Integrated automated filtration for separation, washing and drying of peptide crystals

Abstract

The present invention describes the integration of preparative crystallization, crystal separation, crystal washing and freeze-drying processes of insulin and insulin analogues into single continuous process using pressure filtration. The process facilitates time reduction and outlines the novel design of using multiple organic solvent washes and nitrogen gas purging for the removal of imbibed water and achieve final drug substance that meets the quality specifications.

Claims

1. An integrated process of crystal separation, washing and drying of insulin or insulin analogues, comprising the steps of: a) preparing a crystal suspension of insulin or insulin analogue using Zn Cl.sub.2 and adjusting the pH of the crystal suspension to between 4.5 and 8.5; b) pouring the crystal suspension obtained from step a) onto a filter fabric placed on an area ranging between 80 cm.sup.2 to 100 cm.sup.2 of a pressure filtration unit; c) purging the pressure filtration unit with nitrogen gas via an inlet port of the pressure filtration unit to obtain a filter cake; d) collecting a filtrate via outlet port of the pressure filtration unit; e) pouring 100% acetonitrile, butanol, ethanol, ethyl acetate or isopropanol as a solvent wash on the filtrate and purging nitrogen gas via an inlet port of the pressure filtration unit at a pressure of 1 to 2.5 bar; f) collecting the filtrate via an outlet port of the pressure filtration unit; g) repeating steps e) and f) to provide further solvent washes on the filtrate and thereby providing a filter cake; h) drying the filter cake by passing nitrogen gas through the filter cake at a pressure range of 1 bar to 2.5 bar; and i) removing the filter cake in the form of a dried powder, wherein the process is conducted as a single continuous process.

2. The integrated process of claim 1, wherein the insulin or insulin analogue has a crystal size in the range of 1-40 μm.

3. The integrated process of claim 1, wherein the temperature inside the assemblage pressure filtration unit is maintained in the range of 25° C. to 35° C.

4. The integrated process of claim 1, wherein the pressure filtration unit comprises a filter membrane with a pore size in the range of 1-15 μm.

5. The integrated process of claim 1, wherein the single continuous process is completed within 5 to 10 minutes.

6. The integrated process of claim 1, wherein the filter fabric is placed on a stainless steel mesh support of the pressure filtration unit.

7. The integrated process of claim 1, wherein the filter fabric is selected from a poly ester fabric and poly propylene fabric.

8. The integrated process of claim 1 wherein the volume of the 100% organic solvent is 5-10% of the volume of the crystal suspension.

9. The integrated process of claim 1, wherein the 100% organic solvent is acetonitrile.

10. The integrated process of claim 1, wherein the process is conducted without relying on centrifugal force.

Description

BRIEF DESCRIPTION OF DRAWINGS

(1) FIG. 1 represents flowchart of process of pressure filtration for insulin and insulin analogues purification

(2) FIG. 2 represents representative image of lab scale pressure-filter

(3) FIG. 3 represents drug substance from conventional (left) and new (right) process

(4) FIG. 4 represents variability chart for Product loss/recovery as Loss %

(5) FIG. 5 represents variability chart for residual moisture content in Drug Substance as LOD (loss on drying) %

(6) FIG. 6 represents variability chart for High Molecular Weight Products (HMWP) %

(7) FIG. 7 represents variance chart for Purity %

DETAILED DESCRIPTION OF INVENTION

Definitions

(8) The term ‘crystal suspension’ refer to a volume of solution prepared by addition of Zinc Chloride (ZnCl.sub.2) to insulin or insulin analogue followed by pH adjustment in the range of 4.5-8.5 using acid/alkali. Addition of zinc improves the physical stability of insulin by multimerization.

(9) The term ‘filter’ refer to a simple set up that is similar to a ‘Buchner funnel’ with the pore size of filter membrane used varying between ‘1-15 μm’ which does not involve any centrifugal force. The term ‘pressure filtration’ refer to a solid-liquid filtration aided by controlled pressure that pushes the solution against a filter membrane so that the solid substance (crystalline insulin and insulin analogue) is retained as a cake while the (organic solvent) mother liquor passes through as permeate. This exogenous pressure results in increased effective flow rate, high pressure gradient across the membrane and use of inert gas ensures stability to the finished product. Pressure filtration is an instantaneous process with an operating time of between 2-5 minutes on a lab scale and 45-60 minutes on an industrial scale.

(10) The term ‘insulin’ or ‘insulin analogue’ refer to 51 amino acid peptide hormone used for treatment of diabetes. It also refers to other insulin related forms such as insulin precursor or insulin analogue precursor, intermediate insulin or insulin analogue molecules obtained during downstream processes.

DESCRIPTION OF INVENTION

(11) Insulin and insulin analogues were expressed extracellularly using a recombinant Pichia based fermentation process. Transformed Pichia pastoris yeast strain are fermented in culture medium. The secreted Insulin or Insulin analogues were recovered by method of chromatography.

(12) Insulin and insulin analogue was crystallized from the chromatography elution pool. It was further made into crystal suspension in presence of Zinc Chloride (ZnCl.sub.2) and at an appropriate pH (in the range of 4.5 to 8.5) by addition of acid or alkali. Profile of crystal suspension (percentage solids, purity) was determined. The resultant insulin or insulin analogue crystals had high purity. Entire volume of crystal suspension was washed with water to remove the free precipitant used for crystallization and was taken forward for further processing of freeze drying.

(13) The crystal suspension was first processed for separation through pressure filtration process to obtain the dried cake. This separation process is devoid of gravity and centrifugal force. The dried cake had imbibed water, which is removed by steps of washing by organic solvent and purging nitrogen gas through it.

(14) The single continuous process can be depicted in following four steps. 1. Pressure Filtration of crystal suspension (initial separation of crystals) 2. Drying: washing with organic solvents for displacement of imbibed water 3. Drying: final purge of nitrogen to remove residual organic solvents 4. Automatic dispensing of dried powder in a cGMP environment.

(15) Similar integration of crystal separation, washing and drying can be performed at multiple preceding crystallization step during the downstream processing of insulin and insulin analogues. The raw material(s), reagent(s) and equipment(s) were used as per listed below in table 1, 2 and 3 respectively.

(16) TABLE-US-00001 TABLE 1 list of raw material Sr. No. Raw Materials Grade Manufacturer 1. Acetic acid HPLC Rankem 2. Acetonitrile (ACN) HPLC Rankem 3. Acetonitrile (ACN) Commercial Grade ASAHI KASEI 4. L-Arginine Reagent grade Sigma-Aldrich 5. 1-butanol Commercial Grade Deepak fertilizers 6. Ethanol Commercial Grade Changshu Yangyuan Chemicals 7. Ethyl acetate Commercial Grade Lakshmi Organic Industries 8. Filter membrane Not Applicable Sefar Filtration (India) Pvt. Ltd 9. HCl ACS Reagent Sigma-Aldrich 10. Methanol Commercial Grade Deepak fertilizers 11. 2-propanol Commercial Grade Deepak fertilizers 12. Trifluoroacetic acid Chromasolv Sigma-Aldrich 13. Water for injection Not Applicable Biocon

(17) TABLE-US-00002 TABLE 2 list of buffer/reagents and preparation Sr. No. Buffers/Reagents Preparation 1. 0.1% TFA in MilliQ Measure 1000 mL of MilliQ water. water Pipette out 1 mL of water and add 1 mL of HPLC grade TFA. 2. 1N HCl Measure 70 mL of water and then add 8.212 mL of concentrated HCl (HPLC grade). Make up the final volume to 100 mL using MilliQ water in a volumetric flask. 3. 0.01N HCl Measure 245 mL of water. Add 2.5 mL of 1N HCl (HPLC grade). Make up the volume to 250 mL with MilliQ water. 4. HMWP buffer Measure 600 mL of milliQ water. (High Molecular Add 0.65 g of Reagent grade Weight Transfer L-Arginine while on stirring. buffer) Make up the volume to 650 mL. Add 150 mL of Acetic acid and 200 mL of ACN. Filter the solution using a 0.45 μm followed by 0.2 μm filter and sonicate for 20 minutes.

(18) TABLE-US-00003 TABLE 3 List of equipment/instruments Sr. No. Name Model/Make 1. Analytical HPLC Agilent HPLC- 1200/Agilent technologies Agilent HPLC- 1100/Agilent technologies LC-2010CHT/Shimadzu 2. Column Advanced Chromatography Technologies: ACE 300-C18, 5 μm; 4.6 × 250 mm Waters: Insulin HMWP 7.8 × 300 mm 3. Pressure filter Prototype from BHS Sonthofen 4. Vacuum oven Servewell 5. Weighing balance Mettler Toledo 6. Water bath Julabo 7. Water bath Equitron 8. Sonicator Servewell 9. Vortexer Shalom 10. Nitrogen Gas Biocon cylinder 11. Cold room Blue star 12. Deep freezer Vestfrost

(19) The trials were performed in a lab scale pressure filter set up with an approximate filter cross-sectional area of 80 cm.sup.2 to 100 cm.sup.2. It also comprises of a pressure inlet for compressed inert gas (Nitrogen), a pressure gauge and safety release valve, sample inlet and outlet ports and a water jacket (FIG. 2).

(20) Variable parameters were wash solvent, solvent volume, number of washes, time of nitrogen purging for each wash and an optional change of filter media, which is dependent on the size of crystals being processed. Process performance and quality attributes of the obtained drug substance were measured by analysing percentage of loss of crystals in permeate and loss on drying (LOD), purity and molecular weight (HMWP). The temperature of the set up was maintained in range of 25° C. to 35° C.

(21) The cake formed by process of pressure filtration of insulin and insulin analogue crystal suspension, had requisite thickness and texture. However, the goal was to dry the cake to achieve acceptable standards of LOD and very importantly to integrate the crystallization-freeze drying of the conventional process into a single continuous process. To achieve sufficient displacement of imbibed residual water from filter cake, single or multiple wash strategies with different organic solvents were studied. The wash steps were also designed to enable effective removal of precipitants, and residual solvents that are present in reverse phase (chromatography) elution pool and obtain a drug substance that complies with the universal quality and regulatory specifications.

(22) The assembly of instrument(s) for achieving the dried powder in continuous and automated manner was as follows: i. The equipment was an SS cylindrical assemblage with jacketing to maintain required internal temperature as well as provision for passing an (inert) gas. Nitrogen gas cylinder and Julabo water chiller were used for these trials. The temperature was maintained in the range of 25-35° C. ii. A filter fabric of 80 cm.sup.2 to 100 cm.sup.2 area was placed on an SS mesh support and housed in the assemblage. Fabric filter was selected from PET 1703 (Poly ester fabric) and SK-011 (Poly propylene fabric). iii. The assemblage was sealed with a SS disc lined with silicon gasket, and clamps were fastened to make the compartment airtight, followed by pouring the homogenous neat suspension from the top of the vessel to commence pressure filtration. iv. Nitrogen gas at bar pressure in the range of 1 bar to 2.5 was applied via NRV/inlet port. v. Filtrate/permeate was collected via the sample outlet port at the bottom. vi. Steps ‘iii’ and ‘iv’ were repeated for single or multiple cake washes using organic solvent. vii. Drying was performed by addition of an organic solvent and further passing N.sub.2 gas for about three minutes after which the assemblage was unfastened. viii. The fabric containing the cake was carefully removed and gently tapped to recover the cake which was then transferred into a sterile airtight amber colour container and stored at −20° C. until further analysis.

(23) Two trials were conducted. First, was preliminary screening (Trial I) and second (Trail II) was design of experiment (DOE).

(24) Preliminary screening trials (trial 1) were performed in a lab scale prototype of industrial pressure filter (FIG. 2) with different organic solvents using the design described in Table 4. In preferable embodiment, acetonitrile was considered best working organic solvent. Filter fabrics were studied and screened for their pore size, i.e. retention of crystals, texture, workability on commercial scale and resilience/inertness to the pH and temperature conditions of the process. Two types of filter fabric were used viz. poly ester fabric and poly propylene fabric. Organic solvents were chosen based on their ability to displace water from the cake and their volatile nature (table 4). The FDA guidelines to acceptable solvents and their residual limits in final drug substance were also considered for the selection of solvents.

(25) TABLE-US-00004 TABLE 4 Solvent screening design % Solvent volume based Trial on crystal suspension number Solvent volume Washes 1. Acetonitrile (ACN) 5 1 2. 5 2 3. 10 1 4. Butanol 5 1 5. 5 2 6. 10 1 7. Ethanol 5 1 8. 5 2 9. 10 1 10. Ethyl acetate 5 1 11. 5 2 12. 10 1 13. Isopropanol 5 1 14. 5 2 15. 10 1 16. Water 5 1 17. 5 2 18. 10 1

(26) Solvents and wash combinations yielding high percentage LOD (loss on drying) values in the above design were eliminated. Data set is presented in table 5 and table 6.

(27) The captured data set in table 5 and table 6 suggested that the resultant insulin and insulin analogue drug substance has comparable quality attributes to the drug substance made from traditional crystallization-freeze drying process with optimal recovery of product (FIG. 3).

(28) Two washes of acetonitrile planned in Trail I has satisfactory outcome with respect to percentage LOD. However, to fall in-line with regulatory requirements a further reduction in LOD was desired. Parameters that were studied in this experimental setup included a) number of washes; b) drying time and c) temperature during filtration to study the probability of temperature accelerating the evaporation of the residual organic solvent trapped in filter cake. For Trail II (DOE), HMWP was also determined.

(29) TABLE-US-00005 TABLE 5 Solvent Screening Result Crystallization 3 Neat suspension (NS) Wash 1 (W1) Wash 2 (W2) NS NS NS W1 W1 W2 W2 Drying DS conc. volume permeate volume permeate volume permeate time weight Wash Solvent % (g/L) (mL) (mL) (mL) (mL) (mL) (mL) (min) (g) LOD % 16 Jan. 2017 LP1-5%-W1 14.65 300 285 15 21 NAP NAP 1.5 5.52 26 LP1 5% W 1 & 2 14.65 300 290 15 13 15 17.5 1.5 5.26 20.46 LB1-5%-W1 14.65 300 280 15 18 NAP NAP 1.5 6.33 35.14 LB1-5%-W 1 & 2 14.65 300 290 15 13.5 15 20 2.5 5.88 36.22 LA-5%-W1 14.65 300 290 15 14.3 NAP NAP 1.5 4.96 15.99 LA-5%-W 1 & 2 14.65 300 295 15 12.5 15 13.5 1.5 4.88 9.34 LA-10%-W1 14.65 300 290 30 29 NAP NAP 2 5.3 21.25 LP1-10%-W1 14.65 300 290 30 31.5 NAP NAP 2 5.43 21.14 LB1-10%-W 1 14.65 290 275 30 30 NAP NAP 2 5.5 25.1 LP1-5%-W1 14.65 200 190 10 9.7 NAP NAP 1.5 3.81 NAV 19 Jan. 2017 LE-5%-W 1 & 2 14.65 300 290 15 9.8 15 18 2 5.66 22.35 LEA-5%-W1 14.65 300 290 15 21.5 NAP NAP 2 7.07 37.84 LEA-5%-W 1 & 2 14.65 300 290 15 9 15 9.6 2 6.67 34.11 LEA 10% W1 & 2 14.65 300 280 30 25 5 30 28 2 6.87 36.03 LE-5%-W 1 & 2 T1 14.65 300 290 15 18.5 15 16 2 5.44 15.75 LE-5%-W 1 & 2 T2 14.65 300 285 15 13 15 16.5 2 6.06 16.43 LA-5%-W 1 & 2 T1 14.65 300 300 15 15 15 14 2 5.38 10.52 LA-5%-W 1 & 2 T2 14.65 300 305 15 10 15 9.5 2 5.7 11.53 LOD values in BOLD are actual values, rest are theoretical values based on neat suspension product concentration. LA: Acetonitrile, LP1: Isopropanol, LB1: Butanol, LEA: Ethyl acetate, NS: Neat suspension, W1/2; Wash 1 and Wash 2, T1/2; Trial 1 and Trial 2; LOD: Loss on drying; NAP: Not applicable, NAV: Not available

(30) A full factorial experimental design along with the response data is presented in table 6. Fifty experiments were performed discretely as below (table 6).

(31) TABLE-US-00006 TABLE 6 Experimental design with response data Drying Drying Sr. Wash I Wash II Wash III Temp Time % % % % No. (% V/V) (% V/V) (% V/V) (° C) (min) LOD Purity HMWP Loss 1 5 10 5 35 3 19.18 99.20 0.158 1.41 2 10 10 0 35 2 22.23 99.25 0.163 1.89 3 10 5 10 35 2 17.66 99.18 0.159 4.50 4 10 10 5 35 3 16.38 99.22 0.162 1.59 5 10 10 10 25 2 12.77 99.06 0.167 0.19 6 5 5 0 25 2 18.70 98.86 0.173 0.49 7 7.5 7.5 5 30 2.5 13.90 99.25 0.156 1.58 8 10 10 10 35 3 16.37 99.24 0.165 1.54 9 5 10 10 35 3 13.06 99.17 0.158 2.62 10 5 10 10 25 2 14.60 98.72 0.162 0.17 11 5 5 10 25 2 11.99 99.00 0.163 0.24 12 10 10 5 25 2 9.14 98.71 0.158 0.13 13 10 10 5 35 2 13.25 99.26 0.164 3.25 14 5 5 0 35 2 14.01 99.35 0.225 0.23 15 10 5 5 35 3 21.74 99.11 0.218 1.11 16 5 5 5 25 3 15.66 98.89 0.171 0.31 17 10 5 5 25 3 14.60 99.19 0.164 0.25 18 10 5 0 25 3 20.36 99.06 0.169 0.18 19 10 10 0 35 3 16.58 99.20 0.172 1.78 20 5 5 0 25 3 12.13 98.95 0.176 0.34 21 5 5 10 25 3 14.58 98.84 0.181 0.33 22 10 10 10 35 2 14.47 99.22 0.152 1.78 23 10 5 10 25 3 9.56 99.04 0.164 0.15 24 10 10 0 25 2 15.70 99.03 0.158 0.16 25 5 10 0 25 3 15.35 98.97 0.164 0.19 26 10 5 5 35 2 15.54 99.18 0.157 1.99 27 5 10 0 35 3 16.10 99.37 0.159 2.61 28 5 10 0 35 2 12.93 99.19 0.157 1.66 29 10 10 0 25 3 12.66 99.06 0.165 0.18 30 5 5 10 35 2 18.12 99.20 0.160 1.42 31 5 5 0 35 3 15.35 99.27 0.159 5.10 32 10 5 5 25 2 16.88 99.05 0.166 0.28 33 5 5 5 25 2 17.11 98.53 0.170 0.30 34 10 10 10 25 3 9.33 99.14 0.166 0.28 35 10 5 10 25 2 20.35 98.73 0.162 0.22 36 5 10 5 35 2 13.63 99.19 0.161 2.02 37 10 5 0 35 2 18.68 99.35 0.159 1.57 18 10 5 0 25 2 23.54 99.13 0.164 0.16 39 5 5 5 35 3 17.18 99.19 0.158 1.83 40 5 10 10 25 3 15.21 99.08 0.169 0.24 41 5 10 5 25 2 13.80 98.85 0.165 0.27 42 5 10 5 25 3 11.79 99.08 0.168 0.25 43 10 5 10 35 3 18.13 99.25 0.157 1.87 44 5 10 10 35 2 17.66 99.20 0.158 1.39 45 5 10 0 25 2 15.81 98.92 0.171 0.33 46 5 5 10 35 3 19.76 99.21 0.161 1.60 47 7.5 7.5 5 30 2.5 13.81 99.23 0.159 1.32 48 10 10 5 25 3 12.42 99.15 0.162 0.16 49 10 5 0 35 3 20.44 99.22 0.158 5.33 50 5 5 5 35 2 13.68 99.25 0.225 0.83

(32) The drug, substance from Trials 12, 23 and 34 had the lowest LOD values i.e. of less than 10% (table 7), while drug substance from trials 11 and 42 had LOD values less than 12%. There were no significant changes in purity and HMWP profiles, and per-step losses were found to be well within range.

(33) The results of variability are also represented in FIGS. 4, 5, 6 and 7 for percentage loss of product, percentage LOD, percentage HMWP and percentage purity respectively,

(34) TABLE-US-00007 TABLE 7 Summary of results of DOE (Trial II) Trial WASHES (% V/V) number W I W II W III % LOD 12 10 10 5 9.14 23 10 5 10 9.56 34 10 10 10 9.33

(35) From the multivariate studies, it was found out that more number of washes with organic solvent (3>2>1) resulted in ample displacement of imbibed water from filter cake. Three washes including wash 1 at 10%, wash 2 at 5% or 10% and wash 3 at 5% or 10%, were found to be most effective in reducing LOD. Volume of washes can be fine-tuned further as per process requirements. Increased drying time helped in forcing out some solvent imbibed in the cake post-washing. Increasing the temperature during filtration did not reduce the LOD of the drug substance, but it did not significantly affect the profile of the drug substance either. Thus, the temperature was maintained in the range of 25° to 35° C. Based on this multivariate study further trials can be designed and executed to further reduce the LOD to <4%.

(36) The results of this multivariate study were satisfactory and drew immediate attention to a novel process that is shorter, continuous, less resource intensive and automated.