Originally intended for the treatment of B cell chronic lymphocytic leukemia (B-CLL) and rheumatoid arthritis, it is a humanised monoclonal antibody that not only binds to, but is a strong agonist for, the CD28 receptor of the immune system's T cells.

In its first human clinical trials in March 2006, it caused catastrophic systemic organ failure in the subjects, despite being administered at a supposed sub-clinical dose of 0.1 mg per kg; some 500 times lower than the dose found safe in animals. Six volunteers were hospitalized on 13 March 2006, at least four of these suffering from multiple organ dysfunction, and one trial volunteer is said to show signs of developing cancer.[citation needed]

The developing company, TeGenero Immuno Therapeutics, entered into insolvency proceedings later in 2006. Tentative opinions from an as-yet uncompleted inquiry suggest that the problems resulted from "unforeseen biological action in humans", rather than breach of trial protocols, and the case therefore has had important ramifications for future trials of potentially powerful clinical agents.

Scientists in early 2007 put forth the theory that the drug acted in a different fashion in humans as compared with the laboratory animals in which the drug was first tried. The severe reactions in humans could have only occurred, they believe, in animals with memory B lymphocytes. Animals raised in a sterile lab would presumably have no 'memory' of previous illnesses, thus would not exhibit the severe reactions that occurred in the human subjects.

The drug, which was designated as an orphan medical product by the European Medicines Agency in March 2005, was developed by TeGenero Immuno Therapeutics, tested by Parexel and manufactured by Boehringer-Ingelheim. TeGenero announced the first elucidation of the molecular structure of CD28 almost exactly one year prior to commencement of the TGN1412 phase I clinical trial.

Mice of the inbred strain BALB/c were immunized with recombinant human CD28-Fc fusion proteins and boosted with a B lymphoma cell line transfected to express human CD28. Hybridomas were obtained by fusing B cells with the hybridoma partner X63Ag8.653 and screened for reactivity with human CD28 and TCR-independent mitogenic activity. Two monoclonals called 5.11A1 and 9D7 were identified. The more active of the two, 5.11A1, is a mouse IgG1 immunoglobulin. The complementarity determining regions of 5.11A1 were cloned into the framework of human IgG and combined with IgG1 (TGN1112) or IgG4 (TGN1412) constant regions. According to the company's Investigator Brochure, "TGN1412 is a humanised monoclonal antibody directed against the human CD28 antigen. The molecule was genetically engineered by transfer of the complementarity determining regions (CDRs) from heavy and light chain variable region sequences of a monoclonal mouse anti-humanC28 [sic] antibody (5.11A1, Luhder et al., 2003) into human heavy and light chain variable frameworks. Humanised variable regions were subsequently recombined with a human gene coding for the IgG4 gamma chain and with a human gene coding for a human kappa chain, respectively." The recombinant genes were transfected into Chinese hamster ovary cells and the recombinant antibody harvested from culture supernatant.

Activation of T cells normally requires both engagement of the the antigen receptor (signal 1) and co-stimulation (signal 2). Studies of monoclonal antibodies specific for mouse, rat, or human CD28 identified so-called "superagonistic" antibodies that could stimulate T cells without concurrent antigen-receptor stimulation (signal 1). Whether this activity represents a stronger activity or a different activity is uncertain. Two antibodies specific for human CD28 were identified. The more active of the two, TGN1112 (originally called 5.11A1), belonged to the IgG1 class of immunoglobulins. The other, TGN1412 (clone 9D7), belonged to the IgG4 class. The TCR-independent agonism of these antibodies involved binding to a specific part of the CD28 molecule called the C"D loop. It was initially hypothesized that an antibody with this property could be therapeutically useful in stimulating the immune system in immunosuppressed patients. However, in vitro and in vivo data from animal studies later suggested that administration would lead to preferential activation of regulatory T cells, leading to a net effect of T-cell downregulation. On its website, the company writes: "A pronounced T-cell activation and expansion mediated by CD28-SuperMAB in animal models is accompanied by the expression of anti-inflammatory cytokines, like IL-10, rather than by the toxic cytokine storm of pro-inflammatory mediators induced by other agents that address the TCR complex.". As it turned out, the results of the first trial in humans indicate that this may not always be the case.

A new explanation for the trial mishap was suggested by the findings of a recent paper in Clinical Immunology. Pillai et al. found that all T cells that get activated using conventional TCR-mediated stimulation become regulatory for a brief time and express FOXP3. However, eventually most of these cells downregulate their regulatory capabilities and become effector cells. Thus, attempts to induce FOXP3+ T cells might also induce effector cells capable of causing tissue damage.

Other cells activated by CD28 ligation in humans are eosinophil granulocytes. They can release IFN-γ, IL-2, IL-4, and IL-13. However, most in vitro experiments are limited to the use of purified peripheral blood mononuclear cells (PBMN's) that do not contain those cells.

To function as an agonist, it has been suggested that TGN1412 needs to be a whole antibody, including the constant (Fc) region. According to a report by TeGenero, the F(ab)2 is not able to generate the required stimulation. Unlike the related clone TGN1112, an IgG1, TGN1412 is of the subclass IgG4. This choice was made as TGN1112 showed antibody-dependent cellular cytotoxicity on CD28+ Jurkat cells. Thus the function of antibody binding via an Fcγ receptor seems to be a requirement for the immune regulation. However, cell opsonisation by antibody leads normally to phagocytosis of the labeled cells, as seen in the case of HIV.