Computer model clears up immunological conundrum
BERKELEY – A theoretical chemist and chemical engineer at the University of California, Berkeley, has stepped into a budding controversy among immunologists, helping to clear up confusion over a potentially key part of the body's defense against viruses and cancer.
The issue revolves around T cells and how they are triggered to mount an attack on invaders, which could have implications for disorders involving faulty regulation of the immune system, according to Arup Chakraborty, the Warren and Katharine Schlinger Distinguished Professor and Chair of Chemical Engineering at UC Berkeley.
The new findings, to be published online this week in Science Express, "could be the hidden factor in autoimmune disease," said co-author Michael Dustin, the Irene Diamond Associate Professor of Immunology at New York (NYU) University School of Medicine.
The story began five years ago, when scientists at Washington University in St. Louis and NYU discovered a structure, dubbed the immunological synapse, that they thought was an important piece of the body's T cell reaction to infection. The synapse forms at the contact point between a T cell and the immune cell activating it, and remains for more than an hour, as if the immune cell were continually pushing the T cell's "on" button.
At least that's what scientists thought until the same researchers two years ago found conflicting experimental data that gave them second thoughts. They essentially backed off from their previous hypothesis, and instead claimed that the synapse had nothing to do with T cell signaling.
Enter Chakraborty, who maintains a UC Berkeley appointment as professor of chemical engineering and chemistry and is a researcher in the Physical Biosciences and Materials Sciences Divisions at Lawrence Berkeley National Laboratory (LBNL). A theoretician, he had earlier modeled how the immunological synapse forms as big protein molecules migrate along the T cell surface to form a long-lived, doughnut-shaped signaling structure.
He developed a different model of how the complicated signaling could work at the synapse and still give seemingly conflicting results in separate experiments. Running the model on a cluster of microcomputers at LBNL, he and his coworkers were able to show that the synapse is a sophisticated structure, quickly ramping up T cell activity and just as quickly backing off so as not to overstimulate the cell and kill it. Chakraborty's model suggested subsequent experiments that the NYU and St. Louis teams conducted, confirming his hypothesis.
"Together, our three labs discovered that this structure is an adaptive control device, that is, it enhances the sensitivity of T cells to antigen, but, beyond a threshold, it cuts off signaling to prevent T cell death," Chakraborty said. "This is an example of how, in complex biological systems, a complementary and synergistic use of genetic and biochemical experiments with theory and computation can lead to rich dividends."
This synergy between experiment and computation is the premise of the California Institute for Quantitative Biomedical Research (QB3), a center funded by the State of California and private funds and based at UC San Francisco, UC Berkeley and UC Santa Cruz. Chakraborty is a UC Berkeley member of QB3.
"We only realized this with the use of a computational analysis that allowed us to see how all these different variables were playing out," said Washington University researcher Andrey Shaw, professor of pathology and immunology. "There's a lot of talk that goes around about this need for a union between computational biology and what I would call wet biology, and I think it's hard for most of us to imagine how that would work .... But this was a case where I really thought it was beautiful, it worked together so perfectly."
Researchers are hopeful that now that the role of this channel of communication has been identified, it may serve as a potential target for treating diseases - those in which the body attacks itself, such as arthritis, as well as those in which the body doesn't recognize the attacker, such as tumors.
T cells are white blood cells that specialize in attacking viruses and cancer cells, aided by antigen-presenting cells that present evidence of infection - essentially pieces of the invader, called antigens - to the T cells. This presentation was, until recently, thought to be brief, though triggering massive proliferation of T cells to attack the invader. But Shaw and Dustin showed in 1998 that the contact between the two cells was long-lived, giving rise to a doughnut-shaped structure at the contact point.
They showed that the structure, named in analogy with the permanent synapses between nerve cells, was generated by T cell receptor molecules migrating along the surface of the T cell to the contact point, forming a sort of bull's eye. At the center are T cell receptors specific for antigens, ringed by sticky molecules that help the cells adhere to one another. Their assumption was that the immunological synapse served to amplify signals between cells.
Confusion arose last year when the Washington University and NYU teams found that during the height of the immunological synapse formation, when the signaling should have been the strongest, there was only the merest trickle of signal. Most of the T cell receptors were in fact being degraded - in other words, ripped apart and recycled by the body - which only happens after they've finished the majority of their signaling.
"This created a kind of a shock to the field," recalled Dustin, "because everyone had been banking on this central cluster as a site of very active signaling."
In an attempt to resolve the conflict, Chakraborty and his UC Berkeley/LBNL team created a computer model of the synapse that took into account the various competing activities. He calls this an in silico experiment, in contrast to an in vivo experiment in live animals or an in vitro experiment in a glass dish.
"In complex biological systems like this, many different processes which compete with each other occur in concert. It is often very difficult to parse the relative importance of each of these processes and how they influence each other just by looking at one readout of the process, like what one signaling molecule is doing," Chakraborty said.
He subsequently sent a paper explaining the conflicting results to Shaw and Dustin for comment, and they proposed an experiment that would prove or disprove Chakraborty's hypothesis. Their suggestion involved genetically engineering mice so that they could not degrade T cell receptor molecules. If the synapse is important in signaling, as Chakraborty thought, then this genetic modification should produce mice that exhibit sustained signaling and T cell stimulation.
That's exactly what Dustin found when he tested signaling in the altered mice.
"They saw continued triggering in these cells," Chakraborty said. "And because the T cells are triggering so much, they both proliferate more but they also die more, probably because of overstimulation. It is like their control mechanism is not working."
CD2AP, the protein whose levels had been lowered in Shaw's most recent experiments, turned out to be involved in the synapse's ability to dampen signaling by pushing activated receptors on the surface of the T cells toward the lysosome, a kind of cellular garbage can.
"We used the term adaptive controller, an engineering term, to describe the synapse," Shaw said. "It helps to amplify weak signals by concentrating ligands and receptors in the same area of the cells. But at the same time, it prevents strong signals from overpowering the cells - which in most cases would lead to cell death - by rapidly turning off the very strongest signals."
Dustin explained why such a range of sensitivity is necessary for the immune system.
"In some cases, you have pathogens that, perhaps because they're trying to evade the immune response, won't generate many of these antigenic structures. So the T cell has to be very sensitive to detect them," he said. "At the other extreme, in the evolution of these systems there was some pathogen that figured out that if it could swamp this system with one antigenic structure, it would just blind it. So the immunological synapse allows the system to adapt to very strong signals by reducing the T cell receptor density and then arrive at a uniform signaling rate in any situation."
"One important thing about this study," said Mark Davis, professor of microbiology and immunology at the Stanford University School of Medicine and an investigator in the Howard Hughes Medical Institute, "is that it gives us some real clues as to what the function of the immunological synapse might be, and that's been a big question."
Davis, an expert in this field who is familiar with the new study, adds that this research sheds light on some of the basic processes of the immune system.
"Most of the time, when you are sick, it's because your immune system is either doing too much or too little. As we learn more about the basics, it becomes obvious how to intervene and make things more sensitive or less sensitive."
Such treatments, he noted, could lessen the severity of diseases in which the immune system mistakenly attacks its own tissues, as in arthritis or multiple sclerosis, and conversely, thwart those in which the body fails to eradicate or even notice an attacker, such as some tumors.
"The other thing that's important about this paper," added Davis, "is that - almost for the first time - it's a marriage of theoretical biology, using a simulation, with real-time experimental biology. To me, that's a logical evolution of biology, and this is the most striking example I've seen. It's not something that's at all common right now, but it's where things should go."
Other coauthors on the Science Express paper are postdoctoral fellows Aaron R. Dinner and Subhadip Raychaudhuri of UC Berkeley; Kyeong-Hee Lee, Chun Tu, W. Richard Burack, Hui Wu, Julia Wang, Osami Kanagawa, Mary Markiewicz and Paul M. Allen of Washington University; and Gabriele Campi, Tasha N. Sims and Rajat Varma of the Skirball Institute of Biomolecular Medicine at NYU. Dinner is now at the University of Chicago; Lee now is at Genentech Corp. of South San Francisco.
Science Express provides rapid electronic publication of selected Science papers. A print version of the paper will appear in Science in several weeks.
Funding from the National Institutes of Health, the Psoriasis Foundation, the Irene Diamond Foundation, the Burroughs-Wellcome Fund, and the National Science Foundation supported this research.