezela. Distribtors in all major indstrial areas of the world. To locate the Pall office or distribtor nearest yo, visit The information provided in this literatre was reviewed for accracy at the time of pblication. Prodct data may be sbject to change withot notice. For crrent information conslt yor local Pall distribtor or contact Pall directly. 212, Pall Corporation. Pall,, AcroPrep, HyperCel and Spor are trademarks of Pall Corporation. NPage and SimplyBle are trademarks of Life Technologies. Minitab is a trademark of Minitab Inc. Filtration.Separation.Soltion is a service mark of Pall Corporation. indicates a trademark registered in the USA and TM indicates a common law trademark. 1/13, PDF, GN USD 2871