Here Is a Human Being (27 page)

But if the Broad didn’t need the Polonator, perhaps the Polonator didn’t need the Broad quite yet. In 2010 Kevin McCarthy admitted that his company, Dover, was almost certainly fooling itself when it launched the Polonator two years earlier. There had been problems with the manifold that moved the reagents in and out, and as a result the machine was leaking. Its self-cleaning mechanisms were not up to snuff. And the software was becoming increasingly byzantine. “Eventually the Church lab and I agreed that the fluidics and the software needed to be blank-slated,” McCarthy told me. Dover swapped out the manifold in favor of leakproof rotary valves.
²¹
And while Greg Porreca left Team Church in 2009 for a start-up,
²²
in New Mexico Jeremy Edwards stepped in to advance the biochemistry and develop new sequencing protocols. In a few months Dover was shipping Polonators again. McCarthy, who had spent a fair amount of time traveling around and troubleshooting machines in the field, made a user manual available in 2010.
²³
By that time, there were fourteen Polonators installed. “That’s probably more than Helicos,” George pointed out
²⁴
—and for the moment he was right.
²⁵

Now it would be a question of winning over some of his own people. “I think my lab is divided,” he admitted. “One group is pushing the Polonator into extremes that are ridiculous—they are making it do all kinds of amazing new tricks. For example, you can synthesize DNA on it now; you can use it to look at cell morphology and behavior. But the other half of the lab is skeptical and they aren’t using it.” He chuckled into the phone. “I suggested we cut off all of the lab’s supply money for everything except the Polonator.”
²⁶

Across town in Kendall Square, Helicos was still making a precarious crawl across the Valley of Death. In September 2008 the company said that it anticipated fifteen to thirty orders for HeliScopes in the next nine months.
²⁷
In December it reversed course and announced that it would lay off 30 percent of its workforce and retract any and all past predictions of both its future financial results and HeliScope orders.
²⁸
That same week Helicos realigned its senior management: Steve Lombardi was out as CEO, replaced by board member Ron Lowy.
²⁹
Lombardi would remain president and focus all of his efforts on commercializing the HeliScope. “Stepping down as CEO was hard for me,” said Lombardi, though he insisted that “it was the right move to make. We needed to go really, really fast. I couldn’t do both jobs.”
³⁰

The news got worse. Helicos’s stock, which had peaked at sixteen dollars per share in January 2008, bottomed out in December at twenty-two cents. Two thousand and nine began with the company’s first customer, Expression Analysis, deciding to return its HeliScope,
³¹
which had been delivered to the company’s Durham, North Carolina, labs with great fanfare the year before. “The thinking was pretty simple,” explained CEO Steve McPhail, a friendly and rugged-looking guy with a firm handshake, of the decision to send the instrument back to Boston. “We had two large pharma clients that had opted to use Illumina data. Both of them said to us, ‘You
need
to get into this market.’ By then we estimated that Illumina already had over fifty percent of the next-gen market. Whenever that type of standardization happens, there are opportunities for outsourcing.”
³²
Expression Analysis became the first genomics services provider to be fully certified on every platform Illumina offered—genotyping, gene expression, and sequencing.
³³

Not only was the HeliScope getting lapped by the Illumina and ABI sequencing machines in the marketplace, the company was hemorrhaging cash. By mid-2009 it was down to its last $5 million.
³⁴
Helicos brought in an investment banker to advise on “strategic alternatives,” which was construed as a euphemism for trying sell oneself to the highest bidder.
³⁵
Would Stan Lapidus’s insistence on an independent Helicos be trumped by the company’s need for a lifeline?

“It’s inescapable that the other start-ups in next-gen sequencing—454,

Solexa, and Agencourt—all got married to Daddy Warbucks,” Chad Nusbaum said in 2008 with some prescience. “Helicos didn’t … or hasn’t yet. It’s no small thing to invent a machine and make it work. But neither is it a small thing to get that machine out in the field to customers, to support it, to sell it, and to manage a supply chain. Once you start delivering instruments, it’s a whole other ballgame. You’re supporting users, the software doesn’t run, there’s a bug here, a hardware problem there, yadda yadda. Every single one of these companies has had problems shipping reagents. Those are the sorts of growing pains you expect. But big companies like ABI and Illumina, with limitless resources and closets full of intellectual property, can keep pulling old stuff from the attic for years. I’ve always liked the Helicos people. They’re earnest and smart. But they’re a small company.”
³⁶

Tim Harris intimated that Helicos’s engineer-centric culture cost it valuable time in getting the HeliScope out the door. “The scientists knew how to sequence and the engineers didn’t,” he said. “The engineers started over and made every component new; every single part was custom made. That has a large impact on how fast you can go.” A few months after publishing a paper in
Science
in which he and his colleagues used the HeliScope to sequence a virus,
³⁷
Harris left his post as senior director of research for an academic gig at the Howard Hughes Medical Institute campus in rural Virginia.
³⁸
“My departure was amicable, but my role was limited,” he said. “The people who ran Helicos had a specialist view and I was a generalist. I was in a research pigeonhole and I wanted a larger playing field. It was time for me to go.”
³⁹

Eventually a few rays of light began to poke through the clouds hanging over Helicos. In the coming months it would raise additional money.
⁴⁰
The stock price rebounded from its all-time lows.
⁴¹
Steve Quake, the guy who invented the Helicos technology, sequenced his own genome with a HeliScope in a few weeks, with a small number of people’s help, and for a reasonable-at-the-time cost of $48,000.
⁴²
Helicos scientists showed that one could use the HeliScope to sequence RNA directly
⁴³
without first converting it into DNA, something that had never been done before and that none of the other platforms could do. More new applications emerged.
⁴⁴
It sold four instruments to a science institute in Japan.
⁴⁵
A Waltham-based diagnostics company bought a HeliScope and promptly launched a test of three hundred genes related to hereditary cardiovascular disease.
⁴⁶
A few more orders came in.
⁴⁷
It wasn’t a torrential pace exactly, but it was fast enough to buoy spirits in Cambridge, for the moment anyway.

As we finished our Portuguese lunch down the street from his office, Lombardi waxed philosophical. “We’re not out of the woods, but the life of a start-up is a hypothesis. You hope that your technology and product strategy are right, but it’s always going to be a tough environment. In the 1990s I was part of the investor relations group at Affy saying, ‘Why even
worry
about Illumina?’ Now what are they saying about Illumina? ‘It’s over! They’ve won! They’re invincible!’ I don’t think that’s ever true in this business.”

Two months after we spoke, Lombardi was out. On a Friday after the market closed, Helicos quietly filed a form 8-K with the Securities and Exchange Commission announcing his resignation. “Mr. Lombardi did not resign as a member of the Board of Directors as a result of any disagreement with the Company on any matter relating to the Company’s operations, policies or practices,” it said. His successor as CEO and president, Ronald Lowy, personally thanked him for his “ … efforts and professionalism to help Helicos achieve its goals.” Elsewhere in the document, under a section titled “Future Conduct,” there was a standard “non-disparagement” clause—neither side would say anything unpleasant about the other.
⁴⁸

Steve was not about to violate either the spirit or the letter of the agreement. “Things happen,” he told me a few days after his departure became public. If there were any bitterness in his voice, I couldn’t detect it. “It was amicable. It was bittersweet. But it was the right thing to do for the company. Don’t worry about me, Misha … I’ll pop up somewhere.”
⁴⁹

For Helicos, life remained tenuous. In March 2010, investment bank Leerink Swann dropped coverage of the company, citing concerns over the HeliScope’s high price, limited market, and looming competition.

“Time is running out,” warned the analyst.
⁵⁰
A few weeks later the company announced that substantial doubt existed over “its ability to continue as a going concern.”
⁵¹

When I landed in Boston for the second annual PGP-fest, I didn’t have time to dwell on either the plight of sequencing companies or my own existential problems. My nephew Noam was in the hospital. A senior in high school, he was tall and fit, a rower who had somehow developed a rare form of deep-vein thrombosis, Paget-Schroetter syndrome, a condition that causes blood clots in the extremities of otherwise healthy young people.
⁵²
The docs broke up the clot but in the process gave him too much morphine. Noam’s breathing slowed to an erratic crawl. After a rough night, he began to come out of his narcotic haze. He received heparin and would start on other anticoagulants soon, including Coumadin. Coumadin, aka warfarin, is the most widely prescribed blood thinner in North America, but different people respond to different doses and the therapeutic range tends to be narrow—not surprising given that its original use was as rat poison.
⁵³
Warfarin has become the advance guard in the brave new realm of pharmacogenomics (see chapter 6). If we can genotype people for the genes that code for certain metabolic enzymes that break down drugs in the liver, we can predict who will respond to warfarin and at what dose; but we’re still working out how much it costs compared to old-school methods.
⁵⁴
I wondered whether Noam had been genotyped but decided against asking about it as I listened to my agitated brother simultaneously talk at me on his cell phone over the ambient noise of Massachusetts General Hospital, minister to his ailing son, and fumble with the television in the hospital room, desperate to tune in the Boston Red Soxes’ bid to salvage the American League Championship Series against the Tampa Bay Rays. I said I would check in later.

Pharmacogenomic testing is often hailed as the genomic tool that will be the first to move from “bench to bedside.”
⁵⁵
That may be true, but the latter phrase still makes my skin crawl. I imagine it has been used in thousands of PowerPoint presentations to describe the arduous path from research laboratory to clinical practice. It is a mom-and-apple-pie sort of goal, an agreed-upon desire of every stakeholder in the personalized medicine field. It is an implicit call for randomized clinical trials of high-tech postgenomic measures of human individuality à la response to warfarin. It is a sort of “consensus statement.” It is also a cliché, a shorthand that has helped to spawn countless “road-map” and “translational medicine” initiatives and all sorts of other schemes to make personalized medicine the “standard of care.”

But how close were we to that reality? With or without genomics, when would we be in a position to circumvent the inefficiencies woven into the system, to say nothing of our own stubborn humanity: the initial complaints of pain, the visit with the confused and green resident, the referral to the ER, the four-hour wait, the shunting from specialist to specialist, the messy procedure, the middle-of-the-night wakeups, the reams of paperwork, the titrations of so many medications, the crushing disappointment and fear wrought by complications, and the uncertainties of reimbursement? And all of this before we failed to comply with our doctors’ orders anyway. Nor were doctors themselves immune from recalcitrance. In his outstanding 2007
New Yorker
article and subsequent book, surgeon and medical writer Atul Gawande described the quixotic crusade of Dr. Peter Pronovost to have doctors follow a simple five-step checklist to avoid central-line infections in intensive care units.
⁵⁶
In its first fifteen months, the program saved $175 million and 1,500 lives, despite physician resistance to it. In 2009, eight years after it started, the checklist was finally rolled out in twenty-eight states.
⁵⁷
Clearly this quick, simple, and inexpensive intervention could make a difference. What about genotyping for drug response? I didn’t—and don’t—know. What I did know was that if this health-care “system” is what we were getting for 16 percent of our gross domestic product in 2005, and 25 percent of our GDP in 2025,
⁵⁸
or roughly a zillion and a half dollars, it was pretty obvious that we were getting ripped off. President Obama had his work cut out for him—we all did.

Noam was supposed to be released on Monday but developed another clot and had to go back in for more surgery. He would have at least three subsequent procedures over the next week, including the removal of a rib. When he came home he would have to dose himself with warfarin. His doctors would be watching him closely to be sure he stayed within the therapeutic range; otherwise he could bleed to death.

With the help of his family, doctors, physical therapy, blood thinners, and Vicodin, he regained the full use of his arm over the ensuing months. For their part, the Red Sox won Game 6, but fell to Tampa Bay in Game 7 and were eliminated from the playoffs.

I took the train from the bedside (Mass. General) to the bench (George’s lab). It was almost noon and Marilyn Ness and her film crew had been running around for most of the morning. With her dog-eared shooting script in hand, she seemed to be the only one who knew what was going on and where everyone was (“George will be here any minute,” she assured us).
⁵⁹
A
New York Times
photographer brandished two digital cameras with enormous lenses and checked the lighting.