Monoclonal Antibody Production (1999)

Although in vitro techniques can be used for more than 90% of mAb production, it must be recognized that there are situations in which in vitro methods will be ineffective. Because hybridoma characteristics vary and mAb production needs are diverse, in vitro techniques are not suitable in all situations, and requiring their use might impede research, especially if large numbers of mAb have to be screened for efficacy or specificity in the treatment of disease. In some cases, in vitro production of mAb has not met the scientific aims of a project. The National Institutes of Health (NIH) has identified many of these in its response to the American Anti-Vivisection Society (AAVS), as shown in appendix C of the NIH response (Varmus 1997). The committee reviewed appendix C and offers the following explanation for the items listed in the appendix based on the collective experience of its own members.

1. Some hybridoma cell lines do not adapt well to in vitro conditions. Although in vitro methods produce mAb from over 90% of hybridomas, there is a finite and significant failure rate. The NIH response to the first AAVS petition suggested that the failure rate is 4% (Varmus 1997). That is consistent with the 3% failure rate observed by Dutch scientists (Hendriksen and others 1996). A recent European workshop discussed the effects of restrictions on the ascites method in various European countries; each country's laws provide for an exception based on the inability of a hybridoma to grow and produce mAb in vitro. Countries that maintain data banks on requests for exceptions continue to issue such exemptions (Marx and others 1997). Although in vitro conditions are used initially to select mAb-producing hybridomas, the initial culture contains many

normal spleen cells that can act as feeder cells. In some instances, continued in vitro culture does not support hybridoma growth; in these instances, the rising concentration of antibody might adversely affect hybridoma growth or secretion. Transfectomas—myeloma lines transfected with mutated antibody sequences, which are often used to determine structure-function relationships—are notoriously low antibody producers. In general, the only way to obtain adequate amounts of antibody for experimental study from such lines is to use the ascites method.

2. mAb from mouse ascitic fluids might be essential for experiments in which mAb are used in mice. There are, in the committee members' experience, numerous examples to support this observation. The need for the mouse ascites method arises when small volumes of concentrated antibody are needed for a rapid screening in mice in order to select hybridomas with the desired bioreactivity. In vivo studies often examine the ability of an antibody to block a receptor-ligand interaction, to inhibit some aspect of microbial pathogenesis, or to induce the lysis or apoptosis of a particular cell type. To assess antibody function in these situations fully, high concentrations of mAb are often necessary. The mouse ascites method is also required when foreign (nonmouse) proteins could confound results. Halder and others (1998) have stated that mAb produced with an in vitro method should be equally suitable and that ascites contains other factors, such as cytokines, which could render the use of ascites fluids “scientifically wrong.” Although mAb can be produced in vitro, the time required to adapt a hybridoma to media containing 1% or less FBS (which can take several weeks and does not include downstream purification) would severely retard progress directed at selecting a hybridoma that is active in vivo. Because mAb concentration is high in ascitic fluid, only a small volume of the fluid needs to be injected into the mouse to test for effect. Although this small volume might contain small amounts of other factors, such as cytokines, no biologic effect due to these factors is noted. There are three reasons for this observation: the project is not affected by small amounts of contaminants, the contaminant is diluted in the body fluids, and the biologic half life of the contaminant is short (hours) relative to mAb half-life (days). Contaminating antibodies can be avoided by using mice with severe combined immunodeficiency disease syndrome. Semipermeable-membrane-based systems have been developed in which several hybridomas could be grown simultaneously. More experience is needed with this technique to determine whether it will meet the need for rapid screening of many hybridomas to find a cell line that produces a therapeutically effective mAb.

The mouse ascites method for mAb production might be the only choice when contamination of antibody with other mouse proteins does not interfere with the intended scientific goals (especially when the negative controls are also ascites-based). Similarly, the small-scale production of mAb for initial screening as potential diagnostic reagents when several different mAb need to be screened

simultaneously would be hampered if the mouse ascites method could not be used.

Studies can be seriously confounded by purification procedures that alter the native structure of mAb and result in a loss of reactivity with antigen or loss of ability to bind components of the complement system. In many cases, denatured antibodies copurify with active antibodies and interfere with the in vivo function of the active antibodies. Denatured antibodies are more likely to be taken up by phagocytic cells or removed from the circulation by other clearance mechanisms; denaturation can lead to enhanced immunogenicity of the antibody preparation and thus result in a shortening of antibody retention time after in vivo administration. We recognize, however, that when hybridoma selection has been made and large-scale production of pure antibody is needed, in vitro cultures are preferable.

3. Rat hybridoma cell lines do not generate ascites efficiently in rats, usually adapt poorly to in vitro conditions, but usually generate ascites in immunocompromised mice. In some situations mAb to mouse epitopes are required, necessitating the use of another species (usually rat) for immunization. Although some rat hybridomas adapt to in vitro conditions, this often requires tedious manipulation of the culture. When small volumes of concentrated rat mAb are needed and the hybridoma does not easily adapt to culture conditions, the mouse ascites method using immunocompromised mice is required (Wolf 1998). However, if large-scale production (especially of purified antibody) is required, attempts should be made to adapt the rat hybridoma cells to in vitro growth. Other investigators have found that rat-mouse or hamster-mouse fusions yield heterohybridomas that are less stable than rat-rat hybridomas and for that reason have selected the mouse ascites method to obtain high-concentration mAb quickly for testing before extensive recloning procedures are used in preparation for large-scale in vitro production (Ohlin and Borrebaeck 1994).

4. Downstream purification can lead to protein denaturation and decreased antibody activity. When a pure product is not necessary for research goals but maintenance of high affinity and biologic activity is necessary, the mouse ascites method often offers the best option. There are many laboratory situations in which the concentration of antibody obtainable by current in vitro methods is not high enough for experimental studies and absolute purity of the antibody reagent is not essential. Other situations that require the mouse ascites method of producing mAb are related to the need for high binding affinities, the presence of complement-fixing activities, and mAb that are naturally glycosylated. Many of the in vitro-produced antibodies cannot be readily concentrated from culture supernatant, because standard procedures result in losses of antigen binding activity or other antigen-antibody features (Underwood and Bean 1985; Lullau and others 1996), although such a concentration step might not be required with semipermeable-membrane-based systems. For example,

immunoglobulin M (IgM) and immunoglobulin G3 (IgG3) antibodies often undergo denaturation during in vitro purification techniques, resulting in the loss of complement-binding activity (Roggenbuck and others 1994). Random antibodies of other isotypes exhibit similar quirks. OKT3 is an excellent example of an mAb with substantial therapeutic application; it cannot be adequately purified from culture fluids and retain full function, so it must be produced by the ascites method (Stein 1998).

Downstream purification is particularly difficult for immunoglobulin A (IgA) mAb, in which monomeric IgA (with poor antigen-binding abilities) must be separated from dimeric and polymeric IgA (Lullau and others 1996). This problem is alleviated by the mouse ascites method of IgA production.

5. Serum-free or low-serum conditions cannot provide sufficient amounts of mAb for some purposes, such as the evaluation of new vaccines against infectious organisms. Some cell lines can be readily adapted to low-serum or serum-free conditions, but others cannot (Stein 1998; Chandler 1998). More important, it has been noted (Chandler 1998) that some cell lines that appear to be maintained adequately in serum-free or low-serum media, as assessed by viability, but they make less than 10% as much antibody under these conditions compared to their being maintained in higher-serum media. If media with 1% serum result in 10% as much antibody production as media with 10% serum, nothing is gained in purity or yield that warrants the expense and time needed to adapt the cells to the modified culture conditions. The quality of serum can vary from batch to batch and manufacturer to manufacturer, and adapting a cell line to 1% of a particular batch of serum does not guarantee that the same cell line will grow comparably in 1% FBS obtained from another batch.

Those observations are related to manufacturing quality-assurance issues that are especially important to the Food and Drug Administration. Adapting hybridoma cell lines, initially approved for ascites-generated mAb, to serum-free conditions requires the hybridoma owner to demonstrate analytic comparability. Alterations in mAb binding affinity or other biologic functions could result in expenditure of millions of dollars (Maxim 1998).

Some investigators report difficulty in adapting hybridomas that produce IgM or IgA antibodies to serum-free conditions (Varmus 1997). The reason for emphasis on IgM and IgA mAb production is that IgM is a potent complement-fixing antibody generated early in the human immune response in many infectious diseases. IgA is associated with a variety of human diseases (such as Bergers IgA nephropathy, now one of the most common types of glomerulone-phritis and Henoch-Schönlein vasculitis and glomerulitis), in none of which cases is the pathogenesis understood that could lead to effective clinical treatment. These observations indicate the need for production of IgA and IgM isotypes that are biologically active and exhibit high affinity. The committee recognizes that some success has been obtained in the in vitro production of IgA mAb; however,

very few IgA-secreting hybridomas have been tested in vitro, and a high concentration of antibody generally depends on the addition of FBS to the culture medium (Stoll and others 1995; Stoll and others 1996a), prolonged incubation, and critical attention to antibody concentration to avoid production of inactive IgA molecules (Stoll and others 1997). Although Roggenbuck and others (1994) produced milligram quantities of polyreactive IgM mAb with in vitro methods, 1% FBS in the media was required, and reactions between the IgM mAb and other components of the media led to impaired solubility of the antibody and poor reproducibility of purification results. Their two-step purification technique was capable of recovering only 30% of the immunoreactive IgM. Others have observed a loss of up to 99.9% of reactivity during purification of in vitro-produced IgM (Poncet and others 1988).

The mouse ascites method might be required when mAb to infectious agents or tumor antigens are being tested for toxicity and efficacy in mouse models of human diseases. Such testing is usually needed to establish a proof of principle (that is, showing that the mAb in fact is effective therapeutically) or for the preclinical studies required by federal agencies. In those situations, large numbers of mAb of different isotypes and specificities often have to be tested in dose escalation studies before a candidate is chosen for more detailed analysis, and this requires initial production of large amounts of mAb so that enough subjects can be challenged to establish a statistically significant result. Unexpected toxicities or questions of efficacy sometimes require additional batches; in these cases, the presence of nonmouse contaminating proteins and the immune responses to them can distort the results.

6. Culture methods sometimes yield populations of IgG mAb that are glycosylated at positions different from those harvested from mouse ascites fluid, thereby influencing antigen-binding capacity and important biologic functions. Leibiger and others (1995) describe in vitro production of IgG mAb that contained terminal mannose moieties at all glycosylation sites. In some cases, such glycosylation of mAb substantially affected mAb function; in other cases, it was irrelevant. The authors attribute this unusual property to the in vitro culture conditions and speculate that the increased in vivo clearance of such antibodies was due to binding to mannose receptors. It is claimed that culture conditions can be adjusted to achieve the desired terminal sialic acid during glycosylation (Marx and others 1997), but we are unaware of any publication demonstrating this phenomenon. Indeed, manipulating the expression of glycosylation enzymes to achieve the correct in vitro placement of sugars, sialic acids, and so on, on the IgG molecule is a formidable task, extremely expensive, and often not attainable with present technology (Wright and Morrison 1994, 1997, 1998; Matsuuchi and others 1981). In vitro glycosylation patterns might yield mAb with preferred pharmacokinetic characteristics for in vivo applications (Maiorella and others 1993; Monica and others 1993; Patel and others 1992).

7. When hybridoma cells producing mAb are contaminated with infectious agents, such as yeasts or fungi, the cells often must be passed through mice. Yeast, fungal, or mycoplasma contamination of in vitro cultures of hybridoma can be removed by passing cells from the culture through mice. Removal of the organisms cannot be accomplished by current antimicrobial drugs. Thus, one mouse may save a valuable hybridoma which would necessitate more mice to be used to make new hybridomas and, in addition, months of lost time and money. Stein (1998) has independently verified the success of this technique.