Criteria for Selection of IgG Isotype and Glycoform of Antibody Therapeutics

LEVEL: INTERMEDIATE T he efficacy of recombinant antibodies (rMAbs) results from their specificity for target antigens and the biological activities (effector functions) activated by the immune complexes formed. Numerous parameters influence the effector function profile (including antibody class, IgG subclass, and glycoform), and their definition offers opportunities for optimizing a recombinant MAb for a given disease indication. Five classes of human antibodies (immunoglobulins) are defined: IgM, IgG, IgA, IgD, and IgE; four subclasses of IgG and two of IgA are designated IgG1, IgG2, IgG3, or IgG4 and IgA1 or IgA2 according to their relative concentrations in serum. Each antibody class and subclass expresses a unique profile of effector functions (1). All currently licensed rMAbs are of the IgG class, predominantly of the IgG1 subclass, and produced in Chinese hamster ovary (CHO), mouse NS0, or Sp2/0 plasma cell lines. The choice of IgG subclass is increasingly appreciated as a critical decision when developing a recombinant MAb therapeutic (2). In oncology it would seem beneficial to maximize the potential to induce antibody-dependent cellular cytotoxicity (ADCC) and complement-dependent cytotoxicity (CDC) to eliminate targeted cancer cells. However, in chronic diseases, neutralization of a soluble target (a cytokine, for example) may be the central objective, and excessive effector activity could be detrimental. One example is when a target is also expressed as a membrane protein on certain cells, such as targeting of TNFα by infliximab (Remicade, www. centocor.com) in patients with rheumatoid arthritis. Consequently, IgG1 may not be the automatic choice, and the other IgG subclasses are now being evaluated, both in native and engineered forms. Some companies have selected IgG4 as the preferred alternative subclass, and two IgG4 rMAbs are already licensed. However, recent studies suggest that IgG4 may not be the best alternative because it has been shown to activate inflammatory reactions through cellular IgG-Fc receptors (3–5). It should be noted that the “superagonistic” anti-CD28 antibody TGN 1412, which appears to have induced a “cytokine storm” in healthy volunteers, is a humanized IgG4 recombinant antibody (6, 7). In addition IgG4 is secreted as a mixture of covalent and noncovalently assembled molecules; the latter dissociate to “half molecules” in nonreducing SDS gel electrophoresis (5, 8). The proportions of each molecular species vary between cell lines and clones, posing a problem for product consistency. A preferred alternative may be the IgG2 subclass because it is very stable, exhibits a minimum of IgG-Fc effector activities, and has a long serum half-life. Aglycosylated IgG may also be explored because it also exhibits minimal effector activity and retains the long catabolic half-life of glycosylated forms.