Improved method for selecting high producing cell lines

0001The present invention relates to the field of biopharmaceutical cell culture technology. Specifically, the present invention relates to a method for rapidly identifying and selecting a cell line and a cell bank suitable for producing a biological product, and a method for producing the same. More specifically, the method comprises sorting cells without the use of animal-derived proteins. The present invention also relates to a novel gating strategy based on cell backscatter (FSC) properties.

0002Mammalian cells such as CHO (Chinese Hamster Ovary Cell), NSO and PerC6 cells are commonly used in the biopharmaceutical industry to produce biopharmaceuticals. The cells are genetically engineered and then selected to ensure that high titer expression of the desired protein is observed when the resulting cell line is cultured in a bioreactor.

0003Currently, all approaches to selecting the best secretory cell clones include screening hundreds to thousands of transfected cells to identify preferred clones with optimal proliferation and production profiles (eg, Wurm, 2004, Nature Biotechnology 22, 1393-1398). All such methods require considerable time and effort. As a result, there is still a need for improved methods for selecting high yield and stable cell lines. Such methods should be faster, less labor intensive and less resource intensive than current methods. A typical current cell line production protocol is delivery of the desired expression and selective cassette into the desired host cell, followed by serum and animal-derived components-free (ADCF) medium of the genetically engineered cells obtained. Includes seeding, selection and proliferation. First discovered in the early 1990s, such ADCF medium is now routinely used to select and proliferate recombinant cells that produce biopharmaceutical proteins for clinical use. The use of such cell culture media for this purpose ensures that the mammalian cell lines produced are not inadvertently exposed to exogenous or infectious agents of known or unknown origin. Therefore, the use of such ADCF medium is such that the resulting cell lines and their biologic cell banks, processes, production plants and end products are not contaminated with such exogenous and infectious agents, which are of origin. Form part of a strategy that ensures that it is a cell, virus or protein. The use of ADCF medium as part of a strategy to ensure clean manufacturing does not only address the theoretical risks to patients and manufacturing plants: biopharmacy has previously been contaminated with exogenous factors. (Eg, Huang WT et al. Pharmacoepidemiol Drug Safety 2010, 19 (3): 306-310), the manufacturing site has been previously blocked due to contamination with such factors (eg, bioreactors in manufacturing plants). Garnick RL Dev Biol Standard 1998: 93 pp21-9 & See FDA website notice (June 2009)). Further details on these considerations and the best way to reduce the risk can be found in ICH Quality Guide Q5A "Safety Evaluation of Biotechnology Products Derived From Cell Lines of Human or Animal Origin" (Step 4 Version September 1999). Well proven.

0004Specific cell screening methods designed to identify optimal genetically engineered cells utilize flow cytometry to reduce time and effort. Two approaches are adopted when using flow cytometry to select genetically engineered cell lines that secrete high levels of the desired recombinant protein.

0005The first approach involves selecting cells that overexpress an enzyme (when combined with a fluorescent substrate) such as a co-marker (eg, GFP) or DHFR. Attempts must be made to correlate marker expression with expression of the desired protein product in order to be successful with such a selection method. This association is required to ensure that cells expressing high levels of markers also express and secrete high levels of the desired protein product. Examples of such approaches include Yoshikawa T et al. 2001 (Biotech and Bioeng 74: 5 pp435-442), Meng YG et al. 2000 (Gene 25: 242 (1-2) 201-207, de Maria et al. 2007 (Biotechnol Prog). 23, 475-72) and US Patent Application Publication No. 2004/0148647. However, there are drawbacks to this approach, especially the fact that high levels of marker expression are initially selected, but this is not always associated with high expression and secretion of the desired protein product. Also, as a result of this approach, it primarily results in the selection of final cells that express very large amounts of markers. This can be inconvenient. This is because high expression markers compete with the desired protein product for host cell transcription, translation and processing organs, thus reducing the final yield of the desired protein. In addition, such marker proteins can also be toxic (Liu HS et al. 1999 (Biochem Biophys Res Commun 260 (3): 712-7)) and thus inhibit cell line proliferation and production.

0006The second approach involves direct selection of cells that secrete high levels of the desired protein product. For an overview of FACS in general, see Shapiro: Practical Flow Cytometry, 4th Edition, 2003, Wiley-Liss, ISBN # 9780471411253. In a broad sense, there are two FACS-based methods for directly selecting highly expressed clones.

0007The first "direct selection", FACS-based method, utilizes the observed correlation between membrane binding and secretory levels of the desired protein product. An example of this approach can be found in Marder P et al. 1990 Cytometry 11: 498-505. In this study, the authors stained the membranes of hydridoma cells with fluorescently conjugated anti-product antibodies, followed by selection and selection of the highest fluorescent cells. They then demonstrated that the resulting subclones exhibited improved IgG secretion levels compared to unselected cells. Subsequent reports using different cell lines demonstrated similar results (see, eg, Brezinsky S et al. 2003. Jn of Immunol (2003) 141-155).

0008Related, but more complex, alternatives to membrane-level staining of products and subsequent sorting include the use of product capture approaches such as gel microdrop (GMD) techniques or matrix-based secretion assays. .. In such an approach, the secretory antibody is retained either by cross-binding to the cell membrane or within gel microdrops, or is immobilized on an artificial matrix on the cell surface prior to FACS sorting with the anti-product antibody for the level of fluorescence. ..

0009Recent recent cytometrically based cell selection methods that describe both an indirect selection approach (ie, by the use of markers) and a direct selection approach (ie, by labeling the desired product associated with the cell). See Carroll and Al-Rubeai 2004 (Expert Opin Biol Ther 4: 1821-9) and Browne and Al-Rubeai 2007 (Trends Biotech, 25 (9), 425-32) for an overview.

0010However, both direct and indirect labeling methods described above are currently proven to be successful, but both have ever been ideal for use in selecting high-producing cell lines suitable for the production of biologics. Not the target. This is because when an indirect approach is utilized, cells that produce the high marker levels described above are predominantly selected. Alternatively, all direct selection methods have so far utilized animal-derived anti-product antibodies and supporting reagents (eg, fetal bovine serum (FCS) or bovine serum albumin (BSA)), which are biopharmaceuticals. Not desirable for development and manufacturing. Such animal reagents are also regularly utilized to support cell survival during rigorous flow sorting and seeding procedures. This again increases the risk of exposure to exogenous factors for any cell line produced thereby. In addition, animal-derived reagents such as serum or BSA are typically used to block non-specific binding prior to cell incubation with the desired anti-product binding reagent such as antiserum.

0011Recently, Applebaum et al. (US Patent Application Publication No. 2010/0028904) found highly expressed cells by seeding cells in methylcellulose (semi-solid medium) containing fluorescent protein A / G to detect product expression. Methods for screening and selection were described. However, there are still considerable challenges with this method. Importantly, the fluorescent protein A / G does not necessarily bind to the protein of interest on the cell surface, but rather to the pericellular secretory protein (obtained by the "halo" around the cell population). .. In fact, protein A mainly binds to antibodies and Fc fusion proteins in the constant region, thus presenting any such target protein A binding epitope for binding to the cell membrane prior to product separation and complete secretion from the cell. Even if it is available, it remains unknown. Therefore, this method relies heavily on labeling of cells suspended in semi-solid medium that do not contain Fc protein. As a result, the method cannot guarantee the clonality of the selected strain. In addition, the conditions for cell proliferation in semi-solid medium and the amount of fluorescent protein A / G used may need to be established for each time the method is used. Selection of cells from semi-solid medium also also requires a certain amount of time (usually at least 7 days) for cells to be able to secrete a sufficient amount of product into the detectable pericellular medium. Waste time.

0012In addition, the level of target binding and the resulting level of fluorescence achievable is significantly reduced due to the low affinity of protein A compared to higher affinity antisera (such as non-human antisera antisera). .. For these reasons, protein A is typically utilized only as a secondary labeling reagent along with a higher affinity primary reagent.

<p num="0013"> There is a demand for improved cell selection methods in the art.</p>

<p num="0014"> The present invention provides an improved method for identifying and selecting individual cell lines expressing high titers of a desired polypeptide (eg, antibody or antibody-related product). The present invention does not require the use of animal-derived components such as serum at any stage, and is the same as or is the same as using animal-derived antisera and standard protocols to identify optimal high-producing clones. It provides a new direct staining method that is more effective. Therefore, selected cells (including multiple cells) (and cell banks and methods of their manufacture) that can be production cell lines expressing biologics for use in humans are not overexposed to exogenous and infectious agents. .. The present invention overcomes the technical challenge of using an animal-derived component-free FACS selection protocol to select high-producing clones by directly staining products on cell membranes. The present invention also overcomes the need to use a semi-solid medium and the time required for extracellular "halo" formation.</p><p num="0015"> The invention also provides a novel gating strategy based on cell forward scatter (FSC) properties that allows better identification and subsequent isolation of high-producing cells (s) / cell lines.</p><p num="0016"> In addition, as a result of the present invention, high-producing clones are identified from the very early stages of cell line production, that is, as early as the 96-well stage, so there is a need to screen a large number of clones. Reduce. As a result, the effort involved in cell line production is significantly reduced. This is the first report demonstrating the feasibility of identifying high-producing cell lines even from the 96-well stage using a direct labeled FACS approach for single cell selection.</p><p num="0017"> In the first aspect, a method of screening one or a plurality of cells expressing a target polypeptide, the step of providing a cell population containing at least one cell expressing the polypeptide, and a cell population thereof. Was brought into contact with a label that binds to the polypeptide on the surface of the cell, a step of detecting the binding of the label to the polypeptide on the surface of the cell, and binding to the polypeptide on the surface of the cell. Provided is a method comprising selecting the cells based on the presence of the label, wherein the label is not a polyclonal antibody.</p><p num="0018"> In another embodiment, a method of selecting one or more cells expressing a polypeptide of interest, the step of providing a cell population comprising at least one cell expressing the polypeptide, and the cell population thereof. The step of contacting the label with a label that binds to the polypeptide on the surface of the cell, the step of detecting the binding of the label to the polypeptide on the surface of the cell, and the step of binding to the polypeptide on the surface of the cell. Provided is a method comprising the step of selecting the cells based on the presence of the label, wherein the label does not contain animal-derived components.</p><p num="0019"> In another embodiment, a method of selecting highly expressed cells from a population of genetically engineered cells expressing a heterologous polypeptide, comprising providing a population of cells expressing the heterologous polypeptide and specific within the population. A method comprising detecting the relative size of cells with respect to cells, selecting the largest one or more cells from a population, and isolating the selected one or more cells from a population of cells. I will provide a.</p><p num="0020"> In another embodiment, a method of selecting a highly expressing cell from a population of cells expressing a secreted heterologous polypeptide, the step of providing the population of cells expressing the heterologous polypeptide, and the cell population. Detecting the relative amount of the label bound to the polypeptide on the surface of the cell, the relative size of the cell, for the step of contacting the label on the surface of the cell and for a particular cell in the population. And / or selecting at least one cell from the population based on the relative amount of the label bound to the polypeptide on the surface of the cell and / or the relative amount of the cell, and the selection from the cell population. A method comprising the step of isolating at least one cell is provided.</p><p num="0021"> The step of selection may include selecting cells with a higher mean FSC-A or higher FSC-W compared to the mean FSC-A or FSC-W of the living cell population.</p><p num="0022"> The use of non- polyclonal antibody labels allows cells with the desired profile to be selected without the use of animal-derived components such as serum or animal-derived components. In one embodiment, the label is not an antibody. In one embodiment of the invention, the label is recombinant protein A, recombinant protein G or recombinant protein L. Thus, the label is not derived from the animal source and is not, for example, exposed to animal source serum or other potential exogenous factors. This reduces safety concerns and, if the polypeptide is a protein intended for use as a biopharmaceutical, improves the downstream treatment of the polypeptide product.</p><p num="0023"> The label may be conjugated to a marker to facilitate detection of the label bound to the surface of the cell. Suitable markers are fluorescent markers, magnetic markers, or biotin markers. Typically, the marker is a fluorescent marker suitable for detection by FACS devices such as FITC, RPE, DYLIGHT, ALEXA FLUOR, CYDYE, LI-COR, PE, Cy5, Cy7, PerCP and APC. To avoid misunderstanding, marker conjugation may be direct (eg with fluorescent protein A) or indirectly (eg with biotin that can be bound by fluorescent streptavidin). Therefore, typical labels include fluorescent proteins A, G or L.</p><p num="0024"> The method may include detection of a relative amount of a label bound to the surface of the cell, more specifically a label bound to a polypeptide on the surface of the cell. As shown in the anti-product antibody-based approach to FACS (see Marder, P. supra), the binding of the label to the polypeptide found on the surface of the cell is expressed by the cell. An indicator of the amount of polypeptide is provided. Thus, the step of detecting the relative binding of a label to the surface of a cell as compared to other cells in the cell population can provide an indicator that the cell may be an expresser of the polypeptide.</p><p num="0025"> In one embodiment, the detecting step further comprises detecting the relative size of the cells. We find that a subpopulation of cells (which may be mitotic cells) is larger (as measured by FSC and / or SSC characteristics) than other cells within the cell population. (Large volume) was discovered. This subpopulation is different from the main population, for example, when analyzed by forward scatter width (FSC-W) or forward scatter region (FSC-A). When cells from this subpopulation are cultured, the clones surprisingly express more of the polypeptide of interest than the cells of the major cell population, which are smaller in size (as measured by FSC and / or SSC characteristics). .. This observation is particularly surprising as these larger cells are a dynamic subpopulation that returns to average cell size over time. This large cell subpopulation is also referred to herein as the HFA cell subpopulation (and can be identified by FACS measurements of forward scattering properties, ie FSC-A and / or FSC-W), and is smaller ("usually". ") The cell subpopulation is called the LFA cell subpopulation.</p><p num="0026"> In one embodiment, the step of selection comprises selecting one or more cells from the population showing the highest amount of binding label. The number of cells selected depends on the number of cells that are ultimately intended for cloning and / or culturing, and the resources available thereby. One of ordinary skill in the art can define selection criteria to select these cells that have the highest amount of label bound to the cells. For example, if the label is a fluorescent label, one of ordinary skill in the art can define selection criteria to select the cells that fluoresce the most. In the FACS protocol, one of ordinary skill in the art can easily define gating criteria to select the cells that emit the strongest fluorescence. The cells that emit the strongest fluorescence may be the most highly expressed cells in the cell population. In a typical embodiment, 20%, 15%, 10%, 9%, 8%, 7%, 6%, 5%, 4%, 3 of the most bindings (eg, the strongest fluorescence) in the cell population. %, 2% or 1% or less cells are selected.</p><p num="0027"> In one embodiment, the step of selection comprises selecting one or more cells from the population that are larger or largest or proliferating. The number of cells selected depends on the number of cells that are ultimately intended for cloning and / or culturing, and the resources available thereby. One of ordinary skill in the art can define selection criteria to select the larger or largest cells from within the cell population. For example, in the FACS protocol, one of ordinary skill in the art can define gating criteria to select larger or largest cells. It sets the gate for FACS after analyzing a sample of cell population by FSC-H / FSC-A plot, FSC-W / FSC-A plot, FSC-W histogram or FSC-A / SSC-A plot. May include that. As used herein, the terms "larger" and "largest" mean that one or more cells are selected based on their forward scattering properties, namely FSC-A or FSC-W. means. Larger or largest one or more cells have a higher mean FSC-A (if identified in the FSC-A / FSC-H plot or FSC-A / SSC-A plot) and more than the major live cell subpopulation. Any cell (including multiple) within a subpopulation of living cells with high FSC-W (as identified in the FSC-A / FSC-W plot or FSC-W histogram). Similarly, as used herein, the term "highest" describes the forward scattering properties of cells, those in which the cells are higher than the average FSC-A or FSC-W of all living cells in the population. Means selected based on FSC-A or FSC-W. In a typical embodiment, the largest of the cell population, 20% or less, 15% or less, 10% or less, 9% or less, 8% or less, 7% or less, 6 % Or less, 5% or less, 4% or less, 3% or less, 2% or less, 1% or less</p><p num="0028"> In addition to those described herein, there are a wide variety of suitable gating strategies that are used to distinguish and select the same subpopulations described herein. It will be understood that it is also good. A skilled operator of the flow cytometer can therefore identify and gate subpopulations based on the teachings of the present invention. Therefore, such alternative gating strategies for selecting HFA cell subpopulations (or subpopulations thereof) or subpopulations that emit strong fluorescence (or subpopulations thereof) are also contemplated in the present invention.</p><p num="0029"> In addition, different / additional labeling strategies that provide selection and / or isolation of subpopulations (or subpopulations thereof) are also intended in the present invention.</p><p num="0030"> The process of selection typically involves selecting only live cells from the population. Determining that a cell in a population is a living cell is within the normal ability of one of ordinary skill in the art. Thus, for example, in the FACS protocol, one of ordinary skill in the art can define raw gating of cytometers. This typically involves analyzing a sample of a population of cells by FSC-A / SSC-A plot. Further staining with, for example, propidium iodide may also be used.</p><p num="0031"> The method further comprises the step of isolating the selected one or more cells from the cell population. Typically, this involves the selection and isolation of individual cells from a cell population. The individual cells of choice may be precipitated in a separate vessel, such as a 96-well plate well. One or more vessels usually contain a medium suitable for clonal growth of selected and isolated cells. Typically, the medium is a serum-free medium.</p><p num="0032"> In one embodiment, the medium is a medium that does not contain animal-derived components.</p><p num="0033"> It is understood that a small number of cells have difficulty growing in such conditions due to lack of cell-cell contact or essential factors released from the cell population. Thus, in one embodiment, the medium may be conditioned medium (filtered supernatant from a culture of non-transfected cells) or may contain a mixture of conditioned and fresh medium. In an alternative embodiment, the container contains supporting cells that can be essentially self-derived (ie, from the same species, the same tissue or the same genetic origin). In alternative embodiments, the fresh or acclimatized medium contains (optionally recombinant) supplements (s) that support single cell proliferation, such as albumin, transferrin, or insulin growth factors.</p><p num="0034"> In one embodiment, the label is a fluorescent protein A, protein G, or protein L, typically protein A. The step of detecting typically involves determining the relative fluorescence of the label bound to the polypeptide on the surface of the cell, where relatively high fluorescence indicates higher relative expression of said polypeptide. The step of selection typically follows the step of detecting and then selecting one or more cells with a relatively high level of fluorescence from the population detected in the step of detection compared to the overall cell population. including. Isolation steps typically include separating one or more selected cells from the cell population and precipitating the selected cells independently in separate containers, usually in wells of a 96-well plate. including. Under these conditions, selected highly expressed cells can grow under clonal conditions.</p><p num="0035"> The polypeptide may be a non- polyclonal IgG or any polypeptide that can be labeled with a non-antibody label. Polypeptides are, for example, but not limited to, antibodies, hormones, enzymes, growth factors, receptors, fusion proteins, antigens, biologics or any other polypeptide, whether synthetic or natural. It may be any suitable soluble (secreted) or membrane-bound polypeptide or multimeric polypeptide, including. Typically, the polypeptide is a therapeutic protein.</p><p num="0036"> In one embodiment, the polypeptide is a secreted polypeptide. In one embodiment, the polypeptide is an immunoglobulin CH1, CH2 or CH3 domain, CH2 and CH3 domain, or immunoglobulin Fc domain (which may be a native Fc domain or a derivative thereof). including. Typically, the polypeptide is an antibody.</p><p num="0037"> Typically, the polypeptide is a recombinant or heterologous polypeptide. In order to obtain cells expressing the recombinant / heterologous polypeptide, transfection of the cells with a suitable expression vector encoding the expressed gene is usually required. Vectors usually contain marker genes, such as genes encoding antibiotic resistance, to allow selection of cells expressing the polypeptide of interest. Alternatively, the marker gene may be expressed from a second vector that is introduced into the cell at the same time as the vector encoding the polypeptide of interest. In a typical embodiment, the cell line is a dhfr cell line and one or more expression vectors transfected into the cells contain DHFR and antibiotic resistance. However, other transfection methods (or other methods such as transduction) and selection procedures are also suitable and may be optionally selected by one of ordinary skill in the art.</p><p num="0038"> The cell population is not limited, but any of them, including but not limited to mammalian cells, bacterial cells, yeast cells, plant cells, insect cells, bird cells, fish cells, or immortalized or transformed cell populations based on them. It may be any prokaryotic or eukaryotic species such as a derivative. In one embodiment, the cell is a mammalian cell, or a mammalian cell of an immortalized or transformed cell. In one embodiment, the cells are CHO, CHO K1, CHO DG44, NSO, COS-1, COS-7, HEK293, HeLa, HK21, PerC6, 293T, Vero, AGE1.CR, HT1080, TE671, Namalwa or SP2 / Selected from 0. Typical cell lines are, but are not limited to, CHO cell lines, including, but not limited to, CHO K1, CHO K1 Chk2, CHO DG44, CHO DXB-11, or cell lines derived from them. The CHO cell line may be a dhfr cell line such as the CHO DG44 dhfr line. For example, mammalian cell lines engineered to express modified sugar chains are also suitable for use in the present invention. An example of such a cell line is CHO Includes FUT8 knockout cell lines (eg, US Pat. No. 7,214,775, US Pat. No. 6,946,292) or cell lines engineered to upregulate GnTIII expression (eg, WO99 / 55342). In one embodiment, the cell population may consist of cells containing artificial chromosomes such as the CHROMOS-derived ACE lineage (Lindenbaum et al., Nucleic Acids Research, 2004: 32 (21)).</p><p num="0039"> The cell population may be from bulk transfection or may be an established / cloned cell line.</p><p num="0040"> In one embodiment, the method of the invention is a flow cytometry method, more typically a fluorescence activated cell selection (FACS) method. Typically, the detection step is performed in a fluorescence activated cell sorter or analyzer. Typically, the selection step is performed in a fluorescence activated cell sorter or analyzer. Typically, the isolation step is performed in a fluorescence activated cell sorter.</p><p num="0041"> In one embodiment, the method further comprises washing the cells prior to the step of contacting. The washing step removes the polypeptide secreted by the cells in the surrounding medium, thereby ensuring that the label binds to the polypeptide expressed on the cell surface and increases cell-specific staining. The steps of washing the cells are typically, but not limited to, suspending the cells in a buffer such as PBS, followed by centrifugation of the cells to produce cell pellets, and supernatant. Includes separating cell pellets from.</p><p num="0042"> The present invention is also a method for selecting one or more cells expressing a target polypeptide by flow cytometry, which comprises a step of staining cells with a fluorescent binding label and does not utilize animal-derived components. provide.</p><p num="0043"> The present invention is also a method of selecting one or more cells expressing a polypeptide of interest using a fluorescent binding label by flow cytometry, wherein all the components utilized are recombinant and non-animal fed. Provide a method of origin.</p><p num="0044"> According to the present invention, a single selected cell may be grown or cultured under conditions that allow the cell to proliferate and divide, thereby giving a second cell population (cloned population) derived from those cells. ) Is generated. As used herein, it is shown that the step of recloning a single cell from a clonal population can increase the titer (ie, productivity) of the cell. Therefore, in one embodiment, the method comprises the step of growing the selected cells under clonal conditions. This step of growing a is referred to as a clonal population or a further population of cells herein can generate a population of additional cells. The method may further include selecting one or more cells from the further cell population and isolating the one or more cells from the additional cell population. This process may be repeated as many times as desired or necessary to produce stable and / or higher producing cell lines. In each case, the step of isolating cells from the further may be performed according to the method of the invention to select highly expressed cells.</p><p num="0045"> Therefore, in another embodiment, the present invention is a method of cloning cells, wherein the cells are isolated from a population containing a plurality of cells, and the isolation is performed under cloning conditions in order to generate a clone population. Provided is a method including a step of growing cells.</p><p num="0046"> In one embodiment, the growing step is carried out in conditioned medium. In one embodiment, the cells are dhfr cells transfected with the heterologous gene and the DHFR gene, and the step of proliferation is performed in a medium containing methotrexate. In one embodiment, cells are also transfected with a neomycin resistance gene and the step of growing is performed in medium containing G418. In this embodiment, the medium may further comprise methotrexate.</p><p num="0047"> In the present specification, the addition of recombinant serum albumin and / or recombinant transferrin in the medium has been found to increase cloning efficiency (to the extent that the medium does not require a conditioned medium). Therefore, the medium may further comprise recombinant serum albumin and / or recombinant transferrin. Serum albumin concentrations range from 0.1 to 10 mg / mL, more typically 0.5 to 5 mg / mL, more typically further 0.75 to 2.5 mg / mL, and most typically about 0.5 to 1 mg / mL. You may. The transferrin concentration may vary as well, but is typically in the mg / L range, most typically about 5 mg / L.</p><p num="0048"> In the present specification, the viability and / or productivity of a cell line grown under clonal conditions after isolation of a single cell (single cell selection) removes the selective pressure for a certain period of time. It is further found that can be improved by. Thus, in one embodiment, the cell is a dhfr cell transfected with a heterologous gene and a DHFR gene. In this embodiment, after the step of isolating cells from the population, the isolated cells are grown for a first period in a medium which may be an antibiotic-free conditioned medium such as methotrexate or G418. Following the first period, there may be a second period during which methotrexate and G418 are added to the medium. Typically, the first period is 12-100 hours, more typically 24, 26, 48, 64 or 72 hours.</p><p num="0049"> Therefore, in another embodiment, the invention is a method of cloning cells, which comprises the step of isolating cells expressing a heterologous protein and a selective / amplifying agent (including a plurality) during the first growth phase. ), For example, in a medium in the absence of methotrexate and / or antibiotics, and in a second growth phase, in a medium containing a selective / amplifying agent (s). Provide a method to include.</p><p num="0050"> The method of the present invention may further include the step of proliferating the selected cells under clonal conditions to generate a universal cell bank. This method is a step of proliferating the selected cell under clonal conditions, further comprising a step of expressing the polypeptide in the cell, a step of separating the polypeptide from the cell, and a step of purifying the polypeptide. It may be included.</p><p num="0051"> The present invention also provides cells selected by the methods according to the invention and polypeptides produced by such cells.</p><p num="0052"> Other advantages, when read in conjunction with the drawings, will become apparent from the detailed description below.</p>

0053<figref num="1A">It is a cell line preparation flow chart and the vector schematic diagram. (A) Conventional / standard cell line production protocol. SCC: Single cell cloning; SF: Shaking flask; MTX: Methotrexate; DHFR: Dihydrofolate reductase; HC: Heavy chain; LC: Light chain)</figref><figref num="1B">It is a cell line preparation flow chart and the vector schematic diagram. (B) Flow diagram of a novel, FACS-based cell line production protocol. SCC: Single cell cloning; SF: Shaking flask; MTX: Methotrexate; DHFR: Dihydrofolate reductase; HC: Heavy chain; LC: Light chain)</figref><figref num="1C">It is a cell line preparation flow chart and the vector schematic diagram. (C) Schematic diagram of vectors for heavy and light chain constructs. SCC: Single cell cloning; SF: Shaking flask; MTX: Methotrexate; DHFR: Dihydrofolate reductase; HC: Heavy chain; LC: Light chain)</figref><figref num="2">It is a figure which shows the FACS analysis and selection gating logic used in an Example. The flow chart shows the order of gating. An asterisk indicates when the selection could / was performed. Starting from the left side of the figure, to identify living cells, cells are first selected based on their FSC-A vs. SSC-A characteristics. If desired, propidium iodide may be further used to distinguish dead cells. When living cells are gated, two different streams follow as follows: Fluorescence of this live cell population can also be measured at the top (by staining with protein A or IgG as described in the examples below) and gates can be set for the cells with the highest fluorescence (raw-high gates). , Usually the top 2-5%). In parallel, at the bottom, living cells can be further distinguished based on their FSC properties into high forward scatter region (HFA) and low forward scatter region (LFA) cells as described in Example 8. To identify and select HFA-high and LFA-high subpopulations, respectively, fluorescence of either HFA or LFA cells can be measured and a live-high fluorescence gate can be set for the live cell population to be applied.</figref><figref num="3">FIG. 5 shows unstained cell selection for cell surface IgG expression-subcloning of existing cell lines. Batch production model of parent cell line and subcloned daughter cell line for various schemes. The cell line was a single cell cloned by FACS (Plan 5d, 7b and 10) using only live cell gates or by limiting dilution (Plan 6d). In the case of 7b, two rounds of cloning were performed by FACS. The data show a markedly increased titer after cloning. b. Batch production models at different passages (p10, 21, 31, 41 and 51) for single cell selective clones (Plan 7b, strain 84-8 # 158-202) demonstrate cell line stability. ..</figref><figref num="4">It is a figure which shows the detection of the cell surface IgG expression in the bulk-transfected CHO cell which was live cell gated using FITC-labeled anti-human rabbit polyclonal IgG. Fluorescence histogram of untransfected CHO cells or CHO cells transfected with plasmid encoding the antibody of design 10: a) untransfected CHO DG44 cells, b) G418 selected transfected cells, c) G418 Transfected cells after selection and methotrexate amplification. Dotted lines are unlabeled cells and solid lines are anti-human IgG-labeled cells.</figref><figref num="5-1">It is a figure which shows the optimization of the cell staining condition for FACS analysis and selection. Parent CHO-DG44 transfected with plasmid encoding antibody of Plan 10 using protein A Alexa647 (row A), FITC-labeled rabbit anti-human IgG (row B) and DyLight 649-labeled rabbit anti-human IgG (row C). Staining of cells and CHO-DG44 cells. The graph in the left column shows the average fluorescence signals obtained for different antibody / protein A concentrations, and the graph in the right column shows the corresponding signal-to-noise ratio, that is, trans divided by the fluorescence of the DG44 parent cells at the same concentration. Shows the average fluorescence of infected cells. For all graphs, the fluorescence of live gate cells is shown.</figref><figref num="5-2">It is a figure which shows the optimization of the cell staining condition for FACS analysis and selection. Parent CHO-DG44 transfected with plasmid encoding antibody of Plan 10 using protein A Alexa647 (row A), FITC-labeled rabbit anti-human IgG (row B) and DyLight 649-labeled rabbit anti-human IgG (row C). Staining of cells and CHO-DG44 cells. The graph in the left column shows the average fluorescence signals obtained for different antibody / protein A concentrations, and the graph in the right column shows the corresponding signal-to-noise ratio, that is, trans divided by the fluorescence of the DG44 parent cells at the same concentration. Shows the average fluorescence of infected cells. For all graphs, the fluorescence of live gate cells is shown.</figref><figref num="6A">It is a figure which shows the comparison of the IgG (DyLight649 conjugated) label and the protein A (Alexa647 conjugated) label of a control (non-transfected cell), a bulk-transfected cell and a previously cloned strain. (A) Comparison of parental non-transfected CHO DG44 cells with bulk-transfected (after G418 selection and methotrexate amplification t) Plan 14 cells. Analysis after IgG (upper panel) and Protein A (lower panel) staining is shown. The left column shows the live cell gate (FSC-SSC plot) and the right column shows the fluorescence histogram of the live stained cells. The number indicates the average fluorescence intensity of the living cell population.</figref><figref num="6B">It is a figure which shows the comparison of the IgG (DyLight649 conjugated) label and the protein A (Alexa647 conjugated) label of a control (non-transfected cell), a bulk-transfected cell and a previously cloned strain. (B) Comparison of parental untransfected CHO DG44 cells with a previously subcloned cell line (cell line 84, Plan 12). Both IgG (upper panel) and Protein A (lower panel) are shown. The left column shows the live cell gate (FSC-SSC plot) and the right column shows the fluorescence histogram of the live stained cells. The number indicates the average fluorescence intensity of the living cell population.</figref><figref num="7">It is a figure which shows the correlation of the cell selection gating standard and the IgG expression after selection. Four-way selection of CHROMOS ChK2 cells from CHO-K1-stablely transfected with the vector for Plan 17 was performed. Cells were FACS screened after staining with either anti-human IgG or protein A. Four-way gating of (A) protein A (Alexa647 conjugated) fluorescence histogram and (B) anti-human IgG (DyLight649 conjugated) fluorescence histogram of each living cell is shown. The percentage of cells in each gate was about 25% (P3 and P8), 50% (P4 and P9), 20% (P5 and P10) and 5% (P6 and P11). (C) The antibody titer was divided by the number of cells from 5 days after sorting 100,000 cells per sorting gate.</figref><figref num="8">FIG. 5 shows a better response to feed in initial recovery and batch production of clones sorted by FACS when compared to clones derived from the conventional method (Plan 13). (A) T75 cm of clones from either bulk by FACS (gray) or low density seeding (conventional / standard protocol (black)) and selective / amplified in single cell cloning.²Cell viability at the static tissue culture flask stage. (B) Survival of the same clone as measured in the initial shaking flask (SF) stage. (C) Day 14 IgG titers of the same clone in a small growth curve (Duetz) model showing FACS (gray box) and conventional / standard protocol (black box) data.</figref><figref num="9">It is a figure which shows FSC measurement and gating about HFA cell. A&gt; FSC-Histogram of time vs. voltage for detection of laser signals for height, width and area measurements shown. B&gt; FSC-A / SSC-A plot showing live cell gates for strain MTX8 in Plan 10. Note: For the majority of analysis / sorting, a wider range of gates was used (see eg Figure 6). FSC-A / FSC-H plot from the live gate of B above showing C&gt; high forward scatter region (HFA) and low forward scatter region (LFA) cell populations</figref><figref num="10">It is a figure which shows the plot of the further flow cytometry which compares the FSC property of HFA and LFA cells (the representative data shown here was obtained using cell line # 15, plan 12). Plots (A) and (B) show FSC-W histograms or FSC-W / FSC-A dot plots for each of the LFA (dark gray) and HFA (light gray) populations, between the two populations. A clear separation of is observed. (C) Histogram plot for FSC-H for both HFA and LFA populations. No difference was observed between the two populations. (D) Histogram plot for FSC-A for both populations as well. Partial duplication of LFA and HFA cells is observed.</figref><figref num="11-1">FIG. 5 shows a comparison of IgG (DyLight649-conjugated) vs. Protein A (Alexa647-conjugated) labels for bulk-transfected cells (Plan 14) and previously cloned strain (Plan 12, strain # 84). From left to right: Live cell gate (FSC / SSC plot), HFA and LFA gates (FSC-H / FSC-A plot), fluorescence histogram of live cells with the indicated average fluorescence values, with the indicated average fluorescence values Fluorescence histogram of HFA cells and fluorescence histogram of LFA cells with the average fluorescence values shown. From top to bottom: Bulk population stained with rabbit anti-human IgG, bulk population stained with protein A, clone cell line stained with rabbit anti-human IgG, clone cell line stained with protein A.</figref><figref num="11-2">FIG. 5 shows a comparison of IgG (DyLight649-conjugated) vs. Protein A (Alexa647-conjugated) labels for bulk-transfected cells (Plan 14) and previously cloned strain (Plan 12, strain # 84). From left to right: Live cell gate (FSC / SSC plot), HFA and LFA gates (FSC-H / FSC-A plot), fluorescence histogram of live cells with the indicated average fluorescence values, with the indicated average fluorescence values Fluorescence histogram of HFA cells and fluorescence histogram of LFA cells with the average fluorescence values shown. From top to bottom: Bulk population stained with rabbit anti-human IgG, bulk population stained with protein A, clone cell line stained with rabbit anti-human IgG, clone cell line stained with protein A.</figref><figref num="12">It is a figure which shows the statistical analysis about the panel of a bulk population and an established cell line. Bulk population (transfected, selected, 5nM) stained with either DyLight649-IgG or Alexa647-Protein A The data obtained for MTX-amplified) (n = 6) and established cell lines (n = 19) are analyzed as follows: (A)% of HFA cells (black), higher compared to live cells Shows the enrichment ratio of 5% viable fluorescent cells in HFA cells, ie (% of the top 5% viable gates that are HFA) / (% of viable cells that are HFA) (intermediate gray); compared to viable cells Concentration ratio of the top 2% of viable fluorescent cells in HFA cells-(dark gray);, enrichment ratio of the top 5% of HFA cells in viable fluorescent cells compared to live cells, ie (in the top 5% live gate) Shows%) / (5) (light gray) of HFA cells; and enrichment ratio (white) in the top 2% of viable fluorescent cells of HFA cells compared to live cells. The X-axis indicates the type of cells and the type of staining analyzed. (B) Magnification of increase in average fluorescence intensity of HFA cells compared to live cells for parental bulk-transfected populations and established cell lines stained with labeled rabbit anti-human IgG or protein A.</figref><figref num="13">HFA single cell selection for established strains (Plan 10) shows that they produce stable subclones with increased batch production titers (parental titers were 1.6 g / L). Panel A shows the different cloning approaches used, and Panel B shows stable production titers at different passages for the best subclones (MTX8-B7) obtained from HFA sorting.</figref><figref num="14A">It is a figure which shows the increased titer and the improved stability of the clone obtained from the HFA single cell subcloning based on FACS of the design 3M cell line (BP0044). Panel A shows the batch production titers of the parent strain at different passage numbers, with a dramatic decrease in titers observed after passage 20. Panel B shows batch-producing titers for HFA-selected subclones (B2) at different passages, the titers increase and the strain is stable for at least 53 passages.</figref><figref num="14B">It is a figure which shows the increased titer and the improved stability of the clone obtained from the HFA single cell subcloning based on FACS of the design 3M cell line (BP0044). Panel A shows the batch production titers of the parent strain at different passage numbers, with a dramatic decrease in titers observed after passage 20. Panel B shows batch-producing titers for HFA-selected subclones (B2) at different passages, the titers increase and the strain is stable for at least 53 passages.</figref><figref num="15">FIG. 5 shows a comparison of IgG titers in both static and shake cultures for cell clones selected based on four different criteria-plan 3M subcloning (cell line BP0044). Sorting was performed after staining of BP0044 cells with DyLight649 conjugated rabbit polyclonal anti-human IgG. (A) Individual titers at the 6-well stage after single cell subcloning by FACS. The clones have four different selection criteria: -1) live cell gate (black), 2) HFA cells (white square), 3) strong fluorescent cells: live gated cells with top 1.5% fluorescence (rabbit anti). Human IgG staining) (white triangles) and 4) HFA strong fluorescent cells: Derived from FACS selection using HFA cells (black crosses) in the top 1.5% fluorescent gates. (B) Antibody titer and cell number in the shaking flask stage (passage 1). Apply the same color / shape scheme. This data shows that the majority of the large number of clones producing the highest antibody titers and the majority with the highest specific productivity (SPR) are derived from HFA and strong fluorescence selection criteria (black cross). Demonstrate.</figref><figref num="16">FIG. 5 shows the production of clonal cell lines for Plan 10 by FACS using HFA gating and IgG or Protein A staining / gating. CHO-DG44 cells were transfected with a vector encoding the antibody for Design 10, then selected in bulk, amplified, and then four different criteria: (i) Rabbit anti-human IgG-stained cells (FITC Conju). Live cell gating of (gated), (ii) Top 5% fluorescence and HFA gating of rabbit anti-human IgG-stained cells (FITC-conjugated), (iii) Live cell gating of protein A-stained cells, (iv ) Single cells were sorted based on the top 5% fluorescence and HFA gating of protein A stained cells. Panel (A) shows the number and average titers of clones achieved after sorting in the 96-well stage, and the data shown are for clones with a titer of 1 mg / L or higher at this stage. Panel (B) shows the IgG titers of the top clones from their respective sub-sort criteria at different stages (96 wells, 6 wells and SF passages 1 and 2). Panel (C) compares the mean batch-produced IgG titers on day 15 for the top HFA high protein A sorted cell clones (n = 2) and the top HFA high anti-human IgG sorted cell clones (n = 3). Is shown.</figref><figref num="17">FIG. 5 shows the production of clonal cell lines for Plan 14 by FACS using HFA and protein A gating. CHO-DG44 cells were transfected with a vector encoding the antibody for Design 14, then selected in bulk and amplified. The bulk population was then stained with protein A and single cells were screened. On the left is a live cell FSC-H / FSC-A plot with gating for the HFA and LFA populations, and a protein A fluorescence histogram for the LFA and HFA populations is an overview of the gates used for single cell selection. Shown below. The graph on the right shows antibody titers from the highest, 10th, 20th, and 30th highest expressed clones for different selection criteria at the 96-well stage, respectively.</figref><figref num="18">FIG. 5 shows recloning of established cell lines by FACS after either anti-IgG or protein A staining. The batch production titer after continuous recloning is shown for all cases. Panel A shows data for recloning of established cell lines for planned 3M (parent strain, BP0044 strain) after anti-IgG staining. 1st reselection (Strain C1, live high DyLight649 IgG sorting), 2nd reselection (Strain # 8, HFA high FITC IgG sorting) and 3rd reselection (Strain # 8-41, HFA high Dylight649) Best production titer subcloned data shown after IgG selection). Panel B shows (as shown in the graph) data on recloning of the established strain for Plan 12 after staining with DyLight649-conjugated IgG or Alexa647 Protein A. Again, the best productivity shown after the first selection (strain BH83, HFA high protein A selection), the second selection (BH84, HFA high IgG selection) and the third selection (BH87, HFA high protein A selection). Titer subclone data.</figref><figref num="19">It is a figure which shows the comparison of the clone derived from Clonepix (indicated by a gray bar) and the clone derived from FACS (indicated by a black bar) expressing "protein X". Clonepix and FACS-derived clones were stained with PE-conjugated anti-protein X antibody and analyzed by FACS. Each average fluorescence (MFI) is shown on the Y-axis, which is an indicator of expression level.</figref><figref num="20">Clone Cell Line Generation for Plan 14 by FACS Using HFA and Protein A Gating-Graphs showing scale-up data (see Figure 17 for selection gate and 96-well data analysis). The best shaking flask producing titers for the best clones from each selection (HFA, LFA, HFA-high protein A and LFA-high protein A) are shown. Y-axis: cumulative titer, mg / L. X-axis: number of clones / selection type.</figref><figref num="21">It is a figure which shows the production of the clone cell line for another project (M). CHO-DG44 cells were transfected with a vector encoding the antibody for design (M), selected in bulk and amplified. The bulk population was then stained with protein A and single cells were screened. HFA or HFA high protein A gates were used. The results are shown as follows: Upper panel: At 96-well titers, the top, 10, 20, and 30 rank clones are HFA (unstained) and HFA-high protein A (HHA-). High PrA) is shown. Bottom panel: Shaking flask production / growth curves for the top two HFAs (MU1 and MU5, dotted line) and HFA-high protein A (MS1 and MS2, solid line).</figref><figref num="22">FIG. 5 shows an evaluation of optimal FACS selective gating criteria after use of high FSC-A vs. SSC-A gates to define protein A labels and HFA populations. FIG. 22 shows a schematic representation of the gating used to sort single cells. Alexa647 Live cells (A-P1 gates) were analyzed for fluorescence intensity and top 10%, 5 to evaluate optimal selection gating criteria based on% fluorescence after labeling the cell surface with conjugated protein A. The gates for defining%, 2% and 1% protein A stained cells were set (B). The HFA subpopulation of cells was gated using the FSC-A vs. FSC-H dot plot (C), and finally the top% were gated from (B) copied to the fluorescence histogram from the HFA gate (D). Used for single cell sorting. To assess HFA cell gating from within the FSC-A vs. SSC-A dot plot, cells with high FSC-A and high SSC-A properties are gated (A-P9 gate) and then protein. The level of A staining (E) was analyzed. The top 5% of gates from the live population (B) were copied to the high FSC-A vs. SSC-A histogram (E) and this gate was used for single cell selection.</figref><figref num="23">FIG. 5 shows an evaluation of optimal FACS screening gating criteria after use of high FSC-A vs. SSC-A gates to define protein A labels and HFA populations. FIG. 23 shows productivity from single cell clones sorted according to the above criteria at the 96-well stage and in the batch shaking flask production curve. (A) Mean and maximum titers at the 96-well stage were for cells selected from the top 10% (white), 5% (intermediate gray), 2% (light gray) and 1% (black) fluorescent gates. Equivalent. FSC-A vs. FSC-H or FSC-A vs. SSC-A No difference in productivity is observed at this stage between cells stained with the top 5% protein A selected from any of the HFA gates. (B) IgG titers after batch production curve for the highest expressed clones for each of the selection criteria, scaled up to shaking flask culture. Although slight differences in productivity are observed for each of the selection criteria, the data show trends for levels of IgG production that correlate with levels of surface staining. This data also shows that the HFA population can also be identified by directly gated high FSC-A cells using the FSC-A vs. SSC-A dot plot.</figref><figref num="24">It is a figure which shows the comparison with the protein G staining and the protein A. Parental non-transfected cells (upper row), bulk-transfected / selected / amplified pools (middle row) and clonal cell lines (lower row) were stained and analyzed as shown. First column: FSC-SSC plot and live cell gate. Second column: Protein A staining of living cells. Third column: Protein G staining of living cells, Fourth column: Protein G staining of HFA cells. Fifth column: HFA gate.</figref>

0054In the present specification, the present invention is described with reference to embodiments to allow clear and concise description of the specification. It is intended and is to be understood that embodiments may be combined or separated in various ways without departing from the present invention.

0055Unless otherwise defined, all technical and scientific terms used herein are generally understood by those skilled in the art (eg, in cell culture, molecular genetics, nucleic acid chemistry, hybridization techniques and biochemistry). It has the same meaning as the one that exists. Standard techniques are incorporated into molecular, genetic and biochemical methods (generally incorporated herein by reference, Sambrook et al., Molecular Cloning: A Laboratory Manual, 2d ed. (1989) Cold Spring Harbor Laboratory Press, (See Cold Spring Harbor, NY and Ausubel et al., Short Protocols in Molecular Biology (1999) 4th Edition, John Wiley &amp; Sons, Inc.) and used in chemical methods.

0056The method utilizes non-animal-derived non-antibody labels such as recombinant protein A, protein G or protein L instead of animal-derived anti-antibody for product-specific cell membrane staining. Since all of the components used in the methods of the invention were produced without the use of animal-derived substances such as serum or BSA, the risk of exposure of the final product cell line to exogenous and infectious agents is significant. Is reduced to.

0057CHO A typical cell line generation protocol for DG44 cells is outlined in Figure 1A, which states that cell lines are generated by low density seeding of transfected cells. The vectors used herein are shown in Figure 1C. After G418 selection and culture growth, a further step of low density seeding is carried out by the addition of methotrexate for gene amplification. After culture growth, initial stability testing is performed prior to single cell cloning by limiting dilution of the best clones or FACS sorting (without any surface staining). Similar protocols were followed for other CHO cell lines (using appropriate selective pressure; this process can typically involve seeding and analysis of 100 or even 1000 cells (eg Wurm, 2004, supra). Although this process can successfully produce cell lines suitable for commercial production of antibodies, we can select cells stained for surface expression for selection of clones with higher IgG production. In addition to producing cloned cells, we investigated the potential for linking, and we reduced the time required for this process to identify higher secretory strains and the number of clones to be screened. I thought I would get it.

0058An improved protocol according to an embodiment of the invention is outlined in FIG. 1B, wherein selection (by G418 addition and nucleoside removal) and methotrexate gene amplification are performed in a bulk transfect population and such bulk transfection. The cell culture is rapidly transferred from the static culture to the shaking flask. The cells are then stained and the clone with the highest expression of surface IgG is selected from the bulk population as a single cell. Alternatively, or in addition, cells are also sorted based on their forward scattering (FSC) properties, thereby selecting / sorting cells with a higher average FSC-A. As with the isolation of the most highly expressed clones, this novel method allows later isolation of clones that produce high levels of IgG in batch-produced cultures. This method allows the identification of highly expressed cells for further development from as early as the 96-well stage, resulting in the analysis of less than 100 clones, and less than 24 clones, and often less than 24 clones. You get a scale-up of just 10 clones.

0059The following examples are based on data generated by a typical method using an animal-derived protein labeling method (rabbit polyclonal anti-human IgG), which is a novel method (recombinant protein) completely free of animal-derived proteins / components. Compare with (labeled with A). In addition, they demonstrate the advantage of selecting / selecting cells with higher FSC-A characteristics.

0060The flow cytometry device has several parameters such as forward scatter (FSC), side scatter (SSC) and fluorescence (FL) with the maximum number of parameters depending on the number of lasers and detectors available to the device. Parameters can be detected. FSC (forward scattering, also known as small angle forward scattering) is a measure of cell size and is defined as scattering of light from a laser at a small angle, typically 0.5-5 °. Higher FSC correlates with larger cells or events.

0061SSC (orthogonal scattering, 90 ° scattering, lateral scattering, also known as wide-angle scattering or large-angle scattering) is defined as the deflection of light from the laser between 15 and 150 ° and is associated with cell grain size. Be done. Therefore, specific cells with selected characteristics can be selected by gated the data plot generated by the cytometer.

0062Each parameter in flow cytometry is characterized by three values: the most widely used values: signal area (A), signal height (H) and signal width (W). When the signal is plotted against time, its area, width and maximum height are determined as shown in FIG. 9A. Examples of FSCs are FSC-A (also referred to as FSC), FSC-H and FSC-W. For the purposes of this application, if there is no -A, -H or -W next to the name of the parameter, it refers to the area value.

0063As used herein, the term HFA or (high forward scattering region) refers to a larger (larger) cell (also known as an "event") within a population. These cells are examined microscopically and consist primarily of large single cells, which do not aggregate. They can be mitotic / proliferating cells. As used herein, selecting highly expressed cells from this subpopulation is when selecting from smaller cells within the population (or simply when selecting based on high fluorescence / signal in product-specific direct staining). It is shown to result in an increased likelihood of selecting cells that exhibit higher polypeptide production. The exact size of these cells will, of course, vary depending on the cell type, growth conditions and the like. HFA or large cell subpopulations can be established by one of ordinary skill in the art for each particular population of cells. Thus, as used herein, the steps of selection include identifying another cell population (HFA population) within a living cell population and a FACS cytometer or analyzer to select cells from this other population. It may include a step of gated with and a step of selecting cells from this other population. As mentioned above, there are various methods in which cells of this subpopulation can be gated and selected in the cytometer.

0064For any particular population of cells, the HFA cell subpopulation is a living cell that is higher than mean FSC-W or higher than mean FSC-A. As mentioned above, the number of cells selected depends on the number of cells that are ultimately intended to be cloned and cultured, and thereby the available resources and time. One of ordinary skill in the art can define selection criteria to select the largest cell from within the cell population.

0065However, typically, the HFA subpopulation has the highest FSC-W values in the living cell population of 20%, 19%, 18%, 17%, 16%, 15%, 14%, 13%. , 12%, 11%, 10%, 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2% or 1% or less.

0066The mean FSC-A of the HFA cell subpopulation is also higher than the mean FSC-A of all living cells in the population. Typically, the HFA subpopulation has the highest mean FSC-A values of the 20%, 19%, 18%, 17%, 16%, 15%, 14%, 13% of the living cell subpopulations. , 12%, 11%, 10%, 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2% or 1% or less.

0067The average SSC-A can also be higher than the average. The HFA subpopulation can also contain many SSC-A cells.

0068Typically, the FSC-H of the HFA cell subpopulation is substantially identical to the FSC-H of all living cells in the population. In one embodiment, the average FSC-H of the HFA cell subpopulation is approximately 100% (+/-) 10% (+/-) 10%, 5%, 4%, 3%, 2%, 1% of the remaining cells in the living cell population. Or 0% FSC-H.

0069In one embodiment, the HFA cell subpopulation is the top 20%, 19%, 18%, 17%, 16%, 15%, 14 of the living cell populations with the highest FSC-W and highest mean FSC-A. %, 13%, 12%, 11%, 10%, 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2% or 1% cells, the population of which is optional. Has substantially the same FSC-H as the average of.

0070Living cell populations can be determined and gated by those of skill in the art by FSC-A vs. SSC-A plots. Optionally, they can also be determined in additional plots as PI-negative cells upon staining with PI (propidium iodide). The HFA subpopulation of living cells is FSC-H vs. FSC-A plot (Fig. 9C), FSC-W histogram (Fig. 10A), FSC-W vs. FSC-A plot (Fig. 10B), or optionally FSC-A. It can be determined and gated by the vs. SSC-A plot (Figure 22).

0071In one embodiment, the mean FSC-A of the HFA subpopulation is 1.8 (+/- 0.1) times that of the LFA subpopulation. In one embodiment, the mean FSC-H of the HFA subpopulation is 1.1 (+/- 0.1) of the LFA subpopulation.

0072In contrast, the term LFA (Low Forward Scattering Region) refers to smaller (smaller) cells in a population. Consistent with the above, one of ordinary skill in the art can determine the LFA cell population for any given cell population.

0073As described herein, a cell with "high" or "highest" expression is a cell within a cell population that expresses a higher or highest amount of the polypeptide of interest on the cell surface. Typically, such cells are selected for cloning as they are considered more likely to yield high titers of the polypeptide of interest when grown under large-scale culture conditions.

0074The cells in the population may be further divided into HFA and LFA cells. According to one embodiment of the invention, HFA cells are also selected because they are observed to produce high titers of the polypeptide of interest when proliferated under culture conditions. In a further embodiment, high expression HFA is selected.

0075The number (or proportion) of cells selected as a "highly expressive" from a population is, of course, the number of cells in the population as a whole, the range of polypeptide expression within the cells within that population, and the cells selected. Depends on the number. Typically, one of ordinary skill in the art will have 5-20% of the highest expression in the cell population, and optionally 15%, 10%, 5%, 4%, 3%, 2% or 1 of the highest expression in the cell population. Select%. That is, those skilled in the art will appreciate about 20%, 15%, 10%, 5%, 4%, 3%, 2% or 1% of the cell population containing the highest amount of label bound to the polypeptide on the surface of the cell. Or select less than 1%.

0076As used herein in relation to "relative size" or "relative amount" (eg, of a label bound to a polypeptide on the surface of a cell), the term "relative" is presumed for the entire population of cells. It is intended to mean the value of the variable for a particular cell as compared to the value of the variable. Usually, a particular percentage of cells is selected by the operator within the population, such as cell size or fluorescence, to define appropriate gating values for selecting selected cell subpopulations prior to the cell sorting process. It is analyzed with a cytometer that can analyze the variable object. During the sorting process, the cytometer determines the value of the variable (ie, size or fluorescence) for a particular cell as compared to the value of the variable for the cell population as a whole. Cells in FACS are defined as individual events recorded and displayed by flow cytometers.

0077The term "antibody" as used herein in the broadest sense refers to a molecule having an immunoglobulin-like domain, monoclonal, recombinant, polyclonal, chimeric, humanized, bispecific and heteroconjugate. Gate antibody; single variable domain, domain antibody, antigen-binding fragment, immunologically effective fragment, single-stranded Fv, diabodies, Tandabs , etc. (For an overview of alternative "antibody" forms, Holliger and Hudson, Nature Biotechnology, 2005, Vol 23, No. 9, 1126-1136). Conversely, the term "non-antibody" that describes the labeling used in the methods of the invention indicates the absence of any such molecule in the labeling. For the avoidance of doubt, Protein A, Protein G and Protein L are non-antibody (and therefore also non-polyclonal antibody) labels.

0078The phrase "single variable domain" refers to an antigen-binding protein variable domain (eg, V) that specifically binds an antigen or epitope independently of different variable regions or domains._H, V_HH, V_L).

0079A "domain antibody" or "dAb" can be considered similar to a "single variable domain" capable of binding an antigen. The single variable domain may be a human antibody variable domain, but also rodents (eg, as disclosed in WO 00/29004), carpet sharks and camel Vs._HH Contains single antibody variable domains from other species such as dAbs. Camels V_HHIs an immunoglobulin single variable domain polypeptide derived from species including camels, llamas, alpaca, dromedary, and guanaco that produce heavy chain antibodies that naturally lack the light chain. Such V_HHDomains may be humanized according to standard techniques available in the art and such domains are considered "domain antibodies". As used herein, V_HIs a camel V_HHIncludes domain.

0080As used herein, the term "domain" refers to a folded protein structure that has tertiary structure independent of the rest of the protein. In general, domains are involved in the distinct functional properties of proteins and can often be added, removed, or transmitted to other proteins without loss of function of the protein and / or the rest of the domain. it can. A "single variable domain" is a folded polypeptide domain containing a characteristic sequence of an antibody variable domain. Therefore, is the single variable domain a complete antibody variable domain, as well as a modified variable domain in which, for example, one or more loops are replaced by a sequence that is not characteristic of the antibody variable domain, or is the terminal truncated? Alternatively, an antibody variable domain having an N- or C-terminal extension, and a folded fragment of the variable domain that retains at least the binding activity and specificity of the full-length domain are included. Domains can bind antigens or epitopes independently of different variable regions or domains.

0081A clonal population is a population of cells that are descendants of a single common ancestor, such as cells of a single origin isolated from all other cells. As used herein, clonal conditions are suitable conditions to ensure that only one cell can be successfully increased within a clonal population. Typically, this is done by placing only one cell in a single vessel or well in which the culture environment in the vessel or well can support clonal growth.

0082The following method was used in the examples described later.

0083Vectors used in this study: Standard mammalian expression and selection cassettes for expressing biologic antibodies (which can bind proteins A, G or L) and species and other proteins associated with the antibodies in mammalian cells. Was used in all aspects of this study (see Figure 1C for a representative schematic of the vector). Standard oligonucleotides and gene synthesis, cloning, subcloning and sequencing methods were utilized in the design for such expression cassettes. During the ORF design of protein products, codon optimization was performed according to the method described by Kotsopoulou et al. (Kotsopoulou E et al. J Biotechnol 2010; 146 (4): 186-93 and WO 2009/024567 A1). Such codon optimization, with a typical ORF CAI greater than 0.9, helps ensure that expression is high enough.

0084Cell delivery of expression cassettes: There are many well-known methods in the art suitable for delivery of one or more desired expression cassettes to a desired cell host. Some require transfection, while others utilize other methods such as viral vector transduction. Any such method may be utilized for the purposes of the present invention. For the expression of antibodies and antibody-related biopharmaceuticals typically described herein, we utilize a plasmid-based expression cassette, which is then used with lipids (eg, lipofectamine, Invitrogen). , Or by electroporation (eg, Gene Pulser II, Biorad), or most preferably by a combination thereof (eg, Amaxa, Lonza). The Amaxa method can ensure a sufficiently high transfection rate by us to ensure high expression for rapid selection of stable polyclonal pools and preferably for detection with Protein A stain. Although found, electroporation has also been used successfully in this regard.

0085Cell Host and Tissue Culture: Any mammalian cell can be utilized for the purposes of the present invention. Typically, Chinese hamster ovary (CHO) cells have a long and proven track record in producing biopharmaceuticals and therefore utilize them to express biopharmaceutical protein products. There are many different mutant CHO substrains, but most typically the CHO-DG44 or CHO-K1 host is utilized. In addition to CHO cells, other cells utilized to make biologics generally include bacterial cells, yeast cells, immortalized human cells (eg, 293 cells or PerC6 cells), and immortalized non-human mammalian cells. (Eg NS0) is included. Standard ADCF medium was used for culturing such cells as described herein.

0086Electroporation transfection: DNA preparation: Equal volumes (15 μg) of heavy and light chain expression vectors were linearized to completion (using Not I) in a 200 μl volume of Eppendorf reaction, followed by ethanol / sodium acetate precipitation. The pellet was then washed in 70% ethanol, air dried and resuspended in 50 μL of molecular biology grade water.

0087b. Preparation of CHO DG44 cells prior to transfection: healthy growing cells, 1.2 × 10⁷Cells (per transfection) are centrifuged in 15 or 50 ml tubes (2-10 minutes, 1000 rpm), washed in 15 ml ice-cold PBS / sucrose, centrifuged again, and then re-into 800 μl ice-cold PBS sucrose. Suspended. The cell suspension was then added to the pre-prepared DNA, placed on ice for 15 minutes and then transferred to a chilled electroporation cuvette.

0088c. Electroporation: The cuvette containing the prepared DNA and cells was electroporated in a Gene Pulser set at 25 μF and 0.38 kV and then returned to ice for 10 minutes. Cells are then removed and added to 240 ml non-selective medium, then in 40 x 96 well dishes 2-5 x 10 in non-selective medium.³Seeded at cells / well (ie, 50 μL / well). Then plate the plate at 37 ° C and 5% CO₂Incubated for 48 hours.

0089d. Selection, Proliferation and Clone Identification: 48 hours after electroporation, 150 μL of selective medium was added to each well. This selective medium contains G418 and does not contain nucleosides. Optionally, 1 week later, 140 μl of medium was carefully replaced with fresh selective medium so as not to disrupt the colonized cell layer, and 3-4 weeks later, all growing clones (typically 0.1). Colony growth / well, i.e. growth in 10 wells per 96-well plate), was titrated for antibody production. The top ranked clones identified (typically 20-100) were scaled up from 24-well dishes to 6-well dishes in the same selective medium. These clones were then seeded in 96-well dishes (96-wells / clones) at 1000 cells / well and then selected in selective medium also containing 5 nM methotrexate in a volume of 200 μl / well. After an additional 2-3 weeks of incubation, the best clones were scaled up again. Second and sometimes third rounds of amplification were performed with 50 nM and 150 nM methotrexate to obtain high titer cell lines. If the ORF to be expressed was codon-matched (as described in Kotsopoulou E et al. J Biotechnol 2010; 146 (4): 186-93 and WO 2009024567 A1), a single round of amplification at 5 nM was sufficient. .. To evaluate the final clones for productivity, then the best clones in the final methotrexate amplification step were scaled up and the titers and qualities of the products produced were evaluated in a shake flask production model.

0090Amaxa transfection: CHO DG44 (or CHO-K1) cells were passaged the day before transfection. On the day of transfection, cells were counted by ViCell (Beckman-Coulter), pelleted by centrifugation at 1000 rpm for 5 minutes, and the cell pellet was washed once with sterile PBS. CHO cells are then placed in Amaxa nucleofection buffer 10⁷~10⁸Resuspend at cells / mL, then 10⁶~10⁷1-6 μg of cells encoding the antibody of interest using equal volumes of heavy and light chain plasmids, typically using the protocol recommended by the manufacturer for CHO cell suspensions. The light chain plasmid and 1 to 6 μg of the heavy chain plasmid were used to transfect sterile cuvettes.

0091Bulk Selection and Amplification: Immediately after transfection, several separate Amaxa transfections (up to a total of 30) were pooled in 1-3 pools, added to fresh medium and sterile T75 cm.²It was grown in a tissue culture flask. For electroporation, a single electroporation (as described above) was used as a single pool of cells. No selective medium was used at this stage. 48-64 hours after transfection, G418-containing, nucleoside-free medium was added for transfection agent selection. The culture is then counted twice a week and medium added / replaced as needed to adjust volume to maintain viability at cells / ml greater than 0.5e6, if possible. did. As soon as possible, usually between 5-14 days after transfection, depending on the number of viable cells, transfer the cells to a shaking flask 37C, 5% CO_2、It was grown at 130-150 rpm. After harvesting more than 70% of viable cells, methotrexate was added to the medium to a final concentration of 5 nM, usually 2.5-4 weeks after transfection. At this stage, continuation of G418 in the medium is optional. After exposure to methotrexate for a period of 10-25 days, cells were analyzed for surface antibody expression, followed by single cell selection using FACS Aria I (Becton Dickinson).

0092Preparation of fluorescent reagents (IgG and protein A) for use in FACS DyLight649-IgG: Affinity-purified polyclonal anti-IgG (Rockland, catalog number # 609-4102) produced in rabbits was labeled using the commercially available kit DyLight649 Thermo catalog number # 53050 as follows: 50 mM Antibodies in sodium borate buffer (pH 8.5) were added to DyLight649 vials for labeling and mixed gently by inversion 10 times. Samples were incubated in the dark for 60 minutes at room temperature before being passed through a filter column to remove unbound dye. Labeled samples were collected, IgG concentrations were measured, and the samples were aliquoted for storage at -20 ° C.

0093FITC-IgG: A polyclonal anti-IgG conjugated to affinity-purified fluorescein (FITC) produced in rabbits was purchased from Rockland (catalog number # 609-4202).

0094Alexa647-Protein A: This was purchased from Invitrogen (catalog number P21462). Protein A was reconstituted from lyophilized proteins in sterile PBS up to 1 mg / mL and aseptically aliquoted prior to freezing at -20 ° C before use. On the day of use, the frozen aliquots were thawed just prior to staining, further diluted to 0.1 mg / mL working stock with additional sterile PBS, and then added to the cells to be stained to the appropriate amount.

0095As an alternative and specifically described in the examples below (as Repligen Protein A), Protein A was purchased from Repligen and, as per the manufacturer's instructions, the Invitrogen Alexa Fluor 647 Labeling Kit (Catalog No. A20173). ) Was used for labeling.

0096-Although the fluorescent labeling reagent was used as described above, indirect labeling can also be used. For example, the biotinylation reagent can be used with fluorescent dye conjugated avidin or streptavidin as a secondary reagent for FACS analysis / sorting.

0097Cell staining protocol: Staining with FITC or DyLight 649 conjugated anti-IgG: To remove aggregates of cells that could block the flow of FACS, the cells to be stained were passed through a 40 μm sieve of sterile nylon before counting. After counting, cells were washed once or twice in sterile PBS and then incubated with fluorescently labeled anti-product IgG. Washing removes secreted non-cell bound IgG products that can reduce the fluorescent signal. Wash 10 in each FACS tube⁶The exact volume of cells was pipetted through the cells / tubes, then PBS was added to the cells and the cells were centrifuged at 1000 rpm for 5 minutes. After decanting the final wash supernatant, anti-human IgG antibody was added directly to the cell pellet and incubated on ice for 20 minutes in the dark. Optimal staining of cells with these proteins, 5-10 μg of anti-human IgG / 10⁶Achieved in cells (see Figure 5) and very good signal-to-noise ratios were obtained. After staining, one or two washes are performed in PBS or cell proliferation medium, after which the cell pellet is resuspended in growth medium and tubed on ice in the dark until ready for analysis and / or sorting. Was saved.

0098b. Staining with Alexa647 Protein A: To remove populations of cells that could block the flow of FACS, the cells to be stained were passed through a sterile nylon 40 μm sieve before counting. After counting, cells were washed once or twice in sterile PBS and then incubated with fluorescently labeled protein A. Washing removes secreted non-cell bound IgG products that can reduce the fluorescent signal. Wash 10 in each FACS tube⁶The exact volume of cells was pipetted through the cells / tubes, then PBS was added to the cells and the cells were centrifuged at 1000 rpm for 5 minutes. After decanting the second wash supernatant, Protein A was added directly to the cell pellet and incubated on ice for at least 20 minutes in the dark. We obtained more low signals than those achieved with IgG reagents. In addition to the higher background, staining was observed and as a result, the signal-to-noise ratio was lower than that for IgG. However, 0.25 to 1 μg of protein A / 10⁶Optimal staining of cells was achieved when cells were used (see Figure 5). After staining, two washes were performed in PBS or cell proliferation medium, then the cell pellet was resuspended in 600 μl proliferation medium / tube and on ice in the dark until ready for analysis and / or sorting. Stored the tube in.

0099Sorting protocol: Flow cytometry: For all FACS-based sorting, we used FACS Aria (Becton Dickinson, Oxford, UK) to use 70% ethanol to ensure the sterilization of the clones after sorting. A sterilization cycle was performed using. After flushing the equipment with sterile PBS to remove residual ethanol, a drop delay was set using Becton Dickinson Accudrop beads and Diva software v5.0.3 to optimize the equipment for sorting. .. Prior to FACS sorting, unstained and / or stained polyclonal pools of cells or established cell lines were screened through a 40 micron mesh of sterile nylon to remove large populations of cells that blocked flow. Unstained and stained parental (non-transfected) CHO cells were used to set the detection voltage and for accurate gating as described in the sections below.

0100FACS Aria setting: Samples were sorted using a 100 μm nozzle with a low setting for single cell sorting and low sheath fluid pressure. Laser light was passed through a 1.5 neutral density filter for best resolution of CHO cells in the FSC-vs. SSC plot. Signals from a 488 nm blue laser were detected using a 530/30 or 575/25 bandpass filter for FITC and PI, respectively. Fluorescence from APC was captured after excitation from a red diode 633 nm laser using a 660/20 bandpass filter. Stable transfected (either selected or selected and amplified) cells for selection are analyzed by reading and recording sample data and gates for optimized selection on stored data. did. Samples to be sorted were kept on ice in the dark to minimize internalization of staining proteins and degradation of fluorescent signals until required for sorting. The capture plate is prepared on the day of sorting and at high humidity 37 ° C, 5% CO until required.₂Maintained in. The FACS Aria apparatus was optimized for delivery to the center of each well of a 96-well capture plate. Immediately after the plates had been sorted, the sorted plates were returned to the incubator, usually within 1 minute.

0101Sorting Strategy for Identifying and Isolating the Desired Cell Population of the Invention: FIG. 2 provides a flow diagram of the gating strategy determination performed in the following examples.

0102Gates were set using appropriate unstained or stained control cells (untransfected CHO-DG44 cells and / or unstained samples). Samples were recorded and a live cell gate was first established. Propidium iodide staining of cells (50 μg / mL purchased from Becton Dickinson) PI stock 7 μl / tube) was also performed in some cases to distinguish between live and dead cell populations. However, PI staining was not performed in most cases because FSC-SSC gating was found to be sufficient to distinguish between live and dead cells. Further dot plots of FSC-A vs. FSC-H were set up for some experiments and gates were drawn to select the HFA subpopulation as described in Example 8. FITC-A or APC-A histograms were then set for live gate cells and, in some cases, HFA and LFA cells. In most cases, a sorting gate was set up among the top 5% or 2% of fluorescent "live" cells, and if necessary, this gate was copied onto the FITC-A or APC-A histogram of LFA and HFA cells. .. FIG. 11 shows some illustrated sorting plots. FSC-SSC gating is also suitable for selecting HFA subpopulations as shown in Example 14 and FIG.

0103IgG titer analysis: The two methods described below were used for 96-well and some 6-well titer analysis: For medium samples obtained from IGEN: 96-well plates, antibody titers were automated with an IGEN M-Series M8 / 384 analyzer (Bioveris, MaryLand, USA) using manufacturer's recommendations and standard methods 96. Well sandwich was determined by the ELISA style method. The sandwich consisted of streptavidin-coated magnetically coated beads, biotinylated protein A and ruthenium-labeled F (ab) 2 fragments. The signals generated for the test sample were then compared to serial dilutions of antibody reference. Although the assay is sensitive due to assay variability combined with cell proliferation variability in 96-well cultures, the accuracy and reproducibility of the assay intermediates is relatively low for this assay for high yield amplified cell lines.

0104b. MSD ELISA method: Soluble IgG measurement is performed by MSD assay (Mesoscale Discovery,) using a protein A-coated ELISA plate. USA). The assay was performed according to the manufacturer's instructions. Briefly, the plate was first blocked and a special blocker (3% blocker A) was used to reduce non-specific binding for 60 minutes at room temperature. The plates were then washed 3 times in a row with 150 μl / well PBS containing 0.05% Tween-20 and sucked with a paper towel during the wash. IgG control standards were prepared and seeded at 25 μl / well at the following concentrations: 5, 4, 3, 2, 1, 0.5, 0.25, 0.125, 0.0625, 0.031, 0.015 and 0 mg / L. The sample to be tested was also added to 25 μl / well. The Tag-IgG detection antibody was diluted to 0.255 μg / mL in antibody diluent and added at 75 μl / well. The plate was shaken at room temperature for 60 minutes at 750 rpm. More than 3 washes were performed as above, 150 μl / well of 1 × reed buffer was added, and the plates were read with an MSD1200 plate reader. IgG titers for unknown samples were calculated back from the standard curve using MSD plate reader software.

0105c. In the reaction solution using the Beckman Coulter Imaging System (Buckinghamshire, UK) and manufacturer's recommendations and standard methods for media samples and bioreactor production models obtained while shaking the flask. Antibody titers were measured using a turbidimetric analysis method in which the optical signal was scattered by the insoluble immunoprecipitant. The signal generated for the test sample is again compared to the serial dilution of antibody reference criteria.

0106The bioreactor shaking flask production model (expanded batch production model) is also referred to as the growth curve: typically a standard 250 ml tissue culture shaking flask with a vented lid at a volume of 120 ml or a volume of 60 ml. Cells were seeded at 200,000-800,000 cells / ml in medium free of animal-derived components in any of the standard 125 ml tissue culture shake flasks with vented lids. They were then incubated in carbon dioxide enriched air with agitation to a temperature that promoted and maintained cell proliferation. Each clone was tested under different conditions, such as different temperature conditions. The results reported herein illustrate the highest titers for each clone tested (over standard conditions). Typically, the production model endpoint titers, as reported herein, are standard Vi-Cells. Recorded when cell viability dropped to about 50%, as determined by trypan blue dye exclusion-based assay in Vi-Cell (Beckman) using CHO parameter settings and manufacturer-recommended protocols. This endpoint titer is typically generated after 10-20 days of incubation.

0107Small-scale production (Duetz) model (extended batch production model): For the bioreactor model described above. Typically, cells were seeded at 800,000 cells / ml in a final volume of 1.5 mL in a 24-well plate in medium free of animal-derived components.

0108Cell line stability assessment: Incubate the cell line up to 50 passages in the cell line's associated growth medium and use the batch product curve for approximately 10 passages as described in the paragraph above to assess productivity trends. Set for each.

0109Clone proliferation and selection after single cell cloning: Various methods exist in the literature on cloning mammalian cells (see, eg, Brezinsky 2003, ibid. Or Hugin 1997, J Immunol Methods, 205, 211-212). .. Such a method is suitable for use of the present invention. Further novel methods for cloning mammalian cells are described below and are used herein. The following methods / conditions are used for the purposes of the present invention and are used only by way of example.

0110Single cells were precipitated in individual wells in 96-well plates by either serial dilution or FACS sorting as described in the Examples. The plate was pre-filled with either 100-200 μl / well of 25-50% cell-free conditioned medium, unconditioned medium, or support cell-containing medium. The conditioned medium was generated by collecting fresh cell proliferation medium from 3-4 day old cultures of growing untransfected parent cells, centrifuging and mixing with 0.2 μm filtered medium. The unconditioned medium is a fresh medium having the same composition as the medium in which the cells were grown prior to cloning, with or without a supplement such as transferrin or recombinant albumin, as described in Examples. Met. When used, the supporting cells were usually untransfected parental CHO cells precipitated at 500-2500 cells / well 0-24 hours prior to cloning. In the following examples, selection and amplification agents, 400-1000 μg / ml G418 and a final concentration of 5 nM methotrexate, were added 0-64 hours after single cell precipitation and then maintained in medium. Optionally, after precipitation of single cells in 96-well plates, once a week, 140 μl of medium is carefully replaced with fresh medium so as not to disrupt the colonized cell layer, and after 2-4 weeks, All growing clones were scaled up to 24 wells. In some cases clones were randomly selected (mainly for established strains), in other cases up to 300 were titrated for antibody production in 96 wells, only in small amounts, typically 20-100 to 24. Scaled up to the well. Clone evaluations then followed conventional cell line production protocols and production evaluations for scale-up.

<p num="0111"> Unstained Cell Selection for Cell Surface IgG Expression-Subcloning of Existing Cell Lines Based on the (SCC / FSC) +/- (PI / FSC) plots (plans 5d, 7b, 10), early FACS selection was performed exclusively on live cell gating in established cell lines expressing monoclonal antibodies. Carried out. When cells are cloned using standard single cell selection protocols, the results in Figure 3 (a) show the ability to clone out hypersecretory forms in the absence of any anti-IgG cell surface staining. The data show a double titer for Plan 5d, a 1.5-fold increase for Plan 10, and a 3.5-fold increase in titer after two consecutive rounds of planning 7b. Plan 6d was cloned by limiting dilution.</p><p num="0112"> In addition to allowing possible hypersecretory selection, single cell selection also induces strains with good stability over 50 passages as shown in Figure 3 (b) for Plan 7b. To guide. Recommended for data generated from these FACS studies, subsequent studies focus on the highest secretory forms from stably transfected populations and / or established cell lines and attempt to clone them out for surface IgG. I proceeded with the dyeing of.</p>

<p num="0113"> Detection of expression levels by bulk transfection / selection / amplification and staining of cell surface IgG expression.</p><p num="0114"> The latest direct staining FACS sorting protocol for identifying and sorting high-producing cell lines expressing biopharmaceuticals such as antibodies utilizes animal-derived components. Typical protocols are described below.</p><p num="0115"> Affinity-purified anti-human IgG (H + L) (Rockland, Gilbertsville, PA 19525 USA) in a buffer containing bovine serum albumin was used. The antibody was initially produced as rabbit antiserum and purified by immunoaffinity chromatography using human IgG bound to agarose beads prior to final formation in buffer containing bovine serum albumin. The antibody was conjugated to FITC or DyLight 649.</p><p num="0116"> As described in the method section above, CHO DG44 cells were transfected using the Amaxa transfection regime with plasmids encoding the light and heavy chains of Plan 10. Transfected cells were selected and amplified in bulk. FIG. 4 shows the fluorescence levels achieved after antibody staining of non-transfected cells (FIG. 4A) and cells transfected after G418 selection (FIG. 4B) and after both selection and methotrexate (5nM) amplification (FIG. 4C). Shown. This figure demonstrates that cell surface IgG expression is detectable above background levels using this method. In fact, even before amplification, a "shift" of fluorescence is sufficient to separate transfected cells from non-transfected cells.</p>

<p num="0117"> Protein A antiserum staining We then compared surface staining using both Protein A and anti-human IgG (polyclonal). We have observed that staining with protein A results in a reduced signal and a lower signal-to-noise ratio. This was believed to be due to the reduced affinity of protein A for standard IgG staining reagents. However, we have attempted to optimize the staining protocol. To this end, an established cell line (Plan 10) was used to define the optimal amount of protein A to use. The cells were washed twice prior to staining to remove soluble secreted antibody from the transfected cells, then 10⁶Stain a sample of cells with either a variable amount of antibody or protein A, then wash off excess antibody / protein A with two additional washes and on ice until cytometry is performed. Maintained in. The fluorescence levels detected using this staining regime are shown in FIG. Optimal amount for staining is 5-10 μg / 10 for anti-human IgG⁶0.25 ~ 1 μg / 10 for individual cells or protein A⁶Calculated as individual cells, in all cases staining was performed in the dark on ice for 20 minutes.</p><p num="0118"> Even after optimization, the data in FIG. 5 show that protein A gives lower levels of fluorescence than polyclonal anti-IgG antibody. One explanation for this is that due to its polyclonal nature, there is likely to be a higher number of epitopes recognized by the antibody.</p><p num="0119"> Subsequently, the signals produced by IgG vs. Protein A staining were analyzed for both established strains and bulk transfections. Figure 6 shows the results of the examples for each of Plans 12 and 14. They show that protein A staining can be easily visualized in the established strain. Surprisingly, the bulk heterogeneous population was stained to a lesser extent, but the fluorescent staining achieved with protein A was sufficient to isolate the brightest stained cells from the background staining. Given the benefits given for antisera, work was continued on bulk and established cell lines.</p>

<p num="0120"> Correlation between high protein A fluorescence and identification / selection of high level IgG-secreting cells.</p><p num="0121"> Isolation of high IgG-secreting clones after labeling with anti-human IgG antibody is a method already used for biologic cell line production (see Bresinsky et al., 2003, ibid.), Which is expensive. It has been shown that fluorescence corresponds to higher titers. To validate our staining protocol, we transfected with a dAb-Fc fusion molecule (Plan 17), especially with protein A, which actually identifies high-producing cells. Four-way sorting was performed on CHO-K1-based CHROMOS cells (Lindenbaum et al., NAR, 32, e172, 2004) to compare IgG staining with protein A staining.</p><p num="0122"> Rinse the soluble antibody and place the cells in 4 μg / 10 on ice in the dark for 20 minutes.⁶Anti-human IgG or 0.5 μg / 10 in individual cells⁶Incubated with protein A in individual cells. Cells were sorted after two additional washes in PBS. Four-way enhanced sorting was performed based on gating over the fluorescence histogram plots of living cells (Fig. 7). 100,000 cells were sorted per gate and subsequently grown in sterile 24-well culture plates for 5 days. Soluble IgG titers were then measured by MSD and the readings were divided by the number of viable cells in the wells to give an index per cell production. Fluorescent signals were reduced about 10-fold with protein A staining, but this data clearly shows between the level of protein A staining and IgG secretion in cells similar to that seen with anti-human IgG labeling. Demonstrate the correlation.</p><p num="0123"> A second source of protein A (Repligen protein A) labeled in-house as described in the method section was also successfully used (data not shown).</p>

<p num="0124"> Sorting with IgG was successful, but methotrexate-amplified bulk transfection vs. clone growth from pre-cloned strains was reduced when using cell-free capture plates.</p><p num="0125"> In general, when the plate contains supporting cells, when the cells are single cells sorted in the capture plate, the supporting cells promote the proliferation of individually cloned cells (eg Hugin 1997, J Immunol Methods, 205). , 211-212). To maintain the protocol, a simple, regenerative cell-free capture plate was utilized. For this purpose, only the capture plate is a 50% conditioned medium containing G418 and an appropriate amount of MTX (ie, 50% fresh medium and filtration supernatant of a long-term culture of 3-4 days of untransfected CHO cells 50). Consists of%).</p><p num="0126"> After anti-human IgG staining, bulk transfection amplified with methotrexate (5 nM) for antibody design 14 was subjected to single cell FACS selection. The same single cell selection protocol was used for established strains (Plan 3M, strain D16) and the results are shown in Table 1 below. Plate capture data are suitable for conditions to reclone established cell lines and to screen bulk-transfected cells, but optimization is desired to increase the number of clones generated from bulk-transfection. Suggest that it can be done.<tables num="1"><img id="000002" he="71" wi="159" file="JP6039552B2_D0001.tif" img-format="tif" img-content="drawing" /></tables></p>

<p num="0127"> Optimization of capture plate conditions for clonal growth after sorting To increase the number of clones produced after selection of single cell clones from bulk transfection, we compared various capture plate conditions with widely used supporting cell standards. CHO DG44 cells were transfected with a plasmid encoding the antibody for Design 13. Transfected cells were then subjected to G418 selection and nucleoside depletion, followed by 5nM methotrexate for gene amplification. Six different capture plate conditions were compared: 1) 2000 DG44 CHO cells / well in medium containing G418 and methotrexate added 64 hours after sorting; 2) 2000 DG44 CHO cells / wells in medium containing G418 and methotrexate present prior to sorting 3) 50% conditioned medium containing G418 and methotrexate added 64 hours after sorting 4) 50% conditioned medium containing G418 and methotrexate present before sorting 5) 50% conditioned medium containing G418 and methotrexate present prior to sorting with recombinant serum albumin and recombinant transferrin; and 6) Fresh medium containing G418 and methotrexate present prior to sorting with recombinant serum albumin and recombinant transferrin.</p><p num="0128"> G418 was used at 400 μg / mL, methotrexate at 5 nM, recombinant serum albumin at 1 mg / mL and recombinant transferrin at 5 μg / mL. Medium and G418 were sourced from Invitrogen, UK. Methotrexate was prepared from a solid (Sigma, UK) as a 50 μM stock, left free of animal protein, and both recombinant albumin and transferrin were supplied as recombinant protein from Novozyme, UK. Five plates were sorted for conditions 1-4 and 6 and three plates were sorted for condition number 5. Table 2 shows the number of clones generated per plate. The number of colonies generated per plate was counted 20 days after single cell selection.<tables num="2"><img id="000003" he="77" wi="158" file="JP6039552B2_D0001.tif" img-format="tif" img-content="drawing" /></tables></p><p num="0129"> This data shows that the inclusion of G418 and methotrexate in the capture plate medium yields fewer clones both in the supporting cell plate and under cell-free capture plate conditions. The inclusion of recombinant human serum albumin and transferrin in the conditioned medium resulted in an increase in the number of clonal growths throughout the period. Clone proliferation observed in the absence of supporting cells may be replaced by recombinant albumin and transferrin traps and / or cell-cell contact is less important than soluble factors such as those present in conditioned medium. Show that. The absence of supporting cells in the plate improves the visual identification of clones in the 96-well plate and allows for their earlier identification.</p><p num="0130"> The above data indicate that the addition of higher albumin and / or transferrin to the medium containing the capture plate can increase the number of clones and eliminate the need for the use of conditioned medium. These data, in conjunction with previous examples, show that a variety of different conditions can support single-cell cloning of CHO cells throughout, albeit at varying rates of cell proliferation.</p><p num="0131"> Subsequently, the generated colonies were analyzed for antibody expression. The results shown in Table 3 show higher titers when selective pressures (G418 and MTX) are added 2 days after cloning (comparing condition 3 to 4) when using media without a support layer. Indicates that it can be obtained.<tables num="3"><img id="000004" he="51" wi="158" file="JP6039552B2_D0001.tif" img-format="tif" img-content="drawing" /></tables></p><p num="0132"> To further improve cloning conditions and taking into account the promising results obtained above, another experiment was designed, thereby transfecting CHO DG44 cells with a plasmid encoding the antibody. Transfected cells were then subjected to G418 selection and nucleoside depletion, followed by 5nM methotrexate for gene amplification. Three different capture plate conditions were compared (cells were sorted based on viable cell gates and 10 96-well plates were used per condition): 1) 50% conditioned medium containing G418 and methotrexate present prior to sorting with recombinant serum albumin and recombinant transferrin; 2) 50% conditioned medium containing G418 and methotrexate added 48 hours after sorting as well as recombinant serum albumin and recombinant transferrin present prior to sorting; 3) 50% conditioned medium containing methotrexate and G418 added 48 hours after sorting, recombinant serum albumin and recombinant transferrin present prior to sorting; After 3 weeks, single cell sorted colonies were counted and antibody titers were measured. The results summarized in Table 4 again show the clear advantage of applying selective pressure (G418 and MTX) 2 days after sorting.<tables num="4"><img id="000005" he="48" wi="158" file="JP6039552B2_D0001.tif" img-format="tif" img-content="drawing" /></tables></p>

<p num="0133"> Better response for early recovery and batch production of FACS-selected clones when compared to clones derived from conventional methods With respect to Plan 13, we found that (a) low density seeding (conventional protocol, Figure 1 (A)) and (b) single cell selection of highly fluorescent cells from stained amplified bulk pools (new). The recovery of cells selected after the protocol, Figure 1 (B)) was compared. We investigated viability at scale-up and achieved cumulative IgG titers from T75 cm flasks in batch small scale (Duetz) production model settings. The data is shown in Fig. 8. Surprisingly, there was more rapid recovery and was suitable for shake culture of clones derived from the FACS protocol (see Figures 8 (A) and 8 (B)). The majority of FACS clones were recovered with viability greater than 80% at the T75 stage and above 90% at the first passage in the shaking flask, whereas even at the shaking flask stage the majority of conventional The protocol strains still had a survival rate of less than 80%. In addition to excellent cell survival, cells generated from the FACS protocol (gray box) are equivalents derived from those conventional protocols for IgG production (black box) as shown in Figure 8 (C). It was better.</p><p num="0134"> In addition, clones derived from single cell selection of hyperfluorescent cells (by FACS) were compared to clones derived from (conventional methods for their response to feed in batch-produced shaking flask methods). A single bolus feed of 5-10% yeastolate was added on day 7 of culture). Table 5 shows that 96% of FACS clones produced higher titers on feed than 25% of conventional clones. In addition, FACS clones had more average% increase, 45% vs. 19%, and more maximum titer increase of 105% vs. 34% in feed batch titers. These data suggest that FACS cloning is a better way to generate more potent and feed-reactive clones for biologic production.<tables num="5"><img id="000006" he="73" wi="159" file="JP6039552B2_D0001.tif" img-format="tif" img-content="drawing" /></tables></p>

<p num="0135"> Improved fluorescence observed when cells were analyzed according to their FSC-A vs. FSC-H profile Separate subpopulations of viable cell events with higher FSC-A in the FSC-H vs. FSC-A plot were observed with a flow cytometer. This cell subpopulation was named HFA subpopulation due to its high forward scattering region, which is an indicator of larger cell size. This population is within the same FSC-H boundary as the major cell population. In fact, the FSC-H of these HFA cells was 1.1 (+/- 0.1) times that of other cells within the FSC-H boundary (n = 32). This "other cell" subpopulation was named LFA (low forward scatter, an indicator of smaller cell size). Despite the similarities of FSC-H, the mean FSC-A of the HFA subpopulation is 1.8 (+/- 0.1) times that of the LFA subpopulation (n = 32).</p><p num="0136"> However, in the FSC-H vs. FSC-A plot, there is overlap between the HFA and LFA subpopulations on the FSC-A axis. For n = 22, 71% (+/- 25) HFA cells have higher FSC-A than overlapping regions with respect to a minimum of 38% and a maximum of 100%. Nevertheless, HFA subpopulations of cells can be clearly identified by those skilled in the art of flow cytometry, and the gating of each population can be determined accordingly. As a representative example, the diagonal lines in Figure 9B divide the live cell subpopulation into two distinct subpopulations, where the HFA cell subpopulation is the major in the FSC-H (Y-axis) vs. FSC-A (X-axis) plot. It is to the right of the diagonal line drawn to track the bottom edge of the population. When the same data are plotted as a histogram (Figure 10), the FSC-W value of the HFA cell subpopulation is higher than that of the LFA cell subpopulation. Therefore, HFA cells are included in the FSC-H / FSC-A plot (Figure 9) or FSC-W histogram (Figure 10A), or FSC-W / FSC-A (when gated in the FSC-A vs. SSC-A plot). Defined as a subpopulation of living cells with higher FSC-W that can be gated in any of the plots (Figure 10B). It is also possible to directly gate HFA cells in the FSC-A vs. SSC plot (FIG. 22), as shown in Example 14.</p><p num="0137"> Upon microscopic examination after sorting (gated on the FSC-H / FSC-A plot of living cells), this subpopulation of cells consists primarily of larger single cells, not cell aggregates / populations. It should be noted that high FSC-W cells (or "events") are customarily excluded from analytical selection if they form part of a gating strategy to eliminate cell doublet. Should be. This can be done, for example, in the FSC-W or FSC-H vs. FSC-A plot. However, recently, T cells have been shown in which such subpopulations rather contain highly active / proliferating cells (Bohmer et al., Cytometry part A, 2011, 00A: 1-7).</p><p num="0138"> Sorting (gating on FSC-H / FSC-A plots of live cells) and subsequent culture of this HFA subpopulation produces potent cell proliferation and antibody titers that exceed those comparable to those LFA subpopulations. did. The HFA population constitutes about 2-20%, more typically 5-10% of the total population of living cells. We hypothesize that this population represents a dynamic portion, perhaps an actively proliferating portion, of a living cell population that varies in number depending on the time of sampling.</p><p num="0139"> In addition, after sorting, HFA cells return to the general population FSC-A vs. FSC-H profile. However, the HFA subpopulation consistently produced higher levels of fluorescence than the LFA subpopulation (shown in the histogram; Figure 11). Higher fluorescence of this population was observed regardless of the staining method used, protein A or anti-IgG. Importantly for bulk transfection, cloning from the HFA subpopulation alone allows clear isolation of positively stained protein A cells (Fig. 11).</p><p num="0140"> Based on the above, the number of established strains and bulk transfections containing different antibody schemes was analyzed for HFA% and IgG or protein A staining of HFA and LFA cells, the results of which are shown in FIG. They show that the HFA subpopulation forms a higher proportion of higher fluorescent cells when stained with either fluorescently labeled protein A or anti-human IgG. They also show that the average fluorescence of cells in the HFA subpopulation is significantly higher than that in the LFA subpopulation.</p>

<p num="0141"> Selection of HFA cells results in improved selection of high-producing organisms and the generation of cell lines with equal or high stability.</p><p num="0142"> Two different selections presented below were performed to assess the value of using the FSC-A vs. FSC-H properties to identify and isolate by FACS selection of high-producing organisms: 1. Established strain, planned 10-strain MTX8 was a single cell cloned using three different approaches (Fig. 13A): (a) limit dilution (9 x 96 well plate set), (b). ) FACS cloning of cells simply live-gated on the SSC vs. FSC plot (7X 96-well plates set), and (c) raw and then HFA-gated on the FSC-H vs. FSC-A plot of HFA cells. FACS cloning (set 1X 96-well plate). As shown in FIG. 13B, the highest producing clone was that of the HFA clone, and it proved to be stable even after 50 passages.</p><p num="0143">2. Established strains, planned 3M-FACS sorting was performed, and HFA cell subpopulations were sorted at the single cell level. The strains before selection were unstable, and the batch titer decreased with increasing passages, and the decrease began to become noticeable after 20 passages (Fig. 14A). The data obtained show that HFA-derived clone strains have much higher batch titers and stability compared to their parents, with representative data for one clone shown in FIG. 14B.</p>

<p num="0144"> Selection of HFA, or HFA high cells results in improved selection of high producers.</p><p num="0145"> Four different selections presented below were performed to assess value using protein A (and IgG) staining along with FSC-A vs. FSC-H characteristics for identification of high producers: i. Established Strain, Plan 3M, IgG Sorting-Single Cell Sorting, Fluorescent Anti-Human IgG in "Live" Cells, HFA Cells, Live Cells with High Fluorescence (Top 1.5% Fluorescence) or HFA Cells with High Fluorescence It was carried out after staining with. The results (shown in FIGS. 15A and 15B) show the advantages of backscattering properties (ie HFA sorting) and sorting based on high fluorescence to isolate the highly productive. Figure 15A shows the individual clone titers and the number of clones in the 6-well stage after sorting. This data shows that HFA cells with high fluorescence have much higher titers than the other three selection criteria at the 6-well stage, and this tendency is that these clones have the highest number of SPRs (specific production). ) It is shown that it continues during antibody secretion in the shaking flask step of cloning (Fig. 15B). It should be noted that the top few clones from all four selections were further improved and showed higher batch titers and stability compared to the parent strain before selection.</p><p num="0146">ii. Bulk transfection, Plan 10, IgG and Protein A Sorting-Bulk transfection, selection, amplification and staining were performed as described in previous examples. Bulks were sorted after staining with either polyclonal anti-human IgG or protein A. The results in FIG. 16 (A) show the number of clones produced and the average level of IgG secreted at the 96-well plate stage. By sub-selection of HFA cells with high fluorescence (after either anti-IgG or protein A staining) compared to their bioselection equivalents, more clones and consistently higher levels of IgG Produced production. This tendency was retained when cells were examined in the later stages of cell line production (Figure 16 (B) shows the secreted titers for the best clones for each condition at different stages). Also, the superior live cell-selected clones (clone 49, FIG. 16B) produced similar IgG titers in the second shaking flask passage as compared to their superior HFA hyperfluorescence equivalents. When further cultured, the only clone that reached the growth curve stage was the "HFA, highly fluorescent" clone. FIG. 16 (C) shows data for cells modeled in unsupplied batch production, where titers were obtained on day 15 and from anti-IgG sera for selecting high protein A clones. Show potential excellence.</p><p num="0147">iii. A second bulk transfection, Plan 14, Protein A Sorting-Bulk Transfection, Selection, Amplification and Staining was performed as described in previous examples. Staining was performed with protein A. In addition to sorting the HFA and HFA hyperfluorescent cell populations, the LFA and LFA, hyperfluorescent populations were also cloned single cells. FIG. 17 shows associated FACS plots and sorting gates (left) as well as 96-well plate data (right) from single cell clones of this sorting. This data demonstrates that the HFA cell population produces higher expression clones than LFA cells, whether stained with protein A or unstained (similar data obtained by staining with anti-IgG). , Data not shown). However, when staining was used in conjunction with gating on HFA cells, a much higher percentage of clones had higher levels (compared to the titers of ranks 10, 20, and 30 in Figure 17). Had secreted IgG production. The results again show that protein A can be successfully used in place of polyclonal animal-derived labeled antibodies, that applying an HFA gate yields high titers and that an "HFA-high fluorescence" gate is applied. Demonstrate that this consistently yields more high titers. Subsequently, the clones are scaled up and reanalyzed at the 6-well and shaking flask production stage, where applying the HFA gate again yields high titers and applying the "HFA-high fluorescence" gate. Was found to yield many consistently high titers and a higher number of high-producing cell lines (Table 6 and Figure 20).<tables num="6"><img id="000007" he="64" wi="159" file="JP6039552B2_D0001.tif" img-format="tif" img-content="drawing" /></tables></p><p num="0148">iv. Clones were also generated for another scheme as described in (iii) above, where clones of HFA and HFA high protein A only were screened, analyzed and scaled up. Again, both, but more in shaking flask production than in 96 wells, there is a clear advantage to using HFA hyperfluorescent gates to identify high producing cell lines (Fig. 21).</p>

<p num="0149"> Improved predictability of productivity after implementation of the FACS protocol and the need to scale up a very small number of clones All cell line production protocols to date require hundreds of screenings and scale-ups in the absence of thousands of clones to produce cell lines suitable for the biopharmaceutical manufacturing regime. We sought to determine the number of clones needed to scale up from the earliest 96-well plate stage to ensure isolation of the highest production clones in the batch production model. To this end, we analyzed four different experiments presented below, ranked the top 4-5 clones (at the production curve stage), and then the previous stage in culture. I checked their ranking at. The data is presented below.</p><p num="0150"> 1) Plan 14, protein A-labeled cell selection (for top 2% stained [high fluorescence] cells) Bulk transfection, selection, amplification and staining were performed as described in previous examples. The bulk is a single cell sorted after staining with protein A and the results are shown in Tables 7 and 8 below.<tables num="7"><img id="000008" he="57" wi="159" file="JP6039552B2_D0001.tif" img-format="tif" img-content="drawing" /></tables><tables num="8"><img id="000009" he="61" wi="158" file="JP6039552B2_D0001.tif" img-format="tif" img-content="drawing" /></tables></p><p num="0151"> Table 7 shows the number of clones evaluated at the 96-well and shaking flask stage after protein A labeled cells transfected with the vector of Design 14 and sorted according to the four selection criteria shown. This data shows that the number of colonies obtained from HFA or HFA high sorting is significantly higher than that for LFA or LFA high sorting. Table 8 shows the ranking of the highest batch production titers to achieve cloning at different early stages (96 wells, 6 wells and T75). Surprisingly, these data demonstrate that 4 of the top 5 producing clones can be predicted from the top 5 clones even at the 96-well stage.</p><p num="0152"> 2) Plan 14, IgG-labeled cell selection (for the top 2% of IgG-stained cells) Bulk transfection, selection, amplification and staining were performed as described in previous examples. The bulk is single cell selection after staining with IgG and the results are shown in Tables 9 and 10 below.<tables num="9"><img id="000010" he="61" wi="159" file="JP6039552B2_D0001.tif" img-format="tif" img-content="drawing" /></tables><tables num="10"><img id="000011" he="61" wi="159" file="JP6039552B2_D0001.tif" img-format="tif" img-content="drawing" /></tables></p><p num="0153"> Table 9 shows the number of clones evaluated at the 96-well and shaking flask stage after cells transfected with the vector of Design 14 were labeled with anti-IgG and sorted according to the selection criteria shown. Cloning was performed on LFA hyperfluorescent cells using capture plates containing supporting cells (as described in the previous section), whereas for HFA hyperfluorescent cells two different conditions were used: support. Cells or acclimatized medium (CM). From this data, the number of clones produced from HFA cells using high fluorescence subselection was again higher. Four of the top five clones in the batch production model (Table 10) were also "HFA high" sorted cells. The data also demonstrate that 4 of the top 5 growth curve-producing clones can be predicted from the top 5 clones at the 96-well stage. For three of these clones, this ranking is maintained throughout the 6-well and shaking flask steps (pre-growth curve analysis).</p><p num="0154"> 3) Plan 10, cell selection labeled with IgG or protein A (for cells stained with the top 5% protein A) Bulk transfection, selection, amplification and staining were performed as described in previous examples. The bulk was sorted after staining with IgG and the results are shown in Tables 11 and 12 below.<tables num="11"><img id="000012" he="65" wi="159" file="JP6039552B2_D0001.tif" img-format="tif" img-content="drawing" /></tables><tables num="12"><img id="000013" he="67" wi="159" file="JP6039552B2_D0001.tif" img-format="tif" img-content="drawing" /></tables></p><p num="0155"> Table 11 shows 96 after labeling with anti-IgG or protein A transfected with the vector of Plan 10 and sorting them with either two selection criteria, "live gate" or "HFA cells with high fluorescence". The number of clones evaluated in the well and shaking flask steps is shown. From these data, the number of clones generated from the sub-selection of "high-fluorescence HFA cells" was also higher than that for the live-selected cells. In addition, all the top 5 clones in the growth curve (Table 12) are derived from "HFA high" cells, the highest being those selected for HFA high-protein A. The data also demonstrate good predictions from the top 8 clones at the 96-well stage, with 4 of them reaching the top 5 on the growth curve.</p><p num="0156"> 4) Plan 3M HFA cells Select cells labeled with IgG or protein A (for cells stained with the top 5% protein A) Two subclones of the planned 3M strain BP0044 were recloned by FACS (a) cell line C1 (initially a single cell selected as the raw top 1.5% IgG stain of parent BP0044), and (b) cell line D16. (Single cells initially screened for top 1.5% IgG staining of HFA in parent BP0044). C1 and D16 were selected single cells that were reselected from the top 5% fluorescent and HFA cell populations after labeling the cells with either fluorescent protein A or anti-human IgG. The results are shown in Tables 13 and 14 below:<tables num="13"><img id="000014" he="61" wi="159" file="JP6039552B2_D0001.tif" img-format="tif" img-content="drawing" /></tables><tables num="14"><img id="000015" he="63" wi="159" file="JP6039552B2_D0001.tif" img-format="tif" img-content="drawing" /></tables></p><p num="0157"> Table 13 shows the number of clones evaluated at the 6-well and shaking flask stage. 96-well titer data was not obtained in this case and is not shown. From these data, 4 of the top 5 clones in the growth curve stage are derived from the top 10 clones in the 6-well stage (Table 14).</p><p num="0158"> Taken together, the findings of these four examples above demonstrate that the novel protocol according to the invention identifies the highest production clones as early as the 96-well stage. This is believed to be the first protocol for CHO-based biopharmaceutical cell line production that describes the identification of final cell lines from within the top 5-10 of early 96-well screenings.</p><p num="0159"> In addition, HFA cells with high fluorescence consistently produce a large number of highest production clones in the batch production model, in total and compared to other selection criteria.</p><p num="0160"> Further analysis of the conventional CLD protocol previously used (eg, WO 2009024567 and Kotsopoulou et al. J Biotechnol, (As described in 2010) was used to compare FACS-based protocols as described herein. To this end, we analyzed seven different antibody schemes, where bulk transfection / amplification and cloning by the FACS (HFA-High Protein A) protocol were used. We also analyzed seven plans using traditional seeding / amplification protocols. In both cases, the top 1-2 clones ranked in the shaking flask production model were tracked and their ranking was recorded at the 96-well and 6-well stages of their selection. In conventional protocols these rankings were only tracked during the final amplification stage. The results shown in Table 15 show that there is a higher predictability of cloning performance when the FACS protocol (and HFA-high protein A gating) is performed, and only a few clones have high production suitable for production. Indicates that it is necessary to scale up to select a cell line. We also consider, in many cases, conventional protocols, this is not the first ranking, but one or more amplification / selection rounds that precede the last round recorded below. , These results are even more impressive.<tables num="15"><img id="000016" he="61" wi="158" file="JP6039552B2_D0001.tif" img-format="tif" img-content="drawing" /></tables></p>

<p num="0161"> Improvements in batch titer through reselection of clones using anti-IgG and protein A Staining achieved with methotrexate amplified bulk transfection was relatively low for protein A, but the fluorescent signal increased after single cell selection. This facilitates subsequent selection / recloning of highly expressed cells. In addition, the increase in fluorescent signal generated by protein A staining is increased with each separation step of single cell cloning (data not shown). Reselection adds time to the protocol for cell line production, but it has the advantage of producing clones, which not only have high titers, but also exhibit greater stability than their "parents" (typical). Example data is shown in Figure 14 and Example 9). FIG. 18 shows soluble IgG from a growth curve study of multiple selected cells (selected after performing a raw or HFA cell gate with or without anti-IgG or protein A staining as shown in the legend of FIG. 18). Shows the gradual increase observed at the level.</p><p num="0162"> Further analysis of the increase in IgG titer for planned 3M showed that it corresponded to the increase in mean SPR value obtained over 3-5 separate passages. See Table 16 below, except for the initial reselection where the increase in titer was due to an increase in cell number.<tables num="16"><img id="000017" he="72" wi="159" file="JP6039552B2_D0001.tif" img-format="tif" img-content="drawing" /></tables></p><p num="0163"> The data described in this example demonstrate the benefits of multiple recloning rounds to increase cell line productivity (and stability, if necessary). They also support the possibility that more than 3 and more than 3 additional reclones could increase the production titer even more.</p>

<p num="0164"> Comparison of FACS and selection from CLONEPIX (semi-solid medium) for selecting highly expressed cells ARH77 cells were bulk transfected using the Amaxa nucleofection kit with a vector encoding a membrane protein optionally named "Protein X" herein. Stable transformants were subsequently labeled with (A) PE-conjugated anti-protein X antibody and then single-cell FACS-selected (live-LFA cells) or (B) DyLight-649 conjugate. It was one of the clones selected from the semi-solid medium containing the anti-protein X detection antibody. With respect to (A), cells were stained with anti- "protein X" antibody and both the brightest and dimest positive cells were single cells sorted in 96-well format in 50% conditioned medium. For (B), 1e5 cells were seeded in Petri dishes in Genetix semi-solid medium containing supplements and also detection antibodies. The brightest clones were selected 7-10 days after sowing according to the manufacturer's instructions.</p><p num="0165"> The fluorescence of clones selected from Clonepix was compared to both low and high expression clones selected from FACS cloning. FIG. 19 shows the average fluorescence intensity (MFI) of all clones as measured by FACS analysis using the same anti- "protein X" antibody previously used for FACS sorting. The results demonstrate that the FACS method results in the identification / selection of very high expression clones. In fact, the clones from the highest expressed FACS had 6-fold higher MFI than the clones from the highest expressed Clonepix.</p>

<p num="0166"> Use of FSC-A gates and also improvements to sorting% gates Code a monoclonal antibody to investigate the correlation between protein A stained cells showing the highest levels of fluorescence intensity and the resulting clonal productivity, and thereby to determine optimal gating parameters for single cell selection. Sorting was performed using cells bulk-transfected with the plasmid to be used, and the cells were selected and amplified using the previously described protocol. Bulk-transfected cells were first gated for the live subpopulation and then for the HFA subpopulation (in the FSC-H vs. FSC-A plot) and then sorted according to% of protein A cell surface staining. The cells showing the top 10%, 5%, 2% and 1% fluorescence, respectively, were single cells cloned in 96-well plates (Fig. 22).</p><p num="0167"> In addition, as an alternative gating method, subgates were set within the live cell population (in the FSC-A vs. SSC-A plot) to incorporate all cells with high FSC-A characteristics. It contained low and high SSC-A cells. These high FSC-A cells are a subpopulation of the HFA subpopulation. These cells were subsequently evaluated for protein A cell surface staining and the cells showing fluorescence within the top 5% fluorescence range (when set for the live gate) were single cells sorted on a 96-well plate. (Fig. 22).</p><p num="0168"> Productivity analysis at the 96-well stage showed slight differences in mean and maximum titers recorded for cell clones sorted from either different% fluorescence intensity gates or high FSC-A and top 5% protein A stained gates. (Fig. 23). IgG production in supplemented batch shaking flask cultures consisting of the highest producing clones also showed slight differences in different selection criteria, but a trend was observed regarding the level of cell surface staining that correlated with production yield. This data also allows similar productivity clones to be isolated from the gating cell population based on high FSC-A from either the FSC-H vs. FSC-A plot or the FSC-A vs. SSC-A plot. Demonstrate that.</p>

<p num="0169"> Use of protein G We further sought to evaluate the possibility of using other reagents for staining / identifying / sorting high-producing cells. To this end, we evaluated protein G. Protein G Alexa488 was purchased from Invitrogen (catalog number P11065), and after titration measurement, the optimum staining was 10⁶It was found to be achieved with 0.25 to 0.5 μg of protein G per cell. Good labeling (and comparable to protein A) was achieved with protein G, as seen in FIG. The data generated show cloned cell lines (higher producers) that show high levels of fluorescence with labeled protein G (and protein A), as opposed to the low levels observed in the low-producing bulk population. Maintains soluble antibody levels secreted by cells. Higher protein G fluorescence was observed in the HFA gate population, similar to protein A in previous examples. In summary, like protein A, protein G should be suitable for use as a labeling reagent and for sorting high IgG-expressing CHO cells from bulk or pre-cloned strains.</p>