Method for modulating glycosylation of a biological product

Abstract

The present disclosure provides a method for controlling the glycosylation of a biological product having a predetermined target glycosylation profile.

Description

Technical Field

[0001] This disclosure relates to a method for controlling the glycosylation of a biological product having a predetermined target glycosylation profile. Background Technology

[0002] Protein products (such as antibodies) undergo post-translational modifications during their expression from cells, including the linking of sugar moieties. One such modification is N-linked glycosylation of immunoglobulin G (IgG), which occurs at Asn 297 of the CH2 domain of the mammalian IgG heavy chain. N-linked glycosylation is achieved by initially adding a pre-formed oligosaccharide followed by enzymatic modification of that oligosaccharide to remove or add sugars. This modification includes the removal of mannose and glucose residues and the addition of N-acetylglucosamine (GlcNAc), fucose, galactose, or sialic acid. The final oligosaccharide may or may not include a high mannose content, fucose, galactose, or sialic acid content, depending on the enzymatic reactions that occur in the N-glycosylation pathway.

[0003] Glycosylation can significantly affect the biological activity of proteins. In particular, antibody-dependent cytotoxicity (ADCC), a crucial mechanism in many therapeutic antibodies, depends on the type of glycosylation present on the antibody. Therefore, the production of antibody products with modifiable glycosylation is advantageous for some therapeutic approaches, especially in oncology.

[0004] Furthermore, because the type and extent of glycosylation can affect biological activity, the glycan profile of therapeutic antibodies is an important critical quality attribute (CQA) that must be reported to regulatory agencies and consistently reproduced. However, when the production of therapeutic antibodies is shifted from one process to another (or even between production sites), CQAs (such as glycosylation) may change, potentially requiring adjustments to the transferred process to achieve the previously obtained glycosylation profile for the product.

[0005] Therefore, there is a need to design a system to improve existing glycosylation control methods. Summary of the Invention

[0006] In this embodiment, a method for controlling the glycosylation of a biological product having a predetermined target glycosylation profile is provided, the method comprising: A predictive model for the glycosylation of the biological product is provided, the predictive model including data on the concentrations of hexoses, metal ion cofactors, amino monosaccharides and α-mannosidase inhibitors in a nutrient culture medium, the predictive model being configured to predetermine the target glycosylation profile of the biological product; Cells are grown in the nutrient medium to produce the biological product; The first measurement of the glycosylation profile of the biological product was taken; The first measured glycosylation spectrum is input into the prediction model, which calculates the required concentrations of hexose, metal ion cofactors, amino monosaccharides, and α-mannosidase inhibitors in the nutrient culture medium, so as to adjust the first measured glycosylation spectrum to the predetermined target glycosylation spectrum. A culture medium supplement was prepared based on the concentrations of hexose, metal ion cofactors, amino monosaccharides, and α-mannosidase inhibitors calculated by the prediction model; and The culture medium supplement is added to the nutrient culture medium.

[0007] Other features and aspects of this disclosure are discussed in more detail below. Attached Figure Description

[0008] The full and implementable disclosure of this invention is set forth in more detail in the remainder of the specification (including with reference to the accompanying drawings), wherein:

[0009] Figure 1 This is a schematic diagram of one embodiment of a bioreactor system according to the present disclosure. Detailed Implementation

[0010] Those skilled in the art will understand that this discussion is merely a description of exemplary embodiments and is not intended to limit the broader aspects of this disclosure.

[0011] In the embodiments, this document provides a method for modulating the glycosylation of an antibody product, which is generally intended to (i) modify the biological activity and / or half-life of the antibody, for example, to increase antibody-dependent cytotoxicity (ADCC) or (ii) fine-tune the glycosylation during the process to match previously obtained levels.

[0012] In some embodiments, three main methods are provided for controlling glycosylation: cell line engineering, process parameter modification, and cell culture medium supplementation. All of these methods can affect the activity of the active enzymes during glycosylation. Cell line engineering is limited by the time required to develop and screen new cell lines and can have unintended effects on productivity and culture health. Process parameter modification has proven effective, such as influencing glycosylation by lowering the culture temperature. Changes in pH and osmotic pressure have also been shown to affect glycosylation. The main drawbacks of process modification include unintended effects on metabolic activity, a lack of understanding of the underlying mechanisms, and the need to maintain predetermined operating limits during cGMP production.

[0013] In the embodiments, the culture medium supplementation methods described herein explored a variety of nutrients.

[0014] In some embodiments, this disclosure utilizes a variety of supplements for glycosylation control. "Glycosylation" is a post-translational modification, a process of adding sugar units to molecules, including proteins, such as antibodies (also referred to herein as "antibody products"). There are four types of glycosylation: fucosylation, galactosylation, mannosylation, and sialylation. This disclosure focuses on the regulation of fucosylation, galactosylation, and mannosylation.

[0015] "Fucosylation" is a type of glycosylation that involves adding a fucose unit to a molecule, including proteins such as antibody products. As used herein, "fucosylation-free" refers to the absence of a fucose unit on a particular molecule, such as a particular antibody product. In the preparation of antibody products, the level of fucosylation-free is the percentage of antibody molecules lacking a fucose unit.

[0016] "Galactosylation" is a type of glycosylation that involves adding galactose units to molecules, including proteins such as antibody products. As used herein, "galactosylation" refers to the presence of galactose units on a particular molecule, such as a particular antibody product. In the preparation of antibody products, the level of galactosylation is the percentage of antibody molecules containing galactose units.

[0017] "Mannosylation" is a type of glycosylation that involves adding mannose units to molecules, including proteins such as antibody products. As used herein, "mannosylation" refers to the presence of mannose units at the ends of glycans attached to a particular molecule, such as a particular antibody product. In the preparation of antibody products, the level of mannosylation is the percentage of antibody molecules with mannose units at the ends of the attached glycans.

[0018] Glycosylation can be measured by a variety of methods, including mass spectrometry and high-performance liquid chromatography (HPLC). Because there can be some variation between the methods used, in one embodiment, a time-of-flight liquid chromatography-mass spectrometry (TOF LC / MS) system is used to measure the percentage. The antibody is reduced and subsequently loaded directly onto a liquid chromatography-mass spectrometry (LC / MS) system, such as an Agilent 6230B TOF LC / MS system (without PNGase F digestion), where the reduced antibody has passed through a reversed-phase desalting column on HPLC before being injected into the time-of-flight mass spectrometer.

[0019] Other suitable methods include, for example, Tay and Butler, 2015, J. Biol. Methods 2:19; and Mishra et al., 2020, J. Biotechnology X 5:100015, the disclosure of which is incorporated herein by reference in its entirety, particularly for the disclosed glycosylation determination method. In one of these described methods, the glycans can be removed using PNGase F and dried. They are then labeled using 2-AB labeling and analyzed using hydrophilic interaction liquid chromatography-HPLC (HILIC-HPLC).

[0020] The detected types are then analyzed to determine the percentage of glycosylated antibodies. Measurements are typically repeated to improve accuracy.

[0021] In one embodiment, the percentage of glycosylation is calculated only relative to the type of N-linked glycan (such as N-linked glycans linked to the Fc domain of an antibody). N-linked glycan types include G0, G0F, G0FLys, G0F-GlcNAc, G1F, G2F, G1F + NeuAc, G2F + NeuAc, G2F + 2NeuAc, Man5, Man6, Man7, Man8, and Man9.

[0022] In one embodiment, the percentage of fucosylated de-glycans is measured based on G0 / (the sum of all glycan types).

[0023] In one embodiment, the percentage of galactosylation is measured based on (the sum of G1F, G2F, G1FNeuAc, G2FNeuAc, and G2F2NeuAc) / (the sum of all glycan types).

[0024] In one embodiment, the percentage of mannose sylation is measured based on (the sum of M5, M6, M7, M8, and M9) / (the sum of all glycan types).

[0025] The change in glycosylation of the antibody product prepared according to the method of the present invention is measured using methods known in the art, such as mass spectrometry analysis, including liquid chromatography-mass spectrometry and other methods. Also as mentioned above, since there may be some variations between the methods used, in one embodiment, the percentage is measured using a TOF LC / MS system: the antibody is reduced and subsequently directly loaded onto a liquid chromatography-mass spectrometry (LC-MS) system, such as an Agilent 6230B TOF LC / MS system (without PNGase F digestion), wherein the reduced antibody is passed through a reversed-phase desalting column on HPLC before being injected into the time-of-flight mass spectrometer.

[0026] Other suitable methods include, for example, Tay and Butler, 2015, J. Biol. Methods 2:19; Mishra et al., 2020, J. Biotechnology X 5: 100015. In one of these described methods, the glycans can be removed using PNGase F and dried. They are then labeled with 2-AB and analyzed using hydrophilic interaction liquid chromatography-HPLC (HILIC-HPLC).

[0027] A comparison of the amount of glycosylation from one protein population to another provides a percentage change in glycosylation, and is typically provided relative to an antibody population. That is, the measurement of the percentage change in glycosylation from one antibody population to another is typically calculated based on the amount of antibody produced in the order of about 0.5 g / L or more, rather than on a single antibody. Suitably, the amount of antibody produced using the methods described herein is about 1 g / L or more, suitably 5 g / L or more, or about 10 g / L or more. Thus, in the embodiments described herein where there is at least a 0.5% change in glycosylation, this change is measured relative to a total antibody volume of about 0.5 g / L or more.

[0028] In some embodiments, this disclosure provides a method for producing a biological product. The biological product production process includes a cell population for producing the biological product, and a production culture medium or buffer, which suitably includes necessary reagents and supplements, including a suitable nutrient medium, to support cell proliferation and the desired production of the biological product.

[0029] As used in this article, cell culture refers to a population of cells in a nutrient culture medium.

[0030] In this embodiment, the cells are eukaryotic cells, such as mammalian cells. Mammalian cells can be, for example, human, rodent, or bovine cell lines or cell strains. Examples of such cells, cell lines, or cell strains include, for example, mouse myeloma (NSO) cell lines, Chinese hamster ovary (CHO) cell lines, HT1080, H9, HepG2, MCF7, MDBK Jurkat, NIH3T3, PC12, young hamster kidney cells (BHK), VERO, SP2 / 0, YB2 / 0, Y, C127, L cells, COS (e.g., COS1 and COS7), QC1-3, HEK-293, VERO, PER.C6, HeLA, EBI, EB2, EB3, oncolytic, or hybridoma cell lines. Preferably, the mammalian cells are CHO cell lines. In one embodiment, the cells are CHO cells. In one embodiment, the cells are CHO-K1 cells, CHO-K1 SV cells, DG44 CHO cells, DUXB11 CHO cells, CHOS, CHO GS knockout cells, CHO FUT8 GS knockout cells, CHOZN, or CHO-derived cells. The CHO GS knockout cells (e.g., GSKO cells) are, for example, CHO-K1 SV GS knockout cells. The CHO FUT8 knockout cells are, for example, Potelligent® CHOK1 SV (Lonza Biopharmaceuticals). Eukaryotic cells can also be avian cells, cell lines, or cell strains, such as EBx® cells, EB14, EB24, EB26, EB66, or EBv13.

[0031] In some embodiments, the bioproduct may be an antibody or an antibody product. As used herein, “antibody product” and “antibody” are used interchangeably, wherein an antibody product is the result of an antibody production process. As used herein, the terms “antibody” and “immunoglobulin” are used interchangeably and refer to a polypeptide or polypeptide group comprising at least one binding domain formed by the folding of a polypeptide chain having a three-dimensional binding space with an inner surface shape and charge distribution complementary to the characteristics of the antigenic determinants of the antigen. Antibodies typically exist in tetrameric form, having two pairs of polypeptide chains, each pair having one “light” chain and one “heavy” chain. A variable region of each light / heavy chain pair forms an antibody binding site. Each light chain is connected to the heavy chain by a covalent disulfide bond, the number of which varies between heavy chains of different immunoglobulin isotypes. Each heavy and light chain also has regularly spaced intrachain disulfide bridges. Each heavy chain has a variable domain (VH) at one end, followed by a plurality of constant domains (CH). Each light chain has a variable domain (VL) at one end and a constant domain (CL) at the other end; wherein the constant domain of the light chain is aligned with the first constant domain of the heavy chain, and the variable domain of the light chain is aligned with the variable domain of the heavy chain. Based on the amino acid sequence of the constant region of the light chain, the light chains are classified as λ chains or κ chains.

[0032] Immunoglobulin molecules can be any isotype (e.g., IgG, IgE, IgM, IgD, IgA, and IgY), subisotype (e.g., IgG1, IgG2, IgG3, IgG4, IgA1, and IgA2), or allotype (e.g., Gm, such as Glm1 (f, z, a, or x), G2m(n), G3m (g, b, or c), Am, Em, and Km (l, 2, or 3)). Immunoglobulins include, but are not limited to, monoclonal antibodies (mAh) (including full-length monoclonal antibodies), polyclonal antibodies, multispecific antibodies (e.g., bispecific antibodies) formed from at least two different epitope-binding fragments, human antibodies transplanted with CDRs, humanized antibodies, camelified antibodies, chimeric antibodies, anti-idiotype (anti-id) antibodies, intracellular antibodies, and their desired antigen-binding fragments, including recombinant antibody fragments. Examples of recombinable antibody fragments include, but are not limited to, antibody fragments containing variable heavy and light chain domains, such as single-chain Fv (scFv), single-chain antibodies, Fab fragments, Fab' fragments, and F(ab')2 fragments. Antibody fragments may also include epitope-binding fragments or derivatives of any of the antibodies listed above.

[0033] Antibody products include antibody conjugates, wherein the antibody is conjugated to a small molecule via a linker molecule. In suitable embodiments, the antibody product is a monoclonal antibody (mAb), and more preferably a therapeutic antibody product.

[0034] In some embodiments, the bioproduct is prepared from cell cultures grown in a cell culture medium, and this disclosure controls the fucosylation, mannosylation, and galactosylation properties of the bioproduct by controlling the culture medium supplement. In some embodiments, the bioproduct is an antibody product.

[0035] Production methods

[0036] As described herein, antibody production processes are suitably carried out in bioreactors, which are containers suitable for culturing production cells expressing the target antibody. Since such bioreactors are typically used for production-scale operations or pilot-scale operations prior to scaling up production, although smaller bioreactors, such as the AMBR® 250 system with a volume of 100 mL to 250 mL, may be used to test the process, the bioreactors used in the production process typically have a volume of at least 10 L. Therefore, in exemplary embodiments, the bioreactor may have a volume between about 100 mL and about 50,000 L. Non-limiting examples include 100 mL, 250 mL, 500 mL, and 750 L. mL, 1 L, 2 L, 3 L, 4 L, 5 L, 6 L, 7 L, 8 L, 9 L, 10 L, 15 L, 20 L, 25 L, 30 L, 40 L, 50 L, 60 L, 70 L, 80 L, 90 L, 100 L, 150 L, 200 L, 250 L, 300 L, 350 L, 400 L, 450 L, 500 L, 550 L, 600 L, 650 L, 700 L Volumes of 750 liters, 800 liters, 850 liters, 900 liters, 950 liters, 1000 liters, 1500 liters, 2000 liters, 2500 liters, 3000 liters, 3500 liters, 4000 liters, 4500 liters, 5000 liters, 6000 liters, 7000 liters, 8000 liters, 9000 liters, 10,000 liters, 15,000 liters, 20,000 liters, and / or 50,000 liters. Suitable reactors can be reusable, single-use, disposable, or non-disposable, and can be formed from any suitable material, including metal alloys such as stainless steel (e.g., 316L or any other suitable stainless steel) and Inconel, plastics, and / or glass.

[0037] exist Figure 1In the illustrated embodiment, the bioreactor system according to this disclosure includes a bioreactor 10. The bioreactor 10 includes a hollow vessel or container comprising a bioreactor volume 12 for containing cell cultures in a fluid growth medium, a rotatable shaft 14 connected to a stirrer such as dual impellers 16, 18, a distributor 20, and baffles 22. The rotatable shaft 14 may be connected to a motor 24 for rotating the shaft 14 and the impellers 16, 18. The distributor 20 is in fluid communication with a gas source 48 for supplying gases such as carbon dioxide, oxygen, and / or air to the bioreactor 10. Furthermore, the bioreactor system may include various probes for measuring and monitoring pressure, foam, pH, dissolved oxygen, dissolved carbon dioxide, etc. The bioreactor 10 includes a bottom port 26 connected to a discharge end 28 for continuously or periodically removing material from the bioreactor. Additionally, the bioreactor 10 includes multiple top ports, such as ports 30, 32, and 34. Port 30 is in fluid communication with the first fluid feed 36, port 32 is in fluid communication with the second feed 38, and port 34 is in fluid communication with the third feed 40. The feeds 36, 38, and 40 are used to feed various materials, such as nutrient culture media, into the bioreactor 10.

[0038] like Figure 1 As shown, the bioreactor can be connected to multiple nutrient feeds. In this way, nutrient media containing only a single nutrient can be fed into the bioreactor to better control the concentration of that nutrient during the process. Alternatively, different feed lines can be used to separately feed gases and liquids into the bioreactor.

[0039] In addition to the ports at the top and bottom of the bioreactor 10, the bioreactor may also include ports distributed along the sidewalls. For example, Figure 1 The bioreactor 10 shown includes ports 44 and 46.

[0040] Ports 44 and 46 are connected to a monitoring and control system that maintains optimal concentrations of one or more parameters in the bioreactor 10 for cell propagation or other production of biological products. In the illustrated embodiment, for example, port 44 is associated with a pH sensor 52, while port 46 is associated with a dissolved oxygen sensor 54. The pH sensor 52 and dissolved oxygen sensor 54 communicate with a controller 60. The system of this disclosure can be configured to allow the determination and measurement of various parameters within cell cultures contained within the bioreactor 10. Some measurements, such as pH and dissolved oxygen, can be performed online. However, alternatively, measurements can be performed online or offline. For example, in one embodiment, the bioreactor 10 may be connected to a sampling station. Samples of cell cultures can be fed to the sampling station for various measurements. In another embodiment, samples of cell cultures can be removed from the bioreactor and measured offline.

[0041] culture medium

[0042] As used herein, a nutrient medium refers to any fluid, compound, molecule, or substance that can increase the quality of a biological product, such as any substance that can be used by an organism for survival, growth, or otherwise to add biomass. For example, nutrient feeds may include gases used for respiration or any type of metabolism, such as oxygen or carbon dioxide. Other nutrient media may include carbohydrate sources. Carbohydrate sources include complex sugars and monosaccharides, such as glucose, maltose, fructose, galactose, and mixtures thereof. Nutrient media may also include amino acids. The amino acids may include: glycine, alanine, valine, leucine, isoleucine, methionine, proline, phenylalanine, tryptophan, serine, threonine, asparagine, glutamine, tyrosine, cysteine, lysine, arginine, histidine, aspartic acid, and glutamic acid, their single stereoisomers, and racemic mixtures thereof. The term "amino acid" can also refer to known non-standard amino acids, such as 4-hydroxyproline, ε-N,N,N-trimethyllysine, 3-methylhistidine, 5-hydroxylysine, O-phosphoserine, γ-carboxyglutamic acid, γ-N-acetyllysine, ω-N-methylarginine, N-acetylserine, N,N,N-trimethylalanine, N-formylmethionine, γ-aminobutyric acid, histamine, dopamine, thyroxine, citrulline, ornithine, β-cyanoalanine, homocysteine, diazoserine, and S-adenosylmethionine. In some embodiments, the amino acid is glutamic acid, glutamine, lysine, tyrosine, or valine.

[0043] The nutrient medium may also contain one or more vitamins. Vitamins that may be included in the nutrient medium include B vitamins, such as B12. Other vitamins include vitamin A, vitamin E, riboflavin, thiamine, biotin, and mixtures thereof. The nutrient medium may also contain one or more fatty acids and one or more lipids. For example, the nutrient medium feed may include cholesterol, steroids, and mixtures thereof. The nutrient medium may also provide proteins and peptides to the bioreactor. Proteins and peptides include, for example, albumin, transferrin, fibronectin, fetoglobulin, and mixtures thereof. The growth medium in this disclosure may also include growth factors and growth inhibitors, trace elements, inorganic salts, hydrolysis products, and mixtures thereof. Trace elements that may be included in the growth medium include trace metals. Examples of trace metals include cobalt, nickel, etc.

[0044] Polysaccharide supplement

[0045] The glycosylation profile of bioproducts can be controlled by adding specific supplements to the nutrient culture medium. These supplements may include hexoses, metal ion cofactors, amino monosaccharides, and α-mannosidase inhibitors. These supplements can collectively be used to alter the levels of galactosylation, mannosylation, and afucosylation.

[0046] Uniquely, the amino monosaccharide can alleviate the effects of α-mannosidase inhibitors. This effect is unexpected and has not been previously documented in the literature. Currently, no known mechanism can explain this phenomenon.

[0047] As mentioned above, the glycosylation profile of a bioproduct depends on the relative amounts of the hexose, metal ion cofactors, amino monosaccharides, and α-mannosidase inhibitors present in the nutrient medium. To determine the relative concentrations of these supplements required in the nutrient medium to achieve the desired glycosylation profile, a predictive model was developed to correlate the effect of the concentration of each supplement on the glycosylation profile.

[0048] The predictive model represents a functional relationship that summarizes how the culture medium supplement affects the final glycosylation profile. The model generates an estimate of the expected glycosylation profile based on the provided concentration of the culture medium supplement. Such models can be constructed based on prior reference data, first-principles / mechanistic relationships, or a hybrid simulation strategy combining both. The predictive model can use various multivariate methods to predict the expected glycosylation profile. In one embodiment, the predictive model can be trained using collected reference data, including the culture medium concentration of the aforementioned supplement and the associated percentages of galactosylation, mannosylation, and fucosylation. In the simplest case, multiple linear regression can be applied to these reference data, using the culture medium supplement concentration (and / or its interaction) determined to be statistically significant by analysis of variance as input, to establish a functional relationship between the culture medium supplement concentration and the glycosylation profile. Other methods can also be used to establish the relationship between the culture medium supplement concentration and the glycosylation profile, including but not limited to: partial least squares, neural networks, limiting gradient boosting trees, random forests, support vector machines, etc.

[0049] In one embodiment, the predictive model uses the culture medium concentrations of hexoses, metal ion cofactors, amino monosaccharides, and α-mannosidase inhibitors to predict the percentages of future galactosylation, mannosylation, and fucosylation-free levels. The predictive model can be incorporated into an optimization strategy to determine the required concentration of each supplement in the culture medium, thereby minimizing deviations from a predetermined target glycosylation profile in terms of percentage galactosylation, mannosylation, and fucosylation. The current culture medium concentration of each supplement can be varied based on the optimization output, and the resulting percentages of galactosylation, mannosylation, and fucosylation levels can be recorded. The recorded percentages of galactosylation, mannosylation, and fucosylation levels can be compared to predetermined target levels, and the differences between them can be used to compensate for errors in the predictive model.

[0050] In another embodiment, the predictive model can be used to control the glycan spectrum to a target spectrum via a feedback controller, such as a model predictive controller. In this case, a dynamic predictive model is constructed to predict the glycan spectrum at multiple future time points. Such models can be constructed using a variety of different strategies. In one embodiment, the model can predict the glycan spectrum at future time points using culture medium supplement concentration, controllable culture conditions (pH, temperature, etc.), and glycan spectrum values ​​from previous time points. By incorporating the output prediction of glycan spectrum values ​​and preset variations in culture medium supplement concentration and culture conditions (such as variations determined by a control strategy), the model can be extended to a multi-step forward predictor for predicting future glycan spectrum outputs. In another embodiment, reactor mass balance is used to generate a set of differential equations controlling culture conditions (such as viable cell concentration, metabolite concentration, etc.). In this embodiment, a model can be created using culture medium supplement conditions, controllable culture conditions (pH, temperature, etc.), and recorded / simulated culture conditions to predict unknowns in the differential equations, such as growth rates and cell-specific consumption / production rates of each metabolite. Alternatively, parameterized and optimized hypothetical mechanistic relationships (such as Monod-type equations) can be used to replace the unknown terms in the differential equations to optimally fit previous reference data. In the first form of the hybrid model, given initial culture conditions and preset values ​​for future changes in culture medium supplements and manipulable culture conditions, the predictive model can be used to determine the unknown parameters in the differential equations, thereby simulating the differential equations at future time points. In this simulation, the parameters predicted by the model remain constant between different time points. At a future time point, using the simulated culture conditions, combined with preset changes in culture medium supplements and manipulable culture conditions (such as conditions to be determined by a control strategy), new values ​​for each unknown parameter in the differential equations are determined, and the above process is repeated. A second predictive model can be established that uses culture medium supplement concentrations, manipulable culture conditions, and recorded / simulated culture conditions to predict glycan spectra, thereby enabling continuous prediction of the evolution of glycan spectra over time. In some embodiments, soft sensors, such as those developed based on Raman spectroscopy, can be used instead of the physical measurements of cell culture metabolites required for initialization in simulating this embodiment.

[0051] As described above, in one embodiment, the systems and methods of this disclosure are designed to adjust glycan spectra using a set of manipulated variables. In one embodiment, a model predictive controller can specify the values ​​of manipulated variables within a control time domain based on knowledge of the desired glycan spectra and historical values ​​of the recorded manipulated variables and glycan spectra. The model predictive controller can utilize a dynamic model developed based on historical process data to determine the values ​​of the manipulated variables so that the glycan spectra reach the desired values ​​in the future. A predictive model / simulation in the prediction time domain generates glycan spectra predictions in a multi-step manner based on the sequence of manipulated variable values ​​within the control time domain. The optimal values ​​of each manipulated variable are determined within the control time domain to minimize an objective function relating to the deviation between the model output prediction and the desired trajectory within the prediction time domain. Once the optimal sequence of manipulated variables is determined, in one embodiment, only the first value in the sequence may be used in the bioreactor. In this way, at the next sampling time, the glycan spectra are measured and the above process is repeated. Since the recorded glycan spectra are used instead of the predicted values ​​in each subsequent optimization cycle, the impact of prediction errors that may accumulate in multi-step prediction / simulation on the controller implementation is limited.

[0052] In one embodiment, the design of a model predictive controller may include specifying multiple design parameters to compute an objective function that is optimized during controller operation. For example, in one embodiment, the objective function may be expressed as: in: • P represents the number of days in the prediction time domain. • The number of glycan output results • The predicted value of glycan j obtained from the prediction model • The value of glycan j for the desired reference trajectory • The weights applied to each glycan (j) at each time (i) in the prediction domain for the difference between the predicted output and the reference trajectory. • To control the number of variables • To control the value of variable j at a specific time. • The weight applied to the difference between the values ​​of the control variable j at the i-th prediction time domain. • The scaling factor for the j-th control variable is used to handle differences in dimensions among the control variables.

[0053] In one embodiment, the coefficient on the right side of the above equation can be set to zero, resulting in the following simplified equation. Where: P is the number of days in the prediction time domain; The number of glycan output results. The predicted value of glycan j obtained from the prediction model; Let j be the value of the glycan for the desired reference trajectory; The weights applied to each glycan at each time step in the prediction domain to account for the difference between the predicted output and the reference trajectory.

[0054] The objective function penalizes the difference between the predicted output and the reference trajectory value. If concerns arise regarding the prediction accuracy of the model over multiple steps in the distant future, different weights can be applied at different times within the prediction time domain. The optimal values ​​of the manipulated variables in the control time domain can be obtained by minimizing the objective function under boundary and rate constraints.

[0055] Glycosylation control

[0056] In exemplary embodiments, the amount of glycosylation increase resulting from the methods described herein is at least 0.5% increase, or in other embodiments, at least 0.6%, at least 0.7%, at least 0.8%, at least 0.9%, at least 1%, at least 1.1%, at least 1.2%, at least 1.3%, at least 1.4%, at least 1.5%, at least 1.6%, at least 1.7%, at least 1.8%, at least 1.9%, at least 2.0%, at least 2.1%, at least 2.2%, at least 2.3%, at least 2.4%, at least 2.5%, at least 2.6%, at least 2.7%, at least 2.8%, at least 2.9%, at least 3.0%, or 0.5% to about 2.0%, about 0.5% to about 1.5%, or about 0.5% to about 1.0%.

[0057] In another embodiment, the method is used to control the level of glycosylation to match a predetermined level (target value). Thus, in one embodiment, a method is provided for matching the glycosylation of a recombinant antibody to a target glycosylation percentage of a previously obtained same antibody, the method comprising culturing cells expressing the antibody in a bioreactor; and, during antibody production, controlling the addition of hexose, metal ion cofactors, amino monosaccharides, and α-mannosidase inhibitors to the bioreactor to obtain an expressed antibody having the target glycosylation percentage.

[0058] Typically, the glycosylation level is controlled within + / - 0.05%, + / - 0.10%, + / - 0.15%, + / - 0.20%, + / - 0.25%, + / - 0.30%, + / - 0.35%, + / - 0.40%, + / - 0.45%, + / - 0.50%, + / - 0.75%, + / - 1%, + / - 1.50%, + / - 1.75%, + / - 2%, + / - 5%, or + / - 8% of the desired target value (wherein, when the target value is a range, the deviation is relative to the midpoint of that range). In some embodiments, the glycosylation level is controlled within + / - 0.25%. In some embodiments, the glycosylation level is controlled within + / - 0.5%.

[0059] In some embodiments, this disclosure provides a process or method for controlling the glycosylation of a biological product having a predetermined target glycosylation profile. In some embodiments, the method includes: providing a predictive model for the glycosylation of the biological product, dependent on the concentrations of hexoses, metal ion cofactors, amino monosaccharides, and α-mannosidase inhibitors in a nutrient culture medium, the predictive model including information related to the predetermined target glycosylation profile of the biological product; Cells are grown in the nutrient culture medium, and the cells are able to produce the biological product. The first measurement of the glycosylation profile of the biological product was taken; The glycosylation profile measured in the first step is input into the prediction model. The prediction model calculates the required concentrations of hexose, metal ion cofactors, amino monosaccharides, and α-mannosidase inhibitors in the nutrient culture medium in order to change the first measured glycosylation spectrum in a direction toward the predetermined target glycosylation spectrum. A culture medium supplement was prepared based on the concentrations of hexose, metal ion cofactors, amino monosaccharides, and α-mannosidase inhibitors calculated by the prediction model; and The culture medium supplement is added to the nutrient culture medium.

[0060] In some embodiments, the predetermined target glycosylation profile is a percentage or range of glycosylation in the product. In some embodiments, the predetermined target glycosylation profile is pre-input and stored in the prediction model based on product glycosylation measurements from a smaller-scale production bioreactor, experimental bioreactor, or other equipment, wherein the product has high quality in terms of its glycosylation, such as having a low fucosylation level. In some embodiments, the predetermined target glycosylation profile is obtained by measuring the high-quality product obtained from the smaller-scale production bioreactor or experimental bioreactor using mass spectrometry and HPLC or any other suitable method.

[0061] Cells for the production of biological products are cultured in a nutrient medium in a bioreactor. The cells may be mammalian cells as described herein, and the bioreactor may be, for example, […]. Figure 1 The diagram shows a large-scale production bioreactor, and the cells may contain exogenous genes encoding the bioproduct. The bioproduct may be a monoclonal antibody. During culture, the bioproduct is expressed and preferably secreted into a cell culture medium. The glycosylation profile of the product in the culture medium is measured and input into the prediction model. The measured glycosylation profile may deviate from the predetermined target glycosylation profile. In some embodiments, the prediction model compares the measured glycosylation profile with the predetermined target glycosylation profile. When the measured glycosylation profile exceeds the range of the predetermined target glycosylation profile, or deviates from the predetermined target glycosylation profile by a margin greater than a threshold (such as 0.25%), the prediction model then predicts the concentrations of hexoses (e.g., galactose), metal ion cofactors (e.g., manganese), amino monosaccharides (e.g., glucosamine), and α-mannosidase inhibitors (e.g., kifunensine), which can alter the glycosylation profile of the product toward the predetermined glycosylation profile. These four components are also referred to as critical process parameters (CPPs) because their concentrations are critical to the quality of the produced product. Following the prediction, the prediction model then instructs or controls the bioreactor to prepare a culture medium supplement containing CPP based on the predicted desired CPP concentration, and adds the culture medium supplement to the culture medium.

[0062] In some embodiments, the hexose is galactose, the metal ion cofactor is manganese, the amino monosaccharide is glucosamine, and the α-mannosidase inhibitor is kifunensine.

[0063] In some embodiments, the following steps are repeated during cultivation: measuring the glycosylation profile of the product, predicting the CPP concentration required to change the glycosylation profile to the predetermined glycosylation profile, preparing a culture medium supplement containing the predicted CPP concentration, and adding the prepared culture medium supplement to the culture medium. In some embodiments, after the step of measuring the glycosylation profile and before the step of predicting the required CPP concentration, the method further includes a step of comparing the measured glycosylation profile with the predetermined glycosylation profile. If the measured glycosylation profile deviates from the predetermined glycosylation profile, the process continues to the prediction step. If the measured glycosylation profile does not deviate from the predetermined glycosylation profile, the predictor stops the process and waits for the input of the next measured glycosylation profile.

[0064] Example 1 – Modulation of fucosylation, galactosylation, and mannosylation by adding galactose + manganese, glucosamine, and kifunusenine

[0065] Controllability Research

[0066] CHO GS-KO cells were cultured in a chemically defined medium supplemented with different concentrations of various chemicals to test the effect of those chemicals on the N-linked glycan profile of the products produced by the culture. The product produced by this cell line was a model IgG antibody.

[0067] Experimental procedure:

[0068] 1. Preparation of culture media with supplemental chemical substances

[0069] 2. Culture the cells in the medium derived from (1).

[0070] 3. Harvest the mAbs produced in the culture and measure the glycan profiles of the purified, reduced, and desalted mAb products.

[0071] 4. Determine the statistical effect of chemical supplements on the quality of polysaccharide products.

[0072] Preparation of culture media with supplemental chemicals:

[0073] The chemicals added to the culture medium include: - None (control) -galactose + manganese -glucosamine -Kifunensine

[0074] Concentrated stock solutions of galactose, manganese, glucosamine, and kifunensine were prepared and added to the culture medium. Specific volumes of the stock solutions were added to the culture medium to generate the conditions shown in Table 1. As part of performing a full factorial DoE, the conditions were mixed in various combinations to generate the model described herein.

[0075] Table 1:

[0076] Cell culture:

[0077] GS-KO cells were prepared at 5 x 10⁻⁶ 5 Inoculate the cultures at a density of 10 cells / mL into ventilated shake flasks containing the culture medium mixture shown in Table 1. Incubate the cultures in a temperature-, CO2-, and humidity-controlled incubator for five days. Take samples immediately after inoculation and on harvest day to monitor the health of the cultures.

[0078] Product harvesting and polysaccharide measurement:

[0079] mAb products were harvested from shake-flask cultures on day 5 and purified by protein A capture. Purified mAb was prepared for glycan analysis by reducing the mAb to separate the heavy and light chains. The reduced mAb was injected into LC-MS, where it was desalted by passing through a reverse-phase desalting column on the LC, and then injected into a time-of-flight mass spectrometer (TOF MS) (Agilent 6230B). Each sample was injected three times for technical reproducibility.

[0080] Glycan data analysis and statistical analysis:

[0081] The obtained LC-MS data were processed using Protein Metrics software, and the relative abundance of each glycan species was reported. The glycan species measured during the process included: G0, G0F, G0FLys, G0F-GlcNAc, G1F, G2F, G1F + NeuAc, G2F + NeuAc, G2F + 2NeuAc, Man5, Man6, Man7, Man8, and Man9. The percentage of glycosylation was determined using the relative abundance of these glycan species.

[0082] The supplementation conditions tested in the full factorial DoE were categorically coded for analysis of variance (ANOVA) to identify statistically significant items. The `anova_lm` function from the `statsmodels` package in Python was used to perform ANOVA between the categorically coded conditions and the recorded percentage changes in galactosylation, mannosylation, and fucosylation to determine whether there was a significant difference in mAbs produced in cultures with supplemented medium compared to those produced in cultures without supplemented medium (control). This analysis was also used to determine the presence of co-supplementation-related combinatorial effects of the medium supplements tested in these experiments. Variables with statistically significant relationships to the output were determined using p-values ​​calculated through ANOVA, with a p-value threshold of 0.1. To determine system controllability, coefficients of items with p-values ​​less than the threshold were retained, while coefficients of items with p-values ​​greater than the threshold were set to zero. The resulting coefficients were arranged into a process gain matrix (K), which correlated changes in input factors with changes in the glycan profile using the following formula: in and These represent changes in input factors and changes in the output glycan spectrum, respectively. System controllability is determined by establishing the rank of the process gain matrix; a full-rank matrix indicates system controllability. Table 2 lists the statistically significant coefficients in this analysis.

[0083] Results / Discussion:

[0084] Adding the tested supplement to the culture medium caused changes in the baseline glycan profile (see Table 2).

[0085] Table 2

[0086] The quantities in the results table represent the differences between the baseline glycan profile and that produced without supplemented medium. Furthermore, for conditions treated with more than one supplement, the results represent the differences between the baseline glycan profile and the sum of the effects of each individual condition.

[0087] For example, the percentage of galactosylation under the galactose + manganese + kifunensine condition (-6.12%) represents the difference between this result and the sum of the baseline galactosylation (45.09%), the galactosylation under the galactose + manganese condition (+21.25%), and the galactosylation under the kifunensine condition (-36.27%). This means that although the individual effects of galactose, manganese, and kifunensine suggest a result of approximately 30.07% galactosylation, the actual result is 6.12% lower, or approximately 23.95%.

[0088] Example 2 – Further regulation of fucosylation, galactosylation, and mannosylation by adding galactose + manganese, glucosamine, and kifunusensine

[0089] Controllability Research

[0090] Further experiments were conducted using the methods described in the “Controllability Studies” section above. These experiments included other cell clones cultured and supplemented using the same methods, including the GS Xceed® CHOK1SV GS-KO® cell line. The results of this study provide information about the impact of clones and products on the efficacy of the supplement described herein. The insertion method used to construct these clones differed from that used for the clones previously used for data collection. One clone produced a product different from that produced by the clones previously used for data collection, while the other clone produced the same product. The completion of these studies demonstrates the broad applicability of the model system described herein.

[0091] Validation study

[0092] Concentrated stock solutions of galactose, manganese, glucosamine, and kifunensine were prepared and added to the culture medium. Specific volumes of the stock solutions were added to the culture medium to generate the conditions in Table 3. The conditions in Table 3 were determined by estimating the concentrations of the supplements required to achieve the target glycan profile using the model generated in this paper.

[0093] Table 3:

[0094] Exemplary Example:

[0095] Example 1 is a method for controlling the glycosylation of a biological product having a predetermined target glycosylation profile, the method comprising: A predictive model for the glycosylation of the biological product is provided, the predictive model including data on the concentrations of hexoses, metal ion cofactors, amino monosaccharides and α-mannosidase inhibitors in a nutrient culture medium, the predictive model being configured to predetermine the target glycosylation profile of the biological product; Cells are grown in the nutrient medium to produce the biological product; The first measurement of the glycosylation profile of the biological product was taken; The first measured glycosylation spectrum is input into the prediction model, which calculates the required concentrations of hexose, metal ion cofactors, amino monosaccharides, and α-mannosidase inhibitors in the nutrient culture medium, so as to adjust the first measured glycosylation spectrum to the predetermined target glycosylation spectrum; and A culture medium supplement was prepared based on the concentrations of hexose, metal ion cofactors, amino monosaccharides, and α-mannosidase inhibitors calculated by the prediction model; and The culture medium supplement is added to the nutrient culture medium.

[0096] Example 2 includes the method described in Example 1, wherein the concentration of the amino monosaccharide is used to reduce mannosylation caused by the α-mannosidase inhibitor.

[0097] Example 3 includes the method described in any of the foregoing examples, further comprising the steps of: sampling the biological product in the nutrient medium after adding the culture medium supplement, measuring the second glycosylation profile of the biological product, inputting the second glycosylation profile into the prediction model to calculate the second culture medium supplement, and adding the second culture medium supplement to the nutrient medium.

[0098] Example 4 includes the method described in any of the preceding examples, wherein the α-mannosidase inhibitor is kifunensine.

[0099] Example 5 includes the method described in any of the preceding examples, wherein the hexose is a precursor of UDP-Gal.

[0100] Example 6 includes the method described in any of the preceding examples, wherein the precursor of UDP-Gal is galactose.

[0101] Example 7 includes the method described in any of the preceding examples, wherein the metal ion cofactor is a cofactor of galactosyltransferase.

[0102] Example 8 includes the method described in any of the preceding examples, wherein the cofactor of the galactosyltransferase is manganese.

[0103] Example 9 includes the method described in any of the preceding examples, wherein the amino monosaccharide is a competitor of UTP.

[0104] Example 10 includes the method described in any of the preceding examples, wherein the competitor of UTP is glucosamine.

[0105] Example 11 includes the method described in any of the preceding examples, wherein the competitor of the UTP is N-acetylglucosamine.

[0106] Example 12 includes the method described in any of the preceding examples, wherein the prediction model provides outputs that regulate the fucosylation, galactosylation, and mannosylation of the bioproduct.

[0107] Example 13 includes the methods described in any of the preceding examples, wherein the biological product is an antibody.

[0108] Example 14 includes the method described in any of the preceding examples, wherein the cell is a mammalian cell, more preferably a Chinese hamster ovary cell.

[0109] Example 15 includes the method described in any of the foregoing examples, further comprising, after the step of inputting the first measured glycosylation profile into the prediction model:

[0110] The first measured glycosylation spectrum is compared with the predetermined target glycosylation spectrum.

[0111] When the increase in the first measured glycosylation spectrum compared to the predetermined target glycosylation spectrum exceeds a threshold, a concentration calculation step is performed; and when the increase in the first measured glycosylation spectrum compared to the predetermined target glycosylation spectrum is less than the threshold, a second measurement of the glycosylation spectrum is awaited.

[0112] Example 16 includes the method described in Example 15, wherein the threshold is 1% to 5%.

[0113] Example 17 includes the method described in Example 15, wherein the threshold is 2%.

[0114] Example 18 includes a method for controlling the glycosylation of a biological product having a predetermined target glycosylation profile, the method comprising: A predictive model for the glycosylation of the biological product is provided, the predictive model including data selected from the concentrations of hexoses in nutrient culture media, cofactors of enzymes in the Leloir pathway, molecules that can reduce UDP-Gal levels, and α-mannosidase inhibitors, the predictive model being configured to predetermine the target glycosylation profile of the biological product. Cells are grown in the nutrient medium to produce the biological product; The first measurement of the glycosylation profile of the biological product was taken; The first measured glycosylation profile is input into the prediction model, which calculates the required concentrations of hexose, cofactors of enzymes in the Leloir pathway, molecules that can reduce UDP-Gal levels, and α-mannosidase inhibitors in the nutrient medium, so as to adjust the first measured glycosylation profile to the predetermined target glycosylation profile; and A culture medium supplement was prepared based on the concentrations of hexose, cofactors of enzymes in the Leloir pathway, molecules capable of reducing UDP-Gal levels, and α-mannosidase inhibitors calculated by the prediction model; and The culture medium supplement is added to the nutrient culture medium.

[0115] Example 19 comprises the method of any one of the preceding claims, wherein the hexose is a precursor of UDP-Gal; the cofactor of the enzyme in the Leloir pathway is a cofactor of galactosyltransferase; the molecule capable of reducing UDP-Gal levels is a competitor of UTP; and the α-mannosidase inhibitor is kifunensine.

[0116] Example 20 includes the method of any one of the preceding claims, wherein the precursor of UDP-Gal is galactose; the cofactor of the galactosyltransferase is manganese; and the competitor of the UTP is glucosamine or N-acetylglucosamine.

Claims

1. A method for controlling the glycosylation of a biological product having a predetermined target glycosylation profile, the method comprising: A predictive model for the glycosylation of the biological product is provided, the predictive model including data on the concentrations of hexoses, metal ion cofactors, amino monosaccharides and α-mannosidase inhibitors in a nutrient culture medium, the predictive model being configured to predetermine the target glycosylation profile of the biological product; Cells are grown in the nutrient medium to produce the biological product; The first measurement of the glycosylation profile of the biological product was taken; The first measured glycosylation spectrum is input into the prediction model, which calculates the required concentrations of hexose, metal ion cofactors, amino monosaccharides, and α-mannosidase inhibitors in the nutrient culture medium, so as to adjust the first measured glycosylation spectrum to the predetermined target glycosylation spectrum. A culture medium supplement was prepared based on the concentrations of hexose, metal ion cofactors, amino monosaccharides, and α-mannosidase inhibitors calculated by the prediction model; and The culture medium supplement is added to the nutrient culture medium.

2. The method according to claim 1, wherein the concentration of the amino monosaccharide is used to reduce mannosylation caused by the α-mannosidase inhibitor.

3. The method according to any one of the preceding claims further includes the following step: After adding the culture medium supplement, the biological product in the nutrient culture medium is sampled, the second glycosylation spectrum of the biological product is measured, the second glycosylation spectrum is input into the prediction model to calculate the second culture medium supplement, and the second culture medium supplement is added to the nutrient culture medium.

4. The method according to any one of the preceding claims, wherein the α-mannosidase inhibitor is kifunensine.

5. The method according to any one of the preceding claims, wherein the hexose is a precursor of UDP-Gal.

6. The method according to any one of the preceding claims, wherein the precursor of UDP-Gal is galactose.

7. The method according to any one of the preceding claims, wherein the metal ion cofactor is a cofactor of galactosyltransferase.

8. The method according to any one of the preceding claims, wherein the cofactor of the galactosyltransferase is manganese.

9. The method according to any one of the preceding claims, wherein the amino monosaccharide is a competitor of UTP.

10. The method according to any one of the preceding claims, wherein the competitor of the UTP is glucosamine.

11. The method according to any one of the preceding claims, wherein the competitor of the UTP is N-acetylglucosamine.

12. The method according to any one of the preceding claims, wherein the prediction model provides outputs that regulate the fucosylation, galactosylation, and / or mannosylation of the biological product.

13. The method according to any one of the preceding claims, wherein the biological product is an antibody.

14. The method according to any one of the preceding claims, wherein the cell is a mammalian cell, more preferably a Chinese hamster ovary cell.

15. The method according to any one of the preceding claims, after the step of inputting the first measured glycosylation spectrum into the prediction model, further comprising: The first measured glycosylation spectrum is compared with the predetermined target glycosylation spectrum. When the increase in the first measured glycosylation spectrum compared to the predetermined target glycosylation spectrum exceeds a threshold, a concentration calculation step is performed; and when the increase in the first measured glycosylation spectrum compared to the predetermined target glycosylation spectrum is less than the threshold, a second measurement of the glycosylation spectrum is awaited.

16. The method of claim 15, wherein the threshold is 1% to 5%.

17. The method of claim 15, wherein the threshold is 2%.

18. A method for controlling the glycosylation of a biological product having a predetermined target glycosylation profile, the method comprising: A predictive model for the glycosylation of the biological product is provided, the predictive model including data selected from the concentrations of hexoses in nutrient culture media, cofactors of enzymes in the Leloir pathway, molecules that can reduce UDP-Gal levels, and α-mannosidase inhibitors, the predictive model being configured to predetermine the target glycosylation profile of the biological product. Cells are grown in the nutrient medium to produce the biological product; The first measurement of the glycosylation profile of the biological product was taken; The first measured glycosylation spectrum is input into the prediction model, which calculates the concentrations of hexose, cofactors of enzymes in the Leloir pathway, molecules that can reduce UDP-Gal levels, and α-mannosidase inhibitors required in the nutrient culture medium, so as to adjust the first measured glycosylation spectrum to the predetermined target glycosylation spectrum. as well as A culture medium supplement was prepared based on the concentrations of hexose, cofactors of enzymes in the Leloir pathway, molecules capable of reducing UDP-Gal levels, and α-mannosidase inhibitors calculated by the prediction model; and The culture medium supplement is added to the nutrient culture medium.

19. The method according to any one of the preceding claims, wherein the hexose is a precursor of UDP-Gal; the cofactor of the enzyme in the Leloir pathway is a cofactor of galactosyltransferase; the molecule capable of reducing UDP-Gal levels is a competitor of UTP; and the α-mannosidase inhibitor is kifunensine.

20. The method according to any one of the preceding claims, wherein the precursor of UDP-Gal is galactose; the cofactor of the galactosyltransferase is manganese; and the competitor of the UTP is glucosamine or N-acetylglucosamine.

Images