Bringing Good Things To Life (Science)?

Infiltrating ductal carcinoma tissue labeled with both H&E (hematoxylin and eosin) staining and clinical biomarkers.
Image courtesy of GE Global Research

In the spring of 2005, a team of applied physicists and electrical engineers from the General Electric Company filed into pathology laboratories at the Memorial Sloan-Kettering Cancer Center, taking notes on their clipboards and clicking their stopwatches. Day after day, they went from laboratory to laboratory to watch as clinical researchers section and stain tissue - "Click" - examine it under a fluorescent microscope - "Click" - take digital images - "Click" - analyze expression patterns or scribble diagnoses.

"I had worked in these lab environments," says Michael Montalto, a biologist at GE, "but nobody else at GE had. So, I knew it was important that these scientists had an idea as to what the market was going to look like." Montalto's team, working on GE's first Advanced Technology initiative, analyzed the numbers to identify bottlenecks in the scientists' procedures at Sloan-Kettering and to determine how GE's wide-ranging industrial expertise and imaging technologies could be applied to everyday problems in the wet lab.

Jens Rittscher, a computer vision expert.

Back then, GE's new CEO, Jeffrey Immelt, was three years into what BusinessWeek called "a cultural revolution." The previous CEO, tough-talking Jack Welch, had emphasized penny-pinching and deal making during his 20-year tenure. Immelt was on course to inject a little more creativity into the conglomerate's many tentacles, which stretch from Chinese aircraft engines to nuclear reactors. Immelt had just sealed the deal to purchase Amersham Biosciences, the UK diagnostic imaging leader, for $9.5 billion US in shares, but it wasn't supposed to be just another Welch-style acquisition.

That $9.5 billion included a sizeable premium above Amersham's market value because, presumably, GE would cross-leverage these assets with its own businesses and, some hoped, adopt Amersham's research culture, which drove so much internal growth. It made perfect sense, says Peter Ehrenheim, head of GE Healthcare Life Sciences. The life sciences tools market was worth $20 billion and growing more than 7% per year. GE had once been the driver behind such medical imaging technologies as X-rays, magnetic resonance imaging, and computed tomography scanning - technologies requiring Amersham's imaging contrast agents for marking tissues of interest. GE saw the microscopic realm of the pathology lab as an extension of that business. "Basically, GE bought a molecular imaging contrast franchise that was very strong and gave us a footprint in the life sciences," says Ehrenheim. From a business perspective, Immelt wanted GE to snatch up the recently slimmed-down Amersham before competitors such as Philips or Siemens could.

Any holes in the GE game plan were soon filled with subsequent acquisitions, and even the rehiring of former Amersham employees, who had departed in an earlier shake-up. GE purchased Biacore Life Sciences in June 2006, Wave Biotech in August 2007, and Whatman in February 2008 - all part of a push by the world's second largest company to be a top supplier of life sciences equipment. Last year, GE Healthcare had $17 billion in sales, with $1.3 billion coming from the Life Sciences division. Its competitors include Invitrogen and Thermo Fisher Scientific, which respectively claimed $1.3 billion and $9.8 billion in sales in 2007, although the three companies compete only in specific market segments.

Company representative Conor McKechnie says that GE Life Sciences aims to be the first or second in its specialized market segments. But the road isn't going to be easy: A number of Amersham employees left the company, some subsequent deals have fallen through, and investors have criticized the behemoth's eclectic portfolio. This year, the company experienced the biggest sell-off of shares in 20 years. Still, GE Healthcare is one of its strongest divisions, and since the Amersham acquisition, the company claims to have seen accelerated profit growth due to operational efficiencies and leveraging common cost bases. (Representatives declined to provide specific figures).

Jason Castle, a research scientist in GE's Biosciences Global Technology Organization, performs cell maintenance work in a cell culture hood.

Dustin Fenstermacher / wonderful machine

Product-driven research for GE Healthcare's Life Science business division still takes place at Amersham's facilities in New Jersey, Sweden, and the United Kingdom, but, for basic research, GE sunk $125 million into renovating its corporate-wide Global Research Center (GRC) in New York, and erected the "D-wing," a 5-story, 3,700 square meter biosciences laboratory, which houses more than 70 life sciences researchers. It employs 1000 PhDs whose expertise ranges from fluid dynamics to polymer design, and who are all going to play a role as GE seeks to cross-leverage its technical expertise and dominate the life sciences market. In addition, GE's life sciences business employs 450 researchers in GE Healthcare, and R&D spending in the life sciences amounts to $1 billion annually.

Even Invitrogen and Thermo Fisher Scientific are dwarfed by the resources available to one branch of a massive conglomerate like GE, and the arrival of this giant harkens a new era in a life sciences market that big business now dominates. So, can GE's glossy new laboratories foster the same innovative atmosphere its acquired companies once prized?

It's hard to escape the influence of Thomas Edison in the sunny lobby of the GRC. Edison's massive wooden writing desk is enshrined under a plate of glass, right next to the mechanical stock ticker the inventor once obsessively checked for his company's latest valuation. On each table in the nearby cafeteria, a placard with Edison's photo reads: "He had 1093 patents. You have some impressive numbers, too." Employees with 25 or more patents are featured on a bulletin board; the top spots are reserved for those with 100 or more to their name. Since 2002, GE has filed more than 50 patent applications on protein biomarkers, diagnostic pharmaceutical agents, and molecular imaging agents for cancer, and currently has nine clinical trials in these areas, some of which have their origins in preacquisition Amersham projects.

There's a lot less sun just down the hall in the GRC's Visualization Lab. The lights are off and Jens Rittscher, a computer vision expert, sits in front of the meter-long eyeball of a zebrafish projected on the wall above his head. The image is pixelated and noisy, like the static on a television set. Each colored dot represents a potential piece of data about cell-signaling events, untangled using an experimental package for InCell 1000, the high-content screening technology that Amersham originally developed. "This is ongoing research," says Rittscher. "We can segment it now. The next step is what to do with this information. What can you do if you can identify cell populations? What can you do if you measure the density of cells in certain areas?"

Evelina Loghin, a cell biologist at GE Global Research, mixes reagents that will be used in an assay.

Dustin Fenstermacher / wonderful machine

Moments earlier, Rittscher played a movie tracking the replication of three cells, each phase of mitosis automatically labeled with a different color. Then he pulled up a multicolored image from a slice of breast cancer tissue. Back at Sloan-Kettering, such images would have contained no more than two or three dyes, painstakingly prepared and analyzed by human technicians at pharmaceutical labs. Here, however, more than 20 different biomarkers have been labeled, localized in the cytoplasm, membrane, or cell wall, and quantified in a graph below the image. The image is a tour de force of computer vision algorithms, which can identify objects from their curvature and local geometry, and machine learning algorithms, which improve these identifications over time. "These are the sort of things, where we go one or two steps ahead, and we say 'this is possible'," says Rittscher, who came on board just about the time his colleagues were scrutinizing Sloan-Kettering.

Rittscher is part of an interdisciplinary team assembled by Mohan Amaratunga, the business program manager in GRC Biosciences. Amaratunga oversees the six research labs in Biosciences, one of 10 technology organizations that have a home on the GRC campus. Here, individual labs primarily conduct corporate-wide research rather than fulfilling the needs of businesses partners such as the Life Sciences division of GE Healthcare. Infrequently, collaborations occur directly between GRC researchers and outside companies, but such collaborations are increasing with the GRC Biosciences organization and biotechnology or pharmaceutical companies.

An affable biochemist from Sri Lanka, Amaratunga worked on biosynthesizing fatty acids and other chemicals used in GE's plastics business before moving into management. In comparison with the product development labs at the former Amersham facilities, the GRC is more of an academic institution. Make no mistake, however: Basic science at the GRC is not about publishing papers. "We file patents," Amaratunga says.

Nevertheless, the GRC is all about pursuing long-term projects that push into what Amaratunga calls "white spaces" - markets that don't yet exist. To pursue them, GE began its Advanced Technology programs, and molecular pathology was one of the first. After amassing data from their Sloan-Kettering expedition, Amaratunga, who was not part of the tour, sat down with many of those same cancer researchers and discussed the limitations in current imaging technologies. "Pathology people are using 100- to 150-year-old technology," he says, referring to the human grunt work and the venerable staining technologies that are still used today. "People are doing a great job, but it's mechanical; they have to look at every slide one-by-one. It's a slow process, labor-intensive, and it's in black-and-white." Moreover, sharing information is a problem. "If you generate a slide today, somebody at another site can't look at it." Although it is possible to digitize images of slides, it must be done at just the right focal length and resolution, with just the right stain. There is no way to digitally revisit slides in the way, say, Google Earth allows users to explore a three-dimensional terrain, zooming in and out and focusing at different depths.

A breast carcinoma tissue stained for two important clinical markers; Her2 (red), p53 (purple).

Courtesy of GE Global Research

This white space is already turning into a real world collaboration between researchers at GE Healthcare Life Sciences and its pharmaceutical partners. The first step along the way was the development of a technology that would allow researchers to chemically stain and destain samples a hundred times, and to automate this process. Fiona Ginty, a bioinformatics expert in charge of the biological end of the molecular pathology work, has impressed scientists with presentations featuring her images. These imaging technologies depend on radiotracers coming out of the former Amersham operations in Sweden, the United Kingdom, and the United States. In a collaboration announced in November, the GRC is working with Eli Lilly and Company to analyze tissue samples from clinical and preclinical trials for two anticancer drugs, Enzastaurin and a transforming growth factor (TGF-beta) small-molecule inhibitor, in order to develop targeted therapies. "This is unique because we generally don't work on such collaborations with such a large company on the other side," says Jeremy Graff, a research advisor in Eli Lilly's cancer drug unit. "It's very different from what we've tried before."

For Enzastaurin, researchers are primarily looking at biomarkers on the glycogen synthase kinase (GSK-3 beta) pathways that the drug inhibits, along with markers that complement that pathway. "We suspect that [you'll have] more robust information if you think about an entire pathway of signaling and look at not only the first signaling event but [also] two, three, four, or five in combination," says Graff. On top of that, Eli Lilly pressed GE to segment the tumor mass, giving the drug maker an unprecedented view of the tumor at the cellular level. "With high-content imaging," Graff says, "we can not only look at what happens within the tumor but [also] look at endothelial cells, and understand signaling within epithelial cells and tumor cells." Without GE, he says, "I don't think we could have answered these questions."

Amaratunga says GE has no interest in moving into therapeutics, but that the company is watching as the lines between diagnostics and therapeutics continue to blur. In June, for instance, GE Healthcare and the University of Pittsburgh Medical Center announced the formation of Omnyx, a joint company that will provide high-volume slide scanning and digital pathology tools for clinicians.

Of course, setbacks have occurred since the Amersham acquisition. In the spring of 2008, CEO Immelt promised that GE would meet its earnings forecast, but at the end of the first quarter, the company announced a 12% drop in earnings. GE's financial and industrial units were hit hard, but GE Healthcare was also hurting. Welch, the former CEO, criticized Immelt on CNBC for missing his projections. "Here's the screw-up: You made a promise that you'd deliver this, and you missed three weeks later."

There have been a few other hiccups along the way. "We have lost some staff, no doubt about it," says Ehrenheim, but he estimates that 75-80% of the employees who were with Amersham at the time of the acquisition are still with GE, which represents a typical company attrition rate of just over 5% per year. Some Amersham directors, such as William Castell, who left his post to become chair of the Wellcome Trust, had planned to depart from the beginning. Others became redundant with the merging of the two companies. Ehrenheim, who joined Pharmacia in 1983 before it was a part of Amersham, says, "All in all in life sciences, the integration in GE has been good. The big benefit is we have gotten the resources we wanted."

But some say that sacrifices were made along the way. Andrew Campbell, director of Ashridge Business School in the United Kingdom, doubts that GE has adopted Amersham's more flexible approach to research-driven growth, and he wonders if the premium GE paid was worth it. (Amersham's shares leapt from about $11 to $12.61 after confirmation of the takeover, which was valued at $13.28 per share, meaning that GE paid at least a 20% premium for the acquisition.) While locking up Amersham's contrast agents may have been valuable, Campbell points out that the acquisition also prevents Amersham from maintaining external collaborations with GE's competitors in the imaging business, such as Toshiba and Philips. "It would surprise me if the deal had been a big success," he says.

Moreover, it didn't help the science that in the lead-up to the purchase, the six sites that made up Amersham's US research wing had been slashed and burned to boost profitability, according to Stevan Jovanovich, who was the vice president of global research at Amersham until 2003, and now heads up an independent Amersham spin-off called Microchip Biotechnologies in Dublin, California. "I had roughly 120 people in research," he says. "What happened when [Amersham] centralized research to New Jersey is they ended up with something like 12 people." Jovanovich says the company had projects ranging from stem cell differentiation to next-generation sequencers and was planning to focus on systems biology, much of which disappeared with the downsizing. "I've talked to some GE Global Research folks," he says. "They are growing some capabilities but that was a hiccup." According to GE spokesperson Conor McKechnie, the 90% staff reduction that Jovanovich experienced at Amersham was specific to R&D for gene sequencing, which had "matured."

Last year, GE was unable to reach a deal to purchase Abbott Laboratories, which develops products to test blood proteins for heart disease, along with many other products. Although the deal would have provided GE with valuable tools in early-disease detection, some industry analysts say that the acquisition would have been a bad move because parts of Abbott's business were in decline. More generally, investors have criticized GE's eclectic portfolio, which ranges from a film studio to real estate to jet engines, but in many ways this diversity might actually benefit life sciences research, particularly at the GRC. Ger Brophy, a general manager for Life Sciences in GE Healthcare, says that the depth of knowledge at the GRC, coupled with GE's internal manufacturing capabilities, provides an advantage over smaller competitors. "The GRC is almost an academic institution," says Brophy. "Those guys are great at making step changes in technology, but the secret to doing this well is to nurture that research development and at the right stage move it into product development."

Prameela Susarla, for example, is a fast-talking materials scientist at the GRC who works in membranes and separation technology. She thrives on interfaces: her group works with GE Energy, GE Healthcare, and GE Water, designing and testing novel membranes for tasks ranging from reverse-osmosis to protein separation. "What we are able to learn from our interactions with GE Energy and GE Water, we are able to apply immediately to GE Healthcare problems," she says. Since January, Susarla has been running an eight-person team that is designing a new cell-expansion matrix to help companies manufacture stem cells in the disposable plastic bioreactors from Wave Biotech, the company that GE Healthcare purchased in August. While much of the work on bioprocessing of vaccines is taking place at former Amersham labs in Sweden and New Jersey (see sidebar, Vaccine Dreams), GE sees the bioreactors as part of the manufacturing supply chain for a serious "white space" in the next decade: cell therapies.

One problem with growing primary cells in the Wave bioreactors is that cells get injured when the beads they are growing on collide. To study the problem, Susarla assembled not only chemists and biologists, but also an expert in multiphase flows - liquids and particles mixed together - and a computational fluid dynamics expert who has worked on the geometry of automobile engines during an internship with Daimler-Benz.

"It's the type of expertise you might not find at a biotech company," says Susarla. She says the group has already succeeded in reducing shear forces and collisions of the beads in the Wave reactor. She was not, however, willing to discuss details, because that's one patent GE has yet to file.

Correction: When originally posted, the story incorrectly valued GE's purchase of Amersham at $9.5 million. The correct figure is $9.5 billion. The Scientist regrets the error.