Editor's Note:

An article in the Sept. 26, 2003, issue of BioWorld Today on rheumatoid arthritis research being carried out by researchers from Xencor and the University of Texas Southwestern Medical Center contained a number of inaccuracies. The article is being reprinted in its corrected form today. This version has replaced the original in BioWorld Online and in BioWorld's archives.

A good 100,000,000 years ago, a large dinosaur suffered from what we would now call rheumatoid arthritis (RA). Archeologists in 1960 named the beast Comanchean saurian, and described its seven fossilized vertebrae. (The term "Comanchean" denotes an area of the great Western plains in what is now the U.S.)

As far back as 3,000 B.C., Egyptian mummies showed unmistakable signs of arthritis - quite possibly the rheumatoid variety.

Rheumatoid arthritis is a generalized disease, occurring more often in women for reasons unknown. It attacks many of the body's joints, especially those of the hands and feet, and is accompanied by thickening of the soft synovial tissue over articular cartilage, which it erodes. RA's course varies but often is chronic and progressive, leading to deformities and disability.

American physicians treat 2.1 million patients a year with RA. Both in the U.S. and worldwide, the overall epidemiology runs at 1 percent of the population. Unlike the more common osteoarthritis, rheumatoid arthritis rots bone and cartilage to destruction.

Science, dated Sept. 26, 2003, carries an article titled "Inactivation of TNF [tumor necrosis factor] signaling by rationally designed dominant-negative TNF variants." Its senior author is Bassil Dahiyat, president and CEO of Monrovia, Calif.-based Xencor. The lead authors are Paul Steed and Jonathan Zalevsky, researchers at Xencor, and Mal Tansey, formerly at Xencor and currently at the University of Texas Southwestern Medical Center in Dallas.

"We essentially took TNF-alpha, the target in RA and several other inflammatory diseases, and redesigned it into an inhibitor of TNF," Dahiyat told BioWorld Today. "We made it do the opposite of its natural function as an agonist of TNF receptors to make it a drug candidate. Using our proprietary structure-based protein design methods, we engineered variants of TNF-alpha that neither bind to nor stimulate signaling through TNF receptors, but retain the capacity to rapidly bind to native, disease-causing TNF, thereby rendering it biologically inactive. These new molecules, called dominant-negative TNFs (DN-TNF), block the signaling of native TNF by sequestering it from its disease-mediating receptors. This approach now allows us to create inhibitors of TNF and creates a very powerful platform for a wealth of drugs against TNF, and against other targets that are structurally related.

"The current generation of TNF drugs inhibits TNF signaling to both known TNF receptors because they block all TNF-binding activity," Dahiyat said. "The initial molecules that we disclosed in this Science paper also block both TNF receptors, and we have tested these molecules in several preclinical models, which showed efficacy. Specifically, we did a rat collagen-induced arthritis model for TNF-induced joint damage and showed significant inhibition, looking at both joint swelling and histopathology analysis."

The Advantages: Cost and Specificity

"Our first preclinical program is focused on validating our new mechanism, and we selected a dominant negative that blocks both receptors to simplify comparison to existing agents," Dahiyat said. "We are very excited, however, at the potential of developing receptor-specific inhibitors. There is evidence emerging from transgenic models that one receptor, the R1, is pro-inflammatory, while the other, the R2, has anti-inflammatory activity. We have been able to engineer receptor-specific TNF variants and showed some of the data in the Science paper. Because we specifically target the ligand-receptor interface in our design work, we can tune the dominant-negative molecules for one receptor, the other, or both. R1-specific inhibitors could be important for preventing drug-related infections and for treating indications such as multiple sclerosis, where maintaining R2 signaling could be important for using TNF inhibitors. Several TNF inhibitors," Dahiyat continued, "have been investigated in MS without convincing success, and recent reports on the role of TNF in neurologic disease suggest that this lack of efficacy could be due to blocking R2-mediated regression of myelin-specific autoimmunity."

The TNF inhibitors described in the Science paper were modified with polyethylene glycol (PEG), a common tactic to extend the half-life of protein therapeutics in the body.

"We modified the DN-TNF's with PEG to extend their half-life during the rodent efficacy models," Dahiyat said. "PEGylation greatly extended their duration of action and provides us with a way to compete with the long-acting anti-TNF agents currently in use. We are able to tune the pharmacokinetics with different PEG modifications, and can even have a short-acting agent if we don't PEGylate at all. We used our structure-based design tools to make the PEG modification without disrupting the activity of the DN-TNFs."

A Powerful Platform To Inhibit The TNF Superfamily

"The TNF protein assembles into trimers, and requires trimerization for receptor signaling and biological activity," Dahiyat said. "The dominant-negative TNF molecules that we designed are able to intermix with native, disease-causing TNF, thereby pulling the native TNF into nonbinding trimers. This intermixing, or heterotrimerization, is what sequesters the native TNF and prevents it from signaling and causing disease. So we are taking advantage of how TNF assembles into trimers to create a new inhibitory mechanism.

"This mechanism is completely different from either using soluble TNF receptors, like Enbrel, a blockbuster for Amgen Inc., or monoclonal antibodies, such as Remicade, from Centocor Inc. or Abbott Laboratories' Humira. In addition to the potential advantages of specificity and cost, the dominant-negative mechanism can be applied to other proteins that are structurally related to TNF. There are more than 20 such proteins, called the TNF superfamily, and several are exciting targets. The TNF superfamily," Dahiyat went on, "is broadly implicated in immunity and inflammation, and includes proteins like RANK ligand [RANKL] which is important for bone resorption, and BLyS [BAFF] a regulator of B-cell differentiation that plays a role in autoimmune disease. The biology of the superfamily is just beginning to be understood, and it is likely that other members will be important disease targets.

"The dominant-negative approach was enabled by our strong protein-design expertise," Dahiyat said. "We use our Protein Design Automation platform to tease apart the structural determinants of receptor binding from those that control trimerization, and also to ensure that we did not destabilize the protein or cause it to become insoluble. Using this platform will allow us to continue refining dominant-negative TNF candidates and explore other members of the superfamily."