Additional Supporting Information may be found in the online version of this article.

Transcription

1 Cellular and Metabolic Engineering Biotechnology and Bioengineering DOI /bit Glutamine synthetase gene knockout-human embryonic kidney 293E cells for stable production of monoclonal antibodies Running title: Glutamine synthetase gene knockout HEK 293E cells Da Young Yu, Sang Yoon Lee, and Gyun Min Lee* Department of Biological Sciences, KAIST, 291 Daehak-ro, Yuseong-gu, Daejeon 34141, Republic of Korea; telephone: ; fax: ; *Correspondence to: Gyun Min Lee This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: [ /bit.26552] Additional Supporting Information may be found in the online version of this article. Received October 15, 2017; Revision Received December 17, 2017; Accepted January 15, 2018

2 Abstract Previously, it was inferred that a high glutamine synthetase (GS) activity in human embryonic kidney (HEK) 293E cells results in elevated resistance to methionine sulfoximine (MSX) and consequently hampers GS-mediated gene amplification and selection by MSX. To overcome this MSX resistance in HEK293E cells, a GS-knockout HEK293E cell line was generated using the CRISPR/Cas9 system to target the endogenous human GS gene. The GS-knockout in the HEK293E cell line (RK8) was confirmed by Western blot analysis of GS and by observation of glutamine-dependent growth. Unlike the wild type HEK293E cells, the RK8 cells were successfully used as host cells to generate a recombinant HEK293E cell line (rhek293e) producing a monoclonal antibody (mab). When the RK8 cells were transfected with the GS expression vector containing the mab gene, rhek293e cells producing the mab could be selected in the absence as well as in the presence of MSX. The gene copies and mrna expression levels of the mab in rhek293e cells were also quantified using qrt-pcr. Taken together, the GS-knockout HEK293E cell line can be used as host cells to generate stable rhek293e cells producing a mab through GS-mediated gene selection in the absence as well as in the presence of MSX. Keywords: HEK293E cell, glutamine synthetase knockout, CRISPR/Cas9, methionine sulfoximine, antibody

3 Mammalian cells are the dominant host cells for the production of therapeutic glycoproteins, mainly because of their ability to synthesize proteins that are similar to those naturally occurring in humans with respect to post-translational modifications and molecular structures (Zhu, 2012). Although Chinese hamster ovary (CHO) cells have been the most widely used mammalian host cells for the commercial production of therapeutic glycoproteins, non-human glycan structures at the terminal and the absence of γ-carboxylation are concerns (Ghaderi et al., 2012; Walsh, 2010). Therefore, human cell lines such as human embryonic kidney 293 (HEK293) cells have emerged as an alternative for the production of therapeutic glycoproteins most similar to those synthesized naturally in humans (Bandaranayake & Almo, 2014). In fact, several recombinant blood clotting proteins produced in HEK cells have been approved by the US Food and Drug Administration (Dumont et al., 2016). For high-level, stable production of therapeutic proteins in CHO cells, a glutamine synthetase (GS)-mediated gene amplification and selection system is increasingly used because the GS system typically requires only a single round of selection in the presence of a GS inhibitor, methionine sulfoximine (MSX) (Noh et al., 2013). With the availability of an efficient gene amplification and selection system like the GS system, HEK293 cells may be further used for the commercial production of a broader range of therapeutic proteins. Recently, the potential of the GS system in HEK293 cells for the stable production of therapeutic proteins was examined (Yu et al., 2016). Unlike CHOK1 cells, HEK293E cells producing therapeutic proteins were not selected at all even at a MSX concentration of 500 μm because the high GS activity of the HEK293E cells resulted in an elevated resistance against MSX and therefore, hampered GS-mediated gene amplification and selection by MSX. Therefore, to use GSmediated gene amplification and selection in HEK293 cells, the endogenous GS expression level in HEK293 cells should be minimized.

4 In this study, GS-knockout HEK293E cell lines were generated by disrupting the GS gene using the clustered regularly interspaced short palindromic repeats (CRISPR) technology (Cho et al., 2013). A resulting GS-knockout HEK293E cell line was characterized confirming the GS gene disruption and subsequently, evaluated the potential of the GS system in the recombinant HEK293E (rhek293e) cells for the stable production of monoclonal antibodies (mabs). The GS-mediated gene amplification and selection system is based on the GS enzyme which provides the critical pathway for the synthesis of glutamine (Liaw et al., 1995). Selection in the absence of glutamine has been successfully used in cell lines lacking GS activity (NSO and Ag14) as well as in cell lines with an endogenous GS gene (CHOK1) in the presence of MSX (Brown et al., 1992; Noh et al., 2013). Cell lines containing the transfected GS gene can develop resistance to MSX by the GS gene. The specific gene of interest, which is co-linked in the same expression vector, is co-amplified with the GS gene resulting in an increased specific productivity (Bebbington et al., 1992). However, the GS-mediated gene amplification and selection system in the presence of MSX cannot be used in HEK293E cells because of a high endogenous GS expression level in the HEK293E cells (Yu et al., 2016). To generate GS-knockout HEK293E cells, HEK293E cells were co-transfected with the Cas9 gene, hygromycin reporter gene, and single guide RNA (sgrna) targeting GS gene exon 4 (Fig. 1A) expression vectors (RGENs) using the Lipofectamine 2000 reagent, followed by hygromycin selection to enrich the RGEN-transfected cells. Targeted modifications in the transfected cells and hygromycin-selected cells were confirmed by T7 endonuclease 1 (T7E1) assay. The frequency of genetic mutation in the transfected cells increased from 26% to 57% by hygromycin selection (Fig. 1B). From the hygromycin-selected cells, 159 single cell clones were obtained in 24-well plates by limiting dilution method into 96-well plates in the absence

5 of hygromycin. The potential GS knockout clones were identified by normal growth in medium with glutamine and significantly reduced or no growth in medium without glutamine. Thirtyseven of the 159 clones isolated from the hygromycin-selected cells were identified as potential GS knockout clones (data not shown). Among the 37 clones, 30 fast growing clones were subjected to Western blot analysis of GS. Twenty-two of the 30 clones showed no detectable GS protein expression level (Supplementary Fig. 1). Eight of the 22 clones could grow in suspension cultures without severe cell aggregation, which is preferred over adherent cultures for large-scale commercial production of therapeutic glycoproteins. To confirm the GS enzyme function, the 8 clones, together with HEK293E cells as a control, were cultivated in 6-well plates. Unlike the 293E cells, these 8 clones had no detectable GS protein expression level by Western blot analysis (Fig. 1C) and could not grow in medium lacking glutamine (Fig. 1D). To identify the specific genetic modification in the 8 potential GS knockout clones, the human GS exon 4 gene region was amplified by PCR from these clones, and the sequencing results are summarized in Supplementary Table 1. The existence of multiple sequences for the different clones reflected the natural heterogeneous ploidy of the HEK293 cells (Lin et al., 2014). Among the 8 clones, 5 clones showed triploidy reflecting the cytogenetic characteristics of the HEK293 cells which are near triploid (Bylund et al., 2004). Although all 8 clones had deletions, insertions, and substitutions, only three clones (RK1, RK7 and RK8) showed no wild-type sequence and had frame shift mutations on every allele. Therefore, RK8 showing triploidy was selected for further characterization. To construct rhek293e cells producing Rituximab, the GS knockout RK8 cells were transfected with a GS expression vector containing mab genes (Yu et al., 2016). For comparison, HEK293E cells were also transfected with the same expression vector. Selection

6 was performed by seeding cells/ml in 96-well plates containing selection media (200 μl) with various MSX concentrations. The term cell pool indicates the cells which survived in the 96-wells after selection, and selection efficiency indicates a percentage of wells with a cell pool out of the total number of wells that were seeded with transfected cells. As observed previously (Yu et al., 2016), for the HEK293E cells, the number of wells with a cell pool did not decrease with an increase in the MSX concentration up to 50 μm (Fig. 2A). Regardless of the MSX concentration used, the mab concentration was almost zero confirming that the rhek293e cells cannot be established using the GS system in the HEK293E cells (Fig. 2B). In contrast, for the RK8 cells, the number of wells with a cell pool decreased significantly with an increase of the MSX concentration. In the absence of MSX, all the wells contained a cell pool, whereas 8 ± 2 % and 2 ± 0.6 % of the wells contained a cell pool at 12.5 and 25 μm MSX, respectively (Fig. 2A). When the MSX concentration was further increased to 50 μm, the transfected RK8 cells could not survive. Furthermore, mabs were produced even in the absence of MSX although the highest mab production was achieved at a MSX concentration of 12.5 μm (Fig. 2B). The average mab concentrations in the top three wells with a cell pool at 0, 12.5 and 25 μm MSX were 2.4, 6.9 and 1.5 μg/ml, respectively. Thus, relatively highproducing cell pools by MSX selection could be generated by removing the endogenous GS gene in HEK293E cells. To characterize the rhek293e cells, cell pools in randomly selected wells were expanded in selection medium with a corresponding concentration of MSX. Then, exponentially growing cells from each pool were seeded at a concentration of cells/ml into 24-well plates containing 1 ml of selection media with a corresponding concentration of MSX. To determine the specific growth rate (μ) and specific mab productivity (qmab) of the cell pools, two wells in 24-well plates were sacrificed every other day. Generally, the μ in the presence of MSX was

7 significantly lower than that in the absence of MSX (P < 0.05) (Fig. 3A). Many cell pools selected in the presence of MSX did not exhibit an elevated qmab, compared with those selected in the absence of MSX (Fig. 3B). However, when the cell pools with the highest qmab at each MSX concentration were compared, the qmab of cell pool #14 at 12.5 μm MSX and cell pool #10 at 25 μm MSX was 3.0 and 2.1 times higher than that of cell pool #7 in the absence of MSX, respectively. This enhanced qmab at 12.5 and 25 μm MSX was not due to the presence of MSX. The presence of MSX up to 25 μm affected neither the μ nor qmab of the cell pools (Supplementary Fig. 2). For further characterization of the rhek293e cells, cell pools on day 4 in the cultures shown in Fig. 3 were subjected to qrt-pcr analyses of the gene copy number and mrna expression levels of the mab heavy chain (HC) and light chain (LC) and the selectable marker (GS). The primer sequences for the qrt-pcr are listed in Supplementary Table S2. Like the qmab, there was no correlation between the MSX concentration and gene copy number or mrna expression level of the HC and LC (Fig. 4 and Supplementary Fig. 3). Cell pool #14 at 12.5 μm MSX that showed the highest qmab among all the cell pools at various MSX concentrations had high levels of relative gene copies and mrnas of the HC and LC (Fig. 4A, 4B, 4D and 4E). In contrast, all non-producing cell pools at 12.5 and 25 μm MSX had GS gene copies and expressed GS (Fig. 4C and 4F). However, many of these non-producing cell pools lost the mrna expression of the HC or LC, suggesting molecular defects in the transfected plasmids during the integration process into the cell chromosome. In conclusion, the GS-knockout HEK293E cell line can be successfully used as host cells for generating stable rhek293e cells producing mab through GS-mediated gene selection.

8 MATERIALS AND METHODS Generation of the GS-knockout HEK293E cells by CRISPR/Cas9 technology and culture maintenance CRISPR/Cas9-mediated GS gene knockout was performed in HEK293E cells (ATCC number: CRL-10852). A total of cells were cotransfected with the Cas9 gene, hygromycin reporter gene, and single guide RNA (sgrna) targeting GS gene exon 4 (AAATTCCACTCAGGCAACTCTGG) expression vectors (RGENs, ToolGen, Seoul, Korea) at a weight ratio of 10:1:10 using the Lipofectamine 2000 reagent (Life technology, Carlsbad, CA) according to the manufacturer s instructions. Two days after transfection, hygromycin selection was performed in T-25 flasks (Nunc, Rosklide, Denmark) by culturing the transfected cells in the presence of 2 mg/ml of hygromycin B (Clontech, Palo Alto, CA) for two days. Single cell clones from the hygromycin-selected cells were generated by limiting dilution into 96-well plates (Nunc) at a density of 0.5 cells/well in the absence of hygromycin. The GSknockout clones with a triallelic mutation on the endogenous GS gene were then selected by glutamine dependency screening, Western blot analysis of the GS, and DNA sequencing of the GS gene. HEK293E and GS-knockout clones were maintained in T-25 flasks in a 5% CO2/air mixture, humidified at 37. The medium for culture maintenance was Dulbecco s modified Eagle s medium (DMEM, Gibco, Grand Island, NY) supplemented with 7% fetal bovine serum (FBS, Biotechnics Research, Lake forest, CA) and 1 GS expression medium supplement (GSEM, Sigma-Aldrich, St. Louis, MO).

9 T7E1 assay for the targeting efficiency of CRISPR and sequence confirmation of the GS gene modification Genomic DNA was extracted from the RGENs-transfected cells using the Exgene TM Kit (GeneAll Biotechnology, Seoul, Korea). The genomic region containing the RGEN target site was amplified by PCR using PrimeSTAR HS DNA polymerase (Takara Bio, Shiga, Japan). The primers used for this study were 5 - CAATGGTCTTACTTGGAACTCAAA 3 (forward) and 5 - GTTCAGGGAAAAGAATCACTCAGA -3 (reverse) for the human GS exon 4 gene. Approximately 400 ng of the PCR products (510 bp) purified using the MEGAquick-spin total fragment DNA purification kit (intron Biotechnology, Seoul, Korea) were subjected to a re-annealing process to enable heteroduplex formation (95 for 10 min; ramped down to 85 at -1.5 /s; ramped down to 25 at -0.1 /s; and held at 4 ). Subsequently, the annealed PCR products were digested with T7E1 (New England Biolabs) at 37 for 90 min and analyzed on a 2% agarose gel. The percentage of the indel frequency was calculated using the ImageJ software (NIH, ver. 1.46) as described previously (Ramakrishna et al., 2014). PCR products prepared from potential GS gene knockout clones were cloned into the T&A cloning vector (Real Biotech Corporation, Banqiao City, Taiwan). Transformants were picked and plasmid DNA was purified using Hybrid-Q Plasmid Rapidprep (GeneAll) for DNA sequencing. Glutamine dependency screening To identify successful GS gene knockout clones, two rounds of screening with media with or without (+/-) glutamine were performed in 24-well plates and 6-well plates (Nunc). For each round, a duplicate set of cultures for each individual clone was used to test cell growth in media +/- glutamine. Exponentially growing cells were seeded at a concentration of cells/ml in 24-well plates or 6-well plates (Nunc) containing 1 or 3 ml of DMEM, +/- glutamine,

10 supplemented with 7% dialyzed FBS (dfbs, Gibco) and 1 GSEM, respectively. Cultures were grown in a 5% CO2/air mixture humidified at 37. To determine the cell concentration, one well in the plates was sacrificed daily. The cell concentration was estimated using a CountessII FL automated cell counter (Invitrogen, Waltham, MA), and viable cells were distinguished from dead cells using the trypan blue dye exclusion method. Acknowledgments This research was supported in part by a grant from the Bio & Medical Technology Development Program of the NRF funded by the Korean government (2013M3A9B ). DYY and GML are the co-inventors of patent US

11 References Bandaranayake, A. D., Almo, S. C., (2014). Recent advances in mammalian protein production. FEBS Letters, 588(2), Bebbington, C. R., Renner, G., Thomson, S., King, D., Abrams, D., Yarranton, G. T., (1992). High-level expression of a recombinant antibody from myeloma cells using a glutamine synthetase gene as an amplifiable selectable marker. Nature Biotechnology, 10(2), Brown, M. E., Renner, G., Field, R. P., Hassell, T., (1992). Process development for the production of recombinant antibodies using the glutamine synthetase (GS) system. Cytotechnology 9(1-3), Bylund, L., Kytola, S., Lui, W. O., Larsson, C., Weber, G., (2004). Analysis of the cytogenetic stability of the human embryonal kidney cell line 293 by cytogenetic and STR profiling approaches. Cytogenetic and Genome Research, 106(1), Cho, S. W., Kim, S., Kim, J. M., Kim, J. S., (2013). Targeted genome engineering in human cells with the Cas9 RNA-guided endonuclease. Nature Biotechnoly, 31(3), Dumont, J., Euwart, D., Mei, B., Estes, S., Kshirsagar, R., (2016). Human cell lines for biopharmaceutical manufacturing: history, status, and future perspectives. Critical Reviews in Biotechnology, 36(6), Ghaderi, D., Zhang, M., Hurtado-Ziola, N., Varki, A., (2012). Production platforms for biotherapeutic glycoproteins. Occurrence, impact, and challenges of non-human sialylation. Biotechnology and Genetic Engineering Reviews, 28, Jeon, M. K., Yu, D. Y., Lee, G. M., (2011). Combinatorial engineering of ldh-a and bcl-2 for reducing lactate production and improving cell growth in dihydrofolate reductasedeficient Chinese hamster ovary cells. Applied Microbiology and Biotechnology, 92(4), Liaw, S. H., Kuo, I., Eisenberg, D., (1995). Discovery of the ammonium substrate site on glutamine synthetase, A third cation binding site. Protein Science, 4(11), Lin, Y. C., Boone, M., Meuris, L., Lemmens, I., Van Roy, N., Soete, A., Reumers, J., Moisse, M., Plaisance, S., Drmanac, R., Tavernier, J., & Callewaert, N. (2014). Genome dynamics of the human embryonic kidney 293 lineage in response to cell biology manipulations. Nature Communications, 5:4767, doi: /ncomms5767. Noh, S. M., Sathyamurthy, M., Lee, G. M., (2013). Development of recombinant Chinese hamster ovary cell lines for therapeutic protein production. Current Opinion in Chemical Engineering, 2(4), Ramakrishna, S., Cho, S. W., Kim, S., Song, M., Gopalappa, R., Kim, J. S., Kim, H., (2014). Surrogate reporter-based enrichment of cells containing RNA-guided Cas9 nucleaseinduced mutations. Nature Communications, 5:3378, doi: /ncomms4378. Walsh, G., (2010). Post-translational modifications of protein biopharmaceuticals. Drug Discovery Today, 15(17-18), Yu, D. Y., Noh, S. M., Lee, G. M., (2016). Limitations to the development of recombinant human embryonic kidney 293E cells using glutamine synthetase-mediated gene amplification: Methionine sulfoximine resistance. Journal of Biotechnology, 231, Zhu, J., (2012). Mammalian cell protein expression for biopharmaceutical production. Biotechnology Advances, 30(5),

12 Figures legends Figure 1. Generation of the GS knockout HEK293E cells. A: Structure of the human GS gene: The human GS gene has 8 exons. The sgrna (AAATTCCACTCAGGCAACTCTGG) target site is in exon 4. B: T7E1 assay to detect the target sequence mutation in RGEN-transfected HEK293E cells and subsequently hygromycin-selected cells. The predicted position of DNA bands cleaved by T7E1 are indicated by asterisks. C: Western blot analysis of the total cellular proteins in putative GS knockout clones. An equal amount of cellular proteins was separated on 4-12% Bis-Tris NuPAGE gel. β-actin was used as a loading control. D: Profiles of (a) cell growth, (b) cell viability during glutamine dependency screening in putative GS knockout clones. Closed symbol, glutamine-containing medium; open symbol, glutamine-free medium. HEK293E cells that were not transfected by RGEN were used as a control. The error bars represent the standard deviations calculated from two independent experiments. Figure 2. Selection efficiency (A) and mab concentration in the wells with a cell pool (B) at various MSX concentrations. The selection medium for both HEK293E and RK8 was modified DMEM without glutamine (M-DMEM, Gibco) which was supplemented with 7% dfbs and 1 GSEM. Upon exceeding 30 % confluence, the culture supernatants in the wells with a cell pool were harvested to quantify the mab concentration with an enzyme-linked immunosorbent assay (Jeon et al., 2011). The error bars in Fig. 2A represent the standard deviation calculated from data obtained in triplicate experiments. The box plots in Fig. 2B represent the mab concentration in the wells with a cell pool obtained from triplicate experiments. A line within the box represents the median.

13 Figure 3. The μ (A) and qmab (B) of cell pools randomly selected at various MSX concentrations. The numbers of cell pools tested were 7 at 0 MSX, 14 at 12.5 μm MSX, and 10 at 25 μm MSX, respectively. The μ and qmab were determined based the data from day 2 to day 4. The asterisk represents that the qmab is almost zero. Figure 4. Relative gene copy number and mrna expression levels of the HC, LC, and GS in the cell pools shown in Fig. 3. The relative gene copy number of A: HC, B: LC, and C: GS. The relative mrna expression levels of D: HC, E: LC, and F: GS. The values of the gene copy number and mrna expression level of the cell pools were normalized to those of cell pool #1 in the absence of MSX. The total RNAs from the cells on day 4 were isolated using Hybrid-R (GeneAll), and the cdnas were prepared using the High-Capacity cdna Reverse Transcription Kit (Applied Biosystems, Cheshire, UK) according to the manufacturer s instructions. The primers for the qrt-pcr were designed using PrimerQuest (Integrated DNA Technologies, Coralville, IA). For each gdna and cdna sample, qrt-pcr was carried out in triplicate using iqtm SYBR gene Supermix (Biorad, Hercules, CA) in a Bio-Rad CFX96 machine according to the manufacturer s instructions. Each PCR reaction included a reaction mixture without a template to check for possible reagent contamination. The Ct values of the HC, LC and GS were normalized to those of the GAPDH and VCP level and the relative levels of gdna and mrna were calculated using the 2-ΔCt method.

14

15

16

17

18 本文献由 学霸图书馆 - 文献云下载 收集自网络, 仅供学习交流使用 学霸图书馆 ( 是一个 整合众多图书馆数据库资源, 提供一站式文献检索和下载服务 的 24 小时在线不限 IP 图书馆 图书馆致力于便利 促进学习与科研, 提供最强文献下载服务 图书馆导航 : 图书馆首页文献云下载图书馆入口外文数据库大全疑难文献辅助工具

19 本文献由 学霸图书馆 - 文献云下载 收集自网络, 仅供学习交流使用 学霸图书馆 ( 是一个 整合众多图书馆数据库资源, 提供一站式文献检索和下载服务 的 24 小时在线不限 IP 图书馆 图书馆致力于便利 促进学习与科研, 提供最强文献下载服务 图书馆导航 : 图书馆首页文献云下载图书馆入口外文数据库大全疑难文献辅助工具