Clonal variability within dihydrofolate reductase-mediated gene amplified Chinese hamster ovary cells: stability in the absence of selective pressure

Recombinant Chinese hamster ovary (rCHO) cells expressing a high level of chimeric antibody were obtained by cotransfection of heavy- and light-chain cDNA expression vectors into dihydrofolate reductase-deficient CHO cells and subsequent gene amplification in medium containing stepwise increments in methotrexate (MTX) level up to 1.0 microM. To determine the clonal variability within the amplified cell population in regard to antibody production stability, 20 subclones were randomly isolated from the amplified cell population at 1.0 microM MTX (CS13-1.0 cells). Clonal analysis showed that CS13-1.0 cells were heterogeneous with regard to specific growth rate (mu) and specific antibody productivity (qAb), although they were derived from a single clone. The mu and qAb of 20 subclones were in the range of 0.51 to 0.72 day-1 and 10.9 to 19.1 microgram/10(6) cells/day, respectively. During 8 weeks of cultivation in the absence of selective pressure, the mu of most subclones did not change significantly. On the other hand, their qAb decreased significantly. Furthermore, the relative decrease in qAb varied among subclones, ranging from 30% to 80%. Southern and Northern blot analyses showed that this decreased qAb resulted mainly from the loss of amplified immunoglobulin (Ig) gene copies and their respective cytoplasmic mRNAs. For the sake of screening convenience, an attempted was made to correlate the initial properties of subclones (such as mu, qAb, and Ig gene copies) with their antibody production stability during long-term culture. Among these initial properties examined, only qAb of subclones could help to predict their stability to some extent. The subclones with high qAb were relatively stable with regard to antibody production during long-term culture in the absence of selective pressure (P < 0. 005, ANOVA). Taken together, the clonal heterogeneity in an amplified CHO cell population necessitates clonal analysis for screening stable clones with high qAb.     

Cytogenetic analysis of chimeric antibody-producing CHO cells in the course of dihydrofolate reductase-mediated gene amplification and their stability in the absence of selective pressure.

Previously, the highest producing (HP) recombinant CHO subclones isolated at various methotrexate (MTX) levels showed different antibody production stability during long-term culture, although they were clonally derived from CS13 transformant. In this study, genetic basis for their difference in antibody production stability was investigated using southern blot hybridization and fluorescence in situ hybridization (FISH) techniques. Southern analysis of HP subclones revealed that light-chain (LC) and heavy-chain (HC) cDNAs were located closely within 23 kb on an amplification unit, and the configuration of LC and HC cDNAs within this amplification unit was not disrupted during long-term culture in the absence of MTX. However, when LC and HC genes were localized on the metaphase chromosomes of HP subclones using FISH, the amplified sequences were present as an extended array on diverse marker chromosomes. HP subclones selected at higher MTX level had more kinds of marker chromosomes. CS13*-002 isolated at 0.02 microM MTX had only one marker chromosome (m002), whereas CS13*-1.0 isolated at 1 microM MTX had five different ones (m10A, m10B, m10C, m10D, and m10E). Each marker chromosome showed different fate during long-term culture of HP subclones in the absence of MTX, resulting in different degrees of stability among the HP subclones. The m10A and m10B remained unchanged, whereas the others disappeared or evolved to variants with shortened amplified arrays. The cells containing stable marker chromosomes constituted dominant subpopulations in CS13*-1.0, and thereby CS13*-1.0 became most stable in regard to antibody production during long-term culture. Furthermore, our dual-color FISH showed that the telomeric ends of amplified arrays on the stable marker chromosomes were always surrounded by (TTAGGG)(n) sequences, indicating that (TTAGGG)(n) sequences are closely related to the stability and evolution of amplified sequences. Taken together, our data show that the assessment of genotypic stability of amplified CHO cells is a prerequisite for understanding their production stability during long-term culture in the absence of selection pressure.     

Enhancing effect of low culture temperature on specific antibody productivity of recombinant Chinese hamster ovary cells: clonal variation

To understand the different responses of recombinant Chinese hamster ovary (rCHO) cells to low culture temperature regarding specific productivity (q), 12 parental clones and their corresponding amplified clones producing a humanized antibody were cultivated at 32 and 37 degrees C. The specific growth rate of all clones, including both parental and amplified clones, decreased by 30-63% at 32 degrees C, compared to rates at 37 degrees C. In contrast, their specific antibody productivity (qAb) was significantly enhanced at 32 degrees C. Furthermore, the degree of qAb enhancement at 32 degrees C varied a lot from 4- to 25-fold among the parental clones. At 32 degrees C, most of the amplified clones, regardless of methotrexate (MTX) levels, also showed enhanced qAb but to a lesser extent than their parental clones. However, clone 14 amplified at 0.32 microM MTX (clone 14-0.32) and clone 20 amplified at 1 microM MTX (clone 20-1.00), unlike their parental clones, did not show enhanced qAb at 32 degrees C. Thus, it was found that the enhancing effect of low culture temperature on q of rCHO cells depends on clones. Taken together, the results obtained here emphasize the importance of clonal selection for the successful application of low culture temperature to the enhanced foreign protein production in rCHO cells.     

Key determinants in the occurrence of clonal variation in humanized antibody expression of cho cells during dihydrofolate reductase mediated gene amplification

Recombinant Chinese hamster ovary (CHO) parental clones expressing a humanized antibody against S surface antigen of hepatitis B virus were obtained by cotransfection of heavy chain (HC) and light chain (LC) cDNA expression vectors into dihydrofolate reductase (DHFR)-deficient CHO cells. When 23 representative parental clones were subjected to stepwise selection for increasing methotrexate (MTX) resistance, such as 0.02, 0.08, 0.32, and 1.0 microM, their clonal variations in regard to antibody expression were found to be significant. Among 23 parental clones, only one clone (hu17) showed the significant increment of specific antibody productivity (q(Ab)) with increasing MTX concentration up to 0.32 microM. Compared with the parental clone (hu17), the q(Ab) of hu17 resistant at 0.32 microM MTX (hu17-0.32) was enhanced approximately 12.5-fold. To clarify the reason for the occurrence of clonal variations, Southern blot analyses of chromosomal DNAs derived from each amplified clone at 0.32 microM MTX were performed. Only the hu17-0.32 clone did not experience severe genetic rearrangement during gene amplification, and it had only one 49-kb amplification unit including the LC and HC cDNAs. A fluorescent MTX competition assay showed that the resistance against MTX toxicity of the other clones without enhanced q(Ab) at 0.32 microM MTX was obtained by mechanisms such as an impaired MTX transport system. Taken together, the data obtained here show that clonal variations in regard to antibody expression are found to be significant because clones can acquire MTX resistance by mechanisms other than DHFR-mediated gene amplification despite the stepwise selection.     

Limitations to the development of humanized antibody producing Chinese hamster ovary cells using glutamine synthetase-mediated gene amplification

Recombinant Chinese hamster ovary (CHO) cells expressing a humanized antibody were obtained by transfection of an antibody expression vector (pKC-GS-HC-huS) into CHO-K1 cells and subsequent glutamine synthetase (GS)-mediated gene amplification in media containing different concentrations of methionine sulfoximine (MSX). Concentrations consisted of 25, 200, 500, and 1000 microM of MSX. The highest producer (HP) subclones were isolated from each MSX level by the limiting dilution method and were characterized with respect to antibody production. No positive relationship was observed between specific antibody productivity (q(Ab)) and MSX concentration. Furthermore, it was found that the antibody production stability of these subclones was very poor even in the presence of selection pressure. During long-term cultures in the presence of the corresponding concentrations of MSX, q(Ab) of all HP subclones significantly decreased for the first six passages and thereafter stabilized. Southern and slot blot analyses showed that the loss of antibody gene copies was only partially responsible for the decreased q(Ab). Fluorescence in situ hybridization (FISH) analysis revealed some cytogenetic features indicative of antibody production instability. Unstable chromosomal structures including dicentrics, rings, and extremely long chromosomes were observed. Amplified sequences enclosed in nuclear projections were often observed. The telomeric repeat sequence, which may be involved in the stabilization of amplified arrays, was found to be absent at the ends of most marker chromosomes. Furthermore, FISH analysis revealed that the overall chromosome content was duplicated in some HP subclones. When metaphase of 12 high producing parental clones was examined, the frequency of occurrence of the polyploidy was 25%. Taken together, the data obtained here suggests that instability could be a concern in the development of CHO cells with GS-mediated gene amplification.     

用体外免疫的人淋巴细胞建立抗破伤风类毒素的...

Human tonsil cells in vitro immunized with tetanus toxoid were fused with human-mouse heteromyeloma line RF to generate human-mouse hybridomas. Hybridoma 891112-50 was cloned and 2 subclones (891112-50-3 and -4) were obtained. The secreted antibodies from the subclones were antigen specific, since they did not cross react with three irrelevant antigens (OVA, TCS and F gamma G). The hybridomas were quite stable. After 13 passages in tissue culture flasks, they still retained their antibody secreting ability. From flow cytometry analysis the subclone 50-3 was more stable than the subclone 50-4. The human immunoglobulin contained in supernatant collected during regular passages was equivalent to 69.6 micrograms/ml.     

Decreased chimeric antibody productivity of KR12H-1 transfectoma during long-term culture results from decreased antibody gene copy number

The stability of KR12H-1 transfectoma in regard to chimeric antibody production was examined during long-term, repeated batch culture without selection pressure using antibiotics. Both serum-supplemented and serum-free media were used. Regardless of the medium used, the specific antibody productivity (q(Ab)) of transfectoma decreased by 60% to 88% during 70-day culture. This loss of antibody productivity was not due mainly to the appearance of a nonproducing population (NP) of transfectoma. The percentage of a producing population (P), which was monitored by the limiting dilution method, remained over 90% until the end of culture, indicating that the q(Ab) of P decreased during the culture. Flow cytometric data also showed the increase of cell population with low fluorescence intensity during culture, indicating that the intracellular antibody content of P decreased. The subclones of P obtained at the end of long-term culture were further characterized. Compared with the q(Ab) of P at the beginning of long-term culture, the q(Ab) of most P subclones was significantly low, confirming that the loss of antibody productivity was due mainly to the decreased q(Ab) of P during long-term culture. The decreased antibody gene copy number of P subclones was found to be partly responsible for the decreased q(Ab) of P during long-term culture. (c) 1996 John Wiley & Sons, Inc.     

Comprehensive characterization of dihydrofolate reductase-mediated gene amplification for the establishment of recombinant human embryonic kidney 293 cells producing monoclonal antibodies

Human embryonic kidney 293 (HEK293) cells with glycosylation machinery have emerged as an alternative host cell line for stable expression of therapeutic glycoproteins. To characterize dihydrofolate reductase/methotrexate (DHFR/MTX)-mediated gene amplification in HEK293 cells, an expression vector containing dhfr and monoclonal antibody (mAb) gene was transfected into dhfr-deficient HEK293 cells generated by knocking out dhfr and dhfrl1 in HEK293E cells. Due to the improved selection stringency, mAb-producing parental cell pools could be generated in the absence of MTX. When subjected to stepwise selection for increasing MTX concentrations such as 1, 10, and 100 nM, there was an increase in the specific mAb productivity (q_mAb) of the parental cell pool upon DHFR/MTX-mediated gene amplification. High producing (HP) clones with a q_mAb of more than 2-fold of the corresponding cell pool could be obtained using the limiting dilution method. The q_mAb of most HP clones obtained from cell pools at elevated MTX concentrations significantly decreased during long-term culture (3 months) in the absence of selection pressure. However, some HP clones could maintain high q_mAb during long-term culture. Taken together, a stable HP recombinant HEK293 cell line can be established using DHFR/MTX-mediated gene amplification together with dhfr⁻ HEK293 host cells.     

Clonal variation in response to adrenocorticotropin in cultured bovine adrenocortical cells: relationship to senescence

During cellular senescence, non-clonal cultures of bovine adrenocortical cells show a continuous decline in the rate of production of cyclic AMP (cAMP) stimulated by adrenocorticotropin (ACTH), without changes in the rate of forskolin- or prostaglandin E1-stimulated cAMP production. We investigated the possible mechanisms for loss of response to ACTH by examining the properties of clones of bovine adrenocortical cells. ACTH-stimulated cAMP production rates were measured in clones immediately after isolation, during long-term growth following isolation, and after subcloning. ACTH-stimulated rates were compared with cAMP production in response to forskolin, which acts directly on the catalytic subunit of adenylate cyclase. The results show that cloning is not necessarily associated with a loss of response to ACTH, but that clones with high ACTH response can give rise to subclones with low response. Clones of adrenocortical cells, at the same approximate population doubling level (PDL), showed ACTH response levels that ranged from 12 to 135 pmol cAMP/10(6) cells/min, whereas mass cultures at this PDL showed approximately 50 pmol/10(6) cells/min. Forskolin-stimulated cAMP production rates in clones varied only over the range of 59-119 pmol/10(6) cells/min and showed no correlation with the ACTH-stimulated rates. All clones were adrenocortical cells, as shown by mitogenic response to angiotensin II and cAMP-inducible 17 alpha-hydroxylase activity. The replicative potential of clones varied widely, and there was no apparent correlation between ACTH response levels and growth potential. The level of ACTH response in each clone was stable during proliferation through at least 25 PD beyond the stage at which the clone was isolated. When clones were subcloned, a clone with a high ACTH response level produced sister subclones that had ACTH response levels ranging from 3% of that of the parent clone to a level slightly greater than that of the parent clone. The growth potential of sister subclones varied widely, as for the parent clones, and there was no obvious correlation between growth potential and ACTH response. Two subclones were cloned; in sub-subclones, levels of ACTH response were again different from each other and also from the parent subclone; in one case, the level of ACTH response was approximately eight-fold higher than that of the parent subclone. These experiments show that clonal variation in the extent of expression of a differentiated property may occur in a normal differentiated cell in culture. The loss of ACTH response and the loss of proliferative potential appear to be independent stochastic events.     

Efficient selection of high-producing subclones during gene amplification of recombinant Chinese hamster ovary cells by flow cytometry and cell sorting

The screening procedure for high-producing cell lines is extremely time- and labor-intensive and costly, and is at present guided by an empirical approach based on individual experience. Flow cytometry and cell sorting, with its ability to analyze and separate single cells, an ideal method in the selection of such rare cells. The isolation of recombinant cell lines is especially difficult due to repeated gene amplification, which introduces high mutational variation into the population. We have established and evaluated a modification of a previous method that traps secreted product on the surface of the secreting cell, thus allowing direct analysis of single cell specific production rates. This method was used to select for high-producing subclones of a recombinant Chinese hamster ovary (CHO) cell line producing a human antibody against HIV-1 by repeated rounds of gene amplification and cell sorting. This cell line has been amplified in previous investigations, so that the amount of work and testing required by traditional methods can be compared with the protocol described herein. Forty-five 96-well plates were necessary to obtain a high-producing subclone by limited dilution methods, whereas only five plates were required when cell sorting was used. The specific production rate of the best clone obtained by sorting, however, was five times that of the clone obtained by traditional methods. In contrast to the clones obtained by limited dilution, which consisted of several populations of low- and high-producing cells even at high methotrexate concentrations (6.4 microM), the clones isolated by sorting were already homogeneous at 0.8 microM methotrexate.     

Effect of copper sulfate on performance of a serum-free CHO cell culture process and the level of free thiol in the recombinant antibody expressed

A recombinant Chinese hamster ovary (CHO) cell line was used to express a humanized antibody. Product quality analysis of this humanized antibody showed the presence of free thiol, due to unpaired cysteine residues in the Fab region. Decreased potency of this thiol Fab made it critical to minimize the levels of free thiol. In an effort to do this, we evaluated the effect of copper sulfate addition to the cell culture production medium. As a component of the production medium, copper sulfate can act as an oxidizing agent, thereby facilitating disulfide bond formation. Four concentrations of copper sulfate were added at the beginning of 2-L benchtop production cultures of the recombinant CHO cell line: 0, 5, 50, and 100 microM. We found that these copper sulfate additions had no effect on cell growth or antibody production. However, a slight dose-dependent depression in culture viability was observed. Analysis of the purified antibody showed that either the 50 or 100 microM copper sulfate additions reduced the level of free thiol by more than 10-fold.     

Comparison of two aposymbiotic ciliate clones of Climacostomum virens by their ability for reinfection

The ability of two aposymbiotic (algae-free) subclones of the same green clone of C. virens to establish a stable symbiotic association with Chlorella sp. has been studied by light and electron microscopy. Alga-free subclone No. 1 was obtained from the original green clone by a long-term cultivation in darkness, while subclone No. 2 originated from one cell that spontaneously lost the algae and was found among normal green cells during daily inspection. For infection, algae isolated from ciliates with chlorellae of parental clone of C. virens were used. 5-10 minutes after feeding with Chlorella, specimens of both subclones show numerous algae mostly inside food vacuoles, but some rare algae (3-4 per cell) may occur in individual perialgal vacuoles. Later on, the number of symbiotic chlorellae in ciliates of subclone No. 1 increased, and a stable symbiotic association was reestablished. Unlike, in specimens of subclone No. 2 all newly ingested algae were seen digested within food vacuoles. Within 24-28 h all the ciliates investigated appeared free of algae. However, obviously stable symbiotic ciliate-algae systems in this subclone were obtained after improving the microinjection technique. Injection of algae into alga-free ciliates resulted in maintenance of intact chlorellae in these ciliates. The algae were seen to be located individually within perialgal vacuoles, being presumably protected against host lytic enzyme attack. The endosymbiont population in ciliates was established from as many as 3-5 originally injected algae. The number of symbiotic chlorellae increased steadily reaching the value equal to that in the parental clone 28-30 days after the start of experiment.     

Expression of functional alpha beta T cell receptor gene rearrangements in suppressor T cell hybridomas correlates with antigen binding, but not with suppressor cell function

We have examined the expression of TCR genes in 4-hydroxy-3-nitrophenyl-acetyl (NP)-specific Ts cell hybridomas. Each of three independently isolated hybridomas expressed in-frame TCR alpha-chain rearrangements derived from the original suppressor Ts cell. Different V alpha and J alpha gene segments were rearranged and expressed in each Ts cell line. The only TCR beta-chain expressed in these cells was derived from the BW5147 fusion partner. Expression of the BW5147 beta-chain was found to correlate with cell surface Ag binding, inasmuch as subclones derived from one of the original Ts lines expressed greatly reduced levels of beta-chain mRNA and no longer bound to NP-coupled RBC. Subclones that continued to express beta-chain mRNA did bind to NP-coupled RBC. This suggests that the Ag receptor on Ts hybridomas is a TCR-alpha beta dimer composed of a unique alpha-chain and the BW5147 beta-chain. Ag binding could be modulated by preincubation of Ts hybridoma cells with anti-TCR-alpha beta antibody, thereby supporting this conclusion. Suppressor factor activity was measured in the conditioned media of Ts subclones that differed by 250-fold in levels of beta-chain mRNA expression. No difference in suppressor factor activity was found; conditioned media from these subclones suppressed both plaque-forming cell responses and delayed-type hypersensitivity responses at approximately equivalent dilutions. Suppressor factor activity in the conditioned media of both a beta-chain negative subclone and a beta-chain positive subclone could be absorbed with an antibody that recognizes the TCR alpha-chain, but not with an antibody that recognizes the TCR beta-chain. We conclude that suppressor factor activity in the conditioned media of these Ts hybridomas is not derived from surface TCR-alpha beta receptors, although it does share TCR alpha-chain determinants.     

Loss of antibody productivity is highly reproducible in multiple hybridoma subclones

An immunoglobulin G (IgG(2b)) producing hybridoma cell line (S3H5/gamma2bA2) was cloned and subcloned. Twenty subclones were grown in parallel while being adapted in a stepwise fashion to serum-free medium. Following adaptation to serum-free medium, it was found that 16 of the 20 subclones remained at a relatively constant proportion of nonproducing cells. Three of the remaining subclones transiently deviated from this balance but eventually returned toward this population composition. One subclone continued to lose productivity. A population balance was reached at approximately 8% of the population being nonproducers. The loss of antibody productivity was thus highly reproducible. (c) 1993 John Wiley & Sons, Inc.     

Characterization of prion susceptibility in Neuro2a mouse neuroblastoma cell subclones

In this study, we established Neuro2a (N2a) neuroblastoma subclones and characterized their susceptibility to prion infection. The N2a cells were treated with brain homogenates from mice infected with mouse prion strain Chandler. Of 31 N2a subclones, 19 were susceptible to prion as those cells became positive for abnormal isoform of prion protein (PrP(Sc)) for up to 9 serial passages, and the remaining 12 subclones were classified as unsusceptible. The susceptible N2a subclones expressed cellular prion protein (PrP(C)) at levels similar to the parental N2a cells. In contrast, there was a variation in PrP(C) expression in unsusceptible N2a subclones. For example, subclone N2a-1 expressed PrP(C) at the same level as the parental N2a cells and prion-susceptible subclones, whereas subclone N2a-24 expressed much lower levels of PrP mRNA and PrP(C) than the parental N2a cells. There was no difference in the binding of PrP(Sc) to prion-susceptible and unsusceptible N2a subclones regardless of their PrP(C) expression level, suggesting that the binding of PrP(Sc) to cells is not a major determinant for prion susceptibility. Stable expression of PrP(C) did not confer susceptibility to prion in unsusceptible subclones. Furthermore, the existence of prion-unsusceptible N2a subclones that expressed PrP(C) at levels similar to prion-susceptible subclones, indicated that a host factor(s) other than PrP(C) and/or specific cellular microenvironments are required for the propagation of prion in N2a cells. The prion-susceptible and -unsusceptible N2a subclones established in this study should be useful for identifying the host factor(s) involved in the prion propagation.     

Limitations to the amplification and stability of human tissue-type plasminogen activator expression by Chinese hamster ovary cells

Chinese hamster ovary cell production of recombinant tissue-type plasminogen activator (t-PA) was increased by amplification of cotransfected dihydrofolate reductase cDNA using stepwise adaptation to increasing methotrexate (MTX) concentrations. The highest producing clones were isolated at 5 microM MTX and yielded 26,000 U/10(6) cells/day t-PA (43 microgram/10(6) cells/day). Above 25 microM MTX, cell specific t-PA production rates became increasingly variable and the cDNA copynumbers decreased. No apparent correlation between the cell specific t-PA production rate and the growth rate was observed upon subcloning of the amplified cells. When MTX selection was removed, the t-PA production rate decreased up to tenfold within 40 days; this was accompanied by an up to 60% drop in cDNA copynumber. Subclones isolated after 108 days of culture in the absence of MTX were, on average, sixfold more stable than their parental cells. In culture without MTX, the maximum stable t-PA production rate obtained (over 250 days) was 7000 +/- 750 U/10(6) cells/day (approximately 12 microgram/10(6) cells/day), approximately threefold lower than the maximum unstable levels of production reached under selective pressure. Taken together, these results define a wide range of the highest t-PA expression rates obtained under MTX selection, for which stable expression without selection has not been reported.