An Elegant Defense

by Matt Richtel

  What Allison and Krummel and others in the lab revealed was that tumors appeared to stymie the immune system by taking advantage of this molecule that is crucial for survival. The tumor stimulated a braking system that keeps our elegant defense from going berserk and overheating, causing inflammation, fever, autoimmune disorder, and so on. But the cancer in mice, the Berkeley scientists discovered, was sending a signal to activate CTLA-4 and thus cause the immune system to come to a standstill. In this way, the cancer could grow unchecked by the immune system.

  CTLA-4 turns out not to be the only such brake. One of the others is known as PD-1. The PD stands for programmed death, which I’ve briefly described already. It is a molecule on a T cell that causes the immune system to self-destruct—in effect, to commit suicide.

  The notion seems incredible on its face. But it’s very common. The discovery was made in 1992 in Japan by Dr. Tasuku Honjo at the Kyoto University School of Medicine. He hadn’t been looking for a discovery nearly as profound as the one he made. So profound was the work that Dr. Honjo would share the Nobel with Allison in 2018. Dr. Honjo and his team had been trying to understand what the Cancer Research Institute described as a “normal cellular housekeeping.” The researchers scoured genetic material until they found what looked like a gene involved in prompting some cells to die when they were no longer valuable. This was called programmed death, a kind of suicide by cells no longer useful to the body. Dr. Honjo and his team went deeper into the origins and function of programmed death and found that when they disrupted or knocked out the PD-1 gene in mice, a huge portion of the rodents developed autoimmune disorders approximating lupus.

  In other words, the programmed death gene appeared to be involved in suppressing the immune function.

  Why would it make sense for an immune cell to commit suicide? For the same reason that there are so many brakes in our body’s defense network. It’s one more fail-safe process, another way to keep the most powerful, free-ranging system in our body from going rogue.

  Across the Pacific Ocean from Dr. Honjo’s lab, in Silicon Valley, the initial discovery of programmed death was met with great interest by Nils Lonberg, a scientist and entrepreneur who thought he could use it to cure cancer. He’d been planning for this moment for years, ever since he started milking mice.

  Lonberg, born in 1956 in Berkeley to a chemist father and a psychologist mother, began his own pioneering cancer work indirectly, with a dream of making transgenic mice. This involves genetically engineering mice to harbor human genes. Off topic as this might sound, it happens to be directly in line with the field of immunology, going back even before Jacques Miller was toying with mice to discover the role of the thymus, and without exaggeration, it extends forward to saving Jason’s life.

  By the mid-1980s, technology had moved far beyond Jacques Miller’s shed. By this time, the idea was to use sophisticated genetic techniques to create mice that were mice, mostly, but with key human DNA spliced in. That way, it would be possible to see the effects of a particular molecule or drug on human DNA without killing a human subject.

  But putting human DNA into a mouse isn’t easy, or it wasn’t at the time—“crude and physical,” Lonberg described it to me. He was at Memorial Sloan Kettering in New York. He’d mate two mice around midnight. Then early in the morning, he’d take embryos from the female, inject the human DNA he wanted to embed, and then transfer the embryos into a “pseudo-pregnant mouse,” meaning one that was primed to give birth. “Three weeks later, you’d get pups, little baby mice, with human DNA,” Lonberg said. Then, through subsequent breeding, you’d get a purer form of the DNA-mouse.

  (As an aside, Lonberg was in the lab one night, milking a mouse, when his wife entered. She is a scientist too. “She walked in. I had a mouse hooked up to the vacuum, milking it. She just looked at me,” he said, laughing.)

  Lonberg figured that if you could make a mouse with fully human DNA, could you make a mouse with human antibodies? If so, what would you do with those antibodies? Could you turn a mouse into a factory for building specific molecules of the human immune system?

  If you could, then just maybe you could inject those antibodies into a human to support the person’s immune system without risking the person’s rejection of the molecule as other.

  Lonberg was helping give birth to a new class of drug called monoclonal antibody therapeutics. It is the most important class of drugs of the last twenty years, and at this rate is likely to impact most of our lives before we die. It was life-changing for Jason, Linda, and Merredith, and for many others. Sales of monoclonal antibody drugs hit $87 billion a year by 2015 and are projected to reach $246 billion a year by 2024.

  As a recap, monoclonal antibodies are exact copies of antibodies. Antibodies are essential pieces of the immune system. They sniff out and bind to antigens on other cells, including bad actors. If you know what an antibody does and create lots of copies of it, you can theoretically create a drug that fills a human being with the correct antibody and then prompts a targeted immune response.

  This might sound logical after all you’ve read, but it still is insanely complicated, and requires high levels of both innovation and technology. So perhaps it’s no wonder that Lonberg relocated to Silicon Valley, where the biotechnology business—pairing medicine and high tech—was exploding.

  Lonberg’s contributions wound up being significant because he helped solve the vexing challenge of how to manufacture lots of human antibodies. Lonberg’s solution took years to develop, until the mid-1990s, and entailed creating what he called a “frankenmouse”—part mouse and part human. The part that was human was the immune system. Lonberg and his team could inject said frankenmouse with a particular molecule and prompt the reaction and production of antibodies. In cinematic terms, a molecule was injected into the mouse and began circulating in its Festival of Life. This would prompt the immune system to react. As part of that reaction, the mouse would generate antibodies targeted at the molecule that had been injected. In this way, the mouse was turned into a monoclonal antibody manufacturing plant, a robo-immune system, a prosthetic or synthetic elegant defense, a targeted therapy to do inside a human body what the body seemed unable to do on its own. From this, a drug could be developed built on the extracted monoclonal antibody.

  But there was a twist, one essential to saving Jason. The antibody they ultimately harvested didn’t target the cancer. It targeted the immune system.

  For centuries, the fight against cancer had been built on the idea of attacking the cancer. But Lonberg and the company he worked for (through acquisitions, he by then was employed by Bristol-Myers Squibb) were developing an antibody that did not rely on this core idea, at least not directly. The specific antibody they were developing was aimed at attaching itself—binding—to cells in the immune systems of people like Jason.

  As counterintuitive as it sounds at first, it makes a ton of sense. After all, one of the key reasons that Jason’s cancer was out of control was that his immune system was standing down. It had received a signal to stop from the cancer. The drugmakers wanted to interrupt that signal in a systematic way, block it, by shielding the T cell receptor from receiving the signal to stand down.

  Lonberg offers his own cinematic description of this process. He pictures a T cell, roaming the body, and it has powerful cannons on its surface. The job of this artillery is to take out dangerous organisms. But the surface of the T cell also has many antennae. The antennae receive signals from other parts of the immune system authorizing the T cell to fire or, as often as not, telling it not to fire. The cancer had succeeded in connecting to an important antenna, or maybe several, that had hit pause on the cannon.

  So Lonberg and his cohorts wondered if they could use an antibody to block that antenna from getting a signal.

  Their technique built on the work of others, like the discoveries of Allison and Krummel at Berkeley. Recall that these researchers had discovered that T cells could be sent into attack mode or slowe
d down, depending on what signal they received. The researchers had also found specific places on the T cell that were involved in receiving these communications and specific molecules responsible for sending the communications.

  One way to think about the research is to picture a simplistic version of the immune system’s interaction with a cancer cell.

  After the cell develops, it might well have contact with a dendritic cell. This is a cell in the immune system that carries pieces of a foreign organism back to T cells for examination. The dendritic cell acts as an intermediary between a potential pathogen and a T cell. In many cases of malignancy, the dendritic cell carries back a signal that is interpreted by the T cell as a “go” or “attack” signal. The T cell then proceeds with an attack.

  But some cancers, like Jason’s, wind up getting a signal to the T cell that instructs it to stand down. Plus, it appears that these cancers manage to send such a powerful signal that they overwhelm the communications system; the T cell isn’t really able to pick up any “go” signal.

  Lonberg, among others, wondered if it was possible to displace the “stop” signal by in effect sending a louder one—swamping the T cell’s “go” antennae such that it received the signal to attack. With help from the mouse, they would send molecules in to take back the T cell’s antennae, prevent it from being hogged by the cancer’s insidious signal, and allow it to proceed.

  (For those interested in the details, Lonberg and his peers, in the late 1990s, were figuring out how to cause the T cell to receive its signal at CD28, which is the spot where the “go” signal is received, and not at CTLA-4, where the “stop” signal arrives. Both receive their signal from the molecule B7-1; if B7-1 binds to CTLA-4, the immune system stops, and if it binds to CD28, the attack goes forward. In some cancers, “CTLA-4 is hogging B7,” Lonberg said. So the goal was to “displace” B7-1 from the CTLA-4 so that CD28 could bind. They did this by creating an ultra-specific antibody to bind to CTLA-4. When the antibody bound to CTLA-4, it pried loose the B7-1. Now the brakes would have been turned off. The immune system could attack the tumor as if it were foreign and dangerous, not innocuous and self.)

  If it worked, this would unleash the immune system to operate the way it was intended to work. The theory is a marvel. Within days or weeks, the body’s own defenses could destroy a tumor that toxic chemotherapy couldn’t kill over months or years. The locks would be taken off the T cell guns, the cannons unshackled, cancer’s magic trick exposed.

  The clinical tests that took place in 2007 were reported in a New England Journal of Medicine article published in September 2010. The drug was given to 676 patients with stage 3 or 4 metastatic melanoma, a cancer that is more or less fatal. The drug had extended the average life expectancy to 10 months, up from 6.4 months. That may not sound like much. But it’s 40 percent more life!

  There was a catch.

  The study published in the New England Journal of Medicine in 2010 alluded to side effects that showed up in 10 to 15 percent of patients. Serious, serious side effects. Seven patients died, several “associated with immune-related adverse effects.”

  The drug, called Yervoy (ipilimumab), took the brakes off the T cell. But remember, there are lots of good reasons those brakes are there. Now, with the immune system unleashed, it could go off half-cocked and attack much more than the cancer.

  It wasn’t the first time that researchers had monkeyed with the immune system and paid the ultimate price.

  In the spring of 2006, at a hospital in London, a handful of patients “took part in a study that has sent shock waves through the research world,” noted an article in the New York Times by my then-colleague Elisabeth Rosenthal. The phase I clinical trial was for a monoclonal antibody that also worked on CD28. The goal of a phase I trial is to test for safety. So the volunteers at Northwick Park Hospital were all healthy, but also were selected because they had CD28 receptors similar to those of people with rheumatoid arthritis and B cell cancer.

  Let me pause a moment to underscore the fact that these drugs are designed to work on two classes of diseases that, on their faces, have no relation—cancer and autoimmunity. Of course now it’s clear that they are highly related. One dupes and slows or stymies the immune system, as in Jason. The other overheats the immune system, as in Linda or Merredith. Could the same kind of drug target the immune cells to restore balance?

  Not in the case of TeGenero. That was the name of a drug involved in an infamous clinical trial.

  Six healthy individuals entered the phase I clinical trial. They were given an infusion of the drug at a tiny dose—500 times smaller than was shown safe in animals.

  “Minutes after the first infusion,” one case study reads, “all patients started suffering from severe adverse reaction resulting from rapid release of cytokines by activated T cells.”

  Time to define a rightfully scary term: cytokine storm.

  Remember cytokines? They are proteins that send signals to the immune system, creating a powerful, virtually instantaneous telecommunications network, the envy of even the fastest Internet provider or connection. The commands they send can prompt a range of responses, including cell growth and inflammation. They call on the interferons, which are central to the innate immune system, and the interleukins, which have an even broader charge, and the chemokines, which can recruit macrophages and neutrophils. A cytokine storm occurs when the network begins sending a flurry of messages, an out of control torrent of signals. The term cytokine storm actually understates how dangerous it is. Cytokine typhoon or hurricane might be more accurate. It is deadly.

  Within eight hours, all six patients involved in the TeGenero clinical trial were in the intensive care unit.

  Five of them died.

  It was, to say the least, a bridge too far, a poorly constructed trial that showed how close the guardrails are when innovation gives momentum to our elegant defenses. You mess with the immune system at your own risk.

  By the time Jason got sick, the drug developers had made major strides.

  41

  The Breakthrough

  Years ago, when the New York Times first started putting color pictures in its paper, I joked that people should not worry: The writing in the Gray Lady, I told them, would remain dry and lifeless.

  I meant it with love, of course. There’s a time and place for hyperbole and exciting bold adjectives, and newspaper writing about serious subjects is not one of them. So perhaps it’s understandable the way that the Times described with due caution what might, eventually, be seen as cancer’s version of the Apollo missions. One took place on March 25, 2011. That day, the Food and Drug Administration approved for use in people with melanoma, that deadly skin cancer, the drug called Yervoy I mentioned a few pages earlier, made by Bristol-Myers.

  An article ran in the Times in the business section, written by an encyclopedic colleague of mine, since retired, named Andrew Pollack. The story explained that Yervoy had been approved for use in metastatic melanoma, a major breakthrough. The article explained that 20 percent of people in the trial who took the drug lived two years or more. Yes, there were side effects, but not treating metastatic melanoma came with its own likely terminal side effect.

  So for people dying of melanoma, Andrew’s article might just as well have read: WE CAN BRING YOU BACK FROM THE DEAD!

  Looking back, too, there’s just no way to downplay the wording Andrew used to describe Yervoy: “a novel type of cancer drug that works by unleashing the body’s own immune system to fight tumors.”

  This is where all the scientific study had been leading, from Metchnikoff and Ehrlich to Jacques Miller and Max Cooper, Peter Doherty, Tonegawa, and on and on. One discovery on top of another, one technique following the next, one painstaking failure leading to tiny breakthroughs, all on the backs of patients who willingly took their chances, let themselves be transplanted (begged for the treatment!) or tested with new medications, so that the immune system might not just be understood, but “unleash
ed.”

  Science and market forces had collaborated to bring a seeming miracle cure to market. Just in time for Jason.

  42

  Jason Races Time

  After Jason’s first round of chemo failed to work, his care was moved to Colorado Blood Cancer Institute, and came to be overseen by Dr. Brunvand, Jason’s oncologist. The second level of treatment is called salvage. It’s more toxic than the first kind of chemo. Jason responded. But there is another step in this second stage, and it is brutal.

  What follows is a bone marrow transplant known as an autologous hematopoietic stem cell transplant. This transplant replaces stem cells in the patient’s bone marrow that have been damaged by the chemotherapy. In a very real sense, it involves pulling out the patient’s immune system and then restarting it.

  That’s not the horrible part. What makes this process so devastating is an interim step known as BEAM. This is another level of chemotherapy—a terrible, evil nuclear-winter-level therapy—that is used to wipe out the last of the cancer cells that are left behind by salvage. Typically, the salvage therapy leaves behind about a million such cancer cells. BEAM is toxic enough to get these remaining dogged cancer cells, but it is also so toxic that it wipes out the patient’s own stem cells.

  “All his stem cells are sacrificed on the altar of killing the last cancer cells,” Dr. Brunvand explained.

  Dr. Anthony Fauci, head of the National Institute of Allergy and Infectious Diseases at the NIH, and Dr. Mark Brunvand, Jason Greenstein’s oncologist. (Courtesy of the author)

  BEAM, coupled with the emotional challenge of transplantation itself, is so intense that the procedures don’t go forward until a patient is assessed on three levels: Is he responding to the chemo, and is he both physically and emotionally able to survive the experience?

  It was time for Jason to be assessed by a psychologist.