Adjusting control of cell culture in production scale vessels relative to starting media

Abstract

Computer-implemented methods and systems for adjusting control of cell culture in a production scale vessel relative to a starting media are provided. The methods include providing a plurality of production scale process trajectories, each production scale process trajectory resulting from a successfully controlled cell culture. The methods also include receiving a media set for a cell culture. The methods further include sampling a first media from the media set that can be used in the production scale vessel. In addition, the methods include starting a seed train using the first media to achieve inoculation of a production scale vessel. The methods also include providing a plurality of micro scale vessels in a process control device, wherein the production scale is larger than the micro scale. The methods further include sampling a second media from the media set for the micro scale vessels, wherein each micro scale vessel receives a representative portion of the media set. In addition, the methods include introducing cells from the seed train into the micro scale vessels to start a cell culture in each micro scale vessel.

Description

[0001] This disclosure relates to the control of chemical, pharmaceutical, biological, and / or biological processes. More specifically, the disclosed aspects relate to the control of cell culture in production-scale vessels relative to the adjustment of the starting culture medium.

[0002] Cell culture media can be a significant source of variability in biopharmaceutical manufacturing processes and can adversely affect cell growth, viability, and specific productivity, or alter the quality profile of the resulting products (e.g., therapeutic proteins). Therefore, the consistency of cell culture media is increasingly important. Furthermore, the culture medium in mammalian cell culture is a critical raw material (CRM), which can influence some key quality attributes (CQA) of the products generated during cell culture or those produced using cell culture.

[0003] Given the uncertain nature of animal serum and the batch-to-batch variability leading to differences in cell growth and productivity, serum-supplemented media can be particularly unpredictable. Furthermore, contaminants can be introduced into the final product, potentially raising ethical concerns, especially in the case of fetal bovine serum (FBS). Therefore, serum-free media may be preferable. Serum-free media can consist of cell growth factors, including hydrolysates, amino acids, vitamins, and inorganic salts. Serum-free media may also contain products of uncertain animal origin, such as serum albumin, hormones, carrier proteins, and adhesion factors, which may contain contaminants.

[0004] Typically, culture media may contain one or more of the following: amino acids (e.g., about 20 different amino acids), purines (e.g., hypoxanthine), pyrimidines (e.g., thymidine), choline, inositol, thiamine, folic acid, biotin, calcium, nicotinamide, pyridoxine, riboflavin, thymidine, cyanocobalamin, pyruvate, lipoic acid, magnesium, glucose, sodium, potassium, iron, copper, zinc, and sodium bicarbonate. In some cases, the culture medium may contain serum from mammals. In other cases, the culture medium may be serum-free. The culture medium may contain trace metals, mammalian growth hormones, and / or mammalian growth factors.

[0005] In some cases, the culture medium can contain more than 100 components. In other cases, the medium may be limited to fewer than 10 components, such as inorganic salts, carbon sources, and water. In the latter case, cell cultures can be maintained in a basal medium without further supplementation.

[0006] Serum-free media can be inconsistent due to less defined components, such as peptones and hydrolysates (which can be added to mimic growth achieved with serum-containing media). Hydrolysates can be complex mixtures containing other low-molar amounts of oligopeptides, amino acids, iron salts, lipids, vitamins, and trace elements, which have proven to be suitable substitutes for recombinant insulin and serum components in media.

[0007] The culture medium can be chemically defined. Therefore, all components are identified as having known concentrations. Consequently, a chemically defined culture medium can be completely free of animal-derived components, such as bovine serum albumin, human serum albumin, or fetal bovine serum.

[0008] Therefore, chemically defined culture media can be advantageous for controlling cell culture to produce target products. However, chemically defined culture media can exhibit significant variations in some components, particularly trace components such as vitamins, amino acids, and metals such as zinc, iron, and copper. Furthermore, variations may exist in impurities in the chemical compounds, such as trace element residues.

[0009] Providing a culture medium limited to chemically defined components can be challenging. Therefore, chemically defined culture media may be supplemented with recombinant forms of albumin and / or growth factors. Chemically defined culture media may include mixtures of nutrients and / or basal preparations. Furthermore, chemically defined culture media may be supplemented with less defined components, such as hydrolysates. Chemically defined culture media may also include the aforementioned elements to meet specific requirements. Such supplemented culture media can still be serum-free.

[0010] Cell culture can be used to produce recombinant proteins with clinical applications. Therefore, mammalian cell culture may be suitable. However, other cell cultures can also be used.

[0011] Culture media can be provided in liquid or powder form.

[0012] For a variety of reasons, powdered culture media may be preferable (e.g., more preferable than liquid media). Specifically, high transportation costs and limited shelf life can limit the applicability of liquid culture medium formulations. For example, some culture medium components (e.g., amino acids) in liquid form become chemically unstable after a few days or weeks. These components may be more stable in powder form (i.e., they can last longer). Furthermore, powder can offer flexibility in scaling up or down production processes without requiring significant investment to store large quantities of unused liquid.

[0013] Powdered culture media can also introduce problems. Specifically, powders can have natural variations based on factors such as weighing errors, uniform mixing and preparation, and varying levels of humidity in the environment.

[0014] In light of the above, inconsistencies or variations in the culture medium can be a significant (or even the most significant) source of process variations and project rejections, leading to quality control failures.

[0015] To detect and adapt to changes in culture media, analytical techniques can be applied to determine the composition of the media. Analytical techniques can include separation methods and spectroscopy. Spectroscopic techniques can include near-infrared (NIR), mid-infrared (MIR), Raman, fluorescence, and nuclear magnetic resonance (NMR). Furthermore, spectroscopic techniques can be used to quantify trace elements; these techniques can include atomic emission spectroscopy (AES), laser-induced breakdown spectroscopy (LIBS), and X-ray fluorescence (XRF). Additionally, mass spectrometry techniques, such as tandem mass spectrometry, can be used. Chromatographic methods can also be applied. Other techniques may also be used.

[0016] Analytical techniques can have drawbacks. For example, chromatography combined with spectroscopic detection methods can be particularly labor-intensive because it typically requires an initial pre-concentration extraction step, followed by pre-column derivatization of monosaccharides.

[0017] Furthermore, due to limited sensitivity, spectroscopic analysis of materials may only reveal partial information about a sample. While mass spectrometry can reveal additional information, these techniques are also limited by the completeness of the reference library used in the identification process. Many spectroscopic techniques, such as infrared or Raman, may have difficulty with low concentrations of components. More sensitive techniques, such as surface-enhanced Raman spectroscopy (SERS), may be difficult to achieve reproducibly or may not be developed for commercial validation. Some techniques may also be affected by the turbidity of the culture medium.

[0018] In addition, some analytical techniques, such as nuclear magnetic resonance spectroscopy, may have high capital and operating costs and significant space requirements.

[0019] Therefore, complete analysis of the culture medium can be challenging.

[0020] Furthermore, providing a complete analytical profile of the culture medium may conflict with the strategies of the culture medium manufacturer. Specifically, many culture medium manufacturers may choose to keep the exact composition of their culture medium a trade secret and may not wish to disclose it.

[0021] Furthermore, the composition of the culture medium can change over time. Specifically, during loading and / or transportation, the medium may be affected by decomposition through aging and stratification (unintended separation). Stratification is particularly relevant to powdered culture media.

[0022] Therefore, managing changes in the culture medium and adjusting it relative to the starting medium for control of cell culture at production scale can be a problem.

[0023] According to one aspect, a computer-implemented method is provided for adjusting control of cell culture in production-scale containers relative to a starting culture medium. The method includes providing a plurality of production-scale process trajectories, each production-scale process trajectory derived from a successfully controlled cell culture. The method also includes receiving a collection of culture media for cell culture. The method further includes sampling a first culture medium from the collection of culture media potentially used in production-scale containers. Furthermore, the method includes initiating seed training using the first culture medium to achieve inoculation of the production-scale containers. The method also includes providing a plurality of micro-scale containers in a process control device, wherein the production scale is larger than the micro-scale. The method further includes sampling a second culture medium from the collection of culture media for micro-scale containers, wherein each micro-scale container receives a representative portion of the culture medium collection. Furthermore, the method includes introducing cells from seed training into the micro-scale containers to initiate cell culture in each micro-scale container. The method also includes controlling and monitoring the cell culture in each micro-scale container via a process control device and at least in partial parallel. The method further includes determining a process trajectory from each micro-scale container for the micro-scale containers during or at the end of a cell culture run. Furthermore, the method includes determining a statistically representative trajectory from the process trajectory of a microscale container, wherein the statistically representative trajectory represents the effect of a second culture medium on cell culture. Additionally, the method includes determining whether a first culture medium is suitable for controlling cell culture in a production-scale container based on the effect of the second culture medium on cell culture. When the first culture medium is determined to be suitable, the method further includes comparing the statistically representative trajectory with a provided production-scale process trajectory and using the first culture medium to control cell culture in a production-scale container based on the comparison.

[0024] The regulation of cell culture control can refer to the selection of process data for controlling cell culture based on the analysis of the culture medium. Process data may include production-scale process trajectories, formulations, and set points. In some cases, adjusting the control of cell culture may include controlling cell culture to produce a biopharmaceutical product that meets specified criteria (e.g., process outputs with specified values).

[0025] A recipe can be a sequence of one or more steps performed in a defined order to produce a target product. Therefore, the recipe can provide instructions for performing the process.

[0026] In some instances, the starting medium can refer to a portion of a set of culture media. Specifically, the starting medium can refer to a culture medium used for inoculating production-scale containers. In other words, the starting medium can be synonymous with the first culture medium. A culture medium can be a substrate that provides one or more of the following substances: nutrients (e.g., as listed above), growth factors, and hormones.

[0027] In some instances, cell culture can refer to the process of growing cells outside their natural environment under controlled conditions. Cells can be isolated from living tissue. Cells can be maintained under controlled conditions. Cells can be on a surface or artificial matrix, or they can float freely in a culture medium.

[0028] A culture medium ensemble can refer to a specified amount or a significant amount of culture medium. For example, in liquid form, the volume of a culture medium ensemble can be up to 10,000 liters. In powder form, the volume of a culture medium ensemble can be up to 6,500 kg.

[0029] A representative portion of the culture medium set can be obtained using laboratory sampling techniques. For example, a grab sampling method can be used.

[0030] In some instances, a container can be a vessel or a receptacle. Production-scale containers can be at least an order of magnitude larger or more than micro-scale containers.

[0031] Production scale can also be referred to as large-scale, industrial-scale, or first-scale. Micro-scale containers can also be referred to as laboratory-scale, experimental-scale, or second-scale. Production scale can correspond to the first order of magnitude, while micro-scale can correspond to the second order of magnitude, where the first order of magnitude differs from the second order of magnitude.

[0032] The number of micro-sized containers can be a multiple of 2. For example, a process control unit can have at least 4, 6, 8, 12, 24, or 48 micro-sized containers.

[0033] Inoculation can describe the introduction of cells into an environment suitable for their growth.

[0034] In some cases, the final stage of seed training can be carried out in production-scale containers. Therefore, the same containers from seed training can be used for the next stage of seed training and for miniature-scale containers in process control devices, inoculating at a target seeding density. The target seeding density can be 2 × 10⁻⁶. 6Cells / mL. Seed training is a multi-step process in which the initial number of cells in a first cell culture is expanded to n-1 cell cultures containing a sufficient number of cells to be seeded into production-scale containers at an initial cell density of the target value. The purpose of seed training is to generate a sufficient number of cells at an appropriate cell density for seeding production-scale containers. Seed training provides room for optimization, such as selecting the optimal time point for cell passage from one scale to a larger scale. Furthermore, seed training allows for selection of both seeding cell density and culture volume.

[0035] In some instances, cells from any stage of seed training, from stage 2 (after freezing) to stage n-1, can be introduced into miniature-scale containers.

[0036] Specifically, cryovials may be too small, and obtaining information about the starting medium before inoculating the production container can be useful.

[0037] Specifically, cells can be separated from seed training stage n-2 to run cell culture in microscale containers in parallel with seed training stage n-1. If n is relatively high (e.g., 7 or greater), it may be desirable to collect cells from stages prior to n-1 (e.g., n-2 or earlier). Collecting cells from earlier stages makes it easier to separate the influence of culture medium on the cells from the effects of passage (i.e., the effect of transferring cells from one stage to the next). Advantageously, the biological behavior of cells in microscale containers can be more similar to the biological behavior of cells in the corresponding stages of seed training compared to cells separated from different stages.

[0038] In some instances, the same container can be used for both the seed training stage and the micro-scale container. For example, a container used for the n-2 stage of seed training can be used for both the micro-scale container and the n-1 stage of seed training.

[0039] At least partially parallel operation can refer to the simultaneous execution of multiple processes by a process control device. These processes can be executed in complete parallel. Alternatively, the processes can be interleaved, such that a process in one culture station of the process control device begins during the execution of a process in a vessel at another cell culture station. Specifically, the execution of processes within different culture stations of the process control device can be interleaved.

[0040] In some instances, process trajectories can be determined throughout the entire cell culture run. A statistically representative trajectory can be determined from the process trajectory at the end of the cell culture run or at a specific point during the run. This specific point can be based on time, for example, chosen midway through the cell culture run, or it can be an interval along the path of the cell culture run (e.g., 1 / 5, 2 / 5, 3 / 5, or 4 / 5). Alternatively, the specific point can be a phase of a fed-batch process. For example, a fed-batch process can be divided into approximately 4 to 8 (e.g., 5) phases. These phases can include one or more of the following: final, rapid (or exponential) growth, declining growth, quiescence, and death. More specifically, a fed-batch process can have the phases described in “Modelling of a Fed-Batch Fermentation Process,” Ulla Saarela et al., 2003. Process trajectories and statistically representative trajectories can be determined at the end of any (or all) phases. Alternatively, a process trajectory and a statistically representative trajectory may be determined after the depletion of major nutrients or after a specific reduction in cell viability (e.g., at least 10% or at least 20%) is detected.

[0041] A statistically representative trajectory can be either the mean or the median. A statistically representative trajectory can be determined after eliminating outliers from the process trajectories, and / or may include weighting some process trajectories more heavily than others. Processes within microscale containers can be controlled in such a way that the statistically significant variations inherent in the process are separated from those introduced by the second culture medium. This can be achieved by varying the process parameters for each microscale container according to the experimental design to reflect the inherent variations in the process parameters. In this way, a statistically representative trajectory represents the effect of the second culture medium on cell culture in microscale containers. Furthermore, because both the second and first cultures originate from the same set of cultures (i.e., both are representative portions of the culture medium set), a statistically representative trajectory also represents the effect of the first culture medium on cell culture in production-scale containers.

[0042] When it is determined that the first culture medium is unsuitable for cell culture in a production-scale vessel (e.g., the second culture medium has too much of a negative impact on cell culture), the culture medium set can be discarded. Alternatively, the deviation in cell culture in a production-scale vessel can be recorded. If a cell culture run for a microscale vessel can be classified as successful, it is possible to record the deviation even if the statistically representative process trajectory is outside the control limits. A cell culture run for a microscale vessel can be classified as successful based on process outputs (e.g., key performance indicators or key quality attributes). Control limits can refer to the upper and lower limits of a reference multivariate process diagram used for the process (discussed in more detail below).

[0043] Several examples are discussed below; these examples may be combined with aspects disclosed in this application, including the foregoing aspects.

[0044] The culture medium set may consist of powdered or liquid culture media. When the culture medium set consists of powdered culture media, sampling may be performed considering the characteristics of the powdered culture media (for the first and / or second culture media), including:

[0045] - The particle size range of powdered culture media.

[0046] - The shape of the particles

[0047] - Changes in particle composition,

[0048] - The amount of culture medium aggregate,

[0049] - The amount of culture medium sampled.

[0050] When the culture medium assembly consists of powdered culture medium, sampling can be performed according to Pierre Gy's sampling theory.

[0051] Culture medium sampling can be performed based on the principles of drug sampling guidelines. More specifically, the World Health Organization guidelines can be used (http: / / apps.who.int / medicinedocs / documents / s21440en / s21440en.pdf).

[0052] The method may further include receiving process parameters to be controlled. Cell culture in microscale containers can be controlled and monitored based on the process parameters to be controlled and the corresponding setpoints. The method may also include setting at least one subset of setpoints differently for at least a portion of the microscale containers. The subset of setpoints can be set according to the experimental design.

[0053] Process parameters can also be called process variables. Process parameters can be process conditions that can change the process in some way.

[0054] The process parameters to be controlled can constrain or regulate the process. These parameters can be set at the start of the process and controlled throughout to manage the environment within a microscale vessel. Examples include temperature and stirring rate. The process parameters to be measured can be determined from a sample of the fluid or via a first process control device. A process parameter can be a process parameter to be controlled, a process parameter to be measured, or both.

[0055] The process parameters to be controlled and measured can each include one or more of the following: dissolved oxygen (DO), dissolved carbon dioxide, pressure, pH, flow rate, temperature, nutrient level, stirring speed, substrate concentration (e.g., glucose, glutamine, or other chemically determined carbon or nitrogen sources), metabolite concentration (e.g., lactate, ammonia), redox potential, and turbidity. Flow rate can include carbon dioxide, oxygen, and acid / base. Nutrient level can refer to organic nutrients (e.g., glucose) or inorganic nutrients (e.g., trace minerals). The values ​​of the process parameters can include time-series values.

[0056] Each setpoint can be a value of a process parameter that is desired to be maintained. For example, if the process temperature needs to be maintained within a 0.5°K interval from 37°C, then the setpoint is 37°C. Setpoints may include a base setpoint. Alternatively, ranges may be used. For example, a temperature process parameter may have a corresponding range of 36.5°C to 37.5°C.

[0057] The method may further include dividing the microscale containers into culture stations, wherein each culture station includes a portion of the microscale containers. Each culture station may include approximately one-third, one-quarter, or one-sixth of the microscale containers. A subset of setpoints that are set differently for a portion of the microscale containers may include setting setpoints for each culture station according to the experimental design.

[0058] The method may further include receiving an acceptance range for each process parameter to be controlled. Each acceptance range may reflect an acceptable variation in the corresponding process parameter to be controlled. The experimental design may reflect the variation of the process parameter to be controlled within the acceptance range. The method may further include receiving a reference multivariate process plot representing the acceptance range. Determining whether a first culture medium is suitable for controlling cell culture may include determining whether a statistically representative trajectory falls within the upper and lower limits of the reference multivariate process plot.

[0059] Each acceptance range may include a normal operating range. The normal operating range may be part of the design space used to control cell culture.

[0060] Typically, a multivariate process diagram (e.g., a reference multivariate process diagram) can be used to control the process. A multivariate process diagram can be implemented as a control chart, i.e., a process behavior diagram. The multivariate process diagram may include a process trajectory that shows the expected path of the process over time, and upper and lower limits (i.e., warning or control limits) defined relative to that trajectory. Values ​​of process parameters from the actual or current process (e.g., measurements can be made) can be determined and compared with the multivariate process diagram to ensure that the process will result in the production of a usable product, i.e., a product that meets at least one specified (e.g., predetermined) condition. The specified condition may be related to at least one process output or process parameter.

[0061] The advantage of using multivariate process diagrams to control a process is that the actual trajectory derived from the process parameter values ​​(i.e., the process trajectory of the controlled process) can be compared with the multivariate process diagram to identify deviations between the actual trajectory and the multivariate process diagram's trajectory, and to correct these deviations as early as possible. Early correction of process deviations, especially those outside the upper or lower limits of the multivariate process diagram, can lead to the production of the product meeting specified conditions.

[0062] The design space can be determined based on hundreds of small-scale runs of the process. The design space can include a tolerable degree of parameter variation without adversely affecting the process results.

[0063] Cell culture can be controlled based on processes that have been performed multiple times previously, for example, at least 5-10 times. The described technique can be particularly useful when receiving a new set of culture media whose characteristics are unknown or only partially known. The process can be performed multiple times using media from different sets of media. Therefore, a multivariate process diagram can be derived from previous successful runs of the process using different media that are not from the sets of media from which first and second media were sampled.

[0064] The process trajectory of each microscale container can be determined at the end of the cell culture run (i.e., at the end or completion) using process data, process-related data, and derivatives of process data or process-related data collected during the cell culture run.

[0065] Data-driven modeling, including deep learning, machine learning, and / or multivariate statistical techniques, can be used to determine process trajectories. Multivariate statistical techniques may include principal component analysis (PCA) and / or projection on the underlying structure (PLS).

[0066] The method may further include determining that a first culture medium is unsuitable for a production-scale container when a statistically representative trajectory exceeds the upper or lower limit of a reference multivariate process graph. When a statistically representative trajectory does not exceed the upper or lower limit of a reference multivariate process graph, the method may further include determining that a first culture medium is suitable for a production-scale container.

[0067] Comparing statistically representative trajectories with provided production-scale process trajectories may include calculating trajectory similarity, which can be calculated using multivariate distance metrics. Multivariate distance metrics may include, but are not limited to, one or more of the following: Euclidean distance, Mahalanobis distance, Hotellings T2 statistic, Q statistic, and distance to model (DModX) statistic.

[0068] The method may further include determining a production-scale process trajectory based on the comparison to most closely resemble a statistically representative trajectory. A production-scale process trajectory can be determined based on multiple production-scale process trajectories. Controlling cell culture in a container based on the comparison may include using a production-scale process trajectory to control cell culture.

[0069] Therefore, an adaptive control loop is formed for variations in the starting medium without requiring a complete analytical profile of the starting medium (i.e., an analysis indicating all components). Specifically, the control of cell culture in production-scale containers is adjusted based on the influencing factors of the starting medium, determined by monitoring its effects on cell culture in microscale containers.

[0070] Controlling and monitoring cell cultures in each microscale container may include periodically determining process parameter values ​​from the cell cultures in each microscale container, at least in part, by a process control device. The method may also include defining groups of process parameter values ​​based on the time intervals at which the corresponding process parameter values ​​are determined. Each group may include process parameter values ​​determined from multiple microscale containers.

[0071] The time interval may correspond to the duration required for the process control device to obtain a fluid sample from each microscale container. The time interval may also correspond to the additional time required to determine process parameter values ​​from the fluid. More specifically, the first group may be defined based on process parameter values ​​determined from a first group of samples, which includes individual samples from each microscale container.

[0072] Furthermore, techniques for smoothing local fluctuations in process parameter values ​​can be used to define groups of process parameter values. More specifically, moving averages can be used to define groups. Thus, a first value obtained from a first micro-scale container can be excluded from a second group, and a second value determined from the first micro-scale container can be added. In other words, adjacent groups can be formed by shifting forward, i.e., excluding a first process parameter value determined from a micro-scale container and including the next value determined from the same micro-scale container. Other smoothing techniques, such as interpolation, can also be applied. Additionally, weighting can be applied to multiple recently obtained values, for example, by assigning higher weights to those values ​​or by assigning lower weights to values ​​that are substantially different from other process parameter values.

[0073] Generally, groups of process parameter values ​​can be defined based on any common characteristics of their values. A common characteristic (e.g., time interval) reflects process maturity, that is, the degree or level of process maturity. Specifically, groups can be defined based on similar process maturity. In other words, process parameter values ​​can be grouped based on process maturity (e.g., the time it took to determine these values). Specifically, a common characteristic can be one of the following:

[0074] - Process output values ​​determined by a miniature container with the same set of corresponding process parameter values. For example, multiple process outputs can be determined periodically from the miniature container. More specifically, the percentage of live cells can be determined as the process output from each miniature container every 6 hours. Thus, process parameter values ​​determined from a miniature container with a first process output value (e.g., a first percentage of live cells) can be in a first set, while process parameter values ​​determined from a miniature container with a second process output value (e.g., a second percentage of live cells) can be in a second set. The first process output value can be within a first specified range, and the second process output value can be within a second specified range. The first specified range can overlap with the second specified range.

[0075] - A range of values ​​for one of the process parameters. For example, a process parameter could be a nutrient concentration, and a first set could include process parameter values ​​determined when the nutrient concentration has a specified value. Similarly, a second set of process parameter values ​​could include values ​​determined when the nutrient concentration has a second specified value. The specified values ​​can each be within different, possibly overlapping, ranges of values.

[0076] These groups can be defined after sufficient data has been collected from each process parameter container. For example, these groups can be defined after the process control unit has accessed each container.

[0077] Defining groups of process parameter values ​​can include generating batch evolutionary models from process parameter values. Each group can be a subset (e.g., an appropriate subset) or a portion of the process parameter values. Groups of process parameter values ​​can be defined after cell culture runs for microscale containers have ended, i.e., after the process in each container has been performed until completion and / or after all steps in the formulation have been performed.

[0078] The group may include multivariate scores derived from process parameter values, rather than individual process parameter values, where each score represents multiple process parameter values. Multivariate scores can be determined using multivariate statistical process control techniques such as PCA and / or PLS.

[0079] Determining a statistically representative trajectory may further include determining an average multivariate process plot from the process trajectory of a microscale container. The average multivariate process plot may include a statistically representative trajectory, an upper limit for the statistically representative trajectory, and a lower limit for the statistically representative trajectory. Determining a statistically representative trajectory may include determining a mean from each of the process parameter value groups, establishing a statistically representative trajectory from said mean, and determining the upper and lower limits based on a measure of variation within each group.

[0080] The method may also include removing irrelevant process trajectories of microscale containers before determining statistically representative trajectories from process trajectories.

[0081] In the case of a batch process, a statistically representative trajectory can be determined using one of the following methods.

[0082] A statistically representative trajectory can be determined by using batch-wise-unfolding (BWU) data, i.e., unfolding the process trajectory (represented as a three-dimensional array), to form a matrix such that all information for each batch is contained in a single row of the matrix (Nomikos, P., MacGregor, JF, "Monitoring of batch processes using multi-way principal component analysis", 1994; Nomikos, P., MacGregor, JF, "Multivariate SPC charts for monitoring batchprocesses", 1995; and Nomikos, P., MacGregor, JF, "Multi-way partial leastsquares in monitoring batch processes", 1995). Data-driven modeling is then performed on the matrix to summarize the main sources of variation between different batches and allows for effective inter-batch comparisons. When performing data-driven modeling, this method also allows for the incorporation of initial conditions (Z) and product quality attributes (Y) associated with each batch.

[0083] Statistically representative trajectories can be determined by observation-wise-unfolding (Owu), that is, by using each row corresponding to an observation at a certain time in each batch and each column corresponding to the measured variable (Wold, S. et al., "Modelling and diagnostics of batch processes and analogous kinetic experiments", 1998; Nomikos, "PhD thesis: Statistical control of batch processes", 1994). Performing data-driven modeling on this Owu data (which may include using local batch times or mature variables as response variables) results in a reduced number of variables (so-called model components (T)) that summarize the significant behavior of the process trajectory of the original process variables.

[0084] To analyze differences between batches, a reduced number of model components (T) from the OWU method can be used, or alternatively, expanded OWU data from which model components are derived can be used. When performing data-driven modeling, this method also allows the incorporation of initial conditions (Z) and product quality attributes (Y) associated with each batch.

[0085] The method may also include receiving a process parameter to be measured. Monitoring cell culture in a microscale container may include determining the process parameter value from the process parameter to be measured.

[0086] Determining process parameter values ​​may include collecting samples from one or more microscale containers and analyzing them using one or more scientific instruments. Scientific instruments may include at least one of the following: molecular identification instruments, metabolite measuring instruments, and nutrient measuring instruments. Scientific instruments may also be referred to as analytical devices and may be adapted to identify substances or components of the material to be analyzed (e.g., cell cultures and / or culture media). For example, scientific instruments may include one or more of the following: spectrometers, mass spectrometers, and liquid chromatography systems (e.g., ultra-high performance or high performance liquid chromatography systems).

[0087] The process control device may include one or more sensors for measuring at least one of the following: pH, dissolved oxygen, and temperature.

[0088] Culture media may have one or more of the following characteristics: they are chemically determined, they are animal-free and / or serum-free. Specifically, culture media sets may be chemically determined, animal-free and / or serum-free.

[0089] Cell culture can be mammalian cell culture, especially cell lines derived from Chinese hamster ovary cells.

[0090] In some instances, production-scale containers can be production bioreactors. More specifically, production-scale containers can be stirred tank reactors (STRs) or oscillating bioreactors. Production bioreactors can be used in batch, fed-batch, or pouring modes.

[0091] Each successfully controlled cell culture can meet at least one success criterion. Success criteria may include at least one process output that meets a specified threshold. Therefore, a production-scale process trajectory derived from successfully controlled cell cultures can be referred to as a gold batch trajectory.

[0092] In the context of cells, process outputs can include the total number of cells, the number of cells per unit volume of input fluid, the chemical composition of the cells, the amount of cell debris, the amount of shear or chemical damage, and cell viability.

[0093] Each process output is a product quality attribute (e.g., CQA) or key performance indicator. Process outputs can be distinguished from process parameters because process parameters can be directly affected or controlled (e.g., a temperature process parameter can be directly controlled by raising the temperature in a container), while process outputs may not be directly controllable. Instead, process outputs can be indirectly affected by controlling process parameters.

[0094] Further details and examples of process outputs, key performance indicators and product quality attributes can be found in "Cellculture processes for monoclonal antibody production", Feng Li et al., 2010.

[0095] Process outputs may include one or more of the following: total amount of product, amount of input fluid or starting material per unit volume, specified characteristics such as chemical composition of product, purity of product, cost of starting material, energy cost of process, glycosylation or glycan profile, charge variants or isotypes, including acidic and basic variants, low molecular weight variants, potency or bioactivity, polymerization or polymerization level, fragmentation.

[0096] Seed training can have n stages, and the final stage of seed training can be performed in a production-scale container. Control and monitoring can be performed in parallel with either stage n-1 or n-2 of seed training.

[0097] In some cases, n can be 6 to 8, resulting in 6 to 8 stages of seed training. Therefore, the first stage of seed training can be performed in a container with a volume of approximately 1 mL to approximately 2.5 mL. The container used for the first stage of seed training can be a cryotube (i.e., a cryopreservation tube) or a shake flask. Thus, the first stage can be performed in a small plastic tube (e.g., a Falcon tube) with centrifuged cells and a preservative to keep the cells in nitrogen. The preservative can be dimethyl sulfoxide (DMSO) or glycerol.

[0098] The second stage of seed training can be carried out in a container with a volume of approximately 1 L to approximately 3 L. The third stage of seed training can be carried out in a container with a volume of approximately 5 L to approximately 15 L. The fourth stage of seed training can be carried out in a container with a volume of approximately 15 L to approximately 25 L. When stage n-1 is carried out in a container with a volume of approximately 150 L to approximately 250 L, stage n is carried out in a container with a volume of approximately 400 L to approximately 600 L. When stage n-1 is carried out in a container with a volume of approximately 400 L to approximately 600 L, stage n can be carried out in a container with a volume of approximately 800 L to approximately 1,200 L or approximately 1,800 L to approximately 2,200 L.

[0099] Therefore, the seed training phase can be carried out in containers with sizes of 1 mL, 2 L, 10 L, 20 L, 500 L, and 1,000 L. As another example, the seed training phase can be carried out in containers with volumes of 2 mL, 2 L, 10 L, 20 L, 1,000 L, and 2,000 L. Additional phases of 20 L to 500 L and 20 L to 1,000 L can be included. Specifically, a 50 L container can also be part of the seed training.

[0100] According to another aspect, a computer program can be provided. The computer program includes instructions that, when executed by a computer, cause the computer to perform any of the methods described above. The computer program can be embodied in a product. The computer program can be embodied in a computer-readable medium. More specifically, the computer program can be tangibly embodied in a computer-readable medium.

[0101] According to another aspect, a system is provided for adjusting control of cell culture in a production-scale container relative to a starting culture medium. The system includes a database storing production-scale process trajectories, each from a successfully controlled cell culture. The system also includes a first process control device for controlling cell culture in a production-scale container. The first process control device includes a production-scale container configured to receive a first culture medium from a culture medium set. The first process control device further includes a controller operable to receive output from seed training initiated using the first culture medium, thereby enabling inoculation of the production-scale container. The system also includes a second process control device. The second process control device includes a plurality of microscale containers, each microscale container configured to contain a representative portion of the culture medium set. The production scale is larger than the microscale. The system also includes a robot capable of addressing each container, dispensing fluid to each container, and extracting a fluid sample from each container. The system also includes a controller operable to dispense a second culture medium sampled from the culture medium set into the microscale containers and receive cells from seed training for the microscale containers to initiate cell culture in each microscale container. The controller is also operable to control and monitor cell culture in each microscale container at least in parallel. The controller is also operable to determine a process trajectory from each microscale container at the end of a cell culture run. The controller is also operable to determine a statistically representative trajectory from the process trajectories of the microscale containers, wherein the statistically representative trajectory represents the effect of a second culture medium on the cell culture. The controller is also operable to determine whether a first culture medium is suitable for controlling cell culture in production-scale containers based on the effect of the second culture medium on the cell culture. When the first culture medium is determined to be suitable, the controller operablely compares the statistically representative trajectory with a stored production-scale process trajectory and, based on the comparison, controls cell culture in production-scale containers using the first culture medium.

[0102] According to another aspect, the system is provided for use in adjusting control of cell culture in a production-scale vessel relative to a starting culture medium.

[0103] The subject matter described in this disclosure can be implemented as a method or apparatus, possibly in the form of one or more computer programs (e.g., computer program products). Such computer programs can cause data processing equipment to perform one or more operations described in this disclosure.

[0104] The subject matter described in this disclosure can be implemented as a data signal or a machine-readable medium, wherein the medium is embodied in one or more information carriers, such as CD-ROM, DVD-ROM, semiconductor memory or hard disk.

[0105] Furthermore, the subjects described in this disclosure can be implemented as a system including a processor and memory coupled to the processor. The memory can encode one or more programs to cause the processor to execute one or more methods described in this application. Other subjects described in this disclosure can be implemented using various machines.

[0106] Details of one or more implementations are set forth in the exemplary drawings and description below. Other features will be apparent from the specification, drawings, and claims. Brief description of the attached diagram

[0108] Figure 1 The first part of a flowchart shows a method for adjusting controls of cell culture in a production-scale vessel relative to the starting culture medium.

[0109] Figure 2 The second part of the flowchart shows the control of cell culture in a production-scale vessel relative to the starting culture medium.

[0110] Figure 3 The third part of the flowchart shows the control of cell culture in a production-scale vessel relative to the starting culture medium.

[0111] Figure 4 Part four of the flowchart shows the control of cell culture in a production-scale vessel relative to the starting culture medium.

[0112] Figure 5 Part five of the flowchart shows the control of cell culture in a production-scale vessel relative to the starting culture medium.

[0113] Figure 6 A perspective view of a process control device comprising multiple miniature containers is shown.

[0114] Figure 7 Another perspective view of the process control device is shown.

[0115] Detailed description

[0116] The following text provides a detailed description of the examples with reference to the accompanying drawings. Various modifications can be made to the examples. Specifically, one or more elements of an example can be combined and used in other examples to form new examples.

[0117] Figures 1 to 5 A portion of a single flowchart is shown, such that each diagram includes a part of the steps of the entire flowchart.

[0118] Figure 1 Step S101 initiates a method for adjusting the control of cell culture in a production-scale vessel relative to the starting culture medium. Cell culture can be controlled during the execution of chemical, pharmaceutical, biopharmaceutical, and / or biological processes.

[0119] More specifically, the process can be a feed-in batch process. Therefore, the process can have an initial batch phase and subsequent phases in which a constant feed of the substrate is performed. In one example, the entire process lasts approximately 14 days, with the initial batch phase lasting approximately 3 days.

[0120] For various reasons, control over cell culture in production-scale containers relative to the starting medium may be desirable. For example, even when receiving culture medium sets from the same source, the composition (or amount of composition) can vary significantly between different culture medium sets. Culture medium sets can be complex mixtures containing multiple components, making it potentially surprisingly difficult to identify and quantify each component present in a given culture medium set. Furthermore, culture medium set manufacturers may protect the amount of components or the components themselves as trade secrets.

[0121] Furthermore, obtaining a complete analytical profile of a given culture medium set may be difficult. Analytical techniques may not provide all the information about the components of the culture medium set. Specifically, no single spectroscopic technique can capture all the relevant information for each culture medium. In addition, some analytical techniques (e.g., spectroscopic techniques) may be too expensive or labor-intensive.

[0122] In step S103, a starting culture medium may be received. The starting culture medium may be derived from a collection of culture media used for cell culture. Cell culture may be mammalian cell culture, such as a cell line derived from the ovary of a Chinese hamster. The collection of culture media may include chemically defined culture media. The collection of culture media may be serum-free and / or free of animal-derived components.

[0123] In step S105, it can be determined whether the culture medium is a readily available liquid. If the culture medium is a readily available liquid, it is sampled in step S107. The sample of the culture medium can be used for initial seed training to inoculate production-scale containers as well as micro-scale containers. The liquid culture medium can be mixed to ensure that each micro-scale container receives a representative portion of the culture medium set.

[0124] Miniature containers can be provided in process control devices, for example, such as... Figure 6 and 7 As shown.

[0125] If the culture medium is not a ready-made liquid, further processing may be required. For example, the culture medium assembly may be provided in powder form. If the culture medium assembly is provided in powder form, a grab sampling method can be used to obtain a representative portion of the culture medium assembly. If the culture medium is received in liquid form but is not ready-made for production, a culture medium for production can be prepared in step S111.

[0126] exist Figure 2 In this process, step S113 can be performed directly after step S109. Step S117 can be performed directly after either step S107 or step S111.

[0127] In step S113, a liquid culture medium for microscale containers can be prepared from the powdered culture medium discussed in step S109. The liquid culture medium can be prepared using conventional techniques.

[0128] In step S115, the experimental design can be defined. The experimental design can be determined based on values ​​already developed during the process development phase (e.g., using different sets of culture media). The process parameters to be measured and controlled can be received or determined during process development. Furthermore, a process development phase can be performed to define the design space, particularly for critical process parameters, in order to obtain key quality attributes with desired values. The normal operating range for each process parameter can be defined during process development.

[0129] Therefore, the experimental design can be defined within the design space to reflect the naturally occurring variations of each process parameter within its normal operating range. For example, temperature can vary by ±0.1℃, and pH can be defined as varying by ±0.1℃. Variations can also be defined for other process parameters.

[0130] The miniature containers of the process control unit can be separated between different culture stations. For example, when the process control unit includes 24 containers, each culture station can include 6 containers. Therefore, each culture station can be configured with a different setpoint. For example, the temperature of the first group of culture stations can be set 0.1°C lower than the temperature setpoint of the second group of culture stations.

[0131] In step S117, seed training can be initiated to inoculate production-scale containers. Seed training can be initiated using a first culture medium sampled from a culture medium collection. Seed training can range from microscale (e.g., shake flasks or cryovials) to production-scale containers (e.g., 2,000 L production bioreactors). The seed training phase can be performed in parallel with the treatment in the microscale containers.

[0132] Seed training can be used to inoculate production-scale containers with an initial cell density of at least the inoculation threshold amount. In this paper, cell density may refer to viable cell density.

[0133] The cells in the container during the first stage of seed training can have an initial threshold cell density. The initial threshold can be at least 2 × 10⁻⁶ cells / year. 5 Cells / mL. In some cases (depending on the process and / or cell line), the initial threshold quantity may be at least 3 × 10⁻⁶ cells / mL. 6 c / mL and up to 10×10 6 c / mL.

[0134] Production containers can be seeded using cell densities at least at the seeding threshold. The seeding threshold can be at least 2 × 10⁻⁶. 6 C / mL (cells / mL). Depending on the process and / or cell line, the seeding threshold can be as high as 75 × 10⁻⁶. 6 Or up to 100×10 6 Cells / mL. In some cases, the n-1 phase of seed training can have a range of 10-150 × 10⁻⁶ cells / mL. 6 Cell density of c / mL.

[0135] Production-scale containers can be (continuous) stirred tank reactors and / or perfusion bioreactors. Production-scale containers can also be implemented as gyratory platforms or gyratory motion containers. Containers within the seed training can be implemented as stirred tank reactors, single-use bioreactors, perfusion bioreactors, or gyratory motion reactors. Other implementations or combinations are also possible.

[0136] In step S119, cells from seed training can be introduced into microscale containers to initiate cell culture in the microscale containers. Specifically, cells can be isolated from seed training, expanded along with a target, and used to inoculate multiple microscale containers in parallel to initiate parallel operation of the microscale containers, allowing for control and monitoring of the microscale containers in parallel with the n-2 or n-1 phase of seed training. Therefore, any stage from the second to the n-1 phase of seed training can provide a cell source suitable for inoculating microscale containers. In some cases, it may be necessary to control and monitor cell culture in the microscale containers in parallel with the n-1 or n-2 phase of seed training.

[0137] In some instances, the same set of culture media sampled for seed training (including production-scale containers) can also be the source of culture media for microscale containers.

[0138] In step S121, cell culture in each microscale container can be controlled and monitored via a process control device. Therefore, process data can be collected from the microscale containers. Furthermore, scientific instruments (e.g., instruments capable of spectroscopic measurements) and inline sensors (e.g., pH, dissolved oxygen, biomass) can be used to analyze process parameters, key performance indicators, and key quality attributes, as well as connected online or offline data. Multivariate analysis can be performed to draw conclusions from the process data.

[0139] Process parameter values ​​can be collected for at least the following: pH, temperature, dissolved oxygen, stirring speed, and venting rate. Online process parameter values ​​may include values ​​for the following: oxygen and carbon dioxide. Online process parameter values ​​may include values ​​for nutrients such as glucose and lactose. Offline data (e.g., data collected using scientific instruments such as spectrometers) may include glucose, ammonia, titers, amino acids, and nicotinamide adenine dinucleotide (NADH). Multivariate data may also be determined, for example, based on principal components derived from values ​​of multiple process parameters.

[0140] In step S123, the seed training stage n-1 can be reached. Stage n-1 can be a 500L container or a 200L container. Other container sizes are also possible. Stage n-1 can be the stage just before the production-scale container. Therefore, the size of the container used in the seed training stage n-1 can be approximately half the size of the production-scale container.

[0141] Process data may include values ​​of process parameters determined during the process. These determined process parameter values ​​may be obtained from the process parameter to be measured (e.g., measured temperature or pH), or determined through analysis (e.g., spectroscopic), or through a hybrid system model (e.g., metabolic rate). A hybrid system model is one in which experimentally measured data are combined with theoretical process models (e.g., metabolic flow equilibrium analysis) to create different state observers (e.g., metabolic state observers).

[0142] Figure 3 This includes steps S125 to S133. Step S125 immediately follows step S121. Step S133 immediately follows step S123.

[0143] In step S125, a statistically representative trajectory can be determined from the process trajectory of the microscale vessel. The process trajectory from the microscale vessel can be determined at the end of the cell culture run used in the microscale vessel. The end of the cell culture run can occur as follows:

[0144] -After a specified amount of time has elapsed

[0145] -After measuring or determining certain process outputs from the process, or

[0146] - After the multivariate analysis of process parameter values ​​and / or quality attributes reaches the specified results.

[0147] The end of a cell culture run can refer to the end or completion of the process.

[0148] A statistically representative trajectory can be the average of process trajectories from microscale containers, for example, after outlier removal. A process trajectory can also be referred to as a process data trajectory. Alternatively, a statistically representative trajectory can be determined by defining groups of process parameter values ​​within each microscale container's cell culture. These groups of process parameter values ​​can be defined based on the time interval at which those values ​​were determined. Each group of process parameter values ​​can be determined from multiple microscale containers.

[0149] At step S127, the stored process data from database 301 can be used to validate a statistically representative process trajectory. The stored process data may include acceptable ranges for each process parameter (e.g., reference trajectories and known operating ranges). The stored process data may be obtained during process development, for example, using media from other media collections. Validation may include determining whether a statistically representative process trajectory falls within acceptable ranges represented by a reference multivariate process plot (e.g., whether a statistically representative process trajectory falls within the upper and lower limits of the multivariate process plot).

[0150] In step S129, it can be determined whether a statistically representative trajectory exceeds the upper or lower limit or boundary of the reference multivariate process diagram. If the statistically representative trajectory is within the boundary of the reference multivariate process diagram, it can be determined that the culture medium is suitable for cell culture in a container for controlling the scale of production.

[0151] Step S133 can be performed after step S129. Therefore, the process can continue as follows. Figure 4 As shown.

[0152] Alternatively, seed training may be stopped in step S133 if a statistically representative trajectory falls outside the limits of the reference multivariate process diagram. The determination that a statistically representative trajectory falls outside the limits of the reference multivariate process diagram can be a sufficient indication that the culture medium is unsuitable for use in production-scale vessels. Therefore, the process can be stopped in step S133. Alternatively, further analysis may be performed to determine how to adjust controls for cell culture in production-scale vessels relative to the starting culture medium.

[0153] If further analysis reveals a path forward, the process can continue. Figure 4 Step S141.

[0154] Figure 4 A database 401 for storing production process data is shown. Database 401 may be part of database 301. Alternatively, database 401 may be implemented as a separate database. Considering the influence of the second culture medium on cell culture in production-scale containers, step S137 may be performed when it is determined that the first culture medium is suitable for controlling cell culture in production-scale containers. Step S137 may include finding the production-scale process trajectory that most closely resembles a statistically representative trajectory. Therefore, step S137 may be performed by searching database 401 and comparing the production process trajectory stored in database 401 with a statistically representative trajectory. The trajectory comparison may be performed using a multivariate distance metric. The multivariate distance metric may include one or more of the following: Euclidean distance, Hotellings T2 range, distance to the model (e.g., DModX), Mahalanobis distance. Other techniques may also be used.

[0155] In step S139, the production (i.e., production scale) process trajectory most similar to the statistically representative trajectory can be extracted from the database. Furthermore, the recipe and corresponding setpoint can be associated with the production process trajectory in database 401 and can be extracted together with the production process trajectory.

[0156] In step S141, the final stage of seed training can be initiated. Therefore, cell culture can be initiated in a production-scale container. In step S143, cell culture in the production-scale container can be controlled and monitored. Cell cultures in the production-scale container can be cultured using a first culture medium. Cell culture in the production-scale container can be controlled and monitored based on a comparison of statistically representative trajectories with production-scale process trajectories in database 401. Specifically, trajectories stored in database 401, most similar to statistically representative trajectories, can be used to control and monitor cell culture in the production-scale container. Database 401 may include multiple production process trajectories that may come from multiple users. Specifically, database 401 may include at least 5 production process trajectories, at least 10 production process trajectories, or at least 20 production process trajectories.

[0157] continue Figure 5In the process described in step S105, process data can be collected from the production process. Process data can be collected simultaneously with controlling and monitoring the production process. The process data collected in step S145 can be stored in database 501. Furthermore, in step S145, the recipe and setpoint extracted from database 401 can be modified and stored in database 501. Database 501 can be separate from database 401 or as part of database 401. Specifically, databases 301, 401, and 501 can be separate databases (e.g., on a separate computer system) or can be implemented as a single database with multiple tables. In step S147, the trajectory of the production process can be determined. The trajectory can be formed during the production process and can be finalized after the production process ends. The determined trajectory of the production process can be stored in database 501. In addition to this trajectory, updated recipes and updated setpoints can be stored for the production process. Associations between the production process trajectory, updated setpoints, and updated recipes can be created within database 501.

[0158] In step S149, the process ends.

[0159] The process control device 10 (possibly implemented as a bioreactor system) includes a container array (e.g., a microscale bioreactor). Figure 6 As shown in the diagram. The container is configured to contain fluids (other types of fluids are also possible) for producing biopharmaceutical products from cell cultures. Process control device 10 may correspond to the process control device described above. The container may be located in container station 11 (also called receiving station). Container station 11 is configured to receive a specified number of containers, such as 6, 12, or 24. Process control device 10 is operable to control and / or monitor the containers in container station 11 in at least partially parallel, and possibly fully parallel, manner. Process control device 10 is operable to determine or set process parameters to be controlled (i.e., process control parameters). Examples of process parameters to be controlled include stirring speed and gas supply rate.

[0160] Process control device 10 is operable to periodically determine process parameter values ​​for a process parameter (e.g., the process parameter to be measured). Process parameter values ​​can be determined directly from the container (e.g., via a sensor point) or from a sample removed from the container. More specifically, analysis module 12 can be used to process fluid (e.g., a sample) from the container to determine process parameter values. Thus, analysis module 12 can direct fluid from the container to scientific instruments (e.g., analytical instruments) to determine values ​​for process parameters such as pH, cell counts, metabolite levels, and nutrient levels. The pH value determined by the analysis module can be used for sensor calibration. Analysis module 12 can also support sample preparation and cleaning and rinsing after sample collection.

[0161] The process control device 10 includes a robot, which may be implemented as a liquid processor 13. The robot is capable of addressing each container of the first scale, and dispensing and extracting fluid from the containers. The liquid processor 13 performs automated process control and sampling. The liquid processor 13 collects (or extracts) samples from each individual container in the container station 11 and supplies nutrients or detergents (e.g., acids, alkalis, defoamers, etc.) to each individual container. These tasks may also be performed by the robot in embodiments other than the liquid processor 13.

[0162] The process control device 10 may include a process control module 14 (also referred to as a workstation). The process control module 14 includes a user interface (e.g., input devices such as a keyboard, output devices such as a display, processing devices, and memory). The process control module may store process control strategies to control the process control device 10, and more specifically, to control the liquid processor 13 and the analysis module 12. Specifically, the process control device may store values ​​of process parameters to be controlled (i.e., control setpoints). Furthermore, the process control device may store recipes for the process.

[0163] The process control device 10 may include a sampling device 15. More specifically, the liquid processor 13 may include a sampling device 15. The sampling device 15 may enable an automated pipetting system and / or carry a pipette tip.

[0164] The process control device 10 may include a liquid 16 to be supplied to the analysis module 12. The liquid 16 may include cleaning and rinsing agents, pH buffers, calibration solutions, etc.

[0165] The analysis module 12 and the process control module 14 can be combined in the controller.

[0166] Storage vessel 17 can be used to store liquids to be supplied to containers. Liquids from storage vessel 17 can be supplied by process control device 10, particularly liquid processor 13. Liquids may include glucose feed, acids, alkalis, defoamer solutions, etc.

[0167] The process control device 10 may include a sample holder or apparatus, possibly implemented as a sample cup 18. More specifically, the sample cup 18 may be part of the analysis module 12. The sample cup 18 may be configured to receive samples acquired by the liquid processor 13 and / or the sampling device 15, and to feed the samples to the analysis module 12 and further analysis devices.

[0168] Process control device 10 may include scientific instruments, possibly in the form of analytical device 20. Analytical device 20 may be implemented as a Raman measurement system (i.e., a spectrometer), a high-performance liquid chromatography (HPLC) device, or a mass spectrometry device. Multiple analytical devices (not shown) may be present. Analytical device 20 may be configured to receive samples from analytical module 12 and perform analytical measurements to determine process parameter values ​​or process outputs. Process outputs may include product quality attributes, such as glycosylation.

[0169] One or more heaters or coolers (not shown) may be located near container station 11 to control the temperature of the container.

[0170] Figure 7 Shown from a top-down perspective Figure 6 The process control device.

Claims

1. A computer-implemented method for adjusting control of cell culture in a production-scale vessel relative to a starting culture medium, comprising: It provides process trajectories for multiple production scales, each derived from successfully controlled cell cultures; Receive a collection of culture media for cell culture, wherein the collection of culture media consists of powdered culture media or liquid culture media; A first culture medium that may be used in the production-scale container is sampled from the culture medium set; The first culture medium is used to start seed training to achieve inoculation of the production-scale containers; The process control unit provides multiple micro-scale containers, the production scale being larger than the micro-scale; A second culture medium is sampled from the culture medium set for use in the microscale containers, wherein each of the microscale containers receives a representative portion of the culture medium set; Cells from seed training are introduced into microscale containers to initiate cell culture in each of the microscale containers; Receive process parameters to be controlled; Receive the acceptable range for each process parameter to be controlled. The acceptable ranges mentioned therein each reflect the acceptable variation of the corresponding process parameter to be controlled; Cell culture in each of the microscale containers is controlled and monitored via the process control device and at least in part in parallel, based on the process parameters to be controlled and the corresponding setpoints, wherein at least one subset of setpoints is set differently for at least a portion of the microscale containers, wherein the at least one subset of setpoints is set according to the design of the experiment. The design of the experiments described above reflects the variation of the process parameters to be controlled within the acceptable range. A reference multivariate process diagram representing the scope of acceptance; During or at the end of the cell culture run, a process trajectory from each of the microscale containers is determined for each of the microscale containers. A statistically representative trajectory is determined from the process trajectory of the microscale container, wherein the statistically representative trajectory represents the effect of the second culture medium on the cell culture; Based on the effect of the second culture medium on cell culture, determine whether the first culture medium is suitable for cell culture in a container to control the production scale. Determining whether the first culture medium is suitable for controlling the cell culture includes determining whether the statistically representative trajectory is within the upper and lower limits of the reference multivariate process graph; and When the first culture medium is determined to be suitable Compare the statistically representative trajectory with the provided production-scale process trajectory; and Based on the comparison, cell culture in the container at the production scale is controlled using the first culture medium.

2. The method as described in claim 1, When the culture medium assembly consists of powdered culture media, sampling is performed considering one or more characteristics of the powdered culture media, including: The particle size range of powdered culture media. The shape of the particles, Changes in particle composition, The amount of the culture medium assembly, The amount of culture medium sampled.

3. The method of claim 1 or 2, further comprising: The microscale containers are divided into culture stations, wherein each culture station includes a portion of the microscale containers; A subset of setting the setpoints differently for a portion of the microscale containers includes setting the setpoints for each of the culture stations according to the design of the experiment.

4. The method of claim 1, further comprising: When the statistically representative trajectory exceeds the upper or lower limit of the reference multivariate process graph, it is determined that the first culture medium is not suitable for the container of the production scale. When the statistically representative trajectory does not exceed the upper or lower limit of the reference multivariate process graph, the container for the first culture medium is determined to be suitable for the production scale.

5. The method of claim 1, wherein the comparison includes calculating the similarity of the trajectories, wherein the similarity of the trajectories is calculated based on a multivariate distance metric.

6. The method of claim 1, further comprising: Based on the comparison, the production-scale process trajectory that is most similar to the statistically representative trajectory is determined; Cell culture in a container controlled by the comparison includes controlling the cell culture using a process trajectory at the production scale.

7. The method according to any one of claims 1 to 2 and 4 to 6, The control and monitoring of cell culture in each microscale container includes: The process parameter values ​​are determined periodically, at least in part, from the cell culture in each microscale container by the process control device. The process parameter values ​​are defined as groups within each group based on the time interval at which the corresponding process parameter values ​​are determined, wherein each group includes process parameter values ​​determined from multiple microscale containers.

8. The method of claim 7, wherein determining the statistically representative trajectory further comprises determining an average multivariate process plot from the process trajectory of the microscale container, the average multivariate process plot including the statistically representative trajectory, an upper limit for the statistically representative trajectory, and a lower limit for the statistically representative trajectory, including: Determine the average value from the process parameter values ​​of each group; The statistically representative trajectory is established from the average value; The upper and lower limits are determined based on measurements of changes within each group.

9. The method of claim 1, further comprising: Receive the process parameters to be measured. Monitoring cell culture in microscale containers involves determining the values ​​of process parameters to be measured, including: Samples are collected from one or more miniature containers; The sample is analyzed using one or more scientific instruments, wherein the scientific instruments include at least one of the following: molecular identification instruments, metabolite measuring instruments, and nutrient measuring instruments.

10. The method of claim 1, wherein the culture medium has one or more of the following characteristics: It is chemically determined; It is animal-free and / or serum-free.

11. The method of claim 1, wherein the seed training has n The seed training is conducted in several phases, with the final phase taking place in the production-scale container. The control and monitoring phases are related to the seed training. n-2 or n-1 They proceed in parallel.

12. The method as described in claim 1, The seed training described therein has n Each stage; The first stage is carried out in a container with a volume of 1 ml to 2.5 ml. The second stage is carried out in a container with a volume of 1L to 3L. The third stage is carried out in a container with a volume of 5L to 15L. The fourth stage is carried out in a container with a volume of 15L to 25L. At this stage n-1 When carried out in a container with a volume of 150L to 250L, the stages n The process is carried out in containers with a volume of 400L to 600L; At this stage n-1 When carried out in a container with a volume of 400L to 600L, the stages n It is carried out in containers with a volume of 800L to 1200L or 1800L to 2200L.

13. The method as described in claim 3, The control and monitoring of cell culture in each microscale container includes: The process parameter values ​​are determined periodically, at least in part, from the cell culture in each microscale container by the process control device. The process parameter values ​​are defined as groups within each group based on the time interval at which the corresponding process parameter values ​​are determined, wherein each group includes process parameter values ​​determined from multiple microscale containers.

14. The method of claim 13, wherein determining the statistically representative trajectory further comprises determining an average multivariate process plot from the process trajectory of the microscale container, the average multivariate process plot including the statistically representative trajectory, an upper limit for the statistically representative trajectory, and a lower limit for the statistically representative trajectory, including: Determine the average value from the process parameter values ​​of each group; The statistically representative trajectory is established from the average value; The upper and lower limits are determined based on measurements of changes within each group.

15. A system for adjusting control of cell culture in a production-scale vessel relative to a starting culture medium, the system comprising: A database storing process trajectories at production scale, with each production scale process trajectory derived from successfully controlled cell cultures; A first process control device for controlling the scale of cell culture in a container, the device comprising: The production-scale container is configured to receive a first culture medium from a culture medium set, wherein the culture medium set consists of powdered or liquid culture medium. The controller, which is operated to: Receive the output from seed training starting with the first culture medium to enable inoculation of production-scale containers; A second process control device, the device comprising: Multiple microscale containers, each of the microscale containers being configured to contain a representative portion of the culture medium assembly, wherein the production scale is larger than the microscale; A robot capable of addressing each of the containers, dispensing fluid into each of the containers, and extracting fluid samples from each of the containers; The controller, which is operated to: A second culture medium sampled from the culture medium set is dispensed into the microscale container; Cells for the microscale containers are received from the seed training to initiate cell culture in each of the microscale containers; Receive process parameters to be controlled; Receive the acceptable range for each process parameter to be controlled. The acceptable ranges mentioned therein each reflect the acceptable variation of the corresponding process parameter to be controlled; Cell culture in each of the microscale containers is controlled and monitored in at least partially parallel according to the process parameters to be controlled and the corresponding setpoints, wherein at least one subset of setpoints is set differently for at least a portion of the microscale containers, wherein the at least one subset of setpoints is set according to the design of the experiment. The design of the experiments described above reflects the variation of the process parameters to be controlled within the acceptable range. A reference multivariate process diagram representing the scope of acceptance; During or at the end of a cell culture run, a process trajectory from each of the microscale containers is determined for each microscale container. A statistically representative trajectory is determined from the process trajectory of the microscale container, wherein the statistically representative trajectory represents the effect of the second culture medium on the cell culture; Based on the effect of the second culture medium on the cell culture, determine whether the first culture medium is suitable for cell culture in a container to control the production scale; Determining whether the first culture medium is suitable for controlling the cell culture includes determining whether the statistically representative trajectory is within the upper and lower limits of the reference multivariate process graph; and When the first culture medium is determined to be suitable Compare the statistically representative trajectory with the stored production-scale process trajectory; Based on the comparison, cell culture in the container at the production scale is controlled using the first culture medium.

Images