Published Online: 15 DEC 2009
Copyright © 2009 John Wiley & Sons, Inc. All rights reserved.
Encyclopedia of Industrial Biotechnology
How to Cite
Al-Rubeai, M., Khoo, S. H. G. and Singh, R. P. 2009. Apoptosis. Encyclopedia of Industrial Biotechnology. 1–14.
- Published Online: 15 DEC 2009
- Top of page
- Apoptosis and Animal Cell Biotechnology
- Future Prospects
- Further Reading
In the past, it has generally been assumed that certain cell-culture conditions, such as interactions with specific hormones and deprivation of basic nutrients and growth factors, result in physical damage or metabolic collapse of the cell, leading to a passive, or “necrotic,” death. It was thought that the cell was at the complete mercy of its environment and had no control over its fate. However, it is now clear that the cell does not always respond in such a simple and passive manner. Instead, a range of factors can trigger a highly complex and genetically regulated cellular response during which specific “death proteins” are activated, a phenomenon that has been named apoptosis (1). It is these proteins that are responsible, ultimately, for the death and destruction of the cell. Indeed, in many cases, the cell may have sustained only low levels of damage. Clearly, under these conditions, apoptotic death will be premature and is often referred to as cellular suicide. Apoptosis is also termed type I programmed cell death and is often classified by the eventual activation of the caspase, a family of zymogenic proteases (2).
Apoptosis is now acknowledged as a fundamentally important process that plays an essential role in embryogenesis and in the maintenance and functionality of the highly ordered cell populations that constitute higher organisms. Indeed, there are now few aspects of biomedical research upon which apoptosis has not had an impact. It is as a consequence of its fundamental nature that any failure in its regulatory mechanisms leads to many of the diseases that pose the greatest challenges to medical science. This has resulted in an explosion of research into the genetic basis of apoptosis, the objective of which is to develop novel therapeutics for a wide range of disorders, including cancer, AIDS, and ischemic injury. Consequently, a growing number of proteins that are involved in the induction suppression and execution of the apoptotic program have been identified. It is now clear that many of the cell lines used in the biotechnology industry for the production of therapeutics undergo apoptotic death. Obviously, the advances made in the characterization of the apoptotic pathway may be applied to the suppression of the apoptotic response in industrial culture processes. This should provide the biotechnologist with a new route to the optimization of culture performance.
We begin by describing the biochemical basis of apoptosis, the morphological changes that accompany it, and some of the techniques that have been used to identify apoptotic cells. This is followed by a general discussion of the regulation and induction of apoptosis. We then describe studies that have investigated this phenomenon from a biotechnological perspective. Particular reference is made to the conditions that elicit an apoptotic response, the cell lines that are susceptible, and the cellular engineering approaches to enhance cell survival and productivity in the bioreactor environment.
- Top of page
- Apoptosis and Animal Cell Biotechnology
- Future Prospects
- Further Reading
2.1 Morphology of Apoptosis
Apoptosis is defined by its highly characteristic morphology, which is illustrated in Fig. 1. Early changes include a reduction in cell volume (3, 4) and loss of surface microvilli. Cytoskeletal changes result in the formation of protrusions on the surface of the cell, which are referred to as blebs. These may break away as intact vesicular structures, giving rise to “apoptotic bodies” (5).
One of the most striking changes during apoptosis occurs within the nucleus. Fluorescence microscopy of a typical viable cell following staining with a DNA stain such as acridine orange reveals a large spherical nucleus that may constitute most of the cellular volume. Often, the chromatin is highly diffuse, although occasionally it may be possible to see condensed chromosomes in mitotic cells. During apoptosis, the nucleus undergoes major changes. Initially, the chromatin condenses and marginates to the nuclear membrane, forming crescent or ring-shaped structures that exhibit intense fluorescence. Eventually, the chromatin collapses into two or more particles, which are often highly spherical. Again, the remainder of the cell will be devoid of chromatin and will be almost transparent in appearance. The shape of the cell also undergoes a highly characteristic change. Although viable cells tend to be highly irregular in shape, on entry into apoptosis they become smooth surfaced and in many cases almost spherical.
The changes in nuclear morphology coincide with the activation of a nuclease enzyme that cleaves chromatin first into 300- and/or 50-kbp fragments (6), and then ultimately into multiples of 200 bp (7). The first stage is believed to be responsible for the morphological changes described above. The second stage represents the cleavage of chromatin at the internucleosomal linker regions. This generates a striking ladder-like pattern when DNA samples from apoptotic cells are subjected to DNA gel electrophoresis (Fig. 2). Together with condensation of chromatin, this so-called DNA ladder has become one of the hallmarks of apoptotic death.
Many cell types express a further enzymatic activity—that of the transglutaminase enzyme (8). This enzymatic activity cross-links proteins within the cell, generating a protein scaffold that is believed to hold the dead cell together, thus explaining the relative robustness of dead apoptotic cells as compared with cells that have undergone necrotic death. It is also important to distinguish apoptosis from caspase-independent cell death (CICD), which is preceded/triggered by mitochondrial outer membrane permeabilization. Although CICD shares features with apoptosis such as mitochondrial outer membrane permeabilization, diffusion of some proteins from the intermembrane space of the mitochondria, and the ensuing DNA fragmentation, it does not involve the activation of the caspase cascaden; hence, it is considered a nonapoptotic form of death (9).
These highly controlled changes appear to have a specific task to provide a very clean and rapid method of eliminating dead cells, which is a vital consideration when one considers the extent of apoptotic cell death during, for example, embryogenesis. The cleavage of chromatin does not appear to be responsible, in itself, for the death of the cell. Instead, it has been argued that this ensures the complete destruction of the genetic material. It is suggested that this reduces the possibility of the DNA from the dying cells to lead to the malignant transformation of surrounding cells. The stabilization of the dead cell by the transglutaminase enzyme minimizes the probability of leakage of its contents onto surrounding cells. In vivo, surrounding cells and phagocytes engulf the dead cell before it can cause damage to surrounding tissue. Thus, inflammation of tissue as usually results from necrotic death is avoided. In vitro, in the absence of phagocytes, the apoptotic cell will eventually enter a degenerative phase called secondary necrosis.
2.2 Identification of Apoptosis
The search for simple techniques that allow for the identification of apoptotic cells has resulted in several different methods being available. There are a wide variety of dyes that are now used to detect caspase cleavage products and their activity. This section aims to provide an introduction to some basic methods used for such detection. As already stated, visualization of nuclease-mediated cleavage of DNA has been widely used to identify the presence of apoptotic cells. However, the technique can produce variable quality results that are of a qualitative, rather than a quantitative, nature.
Perhaps the simplest techniques for the study of apoptosis are based on the identification of the morphological features of cell death. For example, fluorescence microscopic analysis of nuclear morphology is a highly effective method for identification and quantification of apoptosis (Fig. 3). There are two ways in which this may be done. If samples cannot be analyzed immediately, cells may be fixed in formaldehyde and stored at 4°C. Analysis involves staining with acridine orange, which reveals the condensation of chromatin in apoptotic cells (8). However, this technique has one major drawback—an inexperienced operator may confuse early necrotic cells with viable cells, which have a very similar nuclear morphology. In order to avoid this difficulty, cells may be analyzed immediately while still in their culture medium by using the acridine orange–propidium iodide (PI) dual-staining technique (10). All cells are permeable to acridine orange, which stains chromatin green. Only membrane-damaged cells take up PI, and as a result exhibit red fluorescence. This technique, therefore, provides information regarding plasma membrane integrity, as well as nuclear morphology, and can consequently be used to simplify identification of necrotic cells. Additionally, the classification of apoptotic cells into early apoptotic and membrane-damaged apoptotic (sometimes referred to as secondary necrotic) cells gives an indication of the cell growth and death kinetics under the influence of a variety of environmental conditions during the cultivation process. The early-membrane-intact phase of death is relatively brief, and, therefore, the presence of a large proportion of cells at this stage indicates that the rate of cell death is very high.
Microscopic techniques have two major drawbacks—they are subjective and time consuming. The development of flow cytometric methods has overcome many of these difficulties and provided a powerful tool for the study of the biochemical features of apoptosis in heterogenous cell populations. Furthermore, it allows a means of rapidly assessing thousands of cells at a time. An extensive review on the use of flow cytometry for apoptosis detection can be found elsewhere (11). At present, there are a number of techniques that are available, and some of the most commonly used are discussed further.
2.2.1 Light-Scattering Properties
The simplest flow cytometric method is based on the changes in light-scattering properties that accompany cell death. Cells can be studied without the need for pretreatment, with the decrease in cell size producing a decrease in forward-scattered light. The increased granularity caused by nuclear condensation produces an increase in orthogonal light scatter (12, 13). However, the latter is a transient stage, and eventually, a reduction in orthogonal light scatter is observed. The technique can also be used for the identification of necrotic cells (at least in their early stages). When necrosis is induced by a permeabilizing agent such as saponin, a reduction in forward scatter is observed, but the increase in side scatter that occurs during apoptosis is not seen.
2.2.2 Changes in DNA Content
When stained with a DNA-specific stain such as PI, apoptotic cells exhibit a characteristically low DNA content that appears as a sub-G1 peak (i.e. it appears below the position of the G1 peak of the cell cycle of viable cells). This is believed to result from the leakage of cleaved DNA from apoptotic cells (14-17). The technique provides a rather good correlation with the fluorescence microscopic technique already described. However, necrotic cells undergoing degradation may also exhibit a reduced DNA content, although the passive and asynchronous nature of this process means that a clear “peak” is not usually observed.
Nucleosmal DNA fragmentation can also be detected by the terminal deoxyribonucleotide transferase-mediated dUTP nick end labeling (TUNEL). This detection method measures the caspase activity, which cleaves genomic DNA and exposes single stranded breaks (‘nicks’) that can be labeled with terminal transferase and subsequently visualized with either enzymatic substrates or fluorescent probes (18, 19). While examination by fluorescent microscopy provides a means of examining specific features of DNA fragmentation, the initial staining with PI and counterstaining with TUNEL allows the distinguishing between apoptotic and necrotic cells. Such a technique also provides a means of quantifying the percentage of apoptosis cells with sensitivities up to 90% (20) and can also be easily implemented in flow cytometry.
In viable cells, phosphatidylserine (PS) is located on the inner leaflet of the plasma membrane (21). During apoptosis, one of the earliest changes is the loss of this asymmetrical distribution (21, 22). Annexin V has a very high affinity for PS. Conjugation of annexin V to a fluorescent tag, such as Fluorescein isothiocyanate (FITC), enables the use of this interaction to identify cells that have lost PS asymmetry. By combining this method with PI staining, it is also possible to classify apoptotic cells into two subpopulations: early (membrane intact and therefore PI negative) and late (membrane damaged and therefore PI positive). The technique has been found to give a good correlation with levels of apoptosis during hybridoma batch cultures, during which apoptosis accounts for around 90% of cell deaths, as revealed by the fluorescence microscopic method described earlier. However, as with other flow cytometric techniques, necrotic cells can also give a false-positive result. This is because damage to the plasma membrane allows the annexin to enter the dead cell and bind to PS residues on the inner surface of the plasma membrane. Thus, the technique is reliable only when used to analyze early apoptotic cells, which can be seen as annexin V positive and PI negative (22, 23).
2.2.4 Mitochondrial Potential
Certain mitochondrial membrane transmembrane potential sensitive dyes can be used for the detection of apoptosis. During the onset of apoptosis, the inner mitochondrial membrane potential is reduced, which represents the detection of apoptosis at an early stage. Cationic lipophilic fluorochromes (e.g. DiOC6, Nonyl acridine orange (NAO), tetramethylrhodaminemethylesters, and JC-1) have been used with flow cytometric or fluorescent microscopy detection to assess the apoptotic state of cells (24). Caspase activity has been shown to be correlated to mitochondrial membrane potential, but membrane potential can also be affected by other factors not related to apoptosis. Dual staining with PI or 7AAD (7-amino-actinomycin D) distinguishes different stages of apoptosis. In the same way that Annexin V positive and PI negative cells expose early apoptotic cells, low membrane potential and PI negative cells also expose early apoptotic cells (25).
2.2.5 General Comments Regarding Identification of Apoptosis
Each of the techniques we described has advantages and drawbacks. When designing experiments to identify and quantify apoptosis, a number of points should be taken into consideration. First, there can be significant variations in the morphology of apoptosis from one cell type to the next, including, for example, absence of chromatin condensation. It is therefore recommended that several methods be used simultaneously to identify the mechanism of cell death. However, quantification of actual levels of apoptosis can significantly vary depending upon the technique used. Thus, in order to draw valid conclusions, it is essential that comparison of quantitative data collected using one technique is not made with data collected using a second technique.
2.3 Inducers of Apoptosis: Physiological Factors
The list of factors that induce apoptosis has grown steadily over the last few years. In the present section, factors that have been the center of purely biological studies will be considered, although where necessary, implications for animal cell technology will be highlighted. Factors that are of specific interest to the process biotechnologist are explored in section titled as Apoptosis and Animal Cell Biotechnology.
2.3.1 Presence/Absence of Receptor–Ligand Interactions
The absence of certain hormones can result in the induction of apoptosis in certain cell types. This has led to the suggestion that apoptosis may play a pivotal role in the maintenance of tissue organisation in vivo. It is thought that all cells are primed to undergo apoptosis and are prevented from doing so through constant stimulation by paracrine survival factors (26). If a cell is removed from its physiologically correct location, the absence of the appropriate survival signal will lead to the induction of apoptosis. This role of survival factors as regulators of cellular distribution in vivo may also have an impact on the development of serum-free media preparations for industrial-scale cell-culture processes, as described later.
The presence of receptor–ligand interactions can also lead to the induction of apoptosis. For instance, in the Fas–FasL system, binding of the Fas ligand to the Fas receptor can trigger apoptosis (27). This mechanism is responsible for the regulation of the immune system. Autoreactive B cells undergoing maturation (28) and autoreactive mature T cells (29) are eliminated by the Fas-mediated induction of apoptosis. Fas also acts as the “off” switch for the immune system by inducing apoptosis in antigen-activated B and T cells (30). Molecular dissection of the Fas–FasL system has provided important insights into the early stages of the signaling cascade that leads to the expression of the death pathway (See section titled Mechanisms of Apoptosis).
2.3.2 Induction by Viruses
A number of viruses have been shown to interact with the cellular apoptotic machinery. An apoptotic response to viral infection would appear to act as a protective mechanism that prevents viral replication by triggering the suicide of the infected cell. However, a number of virally encoded antiapoptotic genes that suppress the expression of the death program, thereby providing the virus with the opportunity to propagate itself, have now been identified. Perhaps one of the most interesting examples of this antideath mechanism from a biotechnology perspective is that of baculovirus, which has been synthesized at production scale by infection of insect cell lines, and has applications as a biological pesticide and, more recently, for the expression of recombinant proteins. A mutant was identified that induced high levels of apoptosis during infection of Sf21 insect cells. This was attributed to a mutation in the p35 viral gene, that, in its wild-type form, acts as an antiapoptosis gene (31).
Arguably, the most important example of virus-induced apoptosis is that mediated by HIV (32-35). A number of reports have indicated that the binding of the viral gp 120 protein to the CD4+ receptor of T cells triggers the induction of apoptosis, thus leading to the depletion of this class of cells during HIV infection (36-38). Interestingly, HIV infection of CD4+ cells appears to provide protection from apoptosis. The viral Nef protein downregulates the expression of CD4+, thus preventing the induction of apoptosis. As a result, virus propagation in the infected cell is not prevented (39).
There have also been reports of virus-induced suicide of bacterial cells. Until recently, it was assumed that altruistic cell death was not possible in single-celled organisms, simply because the genes involved in such a phenomenon would be lost when the cell concerned dies. However, bacteriophage infection in bacterial colonies has been reported to induce a suicide response, thus preventing the spread of the infection to the remainder of the colony (40). The genes that mediate such a response are propagated by other clones in the colony, thus preserving the altruistic nature of cellular suicide. Bacterial and viral expression systems have been used for the production of recombinant proteins. Clearly, the possibility that bacterial cells in such systems exhibit an apoptosis-like response needs to be investigated.
2.3.3 Free Radicals and Apoptosis
Free radical–mediated cellular damage has become an important area of study. In recent years, there have been numerous reports that cell death following oxidative stress occur by apoptosis. Moreover, generation of free radicals has been postulated as being a universal triggering event in the induction of apoptosis (41-46). Indeed, for a period in the early 1990s, it was suggested that the widely studied antiapoptosis gene bcl-2 functioned as an antioxidant (47). However, this theory has been challenged by the demonstration that anoxia-induced apoptosis can also be suppressed by Bcl-2 in the absence of measurable levels of free radicals (10, 48, 49). Furthermore, reactive oxygen species (ROS), which may result from highly active mitochondria, can result in the induction apoptosis (50).
2.3.4 Induction by Toxins and Therapeutic Agents
Exposure to high levels of toxic chemicals results in necrotic death of the cell. However, long-term exposure at a low level can trigger an apoptotic response. Many of the agents used in chemotherapy exert their effect by inducing apoptosis in tumor and normal cells. Overexpression of genes such as bcl-2 has been linked to resistance to chemotherapy. Clearly, establishing the mechanism of induction of apoptosis in response to such agents and providing strategies that minimize the effect of antiapoptotic genes should provide novel and more effective chemotherapeutic strategies (51).
Exposure of cells to ionizing radiation induces high levels of apoptosis in many normal tissues. Particularly susceptible are those cells from tissues that undergo rapid proliferation, such as spermatogonia (52) and lymphocytes (53). Such tissue would be especially prone to malignant transformation, and consequently the induction of apoptosis following DNA damage minimizes the likelihood of such an event. Irradiation of tumors can also lead to the induction of apoptosis (54-56). Notably, tumors that respond least favorably to irradiation exhibit the lowest level of apoptosis under such conditions (54). The induction of apoptosis following DNA damage is regulated by the product of the p53 gene, which has been referred to as the guardian of the genome because of its central role in preventing the propagation of cells that can sustain genetic damage and thus, potentially, malignant transformation (57).
2.3.5 Endoplasmic Reticulum Stress
More recently, it has been established that endoplasmic reticulum (ER) stress can result in cellular apoptosis. ER stress can result in the unfold protein response (UPR) when there is an imbalance between the load of newly synthesized proteins in the ER and the organelle's protein folding capacity. It can also result from ER lipid or glycolipid imbalances and redox or ionic changes of the ER lumen. p53-upregulated modulator of apoptosis (PUMA) is significantly upregulated after prolonged ER stress and this links the ER stress response to the mitochondrial apoptosis pathway independent of p53 expression (58). Although the cell initially attempts to reduce the level of proteins by the activation of proteasome-dependent ER-associated degradation or by modulating the translational polypeptide traffic into the ER, prolonged ER stress can give rise to the transcription-dependent UPR changes that result eventually in apoptosis. This is mediated by a variety of transmembrane proteins such as PERK and IRE1 as well as other proteins like procaspase 4 (59). Despite the determination of ER stress and its link to apoptosis, there still remains much to understand about this form of stress.
2.4 Mechanisms of Apoptosis
The study of apoptosis has been the subject of much research over the years. Although there are several mechanistic pathways leading to apoptosis, two major pathways are worth mentioning here. These can be described as the intrinsic and extrinsic pathways (2) (Fig. 4). The extrinsic signaling pathway involves the binding of extracellular ligands (e.g. FasL) to specific receptors, thereby causing a series of cytosolic adaptor proteins [e.g. Fas-associated death domain (FADD)] to be recruited and the activation of initiator caspases subsequently. These initiator caspases give rise to the activation of downstream effector caspases (e.g. caspase 3, 6, 7), which give rise to the characteristic morphological changes mentioned earlier. These extrinsic signals may be amplified by the induction of the mitochondrial outer membrane permeabilization (which leads to the formation of the apoptosome—a complex of cytochrome c/Apaf-1/procaspase-9 complex). The intrinsic pathway occurs via the mitochondria, which integrates signals from various stressors, including DNA damage, cytoskeletal damage, ER stress, nutrient, and growth factor deprivation. These stimuli invoke a similar mitochondrial outer membrane permeabilization as mentioned above.
These different pathways provide the genetic basis for apoptosis reduction in biopharmaceutical applications. The genes involved in apoptotic death may be classified into three groups. The first group consists of the modulators of the apoptotic pathway that suppress or induce death. The second group comprises the components of the cell death pathway that mediate the cell death signal. The final component is the group of effector caspases/enzymes that is responsible for the death and destruction of the cell. Some examples of these genes are given in Table 1.
One of the most common mutations identified in tumors is that of the p53 gene mentioned earlier. In response to DNA damage, p53 causes cell cycle arrest, allowing the cell to repair damaged DNA, thus ensuring that any potentially carcinogenic mutation is not propagated. However, in some cell types, p53 triggers the induction of apoptosis, again in order to minimize the risk of transformation (57, 72).
Of the modulators of the apoptotic pathway, one group of closely related proteins, the Bcl-2 family (e.g. proapoptotic Bax, Bak, Bcl-x and antiapoptotic Bcl-xl, Bcl-2, Bcl-W, E1B-19K) (73), has attracted particular attention. This family of genes can be divided into proapoptotic and antiapoptotic genes. The bcl-2 gene, the first and best characterized member, was identified at the t(14;18) breakpoint found in human follicular lymphoma (61). It encodes a 24-kDa protein that is located on the outer mitochondrial membrane, the cytosolic face of the nuclear membrane, and ER. Numerous studies have demonstrated the ability of this protein to suppress apoptosis in response to a wide variety of inducers. Table 1 lists other members of this family, some of which are functionally similar to Bcl-2, whereas others act as antagonists of Bcl-2 and thereby trigger apoptosis. Although caspase-8 is predominantly the cause of apoptosis in hepatocytes, bcl-2 was able to suppress apoptosis in such cultures when antioxidants or mitochondrial membrane permeability inhibitor could not (74). Studies of bcl-2 in this context have been instrumental in establishing the role of apoptosis in cancer. Mutations that result in the overexpression of bcl-2 lead to the accumulation of cells due to life span extension. These cells then undergo further mutation, most notably involving the c-myc gene, leading to the formation of high-grade tumors that combine the characteristics of high cellular survival with a high rate of proliferation (75-77). Clearly, these are characteristics that would be desirable in the ideal candidate host for the expression of recombinant proteins.
The well-characterized pattern of development of the nematode Caenorhabtis elegans, which includes the induction of apoptosis in specific cell types, has provided an important insight into the genetic basis of apoptosis. Most notable among the genes identified are ced 9 and ced 3. The former is a bcl-2 homolog that blocks the ability of ced 3 to induce cell death (78). The ced 3 gene is a cysteine protease that shares extensive homology with the mammalian protein called interleukin 1β converting enzyme (ICE) (79, 80). As the name suggests, this enzyme cleaves the active precursor pro-interleukin 1β to generate the active molecule interleukin 1β, and a number of studies have now implicated it in the induction of apoptosis. However, apoptosis can also be induced in macrophages and thymocytes of ICE-negative mice, indicating that ICE is not a universal mediator of apoptosis (81). Indeed, several ICE-related enzymes have been identified (Table 2).
|Interleukin 1β converting enzyme (ICE)||82|
|ICErel II/Transcript X (TX)||(85, 86)|
The molecular mechanism of apoptosis induced by the interaction of the Fas ligand with its receptor centers on the activity of ICE-related proteases. Indeed, important progress has now been made in deciphering the earliest events in the signal cascade that transduces the initial death stimulus to the death machinery of the cell. Activation of the Fas receptor by its ligand or agonist antibodies leads to the binding of the adapter protein MORT1/FADD (88, 89). This, in turn, binds to the ICE homolog FLICE (FADD-like ICE) or MACH (MORT1-associated ced-3 homolog) (90, 91).
p53 can also transactivate several BH3-only members (e.g. Puma, Noxa, Bid) (92-94), the disrupter of mitochondrial function p53AIP (95), the apoptosome component Apaf-1 (96) as well as other proapoptotic genes. There are also several inhibitors of apoptosis (IAPs) (97, 98), the most well known of which is XIAP (X chromosome-linked inhibitor of apoptosis). Other IAPs include Aven (99), Hsp70 (100), Hsp90 (101), APIP (102), TUCAN (103), HBXIP (104), and HSp27 (105). Another prominent molecule that significantly affects the development of apoptosis response in cells is survivin (106). This protein together with aurora-B kinase, the inner centromere protein (INCENP), and the telophase disk antigen TD-60 makes up the chromosomal passenger complex. It is an important regulator of chromosome alignment, histone modification, and cytokinesis (107), which can effectively inhibit apoptosis by binding and inactivating caspase 3 and 7.
In addition to the classic apoptosis inhibitors described above, most cells also express caspase-8 decoys called FLIPS (FLICE inhibitory proteins) (108). c-FLIP has a death effector domain (DED) that binds with a complex containing death protease caspase-8 (FLICE) and prevents the transduction of the apoptosis signal from death receptors (109). Another DED containing molecule, BAR (bifunctional apoptosis regulator), is able to compete with FADD for binding with procaspase 8 and 10, thereby preventing FAS-mediated apoptosis. Importantly, BAR protein can also inhibit the intrinsic mitochondrial pathway by interacting and enhancing the antiapoptotic activity of the Bcl-2 and Bcl-xL proteins (110, 111).
3 Apoptosis and Animal Cell Biotechnology
- Top of page
- Apoptosis and Animal Cell Biotechnology
- Future Prospects
- Further Reading
At present, there are 18 approved antibodies and more than 150 recombinant proteins in clinic (112). Production of biologically active therapeutic proteins and antibodies necessitates, in many cases, the expression of the protein in mammalian cell lines. As a result of intense commercial pressure, tried and tested production processes are often adopted in order to ensure that the product reaches the market in the shortest possible time. Initially, such production processes were relatively inefficient and there was tremendous scope for process optimization. Advances in cell biology, in addition to improvements in process monitoring, control, and optimization, and the development of novel bioreactor designs are making the transition from the research laboratory to industry standard practice. Clearly, this will lead to a reduction in the cost and complexity of the process of recombinant protein production using mammalian cell lines by placing the technology on a firmer scientific footing.
In order to optimize the viable cell number and protein productivity, a detailed understanding of the factors that lead to cell death in the bioreactor is required (Fig. 5). These factors may be classified into four groups:
The hydrodynamic environment of the cell.
The accumulation of toxic metabolites.
The exhaustion or local limitation of nutrients and oxygen.
Cellular factors giving rise to stress.
Although each of these areas has now been investigated to varying degrees, recent studies have demonstrated that at least some of the cell lines used in bioreactors undergo apoptotic death rather than necrosis, indicating that a reassessment of the subject is required. As described later, the greater understanding of the mechanism of cell death in commercial cultures should provide new routes to culture optimization. The resultant enhancement in culture efficiency would be expected to manifest itself in three forms:
The nutrients and culture time invested in generating a viable cell will be wasted if that cell should die prematurely. If the survival time of the cell can be enhanced, the proportion of culture resources utilized for production of the biopharmaceutical of interest can be increased by eliminating the need for regeneration of cellular biomass.
As a cell dies, it releases proteolytic enzymes into the culture medium, which can lead to degradation of the product. Thus, product stability and even glycosylation profiles should be enhanced by minimization of cell death.
High levels of cellular debris in the culture medium can complicate the recovery of the target protein. This will add to the cost of downstream processing and lead to a reduction in the efficiency of target protein recovery.
3.1 A Brief History
The first suggestion that apoptosis may account for cell death during the cultivation of hybridoma cells came from an electron microscopic study conducted by Al-Rubeai et al. (86). Further studies by Franek and Dolnikova (87) demonstrated the accumulation of nucleosomal DNA fragments in culture medium during the death phase of batch hybridoma cultures. On the basis of this observation, they estimated that around 30% of cells had undergone apoptotic death. Further evidence of apoptosis during hybridoma cultures was provided by the studies of Mercille and Massie (113) and Singh et al. (114). Upon DNA gel electrophoresis, both studies revealed the laddering pattern that, as mentioned earlier, is characteristic of apoptosis. Morphological analysis of the nuclei of the cells indicated that apoptosis accounted for 90% of the dead cells. Both groups also found high levels of apoptosis during the cultivation of murine plasmacytoma cell lines (sometimes incorrectly referred to as myeloma cells). Furthermore, Singh et al. (115) reported absence of apoptosis during the death phase of CHO and Sf-9 batch cultures. However, studies by Moore et al. (90) indicated significant levels of apoptosis during the death phase of serum-free batch cultures of CHO cells. Recent studies in our laboratory on a CHO 320 cell line overexpressing interferon indicate that this cell line may also be susceptible to apoptosis, although the morphology was not typical, and the frequency was much lower than that seen during hybridoma cultures under comparable conditions.
Studies are now required to give an indication of variability in susceptibility to apoptosis between clones of the same cell type, and between different unrelated cell lines. The objective will be to produce a correlation between susceptibility to apoptosis and general cell robustness. Clearly, such a correlation would provide a simple and easily identifiable predictor of robustness following exposure of cells to a range of stresses, thus simplifying the process of cell line selection.
3.2 Induction and Suppression of Apoptosis in the Bioreactor Environment
The features of the bioreactor environment that may result in cell death were outlined in the introduction. In the present section, studies that have considered these factors in terms of their ability to induce apoptosis are described and the implications of suppression of apoptosis are discussed.
3.2.1 Death during Batch and Fed-Batch Cultivation
The nutrient limitations encountered by cells in the bioreactor may be classed into two groups: cycling or terminal limitation. The former may be encountered at all stages of large-scale or intensive culture systems. For example, in large-scale stirred tank reactors, the cells may be exposed to fluctuating nutrient levels because of inhomogeneity due to poor mixing. In intensive culture systems, the high cell densities reached will result in low local-nutrient concentrations. As a result, the level of cell death in such systems is relatively high. Terminal nutrient limitations will occur at the end of batch cultivation and, as a result, will become more extreme with time, invariably leading to cell death. In the case of hybridoma batch cultures, the first nutrient to become limiting is glutamine, and its exhaustion coincides with the onset of the death phase of the culture. As stated above, cell death under these conditions is almost exclusively by apoptosis. It was also found that in batch cultures of Chinese hamster ovary cells grown in serum-free media, more than 80% of the cells died of apoptosis (116).
Two studies have reported on the effect of bcl-2 overexpression on cell survival during the death phase of hybridoma cultures. Itoh et al. (91) found that bcl-2 overexpression significantly extended the duration of the culture by reducing the rate of cell death. Moreover, they reported a fourfold increase in the Mab (monoclonal antibody) titer in the culture medium. Simpson et al. (10) have also reported an extension in culture duration, although there was only a 40% improvement in Mab titer. Necrosis became the predominant mechanism of cell death in the bcl-2-transfected cell line, indicating near-complete suppression of apoptosis under batch culture conditions. Suzuki et al. (117) have reported that transfection of COS-1 cells with bcl-2 and then with the vector pcDNA-λ carrying the immunoglobulin λ gene for transient expression of λ protein have resulted in higher expression of the protein when compared to the control transfectant (i.e. bcl-2 negative). In the same study, the mouse plasmacytoma p3-X63-Ag.8.653, which is used as a fusion partner in the generation of hybridomas, and the hybridoma cell line 2E3 were transfected with the human bcl-2 gene. In both cases, an extension of batch culture duration was reported. The bcl-2-transfected 2E3 cells survived 2–4 days longer in culture, producing a 1.5- to 4-fold larger amount of antibody in comparison with the control vector transfectants. A further enhancement in survival and antibody production in hybridoma 2E3 cultures was observed when cells were cotransfected with bcl-2 and bag-1. Jung et al. suggested that the constitutive high-level expression of antiapoptotic genes had detrimental effects on the genomic stability of cultured cells and thus implemented an inducible bcl-xL hybridoma cell line (118). These cells were shown to be able to lengthen the culture time when bcl-xL was induced to expression at the stationary phase of a batch culture. Other genes demonstrating apoptosis protection in cell culture include XIAP, a potent caspase inhibitor (119).
In contrast to these promising studies, Murray et al. (120) found that bcl-2 transfection of the murine plasmacytoma NS0 failed to provide any protection from apoptosis. Although there was no endogenous Bcl-2 expression, they did report expression of Bcl-xL, a functional homolog of bcl-2. Thus, they suggested that Bcl-2 may be functionally redundant in this cell line. In other cases, although apoptosis was suppressed in NS0 cells and growth was enhanced (121), antibody productivity was decreased (122).
With the introduction of fed-batch feeding in most bioreactor systems, nutrient limitation is mostly eliminated. Despite this, apoptosis continues to be a common feature in these fed-batch cultures. To understand why this is so, transcriptional profiling of Chinese hamster ovary cells have been carried out on batch and fed-batch cultures (123). It was found that in batch cultures, the upregulation of proapoptotic genes (FasL, Rip1, Bak, Caspase-8) occurred at the stationary phase of the culture due to the nutrient limitation. However, in fed-batch cultures where nutrient limitation was not evident, apoptosis genes (Fadd, Bim, Bad, Requiem, and Alg-2) were upregulated in the exponential and stationary phases of culture. These represented the signaling primarily via death receptors and mitochondria rather than the ER. The subsequent inhibition of apoptosis in these cultures by using small interfering RNA (siRNA) technology led to higher viabilities, increased productivity (over twofold) and better product sialylation (124). The knock-down of Bax and Bak using similar siRNA technology also resulted in increased productivity and viability of Chinese hamster ovary batch and fed-batch cultures (125).
3.2.2 Nutrient Limitation
184.108.40.206 Amino Acids, Inorganic Salts, Vitamins, and Glucose
The link between the onset of apoptosis and exhaustion of glutamine during batch cultures of hybridoma cells has prompted systematic studies of the role of the various nutrients used in culture medium. Initial studies indicated that deprivation of glucose, serum, glutamine, cysteine, and methionine (126-128) could all individually induce high levels of apoptosis. Studies in our laboratory (129) indicate that this is not a feature of these particular nutrients alone. Deprivation of each amino acid individually from the commonly used RPMI 1640 culture medium was found to result in the induction of apoptosis, with particularly high levels observed following deprivation of essential amino acids.
How might the deprivation of nutrients trigger an apoptotic response? Perreault and Lemieux (130) have found that hybridoma cells undergo apoptosis when their protein biosynthetic machinery is compromised. It may be that deprivation of nutrients such as amino acids has the same effect, possibly resulting in the failure of the synthesis of a critical regulatory protein required to keep the apoptotic pathway in check. Vaux and Strasser (131) suggest that some agents or treatments lead to a reduction in the cellular adenosine triphosphate (ATP) pool, and that this may be a trigger of apoptosis. They propose that such changes may be interpreted by the cell as being a consequence of viral infection, and the cell responds by inducing the apoptotic pathway. Overexpression of Bcl-2 was found to offer a high degree of protection following deprivation of each individual amino acid, with two exceptions, glutamine and threonine, exhibiting relatively less protection (129). This may indicate that these two amino acids either play a particularly important role in cellular metabolism and biosynthesis or that they are essential components of the mechanism by which Bcl-2 protects the cell.
The survival of bcl-2-transfected cells, even in the absence of supposedly essential amino acids, suggests a reduction in amino acid utilization due to downregulation of nonessential cellular functions. Indeed, metabolic arrest has been reported following withdrawal of interleukin 3 (IL-3) from the IL-3-dependent murine cell line, which consequently underwent apoptosis. This state was stabilized by bcl-2 transfection of the cells, thus extending survival time by 300% (132). Presumably, Bcl-2 also stabilizes the metabolic arrest caused by amino acid starvation in murine hybridoma cells, possibly by maintaining the ATP pool above a threshold level. Moreover, studies by Simpson et al. (10) suggest that this state may be reversible by feeding the cells with fresh medium. It was also found that the increase in apoptosis was paralleled by an increase in the expression of the growth arrest—and DNA damage inducible gene 153 (gadd153) in NS0 cells—and that the addition of glutamine was therefore able to delay the onset of apoptosis (133). Clearly, such a characteristic is far more desirable than a rapid, and obviously nonreversible, entry into apoptosis that occurs in apoptosis-susceptible cell lines.
Overexpressing other antiapoptotic genes can also protect cells from amino acid or glucose deprivation. Aven and Bcl-xL have shown synergistic effects in protecting Chinese hamster ovary cells in various insults including growth in spent media (134) while the MDM2 (an E3 ubiquitin ligase for p53) overexpression could do the same in HEK293 and Chinese hamster ovary cultures (135). E1B-19K overexpression in NS0 cells grown in glutamine-free medium in batch and perfusion cultures resulted in a decrease in the death rate and a higher overall productivity (136). Oxygen uptake rates were seen to decrease in the case of E1B-19K overexpression and the sustained viabilities were ultimately responsible for the increase in overall yield. Bcl-2 can also suppress apoptosis as a result of the deprivation of one or several B-group media vitamins (D-CaPantothenate, choline chloride, and riboflavin) (137). Out of three different ions (Ca2+, Mg2+, and K+ ions), the deprivation of K+ ions was found to be most effective in suppressing hybridoma cell growth and viability, while Mg2+ caused the cells to die mostly of necrosis and Ca2+ caused the greatest amount of apoptosis (138).
The role of serum in the suppression of apoptosis has been well documented, and consequently, it was no surprise when it was reported that commercially important hybridoma and plasmacytoma cell lines also undergo apoptosis on withdrawal of serum (127, 139). This is expected to have important consequences for the development of new serum-free media formulations. Previously, the rational behind the design of such media was not particularly scientific and often involved the inclusion of chemicals that were identified by empirical studies. However, demonstration of the role of serum in the regulation of apoptosis may provide a new avenue of research for the development of novel, and perhaps cheaper, serum-free media. The Bcl-2-mediated suppression of apoptosis following serum withdrawal was the earliest demonstration of the antiapoptosis activity of this gene. A Burkitt's lymphoma cell line transfected with bcl-2 has been reported to grow better than control vector-transfected cells in commercially available serum-free media without the need for adaptation (140). Similar results were obtained using a bcl-2 murine hybridoma cell line.
220.127.116.11 Media Supplementation
The supplementation of the media with various molecules not naturally present has also been shown to reduce apoptosis. Synthetic peptides consisting of three to six amino acid residues have been shown to have survival factor-like activity in Chinese hamster ovary cells, resulting in an over-increase in productivity (141). Rapamycin, which affected the mTOR pathway, was able to reduce hybridoma cell death and enhance antibody productivity, while increasing the percentage of cells in the G1 phase (142). Some other media additives also act to block events within the apoptosis signaling cascade. The use of caspase inhibitors such as N-acety-Asp-Glu-Val-Asp-aldehyde or benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone have been shown to suppress apoptosis particularly under nutrient deprivation conditions (143). However, the inhibition of caspase-8 activity by benzyloxycarbonyl-Ile-Glu-Thr-Asp-fluoromethylketone and caspase-9 activity by benzyloxycarbonyl-Leu-Glu-His-Asp-fluoromethylketone (144) does not provide complete cytoprotection as mitochondrial dysfunction and plasma membrane permeabilization may still be observed (145).
3.3 Oxygen Limitation
In large-scale and intensive culture systems, effective aeration of the culture is a major difficulty. Thus, oxygen limitation is often the major limiting factor determining maximum cell number. Studies by Mercille and Massie (146) demonstrated that deprivation of oxygen can induce apoptosis. Subsequently, it has been shown that Bcl-2 overexpression protects hybridoma (10) and Burkitt's lymphoma cells (113) from apoptosis. Clearly, this will provide the cells with a considerable advantage in oxygen-limited intensive culture systems.
3.4 Hydrodynamic Stress
During the very early days of mammalian cell culture, it was generally assumed that cell lines would be far more sensitive to shear damage due to the absence of a cell wall. Considerable progress has been made in our understanding of the exact interactions that result in the greatest damage to the cells. It is now clear that it is not the shear forces generated at the impeller tip that are responsible for cell death induced by the hydrodynamic environment of the reactor. Far more damage is caused by the events that take place at the gas headspace–liquid interface during bubble disengagement.
Despite the importance of the hydrodynamic environment, the biochemical response of the cell to this component of the bioreactor environment has been rather neglected. To address this issue, Al-Rubeai et al. (15) investigated the mechanism of cell death following exposure of cells to very high agitation levels. Flow cytometric and morphological analysis indicated that cell death occurred mostly by apoptosis, although levels of necrosis were also significant. Singh et al. (115) found that a Burkitt's lymphoma cell line that had been routinely passaged in stationary cultures underwent apoptosis when attempts were made to grow the cells in suspension. Bcl-2 transfection of this cell line was found to allow much better cell growth in suspension without the need for adaptation. Simpson et al. (10) reported similar behavior of a hybridoma cell line.
Physiological studies have also investigated the relationship between apoptosis and shear stress. Dimmeler et al. (131) found that shear stress actually prevents induction of apoptosis in endothelial cells in the presence of the inducer tumor necrosis factor α or following growth factor withdrawal. Perani et al. (147) showed that bcl-2 transfected cell line grown in shear stressed condition exhibited a nearly fivefold increase in viable cell number compared to the nontransfected control and suggested that under apoptosis-suppressed conditions, shear stress can stimulate cell growth. Clearly, the possibility that a sublethal level of shear stress in the bioreactor protects the cells from apoptosis-inducing agents needs to be investigated.
3.5 Cellular Stress Leading to Apoptosis
The generation of cellular stress can result from processes within the cell. ROS can be generated as a result of high mitochondrial activity or cellular metabolism in high productivity cells. Even mild oxidative stress can be an inducer of apoptosis, either by the upregulation of the FAS pathway or by stimulating cytochrome c release from the mitochondria (148). Early indications show that ROS generation in NS0 cells with high antibody production is responsible for the higher apoptosis rate seen in those cells. Consistent with this, antioxidants such as N-acetylcysteine have been shown to prevent apoptosis (149, 150). As mentioned earlier, ER stress can also be an inducer of apoptosis. This is particularly important in the biotechnology environment as cells are cultured for their primary ability to produce complex proteins. To date, there have not been many reports on the influence of ER stress on recombinant protein production; however, it is acknowledged that such signaling will have significant impact on the production of recombinant proteins (151). The induction of recombinant protein in HEK293 cells has been shown to coincide with increased expression of ER folding chaperone gene BiP, suggesting that the high-level expression of complex proteins does cause some form of cellular stress (152). In yeast, the coexpression of BiP and PDI (protein disulfide isomerase) was found to be required to reduce the unfolded protein response initiated by ER stress (153).
3.6 Growth Arrest, Specific Productivity, and Apoptosis
Studies of p53 and c-myc have revealed a close relationship between cell proliferation and apoptosis. These studies may have important implications for process optimization strategies that have centered on the control of cellular proliferation. Such an approach allows for improvements in specific recombinant protein productivity during cultures, exhibiting a negative correlation between growth rate and productivity. Furthermore, by controlling maximum cell number at an optimal level, cell death due to limitation of nutrients and oxygen should be minimized, thus simplifying medium clarification during downstream processing. However, such studies have had one major drawback: when attempts were made to control cellular proliferation, the cultures rapidly lost viability (154, 155), and it appears that, at least in the case of hybridoma cultures, this was due to the induction of apoptosis (127). Thus, strategies designed to control cellular proliferation must also incorporate methods that minimize apoptosis. Initial results suggest that such an approach does work. Simpson et al. (10) found that Bcl-2 overexpression delays hybridoma cell death following cell cycle arrest induced by thymidine treatment. Similar results have been reported during Burkitt's lymphoma cultures (114).
4 Future Prospects
- Top of page
- Apoptosis and Animal Cell Biotechnology
- Future Prospects
- Further Reading
The study of apoptosis from an animal-cell-technology standpoint has been well documented over the last decade. Apoptosis is important in industrial mammalian cell lines as well as in bioreactor systems as highlighted in several reviews (156-158). Various means of abrogating such phenomena in biotechnological applications so as to increase overall product yield has led to different antiapoptotic genes, feeding regimes, additives, and siRNA methods being used. It has become an important parameter in determining the viability of such cultures and has led to various ongoing projects for its rapid on-line determination. Hence, in the quest for the optimal production process, it would inevitably result in the need to reduce factors that cause apoptosis (e.g. feed regimes or removal of signaling factors) or engineer cells to be resistant to apoptosis. The recent link between ER stress and apoptosis has created a new interest in understanding the effect of recombinant protein biosynthesis on ER function. We are likely to see more studies being done in this area as the ER represents a critical path for protein folding, glycosylation, and secretion.
MAR gratefully acknowledges Science Foundation Ireland (SFI) for partial funding.
- Top of page
- Apoptosis and Animal Cell Biotechnology
- Future Prospects
- Further Reading
- 5Sanderson CJ Clark WR, Golstein P, editors. Mechanisms of cell mediated cytotoxity. New York: Plenum; 1982. pp. 3–21.
- 20Am J Physiol Cell Physiol 2003; 284: 1309–1318., , , , .
- 22Blood 1994; 8: 1415–1420., , , , , .
- 52Allen DJ, Harmon BV, Kerr JFR Potten CS, editors. Perspectives on mammalian cell death. Oxford: Oxford University Press; 1987. pp 229–258.
- 62J Biol Chem 1991; 266: 580–585., .
- 85Cooney LC Beuvery EC, Giffiths JB, Zeijlemaker WP, editors. Animal cell technology: developments towards the 21st century. Netherlands: Kluwer Academic; 1995. pp. xxv–xxx.
- Top of page
- Apoptosis and Animal Cell Biotechnology
- Future Prospects
- Further Reading
- Cell 1993; 74: 597–608. , , , , , , , , .
- Nature 1995; 374: 731–733. , , , , , , , .
- Proc Natl Acad Sci U S A 1993; 90: 3516–3520. , , , , .
- Curr Opin Biotechnol 2001; 12: 175–179. , .
- Cell 1995; 80: 285–291. , , , , , .