Inside the race to diagnose cancer from a simple blood draw

BOSTON — It’s a medical puzzle that has snagged the attention — and the money — of Bill Gates, Jeff Bezos, and venture capitalists across the nation: Is it possible to diagnose cancer from a simple blood draw?

Surgical biopsies are the norm, but they’re invasive, expensive, and carry the risk of infection. So investors have poured hundreds of millions into the goal of developing “liquid biopsies.”

There has been some progress: Blood analysis is now used to identify the best treatments for certain cancers — and to update treatments as the cancer mutates. But so far, no one has scored the ultimate success: diagnosing incipient cancer from a vial of blood drawn from a patient who looks and feels perfectly healthy.

Some 38 companies in the US alone are working on liquid biopsies. Most are trying to analyze blood for fragments of DNA shed by dying tumor cells.

In a lab overlooking Boston Harbor, an unlikely duo is taking a different path.

Mehmet Toner and Dr. Daniel Haber — a Turkish bioengineer and a Jewish geneticist — joined forces a decade ago, in a partnership forged over tuna sandwiches. While other scientists scan blood for scraps of tumor DNA, Toner and Haber filter out all the healthy components of the blood — then scoop up any whole tumor cells left behind.

The process is costly and time-consuming. And their peers mock it as hopelessly old school.

Mention the idea of capturing whole tumor cells to venture capitalists and “they roll their eyes,” said John Boyce, chief executive of Exosome Diagnostics, which recently launched a liquid biopsy test to identify lung tumor mutations.

But Toner and Haber are undaunted. They’re convinced whole cells will give doctors valuable information about cancer. A fragment from a dying tumor cell, Haber said, “doesn’t tell me anything about the biology of the living tumor.”

A tumultuous market

If liquid biopsies work as hoped, the market in the US alone is projected at $29 billion, according to a 2015 report from investment bank Piper Jaffray.

“It is the one area of oncology you see featured at every single conference,” said Dr. Jorge Villacian, chief medical officer at Janssen Diagnostics, a subsidiary of Johnson & Johnson, which has the sole FDA-approved liquid biopsy on the market. “There’s a great deal of enthusiasm.”

And a great deal of drama.

On-Q-ity, a liquid biopsy startup based in Waltham, Mass., folded in 2013. One of its investors said the technology would likely not be market-ready for another five to 10 years.

And last fall, federal regulators upbraided San Diego-based Pathway Genomics for selling a liquid biopsy kit that purported to screen healthy, at-risk individuals for early signs of cancer. The Food and Drug Administration called it “a high risk test that has not received adequate clinical validation and may harm the public health.”

Nonetheless, Gates and Bezos recently teamed up with Illumina, a leading DNA sequencing company, to launch yet another liquid biopsy startup — this one ambitiously called Grail.

The field is such a roller coaster that Haber and Toner make a point of being cautious and methodical.

“You find strange things, you question yourself, and you test it over and over again to be sure it is right,” Haber said.

Flipping haystacks

In 2006, Haber was fresh off identifying a mutation that makes lung tumors susceptible to a particular class of drugs, while Toner was testing a microfluidics technology to sort rare cells, such as fetal cells in the blood of a pregnant woman.

Each had worked at Massachusetts General Hospital for over a decade without ever crossing paths, until Dr. Kurt Isselbacher, founding director of the MGH Cancer Center, brought them into his office for lunch and suggested they work together to capture tumor cells in the blood.

At the time, there was one FDA-approved liquid biopsy on the market, Janssen Diagnostics’s CellSearch, a device that counts the number of circulating tumor cells (CTCs) in the blood of patients with metastatic breast, colon, and prostate cancers to predict a good or bad prognosis.

But little was known about CTCs in the blood. How many were there? From what types of tumors? Could they be efficiently separated out and analyzed?

“The biology of these cells was a mystery,” said Toner.

Flying without a compass, Toner and Haber worked painstakingly on the same idea that had been successful with CellSearch: “positive selection” or actively pulling cancer cells out of the blood by labeling them with sticky, specific molecules called antibodies.

The first iteration of their test was a thick, rectangular silicon chip. It worked: They published a paper showing it could detect tumor cells in the blood of patients with lung, prostate, pancreatic, breast, and colon cancer. But the technology was prohibitively expensive, more than $1,000 per chip, and painfully slow — it took several days to identify one tumor cell.

CTCs, it turns out, are extremely rare in the blood — an estimated one in a billion.

Haber and Toner brainstormed alternative approaches from 40,000 feet in the air. Both men frequently travel to conferences and talks, so when they fly together, they’re sure to get aisle seats across from one another. That way, they can pass ideas back and forth on — what else? — cocktail napkins.

“We solve a lot of problems that way,” said Toner with a happy shrug.

Unless the flight attendants get in the way, Haber added pointedly.

Soon, they had a solution. “Since we know everything about blood cells, we flipped the problem,” said Toner. “Instead of going after a needle in the haystack, we decided to get rid of the haystack.”

In other words: They’d filter out all the healthy components of the blood and look for tumor cells left behind.

It was about this time that Johnson & Johnson stepped in and offered validation — and a big check — for the duo’s efforts, committing $30 million over five years to construct the next generation of the technology.

The most current device, the “iChip,” is made of thin, lightweight plastic etched by a laser with microfluidic grooves just visible to the naked eye. This time, the chip is round, due to the fact that it is manufactured on old CD equipment.

Blood first passes through a series of short rods on the chip, like a miniaturized pachinko machine. Smaller objects move more quickly, bouncing from post to post, so red blood cells and plasma are quickly sorted out and expelled as a red fluid.

Chunky white blood cells and tumor cells, on the other hand, move more slowly through the channels and are funneled single-file into the next section of the chip. Here, the white blood cells, previously labeled with magnetic beads, are removed via a magnetic field and also expelled, as a milky white solution.

At last, the unmarked, untagged tumor cells drip into a pristine test tube. With those cells, the world is Haber’s oyster.

Vials of blood and tumor cells shuttle back and forth in the lab he and Toner share in MGH’s research facilities at Charlestown Navy Yard. They use the material in a dozen ongoing studies, studying how hormonal drugs affect prostate cancer, how immunotherapy alters tumors, and how breast cancer mutates in response to chemotherapy.

“The information content from whole cells and the ability to look at how each cancer cell is different from the other is very powerful,” Haber said.

The fragment hypothesis

But as more researchers and biotech companies enter the field, most of them have shied away from whole cell capture. Instead, the focus is on small DNA fragments shed from dying tumor cells, called circulating tumor DNA (ctDNA).

One study found that ctDNA was detectable in the blood of anywhere from half to nearly 90 percent of patients, depending on the type of cancer. The study also found that the ctDNA could be used to identify clinically relevant gene mutations, which could help oncologists select which specific medication might work best against a given tumor.

California-based biotech Guardant Health — formed in 2012 and recent recipient of $100 million from eager investors — uses that technique to search for 70 actionable mutations in solid tumors.

Yet another novel approach is also making waves in the field: Detecting tumor mutations via DNA, RNA, and proteins isolated from tiny vesicles called exosomes that are expelled by living cells.

Cambridge-based Exosome Diagnostics just snagged $60 million in investor funding and has already released a lung cancer liquid biopsy to detect tumor mutations via those vesicles. The company has three other tests scheduled for release this year, some of which co-analyze exosomes and ctDNA as a way to gather information from both living and dying cancer cells in the body.

Haber chalks up the industry’s enthusiasm for cell-free tests to how complicated and expensive it has been to isolate rare cells from the blood. But that is changing, he said.

“We’re on the cusp now of having a very standardized, affordable technology that will change the whole equation again,” he said.

It’s been 10 years since Haber and Toner started this journey.

“It’s time to renew our vows,” Haber joked, turning toward his friend. “You look good for an old guy.”

Toner laughed, adjusting his round, black spectacles.

Janssen Diagnostics, which funded the research, is pleased with the iChip emerging from their lab. “The engineering is almost ready,” Villacian, the chief medical officer, said.

Though the grant has now run out, Janssen and MGH are now in talks about starting a biotech company to move the technology toward commercialization, Haber said.

Currently, Haber and Toner are designing several large clinical trials to demonstrate when and where the iChip will be most useful. Their plans should be ready in, oh, two or three more cross-country flights.