Purification method and uses thereof

Abstract

A cyclic chromatographic purification method for the separation of at least two overlapping compounds A and B from a feed mixture (F), wherein the method comprises the following steps: a. a first interconnected step (IC1), wherein a fraction containing A and B in overlap is eluted into the downstream adsorber (1) and b. a first batch step (B1), in which said adsorbers (1,2) are disconnected during a first batch purification timespan wherein the following steps are carried out with the upstream adsorber (2): 1. elution of said compound B 2. optional flushing; 3. regeneration; wherein the following steps are carried out with the downstream adsorber (1): 1. loading of the feed mixture (F); 2. optional washing; 3. elution of said compound A; and wherein after completing the steps a. and b. the adsorber positions are exchanged and the steps repeated.

Description

[0001] F07460 9.1.2026

[0002] 1

[0003] TITLE PURIFICATION METHOD AND USES THEREOF

[0004] TECHNICAL FIELD

[0005] The present invention relates to a cyclic chromatographic method for separating compounds with similar adsorptive properties.

[0006] PRIOR ART

[0007] The manufacturing process of therapeutic substances such as peptides, oligonucleotides and proteins typically relies on chromatography as main purification step. To ensure that the product meets specifications, the chromatography step usually requires a center-cut purification to recover the product that is flanked by product-related impurities in the front and in the back of the product peak. As the product-related impurities have similar adsorptive properties as the compound of interest, their peaks usually overlap with the product peak, leading to only partially pure product-containing side fractions.

[0008] Minimizing the overlaps through gradient chromatography leads to an improved product yield and purity. Due to their robustness, linear gradients are employed frequently. The retention of product and impurities, and thus their overlap, can be influenced by modifying the gradient slope. Running a steep linear gradient leads to high throughput (productivity) but lower product purity due to larger overlaps.

[0009] Running a shallower linear gradient improves the separation and reduces the overlaps, however it also increases retention of product and impurities, thereby extending the chromatographic purification time, which can become prohibitive in a manufacturing scenario that requires a certain throughput, in particular in continuous production processes.

[0010] Continuous countercurrent chromatography processes can be used to improve the separation, thereby keeping throughput high. These processes typically feature automatic recycling of impure side fractions and can be classified into single- and multi-adsorber recycling processes.

[0011] Single-adsorber setups involve the recirculation of the chromatographic profile within the same adsorber. In the steady-state-recycling (SSR) process, fractions from the leading and trailing edges of the circulating chromatographic profile are collected, while fresh samples are injected into the center of the profile.

[0012] In contrast, multi-adsorber recycling processes enable the internal recycling of impure side fractions from one adsorber to another, utilizing counter-current principles. This setup allows a relative opposite movement between the stationary and mobile phases, enhancing theF07460 9.1.2026

[0013] 2

[0014] resolution of product and impurities. Automatic recycling in these processes eliminates the need for external storage, handling, and analysis of side fractions, as only the purified product is collected, thereby maximizing yield.

[0015] Multi-adsorber processes that combine internal recycling with counter-current principles are known as simulated moving bed (SMB) processes. Early SMB processes, however, were limited to separating two compounds (binary separations) and lacked the ability to operate under linear gradient conditions, restricting their use in applications that required center-cut separations within the chromatogram.

[0016] Based on the concept of Simulated Moving Bed (SMB) an efficient center-cut (ternary) separation process with linear solvent gradient capabilities was developed, known as the Multicolumn Countercurrent Solvent Gradient Purification (MCSGP) process.

[0017] WO-A-2006116886 describes this MCSGP process for continuous or quasi-continuous purification of a multi-component mixture by means of individual chromatographic columns through which the mixture is fed by means of at least one solvent. The multi-component mixture at least comprises light impurities, an intermediate product to be purified and heavy impurities, and the columns are grouped into at least four sections, in which the first section is provided with at least one inlet of solvent and at least one outlet for purified intermediate product, such that it washes the purified intermediate product out of the system, but keeps the heavy impurities inside the section, the second section is provided with at least one inlet of solvent and at least one outlet connected to an inlet of the fourth section, such that it washes the intermediate product, which is contaminated with heavy impurities into the fourth section through said outlet, but keeps the pure heavy impurities inside the section. The third section is provided with at least one inlet of solvent and an outlet for heavy impurities, such that it washes out the heavy impurities through said outlet and cleans the chromatographic column(s), the fourth section is provided with at least one inlet to receive output of the outlet of the second section as well as at least one inlet for feeding in the multicomponent mixture and at least one outlet for light impurities, such that it washes the light impurities out of the system, but keeps the intermediate product inside the section. After or within a switch time the last column from the first section is moved to the first position of the second section, the last column of the second section is moved to the first position of the third section, the last column of the third section is moved to the first position of the fourth section and the last column of the fourth section is moved to become the first column of the first section. The process can be carried out also with only two columns.

[0018] This MCSGP method is now well-established in the industry. Other multi-adsorber chromatographic techniques that utilize several adsorbers and internal recycling have also been proposed, including a "gradient with steady state recycle" (GSSR) process.F07460 9.1.2026

[0019] 3

[0020] Although MCSGP has been described for configurations with 2 to 8 columns, in practice, the 2-adsorber setup is predominantly used due to its lower equipment complexity and greater operational flexibility compared to configurations with more columns. The MCSGP process is designed based on a single-column batch chromatogram. Generally, MCSGP can achieve product purity equivalent to the purest fraction from the batch chromatogram. A control method for sequential loading processes based on continuously determining the binding capacity of each column in cyclic steady state has been described in WO-A-2010151214. The method comprises measuring a feed signal, representative of the composition of a feed material supplied to the inlet of the column, measuring an effluent signal representative of the composition of the effluent from the column, and using the feed signal and the effluent signal, to determine binding capacities from a so called "deltasignal". A major drawback of this method is that the feed signal is determined using a first detector and the effluent signal is determined using a second detector. Despite the detectors being of the same type, the detectors must be calibrated accurately and on a regular basis in order to derive useful information from the deltasignal such as breakthrough and saturation points that are ultimately used to derive control actions. The method requires the determination of a feed signal by a separate detector whose only purpose is the determination of the feed signal. The number of required detectors is therefore equal to the number of columns +1. The method cannot be used in cases the breakthrough curves cannot be accurately monitored, for instance due to a high impurity signal or due to a low product signal in the feed or both.

[0021] Such a process with only two columns is described in WO-A-2014166799 including methods for setting up such a process, and also control and / or monitoring and / or optimization in this context. The method comprises, within one cycle to be carried out at least once, the following steps: a first batch step, wherein during a batch timespan said columns are disconnected and a first column is loaded with feed via its inlet using a first flow rate and its outlet is directed to waste, and from a second column desired product is recovered via its outlet and subsequently the second column is regenerated; a first interconnected step, wherein the outlet of the first column is connected to the inlet of the second column during an interconnected timespan, the first column is loaded beyond its dynamic breakthrough capacity with feed via its inlet using a second flow rate which is the same or larger than the first flow rate, and the outlet of the second column is directed to waste, a second batch step analogous to the first batch step but with exchanged columns; a second interconnected step, analogous to the first interconnected step but with exchanged columns.

[0022] WO-A-2019096622 proposes a method for monitoring, evaluating and controlling a cyclicF07460 9.1.2026

[0023] 4

[0024] chromatographic purification process involving at least two adsorbers, wherein the method comprises at least the following steps: a) monitoring of the chromatogram including the measurement of at least one current concentration-proportional signal in the liquid; b) evaluation of the chromatogram including a comparison of at least one of said current concentration-proportional signals measured in step (a) with a threshold value thereof; c) controlling the chromatographic purification process by adapting the termination of the currently running phase as a function of the comparison of step (b) and initiating the next phase, wherein the sequence a)-c) is carried out in given order at least twice.

[0025] Some purification challenges encountered in the industry are pseudo-ternary, i.e. the product peak only overlaps with impurities on the front side of the peak or at the tail side, while a third or even fourth impurity group may be present, eluting without overlap (it is base-line separated).

[0026] In such cases a novel continuous countercurrent chromatography process can be derived from the MCSGP process by simplification and reorganisation of the process tasks, leading to a downstream process with further improved throughput.

[0027] As opposed to traditional SMB, the process can use linear gradients while operating with only two adsorbers.

[0028] Abel et al (S. Abel et al. I J. Chromatogr. A 1043 (2004) 201-210, https: / / doi.Org / 10.1016 / j.chroma.2004.05.094) have demonstrated an SMB process for the abovementioned pseudo-ternary purifications, termed “3F-SMB”. However, this process requires a minimum of 4 columns and has no linear gradient capabilities.

[0029] WO-A-2014131719 discloses a chromatographic process for the enrichment of at least one compound of interest from a mixture comprising said at least one compound of interest as well as at least one further compound, which is not of interest, using at least two chromatographic columns, wherein said process at least involves a sequence of the following steps: (i) a cyclic accumulation phase, in which the at least two chromatographic columns are alternatingly operated in an interconnected phase, followed by a disconnected phase, wherein after these two phases the first and second column or group of columns exchange places; wherein said two phases are carried out sequentially M times with M > 1 ; (ii) a cyclic separation phase, in which the at least two chromatographic columns are alternatingly operated in an interconnected phase, followed by a disconnected phase wherein after these two phases the first and second column or group of columns exchange places to undergo the next interconnected and disconnected phases (B); wherein said two phases are carried out sequentially N times with N > 0; (iii) an elution phase, in which from the column or group of columns, which at the end of phase (i) or, in case of N>0 at the end of phase (ii) contains the at least one compound of interest, said at least one compound ofF07460 9.1.2026

[0030] 5

[0031] interest is extracted via the outlet.

[0032] WO-A-2018011102 discloses a cyclic chromatographic purification method for the isolation of a product from a feed mixture consisting of the product and at least one further component representing impurities, which impurities bind stronger to the chromatographic stationary phase than the product is given. The method uses at least two chromatographic adsorbers as chromatographic stationary phase, grouped into only one first adsorber section (1) and one second adsorber section (2), wherein if an adsorber section comprises more than one chromatographic adsorber these are permanently connected in series, wherein the first adsorber section (1) has a first adsorber section inlet and a first adsorber section outlet, and the second adsorber section (1) has a second adsorber section inlet and a second adsorber section outlet.

[0033] SUMMARY OF THE INVENTION

[0034] The present invention aims at providing a simplified MCSGP-like process that can reach higher product purity than a regular MCSGP process for cases of pseudo-ternary or binary purifications.

[0035] It was surprisingly found that in case of simpler purification challenges i.e. of pseudo-ternary or binary purifications instead of ternary purifications, the process scheme shown in WO-A-2006116886 (Fig. 8 thereof and for using two columns only in particular Fig. 28 thereof) can be simplified and re-arranged, leading to significant process time savings and increased throughput.

[0036] As binary and pseudo-ternary purification chromatograms display only one region of overlap of the compounds to be separated (in the examples given here A and B), only one step of internal recycling for the overlapping portion is required to keep the compounds contained therein inside the process. Therefore, generally speaking, the p section of the process in WO-A-2006116886 (so the interconnected step, in which the two columns are interconnected and in which the second overlapping fractions, B and C in WO-A-2006116886, are recycled, see in particular step 3. in Fig. 28 of WO-A-2006116886) is not required.

[0037] Consequently, the batch phases a (see step 2. in Fig. 28 of WO-A-2006116886) and y (see step 4. in Fig. 28 of WO-A-2006116886), which both operate without adsorber interconnection, can be combined into one phase of one adsorber and can be carried out in parallel to the combined phases 5f(loading of new feed mixture F to be purified and starting to take out A) and 8g(potentially taking out more A) with the other adsorber.

[0038] From the adsorber combining previous phases a and y, one compound (compound B in case of binary mixture A, B separation) or two baseline-separated compounds (B and B’)F07460 9.1.2026

[0039] 6

[0040] can be eluted in pure form. Following the elution of B or B / B’, a regeneration step can be carried out with this adsorber to clean and / or equilibrate the adsorber in preparation of the next process step.

[0041] In parallel in the other adsorber the phase 5fis carried out, so following the load of F, pure (base line separated) compound A’ can be eluted, if present, followed by an elution of pure compound A.

[0042] It was found that operation of the abovementioned sections for B / B’ elution and regeneration on one hand with one adsorber, and on the other hand with the other adsorber load of F and elution of A7A, can be parallelized, in the so-called batch phase.

[0043] After completion of such a new batch step, the adsorbers are interconnected. In the following interconnected phase, the region of the chromatogram containing the overlap of A and B is internally recycled from the adsorber that had previously eluted pure compound A into the adsorber that had previously been regenerated, corresponding to the sections 8gand 8rin WO-A-2006116886, respectively (see step 1. in Fig. 28 of WO-A-2006116886). According to a first aspect of the present invention, it relates to a method of cyclic chromatographic purification method for the separation of at least two overlapping compounds, a more weakly adsorbing compound A and a more strongly adsorbing compound B, from a feed mixture F. The method is using at least two chromatographic adsorbers as stationary phase, wherein each adsorber has an adsorber inlet and an adsorber outlet. The compounds A and B (or at least one of them) are the compounds of interest and are individually collected for further use when in the following stated to be eluted in a batch step from an adsorber.

[0044] Typically, the method is carried out with two adsorbers (typically columns) only. It can however also be carried out with more than two adsorbers. If mention is made of only using two adsorbers, this includes the situation where there are for example two physical columns for each adsorber which are always interconnected and fulfil the function of one single such adsorber. Typical adsorbers used in the processes described here, are packed bed columns, membrane adsorbers or monolithic stationary phase.

[0045] The method comprises the following steps:

[0046] a. a first interconnected step IC1, in which two adsorbers are interconnected defining an upstream absorber and a downstream adsorber, in that the upstream adsorber outlet is connected to the downstream adsorber inlet, during a first interconnected purification timespan tici ,

[0047] wherein the upstream adsorber is loaded via the upstream adsorber inlet with eluent and wherein a fraction containing said more weakly adsorbing compound A and said more strongly adsorbing compound B in overlap isF07460 9.1.2026

[0048] 7

[0049] eluted from the upstream adsorber into the downstream adsorber and wherein the stream exiting the upstream adsorber outlet is adjusted inline before entering the downstream adsorber inlet;

[0050] b. a first batch step B1 , in which said adsorbers are disconnected during a first batch purification timespan tsi

[0051] wherein the following steps (substeps of B1) are carried out with the upstream adsorber of said first interconnected step IC1 in given order: 1. elution (and correspondingly individual collection) of said more strongly adsorbing compound B and, if present, elution (and correspondingly individual collection) of at least one further strongly adsorbing compound B’ that is even more strongly adsorbing than said more strongly adsorbing compound B and eluting later than said more strongly adsorbing compound B and is baseline-separated from said more strongly adsorbing compound B;

[0052] 2. optional flushing;

[0053] 3. regeneration;

[0054] wherein the following steps (substeps of B1) are carried out with the downstream adsorber of said first interconnected step IC1 in given order: 1. loading of the feed mixture F;

[0055] 2. optional washing;

[0056] 3. elution (and correspondingly individual collection) of said more weakly adsorbing compound A, preceded, if present, by elution (and correspondingly individual collection) of at least one further weakly adsorbing compound A’, that is even more weakly adsorbing than said more weakly adsorbing compound A and that is eluting earlier than said more weakly adsorbing compound A and is baseline- separated from said more weakly adsorbing compound A.

[0057] After completing the steps a. and b. the adsorber positions are exchanged such that the previously upstream adsorber becomes the downstream adsorber and the previously downstream adsorber becomes the upstream adsorber, and the steps a. and b. are repeated as second interconnected step IC2 and second batch step B2, respectively, with the adsorbers in this exchanged position such that the previously upstream adsorber performs the tasks of the downstream adsorber, and the previously downstream adsorber performs the tasks of the upstream adsorber.

[0058] The above sequence of first interconnected step IC1, first batch step B1, second interconnected step IC2, second batch step B2 is carried out at least twice in a row.F07460 9.1.2026

[0059] 8

[0060] Key of the proposed process, in particular in relation with the above-mentioned prior art document WO-A-2014131719, is that in the proposed process here compounds A and B, both of interest and collected individually, are recovered in steps b.1 of the upstream adsorber and b.3 of the downstream adsorber regularly, i.e. in every cycle. So, a cyclic steady state is reached, wherein compounds of interest A and B are recovered with constant concentration and purity from cycle to cycle.

[0061] In contrast, in WO-A-2014131719 the aim of the process is the accumulation of the compounds of interest (X) for the purpose of enrichment to recover them in a final elution step. Due to the accumulation, the process does not reach a cyclic steady state.

[0062] In a preferred process, these four steps are not only carried out twice in a row, but at least five times, or at least 20 times, or even 100 times, also possible is an essentially continuous repetition until the adsorber material is degraded unacceptably. This is in particular of benefit if there is a continuous production process.

[0063] The presented process can use eluent gradients as for example linear or non-linear gradients in any of the phases, (interconnected or parallel) with a gradient in the form of a temporally changing modifier concentration increase or decrease. Linear gradients improve process robustness, by warranting consistent elution times of the compounds of interest from cycle to cycle in response to process parameter changes, such as temperature or eluent concentration variations. So preferably eluent gradients are used, preferably linear eluent gradients, in at least one or all of the steps (normally linear eluent gradients in all steps except in the flushing, regeneration, loading and washing substeps of B1 and B2), with a gradient in the form of a temporally changing modifier concentration increase or decrease, wherein preferably such linear eluent gradients are applied in the interconnected steps and in the elution steps, and preferably not in flushing, regeneration, loading and washing steps. Preferably essentially the same gradient slope in the interconnected steps is selected as for the batch steps.

[0064] So linear elution gradients are preferentially used in the phases of A, A / B and B (i.e. in IC1, IC2, and A, B elution). Elution gradients can also be used in the flushing, regeneration, loading and washing substeps, but usually not in the form of linear elution gradients, but for example as step elution gradients.

[0065] Preferably, in batch substeps of the method without more strongly adsorbing compound (B) elution, the previously upstream adsorber is cleaned, preferably with eluent with a higher modifier concentration than at the end of the preceding interconnected recycling step or with a different modifier or with a cleaning solution, and / or re-equilibrated.

[0066] The presented process generally uses inline adjustment in the interconnected steps whereby the upstream adsorber inlet is loaded via the upstream adsorber inlet with eluentF07460 9.1.2026

[0067] 9

[0068] with a gradient in the form of a temporally changing modifier concentration and the stream exiting the upstream adsorber outlet is adjusted inline before entering the downstream adsorber inlet with eluent without modifier or with a different modifier concentration than at the inlet of the upstream adsorber.

[0069] By means of the inline adjustment during the interconnected phase, a rapid re-adsorption of the compounds of interest in the downstream adsorber is ensured, preventing a further elution from the downstream adsorber following the elution from the upstream adsorber. Following the elution of the compound of interest, in the batch step of the process the previous upstream adsorber is cleaned, preferably with eluent with a higher modifier concentration than at the end of the preceding interconnected recycling step or with a different modifier or with a cleaning solution, and re-equilibrated.

[0070] This step ensures that the adsorber is ready to adsorb compounds of interest from the upstream adsorber in the next interconnected step.

[0071] During this cleaning step, strongly adsorbing compounds B’ may be eluted if present. In the gradient elution the modifier may be selected from the group consisting of an organic or inorganic solvent or mixture thereof different from a base solvent or mixture thereof of the eluent, an electrolyte in such an organic or inorganic solvent (or a mixture thereof), preferably selected from a dissolved salt or a pH, or a combination thereof.

[0072] Preferably said base solvent may be water or a mixture of water with at least one organic solvent or water in a mixture with one or more salts and / or organic solvents one or both in a minor proportion compared with water.

[0073] Preferably said modifier may be an organic solvent or a mixture of water with at least one organic solvent having a higher concentration of said at least one organic solvent than in the base solvent, water or a mixture of water with at least one organic solvent with a different salt or H+concentration than the base solvent.

[0074] For example the base solvent can be a mixture of water 99.9% and trifluoroacetic acid (TFA, 0.1%), and the modifier can be a mixture of 9.9% water, TFA 0.1%, and 90.0% Acetonitrile. Alternatively the modifier can be 100% acetonitrile. In another example the base solvent can be an aqueous 25 mM phosphate buffer, pH 7.0, and the modifier can be a 25 mM phosphate buffer, 500 mM NaCI, pH 7.0.

[0075] The base solvent is, in particular in case of biomolecules to be separated, for example resulting from a biochemical process, normally water or water in a mixture with one or more salts and / or organic solvents in a minor proportion compared with water (for example supplemented with acetonitrile and trifluoroacetic acid), in the following this base solvent will be called solvent A. For establishing the gradient, it is mixed with a further solvent or solvent mixture different from the base solvent (mixture). This further solvent for exampleF07460 9.1.2026

[0076] 10

[0077] can be a mixture of the same solvents as the base solvent but having different proportions (e.g. for the above example water supplemented with a higher proportion of acetonitrile). In particular for biomolecules typically this further solvent is again based on water but has a further increased proportion of organic solvents or salt. For establishing the gradient typically a modifier mixture with a rather low concentration of the constituent (e.g. organic solvent(s), a salt or pH, or a combination thereof, if e.g. the base solvent is water) differing from the base solvent is provided as eluent at the beginning of the gradient and is increasingly mixed by means of a gradient pump with the further solvent leading to a corresponding controlled gradient. The feed mixture can be provided in a different solvent, or it can be provided in the base solvent or the feed mixture can be provided as a mixture of the original solution (typically a water solution) in a mixture with the base solvent of appropriate concentration.

[0078] Preferably linear eluent gradients are used in at least one or all of the phases with a linear gradient in the form of a temporally changing modifier concentration increase.

[0079] The presented process may be operated with a start-up phase and / or a shutdown phase. In the start-up phase, before carrying out the first interconnected recycling step IC1, during a first batch start-up timespan (te-su) said adsorbers are disconnected

[0080] and the adsorber to become the upstream adsorber of the first interconnected IC1, is carrying out the tasks of adsorber section 2 in step B2, i.e. loading of feed mixture F (in this case optionally proceeded by equilibration), washing, and elution of A, wherein the amount of feed mixture F supplied to the adsorber may differ from the amount supplied in later phases IC1, B1, IC2, B2,

[0081] while the adsorber to become the downstream adsorber of the first interconnected step IC1 is either being equilibrated or already equilibrated and inactive or regenerated.

[0082] The start-up step abbreviates the achievement of a cyclic steady state of the presented process, i.e. a state where concentration and quality of the eluted compounds of interest are constant from cycle to cycle.

[0083] Likewise, the presented process may be operated with a shutdown step wherein after termination of a B2 phase, a shut-down sequence is carried out, comprising the phases IC1 and B1-SD wherein the following steps are applied to the previous downstream adsorber:

[0084] 1. Washing

[0085] 2. Elution of compound A, preceded by elution of compound A’, if present, that is eluting earlier than compound A and is baseline-separated from compound A,

[0086] and the following step is applied to the previous upstream adsorber:

[0087] 1. elution of compound B and, if present, compound B’ that is eluting laterF07460 9.1.2026

[0088] 11

[0089] than compound B and is baseline-separated from compound B.

[0090] and wherein both adsorbers are subjected to optional final regeneration and / or storage steps.

[0091] The shutdown step ensures that the maximum amount of the compounds of interest is recovered while no feed mixture is wasted.

[0092] Through the parallelization of the tasks of the two adsorber sections a better flexibility with respect to the duration of the different tasks can be achieved and the scheduling can be optimized. The only operational constraint for proceeding to the following interconnected step is that both operations of the adsorber sections be completed. To achieve an earlier completion, there is high flexibility to balance a high duration of one task (e.g. Load F) by a short duration of another sequential task (e.g. elution of A’), which may be achieved by a flow rate increase. Likewise, in the other adsorber section, a high duration of the elution of B can be balanced by a shorter regeneration time. With respect to the process presented in WO-A-2006116886, this means that in the presented process, the duration of the sum of the tasks a and y needs to match the sum of the durations of tasks 8fand the disconnected part of 8g. This is a much more flexible constraint than the constraints in WO-A-2006116886, which command firstly, that task a and task 5fbe completed within the same time interval, and secondly, that the task y and the of the disconnected part of 8gbe completed within the same time interval. As the batch / parallel phases a and y (8f and the disconnected part of 8g, respectively) are separated by the task p, for each batch / parallel phase a waiting time interval may arise for one of the adsorbers in order to synchronize the adsorber tasks for the next interconnected phase. Therefore, for each half-cycle of WO-A-2006116886, two waiting time intervals may be observed. Instead, in the presented process, only one batch / parallel phase is present per half-cycle, reducing the number of waiting time intervals to one. Consequently, with the presented process, significant process time savings can be achieved, and the productivity of the process is improved.

[0093] Alternatively, to the linear gradient operation described above, the presented process may be operated with an isocratic elution or using at least one step elution gradient for the elution of A, A / B, and B during any of the phases. Operating under such conditions requires fewer eluents to be prepared, yet these conditions require more careful preparation of eluents used, since small variations of the eluent compositions have a large impact on the elution time of the compounds to be separated.

[0094] The presented process may use, a modulation of the flow rates during the phases B1 and B2, respectively, to achieve a completion of tasks of the two adsorber sections within the same or similar duration. The use of modulated flow rates allows optimization of the duration of the parallel (batch) phases and better synchronization of the completion of the tasks ofF07460 9.1.2026

[0095] 12

[0096] the two adsorber sections during the parallel (batch) phases.

[0097] The presented process may be used for the purification of biomolecules, of natural or synthetic origin, preferably selected from the group consisting of nucleic acid molecules, including DNA and RNA molecules, proteins, including antibodies, peptides, carbohydrates, lipids as well as combinations and modifications as well as fragments thereof.

[0098] Purifications of such kind are likely to contain overlapping and baseline-separated compounds

[0099] The presented process is particularly useful for the separation of empty and full viral capsids, such as AAV (Adeno-associated viruses) capsids, which tend to show overlaps in elution chromatograms, corresponding to compounds A and B according to earlier descriptions.

[0100] For the same reason, the presented process can be used for the purification of chiral compounds.

[0101] To further improve the separation in particular of viral capsids, the adsorbers may comprise a macroporous stationary phase.

[0102] Further embodiments of the invention are laid down in the dependent claims.

[0103] BRIEF DESCRIPTION OF THE DRAWINGS

[0104] Preferred embodiments of the invention are described in the following with reference to the drawings, which are for the purpose of illustrating the present preferred embodiments of the invention and not for the purpose of limiting the same. In the drawings,

[0105] Fig. 1 shows the tasks of the two adsorbers in the presented process with reference to a single column chromatogram;

[0106] Fig. 2 shows the tasks of the two adsorbers in the startup phase of the presented process;

[0107] Fig. 3 shows the tasks of the two adsorbers in the shutdown phase of the presented process;

[0108] Fig. 4 shows the tasks of the two adsorbers, performed in a half-cycle, in the form of a Gantt chart, compared with the prior art process;

[0109] Fig. 5 shows the process applied to the separation of AAV2 empty I full capsids.

[0110] DESCRIPTION OF PREFERRED EMBODIMENTS

[0111] Fig. 1 shows the tasks of the two adsorbers 1 and 2 in the presented process with reference to a single column chromatogram, which is shown at the bottom, and which schematically indicates linear gradient segments as thick dashed lines, to be used in the process in case of linear gradient operation. The vertical thin dashed lines schematically indicate the sectionF07460 9.1.2026

[0112] 13

[0113] borders.

[0114] In Figs. 1-3 vertical sections designated as phases and separated by thin vertical dashed lines indicate the different functions of the columns being in the respective zone; on the bottom starting on the left hand side the chromatogram is illustrated with the weakly adsorbing compounds A’, baseline-separated from the overlapping (typically both) target compounds A and B, and a further base-line separated compound B’.

[0115] In the first interconnected step, designated with IC1, the upstream adsorber 2 is supplied via inlet (in each case arrow) with eluent, applying a linear gradient, and what is exiting its outlet is in-line adjusted and transferred to the downstream adsorber 1. In that stream the overlapping fraction A / B is contained. Thereby, the starting concentration of the linear gradient supplied to adsorber 2 corresponds to the final concentration of the linear gradient supplied to adsorber 2 in the previous batch phase B-Sll (in case the process has just started-up), or to the final concentration of the linear gradient supplied to adsorber 2 in the previous batch phase B2 (in case the process is operated in the cyclic phase).

[0116] Then follows the first batch step B1, in which the previous upstream adsorber 2 is supplied with eluent again, continuing the linear gradient such that the starting concentration of the linear gradient corresponds to the final concentration of the linear gradient supplied to adsorber 2 in the previous interconnected phase to elute B via the outlet. Alternatively, if the feed mixture does not contain B’, B may be eluted with a step gradient to a higher eluent concentration (stripping) to speed up elution and save process time. The same applies if the feed mixture does contain B’, but no separation of B and B’ is desired, as these are not considered target compounds. To further speed up the process, in case B and B’ both are not considered target compounds, the process may enter regeneration steps directly, omitting the strip step.

[0117] Generally, regeneration steps comprise column cleaning with a suitable cleaning solution, which differs from the modifier used to form the linear gradient. Common cleaning solutions in purification of biopharmaceuticals in include solutions of caustic soda (NaOH) or alcohols. The cleaning steps may also serve for sanitization whereby typically the cleaning solution is held in the adsorber at low or zero flow for 15, 30 or 60 minutes. Regeneration steps also include re-equilibration of the adsorber to remove the cleaning solution. Re-equilibration is typically performed by applying a buffer solution, for example solvent A. Re-equlibration can be important to ensure that the adsorber is prepared to adsorb A / B in the subsequent interconnected phase. In case the intention is to shut down the process, the re-equilibration may be skipped in favor of a storage step, wherein a storage solution can be applied to the adsorber. Common storage solutions are mixture of alcohols, such as ethanol or 2-propanol. For all regeneration, re-equilibration and storage steps it is recommended to applyF07460 9.1.2026

[0118] 14

[0119] at least the volume equivalent to the void volume of the adsorber, to ensure that the adsorber is completely filled with the respective solution.

[0120] In parallel, the previous downstream adsorber 1 is fed with feed F, and via the outlet, as soon as available, A' as well as A are taken out. The removal of A’ is generally achieved through a washing step, which is typically performed by applying a solution with a lower modifier concentration than the starting concentration of the gradient. In case A’ is not present in the feed mixture, the washing step may be omitted, and the process may continue with the elution of A using the linear gradient elution.

[0121] Then adsorbers 1 and 2 switch positions, and the same steps as the above-mentioned first interconnected and first batch step are carried out mutatis mutandis.

[0122] Fig. 2 shows the tasks of the two adsorbers in the startup phase of the presented process with reference to a single column chromatogram (shown at the bottom). Basically, this step is the same as the above-mentioned second batch step, with the difference that the column 1 to become the downstream column in the following first interconnected step is either idle or subjected to some preparation or regeneration process.

[0123] Fig. 3 shows the tasks of the two adsorbers in the shutdown phase of the presented process with reference to a single column chromatogram (shown at the bottom). Basically, this sequence is the same as the first interconnected and first batch step of a running process, with the key difference that in the respective batch step there is no feeding with feed F.

[0124] Fig. 4 shows the tasks of the two adsorbers, performed in a half-cycle, in the form of a Gantt chart, according to WO-A-2006116886 (top, for the case of using two adsorbers only) and the presented process (bottom). The figure indicates in which process phase the adsorbers are operated in interconnected and in batch / parallel mode, respectively. In the top representation the four steps from left to right correspond to the steps 1. -4. of the first cycle in Fig. 28 of WO-A-2006116886. It also shows the two waiting time intervals for WO-A-2006116886 and the single waiting time interval for the presented process. The arrows indicate internal recycling of overlapping parts of the chromatograms from one adsorber to another (A / B and B / C for WO-A-2006116886, and A / B for the presented process).

[0125] The top part of the figure shows that in WO-A-2006116886, the time interval available to complete tasks a and 5fhas to be the same, and secondly, that the time interval available to complete y and the disconnected part of 8ghas to be the same. The batch / parallel phases carrying out the tasks a and y on one adsorber (8f and the disconnected part of 8g, on the other adsorber, respectively) are separated by the interconnected phase, carrying out task p employing both adsorbers.

[0126] Therefore, for each batch / parallel phase a waiting time interval may arise for one of theF07460 9.1.2026

[0127] 15

[0128] adsorbers in order to synchronize the adsorber tasks for the next interconnected phase. In the top part, the Figure shows the presence of the two waiting time intervals, one for adsorber 1 and one for adsorber 2. It is important to note that the two waiting time intervals may also occur for the same adsorber. The longer one of the tasks a and 5fdictates the waiting time for the adsorber with the shorter task. The same applies to the tasks y and 5g. Thus, in WO-A-2006116886, the sum of the duration of the parallel phases is equal to the duration of the sum of the longer one of the two tasks of a or y and the longer one of the two tasks 5f or 5g.

[0129] The lower part of the figure shows the presented process. Here, through absence of an intermediate interconnected phase (p section), the tasks of the parallel phases (corresponding to a and y, and corresponding to 5fand 5g) can be operated sequentially without interruption. Therefore, in the presented process, only one waiting time interval may be required associated with B elution and B’ elution (if present), or associated with F, A’ elution (if present) and A elution, respectively. Fig. 4 shows the case where the sum of the duration of B and B’ elution is larger than the sum of the duration of F, A’ and A elution, introducing a waiting time for adsorber 2 where the latter tasks are operated. Importantly, in this case, the waiting time is not required to be located after completion of A elution, which would interrupt the linear gradient elution started for A elution. Instead, the waiting time may be preferably introduced before carrying out F, after carrying out F or A’ before the linear gradient has started, to avoid pausing of the gradient after A elution potentially leading to peak broadening by diffusion effects.

[0130] If read with respect to WO-A-2006116886, in the presented process, the sum of the duration of the parallel phases is equal to the larger one of the sum of the durations (a + y) and (8f+8g).

[0131] Mathematically, this comparison between the parallel phases of the presented process and WO-A-2006116886 can be expressed by the following inequality:

[0132]

[0133] < With the left-hand side of the equation showing the duration of the parallel phase of the presented process and the right-hand side showing the cumulative duration of the parallel phases of WO-A-2006116886.

[0134] The Figure (and the inequation) show that with the presented process, for each adsorber, a long duration of one element of the parallel phase (for example F in case of adsorber 2), can be compensated for by a short duration of another element of the parallel phase (for example A’ in case of adsorber 2), to achieve an earlier completion of the parallel phase. Such balancing of a high duration of one task (e.g. Load F) by a short duration of anotherF07460 9.1.2026

[0135] 16

[0136] sequential task (e.g. elution of A’) may be achieved by a flow rate increase for adsorber 2. Likewise, for adsorber 1, a high duration of the elution of B can be balanced by a shorter regeneration time (removal of B’), and vice versa.

[0137] In WO-A-2006116886, the modulation of the flow rates can be carried out too, but it will reduce waiting times only if the duration of the longer one of the two parallel individual tasks is abbreviated. Only an abbreviation of the rate-limiting steps results in a shortening of the parallel phases (see example 2).

[0138] In summary, in the presented process, only one batch / parallel phase is present per halfcycle, reducing the number of waiting time intervals to one. The presented process offers great flexibility with regards to flow rate selection to abbreviate the duration of the parallel phase. Consequently, with the presented process, significant process time savings can be achieved, and the productivity of the process is improved.

[0139] Fig. 5 shows the process applied to the separation of AAV2 empty I full capsids according to Example 1 detailed below including the start-up and shut-down phase for column 1 (Fig.

[0140] 5A) and column 2 (Fig. 5B); empty capsids are denoted with “A” and full capsids with “B”; the elution phases for A elution, A / B recycling and B elution, operating the linear elution gradient, are shaded in grey; the superimposed elution peaks are shown in Fig. 5C (for column 1) and in Fig. 5D (for column 2).

[0141] Example 1 : Separation of AAV2 empty / full capsids

[0142] The presented process was tested and operated with two chromatography columns (BioPro MacroSep I EX Q30, 100x4.6 mm), representing adsorbers 1 and 2, on a Contichrom CUBE chromatography system (ChromaCon). In the start-up phase, the second column was loaded with diluted AAV2 capture eluate (conductivity 4.6 mS / cm) before moving into the cyclic phase. This phase started with a linear gradient elution from 15-50 % elution buffer. The first weak fraction (empty particles) was discarded during the start-up phase. Subsequently, the main process was started with a first interconnected state, wherein the linear gradient was continued and a fraction containing a mixture of empty and full AAV2 particles (“overlapping” fraction) was recycled from the second column and inline diluted before being loaded onto the first column in an interconnected state. After the internal recycling of the “overlapping” fraction was completed, the next batch operation was started, to collect the full particle fraction (“B”) from column 2, followed by regeneration of column 2, while column 1 was loaded with feed and the linear gradient was started to remove empty capsids (“A”). After completing the batch phase, the columns changed their relative position, and the interconnected and batch phases were repeated with the columns in exchanged order to complete the cycle. The presented process was operated for 6 cycles in total,F07460 9.1.2026

[0143] 17

[0144] corresponding to 12 column elutions of empty and full AAV capsids. The cyclic phase of the presented process was concluded by a shutdown phase to elute the last AAVs.

[0145] The process was monitored online by measuring UV absorption at 280 nm and 260 nm. Quantitative AAV capsid detection was performed using an ELISA kit (PROGEN AAV2 Titration ELISA 2. OR) according to the manufacturer’s instructions.

[0146] Cryo-TEM and mass photometry were used to confirm the empty / full ratio of the eluates from the presented process.

[0147] Analysis of the eluate pool showed that full AAV2 particles were enriched from 27 % in the load to 61 %. Compared to a batch (single-column) chromatography run based on the same linear gradient, a higher full AAV2 enrichment was achieved using the presented process. According to the analytics it was shown that applying the presented process to AAV2 purification increased the yield by 18 % for a comparable purity of 61 % full particles.

[0148] Table 1 - Operating parameters of the presented process described in Example 1

[0149]

[0150] F07460 9.1.2026

[0151] 18

[0152]

[0153] Example 2: Optimization of Processing time of presented process in comparison to WO-A-2006116886

[0154] The optimization of WO-A-2006116886 and the presented process by flow rate increase, leading to abbreviation of process tasks, was investigated and compared.

[0155] Generally, many process tasks in chromatography are completed as a function of the volume that is flowing through the adsorber. Therefore, with a higher volumetric flow rate, the process task can be completed faster. For example, a step requiring 20 min at a flow rate of 1 mL / min, can be completed in 10 min at a flow rate of 2 mL / min. (In both cases, the volume supplied to the adsorber is 20 mL).

[0156] The following example shows the effect of an increase of the flow rate of task a (WO-A-2006116886) and the equivalent section for elution of B in the presented process. For better comparison, the duration of the p is ignored for WO-A-2006116886, and only the overall duration of the parallel phases is evaluated.

[0157] According to Fig.4 (top part), showing the process WO-A-2006116886, a flow rate increase of task a would result to an earlier completion of task a but this would not result in a shortening of the parallel phase because task 8f, operating in parallel, is the longer one of the two tasks and is rate-limiting. The flow rate increase of task a just increases the waiting time following task a and would not allow an earlier start of any subsequent tasks (p, y). Instead, in the presented process shown in Fig. 4 (lower part), an abbreviation of elution of B on adsorber 1 (corresponding to task a) would enable an earlier start of the subsequent elution of B’ (corresponding to task y) leading to earlier completion of the parallel phase and reduction of the waiting time on adsorber 2.

[0158] The numbers for this comparison are displayed in Table 2 and show that a reduction of the duration of a (elution of B) from 20 min to 10 min, achieved by doubling the flow rate, has no effect on the sum of the durations of the parallel phase in WO-A-2006116886. This is because 5fis rate-limiting and its duration is 20 min. Together with the longer one of the two tasks y (30 min) and the disconnected part of 8g(15 min), the cumulative duration of the parallel phase is 50 min. Instead, for the presented process the duration of the parallel phase corresponds to the larger one of the sum a (elution of B) + y (elution of B’) = 10 minF07460 9.1.2026

[0159] 19

[0160] + 30 min = 40 min and the sum of 5f(feed) + 8g(parallel part, elution of A) = 20 min + 15 min = 35 min. Comparing the two durations of the parallel phases of WO-A-2006116886 and the presented process, time savings through the presented process of 20% can be observed.

[0161] This example illustrates that in the presented process the different tasks of the parallel phase can each contribute individually to an optimization of the parallel phase duration, while in WO-A-2006116886, only a reduction of the duration of the rate-limiting tasks results in such optimization.

[0162] Table 2 - Optimization schedule for WO-A-2006116886 in comparison to the presented process as described in Example 2

[0163]

[0164] LIST OF REFERENCE SIGNS

[0165] 1,2 adsorber, column

[0166] IC1 first interconnected step

[0167] IC2 second interconnected step

[0168] B1 first batch step (parallel, non-interconnected step)F07460 9.1.2026

[0169] 20

[0170] B2 second batch step (parallel, non-interconnected step)

[0171] B1-SD batch step (parallel, non-interconnected step) of the Shutdown Phase B-SU batch step (parallel, non-interconnected step) of Startup Phase tici duration of phase IC1

[0172] tic2 duration of phase IC2

[0173] tsi duration of step B1

[0174] tB2 duration of step B2

[0175] ts-su duration of step B-SU

[0176] taduration of task a

[0177] tYduration of task y

[0178] t5f duration of task 5f

[0179] t5gduration of task 8g(parallel / batch part)

[0180] A’ baseline-separated even more weakly adsorbing compounds A more weakly adsorbing compounds overlapping with B

[0181] F feed mixture containing A, B and optionally A’ and B’

[0182] B more strongly adsorbing compounds overlapping with A

[0183] B’ baseline-separated even more strongly adsorbing compounds

Claims

F07460 9.1.202621CLAIMS1. A cyclic chromatographic purification method for the separation of at least two overlapping compounds, a more weakly adsorbing compound (A) and a more strongly adsorbing compound (B), from a feed mixture (F),the method using at least two chromatographic adsorbers as stationary phase, wherein each adsorber has an adsorber inlet and an adsorber outlet,wherein the method comprises the following steps:a. a first interconnected step (IC1), in which two adsorbers are interconnected defining an upstream absorber (2) and a downstream adsorber (1), in that the upstream adsorber outlet is connected to the downstream adsorber inlet, during a first interconnected purification timespan (tici),wherein the upstream adsorber (2) is loaded via the upstream adsorber inlet with eluent and wherein a fraction containing said more weakly adsorbing compound (A) and said more strongly adsorbing compound (B) in overlap is eluted from the upstream adsorber (2) into the downstream adsorber (1) andwherein the stream exiting the upstream adsorber outlet is adjusted inline before entering the downstream adsorber inlet;b. a first batch step (B1), in which said adsorbers (1,2) are disconnected during a first batch purification timespan (tsi)wherein the following steps are carried out with the upstream adsorber (2) of said first interconnected step (IC1) in given order:

1. elution of said more strongly adsorbing compound (B) and, if present, elution of at least one further strongly adsorbing compound (B’) that is even more strongly adsorbing than said more strongly adsorbing compound (B) and eluting later than said more strongly adsorbing compound (B) and is baseline-separated from said more strongly adsorbing compound (B); 2. optional flushing;3. regeneration;wherein the following steps are carried out with the downstream adsorber (1) of said first interconnected step (IC1) in given order:F07460 9.1.2026221. loading of the feed mixture (F);2. optional washing;3. elution of said more weakly adsorbing compound (A), preceded, if present, by elution of at least one further weakly adsorbing compound (A’), that is even more weakly adsorbing than said more weakly adsorbing compound (A) and that is eluting earlier than said more weakly adsorbing compound (A) and is baseline- separated from said more weakly adsorbing compound (A);and wherein after completing the steps a. and b. the adsorber positions are exchanged such that the previously upstream adsorber (2) becomes the downstream adsorber and the previously downstream adsorber (1) becomes the upstream adsorber, and the steps a. and b. are repeated as second interconnected step (IC2) and second batch step (B2), respectively, with the adsorber sections in this exchanged position such that the previously upstream adsorber section performs the tasks of the downstream adsorber, and the previously downstream adsorber performs the tasks of the upstream adsorber;and wherein the above sequence of first interconnected step (IC1), first batch step (B1), second interconnected step (IC2), second batch step (B2) is carried out at least twice in a row.

2. Method according to claim 1, wherein eluent gradients, preferably linear eluent gradients, are used in at least one or all of the steps, with a gradient in the form of a temporally changing modifier concentration, wherein preferably linear eluent gradients are applied in the interconnected steps (IC1, IC2) and in the elution steps, and preferably not in flushing, regeneration, loading and washing steps.

3. Method according to any of the preceding claims, wherein in interconnected steps (IC1, IC2) of the method the upstream adsorber inlet is loaded via its upstream adsorber inlet with eluent, preferably with eluent base solvent or eluent base solvent mixture with or without gradient, and wherein the stream exiting the upstream adsorber section outlet is adjusted inline before entering the downstream adsorber section inlet with eluent without modifier or with a different modifier concentration than at the inlet of the upstream adsorber section.

4. Method according to any of the preceding claims, wherein in batch substepsF07460 9.1.202623of the method without more strongly adsorbing compound (B) elution, the previously upstream adsorber is cleaned, preferably with eluent with a higher modifier concentration than at the end of the preceding interconnected recycling step or with a different modifier or with a cleaning solution, and / or re-equilibrated.

5. Method according to any of the preceding claims, wherein said further strongly adsorbing compound (B’) is eluted during the regeneration step.

6. Method according to any of the preceding claims, wherein eluent gradients, preferably linear eluent gradients, are used in at least one or all of the steps, with a gradient in the form of a temporally changing modifier concentration, and wherein the modifier is selected from the group consisting of an organic or inorganic solvent or mixture thereof different from an eluent base solvent or eluent base solvent mixture, an electrolyte in such an organic or inorganic solvent, or a mixture thereof, preferably selected from a dissolved salt or a pH, or a combination thereof, whereinpreferably said eluent base solvent or eluent base solvent mixture is water or a mixture of water with at least one organic solvent or water in a mixture with one or more salts and / or organic solvents one or both in a smaller or minor proportion compared with water, and whereinfurther preferably said modifier is an organic solvent or a mixture of water with at least one organic solvent having a higher concentration of said at least one organic solvent than in the eluent base solvent, water or a mixture of water with at least one organic solvent with a different salt or H+concentration than the eluent base solvent.

7. Method according to any of the preceding claims, wherein before carrying out said first interconnected step (IC1), a Start-up Phase (B-Sll) is carried out, in which during a first batch start-up timespan (te-su) said adsorbers (1,2) are disconnected and the adsorber to become the upstream adsorber of the first interconnected step (IC1), is carrying out the following steps in given order:

0. optional equilibration1. loading of the feed mixture (F);2. optional washing;3. elution of said more weakly adsorbing compound (A), preceded, if present, by elution of at least one further weakly adsorbing compound (A’), that is even more weakly adsorbing than said more weakly adsorbingF07460 9.1.202624compound (A) and that is eluting earlier than said more weakly adsorbing compound (A) and is baseline- separated from said more weakly adsorbing compound (A);while the adsorber to become the downstream adsorber section of the first interconnected step (IC1) is either being equilibrated or already equilibrated and inactive or regenerated.

8. Method according to any of the preceding claims, wherein after termination of a final batch second step (B2), a Shut-down sequence is carried out, comprising said first interconnected step (IC1)followed by a Shut-Down batch step (B1-SD) like said first batch step (B1) but omitting loading of the previously downstream adsorber (1) with feed mixture (F);and wherein preferably both adsorbers are subjected to optional final regeneration and / or storage steps.

9. Method according to any of the preceding claims, wherein an isocratic elution or at least one step gradient is carried out for the elution of the more weakly adsorbing compound (A), of the more strongly adsorbing compound (B) or of overlapping more weakly adsorbing compound (A) and more strongly adsorbing compound (B) during at least one or all of the respective phases.

10. Method according to any of the preceding claims, wherein a modulation of the flow rates during the batch phases (B1) and (B2), respectively, is used to achieve a completion of tasks of the two adsorbers within the same or similar duration.

11. Method according to any of the preceding claims, wherein the cyclic chromatographic process uses only two adsorbers.

12. Method according to any of the preceding claims, wherein the adsorbers contain a macroporous stationary phase.

13. Use of a method according to any of the preceding claims for the purification of biomolecules, of natural or synthetic origin, preferably selected from the group consisting of nucleic acid molecules, including DNA and RNA molecules, proteins, including antibodies, peptides, carbohydrates, lipids as well as combinations and modifications asF07460 9.1.202625well as fragments thereof.

14. Use of a method according to any of the preceding claims for the purification of empty and full viral capsids, such as AAV (Adeno-associated virus) capsids.

15. Use of a method according to any of the preceding claims for the purification of chiral compounds.

Images