Editor’s Note: This is the first part of a series investigating the TeGenero AG Phase I trial of TGN1412. It will continue Friday.

In mid-March, the monoclonal antibody TGN1412, which was being tested by contract research organization Parexel International Inc. for TeGenero AG, sent six previously healthy volunteers into intensive care, hours after they were given the drug in a Phase I trial.

All six patients have shown at least some improvement since that time, and two were released from the hospital March 28. But what exactly happened to them in the first place still is unclear. (See BioWorld Today, March 16, 2006.)

Perhaps the side effects were due to contamination and had nothing to do with the antibody itself. As Carl June, a professor at the University of Pennsylvania, told BioWorld Today, "It’s always possible to have a pharmaceutical misadventure - like someone using the wrong diluent."

But TGN1412 targeted a complex immune system process known as co-stimulation, and a more likely explanation for the disastrous clinical trial is that the end clinical response represented both excitatory and inhibitory effects of several molecular interactions, and predicting their balance in the clinic is - to say the least - tricky.

TGN1412 was to be tested for rheumatoid arthritis and cancers. Prior to the FDA’s approval of Bristol-Myers Squibb Co.’s Orencia (abatacept) last December - the first FDA-approved molecule to target the co-stimulation pathway - TNF blockers were pretty much the only game in town for treating rheumatoid arthritis. (Though the FDA did approve Genentech Inc. and Biogen Idec Inc.’s Rituxan [rituximab], an antibody that targets B cells, for rheumatoid arthritis earlier this month.) (See BioWorld Today, March 2, 2006.)

It isn’t surprising that TNF blockers do not work equally well for everyone, and sometimes not at all. In a paper in the March 2006 issue of Immunity, researchers from the University of California at San Francisco described Orencia’s path to FDA approval: They stated that "T-cell co-stimulation is among the most important concepts to emerge in basic immunology in the past 30 years."

Targeting Immune System ‘Safety Deposit Box’

To be activated, T cells need to be stimulated by two signals at the same time - a highly specific antigen, which can be any substance that is foreign to the body, and a co-stimulatory signal, which is less specific but can be induced by microbes. Joaquin Madrenas, professor of microbiology, immunology and medicine at the University of Western Ontario and an expert on the co-stimulation molecule CTLA-4, likened co-stimulation to opening a safety deposit box at the bank.

"To open your safety deposit box, you need a unique key. But that key alone will not open the box. You need the bank manager’s help; he has a key that will fit any box, but cannot open any of them by itself. And together, you can open the box."

Likewise, only when both the specific antigen and the general co-stimulatory signal are present can a T cell do its job.

The main surface molecule delivering co-stimulatory signals, and the target of TGN1412, is CD28, which is expressed on most T cells and binds to a counter-receptor known as B7 on antigen-presenting cells.

But TeGenero has labeled TGN1412 a "superagonistic" monoclonal antibody, suggesting it latches to CD28 so strongly that once it is bound, T cells no longer require an antigen in order to be activated. In Madrenas’ analogy, "it is as if the bank manager alone can suddenly open your safety deposit box without your being there."

It’s a complicated system, to be sure. For one thing, T cells come in a variety of types that can either activate or inhibit the immune system. Killer and helper T cells activate the immune response; regulatory T cells shut it off.

Besides regulatory T cells, there are two other ways to brake the immune system once it’s in gear: Hyperactivity also can do T cells through apoptosis, and T cells can express the inhibitory receptor CTLA-4.

Just to add to the complexity of the co-stimulatory system, CTLA-4, the receptor that inhibits T cells, is very similar to the stimulatory CD28 - right down to binding the same B7 molecule on antigen-presenting cells, with a partially identical protein-binding motif.

The reason that inhibitory and activating binding doesn’t turn into a completely ineffective molecular free-for-all is that CTLA-4 binds more strongly to the antigen-presenting cell’s B7 receptors.

"CD28 is expressed all the time by most T cells," Madrenas said. "CTLA-4 is expressed only once the T cell is activated. But CTLA-4 binds much more strongly to B7 than CD28 does" - so even though CTLA-4 expression comes later, once it occurs the net interactions are normally favorable for putting the brakes on the immune response. In that sense, "CTLA-4 is not truly a co-stimulatory molecule. It is a co-inhibitory molecule," Madrenas said.

Nils Lonberg, senior vice president and scientific director of Princeton, N.J.-based Medarex Inc., elaborated on the differences between targeting CD28 and CTLA-4: Though their extracellular domains are similar, "the intracellular portions of the molecules are completely different. So it is incorrect to say that they act on the same signaling pathways." Furthermore, antibodies to CTLA-4 do not recognize CD28, and vice versa.

"All that homology means is that they respond to the same signals," Lonberg said. "But they respond in very different ways." And those different ways are probably part of the explanation for why the anti-CTLA-4 antibody MDX-010 now is in pivotal trials, while TGN1412 has been relegated to the trash heap; TeGenero states on its website that "there is of course no further human testing of the drug being conducted or planned at this time."