MRNA PURIFICATION BY TANGENTIAL FLOW FILTRATION

Abstract

The present disclosure provides a method of purifying mRNA molecules comprising (Ia) purifying precipitated mRNA molecules from a suspension comprising precipitated mRNA molecules, (Ib) washing and dissolving the purified precipitated mRNA molecules, (IIa) purifying the mRNA molecules using a solution comprising a chelating agent, followed by (IIb) washing the purified mRNA molecules, wherein steps (Ia) to (IIb) are performed using tangential flow filtration.

Claims

1. A method of purifying mRNA molecules, said method comprising (Ia) purifying precipitated mRNA molecules from a suspension comprising precipitated mRNA molecules using a first solution, (Ib) washing and dissolving the purified precipitated mRNA molecules obtained from step (Ia) using a second solution, (IIa) purifying the mRNA molecules from the dissolved mRNA molecules obtained from step (Ib) using a third solution comprising a chelating agent, followed by (IIb) washing the purified mRNA molecules obtained from step (IIa) using a fourth solution, wherein steps (Ia) to (IIb) are performed using tangential flow filtration.

2. The method according to claim 1, wherein the chelating agent is EDTA.

3. The method according to claim 1 or 2, wherein the third solution has a pH between 1 and 10.

4. The method according to any of the preceding claims, wherein the third solution comprises MOPS buffer.

5. The method according to any of the preceding claims, wherein steps (Ia) to (IIb) are performed at a temperature between 0° C. and 25° C.

6. The method according to any of the preceding claims, wherein before step (Ia) said suspension is obtained using ammonium acetate for precipitating the mRNA molecules, and wherein the first solution contains ammonium acetate.

7. The method according to any of the preceding claims, wherein the second solution is water and/or wherein the fourth solution is water or comprises sodium chloride and/or citrate.

8. The method according to any of the preceding claims, wherein the mRNA molecules are comprised in a retentate after tangential flow filtration.

9. The method according to claim 8, wherein the retentate obtained in step (Ia) is used as feed solution for tangential flow filtration in step (Ib), the retentate obtained from step (Ib) as feed solution in step (IIa), and the retentate obtained from step (IIa) as feed solution in step (IIb).

10. The method according to any of the preceding claims, wherein the mRNA molecules comprised in the suspension are obtained by in vitro transcription.

11. The method according to any of the preceding claims, wherein the method further comprises dephosphorylating and/or polyadenylating and/or post-capping the mRNA molecules.

12. The method according to claim 11, wherein the method comprises dephosphorylating the mRNA molecules obtained from step (Ib), followed by performing steps (Ia) to (IIb), followed by polyadenylating the obtained mRNA molecules, followed by performing steps (Ia) to (IIb).

13. The method according to any of the preceding claims, wherein in case of step (Ia) a filter membrane with a molecular weight cut-off between 300 kDa and 0.65 μm is used for tangential flow filtration, and/or wherein in case of steps (Ib) and (IIb) a filter membrane with a molecular weight cut-off of between 1 kDa and 0.65 μm is used, and/or wherein in case of step (IIa) a filter membrane with a molecular weight cut-off of at least 70 kDa is used.

14. The method according to any of the preceding claims, wherein the diafiltration volume of any of the first, second, third, and/or fourth solution is at least 1-fold the volume of the suspension of step (Ia).

15. A method for producing a pharmaceutical composition comprising (a) purifying mRNA molecules according to a method of any one of claims 1 to 14, and (b) formulating the thus obtained mRNA molecules into a pharmaceutical composition.

Description

[0144] FIG. 1: Overview of a continuous tangential flow filtration (TFF) system with circulating mRNA molecules according to the disclosed method. The mRNA molecules are in suspension in step (Ia), dissolved in step (Ib), and in solution in case of steps (IIa) and (IIb), respectively.

[0145] FIG. 2: Overview of an exemplary system for purifying mRNA molecules according to the disclosed method (with exemplarily depicted positions of valves, (auxiliary) pumps, and/or (further) vessels).

[0146] FIG. 3: SDS-PAGE followed by colloidal coomassie staining of the TFF retentate comprising unmodified tdTomato encoding mRNA molecules after purification of the respective in vitro transcription (IVT) mix according to steps (Ia) and (Ib) at RT. Lane 1: enzyme mix in ammoniumacetate (NH.sub.4OAc) as control; lane 2: enzyme mix after priming the TFF system; lane 3: enzyme mix after TFF in nuclease free water; lane 4: IVT mix; lane 5: IVT mix after priming the TFF system; lane 6: IVT mix after TFF in nuclease free water; lane 7: enzyme mix in nuclease free water as control. The enzyme mix consisted of RNase inhibitor, T7 RNA polymerase, inorganic pyrophosphatase, and DNase I.

[0147] FIG. 4: SDS-PAGE followed by colloidal coomassie staining of the TFF retentate comprising unmodified tdTomato encoding mRNA molecules after purification of the respective in vitro transcription (IVT) mix according to steps (Ia) and (b) at RT. Lane 1: enzyme mix in nuclease free water as control; lane 2: IVT mix before TFF; lane 3: IVT mix after priming the TFF system; lane 4: IVT mix after TFF in nuclease free water; lane 5: IVT mix after TFF in nuclease free water (after diafiltration with 2.5 M ammoniumacetate and nuclease free water, the respective mix was concentrated 2.5 fold to efficiently detect protein removal). The enzyme mix consisted of RNase inhibitor, T7 RNA polymerase, inorganic pyrophosphatase, and DNase I.

[0148] FIG. 5: SDS-PAGE followed by colloidal coomassie staining of the TFF retentate comprising unmodified tdTomato encoding mRNA molecules after purification of the respective in vitro transcription (IVT) mix according to steps (Ia) and (b) at 4° C. Lane 1: enzyme mix in nuclease free water as control; lane 2: IVT mix before TFF; lane 3: IVT mix after one wash volume (equal to initial feed volume); lane 4: IVT mix after TFF in nuclease free water; lane 5: IVT mix after TFF in nuclease free water (after diafiltration with 2.5 M ammoniumacetate and nuclease free water, the respective mix was concentrated 2.5 fold to efficiently detect protein removal). The enzyme mix consisted of RNase inhibitor, T7 RNA polymerase, inorganic pyrophosphatase, and DNase I.

[0149] FIG. 6: A) Western Blot of hCFTR protein translated in HEK293 cells after transfection with TFF purified hCFTR mRNA and hCFTR mRNA purified by ammoniumacetate precipitation followed by ethanol 70% wash, formulated using Lipofectamine Messenger Max. As house keeper, Hsp90 protein bands are shown. Untransfected cells were used as negative control. B) Expressed hCFTR was quantified by densitometric analysis using Image Lab Software and normalized to HSP90 protein. Untransfected cells (UT) were used as negative control.

[0150] FIG. 7: Residual amounts of different spiked in mRNA molecules not being hCFTR compared to the amount of the hCFTR target mRNA molecules after varying wash volumes in percentage. Tangential flow filtration was performed at RT.

[0151] FIG. 8: Residual amounts of different spiked in mRNA molecules not being hCFTR compared to the amount of the hCFTR target mRNA molecules after varying wash volumes in percentage. Tangential flow filtration was performed at 4° C.

[0152] FIG. 9: Capillary gel electrophoresis using Fragment Analyzer was performed from tdTomato mRNA spiked with different DNA oligonucleotides (15 nt, 25 nt and 120 nt) containing a 25 nt antisense DNA oligonucleotide before (A) and after (B) TFF purification at 4° C. The lower marker is an internal standard provided in the reagent kit of the Fragment Analyzer to normalize the retention times of the peak in each capillary run (15 nt long RNA oligonucleotide).

[0153] FIG. 10: Representatively, permeate flux was measured at different transmembrane pressure set points.

[0154] FIG. 11: Capillary gel electrophoresis using Fragment Analyzer was performed from hCFTR mRNA spiked with 25 nt and 120 nt long DNA oligonucleotides and 256 nt long GLP-1 mRNA molecules before purification according to Approach 1 (A), after step (Ib) (B), after step (IIa) (C), and after step (IIb) (D) at RT. The lower marker is an internal standard provided in the reagent kit of the Fragment Analyzer to normalize the retention times of the peak in each capillary run (15 nt long RNA oligonucleotide).

[0155] FIG. 12: Capillary gel electrophoresis using Fragment Analyzer was performed from hCFTR mRNA spiked with 25 nt and 120 nt long DNA oligonucleotides and 256 nt long GLP-1 mRNA molecules before (A) and after (B) purification according to Approach 2 at RT. The lower marker is an internal standard provided in the reagent kit of the Fragment Analyzer to normalize the retention times of the peak in each capillary run (15 nt long RNA oligonucleotide).

[0156] Other aspects and advantages of the invention will be described in the following examples, which are given for purposes of illustration and not by way of limitation. Each publication, patent, patent application or other document cited in this application is hereby incorporated by reference in its entirety.

EXAMPLES

[0157] Methods and materials are described herein for use in the present disclosure; other, suitable methods and materials known in the art can also be used. The materials, methods, and examples are illustrative only and not intended to be limiting.

[0158] Abbreviations used herein and their respective descriptions are listed in Table 3.

TABLE-US-00003 TABLE 3 Abbreviation Description ° C. Celsius Degree CFTR Cystic Fibrosis Transmembrane Conductance Regulator EDTA Ethylenediaminetetraacetic acid HPLC High performance liquid chromatography IVT In vitro transcription kDa Kilodalton I // ml Litre // Millilitre M // mM Molar // Millimolar mbar Millibar min Minute mg Milligram MOPS 3-(N-morpholino)propanesulfonic acid mPES Modified Polyethersulfone mRNA Messenger ribonucleic acid (RNA) MWCO Molecular weight cut-off NaOH Sodium hydroxide nt Nucleotide(s) NH.sub.4OAc Ammonium acetate, CH.sub.3COONH.sub.4 RT Room temperature, e.g. between 20° C. to 25° C., e.g. 22° C. SDS-PAGE Sodium dodecyl sulfate polyacrylamide gel electrophoresis TFF Tangential flow filtration TMP Transmembrane pressure x Fold μl Micro litre % Percentage

[0159] Material and Methods

[0160] Materials, Devices, Software, and Test System Used

[0161] Materials are listed in Table 4.

TABLE-US-00004 TABLE 4 Material Supplier Cat# 5M Ammoniumacetate solution pH 7 Sigma Aldrich 09691-1L 50 kDa mPES TFF filter Spectrumlabs C02-E050-05-N 100 kDA mPES TFF filter Spectrumlabs C02-E100-05-N 300 kDa mPES TFF filter Spectrumlabs C02-E300-05-N 500 kDA mPES TFF flter Spectrumlabs C02-E500-05-N HiPerSolv Chromanorm Water for VWR 83650.320 HPLC Aqua ad iniectabilia B. Braun 3703444 EDTA disodium salt solution pH 7 Sigma Aldrich E7889 MOPS BioUltra for molecular biology Sigma Aldrich 6947 Sodiumhydroxide 32% solution Carl Roth T197.1 Standard Sensitivity RNA Analysis Advanced DNF-471 Kit (15 nt) Analytical Lipofectamine® MessengerMax™ Thermo Fisher LMRNA015 Scientific Luminata Classico Millipore WBLUC0500 MEM, GlutaMAX™ Supplement Thermo Fisher 41090028 Scientific Precision Plus Protein Dual Color Bio-Rad 161-0374 Standard PVDF Pre-cut Blotting Membranes, Thermo Fisher LC2002 0.2 μm Scientific Roti-Load Carl Roth K930.1 SuperSignal™ West Femto Thermo Fisher 34095 Scientific TRIS-Acetate 3-8%, 1 mm, 15-well Thermo Fisher EA03755BOX Scientific TRIS-Acetate Running Buffer Thermo Fisher LA0041 Scientific TRYPSIN 0.05% EDTA Thermo Fisher 25300054 Scientific WesternBreeze® Blocker/Diluent Thermo Fisher WB7050 Scientific WesternBreeze® Wash Solution Thermo Fisher WB7003 (16X Scientific mouse anti-humanCFTR mAb R&D Systems MAB25031 mouse anti-HSP90 mAb Origene TA500494 cOmpleete™, EDTA-free Protease Sigma-Aldrich 11873580001 Inhibitor Dithiotreitol (DTT) GE Healthcare 17-1318-02 DNase I Solution (2500 U/mL) Thermo Fisher 90083 Scientific DPBS 1x without Ca and Mg Thermo Fisher 14190-169 Scientific Fetal Bovine Serum, heatinactivated Thermo Fisher 10500064 Scientific donkey anti-mouse IgG-HRP Abcam ab6820 Colloidal Coomassie Roth A152.1 o-Phosphoric acid, 85% Rotipuran, Carl Roth 63661. p.a., ACS, ISO RiboLick Rnase Inhibitor Thermo Fisher Scientific DNAse I (Rnase-free) Thermo Fisher Scientific Inorganic Pyrophosphatase Thermo Fisher Scientific T7 RNA Polymerase Thermo Fisher Scientific Bolt 4 - 12% Bis-Tris Plus 10 wells Invitrogen NW04120BOX Bolt LDS sample loading buffer (4x) Life technologies B0007 Precision Plus Protein Dual Color BioRad 161-0374 Standards Bolt MES SDS running buffer 20x Life technologies B0002 Magnesiumchlorde solution Sigma Aldrich M1028 Spermidine Sigma Aldrich 85578 Dithiotreithol (DTT) solution Sigma Aldrich 43816 GTP Jena Bioscience NU-1012 ATP Jena Bioscience NU-1010 CTP Jena Bioscience NU-1011 UTP Jena Bioscience NU-1013 ARCA cap analogue Jena Bioscience NU-855

[0162] Devices are listed in Table 5.

TABLE-US-00005 TABLE 5 Device Supplier KR2i TFF System Spectrum Laboratories, Inc. Fragment Analyzer Advanced Analytical Chemidoc XRS BioRad Laboratories Novex Bolt Mini Gel Tank Life technologies

[0163] Software is listed in Table 6.

TABLE-US-00006 TABLE 6 Software Provider ProSize 3.0 Advanced Analytical Excel Plug In Spectrum Laboratories, Inc. MS Excel Microsoft Image Lab BioRad Laboratories Open Lab Chem Station Agilent

[0164] The test system is listed in Table 7.

TABLE-US-00007 TABLE 7 Test System Species Strain HEK 293 human N.A.

[0165] Purification of an IVT Mix Via TFF—Steps (Ia) and (Ib)

[0166] In the following, an in vitro transcription (IVT) mix was used comprising unmodified, in vitro transcribed unmodified tdTomato target mRNA molecules with 1644 nucleotides (nt) in length. Furthermore, an enzyme mix was used comprising T7 RNA Polymerase, inorganic pyrophosphatase, RNAse inhibitor, and DNase I.

[0167] RNA Precipitation

[0168] The IVT mix and the enzyme mix were each precipitated with an equal amount of ice cold 5 M NH.sub.4OAc with a pH of 7 to a final concentration of 2.5 M NH.sub.4OAc and incubated on ice for at least 30 minutes. Prior to TFF, the respective mix was diluted 1:1 with 2.5 M NH.sub.4OAc pH 7 to a final concentration of about 0.5 mg/ml mRNA in the mix. The IVT or enzyme mix was connected to the TFF system, which was initially primed for 10 minutes by circulating the IVT mix at 50 ml/min with the permeate clamp closed.

[0169] Step (Ia): Removal of Proteins, Nucleotides, and Salts

[0170] Diafiltration of the respective mix was performed at 50 ml/min at a constant TMP of about 200-300 mbar using a 500 kDa MWCO mPES filter column. The respective mix was diafiltered with 10 wash volumes of 2.5 M NH.sub.4OAc pH 7. This step efficiently removed enzymes, nucleotides as well as buffer components from the in vitro transcription. Step (Ia) can be used to remove any other proteins and/or enzymes from any other intermediate production steps (e.g. dephosphorylation, post capping, polyadenylation, . . . ). For some enzymes (e.g. poly(A) polymerase, which binds with high affinity to mRNA molecules of interest) a detergent (e.g. SDS, LDS, . . . ) may be added to the ammoniumacetate buffer pH 7. In case poly(A) polymerase is used for polyadenylation, the subsequent step (Ia) is preferably performed at a temperature between 20° C. and 30° C., preferably between 23° C. and 27° C., more preferably at 25° C. Such a temperature is advantageous for efficient poly(A) polymerase removal.

[0171] Step (Ib): Removal of NH.sub.4OAc from Step (Ia)

[0172] In order remove NH.sub.4OAc and resolve the mRNA molecules, a 100 kDA MWCO mPES filter column was used for diafiltration with 10 wash volumes using nuclease free water at a flow rate of 50 ml/min and a TMP of about 200-300 mbar. Alternatively, a 50 kDa MWCO mPES filter column may be used depending on the size of the mRNA molecule.

[0173] For investigating the purification efficiency of steps (Ia) and (Ib), the TFF retentate comprising the mRNA molecules can be sampled before and after TFF (i.e. step (Ib)) and analyzed for example using UV-measurement, fragment analyzer, sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) followed by e.g. colloidal coomassie staining.

[0174] As shown in FIGS. 3 to 5, enzymes can be efficiently removed from mRNA molecules by applying steps (Ia) and (Ib) as described above. Thereby mRNA molecules were retained by the TFF column, whereas the enzymes and/or proteins were efficiently removed. Smear peak analysis revealed that the smear pre-peak was 4.7% before TFF purification, and had a deviation of +1.3% for TFF purification at 4° C. and +1.8% after TFF purification performed at RT. Therefore no degradation of mRNA could be observed.

[0175] Purification of Precipitated and Dissolved mRNA Via TFF—Development of Steps (IIa) and (IIb)

[0176] In the following, the precipitated and dissolved mRNA was used for further purification.

[0177] Testing Different Filtration Membranes and the Effect of EDTA on mRNA Transmission

[0178] Different filter membranes having varying pore sizes were tested with regard to the transmission of mRNA molecules using either nuclease free water alone or nuclease free water with additional 10 mM ethylenediaminetetraacetic acid (EDTA). Four mPES columns were tested having a molecular weight cut-off (MWCO) of 500 kDa, 300 kDa, 100 kDa and 50 kDa, respectively. The transmembrane pressure (TMP) was adjusted using the retentate clamp on the TFF system and kept at 40 mbar for the columns with a MWCO of 500 kDa and 50 kDa, respectively, and at 100 mbar for the columns with a MWCO of 100 kDa and 300 kDa, respectively. The main pump was set to a flow rate of 15 ml/min and the TFF was performed at RT. The initial mRNA molecule concentration in the feed was 0.1 mg/ml and the mRNA molecule concentrations in retentate and permeate were measured using UV-measurements. Each column was tested with 10× wash volumes of the original sample volume. Without EDTA, mRNA molecules were retained by the tested column having a MWCO of 500 kDa. It was assumed that smaller pore sizes (i.e. 300 kDa, 100 kDa and 50 kDa) will not allow passage of the mRNA molecules either and were therefore not further tested without EDTA present.

[0179] With EDTA, the mRNA molecules surprisingly moved through the pores of the 500 kDa MWCO filter membrane and were also partly lost using the 300 kDa MWCO filter membrane. However, the mRNA molecules could be successfully retained using the 100 and the 50 kDa MWCO filter membranes in the presence of 10 mM EDTA. Hence, the 100 kDa MWCO column was chosen for the filtration of mRNA molecules in the experiments shown below.

[0180] It has to be noted that sporadic precipitation of the mRNA molecules was observed in some cases after filtration with 10 mM EDTA. However, this precipitation could be efficiently prevented by using a buffer comprising 40 mM MOPS and 10 mM EDTA.

[0181] Preparation of a 40 mM MOPS and 10 mM EDTA Diafiltration Buffer

[0182] An 1 M 3-(N-morpholino)propanesulfonic acid (MOPS) buffer was prepared, wherein the MOPS was dissolved in nuclease free water. The pH of the 1 M MOPS buffer was then adjusted to a pH of 7 using 32% NaOH. The obtained 1 M MOPS buffer having a pH of 7 was mixed with 0.5 M EDTA and nuclease free water to finally obtain a diafiltration buffer comprising 40 mM MOPS and 10 mM EDTA.

[0183] Determination of Wash Buffer Volumes to Remove Nucleic Acid Oligonucleotides

[0184] The efficiency of the TFF process, i.e. the minimal number of wash cycles needed for complete removal of impurities such as nucleic acid oligonucleotides representing abortive transcripts and/or hydrolysis products, strongly depends on the rejection of the molecule by the membrane filter and its pore size. Hence, the respective effect of wash buffer volumes was investigated to determine the number of wash cycles required to remove spiked nucleic acid oligonucleotides from precipitated and dissolved mRNA comprising mRNA molecules of interest.

[0185] For this experiment, the TFF flow rate was set to 15 ml/min and the TMP was kept constant at about 40 mbar using a 100 kDa MWCO mPES filter column. In a total feed volume of 5 ml, 500 μL of mRNA molecules were spiked with 10 nt long DNA oligonucleotide representing 2.5% of the target mRNA molecule (m/m), 50 nt long DNA oligonucleotide representing 2.5% of the target mRNA molecule (m/m), and 120 nt long DNA oligonucleotides representing 5% of the target mRNA molecule (m/m). Samples for reversed-phase HPLC analysis were drawn after 5×, 10×, 15×, and 20× wash volumes of the original feed volume with the diafiltration buffer comprising 40 mM MOPS and 10 mM EDTA. Before sample drawing the permeate clamp was closed and the retentate valve opened for circulation of the retentate at 50 ml/min for 5 minutes. For each wash cycle 100 μL sample were drawn using a sterile syringe and the samples were kept on ice until analysis. TFF was performed at room temperature. Samples were investigated using reversed-phase HPLC analysis with the area detected at 260 nm being representative for the relative amounts of DNA oligonucleotide in the sample. The corresponding areas before TFF (control) were set to 100% oligonucleotide.

[0186] In all cases, no nucleic acid oligonucleotides could be detected after applying 5×, 10×, 15×, and 20× wash volumes to the feed volume, respectively. Hence, results showed that already after 5× wash volumes with washing buffer no more nucleic acid oligonucleotides could be detected in the retentate. A minimum of 10× wash volume with washing buffer was determined to be sufficient for the removal of the nucleic acid oligonucleotides from mRNA molecules.

[0187] Determination of Wash Volumes with Nuclease Free Water for Removal of EDTA

[0188] Since removal of the DNA oligonucleotides by TFF is performed using washing buffer, i.e. diafiltration buffer comprising 40 mM MOPS and 10 mM EDTA, a subsequent filtration step is advantageous for exchanging the washing buffer with nuclease free water. Hence, an experiment was performed to determine the amount of wash cycles required for the removal of the washing buffer from the mRNA molecules. From the previous experiment the feed solution comprising the mRNA molecules and having a total volume of about 5 ml that has previously been filtered with 20× wash volumes of diafiltration buffer was diafiltered with nuclease free water at a flow rate of 15 ml/min and a constant TMP of 40 mbar using a 100 kDA MWCO mPES filter column. Samples for the reversed-phase HPLC analysis were taken after 5×, 10×, 15× and 20× wash volumes of the original feed volume with nuclease free water. In order to quantify the amount of residual EDTA after each washing cycle with nuclease free water the diafiltration buffer comprising 40 mM MOPS and 10 mM EDTA was titrated and a calibration standard curve recorded (cf. Table 6; MOPS-EDTA calibration standard curve: log(y)=0.6467 log(x)+1.9128; r=0.99462; r.sup.2=0.99877; curve model: log/log). Since MOPS buffer alone shows no absorption signal at 260 nm, the concentrations stated in Table 8 and Table 9 correspond to the concentration of EDTA. The experiment was performed at 22° C. RT. In Table 9 data are shown after EDTA removal after each wash cycle.

[0189] As shown in Table 9, the decreasing amounts of EDTA according to the peak area detected by reversed-phase HPLC at 260 nm after increasing wash volumes revealed that a minimum of 10× wash volumes with nuclease free water were necessary for partly removal of residual EDTA.

TABLE-US-00008 TABLE 8 EDTA cSOLL Peak Area 260 EDTA calculated Accuracy Absolute [mM] nm [mM] accuracy 1.00 74.864 0.92 92% 8% 0.80 68.269 0.79 99% 1% 0.50 56.447 0.57 115%  15%  0.10 19.282 0.10 96% 4% 0.05 13.057 0.05 100%  0%

TABLE-US-00009 TABLE 9 AREA Residual MOPS- EDTA ID EDTA [mM] 20x washing buffer 0x nuclease free water 205.65 Above ULOQ 20x washing buffer 5x nuclease free water 60.7 0.65 20x washing buffer 10x nuclease free water 24.3 0.14 20x washing buffer 15x nucleasre free water 7.8 Below LOQ 20x washing buffer 20x nuclease free water 13.8 0.05

[0190] In the previous experiments that were carried out at RT, the minimum number of wash cycles using diafiltration buffer required for removing DNA oligonucleotides in step (IIa) as well as the minimum number of wash cycles using nuclease free water required for removing residual EDTA in step (IIb) were determined. For testing the determined parameters, 500 μg of mRNA were spiked with 10 nt long nucleic acid oligonucleotides representing 2.5% of the target mRNA molecule volume, 50 nt long DNA oligonucleotide representing 2.5% of the target mRNA molecule volume, and 120 nt long DNA oligonucleotide representing 5% of the target mRNA molecule volume in a total feed volume of 5 ml (referred to as “Before TFF” in Table 8 and Table 9).

[0191] For experiments described below the following set up was used if not stated otherwise: The feed solution comprising 0.1 mg/ml mRNA molecules was diafiltered using a 100 kDa MWCO mPES filter column at a flow rate of 15 ml/min and a TMP of about 40 mbar. The applied wash volumes were in both cases 10×, i.e. 10× wash volume with diafiltration buffer comprising 40 mM MOPS and 10 mM EDTA followed by 10× wash volume with nuclease free water (referred to as ‘After TFF’ in Table 8 and Table 9). Since mRNA molecules are prone to degradation by hydrolysis especially at elevated temperatures, the experiments were carried out at 4° C., i.e. using ice water, to keep mRNA molecule hydrolysis at a minimum.

[0192] This set up resulted in the efficient removal of DNA oligonucleotides of all three lengths tested (Table 10) as well as of residual diafiltration buffer (Table 11) as determines using reversed-phase HPLC. Furthermore, smear peak analysis were performed after capillary gel electrophoresis using Fragment Analyzer before and after purification (Table 12). Smear peak analyses indicated that the mRNA integrity did not vary due to the TFF purification and thus, was not affected by the method.

TABLE-US-00010 TABLE 10 Peak area of DNA oligonucleotides Length 10 nt Length 50 nt Length 120 nt Before TFF 100% 100% 100% After TFF  0.0%  0.0%  0.0%

TABLE-US-00011 TABLE 11 Peak area MOPS-EDTA Residual MOPS-EDTA [mM] Before TFF 193.0 Above upper limit of quantification After TFF 27.4 0.17

TABLE-US-00012 TABLE 12 Smear pre-peak [%] Reference (without MOPS-EDTA) 20.0 Before TFF with MOPS-EDTA, pH 7; Duplicate 1 18.2 Before TFF with MOPS-EDTA, pH 7; Duplicate 2 15.4 10x wash buffer 10x nuclease free water; Duplicate 1 17.4 10x wash buffer 10x nuclease free water; Duplicate 2 16.5

[0193] Determination of the Translation Efficiency of Purified mRNA Molecules

[0194] Furthermore, it was of great interest to determine the translation efficiency of the obtained purified mRNA molecules. Hence, 1.4×10.sup.6 HEK293 cells were seeded in 6-well plates and transfected with 3.75 μg mRNA molecules purified as described above and according to a standard procedure, respectively. Transfection was performed using MessengerMax (1:15). After 24 h, the cells were lysed and investigated using SDS-page and Western Blot, respectively, using 50 μg cell lysate each.

[0195] As it can be seen in FIG. 6, a comparable amount of mRNA molecules was detected using Western Blot (A) and quantified (B).

[0196] Determination of a Threshold for the Removal of mRNA with Different Lengths

[0197] A total of 300 μg of mRNA target hCFTR mRNA was spiked with different mRNA molecules varying in length using a total volume of 5 ml. In particular, 6 different types of mRNA molecules were used for spiking with each type representing 16.7% of the target mRNA molecule volume. Three types of mRNA molecules exhibited a cap and a poly(A) tail and had a total length of 3,632 nt, 1,864 nt, and 1,111 nt, respectively. Two types had neither a cap nor a poly(A) tail and exhibited a total length of 494 nt and 256 nt, respectively. The last type referred to a 120 nt DNA long oligonucleotide that served as a positive control.

[0198] The spiked mRNA molecule comprising solution was filtered as described above at RT (FIG. 7) and 4° C. (FIG. 8), respectively. Samples were drawn before TFF (directly after sample preparation), after 5×, 10×, 15× and 20× wash volumes of the original sample volume using the diafiltration buffer described above, and after additional washing using 10× wash volumes of nuclease free water. For each wash cycle, 100 μl sample was drawn using a sterile syringe and kept on ice until analysis. Transmission of the molecules through the filter pores was assessed via capillary gel electrophoresis using a Fragment Analyzer.

[0199] As shown in FIG. 7 and FIG. 8, a cut-off for mRNA molecules having at least a length of 951 nt could be observed using a 100 kDa mPES column. For mRNA molecules having a length of less than 951 nt, columns having a lower MWCO might be advantageous such as for example 50 kDa or 70 kDa mPES columns.

[0200] Increasing the mRNA Molecule Concentration from 0.1 mg/ml to 1 mg/ml

[0201] In the following final set of experiments (FIG. 9), the filtering procedure described in the following was applied to a solution comprising unmodified, 1644 nt long unmodified tdTomato mRNA as target mRNA molecule. In particular, a total of 5 ml feed solution was analyzed comprising 1 mg/ml mRNA molecules, corresponding to a total of 5 mg of mRNA molecules, as well as a 15 nt DNA oligonucleotide and an antisense 25 nt long oligonucleotide—which complementary binds to the mRNA molecule of interest—representing 2.5% of the target mRNA molecule (m/m), and 120 nt long oligonucleotides representing 5% of the target mRNA molecule (m/m). This spiked feed solution was diafiltered at 4° C. using a 100 kDa MWCO mPES filter column at a flow rate of 15 ml/min and a TMP of about 40 mbar. The applied wash volumes were i) 10× wash volume with nuclease free water, followed by ii) 10× wash volume with diafiltration buffer comprising 40 mM MOPS and 10 mM EDTA, followed by iii) 10× wash volume with nuclease free water.

[0202] Smear values were investigated as an indicator for mRNA integrity. In particular, the following smear values were obtained: a smear pre-peak of 6.4% before TFF as well as a smear pre-peak of 5.9% after TFF. The smear pre-peak areas reflect the proportion of mRNA related hydrolysed products. Thus, no hydrolysis products could be measured by Fragment Analyzer analysis and it could be shown that TFF purification does not affect mRNA integrity.

[0203] Exemplary Outline for the Determination of Optimal TFF Parameters Applying Routine Measures

[0204] The following is a brief description how the skilled person can optimize the TFF method described above by applying routine measures.

[0205] Step (Ia)

[0206] The TFF flow rate can be adjusted so as to achieve a sufficient shear rate in order to re-circulate the precipitated mRNA molecules through the TFF system and to sweep away the precipitated mRNA molecules from the filter surface. By opening the permeate clamp a sufficient permeate flux can be achieved. Enzyme removal can be determined e.g. by SDS-Page analysis and required wash volumes can be adjusted by routine measures (preferably 1-20 wash volumes; more preferably 5-15; most preferably 9-11). One wash volume is defined as an equal volume of diafiltration medium, e.g. ammonium acetate buffer with a pH of 7, of the initial feed volume.

[0207] Step (Ib)

[0208] Same parameters as described in step (Ia) can be applied. Hereby buffer can, e.g., be exchanged by nuclease free water in order to resolve mRNA molecules and to remove buffer (e.g. ammonium acetate). Buffer removal can be determined by e.g. conductivity, UV-measurement and required wash volumes can be adjusted by routine measures (preferably 1-20 wash volumes; more preferably 5-15; most preferably 9-11). One wash volume is defined as an equal volume of diafiltration medium, e.g. nuclease free water, of the initial feed volume.

[0209] Step (IIa)

[0210] The TFF flow rate can be adjusted so as to achieve a sufficient shear rate in order to re-circulate the mRNA molecules through the TFF system and to sweep away the mRNA molecules from the filter surface. By opening the permeate clamp a sufficient permeate flux can be achieved. Loss of mRNA of interest can be monitored e.g. by UV-measurement (e.g. online/offline measurement) in the permeate line. If loss of mRNA of interest is observed flow rate and/or TMP can be adjusted by routine measures until no further loss of mRNA of interest is observed. Required wash volumes can be adjusted by routine measures dependent from the amount of impurities (e.g. abortive transcripts, and/or hydrolysis products); (preferably 1-20 wash volumes; more preferably 5-15; most preferably 9-11). One wash volume is defined as an equal volume of diafiltration medium, e.g. MOPS-EDTA at a pH of 7, of the initial feed volume.

[0211] Step (IIb)

[0212] Same parameters as described in step (IIa) can be applied. Hereby buffer can, e.g., be exchanged by nuclease free water in order to remove buffer, e.g. MOPS-EDTA. Buffer removal can be determined by e.g. conductivity, UV-measurement and required wash volumes can be adjusted by routine measures (preferably 1-20 wash volumes; more preferably 5-15; most preferably 9-11). One wash volume is defined as an equal volume of diafiltration medium, e.g. nuclease free water, of the initial feed volume.

[0213] TMP excursion experiments can be helpful to determine optimal conditions for mRNA molecule purification. Exemplarily, a total of 5 ml feed solution comprising a target mRNA molecule concentration of 1.0 mg/ml in the diafiltration buffer used in step (IIa) as described above was diafiltrated using a 100 kDa MWCO mPES filter column and a flow rate of 15 ml/min. The permeate flux was measured in ml/min at 6 different TMP values ranging from about 50 mbar to about 150 mbar. As it can be seen in FIG. 10 for tdTomato mRNA, the optimal TMP was determined to be about 50 mbar at a flow rate of 15 ml/min but this might dependent on the experimental setting.

Comparative Example

[0214] In the following, an example is described that was performed in order to compare results obtained by the method disclosed herein (approach 1) and by the method disclosed in WO 2015/164773 A1 (approach 2).

[0215] Approach 1

[0216] In vitro transcribed, unmodified hCFTR mRNA (cf. U.S. Pat. No. 9,713,626 B2; 5 mg pellet) was spiked with DNA oligonucleotides of 25 nucleotides (nt) and 120 nt length as well as with 256 nt long glucagon-like peptide 1 molecules (GLP-1 mRNA molecules (SEQ ID NO: 1). The 25 nt long DNA oligonucleotides were spiked into the mRNA at a ratio of 2.5% referring to the total amount of mRNA in μg. The 120 nt long DNA oligonucleotides were spiked into the mRNA at a ratio of 5.0% referring to the total amount of mRNA in μg. And the 256 nt long GLP-1 mRNA molecules were spiked into the mRNA at a ratio of 16% referring to total mRNA amount in μg.

[0217] The spiked mRNA molecules were precipitated using 2.5 M NH.sub.4OAc (pH 7).

[0218] For removing proteins, salts and abortive transcripts step (Ia) was performed with 10× wash volumes 2.5 M NH.sub.4OAc using a 500 kDa MWCO mPES with 50 mL/min flow rate, and ˜200-300 mbar TMP (mRNA concentration: 0.5 mg/mL).

[0219] For the removal of NH.sub.4OAc and for resolving the mRNA molecules, step (Ib) was performed with 10× wash volumes nuclease-free water using a 50 kDa MWCO mPES with 50 mL/min flow rate, and ˜200-300 mbar TMP (mRNA concentration: 0.5 mg/mL).

[0220] For removing divalent cations, abortive transcripts, and/or hydrolysis products, step (IIa) was performed with 10× wash volumes of 40 mM MOPS and 10 mM EDTA using a 100 kDa MWCO mPES with 15 mL/min flow rate and ˜20 mbar TMP (mRNA concentration: 1.0 mg/mL).

[0221] For the removal of the MOPS-EDTA diafiltration buffer, step (IIb) was performed using 10× wash volumes nuclease-free water using a 100 kDa MWCO mPES with 15 mL/min flow rate and ˜20 mbar TMP (mRNA concentration: 1.0 mg/mL).

[0222] Obtained mRNA molecules were investigated using a Fragment Analyzer/reversed-phase HPLC in view of the removal of spiked oligonucleotides, and a Nanodrop for the determination of mRNA molecule recovery.

[0223] Approach 2 (as Described in WO 2015/164773 A1)

[0224] In vitro transcribed, unmodified hCFTR mRNA (cf. U.S. Pat. No. 9,713,626 B2; 10.26 mg pellet) was spiked with DNA oligonucleotides of 25 nucleotides (nt) and 120 nt length as well as with 256 nt long GLP-1 mRNA molecules (SEQ ID NO: 1). The 25 nt long DNA oligonucleotides were spiked into the mRNA at a ratio of 2.5% referring to the total amount of mRNA in μg. The 120 nt long DNA oligonucleotides were spiked into the mRNA at a ratio of 5.0% referring to the total amount of mRNA in μg. And the 256 nt long GLP-1 mRNA molecules were spiked into the mRNA at a ratio of 16% referring to total mRNA amount in μg.

[0225] The spiked mRNA molecules were precipitated using i) guanidinium thiocyanate; sodium lauryl sarcosyl, and sodium citrate to final concentrations of 2.09 M guanidinium thiocyanate; 0.26% sodium lauryl sarcosyl, and 13.0 mM sodium citrate and ii) absolute ethanol to final concentration of ˜38% EtOH, and incubated for 5 minutes at RT.

[0226] Loading was performed using ˜22 mL precipitated mRNA with a flow rate of ˜6 mL/min using a 500 kDa MWCO mPES (mRNA concentration: 0.52 mg/mL).

[0227] Washing was performed by repeating the flowing two steps >5 times with ˜6 mL/min flow rate using a 500 kDa MWCO mPES (mRNA concentration: 0.52 mg/mL): step a) washing using 5 mL of 2.09 M guanidinium thiocyanate; 0.26% sodium lauryl sarcosyl; 13.0 mM sodium citrate; ˜38% EtOH, and step b) washing using 5 mL 80% ethanol.

[0228] Elution (step C) was performed by treating the obtained solid mRNA with 5 mL nuclease-free water and re-circulating for 5-10 minutes (permeate closed) to ensure dissolution. This procedure was repeated until no more mRNA molecules were recovered using ˜6 mL/min flow rate and a 500 kDa MWCO mPES (mRNA concentration: 0.52 mg/mL).

[0229] Dialysis (step D) was performed using ˜5 wash volumes with 1 mM sodium citrate (pH 6.4), ˜6 mL/min flow rate and a 100 kDa MWCO mPES (mRNA concentration: 0.52 mg/mL).

[0230] Obtained mRNA molecules were investigated using a Fragment Analyzer/reversed-phase HPLC in view of the removal of spiked oligonucleotides, and a Nanodrop for the determination of mRNA molecule recovery.

[0231] Of note, the following change was done in case of Approach 2 described above compared to the method described in WO 2015/164773 A1: After linear downscaling for keeping the shear rate on the filter column constant, the initially planned flow rate was 6 mL/min. However, this flow rate resulted in a very low TMP, i.e. the effective permeate flow rate was 0 mL/min and diafiltration was not possible. Therefore, the flow rate was step wise increased until a sufficient TMP was observed leading to a considerable permeate flux. Thus, the final flow rate used for steps a) to d) ranged from 20-24 mL/min (permeate flow rate between 1.1 and 1.25 mL/min).

[0232] The results shown in Table 13 were obtained by the Fragment Analyzer analysis that was performed for non purified and purified hCFTR mRNA samples. The respective electropherograms are shown in FIG. 11 for Approach 1 and in FIG. 12 for Approach 2, respectively.

TABLE-US-00013 TABLE 13 Residual amount Residual amount Residual amount of 25 nt DNA of 120 nt DNA of 256 nt Process step oligo [%] oligo [%] mRNA [%] Before 100.0% 100.0% 100.0% purification Ia/Ib 0.0% 1.1% 81.6% IIa 0.0% 0.0% 2.3% IIb 0.0% 0.0% 2.4% Step C 0.0% 25.6% 86.6% Step D 0.0% 5.7% 91.4%

[0233] In case of Approach 1, removal of spiked DNA oligonucleotides (25 nt and 120 nt) was mainly achieved by step (Ia)/(Ib). Step (IIa) was required to remove the 256 nt mRNA representing longer hydrolytic products. Step (IIb) did not further remove the 256 nt mRNA, showing that for the removal of hydrolytic products and abortive sequences step (IIa) using a potent chelating agent such as EDTA is essential.

[0234] In case of Approach 2, removal of spiked DNA oligonucleotides (25 nt and 120 nt) was only partially achieved by step A to C. Step D did not completely remove the 120 nt DNA oligonucleotides. The 256 nt mRNA representing longer hydrolytic products almost completely remained in the sample after TFF purification. Thus, performing a dialysis using 1 mM sodium citrate (step D) had no strong effect on further elimination of the 120 nt long DNA oligonucleotides representing abortive mRNA molecules and especially of the 256 nt long mRNA representing longer hydrolytic products.

[0235] In the following, the 256 nt long mRNA sequence used herein is described in further detail.

TABLE-US-00014 Codon optimized GLP-1 sequence SEQ ID No: 1 .sub.GGGAGACU AUGAAGAUCAUCCUGUGGCUGUGCGUGUUCGGCCUG UUCCUGGCCACCCUGUUCCCCAUCAGCUGGCAGAUGCCUGUGGAAAGC GGCCUGAGCAGCGAGGAUAGCGCCAGCAGCGAGAGCUUCGCCAAGCGG AUCAAGAGACACGGCGAGGGCACCUUCACCAGCGACGUGUCCAGCUAC CUGGAAGGCCAGGCCGCCAAAGAGUUUAUCGCCUGGCUCGUGAAGGGC AGAGGCUGA.sup.GAAUU

[0236] Part of the T7 promoter, C: Ethris minimal 5′UTR, followed by an additional U nucleotide, TISU 5′UTR, start codon, codon optimized GLP-1 mRNA sequence, stop codon, Part of EcoRI Restriction Site

[0237] Non-polyadenylated mRNA was used for spiking experiments