Understanding factors that limit the productivity of suspension-based perfusion cultures operated at high medium renewal rates.

One of the key parameters in perfusion culture is the rate of medium replacement (D). Intensifying D results in enhanced provision of nutrients, which can lead to an increase in the viable cell density (X(v)). The daily MAb production of hybridoma cells can thus be increased proportionally without modifying the bioreactor scale, provided that both viable cell yield per perfusion rate (Y(Xv/D)) and specific MAb productivity (q(MAb)) remain constant at higher D. To identify factors prone to limit productivity in perfusion, a detailed kinetic analysis was carried out on a series of cultures operated within a D range of 0.48/4.34 vvd (volumes of medium/reactor volume/day) in two different suspension-based systems. In the Celligen/vortex-flow filter system, significant reductions in Y(Xv/D) and q(MAb) resulting from the use of gas sparging were observed at D > 1.57 vvd (X(v) > 15 x 10(6) cells/mL). Through glucose supplementation, we have shown that the decrease in Y(Xv/D) encountered in presence of sparging was not resulting from increased cellular destruction or reduced cell growth, but rather from glucose limitation. Thus, increases in hydrodynamic shear stress imparted to the culture via intensification of gas sparging resulted in a gradual increase in specific glucose consumption (q(glc)) and lactate production rates (q(lac)), while no variations were observed in glutamine-consumption rates. As a result, while glutamine was the sole limiting-nutrient under non-sparging conditions, both glutamine and glucose became limiting under sparging conditions. Although a reduction in q(MAb) was observed at high-sparging rates, inhibition of MAb synthesis did not result from direct impact of bubbles, but was rather associated with elevated lactate levels (25-30 mM), resulting from shear stress-induced increases in q(lac), q(glc), and Y(lac/glc). Deleterious effects of sparging on Y(Xv/D) and q(MAb) encountered in the Celligen/vortex-flow filter system were eliminated in the sparging-free low-shear environment of the Chemap-HRI/ultrasonic filter system, allowing for the maintenance of up to 37 x 10(6) viable cells/mL. A strategy aimed at reducing requirements for sparging in large-scale perfusion cultures by way of a reduction in the oxygen demand using cellular engineering is discussed.

Apoptosis-resistant E1B-19K-expressing NS/0 myeloma cells exhibit increased viability and chimeric antibody productivity under perfusion culture conditions.

We have shown previously that recombinant NS/0 myelomas expressing sufficient amounts of E1B-19K were resistant to apoptosis occurring in the late phase of batch culture and under stressful conditions such as cultivation in glutamine-free medium or following heat shock. However, no significant increase in monoclonal antibodies (MAb) was observed during the prolonged stationary phase of these batch cultures. Here, we show that E1B-19K can enhance cell survival and improve MAb productivity in high cell density perfusion culture. Typically, lymphoid cells grown under steady state in perfusion exhibit decreasing viabilities with concomitant accumulation of apoptotic cells. By modulating the ability of these cells to resist to induction of apoptosis in low nutrient environment, a 3-fold decrease in specific death rate from 0.22 day-1 for NS/0 control to 0.07 day-1 for E1B-19K cells was achieved, resulting in a significant improvement in cell viability throughout perfusion. E1B-19K cells at the perfusion plateau phase also exhibited a 3-fold reduction in specific growth rate concomitant with a lower percentage of S and higher percentage of G1 phase cells. This was associated with a 40% decrease in specific oxygen consumption rate, likely related to a reduction in the specific consumption rates of limiting nutrient(s). Expression of E1B-19K consequently had a significant impact on the steady-state viable cell density, allowing maintenance of 11.5 x 10(6) E1B-19K cells/mL versus 5.9 x 10(6) control NS/0 cells/mL for the same amount of fresh medium brought into the system (half a volume per day). Whereas MAb concentrations found in perfusion culture of control NS/0 myelomas were almost 3-fold higher than those found in batch culture; in the case of E1B-19K-expressing myelomas, the MAb concentration in perfusion was more than 7-fold higher than in batch. This was attributable to the 2-fold increase in viable cell plateau and to a 40% increase in the perfusion to batch ratio of specific MAb productivity (2.2-fold for E1B-19K myelomas versus 1.6-fold for NS/0 control).

Dose-dependent reduction of apoptosis in nutrient-limited cultures of NS/0 myeloma cells transfected with the E1B-19K adenoviral gene.

It is now well documented that apoptosis represents the prevalent mode of death in lymphoid cultures and occurs spontaneously in late-exponential phase of batch cultures following nutrient exhaustion. In an attempt to enhance the cell survival of these cell lines, we have initially engineered nonproducing NS/0 myeloma cells with a vector expressing the adenoviral E1B-19K protein. NS/0 cells transfected with E1B-19K were found to be more resistant to apoptosis occurring in the late phase of batch culture and under stressful conditions such as cultivation in glutamine-free medium or following heat shock. In this study, we have characterised a number of NS/0 subclones constitutively expressing different levels of E1B-19K, as well as several subclones in which the expression of E1B-19K was regulated by a tetracycline-controllable gene switch. We have found that a threshold E1B-19K level was required in order to achieve protection against apoptosis. The extent of resistance against cell death induced by nutrient deprivation in glutamine-free medium and in the late phase of batch cultures correlated with the level of E1B-19K expression up to an optimal level where further increases in E1B-19K levels did not result in significant additional protection. To assess the effects of E1B-19K on antibody productivity, an apoptosis-resistant NS/0 clone was then transfected with a chimeric antibody construct. Despite their improved viability, the antibody productivity of E1B-19K clones in batch culture was not significantly improved. Moreover, while the use of E1B-19K considerably delayed cell death, cells eventually died by apoptosis. Surprisingly, E1B-19K had no beneficial effect on the efficiency of fusion of NS/0 myelomas and splenocytes for the generation of hybridoma cells. Furthermore, the resulting hybridomas, although expressing E1B-19K at levels comparable to the myeloma parent, were no longer resistant to apoptosis. This indicates that the ability of E1B-19K to prevent apoptosis is not only dose-dependent but also seems to be cell-type dependent.

Apoptosis-resistant NS/0 E1B-19K myelomas exhibit increased viability and chimeric antibody productivity under cell cycle modulating conditions.

Lymphoid cells expressing sufficient levels of Bcl-2 or E1B-19K are known to resist to induction of apoptosis in glutamine-free or nutrient-limited batch cultures. However, despite the increased viability and prolonged stationary phase achieved in batch culture, product yields are not necessarily improved. Here we have found that expression of E1B-19K in NS/0 myeloma cells cultivated in the presence of certain cell cycle modulators could result in a significant increase in MAb productivity as compared to untransfected control cells. The use of E1B-19K significantly enhanced cell survival in the presence of osmolytes (sorbitol, NaCl), DNA synthesis inhibitors (hydroxyurea, excess thymidine), and the cell culture additive OptiMAbtrade mark. E1B-19K myelomas cultivated in the presence of NaCl or OptiMAbtrade mark accumulated in the G1 phase, while those arrested with excess thymidine were blocked in all phases. Interestingly, control NS/0 cells treated with these agents were found to die in a cell-cycle specific manner. Thus, while all G1 and most S phase cells quickly underwent apoptosis, G2/M cells remained alive and maintained MAb secretion for more than 10 days if supplied with adequate nutrients. For both control and E1B-19K cells, incubation with sorbitol or hydroxyurea was detrimental for MAb secretion, while addition of NaCl, excess thymidine and OptiMAbtrade mark resulted in an increased specific MAb productivity as compared to the batch culture. However, this increase resulted in an improvement of final MAb yields only in the case of OptiMAbtrade mark. The extension of viability conferred by E1B-19K allowed to further improve the final MAb yield obtained using OptiMAbtrade mark with a 3.3-fold increase for E1B-19K cells as compared to 1.8-fold for control NS/0 cells.

Induction of apoptosis in nutrient-deprived cultures of hybridoma and myeloma cells.

In the present study, cell death was investigated in cultures of NS/0 myelomas and SP2/0-derived D5 hybridomas through morphological examination of cells stained with acridine orange and ethidium bromide. The relative contribution of elevated levels of lactic acid and ammonia, as well as deprivation of glutamine, cystine, and glucose on the induction of necrosis or apoptosis, was investigated. In batch culture of D5 hybridoma cells, induction of apoptotic cell death correlated with the exhaustion of glutamine, while in the case of NS/0 myelomas, it coincided with exhaustion of cystine. To determine whether limiting nutrients were the actual triggering factors for apoptosis in batch culture, exponentially growing cells were resuspended in glutamine or cystine-free media. Within 30 to 40 h, viability decreased to 50% and the nonviable cell population displayed typical apoptotic morphology, with crescents of condensed chromatin around the periphery of the nucleus, or with the entire nucleus present as one or a group of featureless, brightly staining spherical beads. Similarly, D5 hybridomas and NS/0 myelomas cultivated in glucose-free medium died mainly from apoptosis. Cells were also cultivated in fresh medium supplemented with elevated concentrations of ammonia (3.0 mM) and/or lactate (35 mM, 50 mM). This resulted in decreased viabilities and necrotic death in both cell lines. From these results, we conclude that D5 hybridomas and NS/0 myelomas deprived of essential nutrients die by apoptosis, whereas incubation in the presence of elevated levels of metabolic byproducts such as ammonia and lactate will induce necrotic cell death in these cells.

Filtration-based perfusion of hybridoma cultures in protein-free medium: Reduction of membrane fouling by medium supplementation with DNase I.

In this study, a filtration-based perfusion process was developed for the production of monoclonal antibodies (IgM) by suspended hybridoma cells grown in protein-free medium. It was found that the use of protein-free medium for perfusion culture generated the formation of numerous visible suspended particles consisting of dead cells and cellular debris aggregated into fibrous material. Surprisingly high apparent viabilities were observed in such protein-free cultures. In addition, membrane fouling occurred more rapidly in protein-free medium than in conventional serum-supplemented medium. By the addition of deoxyribonuclease I (DNase I) to the protein-free medium, it was possible to prevent the formation of aggregates and to follow the evolution of the total cell population more accurately. Moreover, DNase I significantly reduced the fouling of filtration membranes, and that, for two different types of separation systems (cross-flow and vortex-flow filtration) and two different types of membranes (polycarbonate and hydrophilized polysultone). From these results, it is clear that the presence of DNA fragments liberated following cellular death is playing an important role in membrane fouling. Longevity of filtration membranes was found to be considerably greater using a vortex-flow filtration module than with a static plate-and-frame cross-flow filtration module. The use of vortex-flow filtration of conjuction with DNase I allowed maintenance of perfusion cultures for more than 1 month without membrane fouling or antibody retention and with a constant permeate IgM concentration of 250 mg/L. Hybridomacells appeared to gradually adapt to increasing rotational speed in the vortex-flow filtration module.

Induction of apoptosis in oxygen-deprived cultures of hybridoma cells.

It is now well documented that apoptosis represents the prevalent mode of cell death in hybridoma cultures. Apoptotic or programmed cell death occurs spontaneously in late exponential phase of batch cultures. Until lately, no specific triggering factors had been identified. Recently, we observed that glutamine, cystine or glucose deprivation induced apoptosis in both hybridoma and myeloma cell lines whereas accumulation of toxic metabolites induced necrotic cell death in these cells. Other triggering factors such as oxygen deprivation might also be responsible for induction of apoptosis. In the present study, induction of cell death by exposure to anoxia was examined in batch culture of the SP2/0-derived hybridoma D5 clone. The mode of cell death was studied by morphological examination of acridine orange-ethidium bromide stained cells in a 1.5 L bioreactor culture grown under anoxic conditions for 75 hours. Under such conditions, viable cell density levelled off rapidly and remained constant for 25 hours. After 45 hours of anoxia, cell viability had decreased to 30% and the dead cell population was found to be 90% apoptotic. In terms of cellular metabolism, anoxia resulted in an increase in the utilization rates of glucose and arginine, and in a decrease in the utilization rate of glutamine. The lactate production rate and the yield of lactate on glucose increased significantly while the MAb production rate decreased. These results demonstrate that glycolysis becomes the main source of energy under anoxic conditions. Cells incubated for 10 hours or less under anoxic conditions were able to recuperate almost immediately and displayed normal growth rates when reincubated in oxic conditions whereas cells incubated for 22 hours or more displayed reduced growth rates. Nonetheless, even after 22 h or 29 h of anoxia, cells reincubated in oxic conditions showed no further progression into apoptosis. Therefore, upon removal of the triggering signal, induction of apoptosis ceased.

Monitoring glutamine in mammalian cell cultures using an amperometric biosensor.

An amperometric biosensor has been developed for monitoring glutamine in the pulsed-batch cultivation of murine hybridoma cells. Glutamine oxidase was cross-linked with bovine serum albumin (BSA) via glutaraldehyde activation and deposited on a preactivated nylon membrane. Glutaminase was then immobilized on the protein layer and the resulting membrane was attached to the sensing area of a hydrogen peroxide probe (platinum vs silver/silver chloride polarized at +0.7 V). An orthogonal test was performed to optimize the activity of the membrane for glutamine with respect to the concentrations of glutamate oxidase, BSA, glutaminase and glutaraldehyde. There was an excellent linear relationship between the biosensor's response and glutamine in the range 0.1-3 mM. The determination of glutamine could be performed in 2 min and each membrane was reused for at least 300 consecutive analyses. The data obtained also agreed well with those high-performance liquid chromatography, thus validating the applicability of the biosensor.

Tissue-specific dopaminergic regulation of the glucocorticoid receptor in the rat pituitary.

The rat intermediate pituitary lobe is one of the rare tissues that is not a known glucocorticoid target and is devoid of immunoreactive glucocorticoid receptor. The intermediate lobe is poorly vascularized and receives a dopaminergic and serotonergic innervation from the hypothalamus. In previous studies we demonstrated that removal of this hypothalamic input results in the appearance of immunoreactive glucocorticoid receptor in the intermediate lobe, as demonstrated with the use of in vitro intermediate pituitary cultures and two in vivo experimental situations. We now show that this appearance of the glucocorticoid receptor is presumably due to removal of hypothalamic dopamine from the intermediate lobe cells, since in this study dopamine (or its potent agonist bromocriptine) inhibits expression of the glucocorticoid receptor in intermediate pituitary cells in primary culture, as demonstrated by [3H] dexamethasone binding and immunocytochemistry. The dopamine antagonist haloperidol blocks the inhibitory effects of the dopamine agonist. In contrast to the intermediate pituitary cells, bromocriptine does not affect glucocorticoid receptor expression in anterior pituitary cells in culture. The differential dopaminergic regulation of glucocorticoid receptor expression in the pituitary gland raises questions about possible effects of dopamine on glucocorticoid receptor levels and glucocorticoid response in other dopamine target tissues, especially in the brain.