HK1238597A1 - Process control systems and methods for use with filters and filtration processes - Google Patents

Description

Process control system and method for filter and filtering process

Cross referencing

The present invention claims priority from united states provisional application No. 61/992,595, filed on 5/13/2014, and which is incorporated herein in its entirety.

Technical Field

This patent relates to process control systems and methods, and in particular, to process control systems and methods for filter and filtration processes.

Background

Many products, such as antibodies, and more specifically monoclonal antibodies, are of cellular origin. To prepare a cell-derived product, one or more initial unit operations may be performed to remove cells and any associated cell debris to achieve purification. After purification, one or more subsequent unit operations may be performed to prepare the product for administration. Filtration can be included in both initial and subsequent unit operations, as set forth below in the context of the general description of the overall process for preparing the cell-derived product.

Referring to commercial scale unit operations that may be used to prepare medical grade extracellular expression products such as antibodies or immunoglobulins, initial separation operations such as centrifugation or filtration may be used to remove cells and cell debris. Centrifugation involves applying centrifugal force (relative to the axis) to a liquid solution or suspension such that denser components of the solution or suspension are displaced further away from the axis, while less dense components of the solution or suspension are displaced toward the axis (or at least are displaced less far away from the axis than dense components). Filtration is a pressure-driven process that uses membranes to separate components in a liquid solution or suspension based on size differences between the components. When used in cell separation collection applications, filtration may be referred to as microfiltration. Depending on the amount of the initial cells and/or cell debris in the solution or suspension, or depending on the extent to which the cells and/or cell debris have been separated off by the centrifugation or primary filtration process, any of the above mentioned centrifugation or filtration may be preceded or followed by one or more (additional) filtration unit operations.

Once the cells and cell debris have been satisfactorily removed, purification can be performed in one or more devices (which may be in the form of one or more columns) using a process known as chromatography. Chromatography involves interaction between a first phase, called the mobile phase, and a second phase, called the stationary phase. Oftentimes, the product of interest in the mobile phase is bound to the stationary phase and then the product is separated from the stationary phase using a solvent (referred to as an eluent). At other times, the product of interest flows through the mobile phase while the contaminants bind to the stationary phase.

The exact nature of the interaction between the mobile and stationary phases is different from the type of chromatography used. Ion exchange chromatography relies on the attractive force between a charged molecule (or contaminant) of the product of interest and a solid phase of opposite charge. For example, in cation exchange chromatography, positively charged molecules are attached to a negatively charged solid phase. Affinity chromatography involves the use of ligands that specifically bind to the product (i.e., the target molecule) or contaminant. With respect to the antibody or immunoglobulin product of interest, the ligand may be an associated antigen.

Once purification has been completed, the product eluted from the chromatographic device may be transported for further processing, including, for example, formulating the protein in the form of a pharmaceutically acceptable excipient and/or performing filtration, prior to administration to a patient. For example, filtration may be performed to remove any viruses present, thereby ensuring viral safety of biotechnologically derived therapeutic agents. Additionally, filtration may be performed on the product to concentrate the product to a medical grade and to desalt the product. While the purpose of filtration is still to separate larger components from smaller components, unlike pre-purification filtration, which is performed to remove cells and cell debris from the product, post-purification filtration removes small peptides and salts from the product in order to increase the concentration of the product. This filtration can also be used to desalt the product, or to introduce a stable pharmaceutical formulation for storage of the product prior to filling (i.e., buffer exchange or replacement). When used in this context, filtering may be referred to as hyperfiltration.

The filtration described above may be a dead-end process or a cross-flow or tangential-flow process. In a dead-end filter, the flow of the liquid solution or suspension (or feed liquid) to be separated is perpendicular to the membrane. Most of the feed flow in a cross-flow filter is tangential to or across the membrane surface.

Tangential flow filtration (or TFF) has certain advantages over dead-end filtration. In particular, the accumulation of material on the membrane surface (also referred to as a retentate membrane layer) during tangential flow filtration is minimized, thereby increasing the length of time that the filter can operate. Tangential flow filtration can therefore be applied to continuous process applications, since the feed liquid can be continuously supplied into and through the filter.

A particular type of tangential flow filtration known as single pass tangential flow filtration may be used in certain applications. Conventional TFFs include directing the feed liquid through a filter device multiple times (or multiple filter devices arranged in parallel), while SPTFFs include directing the feed liquid through a filter device in a single pass. According to certain embodiments, the SPTFF filtration device may comprise a single membrane. According to other embodiments, the SPTFF filtration device may be defined by a plurality of membrane cartridges connected in series, the retentate of one stage being directed as a feed to the following stage. The cartridge may be connected to a plurality of brackets or diverter plates. Alternatively, the housing may be designed to receive multiple membranes, the housing providing a way for connecting individual membranes.

As set forth in detail below, the present invention sets forth an improved process control system and method for filters and filtration, particularly tangential flow filters and filtration, including advantageous alternatives to conventional devices and methods.

Disclosure of Invention

According to an aspect of the invention, a process control system includes: one or more upstream processing units each operating a flow rate; a tank connected to the one or more upstream processing units; a filter having an inlet, a permeate outlet, and a retentate outlet connected to the sump; a fluid supply pump having an inlet connected to the sump and an outlet connected to the inlet of the filter; and a sensor disposed at the permeate outlet for determining a flow rate at the permeate outlet. The system also includes a control system coupled to the sensor and the upstream process and adapted to control a flow rate of one or more of the one or more upstream processing units as a function of a flow rate at the permeate outlet.

According to another aspect of the invention, a process control method is provided for use in conjunction with one or more upstream processing units, a tank connected to the one or more upstream processing units, and a tangential flow filter having an inlet, a permeate outlet, and a retentate outlet connected to the tank. The method includes sensing a flow rate at the permeate outlet and controlling the flow rate of one or more of the one or more upstream processing units based on the flow rate at the permeate outlet.

According to yet another aspect of the present invention, a process control system includes: one or more upstream processing units each operating a flow rate; a tank connected to the one or more upstream processing units; a filter having an inlet, a permeate outlet, and a retentate outlet connected to the sump; a fluid supply pump having an inlet connected to the sump and an outlet connected to the inlet of the filter; and a sensor disposed at the permeate outlet for determining a flow rate at the permeate outlet. The system also includes a control system coupled to the sensor and the feed pump and adapted to control the feed pump according to a flow rate at the permeate outlet.

According to yet another aspect of the invention, a process control method is provided for use in conjunction with one or more upstream processing units, a tank connected to the one or more upstream processing units, and a tangential flow filter having an inlet, a permeate outlet, and a retentate outlet connected to the tank. The method includes sensing a flow rate at the permeate outlet and pumping material from the tank into the filter according to the flow rate at the permeate outlet.

According to yet another aspect of the present invention, a process control system includes: one or more upstream processing units each operating a flow rate; a tank connected to the one or more upstream processing units; a filter having an inlet, a permeate outlet, and a retentate outlet connected to the sump; a fluid supply pump having an inlet connected to the sump and an outlet connected to the inlet of the filter; and a sensor disposed at the permeate outlet for determining a flow rate at the permeate outlet. The system also includes a control system coupled to the sensor and the feed pump and adapted to control the feed pump according to the flow rate at the permeate outlet until a predetermined flow rate for the feed pump is reached, and adapted to control the flow rate of one or more of the one or more upstream processing units according to the flow rate at the permeate outlet after the predetermined flow rate for the feed pump is reached.

According to another aspect of the invention, a process control method is provided for use in conjunction with one or more upstream processing units each having a flow rate, a tank connected to the one or more upstream processing units, and a filter having an inlet, a permeate outlet, and a retentate outlet connected to the tank. The method includes sensing a flow rate at the permeate outlet, pumping material from the tank into the filter according to the flow rate at the permeate outlet until a predetermined pumped flow rate is reached, and then controlling the flow rate of one or more of the one or more upstream processing units once the predetermined pumped flow rate is reached.

According to yet another aspect of the present invention, a process control system includes: one or more upstream processing units each operating a flow rate; a tank connected to the one or more upstream processing units; a filter having an inlet, a permeate outlet, and a retentate outlet connected to the sump; a fluid supply pump having an inlet connected to the sump and an outlet connected to the inlet of the filter; and a sensor disposed at the permeate outlet for determining a flow rate at the permeate outlet. The system also includes a control system coupled to the sensor and the feed pump and adapted to control the feed pump according to the flow rate at the permeate outlet until a predetermined flow rate for the feed pump is reached and adapted to allow a mismatch between the flow rate of the one or more upstream processing units and the flow rate of the feed pump after the predetermined flow rate for the feed pump is reached.

According to yet another aspect of the invention, a process control method is provided for use in conjunction with one or more upstream processing units, a tank connected to the one or more upstream processing units, and a filter having an inlet, a permeate outlet, and a retentate outlet connected to the tank. The method includes sensing a flow rate at the permeate outlet, pumping material from the tank into the filter according to the flow rate at the permeate outlet until a predetermined pumping flow rate is reached, and then pumping material from the tank into the filter according to the predetermined pumping flow rate, thereby allowing a volume change in the tank.

According to yet another aspect of the present invention, a process control system includes: a microfiltration unit; a single pass tangential flow filter having an inlet, a permeate outlet, and a retentate outlet; a feed pump having an inlet connected to the microfiltration unit and an outlet connected to the inlet of the filter; and a permeate pump having an inlet connected to the permeate outlet of the filter. The system also includes a control system coupled to the permeate pump and adapted to control the permeate pump to vary a flow reduction factor, wherein the flow reduction factor is a ratio of the feed liquid flow rate to the retentate flow rate.

In accordance with another aspect of the invention, a process control method comprises pumping material through a single pass tangential flow filter having an inlet, a permeate outlet, and a retentate outlet; and pumping permeate from the permeate outlet of the filter to vary a flow rate reduction factor, wherein the flow rate reduction factor is a ratio of the feed liquid flow rate to the retentate flow rate.

According to another aspect of the invention, a process for purifying a protein is provided. The process utilizes one or more upstream processing units, a tank connected to the one or more upstream processing units, and a tangential flow filter having an inlet, a permeate outlet, and a retentate outlet connected to the tank. The process comprises sensing the flow rate at the permeate outlet as protein flows from the retentate outlet back to the storage tank. Then, the process includes performing one of (i) through (iv). According to (i), the process may comprise controlling the flow rate of one or more of the one or more upstream processing units according to the flow rate at the permeate outlet, the flow rate being the flow rate of the material at least partially comprising protein. According to (ii), the process may comprise pumping at least part of the protein-containing material from the reservoir into the filter according to the flow rate at the permeate outlet. According to (iii), the process may comprise pumping at least part of the protein-containing material from the tank into the filter according to the flow rate at the permeate outlet until a predetermined pumping flow rate is reached, and then controlling the flow rate of one or more of the one or more upstream processing units once the predetermined pumping flow rate is reached. According to (iv), the process may comprise pumping at least part of the protein containing material from the reservoir into the filter according to the flow rate at the permeate outlet until a predetermined pumping flow rate is reached, and then pumping material from the reservoir into the filter according to the predetermined pumping flow rate, thereby allowing the volume in the reservoir to change. Finally, after any of (i) to (iv), the process comprises purifying the protein in an eluent, and optionally formulating the protein in the form of a pharmaceutically acceptable excipient.

Drawings

It is believed that the invention will be more fully understood from the following description in conjunction with the accompanying drawings. Some of the figures may have been simplified by the omission of selected elements for the purpose of more clearly showing other elements. Such omission of elements from some figures does not necessarily indicate the presence or absence of particular elements in any exemplary embodiment, unless may be explicitly so recited in the corresponding written description. The drawings are not necessarily to scale.

FIG. 1 is a schematic diagram of a control system used in conjunction with continuous Single Pass Tangential Flow Filtration (SPTFF);

FIG. 2 is a block diagram of a stepped or tiered control method implemented by the control system of FIG. 1;

FIG. 3 is a graph of Volume Reduction Factor (VRF) over time for an example of a stepped control method according to the embodiment of FIG. 2;

FIG. 4 is a graph of Volume Reduction Factor (VRF) over time for an example of a stepped control method and a continuously variable control method;

FIG. 5 is a schematic diagram of a control system used in conjunction with the connection handling system;

FIG. 6 is a block diagram of a variable flow rate control method that may be implemented by the control system of FIG. 5;

FIG. 7 is a block diagram of a constant flow rate control method that may be implemented by the control system of FIG. 5;

FIG. 8 is a block diagram of a hybrid control method that may be implemented by the control system of FIG. 5;

FIG. 9 is a block diagram of a surge control method that may be implemented by the control system of FIG. 5;

FIG. 10 is a block diagram of a high volume setpoint, fixed flux control method that may be implemented by the control system of FIG. 5;

fig. 11 shows a three column mAb purification process, where columns 2 and 3 are the purification steps, where column 2 is typically cation exchange chromatography operating in bind and elute mode, and column 3 is generally designated as anion exchange chromatography operating in flow-through mode, and where the boxes indicate steps operating simultaneously in a linking process, where a large cell can be converted into a small buffer vessel;

FIG. 12 is a schematic illustration of the design of the connection process showing the sequence of operations, connectivity of steps, location of buffer containers, split sampling and on-line titration;

FIG. 13 is a table showing TFF control strategies for connecting downstream processes;

FIG. 14a is a data plot showing the effect of loading conductivity on CHOp removal when performing STIC thin film chromatography (mAb A), with salt limited and high salt experiments having loading conductivities of 16mS/cm and 28mS/cm, respectively, showing that loading conductivity has minimal effect on CHOp levels (vertical bars) and% CHOp reduction (inclined lines) in the cell;

FIG. 14b is a data plot showing the effect of pH on CHOp reduction when performing STIC thin film chromatography of mAb A, where the load material for these four experiments was prepared by titrating a CEX cell (pH 5.0, conductivity 16mS/cm) with 2M Tris to its corresponding pH, the CHOp levels (vertical bars) and% CHOp reduction (curves) in the cell showing that better host cell protein removal was obtained at higher pH of STIC;

FIG. 15 is a data plot showing an on-line pH titration from pH 5.0 to pH 8.0 of mAb A CEX elution at a titrant of 400mM Tris, pH 8.3 at a volume ratio of 0.1, showing that the target pH (horizontal straight line) is obtained throughout the elution peak (protein concentration, bell curve; conductivity, inclined line);

FIG. 16 shows two data plots (upper: Viresolve Shield; lower: Viresolve prefilter) to evaluate the effect of loading pH in batch mode with mABD using two different prefilters for Viresolve Pro, where the test conditions are protein concentration of 25g/L, 0.1M Naacetate 0.15M NaCl, 30psi constant pressure, and where the conditions are depicted as: pH 5.0 (diamonds), pH 6.5 (squares), pH 7.5 (triangles);

FIGS. 17a and 17b are graphs showing the virus filtration profile for the attached CEX-VF (VPF-VPro) operation with mAb C (0.1M Naacetate, pH 5) with the residence time of the intermediate buffer vessel: 17a)5 minutes, 17b)25 minutes, wherein a trend plot is shown for protein concentration (diamonds), conductivity (squares), pressure (triangles), permeability (Xs);

fig. 18 is a comparative graph showing VF permeability trends for batch versus ligation (CEX-VF) mode using VPF-VPro with mAb C (0.1M NaAcetate, pH 5), where the ligation data is shown as clustered along the diagonal, the batch data is shown as squares, circles and diamonds located below the clustered line, where the batch data is shown as the average permeability for each independent experiment: squares represent high pressure, circles represent low pressure, diamonds represent center points, and open shapes represent low salt conditions, while filled shapes represent high salt conditions;

FIG. 19 is a graph showing TFF permeate flux versus concentration for low salt and high salt conditions at a 20PSI TMP with mAb C (100mM acetic acid, pH 5, low salt 30mM NaCl with open symbols, or high salt 150mM NaCl with closed symbols) where the data points are symbolized, the lines indicate model matching, and the arrows indicate the feed cross flow ramp up required to maintain a constant permeate flux of 38LMH to the end of the ligation process (end of UF1 a);

FIG. 20 shows a table of an example of a connection process developed;

figure 21a is a data diagram showing the progression of mass during the entire mAb B-linked downstream process, with the following steps operating in sequence: CEX bind/elute Chromatography (CEX), HIC flow-through chromatography (HIC FT), Viral Filtration (VF), and TFF (UF1a), where the linked portion of the process ends at UF1 a; operating UF1b, DF and OC in discrete mode;

FIG. 21b is a data plot showing the trend of linked virus filtration, where the change in flow rate (diamonds) over the set point (dashed line) and hence the pressure drop across ViResolve Pro (triangles) is the result of automatic level control of the TFF retentate sump, and the decrease in VPro filter permeability (squares) corresponds to an increase in peak protein concentration (Xs) on the filter;

FIG. 21c is a data plot showing connected TFF trends (operating parameters shown in Table 2 of FIG. 20) where the feed cross flow rate (triangles) and TMP (squares) were increased during UF1a to maintain a constant permeate flow rate (. times.) and the VF flow rate setpoint and TFF permeate flow rate setpoint match and are represented by a horizontal dashed line slightly above 8LPM, the slight float in VF flow rate (+ s) and TFF tank level (Xs) being due to automated control; and

fig. 22 is a graph showing projection cell volumes for the B/E (left-most vertical bar in each three cluster), FT (middle vertical bar in each three cluster), and VF (right-most vertical bar in each three cluster) steps compared to the linking process using 100L buffer bins for mabs a-E operating in discrete mode.

Detailed Description

The present invention uses the following terms, the definitions of which are provided below:

and (3) filtering: pressure-driven separation processes use membranes to separate components in liquid solutions or suspensions based on size differences between the components.

Liquid supply: liquid solution or suspension entering the filter.

Filtrate: through one or more components of the film. Also known as permeate.

Retention solution: one or more components that do not pass through the membrane but are actually retained by the membrane.

Tangential Flow Filtration (TFF): in TFF, a liquid solution or suspension is pumped tangentially along the surface of the membrane. Also known as cross-flow filtration.

Single Pass Tangential Flow Filtration (SPTFF): one type of TFF is where the feed liquid is directed through the filtration device in a single pass without recirculation.

And (3) microfiltration: filtration to separate intact cells and relatively large cell debris/lysate from the rest of the components, such as colloidal materials, proteins (including products of interest), and salts. For this type of separation, the membrane pore size may be in the range of, for example, 0.05 μm to 1.0 μm. The filtrate or permeate from the microfiltration process may be referred to as a microfiltration harvest fluid.

And (3) ultrafiltration: for example, filtration to separate proteins (including products of interest) from, for example, relatively small peptides and buffer components, e.g., upon desalting or concentration. The membrane grade used for this type of separation can be expressed with nominal molecular weight limits and can be in the range of, for example, 1kD to 1000 kD.

And (3) filtration: filtration processes that may be performed in conjunction with other separation categories to enhance, for example, product yield or purity. The buffer was introduced into the recycle tank and the filtrate was removed from the unit operation.

Transmembrane pressure (TMP): TMP is the average applied pressure from the feed to the filtrate side of the membrane.

And (3) connecting: an upstream process and a downstream process are connected, wherein the downstream process is used concurrently with the upstream process. That is, the operations of the upstream and downstream processes overlap at least in time.

The present invention relates to various process control methods and systems for filters and filtration systems. First, a process control method and system for concentration of a microfiltration collected fluid using single pass tangential flow filtration with filtrate (permeate) flow rate control is described. Additionally, process control methods and systems for operation of an ultrafiltration element used concurrently with (i.e., connected to) one or more upstream unit operations are described herein.

As described above, microfiltration is used to separate cells and cell debris from the product of interest. In particular, the microfiltration element is positioned in line with the flow of the collected liquid from the bioreactor. The microfiltration element returns the cells and cell debris to the bioreactor while collecting the filtrate for further downstream processing.

Microfiltration may be combined with diafiltration to improve product yield. However, diafiltration increases the liquid volume of the filtrate collected from the microfiltration element. To achieve a product yield of greater than 80% to 90%, the liquid volume of the filtrate collected from the microfiltration element may be at least three times the working volume of the bioreactor. The large volume of liquid collected can limit the usefulness of diafiltration as scale increases.

To allow diafiltration to be used in conjunction with microfiltration to improve product yield for large scale operations, process control methods and systems for concentration of permeate from microfiltration elements (referred to herein as microfiltration collection fluids) are described herein. In particular, these process control methods and systems use Single Pass Tangential Flow Filtration (SPTFF) with permeate flow rate control.

During microfiltration operation in constant volume diafiltration mode, the product concentration is initially high due to accumulation of product in the bioreactor during the manufacturing stage. That is, the product concentration is initially high because there has been no removal of product to date and no buffer has been added as part of the diafiltration process. As the product passes through the microfiltration element and media is added as part of the diafiltration process, the product concentration in the bioreactor (and in the filtrate of the microfiltration element) will decrease. The varying product concentration will work with the use of SPTFF downstream of the concentrated microfiltration harvest fluid, as the conversion of the feed liquid to permeate via SPTFF depends on the feed liquid concentration as well as the trans-flow rate and trans-membrane pressure. Varying product concentrations in the microfiltration element filtrate will result in varying feed to permeate conversion in SPTFF.

In accordance with the present invention, Single Pass Tangential Flow Filtration (SPTFF) is used in conjunction with control systems and methods to achieve concentration of the microfiltration collected fluid.

With respect to hardware, fig. 1 shows a treatment system 50 comprising a microfiltration unit 51, an optional first tank 52 that receives filtrate from the microfiltration unit 51, a feed pump 54 (which according to other embodiments may be coupled directly to the microfiltration unit 51 and serve the dual function of removing filtrate from the microfiltration unit 51 and driving feed liquid through downstream elements such as SPTFF 56), SPTFF56, a permeate (or filtrate) pump 58, and a second tank 60 that contains retentate. Line 62 connects an outlet 64 of the first sump 52 to an inlet 66 of the feed pump 54, and line 68 connects an outlet 70 of the feed pump 54 with an inlet 72 of the SPTFF 56. Line 74 connects the retentate outlet 76 to the second tank 60, while line 78 connects the permeate outlet 80 to the inlet of the permeate pump 58. A back pressure control valve 82 may be provided in the line 74 between the retentate outlet 76 and the second sump 60. Lines 62, 68, 74, and 78 may further contain connectors, clamps, and other equipment not shown in fig. 1.

As also shown in fig. 1, a control system 120 is provided. The feed pump 54 and valve 82 may be manually set while the permeate pump 58 is controlled by the control system 120 according to the control method shown in fig. 2. According to other embodiments, the control system 120 may be coupled to the feed pump 54, the permeate pump 58, and the valve 82, and may be configured or adapted to control the feed pump 54, the permeate pump 58, and the valve 82.

According to some embodiments, the control system 120 may include one or more processors 122 and a memory 124, the memory 124 being coupled to the one or more processors 122. The one or more processors 122 may be programmed to control the permeate pump 58, and optionally the feed pump 54 and valves 82, according to the control method shown in fig. 2. The instructions executed by the one or more processors 122 may be stored on a memory 124, which memory 124 may include tangible, non-transitory computer-readable media or storage media, such as Read Only Memory (ROM) or Random Access Memory (RAM) in various forms (e.g., hard disk, optical/magnetic media, etc.).

Control systems and methods according to the present invention utilize a variable flow rate reduction factor (FRF) strategy to achieve a target Volume Reduction Factor (VRF). FRF is defined as the ratio of the feed liquid flow rate to the retentate flow rate (feed liquid flow rate/retentate flow rate). VRF is defined as the ratio of the cumulative feed volume to the cumulative retentate volume (feed volume/retentate volume). To achieve the desired target VRF with variable flow rate switching, the control system and method according to the present invention implements a permeate flow rate control strategy in which the FRF is varied during the acquisition. In particular, lower target FRFs are utilized when the product concentration is high (i.e., at the beginning of the collection process). In contrast, higher target FRFs are used when the product concentration is low. The target FRF was changed when the product concentration was changed from high to low.

According to the first embodiment of the present invention, the target FRF is changed in a series of stepwise changes. The permeate flow rate control strategy can be expressed as follows:

total VRF ═ Δ T_total/(Δt₁/FRF₁+…Δt_n/FRF_n) (equation 1)

Wherein total VRF is a cumulative volume reduction factor;

ΔT_totaltotal treatment time;

Δ t is the time interval of the step; and

FRF is the volume reduction factor of the step.

Fig. 2 illustrates an embodiment of a control method, denoted as control method 150. Method 150 begins at block 152 where feed pump 54 is set to operate at a desired bioreactor fill rate. The method 150 continues to block 154 where the backpressure control valve 82 is set to the desired backpressure of the SPTFF 56. It should be recognized that the actions at blocks 152, 154 may be performed sequentially or simultaneously. The method 150 continues to block 156 where the permeate pump 58 is set at a specified pump speed to achieve the target FRF. The method 150 then continues to block 158 where the permeate pump 58 is operated at the specified pump speed. The method proceeds to block 160 where a determination is made whether the operation of the permeate pump 58 should be adjusted to change the target FRF. If it is determined at block 160 that it is not time to change the permeate pump speed to cause the target FRF to change (and, with reference to the specific embodiment, increase), the method 150 returns to 158. If it is determined at block 160 that the pump speed should be changed, method 150 proceeds to block 162, where the pump speed is changed to achieve the new target FRF.

Fig. 3 illustrates an example of a FRF that can be implemented using a SPTFF system according to the present invention in conjunction with a stepped control system and method. It should be appreciated that the method is carried out in three steps over a 72 hour period of time. Each for a period of 24 hours of execution. According to the above discussion, the FRF for the first step is low when the product concentration is high, and the FRF for the third and last step is high when the product concentration is low. Specifically, the FRF for the first step is 2.1, the FRF for the second step is 2.7, and the FRF for the third step is 3.4. Using equation 1 above, the total VRF is thus 2.6.

While the example of fig. 3 includes three steps, it is recognized that a fewer or greater number of steps (e.g., two steps, four steps) may be used. In fact, fig. 4 shows an embodiment of the present invention, wherein an embodiment implementing the FRF in a stepwise manner is compared with an embodiment continuously changing the FRF. Further, while each step is performed over the same time period, the time period in which the permeate pump speed may be maintained to achieve the target FRF may be varied such that the first step may be longer than the subsequent steps, or vice versa. Furthermore, while the changes (in this case, increases) in the target FRFs are substantially equal in the example of fig. 3, it should be appreciated that the differences between the target FRFs of successive steps need not be substantially the same.

The target VRF may be obtained by sizing the membrane area according to the feed flow rate and specifying the FRF within the pressure limits of the system. Each step change in FRF can be specified to operate within a certain transmembrane pressure (TMP) window to provide the desired total VRF.

Having discussed process control systems and methods for concentrating microfiltration harvest fluid, other process control systems and methods for ultrafiltration and coupling processes may be discussed with reference to fig. 5 through 10. In particular, FIG. 5 shows a connected process system (with associated control system) in which the methods of FIGS. 6 through 10 may be implemented.

As discussed above, ultrafiltration is a separation process that uses membranes to separate products of interest (e.g., proteins) from, for example, smaller peptides and salts. In the case of ultrafiltration, the retentate is collected for possible further treatment, packaging etc., while the permeate or filtrate is removed. Ultrafiltration results in a concentrated product with lower salt content. Thus, ultrafiltration may also be referred to as a desalination process.

In a typical ultrafiltration process, such as for monoclonal antibodies (mabs), the ultrafiltration process is performed as a discrete unit operation performed in batch mode at a fixed feed cross-flow rate. The process is discrete in the sense that the units are not directly connected to an upstream or downstream process, but are actually operated in batch mode. The fixed cross-feed flow rate selected is the maximum cross-feed flow rate that is typically allowed by the system design to maximize process efficiency.

As product concentration increases, permeate flux decreases. This reduction is generally attributed to concentration polarization gradients. That is, as the filtration process progresses, a boundary layer of trapped substantially high concentration of species accumulates on or near the membrane surface. The boundary layer impedes the flow of material through the membrane and thus affects the production of permeate.

Indeed, if the ultrafiltration process is operated in batch mode with the liquid supply sump attached to the filter, the inlet flow rate from the liquid supply sump to the filter will typically be reduced to match the permeate flow rate, thereby maintaining a constant retentate volume per unit time. Since the ultrafiltration is operated as a discrete unit operation, no other unit operation is affected by this flow rate reduction.

However, fig. 5 shows a system 200 in which an ultrafiltration processing unit 202 is connected to an upstream processing unit 204 (e.g., a chromatography processing unit, a viral filtration processing unit). The ultrafiltration processing unit 202 contains a sump (or recirculation vessel) 206 into which the product of the upstream process is fed, a feed pump 208, a Tangential Flow Filter (TFF)210, and a back pressure valve 212. A line 214 connects an outlet 216 of the liquid supply sump 206 and an inlet 218 of the pump 208. Another line 220 is connected to an outlet 222 of pump 208 and an inlet 224 of filter 210. Another retentate return line 226 is connected to an outlet 228 of the filter 210 and the supply sump 206. The permeate exits ultrafiltration treatment unit 202 at permeate outlet 230 of filter 210.

When ultrafiltration treatment unit 202 is connected to an upstream process as in fig. 5, any reduction in permeate flow rate (i.e., at outlet 230) will have an effect on upstream operations. That is, it is typical to reduce the inlet flow rate to the filter to match the reduction in permeate flow rate. On the other hand, upstream processing operations, such as chromatography and viral filtration, are typically performed at a constant flow rate. If an upstream processing operation is connected to ultrafiltration processing unit 202, a solution must be provided to account for the operational differences between ultrafiltration processing unit 202 and upstream operation 204. In accordance with an embodiment of the present invention, a control system and method is needed to meet the need to run the upstream processes 204 (e.g., chromatography processing units) at a constant flow rate and to couple those processes (either directly or indirectly through a viral filtration processing unit) to an ultrafiltration processing unit 202 having a variable permeate flow rate.

As shown in fig. 5, a control system 240 may be provided. Control system 240 may be coupled to upstream process 204 and/or feed pump 208. The control system 240 may be configured or adapted to control the upstream process 204 and/or the pump 208 to implement one or more of the control methods described in fig. 6-10. The control system 240 may also be coupled to at least one sensor 246 from which the permeate flow rate may be determined.

According to some embodiments, control system 240 may include one or more processors 242 and memory 244, memory 244 coupled to the one or more processors 242. The one or more processors 242 may be programmed to control the upstream process 204 and the pump 208 in accordance with the control methods illustrated in one or more of fig. 6 through 10. Instructions executed by the one or more processors 242 may be stored on a memory 244, which memory 244 may include tangible, non-transitory computer-readable media or storage media, such as Read Only Memory (ROM) or Random Access Memory (RAM) in various forms (e.g., hard disk, optical/magnetic media, etc.).

According to a first method 250, illustrated in fig. 6 and referred to as a variable flow rate strategy, the control system 240 alters the operation of the upstream process 204. Specifically, the method 250 begins at block 252 where the control system 240 determines a permeate flow rate decrease, for example, in response to signals received from the sensor 246. The method 250 continues to block 254 where a change in the upstream process is determined based on the sensed decrease in permeate flow rate. In other words, the method 250 determines an appropriate response to the sensed decrease in permeate flow rate at block 254. For example, the determined change may be a scheduled decrease in the flow rate of the upstream process 204 (e.g., the flow rate of the chromatography processing unit) to maintain a constant retentate volume. According to other embodiments, the change may be a decrease calculated according to a formula that relates permeate flow rate to flow rate of the upstream process 204. The method 252 then controls the upstream process 204 at block 256 based on the changes determined at block 254.

According to a second method 260, shown in fig. 7 and referred to as a constant flow rate strategy, the control system 240 varies the operation of the pump 208. Specifically, the method 260 begins at block 262, where the control system 240 determines a permeate flow rate decrease, for example, in response to signals received from the sensor 246. The method 260 continues to block 264 where a change in operation of the pump 208 (i.e., an increase or decrease in flow rate at the outlet of the pump 208) is determined based on the sensed decrease in permeate flow rate. For example, the change may be a change in the feed cross flow rate that changes the permeate flow rate to a constant flow rate that matches the flow rate of the upstream process 204, which should also provide a constant volume in the tank 206. In this regard, it should be noted that permeate flux is primarily dependent on feed cross flow rate; wherein a higher cross-flow rate of feed liquid results in a higher mass exchange coefficient and thus higher permeate flux. The method 260 may then continue to block 266 where the control system 240 controls operation of the pump 208 according to the changes determined at block 264.

Another way to resolve the conflict may also be to allow the flow rate of the upstream process 204 to be mismatched with the flow rate of the permeate at the outlet 230. According to this approach, also referred to as a variable volume strategy, the sump 206 must be sufficiently sized to accommodate fluctuations (i.e., increases or decreases) in retentate volume caused by mismatches. Unlike the methods 250, 260 described in fig. 6 and 7, this method is not an active control method, but a passive control method.

Fig. 8 through 10 illustrate three additional control methods that may be implemented by the control system 240, where the control method of fig. 8 is referred to as a hybrid strategy, the control method of fig. 9 is referred to as a surge strategy (which is different from simply allowing a surge to occur in the sump 206), and the control method of fig. 10 is referred to as a high volume set point, fixed flux strategy.

According to the method 270 shown in FIG. 8, the method 270 initially uses the steps of the method 260. That is, the method 270 determines whether there has been a change in permeate flow rate at block 272, determines a change in the pump 208 at block 274, and implements the change at block 276. The method 270 then determines at block 278 whether a predetermined flow rate has been reached for operation of the pump 208, and the method 270 proceeds to blocks 280, 282, 284, where the method 270 determines whether there has been a decrease in permeate flow rate, determines a change to the upstream process 204, and implements the change. According to some embodiments, the predetermined flow rate may be a maximum system flow rate. Because the operation of the pump 208 is limited according to the predetermined flow rate (and in particular, the maximum system flow rate), the changes determined at block 282 and implemented at block 284 are reduced relative to the method 250. In addition, a fixed retentate volume is maintained, thereby fulfilling the requirements of a smaller sump.

According to the method 290 illustrated in fig. 9, the method 290 initially also uses the steps of the method 260. That is, the method 290 determines whether there has been a change in permeate flow rate at block 292, determines a change in the pump 208 at block 294, and implements the change at block 296. The method 290 then determines at block 298 whether a predetermined flow rate has been reached for operation of the pump 208, and the method 270 proceeds to block 300, wherein the sump volume is allowed to surge according to the passive method described above. According to some embodiments, the predetermined flow rate may be a maximum system flow rate. The method 290 has the advantage of allowing the upstream process 204 to continue operating at a constant flow rate.

According to the method 310 illustrated in fig. 10, the use of a larger sump 206 minimizes the concentration in the ultrafiltration treatment unit 202. As mentioned above, the product concentration is the main cause of the formation of the polarization gradient. By limiting the maximum concentration in unit 202 relative to a conventional TFF batch ultrafiltration treatment unit, the maximum permeation flux rate that can be achieved is increased relative to a conventional batch treatment unit. The result of the higher maximum permeate flux rate is that the upstream processing unit can be maintained at a constant flow rate required to optimize its performance.

Thus, according to the method 310 illustrated in fig. 10, the control system 240 alters the operation of the pump 208. Specifically, the method 310 begins at block 312, where the control system 240 determines a permeate flow rate decrease, for example, in response to signals received from the sensor 246. The method 310 continues to block 314 where a change in operation of the pump 208 is determined based on the sensed decrease in permeate flow rate. For example, the change may be a change in the cross-feed flow rate that changes the permeate flow rate to a constant flow rate that matches the flow rate of the upstream process 204, which should also provide a constant volume. However, the variation also depends on maintaining a volume of the sump 206 that is greater than the volume maintained by the method 260 in fig. 8. The method 310 may then continue to block 316 where the control system 240 controls operation of the pump 208 according to the changes determined at block 314.

It should further be appreciated that for the methods shown in fig. 6-10 to be used continuously throughout the operation of the connection system 200, the upstream processing unit 204 may not provide sufficient quality for each cycle of the processing unit 204. Indeed, the variable volume strategy described above may be used to cause the volume of the sump 206 to surge, thereby collecting and combining multiple connected cycles from the upstream processing unit 204.

It will be appreciated that the system and method according to the present invention may have one or more advantages over conventional techniques, as has been set forth above. Any one or more of these advantages may be present in a particular embodiment in accordance with the features of the present invention included in that embodiment. Other advantages not specifically described herein may also be present.

Experimental testing

By way of example, various advantages and advantages have been achieved through the following experimental activities. Specifically, the following description shows one experimental mAb downstream process connected from a purification column to a final Tangential Flow Filtration (TFF) step. A typical mAb platform procedure is depicted in fig. 11, starting with collection, followed by affinity chromatography for extracted protein a, a low pH virus inactivation step, up to two additional chromatography steps for refinement, a Virus Filtration (VF) step, and finally a TFF step for formulation that performs ultrafiltration/diafiltration (UF/DF). The intermediate cell between the polishing column (in this case the bind/elute (B/E) and Flow Through (FT) steps and TFF) is typically the most dilute cell and therefore has the greatest possibility of its volume exceeding the size of the cell tank. By connecting the B/E column, FT column, VF and TFF steps, three large tanks can be reduced or eliminated. This document reports the proposed configuration of the connection process and flow rate control strategy to achieve connectivity for unit operations. A detailed description of the development of how to implement the connection process and the consideration of the additional process monitoring requirements is provided.

Method and material

Material

Five mAb products (mAb a, mAb B, mAb C, mAb D, mAb e) were produced by standard CHO cell culture methods.

The chromatography resin used on a small scale and a large scale comprisesEMD SO₃ ^-(EMD Millipore, Billerica, Mass.) and Phenyl Sepharose^TM6Fast Flow High Sub (GEHealthcare, Piscataway, N.J.). Small-scale column chromatography was performed at 1.15cm of Vantage from EMD Millipore^TML laboratory columns and Large Scale chromatography in Axichrom 60 or 80cm columns from general electro medical group. AEX membranes Sartobind were used in nano (1mL) or 10 "(180 mL) sizes(sartorius stedim, germany gottingen).Prefilter(5cm²、0.55m²And 1.1m²)、ViresolveShield(3.1cm²And 0.51m²)、Viresolve Pro(3.1cm²And 0.51m²) And330kDa(0.0088m²and 1.14m²) Filters were purchased from EMD Millipore.

AKTAexplorer in general electric medical treatment group^TMSmall scale chromatography and ligation process experiments were performed on 100 systems. For the connection process experiments, a number of AKTAs were connected to each other via a remote connection behind the P-900 pump to allow auxiliary input and output signals to be passed between the instruments. By passing(SciLog, Madison, Wis.) pressure sensors and pressure monitors perform pressure monitoring of a small-scale prefilter and a virus filter. EMD Millipore companyA stirred tank (50mL) was used as a buffer vessel; the buffer vessel was used without the use of a top cover and film, so it could be operated open to atmospheric pressure as a continuous stirred tank placed on a magnetic stirring plate.

Small scale discrete virus filtration experiments were performed by constant pressure equipment comprising a pressure regulator, a pressure vessel (300 or 600mL of polycarbonate), a pressure gauge, a scale connected in series to a computer for data acquisition, and a compressed air supply. In AKTAcrossflow^TMSmall scale TFF experiments were performed on the system.

Large-scale runs were performed on custom automated chromatography, virus filtration, and TFF pallets. The chromatography pallet contains a three-stage pump for gradient and dilution capability, with online monitoring of pressure, flow rate, pH, conductivity and UV. The pallet is also equipped with diverter valves and pumps for collecting spurious product pool samples of the product pools. The virus filtration pallet contains supports for the prefilter and virus filter and monitors pressure, flow rate, pH, conductivity, UV online. The TFF pallet comprises a 200L retention liquid tank, a diaphragm pump for system liquid supply, a peristaltic pump for diafiltration buffer solution, an automatic TMP control valve, and pressure, flow rate, pH value, conductivity and liquid level sensing on-line monitoring on the retention liquid tank. The buffer tank is equipped with level sensing.

Method of producing a composite material

Sartobind STIC thin film chromatography

Sartobind STIC experiments were performed on AKTAexplorer with a bypass mixer. An in-line filter (Minisart from Sartorius, 0.2 μm) was used upstream of the STIC membrane to prevent pressure build-up by filtering out particles that may be generated by the AKTA pump. The loaded material is filtered, recovered by low pH viral inactivation (FVIP) or CEX recovery. The product pool is collected as a single main fraction or as multiple fractions during the flow-through and wash processes. The assays performed on the STIC cell contained CHOp ELISA (for CHO host cell proteins), DNA QPCR and concentration UV a 280.

On-line pH titration

The CEX elution fraction was formed from AKTAexplorer running the entire CEX sequence of operations by an automated program. Each fraction was then used to manually screen pH titrants. After finding the appropriate titrant, an experiment with two aktaexplorers was performed to confirm that the selected titrant can provide an accurate on-line pH titration to the target. The first AKTA run CEX and its eluate was collected into a beaker as a buffer vessel for a residence time of 5 minutes. The second AKTA was loaded with product from the beaker by pump a and titrant by pump B. The two streams were mixed in a mixer and the pH was measured by an in-line pH probe on the second AKTA. The second AKTA was also fractionated and the pH of each fraction was verified using Orion two-star off-line pH values (seemer Scientific, waltham, ma).

Virus filtration

Virus filtration tests were performed in either a discrete or connected mode, with a prefilter and a virus filter placed in series. Discrete tests were performed using the constant pressure equipment described in the materials section, and the filtered volume was collected over time by a homogeneous feed liquid loaded onto the filter. The ligation test was performed using the attached AKTAexplorer apparatus, where the prefilter and virus filter on one AKTA was ligated to the previous chromatography step on the other AKTA, with a buffer vessel between each step. The buffer vessel is operated at a fixed residence time and hence volume, typically 5 to 7 minutes. The Unicorn method was programmed to allow automated signaling to begin between AKTAs and end loading and elution. The AKTA B-pump was used for in-line titration, conditioning or dilution, mixing with the feed stream loaded on the AKTA a-pump. Since small scale equipment uses fixed column diameters and filter areas based on commercial availability, to achieve target loading and flow rates for the intermediate link unit operation similar to large scale operation, split streams were employed, where AKTA sample pumps were used after the chromatography step and before the buffer vessel. This diversion enables control of the flow rate for subsequent unit operations and also controls batch loading since the mass and flow rate are correlated during the connection. The material collected from the shunts is used to create a pool of dummy products for evaluating the yield and contaminant removal performance of each connection step.

TFF flux offset

By applying a cross flow rate in the protein concentration range (typically 10 to 80g/L), the feed cross flow rate range (1 to 6L/min/m)²Or LMM) and TMP range (10 to 25psi) flux offset experiments were performed on AKTAcrossflow to empirically determine the retentate membrane model parameters (see equation below). Flux shifts were performed using proteins in salt buffer from previous unit operations to best simulate performance of the connected UF stage (UF1 a). Product was allowed to recirculate at each concentration, TMP and feed cross flow rate until stable permeate flux and Delta pressure (feed-retentate) were obtained. Data points where the osmotic pressure was greater than 4psi are not included in the analysis. After each set of TMP measurements, the membranes were depolarized by recirculation with the permeate outlet closed. Then, C was compared for the flux (J)_bThis data was plotted against the natural logarithm of (protein concentration tested).

Filter sizing

The virus filter area sizing is dependent on the flow rate and maximum allowable operating pressure of the connection process. The lowest observed virus filter permeability (filtration flux normalized for pressure drop) occurs at the peak of the protein concentration. This lowest observed permeability (k)_VF,min) Can be used to set the maximum pressure limit (P)_VF,max) Maximum flux (J) for internal operation_VF,max) Such as J_VF,max＝k_VF,min×P_VF,maxDescribed herein. The desired filter area (A) can be determined by equation 1_VF) Wherein Q is_VFIs the process flow rate.

Regarding TFF modeling, Ng P, Lundblad J, Mitra G, 1976, "Optimization of solute Separation by diafiltration (Optimization by filtration)," Separation Science (Separation Science 11 (5)): the study at page 499-502 describes TFF permeation flux based on a retentate membrane model. The retentate membrane model may be modified to include a feed cross flow rate (k-k) related to the mass exchange coefficient_ovⁿ) Wherein k is_oIs an empirical constant, v is the cross-flow velocity of the feed, and n is a power term for cross-flow velocity correlation of the feed. This modified retentate membrane model is shown in equation 2, where J_TFFIs the permeation flux, C_wIs the concentration of protein near the membrane wall, and C_bIs the volumetric protein concentration.

Parameters derived from flux offset and combined with input parameters obtained from the process are used to determine the flux offset by solving for C_bEquation 2 of (a) determines the optimal final concentration of the target at the end of the connected portion of the treatment (end of UF1 a). The required permeate flux was determined from the process flow rate at the inlet and the TFF filter area. The cross feed flow rate is set at the upper performance limit of the system and the film, typically 6 LMM. The target retentate tank level set point can then be determined based on the total expected mass m for the process using equation 3.

Results

Design of connection process and flow rate control

The overall design of the connection system is similar to the discrete system in that the main components and functions of the standard cell operation remain largely the same. In the connected system, the large tank is replaced by a small buffer vessel with a short residence time (typically 5 to 7 minutes) which is used as a frac between unit operations (fig. 12). Batch dilution and titration were replaced by on-line addition. The key to designing process control of a linked system is determining how to manage the flow rate differences between unit operations. In the final TFF step, the product was initially concentrated to the desired titration endpoint for performing diafiltration. Managing the reduction in permeate flux that occurs as product concentration increases as mass is added to the retentate tank presents challenges; this reduction in permeate flux is attributed to concentration polarization of the membrane. For discrete fed-batch TFF, the inlet flow rate of the retentate tank will be reduced to match the permeate flow rate to maintain a constant retentate volume. Since the TFF step is operated in a discrete unit operation, there is no effect on any other unit operation due to such a reduction in flow rate. However, during the connection process, the inlet stream is directly connected to the previous upstream operation, and any reduction in permeate flow rate will cause a difference in flow rate. For a process sequence that connects a constant flow chromatography step with a variable permeation flux TFF step, either directly or through a viral filtration step, there is a need to actively manage the mismatched flow rates. This can be done during TFF operation using three different strategies: 1) variable flow rate strategy, 2) constant flow rate strategy, or 3) surge strategy (fig. 13, table 1). It should be noted that the initial concentration is the only stage connected in the final TFF step, since the diafiltration and the rest of the final concentration step can be performed in a standard discrete process once the product quality is fully contained in the retentate tank.

In the variable flow rate strategy, TFF was operated similar to a discrete batch feed operation, in that permeate flux decreased as mass accumulated in the retentate tank. To balance the system flow rates, the flow rate of the upstream unit operation is also reduced to match the permeate flux. This maintains a constant retentate volume, but results in a variable flow rate for the chromatography step. The magnitude of the flow rate variation may result in at least a two-fold reduction, which can potentially impact the performance of the chromatography step.

An alternative is a constant flow strategy, in which both the permeate flow rate and the inlet flow rate are maintained at constant values to maintain both a constant retentate liquid volume and a constant flow rate throughout the previous unit operation. To achieve constant permeate flow rate and inlet flow rate, a novel strategy was developed to actively control permeate flux using both TFF feed cross flow rate and transmembrane pressure (TMP). TFF feed cross flow rates can directly affect mass exchange rates and hence cross-membrane flux. Transmembrane pressure (TMP) also controls osmotic flux, but this parameter has reduced control at higher protein concentrations and higher TMP when flux limited mode is reached. In this control strategy, a lower cross flow rate and TMP are used from the beginning of the connection when the product concentration in the sump is low, and these two parameters are gradually increased as the product concentration increases to maintain a constant permeate flow rate. This method was developed into an automated control system that simultaneously modulates the input parameters of both the feed cross flow rate and the TMP to achieve a constant permeate outlet flow rate, thus enabling a connected process system to operate without flow rate differentiation.

The final control strategy, i.e. the surge strategy, can be described almost as the absence of active flow rate control. In this strategy, the inlet flow rate is still maintained at a constant rate as the permeate flux decreases, which in turn induces a volume surge in the TFF retentate tank. In practice, the TFF system will exhibit some self-modulation because as the volume in the sump ramps up, the rate of increase of product concentration will slow, as flux decreases.

These three described control strategies represent available options for flow rate control, although these strategies can ultimately be used in combination to achieve an optimal overall process that balances the requirements for membrane area and process time, the requirements for flow rate reduction affecting previous unit operations, and the requirements for retentate container volume. The following section describes the development of a connection process using a constant flow rate strategy, with emphasis on describing aspects and parameters specific to the connection process. This strategy was chosen for its ease of operation and process development, as it maintained a constant flow rate for the chromatography and viral filtration steps, and minimized the dynamic effects that need to be studied.

Development of connection procedures

Developing a process to connect two polishing columns, virus filtration and TFF requires additional considerations than developing these unit operations individually. These considerations include: 1) estimating the effect of elution from the B/E column on the subsequent steps; 2) developing an on-line pH titration method when it is desired to operate subsequent steps at a pH different from the pH of the B/E cell; 3) a virus filtration step driven by a variable feed solution composition development flow rate; 4) the TFF step was developed at constant permeate flux during the ligation process.

Development of the first chromatography step

Since the first step in the ligation process chain presents a uniform load, the filtered virus is inactivated in the protein a product pool and thus can be independently developed as a discrete process and will not be discussed in detail herein. However, there are two important considerations for the connection process. First, when a B/E step with gradient elution (e.g., CEX) is used as the first step of the ligation process, all subsequent steps experience a product concentration peak resulting from the first step elution as well as a salt concentration gradient. Depending on the maximum product concentration obtained, such concentration peaks can present downstream challenges, particularly for the virus filtration step. To mitigate the effect of high peak concentrations on subsequent steps, a shallower salt elution gradient may be employed. This will reduce the peak concentration and allow the product to flow through the remaining steps with acceptable back pressure. Second, since all steps are connected, the volumetric flow rate of the first step elution needs to be optimized based on the performance of the remaining steps.

Development of the second chromatography step

The second step, which may be a resin-based or membrane-based chromatography, illustrated in the schematic connection process diagram is operated in flow-through mode. This second step is typically the third and last chromatography step for the entire downstream process, however, this step may not be required when the dual column process exhibits sufficient impurity and virus removal capacity. The purpose of this step for a typical mAb purification process is to remove host cell proteins and potentially further reduce High Molecular Weight (HMW) and DNA. When this flow-through step is connected to the B/E step as the first step, the feed is no longer as homogeneous as it would be operated in discrete mode, but is dynamic in terms of protein concentration and conductivity. The conductivity of the feed stream in the flow-through step increased during the fill due to the previous salt gradient elution and reached a maximum at the end of the fill. Because of this, it is important to select a resin or adsorption film that maintains a stable impurity removal rate over a wide range of conductivities; AEX membrane STIC chromatography is an example of a salt-tolerant adsorption matrix. To estimate the effect of load conductivity on host cell protein removal, several discrete flow-through experiments of load conductivity changes were sufficient to evaluate the effect. FIG. 14a compares the effect of load conductivity on CHOp removal when performing the flow-through STIC step. The results indicate that conductivity does not play a significant role for CHOp removal, as expected based on the high salt tolerance of the ligands optimized for this device. Thus, the eluate from the first step can be supplied directly into the second connection step without dilution.

To remove host cell proteins more efficiently, for example for the AEX FT step, it may be necessary to operate the flow-through step at a higher pH than the pH of the B/E step. Several different pH scout experiments were required to find the optimum operating pH for this flow-through step. Figure 14b shows that host cell protein removal via AEX membrane chromatography at higher pH values provides better clearance. With respect to this example, since all four tested pH values provided acceptable clearance, pH 5 was chosen for STIC operation, which has the advantage that a titration step can be avoided during the ligation process. Of course, for cases where a pH titration is required to achieve the desired host cell protein removal, an on-line pH titration method is required.

Development of on-line pH titration step

Titration of the pH of the intermediate product pool is required when the previous step uses a different operable pH than the subsequent step. In the discrete mode, the pH titration can be easily performed by adding a specified amount of titrant to the homogenous product pool to achieve the target pH. However, in the case of the linking process, an in-line pH titration is required to change the pH of the product stream from the previous step, as the product is continuously charged to the next step. The product stream potentially requiring pH titration during the ligation process is the feed stream for the flow-through or viral filtration step, sometimes the load for the UF step. If the pallet or system for each step has a minimum of a dual pump design for delivering both the feed and titrant streams simultaneously, the in-line pH titration of the feed streams for the flow-through and virus filtration steps can be adjusted without additional pumps, followed by mixing via a passive mixer. Additional pumps may be required to deliver titrant to the TFF retentate tank when the UF load calls for titration.

Regardless of the location of the pH titration on the introduction line, changes in protein concentration and conductivity in the eluate from the binding and elution steps need to be taken into account when selecting the titrant. Furthermore, the process and system design is simplified when the titrant is introduced into the product stream at a constant titrant to product volume ratio. This volume ratio should be low to avoid excessive dilution of the product stream, but high enough in the linear range of pump flow rates. Based on this, a volume or flow rate ratio of 0.1 to 0.2 is generally recommended.

The online pH titration development began with an offline pH titration of the eluted multiple fractions across the binding and elution steps. The product concentration and pH of the titrated fractions were screened to ensure that each fraction reached the target pH with the addition of titrant at the same volume ratio. After confirmation of titrants, the results were verified using a laboratory scale ligation run. FIG. 15 shows an on-line pH titration from pH 5.0 to pH 8.0 of CEX elution performed at a volume ratio of 0.1 at pH 8.3 with a titrant of 400mM Tris. After on-line titration, the product stream is subjected to fractional distillation and each fraction is analyzed using an off-line pH probe. As shown in fig. 15, the target pH was obtained throughout the elution peak.

Development of Virus filtration procedure

Preliminary development of the linked viral filtration step is similar to discrete step development in that molecular and solution properties drive the selection of appropriate viral filters and prefilters and dictate the hydraulic permeability of the membrane. Since the virus filter is connected to a previous unit operation that specifies the flow rate through the filter, it is advantageous to select a virus filter with a high membrane permeability to reduce the required membrane area. In addition, the virus filter must be able to operate effectively when exposed to variable pressures and feed solution compositions that vary over time in both product concentration and conductivity. Here, a Viresolve Pro (VPro) filter is used as an example. This filter has high membrane permeability and generally exhibits stable operation despite molecular, feed composition and pressure variations, especially when a prefilter is used. Conventional prefilters include depth filters and charge-based prefilters (Ng P, Lundblad J, Mitra G, 1976), "Optimization of solute separation by diafiltration" (separation science) 11(5) (499 page 502; Brown A, Bechtel C, Bill J, Liu H, Liu J, McDonald D, Pais, Radhamohan A, Renslow R, Thayer B, Yoghe S, Dod C, 2010), "Pre-filtering with adsorbed ion exchange membranes" used to increase the throughput of monoclonal antibody small virus filters (biological extraction) 627 and Bioextraction technology (106) (Bioextraction) for biological extraction and biological extraction).

Batch filtration experiments using uniform liquid feed can provide a relative performance comparison between prefilters having different adsorption characteristics. In addition, batch experiments can be used to screen optimal pH settings for viral filter loading. Figure 16 shows the effect of pH on VPro flux when two different pre-filters were used: 1) a Viresolve Prefilter (VPF) consisting of diatomaceous earth with a charged binder; and 2) ViResolve Shield, consisting of a negatively charged membrane. For mAb D, the negatively charged Shield prefilter showed better performance at low pH values, as expected based on the high pI of the mAb, or more precisely, the mAb polymeric impurities and the interacting cation exchange mechanism. In contrast, the VPF prefilter showed better performance at high pH values. This may indicate a more hydrophobic interaction mechanism at a pH closer to the pI of the molecule and under high salt solution conditions. This example illustrates the advantage of performing a batch study to evaluate the relative performance of the prefilter and pH conditions.

To evaluate the performance of the virus filter for the ligation process, two different approaches may be considered. As in the previous example, experiments can be conducted on a virus filter alone in batch mode; this may be accomplished by forming a plurality of liquid feed materials with varying product and salt concentrations. These experiments can be performed as a design of experiments (DoE) to study the relative effect of protein concentration, salt concentration and even pressure or flow rate on membrane performance. Ranges can be selected to estimate the end of the product and salt concentrations observed from previous chromatography steps and to support the range of pressures experienced by the virus filter. A second method for estimating the connection performance is to simulate the actual connection process with a scaled down system. Such a method would produce a representative time-variable feed solution that would alter the protein and salt concentrations of the previous chromatography step, which would be loaded directly onto the virus filter. An example of a run-on process with CEX gradient elution connected to VPro with a prefilter is shown in fig. 17 a. Changes in protein concentration and conductivity across the virus filter are the result of elution from the CEX gradient, with the plateau at the end of the loading process due to the emptying of the buffer vessel. As the run was operated at a constant flow rate, an increase in protein concentration resulted in an increase in pressure on the virus filter, which corresponded to an initial decrease in membrane permeability and a subsequent restoration of permeability as protein concentration decreased. The final permeability is similar to the starting permeability, which means that fouling is minimal during operation.

The results of comparing virus filter performance in the link mode and batch mode are shown in fig. 18. Both sets of experiments were performed with mAb C using VPF and VPro filters. Batch factorial design experiments with 10 runs (repeated with 2 center points) were performed at protein concentrations of 5 or 40g/L, sodium chloride concentrations of 0 or 250mM, and pressures of 15 or 45psi (with repeat center points at the middle points of those parameters). The ligation experiment performed was similar to the experiment shown in fig. 17a, where the transformed data showed VPro permeability dependent on the instantaneous protein concentration. The batch conditions were chosen to cover the range conditions present in the ligation process (FIG. 17 a). Batch results show that salt concentration and pressure within the tested range have little effect on permeability, while increasing protein concentration shows a clear tendency to decrease permeability. The results from the connected experiments followed similar permeability trends as the batch experiments, however, the filter permeability in batch mode was overall lower than in connected mode. These results are not unexpected as long as the simultaneous operation of the CEX and VF steps minimizes the duration of the CEX cell maintenance and thus the formation of viral filter fouling components. These results indicate that batch experiments can provide a directed trend of relative impact on operating parameters, as well as a preliminary indication of the minimum expected membrane permeability depending on protein concentration. Finally, a representative linked running process should be used for the final process standard.

Once the hydraulic membrane permeability characteristics of the virus filter have been determined, the virus filter area size can be set appropriately for the connection process. The flow rate through the virus filter is predetermined by the flow rate set point of the previous chromatography step. Since the mode of operation is a constant flow rate, the viral filter is sized based on maintaining the feed pressure below a specified maximum limit. The limits may be specified by a virus filter, a prefilter, or even an operating system. For example, the maximum pressure limit for a VPro filter set by the manufacturer is 60psi and the maximum pressure limit for a VPF is 50psi, so an operating pressure limit of 40 to 45psi may need to be imposed on the virus filter in order to meet the prefilter limits. Experiments performed in batch or linked mode may provide the minimum expected permeability based on the maximum expected protein concentration. Given the inputs of flow rate, maximum pressure, and minimum permeability, the virus filter area can be calculated using equation 1. The virus filter sizing for the various connection procedures is shown in table 2 of fig. 20.

Stability was assessed during discrete virus filtration by estimating the change in feed solution composition over the normal operating range. One parameter that may affect the distribution of the feed liquid loaded onto the virus filter during the connection process is the residence time of the buffer vessel before the virus filter. Minimizing the buffer vessel residence time will allow the prior chromatographic elution profile to be transferred almost directly to the virus filter. In contrast, increasing the buffer vessel residence time to the maximum will pool the entire chromatographic elution pool and thus essentially render the virus filtration step a discrete operation. Experiments were performed to compare the buffer vessel residence times of 5 minutes and 25 minutes (fig. 17a and 17b, respectively). As expected, short residence times correspond to the largest observed peak variation, while long residence times produce a relatively uniform pool.

Development of a connected tangential flow filtration step

As described in the introduction, one control strategy that can be used to connect the previous downstream unit operation to the Tangential Flow Filtration (TFF) step is a constant flow rate strategy, where the TFF retentate volume and permeate flux are maintained constant throughout the connection operation. Mass accumulates in the TFF retentate fluid bath during the ligation process and the highest protein concentration is reached when all mass is in the TFF retentate fluid bath at the end of the ligation process. To maintain a constant permeate flux at the end of the ligation process, the TFF retentate volume setpoint and membrane area need to be specified to accommodate the inlet flow rate of the ligation and the highest expected protein concentration. A laboratory scale flux offset study was performed to map TFF permeate flux responses to varying feed cross-flow rates, transmembrane pressure (TMP) and feed concentration and to fit model parameters to the retentate membrane model (equation 2). This model can be used to calculate specified parameters of the connection process.

An example flux shift dataset for mAb C in saline buffer from a previous unit operation is shown in fig. 19. With low salt (30mM NaC)l) and high salt (150mM NaCl) conditions both performed flux shifts, estimating the effect of salt concentration changes due to previous CEX salt gradient elution. For this flux offset study, a concentration range of 9-88g/L was targeted, with TMP ranging from 5-20psi and feed cross flow rate ranging from 1-6 LMM; the 20psi TMP data is shown in figure 19. For this and subsequent examples a Pellicon 330 kDa regenerated cellulose membrane was used. The results show similar performance between the high salt condition and the low salt condition, indicating that the performance of the TFF film is minimally affected by the variable salt concentration. This same trend has been observed for multiple molecules (data not shown). The permeate flux model parameters based on the high salt flux offset data in fig. 19 are: k is a radical of_o＝8.47、C_w175 and n 0.73. For the connection process, equations 2 and 3 and model parameters can be used to calculate the TFF retentate volume setpoint based on known inputs of permeate flux (connection inlet flow rate/TFF area), the maximum feed cross flow rate based on pump or system pressure limits, and the maximum expected mass. As described in the introductory part, the connected TFF process is started as soon as the hold-up liquid volume set-point is reached. The feed cross-flow rate and TMP are initially at low set points in order to achieve a constant permeate flux target at low protein concentration. As mass accumulates and protein concentration increases, feed cross flow rate and TMP ramps up to maintain a constant flux target, with the maximum feed cross flow rate maintained below the system limit. The superimposed arrows in FIG. 19 conceptually illustrate the ramp-up of cross-feed flow rate; at a constant permeate flux of 38LMH, the feed cross flow rate would start at about 120LMH and end at about 300LMH at the end of the connection process (end of UF1 a). The target end values for UF1a protein concentration for various ligation runs calculated using the retention membrane model parameters for each molecular run (not shown) and equation 2 are shown in table 2 of fig. 20.

Once the initial fill volume parameters for the TFF step are determined, the remainder of the unit operation development (e.g., diafiltration and over-concentration/product recovery steps) is the same as for the development of a standard batch TFF process, and therefore will not be discussed here. The process stability of the linking part of the TFF step can be evaluated in several ways. The effect of changes in expected mass or protein concentration, feed cross flow rate, TMP and inlet flow rate can be studied via sensitivity analysis using changes in model fitting parameters and input conditions. Generally, a safety factor should be used to allow for variations in input conditions and still maintain a constant permeate flux, i.e., setting a more conservative or higher retentate volume set point criterion. Permeate flux can also be affected by the inherent membrane permeability and temperature of operation; experiments can be performed with respect to these input parameters.

Process monitoring

Unit operation during the connection process requires additional process monitoring compared to discrete mode to facilitate unit operation transitions, assist in pressure control, and provide the necessary information about the performance of the operation itself. Buffer tank level monitoring provides important transient signals that are communicated to the respective unit operations in real time. For example, when the post-CEX chromatography buffer tank level reaches its predetermined value, the control system sends this signal to the flow-through chromatography pallet to begin loading from the buffer tank. When the post-CEX chromatography buffer tank level reaches zero, the loading phase on the circulation pallet stops and the washing phase begins. For the virus filtration step, operation was performed at constant flow rate and filter inlet pressure was monitored. The inlet pressure fluctuates as the peak concentration of product passes through the virus filter. If the maximum pressure limit is reached, the control system is triggered to reduce the virus filtration flow rate and allow the liquid level of the flow-through buffer tank to increase. In the TFF step, permeate flow rate is measured during the connection process by a flow meter that not only provides flux information, but also communicates with the control system to maintain permeate flux at a preset value by adjusting the feed cross flow rate and TMP. In addition, the control system monitors the liquid level of the TFF retentate tank and maintains a constant volume by adjusting the inlet or viral filtration step flow rate.

Step yield information for a discrete process is typically obtained by measuring the product concentration throughout a uniform product pool and comparing the concentration to a uniform liquid supply and corresponding volume. Since simultaneous operation of the connected units does not allow collecting the whole cell, a small partial flow is diverted from the main product flow during cell collection. This pseudo product pool then provides samples for concentration measurements and product quality determinations. Yield information can also be obtained in real time on pallets by integrating the UV a280nm or a300nm signals and using experimentally determined product-specific extinction coefficients. The UV integration method can be used on the VF step immediately before the TFF step to calculate the mass accumulated in the TFF retentate tank at the end of the connected process. This accumulated mass corresponds to the TFF loading mass when operating in discrete mode, an important parameter for determining the volume level of the retentate tank for diafiltration and over-concentration.

Large scale performance

As described in the previous section, each unit operation in the connection process was developed primarily in a discrete mode, and then connected together on a laboratory scale for testing and further optimization. The process is then scaled up and transferred to a pilot plant for validation and validation. Table 2 lists the operating parameters for 5 different molecules using the connection procedure CEX-aex (ft) -VF-UF and its variants that have been successfully performed at the pilot plant. The buffer tank is located between the chromatography step and the front of VF, has a size of 100L, and operates with a residence time set point of 5 to 7 minutes. The yields listed in table 2 of fig. 20 have been demonstrated on a pilot scale and are similar to the discrete mode of operation (discrete data not shown).

An example of the operable trend for the ligation downstream process (CEX-hic (ft) -VF-UF) of mAb B is shown in fig. 21. Fig. 21a shows the time progression of the product quality during the connected downstream process over 70 minutes. The remainder of the UF step is performed in a discrete mode, and during this time other unit operations complete their cleanup and balancing steps in preparation for the next connected cycle. The concentration profiles of the hic (ft) and VF steps correspond to the elution peaks of the CEX step, with the on-line titration dilution at the hic (ft) step due to loading having a lower concentration and the line hold up and buffer tank mixing volume at the VF step having a slightly lower concentration. The second half of the concentration profile in the HIC (FT) and VF steps flattens out; this represents the operating phase when the previous connected unit operation has been completed and the buffer tank is empty.

Fig. 21b shows the connected VF profile, including flow rate set point, flow rate, permeability, Vpro pressure drop (dP), and protein concentration. The VF flow rate was set at 8.1L/min, corresponding to the flow rate of the hic (ft) step, however once the UF step was started, the VF flow rate was allowed to vary from its set value to maintain a constant UF retentate sump level (fig. 21 b). The pressure drop across VF corresponds to the flow rate change, but the permeability profile of the filter is relatively flat when the flow rate is normalized by the pressure drop. There was a slight decrease in permeability as the protein concentration on the filter increased, and then it returned to the original permeability as the protein concentration decreased.

Fig. 21c shows the connected UF1a trend, as well as the UF1b (batch concentration to DF setpoint) and the discrete trend of the DF stage of operation. As previously described, the VF flow rate is varied around its setpoint to maintain the retentate tank level setpoint; the result of this process control can be observed in fig. 21 c. The UF permeate flow rate setpoint is the same as the VF flow rate setpoint, where the feed cross flow rate and TMP control the UF permeate flow rate. Both the feed cross flow rate and TMP were observed to increase gradually as the protein concentration in the sump increased and the control strategy was able to maintain the permeate flow rate at its set point.

As discussed in the introduction, one of the main advantages of connecting the downstream unit operations is the reduction of the volume of the intermediate tank and thus the reduction of the factory footprint. Fig. 22 shows a comparison of the sump volume requirements for the discrete process versus the sump volume requirements for the connection process for each of the processes described in table 2 of fig. 20. Since the manufacturing plant will need to design a sump size for the maximum expected sump volume, a discrete process will require at least 1000L of storage sump. This is significantly greater than a 100L buffer tank operating at a 5 to 7 minute residence time for the coupling process. In addition, because the VF reservoir can be connected directly to the UF retentate reservoir, the VF reservoir is completely eliminated during the connection.

Discussion and conclusions

The above experimental study outlines the concept of downstream processes linked from the purification chromatography step by the final TFF step and confirms its successful performance on a pilot scale. Multiple flow rate control strategies can be used to manage the flow rate differential between unit operations (or rather, chromatography steps) and the variable permeate flow rate of the TFF step. A constant flow rate strategy is proposed as a means of maintaining a constant TFF permeate flow rate and thus a constant flow rate for the chromatography and viral filtration steps. This minimizes the number of dynamic effects that need to be studied during the development of the connected process. This control strategy also creates constant buffer tank and TFF retentate tank volumes throughout the connected operations, which allows for simpler process, equipment and automation design.

The development of connected downstream processes is similar in many respects to the development of discrete processes. Resin selection, virus and prefilter selection, and loading conditions such as pH, conductivity, and product concentration can all be studied by standard batch experiments. However, there are a number of unique aspects to consider in the development of the connection process. As an example, when the first linked step is a B/E chromatography step with salt gradient elution, a shallower gradient slope may be beneficial for subsequent unit operations such as virus filtration, and the like, to manage the peak product concentration conducted through the process. A shallower gradient slope may have an advantage of increased selectivity for impurity separation over the B/E chromatography step; there is flexibility in choosing the gradient slope based on process requirements rather than constraints on sump volume limitations. The first chromatography step is also of importance in setting the flow rate of the entire connected process and may therefore need to be optimised to achieve a more economical sizing of the filtration step. For the intermediate chromatography step, the effect of gradient elution needs to be evaluated. Experiments should be performed to investigate the effect of conductivity on the performance of the procedure. Although data will not be presented herein, additional experiments may be performed to investigate the effect of changes in product concentration and impurity profile caused by gradient elution. Operable pH values can also be screened and in the event of pH changes between unit operations, on-line pH titration can be developed and implemented for the connected process.

The examples highlighted here illustrate the approach used to develop the connected filtering step. Once the prefilter and viral filter are selected and the feed conditions are determined, relatively few experiments of the connection process are required to determine filter sizing requirements and evaluate the stability of the viral filter against changes in the feed composition. For the TFF step, the development and implementation of an automated control strategy is necessary to manage constant permeate flow rate operation, however, the development of the step can be accomplished in large part by discrete experiments. A flux offset study performed at laboratory scale and a corresponding flux model were used to specify the parameters required to operate the connected steps in constant flux mode. These can be applied directly to an expanded connection process, bypassing a small scale connected operation process.

Additional process monitoring capabilities, such as buffer tank level control for loading to start and end unit operations, must be considered for the connection process. On-line UV integration can be performed to determine the quality input for the TFF step, thus allowing volume targets to be set for diafiltration and over-concentration and determining step yield. A diverter pump should also be incorporated into the pallet design to allow individual step performance and contaminant removal rates to be evaluated.

As described, processing the entire collection batch requires multiple chromatography cycles of the ligation process, each ligation cycle taking 1 to 2 hours. Chromatography steps are typically cycled to reduce column size requirements and resin costs, and TFF membranes are typically cleaned and reused, so multiple cycles of the coupling process using these steps are straightforward. In contrast, in a discrete process, a viral filter is conventionally employed for a single cycle of product loading followed by a single buffer flush. For the join process, the load on the virus filter is not fully utilized for a single join cycle. The examples shown in Table 2 of FIG. 20 show 2 to 4kg/m²Is typical for a connected virus filtration step, whereas for this filter type at least 20kg/m can be achieved²Load of (Bolton G, Basha J, LaCa)sse D, 2010, "(Achiev high-throughput of therapeutic protein partial coverage filters") in Achieving high-quality throughput of medical proteins through small virus retention filters, "(Biotech Progress) 26 (6): 1671 page 1677). To increase the efficiency of the virus filter for the connected process, the same virus filter may be used for the whole collection batch, with the product loading phase followed by the buffer rinsing phase for each successive cycle. The virus filter will undergo alternating cycles of product and buffer, but filter performance is expected to be determined by the total product charge for all cycles. A subsequent filter integrity test will be performed after all cycles are completed to confirm that the filter remains integrated. This approach provides significant advantages by reducing the requirement for filter area and therefore article cost, eliminating the time consuming installation and preparation steps associated with filter replacement, and minimizing the risk of introducing foreign media into the process stream by keeping the system shut down after the virus filter. As the complexity of plant scheduling and equipment utilization increases with multi-cycle connected processing, it will be necessary to simulate plant resources to ensure proper facility fit within equipment maintenance cycle limits. In addition, evaluation of the virus removal capacity of individual steps in the ligation process, including the cycling of the virus filter, is worth careful consideration. Aspects such as developing an acceptable scale-down model and introducing virus incorporation into the linking process will be addressed as the subject of subsequent documents.

The connected downstream processes presented herein provide the immediate advantage of reducing the tank volume, thus resulting in a newer plant design. The reduction in sump size opens up the possibility of using a moving sump that can be easily reconfigured for multiple products with different process requirements. This results in reduced capital costs and provides manufacturing flexibility. The final goal is to fully link the collection, protein a and downstream steps for a fully continuous manufacturing. This would require the implementation of a continuous protein a extraction step, the development of an alternative to low pH viral inactivation batch operation using continuous multi-column chromatography (SMCC) or Simulated Moving Bed (SMB) technologies, and the implementation of all flow-through purification steps. This may be practicable in the near past as evidenced by the recent review articles focusing on continuous manufacturing and process integration (Konstantinov K and Cooney C, 2014, White papers on continuous bioprocessing (White paper on continuous bioprocessing), "journal of pharmacy DOI (J phase SciDOI): 10.1002/jss.24268; Jungbauer A, 2013," continuous downstream processing of biopharmaceuticals "(Trends in Biotechnology), Trends in Biotechnology, 31 (8): 479-492). The concept and control strategy presented in this document to connect downstream refining steps through a final TFF step pushes this technology one step further towards its goal.

While the foregoing text sets forth a detailed description of numerous embodiments of the invention, it should be understood that the legal scope of the invention is defined by the words of the claims set forth at the end of this patent. The detailed description is to be construed as exemplary only and does not describe every possible embodiment of the invention since describing every possible embodiment would be impractical, if not impossible. Numerous alternative embodiments could be implemented, using either current technology or technology developed after the filing date of this patent, which would still fall within the scope of the claims defining the invention.

It will also be understood that, unless a sentence "as used herein is used in this patent, the term '______' is hereby defined to mean …" or a similar sentence, specifically defining a term, it is not intended to expressly or implicitly limit the meaning of that term beyond its plain or ordinary meaning, and that such term should not be construed as being limited in scope by any statement made in any part of this patent (except the language of the claims, etc.). To the extent that any term recited in the claims at the end of this patent is referred to in this patent in the singular, this is done for sake of brevity and so as not to confuse the reader, and it is not intended that such claim term by limited, by implication or otherwise, to that singular meaning. Finally, unless a claim element is defined by reciting the word "means" and not having any structural recitation of functional limitations, it is not intended that the scope of any claim element be interpreted according to the application of 35u.s.c. § 112, paragraph six.

Claims (50)

1. A process control system, comprising:

(a) one or more upstream processing units, each operating a flow rate;

(b) a tank connected to the one or more upstream processing units;

(c) a filter having an inlet, a permeate outlet, and a retentate outlet connected to the sump;

(d) a fluid supply pump having an inlet connected to the sump and an outlet connected to the inlet of the filter;

(e) a sensor disposed at the permeate outlet to determine a flow rate at the permeate outlet; and

(f) a control system which is one of the following (i) to (iv):

(i) coupled to the sensor and the upstream process, and adapted to control a flow rate of one or more of the one or more upstream processing units as a function of the flow rate at the permeate outlet,

(ii) coupled to the sensor and the feed pump and adapted to control the feed pump according to the flow rate at the permeate outlet,

(iii) coupled to the sensor and the feed pump and adapted to control the feed pump according to the flow rate at the permeate outlet until a predetermined flow rate for the feed pump is reached, and adapted to control the flow rate of one or more of the one or more upstream processing units according to the flow rate at the permeate outlet after the predetermined flow rate for the feed pump is reached, or

(iv) Coupled to the sensor and the feed pump, and adapted to control the feed pump according to the flow rate at the permeate outlet until a predetermined flow rate for the feed pump is reached, and to allow for a mismatch between the flow rate of the one or more upstream processing units and the flow rate of the feed pump after the predetermined flow rate for the feed pump is reached.

2. The process control system according to claim 1, wherein the control system (i) is coupled to the sensor and the upstream process and adapted to control the flow rate of one or more of the one or more upstream processing units according to the flow rate at the permeate outlet.

3. The process control system according to claim 2, wherein the control system is adapted to reduce the flow rate of one or more of the one or more upstream processing units according to a reduction in the flow rate at the permeate outlet.

4. The process control system according to claim 1, wherein the control system (ii) is coupled to the sensor and the feed pump and adapted to control the feed pump according to the flow rate at the permeate outlet.

5. The process control system according to claim 4, wherein the control system is adapted to control the feed pump to maintain a flow rate at the permeate outlet that matches the flow rate of one or more of the one or more upstream processes.

6. The process control system according to claim 1, wherein the control system (iii) is coupled to the sensor and the feed pump and adapted to control the feed pump according to the flow rate at the permeate outlet until a predetermined flow rate for the feed pump is reached, and to control the flow rate of one or more of the one or more upstream processing units according to the flow rate at the permeate outlet after the predetermined flow rate for the feed pump is reached.

7. The process control system according to claim 1, wherein the control system (iv) is coupled to the sensor and the feed pump and adapted to control the feed pump according to the flow rate at the permeate outlet until a predetermined flow rate for the feed pump is reached and to allow for a mismatch between the flow rate of the one or more upstream processing units and the flow rate of the feed pump after the predetermined flow rate for the feed pump is reached.

8. The process control system according to any one of claims 6-7, wherein the predetermined flow rate is a maximum system flow rate.

9. A process control method for use in conjunction with one or more upstream processing units each having a flow rate, a tank connected to the one or more upstream processing units, and a filter having an inlet, a permeate outlet, and a retentate outlet connected to the tank, the method comprising:

(a) sensing a flow rate at the permeate outlet; and

(b) performing one of the following (i) to (iv):

(i) controlling a flow rate of one or more of the one or more upstream processing units according to the flow rate at the permeate outlet,

(ii) pumping material from the sump into the filter according to the flow rate at the permeate outlet,

(iii) pumping material from the tank into the filter according to a flow rate at the permeate outlet until a predetermined pumping flow rate is reached, and then controlling the flow rate of one or more of the one or more upstream processing units once the predetermined pumping flow rate is reached, or

(iv) Pumping material from the tank into the filter according to the flow rate at the permeate outlet until a predetermined pumping flow rate is reached, and then pumping material from the tank into the filter according to the predetermined pumping flow rate, thereby allowing a volume change in the tank.

10. The process control method of claim 9, wherein (b) (i) is performed.

11. The process control method according to claim 10, wherein the flow rate of one or more of the one or more upstream processing units is decreased according to a decrease in the flow rate at the permeate outlet.

12. The process control method according to claim 9, wherein (b) (ii) is performed.

13. The process control method according to claim 12, wherein the flow rate at the permeate outlet is matched to the flow rate of one or more of the one or more upstream processes.

14. The process control method according to claim 9, wherein (b) (iii) is performed.

15. The process control method according to claim 9, wherein (b) (iv) is performed.

16. The process control method according to any one of claims 14 to 15, wherein the predetermined flow rate is a maximum system flow rate.

17. A process for purifying a protein using one or more upstream processing units each having a flow rate, a tank connected to the one or more upstream processing units, and a filter having an inlet, a permeate outlet, and a retentate outlet connected to the tank, the process comprising:

(a) sensing a flow rate at the permeate outlet as the protein flows from the retentate outlet back to the tank;

(b) performing one of (i) to (iv):

(i) controlling a flow rate of one or more of the one or more upstream processing units according to the flow rate at the permeate outlet, the flow rate being a flow rate of material at least partially comprising the protein,

(ii) pumping material at least partially containing the protein from the reservoir into the filter according to the flow rate at the permeate outlet,

(iii) pumping material at least partially containing the protein from the storage tank into the filter according to a flow rate at the permeate outlet until a predetermined pumping flow rate is reached, and then controlling the flow rate of one or more of the one or more upstream processing units once the predetermined pumping flow rate is reached, or

(iv) Pumping material at least partially containing the protein from the tank into the filter according to the flow rate at the permeate outlet until a predetermined pumping flow rate is reached, and subsequently pumping material from the tank into the filter according to the predetermined pumping flow rate, thereby allowing the volume in the tank to change; and

(c) the protein is purified in the eluent.

18. The process control method according to claim 17, further comprising formulating the protein in a pharmaceutically acceptable excipient.

19. The process control method according to any one of claims 17 to 18, wherein (b) (i) is performed.

20. The process control method according to claim 19, wherein the flow rate of one or more of the one or more upstream processing units is decreased according to a decrease in the flow rate at the permeate outlet.

21. The process control method according to any one of claims 17 to 18, wherein (b) (ii) is performed.

22. The process control method according to claim 21, wherein the flow rate at the permeate outlet is matched to the flow rate of one or more of the one or more upstream processes.

23. The process control method according to any one of claims 17 to 18, wherein (b) (iii) is performed.

24. The process control method according to any one of claims 17 to 18, wherein (b) (iv) is performed.

25. The process control method according to any one of claims 23 to 24, wherein the predetermined flow rate is a maximum system flow rate.

26. A process control system, comprising:

a microfiltration unit;

a single pass tangential flow filter having an inlet, a permeate outlet, and a retentate outlet;

a feed pump having an inlet connected to the microfiltration unit and an outlet connected to the inlet of the filter;

a permeate pump having an inlet connected to the permeate outlet of the filter; and

a control system coupled to the permeate pump and adapted to control the permeate pump to vary a flow reduction factor, wherein the flow reduction factor is a ratio of a feed liquid flow rate to a retentate flow rate.

27. The process control system according to claim 26, wherein the control system is adapted to control the permeate pump to vary the flow reduction factor to achieve a target volume reduction factor, wherein the volume reduction factor is a ratio of cumulative feed volume to cumulative retentate volume.

28. The process control system according to claim 26, wherein the control system is adapted to control the permeate pump to vary the flow reduction factor in a series of stepwise changes.

29. The process control system according to claim 28, wherein the control system is adapted to control the permeate pump to vary the flow reduction factor in a series of stepwise increases.

30. The process control system according to claim 26, wherein the control system is adapted to control the permeate pump to continuously vary the flow reduction factor.

31. The process control system according to claim 26, wherein the control system is adapted to operate the permeate pump to provide a first flow reduction factor and then to change to a second flow reduction factor different from the first flow reduction factor and to operate the permeate pump to provide the second flow reduction factor.

32. The process control system according to claim 26, wherein the control system includes at least one processor programmed to control the permeate pump to vary the flow reduction factor.

33. The process control system according to claim 32, wherein the at least one processor is programmed to control the permeate pump to vary the flow rate reduction factor to achieve a target volume reduction factor, wherein the volume reduction factor is a ratio of cumulative feed volume to cumulative retentate volume.

34. The process control system according to claim 32, wherein the at least one processor is programmed to control the permeate pump to vary the flow reduction factor in a series of stepwise changes.

35. The process control system according to claim 34, wherein the at least one processor is programmed to control the permeate pump to vary the flow reduction factor in a series of stepwise increases.

36. The process control system according to claim 32, wherein the at least one processor is programmed to control the permeate pump to continuously vary the flow reduction factor.

37. The process control system according to claim 32, wherein the at least one processor is programmed to operate the permeate pump to provide a first flow reduction factor and then to change to a second flow reduction factor different than the first flow reduction factor and to operate the permeate pump to provide the second flow reduction factor.

38. The process control system according to claim 26, wherein the control system is coupled to the fluid supply pump and adapted to control the fluid supply pump to provide a constant flow rate.

39. The process control system of claim 26, comprising a valve disposed between the retentate outlet and a mixing tank, the controller coupled to the valve and adapted to control the valve to provide backpressure for the filter.

40. A process control method, comprising:

pumping the material through a single pass tangential flow filter having an inlet, a permeate outlet, and a retentate outlet; and

pumping permeate from the permeate outlet of the filter to vary a flow rate reduction factor, wherein the flow rate reduction factor is a ratio of a feed liquid flow rate to a retentate flow rate.

41. The process control method of claim 40, wherein the flow rate reduction factor is varied to achieve a target volume reduction factor, wherein the volume reduction factor is a ratio of an accumulated feed volume to an accumulated retentate volume.

42. The process control method according to claim 40, wherein the flow rate reduction factor is varied in a series of stepwise changes.

43. The process control method according to claim 42, wherein the flow rate reduction factor is varied in a series of stepwise increases.

44. The process control method according to claim 40, wherein the flow rate reduction factor is continuously varied.

45. The process control method according to claim 40, wherein pumping permeate includes pumping permeate according to a first flow rate reduction factor followed by a change to a second flow rate reduction factor different from the first flow rate reduction factor and pumping the permeate according to the second flow rate reduction factor.

46. The process control method according to claim 40, wherein the pumping material comprises pumping material at a constant liquid supply flow rate.

47. The process control method according to claim 40, wherein the material comprises cellular material and protein products.

48. The process control method according to claim 47, wherein the protein product comprises monoclonal antibodies.

49. The process control method according to any one of claims 40 to 48, wherein the material at least partially comprises a protein and the process further comprises purifying the protein with an eluent.

50. The process control method according to claim 49, further comprising formulating the protein in a pharmaceutically acceptable excipient.