Technology vs cancer

It’s war. But we’ve got some big guns

What is cancer, though, really?

WHEN YOU VISIT St Jude Children’s Research Hospital in Memphis, Tennessee, you expect to feel devastated. It starts in the waiting room. Oh, here we go with the little red wagons, you think, observing the cattle herd of them rounded up by the entrance to the Patient Care Centre. Oh, here we go with the crayon drawings of needles. The itch begins at the back of your throat, and you start blinking very fast and mentally researching how much money you could donate without starving. Near a row of arcade games, a preteen curls his face into his mother’s shoulder while she strokes his head. Oh, here we go.

But the more time you spend at St Jude, the more that feeling is replaced with wonder. In a cruel world you’ve found a free hospital for children, started by a Hollywood entertainer as a shrine to the patron saint of lost causes. There is no other place like this. Corporations that have nothing to do with cancer – nothing to do with medicine, even – have donated vast sums of money just to be a part of it. There’s a Chili’s Care Centre. The cafeteria is named for Kay Jewellers.

Scott Newman’s office is in the Brooks Brothers Computational Biology Centre, where a team of researchers is applying computer science and mathematics to the question of why cancer happens to children. Like many computer people, Newman is very smart and a little quiet and doesn’t always exactly meet your eyes when he speaks to you. He works on St Jude’s Genomes for Kids project, which invites newly diagnosed patients to have both their healthy and tumour cells genetically sequenced so researchers can poke around.

“Have you seen a circle plot before?” Newman asks, pulling out a diagram of the genes in a child’s cancer. “If I got a tattoo, it would be one of these.” Around the outside of the circle plot is something that looks like a colourful bar code. Inside, a series of city skylines. Through the centre are coloured arcs like those nail-and-string art projects students make in high school geometry class. The diagram represents everything that has gone wrong within a child’s cells to cause cancer. It’s beautiful.

“These are the genes in this particular tumour that have been hit,” Newman says in a Yorkshire accent that emphasises the t at the end of the word hit in a quietly violent way. “And that’s just one type of thing that’s going on. Chromosomes get gained or lost in cancer.

This one has gained that one, that one, that one, that one,” he taps the page over and over. “And then there are structural rearrangements where little bits of genome get switched around.” He points to the arcs sweeping across the page. “There are no clearly defined rules.”

It’s not like you don’t have cancer and then one day you just do. Cancer – or, really, cancers, because cancer is not a single disease – happens when glitches in genes cause cells to grow out of control until they overtake the body, like a kudzu plant. Genes develop glitches all the time: there are roughly 20 000 genes in the human body, any of which can get misspelt or chopped up. Bits can be inserted or deleted. Whole copies of genes can appear and disappear, or combine to form mutants. The circle plot Newman has shown me is not even the worst the body can do. He whips out another one, a snarl of lines and blocks and colours. This one would not make a good tattoo.

“As a tumour becomes cancerous and grows, it can accumulate many thousands of genetic mutations. When we do whole genome sequencing, we see all of them,” Newman says. To whittle down the complexity, he applies algorithms that pop out gene mutations most likely to be cancerrelated, based on a database of all the mutations researchers have already found. Then, a genome analyst manually determines whether each specific change the algorithm found seems likely to cause problems. Finally, the department brings its list of potentially important changes to a committee of St Jude’s top scientists to discuss and assign a triage score. The mutations that seem most likely to be important get investigated first.

It took 13 years and cost R36 billion to sequence the first genome, which was completed in 2003. Today, it costs R13 000 and takes less than a week. Over the last two decades, as researchers like Newman have uncovered more and more of the individual genetic malfunctions that cause cancer, teams of researchers have begun to tinker with those mutations, trying to reverse the chaos they cause. (The first big success in precision medicine was Gleevec, a drug that treats leukaemias that are positive for a common structural rearrangement called the Philadelphia chromosome. Its launch in 2001 was revolutionary.) Today, there are 11 genes that can be targeted with hyperspecific cancer therapies, and at least 30 more being studied. At Memorial Sloan Kettering Cancer Centre in New York City, 30 to 40 per cent of incoming patients now qualify for precision medicine studies.

Charles Mullighan, a tall, serious Australian who also works at St Jude, is perhaps the ideal person to illustrate how difficult it will be to cure cancer using precision medicine. After patients’ cancer cells are sequenced and the wonky mutations identified, Mullighan’s lab replicates those mutations in mice, then calls St Jude’s chemical library to track down molecules – some of them approved medicines from all over the world, others compounds that can illuminate the biology of tumours – to see if any might help.

If Mullighan is lucky, one of the compounds he finds will benefit the mice, and he’ll have the opportunity to test it in humans. Then he’ll hope there are no unexpected side effects, and that the cancer won’t develop resistance, which it often does when you futz with genetics. There are about 20 subtypes of the leukaemia Mullighan studies, and that leukaemia is one of a hundred different subtypes of cancer. This is the kind of precision required in precision cancer treatment; even if Mullighan succeeds in identifying a treatment that works as well as Gleevec, with the help of an entire, wellfunded hospital, it still will work for only a tiny proportion of patients.

Cancer is not an ordinary disease. Cancer is the disease, a phenomenon that contains the whole of genetics and biology and human life in a single cell. It will take an army of researchers to defeat it.

Luckily, we’ve got one.

INTERLUDE

“I used to do this job out in LA,” says the attendant at the Hertz counter at Houston’s George Bush Intercontinental Airport. “There, everyone is going on vacation. They’re going to the beach or Disneyland or Hollywood or wherever.

“Because of MD Anderson, I see more cancer patients here. They’re so skinny. When they come through this counter, they’re leaning on someone’s arm. They can’t drive themselves. You think, there is no way this person will survive. And then they’re back in three weeks, and in six months, and a year. I’m sure I miss some, who don’t come through any more because they’ve died. But the rest? They come back.”

“I gotta tidy that up sometime,” Allison says.

Allison has just returned to the office from back surgery that fused his L3, L4, and L5 vertebrae, which has slightly diminished his Texas rambunctiousness. Even on painkillers, though, he can explain the work that many of his contemporaries believe will earn him the Nobel Prize: he figured out how to turn the immune system against tumours.

Allison is a basic scientist. He has a PHD, rather than an MD, and works primarily with cells and molecules rather than patients. When T-cells, the most power- ful “killer cells” in the immune system, became better understood in the late 1960s, Allison became fascinated with them. He wanted to know how it was possible that a cell roaming around your body knew to kill infected cells but not healthy ones. In the mid-1990s, both Allison’s lab and the lab of Jeffrey Bluestone at the University of Chicago noticed that a molecule called CTLA-4 acted as a brake on T-cells, preventing them from wildly attacking the body’s own cells, as they do in autoimmune diseases.

Allison’s mother died of lymphoma when he was a child and he has since lost two uncles and a brother to the disease. “Every time I found something new about how the immune system works, I would think, I wonder how this works on cancer?” he says. When the scientific world discovered that CTLA-4 was a brake, Allison alone wondered if it might be important in cancer treatment. He launched an experiment to see if blocking CTLA-4 would allow the immune system to attack cancer tumours in mice. Not only did the mice’s tumours disappear, the mice were thereafter immune to cancer of the same type.

Ipilimumab (“ipi” for short) was the name a small drug company called Medarex gave the compound it created to shut off CTLA-4 in humans. Early trials of the drug, designed just to show whether ipi was safe, succeeded so wildly that Bristol Myers Squibb bought Medarex for R32 billion. Ipilimumab (now marketed as Yervoy) became the first “checkpoint inhibitor”: it blocks one of the brakes, or checkpoints, the immune system has in place to prevent it from attacking healthy cells. Without the brakes the immune system can suddenly, incredibly, recognise cancer as the enemy.

“You see the picture of that woman over there?” Allison points over at his desk. Past his lumbar-support chair, the desk is covered in papers and awards and knickknacks and frames, including one containing a black card with the words “Never never never give up” printed on it. Finally, the photo reveals itself, on a little piece of blue card stock.

“That’s the first patient I met,” Allison says. “She was about 24 years old. She had metastatic melanoma. It was in her brain, her lungs, her liver. She had failed everything. She had just graduated from college and got married. They gave her a month.”

The woman, Sharon Belvin, enrolled in a phase-two trial of ipilimumab at Memorial Sloan Kettering, where Allison worked at the time. Today, Belvin is 35, cancerfree and the mother of two children. When Allison won the Lasker prize, in 2015, the committee flew Belvin to New York City with her husband and her parents to see him receive it. “She picked me up and started squeezing me,” Allison says. “I walked back to my lab and thought, Wow, I cure mice of tumours and all they do is bite me.” He adds, dryly, “Of course, we gave them the tumours in the first place.”

After ipi, Allison could have taken a break and waited for his Nobel, driving his Porsche Boxster with the licence plate CTLA-4 around Houston and playing the occasional harmonica gig. (Allison, who grew up in rural Texas, has played since he was a teenager and once performed “Blue Eyes Crying in the Rain” onstage with Willie Nelson.) Instead, his focus has become one of two serious problems with immunotherapy: it works for only some people.

So far, the beneficiaries of immune checkpoint therapy appear to be those with cancer that develops after repeated genetic mutations: metastatic melanoma, non-small- cell lung cancer, and bladder cancer, for example. These are cancers that often result from bad habits such as smoking and sun exposure. But even within these types of cancer, immune checkpoint therapies improve long-term survival in only about 20 to 25 per cent of patients. In the rest the treatment fails, and researchers have no idea why.

Lately, Allison considers immune checkpoint therapy a “platform” – a menu of treatments that can be amended and combined to increase the percentage of people for whom it works. A newer drug called Keytruda that acts on a different immune checkpoint, PD-1, knocked former US president Jimmy Carter’s metastatic melanoma into remission in 2015. Recent trials that blocked both PD-1 and CTLA-4 in combination improved longterm survival in 60 per cent of melanoma patients. Now, doctors are combining checkpoint therapies with precision cancer drugs, or with radiation, or with chemotherapy. Allison refers to this as “one from column A, and one from column B”.

The thing about checkpoint inhibitor therapy that is so exciting – despite the circumscribed group of patients for whom it works, and despite sometimes mortal side effects from the immune system going buck- wild once the brakes come off – is the length of time it can potentially give people. Before therapies that exploited the immune system, response rates were measured in a few extra months of life. Checkpoint inhibitor therapy helps extremely sick people live for years. So what

if it doesn’t work for everyone? Every cancer patient you can add to the success pile is essentially cured.

Jennifer Wargo is another researcher at MD Anderson who is trying to predict who will respond to checkpoint inhibitor therapy and who will not. Originally a nurse, Wargo got so interested in biology that she went back to school for a bachelor’s degree, then a medical degree, and then a surgical residency at Harvard. It was during her first faculty position, also at Harvard, around 2008, that she started to wonder how the microbiome – the bacteria that live in the human body, of which there are roughly 40 trillion in the average 70-kilogram man – might affect cancer treatment. Wargo was investigating the bacteria that lived near the site of pancreatic cancer, in and around the tumour. Could you target those bacteria with drugs and make the cancer recede more quickly?

In the early 2010s, research about the microbiome in the human gut – the bacteria in humans’ stomachs and intestines that appear to affect immune function, gene expression and mood, among other things – gained traction in journals. Before long, two separate researchers had shown that you could change a mouse’s response to immune checkpoint inhibitor therapy by giving him certain kinds of bacteria. Wargo added the microbiome to her slate of experiments. Along with her team, she collected gut microbiome samples from more than 300 cancer patients who then went on to receive checkpoint inhibitors as treatment. The results were, Wargo says, “night and day”. People who had a higher diversity of gut bacteria had a stronger response to checkpoint inhibitor therapy.

Now, Wargo is transplanting stool samples from patients into germ-free mice with melanoma, to see if she can predict whether the mice will mimic the treatment responses of the people whose bacteria they received. “Can we change the gut microbiome to enhance responses to therapy… or even prevent cancer altogether?” she says. “Ah god, that would be the holy grail, wouldn’t it?” she whispers, as if not to invite bad luck. “It’s gonna take a lot of work to get there, but I think the answer is gonna be yes.”

Immunotherapies do have one other problem worth worrying about, one that underlies the most frustrating experience of having cancer. When a patient is diagnosed, the first therapy is still one of the standards: surgery, radiation or chemo- therapy. Cut, burn, or poison, as the doctors say. Doctors can’t use promising immunotherapies as first-line treatments yet because immunotherapies are still dangerous. No one knows what will happen long-term if you shut off the immune system’s brakes. Does a patient survive cancer just to develop another terrible disease, like amyotrophic lateral sclerosis (ALS), in 15 years?

INTERLUDE

“Just to play devil’s advocate,” says a woman at a margarita bar and restaurant in Santa Fe, New Mexico. “Don’t you think the cure exists somewhere already and the medical industrial complex is hiding it? People stand to lose billions of dollars. Don’t you think they want to keep that money?”

I have been talking to this woman for 20 minutes. She is familiar with cancer. She works with natural cures, is a big fan of neuroscience, and knows some of the prominent names in medical research. I tell her that the conspiracy theory she is referencing – that governments or the pharmaceutical industry is hiding the cure for cancer – can’t be true. Of course, it’s hard to believe that US president Richard Nixon initiated the war on cancer in 1971 and the disease still kills 595 690 people a year. And that the most brilliant minds of our time have turned HIV into a chronic disease, but cancer continues relatively unchecked. Yet I’ve talked to 35 researchers and policymakers and visited seven cancer centres and I haven’t seen a shred of evidence that doctors who treat very sick people – and whose job it is, sometimes, to tell people that they will die – aren’t trying with their very souls to succeed at their jobs. “It’s just that it’s hard,” I say. The woman huffs. Someone more interesting is sitting on the other side of her. And that’s the end of that.

to the story of a girl named Lisa, who is pictured in a photo not far from the bear. Lisa had the same illness as Sam around the same time, but her therapy did work. Lisa’s story lasts more than a minute, with Mackall practically cheering at the end. “So she remained fertile and that’s her little boy!” she yells, gesturing towards Lisa’s photo. Mackall smiles the pained, confused smile of someone who has inexplicably survived a car crash. “You have your ups and your downs,” she says.

Overall, children’s cancer has been one of the great success stories in cancer treatment. In the 1970s, dramatic advances in chemotherapy put most patients with certain types of leukaemia (particularly acute lymphoblastic leukaemia in B-cells, otherwise known as B-ALL) into remission. Today, 84 per cent of children who get ALL

can be cured. But then treatment stalled. “We have made steady progress, by all accounts,” says Mackall. “But it’s been largely incremental. And there’ve been these plateaus that have just driven us crazy.”

In those unfortunate few children who relapsed or didn’t respond to the chemo, or who got a different variety of cancer, like Ewing’s sarcoma, there were few treatments left to try. Mackall’s patients came to her after having had surgery and then chemotherapy once, twice, three times. “You can just see, they’re beat up. They’re making it, but all they do is get their treatments,” she says. “They didn’t have enough energy to do anything else.” Then, if they lasted long enough, they got into a trial.

There are several ways to turn the immune system against cancer. Checkpoint inhibitor therapy is one of them. But it doesn’t work in all patients, especially children, whose cancers generally do not have the vast numbers of mutations needed to attract the attention of a newly brakefree immune system. For a long, dark time, immunotherapists would try other sorts of techniques to get the immune system to respond in these patients, and the patients would die anyway, like Michael did. The treatments were toxic or they damaged the brain or they just didn’t work. The doctors would recommend hospice. Hospice. Hospice.

And then all the research began to pay off. In August 2010, a retired correctional officer named Bill Ludwig walked into the Hospital of the University of Pennsylvania to try a new therapy developed by a researcher named Carl June. Ludwig had chronic lymphocytic leukaemia (CLL), another cancer that affects B- cells. Multiple rounds of chemotherapy had failed to cure it, and he didn’t qualify for a bone marrow transplant. June’s idea, which was so risky that the National Institutes of Health had turned down several grant applications to fund it, was the only option Ludwig had. June had just enough money to try it in three patients. Ludwig went first.

To understand how June’s therapy works, consider the T-cells that Jim Allison found fascinating. They’re cells that kill other cells, but they don’t kill you because they have a built-in targeting mechanism. Each person has millions of T-cells, and each one of those T-cells matches a single virus, like a lock and a key. If a virus enters the body, its own personal T- cell key will find and destroy it, then copy and copy and copy itself until the virus succumbs. “I liken it to a bloodhound,” says Mackall. “What the marker says to the T- cell is: anything that has this thing on it, kill it.”

Previously, researchers had created a fake key called a chimeric antigen receptor,

or CAR, that matched a particular lock, CD19, on B-cells, which is where Ludwig’s leukaemia was. During the trial, Ludwig’s doctors removed as many of Ludwig’s T- cells as they could, and June’s team inserted the CAR using a modified form of HIV, which can edit genes. Then they returned the T- cells to Ludwig.

Ten days later, Ludwig started to have chills and fever, like he had the flu. He was so ill that doctors moved him to the intensive care unit. But then, less than a month later, he was in remission. The T-cells had located and demolished the cancer, the same way they would a virus.

When case studies of the first three patients were published in scientific journals, mainstream media went crazy: “Cancer treated with HIV!” they shouted. But it was a later study that showed that the furore was warranted: when the Penn team partnered with the Children’s Hospital of Pennsylvania to try CAR-T cell therapy against B-ALL in children, the cancer disappeared in 24 out of 27 patients.

Novartis was the drug company that partnered with the University of Pennsylvania to turn June’s treatment into a drug for the general public, and the company submitted results of all three required levels of tests to the US Food and Drug Administration early this year. If the FDA approves the drug, any child who has B-ALL and has failed her first therapy can have her white blood cells removed, frozen, and shipped to Novartis’s processing facility in Morris Plains, New Jersey, where molecular engineers will insert the new “key” and send the T-cells back. The patient gets a one-time infusion, and there is an 83 per cent chance she will be cured.

“We also do a second measure of remissions where we look to see if there’s any measurable disease at all,” says David Lebwohl, Novartis’s global programme head for CAR-T treatments. “A more sensitive test than just looking in the blood. And that was also negative for 83 per cent of the patients.”

An 83 per cent cure rate in children who would otherwise die is a monumental achievement. If there is a moment where a culture hits on an idea that can cure a disease – vaccines, for example, or penicil- lin – we are in it. It is difficult to overstate this: humans have been trying to create a cell therapy for cancer patients for generations. “People said: That can’t be done, You can’t make them from cancer patients, You can’t make them, You can’t get them, It’s too complicated,” says Crystal Mackall. “But it’s happening.” Though Novartis couldn’t confirm an official release date, Mackall suspects the drug will become widely available this year.

Cancer being cancer, of course, there are limitations. Until it clears further FDA hurdles, Novartis’s drug will be available only for children with B-ALL and not for any of the dozens of other types of cancers that affect children and adults. In solid tumours, the CAR-T cells aren’t strong enough to kill the whole thing, or they die before they finish the job. Worse, once attacked, some leukaemia cells will remove their CD19 proteins and go back into hiding. “The thing about cancer is, it’s quite a foe,” Mackall says. “The minute you think you’ve got the one thing for it, it’ll outsmart you.”

Slowly, though, the successes are mounting. At City of Hope National Medical Centre just outside Los Angeles, Behnam Badie, an Iranian-born brain surgeon who has the kind of bedside manner you’d dream of if you ever required a brain surgeon, is developing a surgical device that can continuously infuse CAR-T cells into the brain tumours of cancer patients while he operates. For a while, he was working with the California Institute of Technology to build a magnetic helmet that could move the cells to the correct places, but the project ran out of money. Meanwhile, Crystal Mackall is working on a backup target for the CAR-T cells, CD22, in case a child’s cancer resists the ones targeted to CD19. She is also trying to make the cells live longer. Working with similar but slightly different engineered cells, she has managed to get her therapy to stay alive and working for up to two years in patients with solid sarcomas. One of her patients has since got married and bought a farm. Another went on a volunteer trip to Africa. Mackall likens genetically engineered cells to rudimentary machines. Over the next decade, she says, scientists will refine them until they can control where they go and what they do and when. “We’re going to be in a situation,” she says, “where a doctor can tell a patient to take pills to activate his cells one week and then rest them the next.” In fact, a biotech company based in San Diego called Bioatla has already developed conditionally active markers that could tell a T- cell to kill or not kill based on where it is in the body.

Eventually, programmable cell machines could fight autoimmune diseases, or arthritis. They could be used to rebuild collagen in athletes’ knees. But, because such powerful new technology requires a ton of risk to attempt, none of this would have been developed without an adversary as vile as cancer to require it. “We treated 49 kids at the National Cancer Institute with refractory leukaemia. Every single one of those kids had exhausted every other therapy available. If it weren’t for the CAR-T cells, they were gonna die,” Mackall says. Sixty per cent of those children went into remission, and a sizable fraction of those