Device, use of said device, and method for separating substances with an improved utilization of the capacity of chromatographic media

Abstract

The present invention relates to a device for separating and/or isolating substances in or from a mixture with improved utilization of the capacity of chromatographic media, the device comprising a first chromatography system, a second chromatography matrix downstream of the first chromatography system, and a sensor for detecting the substances present in the fluid. Furthermore, the present invention relates to both the use of said device and a method for separating and/or isolating substances in or from a mixture in a fluid.

Claims

1. A device for separating and/or isolating substances in or from a mixture in a fluid, consisting of a first chromatography system which comprises one or more chromatography matrices, a single sensor downstream of the first chromatography system for detecting the substances present in the fluid, and a second chromatography matrix downstream of the single sensor, the second chromatography matrix being connected downstream of the first chromatography system such that the fluid leaving the first chromatography system is guidable through the second chromatography matrix wherein the single sensor for detecting the substances present in the fluid is interposed between the first chromatography system and the second chromatography matrix and determines on a basis of a change in a slope of a breakthrough signal an overloading of the first chromatography system, and wherein said single sensor for detecting the substances present in the fluid is the only sensor comprised in the device, and wherein the combination of the first chromatography system, the second chromatography matrix and the single sensor is present in a unified body.

2. The device as claimed in claim 1, wherein the second chromatography matrix is a convective chromatography matrix such as a membrane adsorber or a monolith.

3. The device as claimed in claim 1, wherein a measurement principle of the sensor is based on UV or IR absorption spectroscopy, terahertz spectroscopy, manometry, conductometry, potentiometry, refractometry, radiometry, fluorescence spectroscopy especially of surface fluorescence quenching, ATR (attenuated total reflection) infrared spectroscopy especially of surface plasmon resonance spectroscopy (SPRS), potentiometry, polarimetry, impendance spectroscopy, NMR spectroscopy, Raman spectroscopy, turbidimetry, nephelometry and on ultrasound transit time, or comprises a combination of multiple techniques as detection principle for the sensor.

4. The device as claimed in claim 1, wherein the first chromatography system has a ligand and the second chromatography matrix has a ligand, the ligand of the first chromatography system and the ligand of the second chromatography matrix each interacting with at least one substance in the mixture via an identical chemical and/or physical interaction, selected from ion exchangers, salt-tolerant ligands, chelating agents, thiophilic or hydrophobic ligands of various chain lengths and configurations, reversed-phase ligands, reactive dyes and other dyes, of the inorganic molecules and ions, organic and inorganic compounds thereof, affinity ligands, high-molecular-weight ligands, enzymes and subunits and also parts thereof, structural proteins, receptors and effectors and also parts thereof, viruses and parts thereof, xenobiotics, pharmaceuticals and active pharmaceutical ingredients, alkaloids, antibiotics, biomimetics and of the catalysts.

5. The device as claimed in claim 1, wherein the first chromatography system and the second chromatography matrix have at least one identical ligand which interacts with at least one substance in the mixture via at least one chemical and/or physical interaction, selected from ion exchangers, salt tolerant ligands, chelating agents, thiophilic or hydrophobic ligands of various chain lengths and configurations, reversed-phase ligands, reactive dyes and other dyes, of the inorganic molecules and ions, organic and inorganic compounds thereof, affinity ligands, high-molecular-weight ligands, enzymes and subunits and also parts thereof, structural proteins, receptors and effectors and also parts thereof, viruses and parts thereof, xenobiotics, pharmaceuticals and active pharmaceutical ingredients, alkaloids, antibodies, biometrics and of the cataylysts.

6. The device as claimed in claim 1, wherein the first chromatography system and the second chromatography matrix each have at least one different ligand which interacts with at least one substance in the mixture via at least one chemical and/or physical interaction, selected from ion exchangers, salt-tolerant ligands, chelating agents, thiophilic or hydrophobic ligands of various chain lengths and configurations, reversed-phase ligands, reactive dyes and other dyes, of the inorganic molecules and ions, organic and inorganic compounds thereof, affinity ligands, high-molecular-weight ligands, enzymes and subunits and also parts thereof, structural proteins, receptors and effectors and also parts thereof, viruses and parts thereof, xenobiotics, pharmaceuticals and active pharmaceutical ingredients, alkaloids, antibiotics, biomimetics and of the catalysts.

7. The device as claimed in claim 1, wherein the fluid leaving the first chromatography system is completely guidable through the second chromatography matrix.

8. The device as claimed in claim 1, wherein a capacity of the second chromatography matrix is smaller than or identical to a capacity of the first chromatography system.

9. The device as claimed in claim 1, wherein the sensor is nondestructive with regard to the target molecule.

10. The device as claimed in claim 1, wherein the sensor is a UV, IR, pH, pressure or conductivity sensor.

11. A method for separating and/or isolating substances in or from a mixture in a fluid, comprising the steps of guiding the fluid through the device as claimed in claim 1 and collecting the fluid which has been guided through.

12. The method as claimed in claim 11, wherein the mixture is a solution of a synthetically or biologically produced product, or a natural substance.

13. The method as claimed in claim 11, wherein the mixture is a protein-containing solution.

14. The method as claimed in claim 11, wherein the mixture is an antibody-containing solution.

15. The use of the device as claimed in claim 1 for separating and/or isolating a target substance in or from a mixture in a fluid.

16. The use of the device as claimed in claim 1, wherein the use of the device allows an improvement in the utilization of the capacity of the first chromatography system.

17. The use of the device as claimed in claim 1, wherein the use of the device allows an over 60% utilization of the capacity of the first chromatography system.

Description

DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 shows the schematic structure of the present invention comprising main medium (HM; first chromatography system), sensor and second chromatography matrix (MA).

(2) FIG. 2 shows the loading of the HiTrap Protein A unit (first chromatography system) with 140 ml of IgG (2.5 mg/ml) and BSA (1 mg/ml) in buffer A and of a second chromatography matrix with variable bed height (0-4 mm) of Sartobind Protein A (MA). The breakthrough curves shown correspond to the measurement of the absorption at the output of the column (Column) or after the membrane adsorber (Column+MA 0-4 mm).

(3) FIG. 3 shows the change in the first derivative (UV/Volume) or in the UV signal (UV=UV.sub.x+DVUV.sub.x) with differing bed height of the membrane adsorber, which is connected downstream of the first chromatography system. The signals were determined from the UV absorption values of the sensor, which is situated immediately after the first chromatography system and before the membrane adsorber.

(4) FIG. 4 shows the relationship of the loss of product and of the increase in back pressure depending on the bed height of the downstream membrane adsorber.

(5) FIG. 5 shows the loading and elution of the HiTrap Protein A column (first chromatography system) having downstream membrane adsorbers (second chromatography matrix) having a 5 cm.sup.2 (MA1) and 20 cm.sup.2 (MA2) flow area. What is shown are the chromatograms which were detected after the column or after the membrane adsorber (Column+MA1 or MA2) with the aid of a sensor.

(6) FIG. 6 shows the UV absorption (280 nm) during the loading of the structure described in table 5 with lysozyme (5 mg/ml) and bovine serum albumin (BSA, 2 mg/ml). Owing to the additional membrane adsorber, the breakthrough of the target molecule occurs only at higher loading volume (dashed line, detection after the membrane adsorber). The lower line corresponds to the UV signal, which was measured before the membrane adsorber.

DESCRIPTION OF THE INVENTION

Examples

(7) Materials and Methods

(8) All the experiments were carried out on the FPLC system ktaPrime plus from GE Healthcare. UV absorption was measured at 280 nm and chromatograms were evaluated with the aid of the software PrimeView 5.31 (GE Healthcare). Dead volumes were determined by the injection of a nonbinding, UV-active tracer molecule (20% acetone), which was added to the particular equilibration buffer. Samples up to a volume of 150 ml were injected with the aid of a Superloop (GE Healthcare). In the case of loading volumes above a volume of 150 ml, the samples were applied via one of the buffer valves of the ktaPrime. The flow rate for equilibration, loading, washing and elution was 5 ml/min for the experiments with the HiTrap Protein A HP column, and 10 ml/min for experiments with Sartobind Q Nano capsules. All the buffer substances were from Carl Roth (Karlsruhe) or Merck (Darmstadt). Adjustment of the pH of the solutions was done at room temperature.

(9) Protein A Chromatography

(10) For the experiments with protein A chromatography, the following buffers were used: A) 1PBS pH 7.4 (Na.sub.2HPO.sub.4 (10 mM), KH.sub.2PO.sub.4 (1.8 mM), NaCl (137 mM), KCl (2.7 mM), 15 mS/cm, and B) 0.1 M glycine/HCl, pH=3.5, 4 mS/cm for elution

(11) Different concentrations (0.5-5 mg/ml) of human IgG (SeraCare, HS-475) and 1 mg/ml BSA in buffer A (see above), or a prefiltered undiluted cell-free culture supernatant of an IgG.sub.1-expressing CHO (Chinese Hamster Ovary) cell line (DG44 Cellca), were used for loading. The detailed description of the experimental procedures is found in the respective exemplary embodiments.

(12) Anion-Exchange Chromatography

(13) For the experiments with anion-exchange chromatography, the following buffers were used: C) Tris-HCl pH 7.4 (20 mM) 1.8 mS/cm, and D) Tris-HCl pH 7.4 (20 mM), NaCl (1 M), 87 mS/cm for elution

(14) The loading solution was prepared by dissolving bovine serum albumin (BSA) (2 g/l, Kraeber & Co No. 4909 2052) and lysozyme (5 g/l, Roth No. 8259.2) in buffer C and filtering this solution across a Sartolab filter (PESU membrane, 0.2 m pore size, Sartorius Stedim Biotech GmbH). The influence of the membrane adsorber (see examples below) was investigated by flushing the unit (Sartobind Q Nano with and without downstream membrane adsorber) with 30 ml of buffer C, then loading with 120 ml of BSA (2 mg/ml), lysozyme (5 mg/ml) in buffer C, washing with 30 ml of buffer C, then eluting with 30 ml of buffer D and lastly equilibrating again with 30 ml of buffer C.

(15) Determination of the Degree of Breakthrough and of Loss of Product

(16) In all the described experiments, the column material has only a low capacity for contaminants and a distinctly higher capacity for the target molecule; accordingly, the breakthrough of the contaminants occurs first and is visible as a plateau in the UV absorption at the output of the column. Only at higher loading volume is the capacity for the target molecule exhausted and the breakthrough of the target molecule observed. The degree of the breakthrough of the target molecule can be calculated on the basis of the following relationship:

(17) $\begin{array}{c}\mathrm{Breakthrough}\mathrm{of}\\ \mathrm{target}\mathrm{molecule}\left(\%\right)\end{array}=\frac{\left({\mathrm{UV}}_X-{\mathrm{UV}}_{\mathrm{Contaminant}}\right)}{\left({\mathrm{UV}}_{\mathrm{Total}}-{\mathrm{UV}}_{\mathrm{Contaminant}}\right)}100$

(18) where

(19) UV.sub.x is the UV signal at a particular loading volume[mAU], UV.sub.contaminant is the UV signal at complete breakthrough of the contaminant (base line) [mAU] and

(20) UV.sub.Total is the UV signal at complete breakthrough of the contaminant and of the target molecule [mAU].

(21) Relative loss of product can be directly calculated from the breakthrough curve via the ratio of the area integral of the UV absorption of loaded target molecule and unbound target molecule (below the breakthrough curve):

(22) $\mathrm{Loss}\mathrm{of}\mathrm{product}\left(\%\right)=\frac{\mathrm{Area}\mathrm{below}\mathrm{the}\mathrm{curve}}{\mathrm{Total}\mathrm{area}}100$

Example 1-Influence of the Bed Height of the Downstream Membrane Adsorber

(23) In the first example, IgG (SeraCare, HS-475) is purified with the aid of protein A affinity chromatography. Protein A binds with very high affinity to the Fc region of the heavy chain of immunoglobulins, with contaminants binding only very weakly to the column material and being in the flow-through during loading of the column. The main medium used (corresponding to the first chromatography system) is a protein A column (HiTrap Protein A, article No. 17-0403-01, GE Healthcare) having a bed volume of 5 ml and a binding capacity of 30 mg IgG/ml. By means of the downstream connection of a membrane adsorber (MA 1, Sartobind Protein A membrane article No. 93PRAP06HB-12-A, capacity 6.5 mg/ml) with variable dimensioning (0.1-2 ml), the influence of the bed height on the loss of product, the sensitivity of detection and the back pressure during loading of the structure is investigated (Table 1). To this end, the loading experiments are carried out with a mixture of IgG (2.5 mg/ml) as target molecule and BSA (1 mg/ml) as contaminant.

(24) TABLE-US-00001 TABLE 1 Number of layers, and size of the total area, of the volume and of the dead volume of the individual membrane adsorbers (MA 1) Bed Membrane Dead height, No. of Flow Total volume, volume, mm layers area, cm.sup.2 area, cm.sup.2 ml ml 0 0 0 0 0 0 0.2 1 5 5 0.1 0.5 1 5 5 25 0.5 0.6 2 10 5 50 1.0 1.1 3 15 5 75 1.5 1.2 4 20 5 100 2.0 1.5

(25) To this end, each structure is first equilibrated with ml of buffer A, then loaded with 140 ml of IgG (2.5 mg/ml) and BSA (1 mg/ml) in buffer A, washed with 70 ml of buffer A and then eluted with 50 ml of buffer B. In the chromatogram, the early breakthrough of the contaminant (BSA, UV.sub.Contaminant=135 mAU) followed by the breakthrough of the target molecules (40-60 ml) (UV.sub.Total=680 mAU) can be seen (cf. FIG. 2). At the same time, the influence of the downstream membrane adsorber can be seen in an increase in the steepness of the breakthrough curves with increasing size of the adsorber and with an increased capacity (later breakthrough) (cf. FIG. 2).

(26) In the case of practical application, the additional capacity introduced by the membrane adsorber can be used for the optimal utilization of the column together with a minimization of the loss of product. In the case of the breakthrough curve without downstream membrane adsorber, the termination criterion for loading that is used is the maximum of the first derivative of the UV absorption. In the case of the experiments with downstream membrane adsorber, the maximum of the difference of the signal (in mAU) of the sensor in front of the membrane adsorber within a defined volume interval is used as the termination criterion (UV=UV.sub.X+YUV.sub.X). The volume interval can be selected on the basis of the concentration of the target molecule in the feed solution, the capacity of the membrane adsorber, and the sensitivity of the sensor used. In this example and the following examples, the volume interval is set conservatively with the dead volume of the downstream membrane adsorber (UV=UV.sub.X+DVUV.sub.X) without taking into account the capacity of the downstream membrane adsorber.

(27) It is possible to identify a distinct relationship from the size of the selected volume difference and the amplitude of the UV signal, though the position of the maximal values on the x-axis remains the same (cf. FIG. 3). What is noticeable is the larger amplitude of the UV signal in comparison with the first derivative at larger membrane volumes (1.5 ml and 2 ml), and this is an advantage when determining the breakthrough in the case of poorly detectable signals (in the case of high concentration of contaminants and low concentration of target molecules). Using the described termination criteria, it becomes clear that the enlargement of the volume of the downstream membrane adsorber makes it possible to achieve a minimization of the loss of product from 3% to 0% (cf. Tab. 2).

(28) TABLE-US-00002 TABLE 2 UV signal, pressure during loading, breakthrough of the target substance (IgG) in %, and corresponding loss of product Average Bed UV (mAU) pressure Loss of height, or during Breakthrough, product, mm (mAU/ml) loading, MPa % % 0 20.7 0.09 32.4 2.95 0.2 11.3 0.09 29.9 2.14 1 13.2 0.10 14.8 0.63 2 23.4 0.12 0.6 0.20 3 25.1 0.15 0 0.00 4 31.1 0.17 0 0.00

(29) At the same time, it is, however, also possible to observe an increase in the back pressure, which correlates with the size of the downstream membrane adsorber (cf. Tab. 2 and FIG. 4). Therefore, the increasing size of the membrane adsorber yields a contrast between of product (which approaches zero) and the increase in the back pressure during loading. This relationship may be crucial for the dimensioning of the membrane adsorber, since it is ideally integrable into already existing systems having a maximum permissible back pressure.

(30) It was possible to load the 5 ml protein A column used here with 152 mg of IgG until complete breakthrough of the target molecule. However, this capacity is generally not fully utilized. To prevent a loss of the target molecule, column materials on a process scale are only loaded to approx. 60% of their capacity (in the case of the column used here, this would correspond to 91 mg of IgG). In the structure described here, a loading of the column with 127 mg of IgG can be realized until the termination criterion is reached, corresponding to an 83% utilization of the existing binding capacity of the column.

(31) Overall, it can be stated that the downstream membrane adsorber can catch target molecules which, during a reliable detection of the breakthrough, were lost after the column. In this connection, the structure benefits from the property of the steeper breakthrough curve following the membrane adsorber, since it is possible as a result to completely avoid loss, even with relatively small dimensioning of the membrane adsorber (6.5% of the capacity of the main medium, 3 mm bed height).

Reference Example 2-Size of the Flow Area of the Downstream Membrane Adsorber

(32) An important parameter in chromatography is the elution of the target molecules. Here, care has to be taken that the elution volume should not be excessively increased by the fitting of an additional membrane adsorber. Furthermore, peak symmetry during elution is an important parameter. The geometry of the membrane adsorber has a crucial influence on the linear flow rate of the sample through the membrane and can thus have an influence on the binding and the elution of the target molecules. To investigate this effect, a solution of IgG (5 mg/ml) and BSA (1 mg/ml) is loaded onto a protein A column and then eluted (cf. FIG. 5) in this experiment. The column (5 ml) is used on its own in one case; it is used with a downstream membrane adsorber having a 5 cm.sup.2 flow area in the other case (MA1) and with a downstream membrane adsorber having. a 20 cm.sup.2 flow area in the further case (MA2), with the same bed volume of 2 ml in both cases, in order to thus achieve a variation of the flow area by a factor of 4 (cf. Tab. 3).

(33) TABLE-US-00003 TABLE 3 Properties of the membrane adsorbers used MA 1 MA 2 Flow Axial Axial No. of layers 20 5 Bed height 4 mm 1 mm Bed volume 2 ml 2 ml Flow area 5 cm.sup.2 20 cm.sup.2 Diameter 30 mm 57 mm Membrane area 100 cm.sup.2 100 cm.sup.2 Dead volume 1.5 ml 2.8 ml

(34) To test the effects of the altered flow area, the units were now equilibrated with 50 ml of buffer A (see Materials and methods), then loaded with 80 ml (column only) or 100 ml (column+membrane adsorber) of IgG (5 mg/ml) and BSA (1 mg/ml) in buffer A, washed with 70 ml of buffer A, and then eluted with 60 ml of buffer B (cf. FIG. 5).

(35) Besides the already described steeper rise of the breakthrough curve of the target molecule in the presence of the membrane adsorbers (cf. FIG. 5, and also example 1), it is also possible to observe a change in the shape of the elution peak. Especially the tailing factor (T) of the elution peak is an important parameter for describing this circumstance and can be represented by means of a simple equation:

(36) $T=\frac{a+b}{2a}.$

(37) Here, a and b are the distances which describe the extent of the elution peak (fronting and tailing) from the peak center at 1/10 of the maximum peak height (h) (FIG. 5). In the case of a value of T1, the peak is asymmetrical, with a value of T<1 describing fronting and a value of T>1 describing tailing. In all cases, it is possible to observe a tailing of the elution peak, it being possible to observe the least tailing in the case without membrane adsorber, only a slight increase in tailing in the case of MA 1 and a distinct tailing in the case of MA 2 (cf. FIG. 5 and Tab. 4).

(38) Furthermore, the elution volume required in order to elute the target molecule from the column was measured (cf. Tab. 4). In general, the results show a rise in the elution volume in the presence of a membrane adsorber, the membrane adsorber having the larger flow area (MA 2) exhibiting a distinct peak broadening and thus a more distinct increase in the elution volume (factor of 2.2) in comparison with the membrane adsorber having the smaller flow area (MA 1, factor of 1.2).

(39) TABLE-US-00004 TABLE 4 Influence of the different flow areas on the elution of target molecules Column Column + MA 1 Column + MA 2 Tailing 1.31 1.39 1.90 factor (T) Elution 7.6 9.5 16.8 volume (ml) Volume 1.0 1.2 2.2 increase

(40) It has to be assumed that both the dead volume of the membrane adsorber and the linear flow rate along the radial axis have an influence on the binding of the target molecules on the membrane. Therefore, the increase in the elution volume in the case of the unit having a broad flow area can be explained by differences in the residence times in the radial direction of flow. This may be crucial for applications in which a minimization of the elution volume is important (this is generally the case for many applications, since they often involve saving costs for buffers and further dilutions, for example for adjusting the buffer before introduction into a further process step) and must be absolutely observed in the application of the method.

Example 3-Applicability of the Method to Ion-Exchange Chromatography

(41) The general applicability of the method to other chromatographic matrices is demonstrated in the following example by separating a mixture of lysozyme (5 mg/ml) and bovine serum albumin (BSA, 2 mg/ml) by means of anion-exchange chromatography at a pH of 7.4. Whereas lysozyme has an isoelectric point of 11.4, BSA has an isoelectric point of 4.7. Therefore, lysozyme has a positive net charge at the pH used and only binds to a very slight extent to the anion exchanger, whereas BSA is negatively charged and therefore has a strong affinity for the positively charged quaternary ammonium groups RCH.sub.2CH.sub.2N.sup.+ (CH.sub.3).sub.3 of the anion exchanger. Therefore, lysozyme serves as a model for a contaminant and BSA serves as a model for the target molecule. The ligands and the membrane are identical both in the case of the main medium (Sartobind Q 3 ml membrane adsorber) and in the case of the smaller dimensioned membrane adsorber (MA 1, Sartobind Q 0.4 ml); both have a dynamic binding capacity of 30 mg BSA/ml (see Tab. 5). However, the two media differ in the dimensioning; the second membrane adsorber is smaller and only has 13% of the capacity of the main medium.

(42) TABLE-US-00005 TABLE 5 Parameters of the structure for separating BSA and lysozyme on anion-exchanger membranes (MV = membrane volume) Parameter Main medium MA 1 Flow Radial Axial Binding capacity 30 mg/ml MV 30 mg/ml MV (BSA) Bed volume 3 ml 0.4 ml Dead volume 4.3 ml 0.9 ml

(43) The detection of the loading status takes place at the output of the main medium. To this end, the slope of the UV signal due to the UV value is determined as already described. In this example, the volume interval again corresponds to the dead volume of the downstream membrane adsorber (MA 1, 0.9 ml, cf. example 1). The termination criterion for loading that is selected is the maximal value of the UV curve (approx. 36 ml, cf. FIG. 6).

(44) In this method, a 2% loss of product is achieved in the absence of downstream membrane adsorber (cf. Tab. 6).

(45) TABLE-US-00006 TABLE 6 Detection of the breakthrough of BSA Loading mAU Breakthrough, Loss of volume, ml (280 nm) % product, % Main 36 2024 29 2 medium Main 36 1997 0 0 medium + MA 1

(46) However, the downstream membrane adsorber catches the majority of the target molecules which did not bind to the main medium (FIG. 6, dashed line). Thus, the loss of product can be minimized (not detectable) when using the membrane adsorber in the case of loading of the main medium until breakthrough of the target molecule and thus optimal utilization of the capacity. The concept is thus absolutely not restricted to protein A chromatography and is readily transferrable to other systems.