Designer antibodies fight cancer by tethering immune cells to tumor cells

Amy Boland has fought lymphoma for more than a decade, but her cancer vanished after treatment with bispecific antibodies.

Amy Boland has gone through many ups and downs since she noticed lumps under her arms 12 years ago and learned she had cancer of the lymph system. For about 6 years, conventional chemotherapy helped shrink her lymphoma tumors, but they started to grow again. A succession of other cancer therapies, including a bone marrow transplant and a class of drugs called checkpoint inhibitors, either failed or only brought temporary relief. In one elaborate effort, physicians harvested her T cells, engineered those immune cells to kill her lymphoma, and infused them back into her body. The cancer vanished, but 2 years later bounced back. “Nothing was really working,” says oncologist Stephen Schuster of the University of Pennsylvania (UPenn).

So in October 2018, Boland joined a clinical trial testing another way to harness her immune system to kill the tumor cells. The idea of the trial, which Schuster co-leads, was to use a molecular rope known as a bispecific antibody to tether her natural T cells to the tumor cells so the immune warriors would attack. Like the engineered T cells she had received earlier, the experimental infusions sometimes made her sick enough to spend a couple of nights in the hospital. But the antibody rapidly sent her into remission. Today, more than 1 year after going off the Roche drug, mosunetuzumab, Boland, now 60, appears to be free of cancer and leads a normal life. “I’m feeling really good. I’m so grateful,” she says.

The ongoing trial Boland participated in made headlines in December 2019. At a meeting of the American Society of Hematology, Schuster reported that the antibody shrank fast-growing non-Hodgkin lymphoma tumors in 46 of 124 patients in whom other treatments had failed. For some of those people, like Boland, the failures included having their immune cells altered to attack the cancer. Those engineered cells, known as chimeric antigen receptor (CAR) T cells, have achieved remarkable results in some cancers. Yet at the same meeting, data from a small clinical trial suggested a bispecific antibody might work equally well on myeloma, another blood cancer. Bispecific antibodies for cancer are “superhot,” says Janice Reichert, executive director of the Antibody Society, who has tracked their development.

That’s the culmination of a slow boil. Scientists have worked on bispecific cancer drugs for decades and the first clinical success came 12 years ago. That result jump-started the field for a while, but other therapies, including CAR T cells, raced forward, in part because the cancer-fighting antibodies proved challenging to design and produce. But companies have now made those protein drugs safer and more potent. Variants are being tested in dozens of clinical trials, in the hope they can rival or surpass the engineered cells.

“If bispecific antibodies can do what CAR T cells can do, this would represent a big advance” and a “potentially fundamental change,” said Johns Hopkins Medicine hematologist Robert Brodsky at a press preview for the meeting where Schuster presented. A major advantage of bispecific antibodies is that they can be mass-produced in advance. CAR T cells, by contrast, must be prepared for each cancer patient. That process is costly and, for some very sick patients, takes too long.

Those new players in immunotherapy are no panacea yet. For some blood cancers, the bispecific antibodies aren’t giving patients the long-lasting remissions often seen with CAR T cells. As happened with CAR T cells, several patients have died in trials testing bispecific antibodies, possibly from overzealous immune responses sparked by the drugs. And bispecific antibodies may prove less effective against solid tumors, such as those of the colon and lungs, than against blood and lymph cancers—a drawback shared with CAR T cells. “There are a lot of open questions. But it’s also a field moving quickly, and a lot of really smart people are working on it,” says Paul Carter, an antibody researcher at Genentech, a Roche subsidiary.

Antibodies have a long history as a cancer treatment. The Y-shaped proteins are normally pathogen fighters, latching onto an antigen—a protein or a bit of one—on viruses, bacteria, or other microbes. The binding, which takes place at the tips of the Y, can directly disable and clear a pathogen or can signal the immune system to attack it. Cancer researchers first learned to exploit that natural system by making many copies of a specific antibody that latches onto an antigen unique to a particular cancer. This marks the cancer for destruction by components of the immune system other than T cells. Some of the most effective, and bestselling, cancer drugs are such monoclonal antibodies, including the breast cancer drug trastuzumab, better known as Herceptin.

Newer anticancer strategies enlist T cells. Tumor cells can appear foreign enough for the body to sometimes “train” these cells to attack them. But CAR T cells, altered to carry a receptor that targets a cancer cell antigen, can deliver a more powerful response. Checkpoint inhibitors, drugs that release the molecular brakes that can restrain T cells, can also boost the T cell attack.

Bispecific antibodies offer a third way to harness T cells. In the mid-1980s, cancer researchers began to engineer antibodies that had two tips—one matched to a cancer cell antigen and the other to a T cell surface protein called CD3. The idea was to directly link T cells to tumor cells, thereby skipping the need for T cells to learn to attack a cancer. “It’s mimicking what naturally happens, but the advantage is that you can engage all T cells,” not just those trained to attack the tumor, says Dirk Nagorsen, a vice president and cancer researcher at Amgen. In 1985, the field was galvanized by two reports in Nature that such a “bispecific” antibody could destroy cancer cells in a dish; studies soon showed those antibodies could shrink tumors in mice.

The drugs were hard to make. Antibodies are modular, with two identical “heavy” chains, making up the stem and half of each arm of the Y, and two identical “light” chains, each of which completes one arm. Trying to assemble bispecific antibodies from those complex components, protein chemists got 10 versions of each molecule. That outcome meant laborious efforts to sift out the one researchers wanted.

And the excitement faded when tests of bispecific antibodies moved from lab animals and cells to cancer patients. In an early clinical trial, one antibody appeared to shrink lymphoma. But researchers had to stop treating patients before the study yielded definitive results because the antibody’s maker ran out of the drug. The antibodies also sometimes triggered serious side effects, including liver damage and an immune overreaction in which white blood cells pump out signals called cytokines that can be toxic in large quantities. Such cytokine “storms” cause fevers and, in severe cases, organ damage. (CAR T cells can cause the same overreaction.)

Two immunologists at the Ludwig Maximilian University of Munich, Peter Kufer and Gert Riethmüller, pressed ahead anyway with an idea some colleagues thought would fail: a stripped-down bispecific antibody with two tips linked by a flexible peptide instead of the traditional stem. The simplified design made the antibody easier to manufacture, but because the stem was missing, the kidneys cleared it from the blood within 2 hours. In its first clinical test, for non-Hodgkin lymphoma, patients had to wear a pump to continually infuse the antibody. Still, tiny doses of the drug shrank tumors in all seven lymphoma patients in the trial, run by Micromet, a German biotech Riethmüller co-founded. “We thought, ‘Oh my God, there’s something amazing happening here,’” says molecular biologist Patrick Baeuerle, then–chief scientific officer of Micromet, who dubbed the concept a bispecific T cell engager (trademarked as BiTE).

The small trial, published in Science in 2008, sparked interest from companies and academics. “The whole field realized, ‘This is a big deal. We want in on this,’” says John Desjarlais, chief scientific officer of the biotech Xencor. At about the same time, CAR T cells began to show impressive results in some leukemia patients—which boosted interest in other, potentially simpler immunotherapies such as bispecific antibodies. Like CAR T cells, T cells stimulated by BiTEs release toxic molecules called granzymes and perforins that punch holes in tumor cells and cause them to self-destruct. “I see bispecifics [such as BiTEs] as an off-the-shelf CAR T cell,” says Elad Sharon, senior investigator with the National Cancer Institute’s Cancer Therapy Evaluation Program.

The first bispecific antibody for cancer was approved in Europe in 2009. It was meant to mop up the malignant cells that cause abdominal fluid to build up in some cancer patients—but it didn’t work that well, so the drug only stayed on the market a few years. The field regained momentum, however, after Amgen snapped up Micromet in 2012 and later showed that its BiTE drug, blinatumomab (Blincyto), doubled the survival time of patients with advanced acute lymphocytic leukemia. Beginning in 2014, the Food and Drug Administration approved the drug to treat several adult and pediatric forms of the disease. Amgen is now testing BiTEs for other cancers, including myeloma and lung, prostate, and brain cancers.

Others have rushed to improve on BiTEs by using protein engineering tricks to create desired bispecific antibodies. Some firms have restored the antibodies’ stem, known as the Fc receptor, so that the protein stays in the blood longer—but with modifications that make it less toxic to the liver. Cancer patients such as Boland no longer have to wear a fanny pack containing a pump, but can now receive the drug as an intravenous drip every 3 weeks.

Industry scientists have also added a second copy of the tumor antigen–binding site to one tip of the antibody (see graphic, above). Known as a “2+1” bispecific antibody, this particular design is meant to make the antibody more selective for cancer cells and less likely to target healthy cells carrying small amounts of the cancer antigen.

To reduce the risk of triggering a cytokine storm, researchers are also designing bispecifics that, instead of T cells, snag a different kind of immune cells called natural killers (NK). Several companies have started or are readying clinical trials of such antibodies, which bind to an NK cell surface protein called CD16. “NK cells are very potent tumor cell killers if activated, and the cytokine release is significantly reduced with this cell type,” says Dmitri Wiederschain, head of cancer immunology at one such company, Sanofi.

These variants and scores of other bispecifics are easier to synthesize now. “Antibody engineering has become so sophisticated, it’s possible to make these molecules quite efficiently,” says biochemist Christoph Rader of Scripps Research in Jupiter, Florida.

By Reichert’s count, more than 60 T cell–directing bispecific antibodies for cancer are in early- or later-stage clinical trials. One Amgen BiTE has shown hints of shrinking tumors in a few patients with advanced prostate cancer, the company reported last year.

Amy Boland’s cancer scans showed a dramatic disappearance of her lymphoma tumors from the start of her bispecific antibody treatment (left) to 12 weeks later.

Solid tumors are a challenging target for bispecifics in part because tumors often lack a unique antigen for the antibodies to grab. Many tumors are also surrounded by blood vessels, tissue, and immune cells that form a barrier T cells can’t easily penetrate. But findings from mouse studies suggest some bispecific antibodies can drive T cells into tumors, says Nai-Kong Cheung of Memorial Sloan Kettering Cancer Center. His lab has systematically tweaked design factors, such as how binding sites are arranged, to learn what optimizes the molecules’ potency.

And some companies hope to boost the attack on solid tumors with antibodies that bind not only to CD3, but also to another receptor on T cells known as a “second signal,” which stimulates the cells to grow. For years, says Regeneron Senior Vice President Israel Lowy, industry has been “afraid to touch” that protein, called CD28, because of a devastating mishap: An antibody designed to bind to it made six healthy volunteers critically ill from cytokine release syndrome in a 2006 U.K. clinical trial.

Findings from new studies, however, suggest it’s possible to exploit that cell growth trigger safely. Last year in Nature Cancer, a Sanofi team reported that a “trispecific” antibody with arms matched to CD28, CD3, and a cancer antigen wiped out myeloma tumors in mice. Other firms have split up the task by creating two bispecifics. One targets a tumor antigen and CD28 or another growth-signal receptor; the other binds to the tumor antigen plus CD3. “One of our hopes is that this costimulatory bispecific may help us unlock responses in solid tumors,” says Lowy, whose company reported in Science Translational Medicine in January that such a two-drug combination shrank ovarian tumors and slowed prostate tumor growth in mice.

Could those next-generation bispecifics eclipse CAR T cells for some cancers? UPenn’s Carl June, a CAR T cell pioneer, is skeptical. Many leukemia patients who get blinatumomab eventually relapse because cancer cells become resistant to the drug, he notes, so oncologists use it mainly as a “bridge” until a very sick patient can get a stem cell transplant or CAR T cells. June adds that bispecifics may not work in the many cancer patients whose T cells have become depleted or “exhausted,” leaving too few to attack the cancer. CAR T cell treatment, by contrast, replenishes the immune ranks by growing the cells outside the body—something “not possible with bispecifics,” June says.

Schuster, who has a foot in both camps—he runs Boland’s study and has led CAR T trials with June—says bispecifics are still proving themselves. He points to an Amgen report last year that some lymphoma patients who responded well to blinatumomab were still alive after 7 years, suggesting they could stay in remission long term. “I am confident that within the next 2 to 5 years you’re going to see quantum leaps in our ability to target resistant tumors, including solid tumors,” Schuster says.

He and other cancer researchers see CAR T cells, checkpoint inhibitors, and bispecific antibodies as interchangeable. “Why not do all of the above?” asks Schuster, who’s preparing a trial that will give patients CAR T cells and then a bispecific drug. “All these approaches to manipulate the cellular immune system to treat cancer are essentially different means to the same end.” Doubling up the treatments can be risky, however—last year, two patients died in such a trial after developing cytokine release syndrome.

Boland, who has seen all three of her children grow up and go off to college during her cancer fight, welcomes the progress. A bispecific drug is keeping her lymphoma in check for now, she notes. “I hope it’ll last, but if not, I feel confident that there’s always more treatments. You just don’t worry about it. Nobody knows what will happen.”

Jocelyn is a staff writer for Science magazine.