The prominent researcher has been put on administrative leave pending an investigation into unspecified allegations.
In efforts to translate basic-science results into pharmaceuticals and other technologies, success cannot be taken for granted.
March 1, 2018|
Back in the ’90s, immunologist James Allison wasn’t trying to develop a cancer drug. “I was doing just really fundamental research trying to understand T-cell regulation,” he says. But in the course of that work, performed at the University of California, Berkeley, Allison discovered that a protein receptor called CTLA-4 negatively regulated T-cell responses to antigens, and that inhibiting that receptor with an antibody enhanced T-cell activity.
The clinical applications were obvious. “I had the idea that you might be able to exploit that to get immunological responses, T-cell responses, to tumor cells,” says Allison, now chair of immunology and director of immunotherapy at the University of Texas’s MD Anderson Cancer Center. In a 1996 Science paper, he and his colleagues reported that, in mice, this approach worked: rodents treated with an anti-CTLA-4 antibody rejected tumors. “I thought this was pretty cool. We patented it,” says Allison. “I thought everybody would jump at it.” But everybody—in particular pharmaceutical companies—did not.
For more than two years, Allison says, every company he approached turned him down. Then, the first company to express interest in licensing Allison’s patent and attempting to develop an anti-CTLA-4 drug floundered in its translational efforts. That firm, NeXstar Pharmaceuticals, tried in the late ’90s to develop an RNA aptamer to bind and inhibit CTLA-4, but, as Allison describes it, they “just couldn’t get it.”
Eventually, through a friend, Allison connected with the pharmaceutical company Medarex (later acquired by Bristol-Myers Squibb), which took on the project, and, working with Allison, developed an anti-CTLA-4 antibody into a drug, ipilimumab (Yervoy). After years of clinical trials, in 2011 the FDA approved ipilimumab for treatment of late-stage melanoma—about 15 years after the therapy’s conception—and in 2015, Allison won the Lasker-DeBakey Clinical Medical Research Award in honor of the work.
Allison’s wish to translate his research findings into drugs or other technologies is an increasingly common one. According to Nature Index Science Inc. 2017, the number of academic-industry collaborations more than doubled from 12,672 in 2012 to 25,962 in 2016, and half of those 2016 collaborations were in the life sciences. Most labs do not have the resources to take a discovery from the bench to the clinic, and industry offers funding, personnel, and experience in manufacturing products and bringing them to market.
“I think it’s been borne out that really to get treatments to patients . . . it needs to be a partnership,” says Joan W. Miller, chair of ophthalmology at Harvard Medical School, who has worked with industry partners to develop drugs for macular degeneration and other eye diseases.
But, as Allison found out, the success of those partnerships is far from guaranteed. New therapies and technologies can be delayed, or even blocked entirely, on the road to the clinic. Although most universities have technology transfer offices to help researchers file for patents, form start-ups, and collaborate with industry, researchers across the academic-industry divide have to contend with differences in institutional cultures, on top of the scientific challenges of translating academic results into clinical treatments.
For instance, while academic researchers generally receive grants to pursue given projects over several years, industry labs tend to expect faster results and are quicker to abandon projects that don’t show immediate promise. Academic culture encourages openness, with researchers hurrying to publish their results. Industry, meanwhile, has a tendency to want to safeguard, rather than publicize, results from which a company might be able to profit.
But where there’s a will there’s a way, and collaborators have developed a variety of strategies to overcome such differences and facilitate academic-industry partnerships, from forging relationships as soon as possible, to placing emphasis on trust. As Isaac Kohlberg, chief technology development officer of Harvard’s technology transfer office, says, “It has been our experience that if there is the intent that these partnerships should work, they in the end work.”
One of the major problems in translating research from academia to industry is a lack of reproducibility. According to studies by pharmaceutical companies Bayer and Amgen, when industry labs try to reproduce academic results they are unsuccessful approximately 80 percent of the time, as McGill University ophthalmologist Leonard Levin and coauthor Francine Behar-Cohen note in a 2017 Trends in Pharmacological Sciences editorial.
There are likely various reasons for this reproducibility problem, from insufficiently detailed methods in published studies to the use of different animal models and statistical approaches in academic vs. industry labs. The result is that academic results get “lost in translation” when industry labs try to reproduce them, Levin tells The Scientist—“something that is well recognized by industry, but is not so well recognized in academia.”
To solve this problem, Levin recommends that academic scientists, in addition to writing more-detailed methods and holding data to higher statistical standards, should begin their collaborations as early as possible in the development of a product. If a lab and an industry partner are already working together on a project then such translational problems should be less severe.
Starting academic-industry cross-talk early in the career of an academic also promotes reproducibility, Levin says, since scientists who know how industrial research works are more likely to carry out studies that can achieve translation. To educate academics about industry research, Levin and Behar-Cohen recommend the establishment of university programs in which academics spend time working in industry labs (see “Making the Most of School,” The Scientist, May 2016).
Universities such as Yale offer industry fellowships for grad students, postdocs, and junior faculty. “Academics, particularly younger junior faculty, are totally unaware of what it takes to actually do the translational research that’s going to make their observation that they make at their lab bench into a successful product,” says Jon Soderstrom, managing director of Yale’s Office of Cooperative Research. It’s only through experience working with industry that researchers learn to collaborate with industry, he adds. “You can’t teach it as a lecture.”
Such hands-on early career experience is often linked to another aspect of successful collaborations, Soderstrom says: a gradual development. “The most successful collaborations with industry are actually growing out of personal relationships that the faculty already have with somebody at a company or have developed over time with a company, usually based on some relatively small initiatives to start, and then they build up over time.”
Despite the benefits of relations that start early and develop gradually, there are some sources of tension that are likely to be inevitable whatever the nature of the collaboration. Partners may clash over intellectual property or decisions about how much a company will pay to license a patent from a university lab. And then there’s liability: who will be legally responsible for what, and under what circumstances either partner could be sued. “Every single collaboration we enter into, [liability] is an issue that has to be discussed and negotiated,” Soderstrom says. “Suffice it to say each side is interested in protecting its interests from damages resulting from lawsuits,” he adds in an email.
Such tensions can significantly delay a project’s progress. For example, several years ago Stephen Waxman, a neurologist at Yale School of Medicine and Veterans Affairs Connecticut Healthcare System, collaborated with Pfizer to test a drug for “man-on-fire syndrome”—in which pain-sensing neurons overreact to mild stimuli, causing searing skin pain—in human patients (see “Channeling the Pain,” The Scientist, January 2018). Yale and Pfizer investigators had planned to work together to carry out the trial, using Pfizer’s drug and patients recruited by Waxman, at a Pfizer clinical research facility in New Haven. Yet for a while, liability concerns prevented Yale researchers from working at the Pfizer lab. “It slowed things down,” Waxman says. “That frankly was discouraging, and it took some strong urging on my part to move things ahead.”
One common strategy to minimize such logistical difficulties is to draw up master agreements that cover multiple projects. These agreements decrease the strain that per-project negotiations would otherwise place on partnerships by laying out the parameters of the collaboration—for example, whether a company might have first dibs on licensing the resulting intellectual property and to what extent academics are free to publish the results—in writing.
Harvard, for example, engages in long-term collaborations, called strategic alliances, with company partners. The multi-project agreements that govern these collaborations serve to expedite collaborative research, Caroline Perry, director of communications at Harvard’s technology transfer office, tells The Scientist in an email. “Our research alliances encourage ongoing collaboration by establishing a common understanding up front, under which multiple specific projects in various labs can then be initiated more quickly.”
Such long-term agreements can also help resolve the conflict between academics’ wish to publish results and industry’s preference for secrecy. As described in a 2012 article in the MIT Sloan Management Review by the University of Bath’s Ammon Salter and Imperial College London’s Markus Perkmann, who both research academic-industry collaborations, short-term, confidential projects would ordinarily not appeal to academic researchers. But in the context of a longer, more open collaborative project, academics might find short-term secret projects more palatable.
Handled the right way, logistical matters are simply part of the collaborative process. In the case of Waxman’s collaboration with Pfizer, “it wasn’t that there was an issue,” Soderstrom says. Rather, “we had to contractually figure out . . . who was responsible for what,” or who was liable for the patients in the trial. In the end, Yale and Pfizer researchers did work together at the Pfizer research unit, and the resulting study appeared in Science Translational Medicine in 2016. Waxman looks back on the liability-related delay as a “legal administrative wrinkle.”
Perhaps the most insidious threat to collaborations across the academic-industry boundary, however, is a lack of communication. Barriers to information flow might arise for a number of reasons. For example, a firm might withhold information about an invention it doesn’t want the university to take part in, or a university might restrict the flow of information or materials, such as cell lines or plasmids, from its labs out of fear that the companies might use them without giving the institution credit. That “only leads to frustration,” says Soderstrom.
To avoid such outcomes, he recommends keeping in frequent contact with industrial partners. Key to success, he says, is “free flow of information” in both directions. “Regular exchanges of telephone calls, emails, visits, et cetera. . . . If you don’t have full and open transparent communication it tends to make people suspicious and distrustful. And if you have that, it’s going to fail.”
The value of trust in academic-industry collaborations is supported by research. For a pair of studies published in 2010 and 2012, Salter and colleagues surveyed both academic and industry researchers in STEM fields about two types of barriers to collaboration: “orientation-related barriers,” such as academia and industry’s different perspectives on publishing results, and “transaction-related barriers,” or those related to intellectual property and other logistical details. Survey responses revealed that the greater trust industrial and academic partners reported feeling toward their collaborators, the lower both sides perceived both types of barriers to collaboration to be.
But whatever your collaborative strategies may be, patience and perseverance will also be required. “A message for young people thinking about academic-industry interactions is to be prepared for a long ride, sometimes a ride that has bumps in the road, but to be persistent,” says Waxman. Of his collaboration with Pfizer, he says, “It took a long time. It required the coming together of two cultures, the academic culture and the biopharma culture, but neither of us . . . could have done this alone.”
TIPS FOR SUCCESSFUL COLLABORATION
Focus on reproducibility: To increase the likelihood that preclinical findings can be translated, determine study outcomes and statistical analyses in advance of projects, and publish detailed methods. “It’s very possible that an industry laboratory may not be able to reproduce [a paper’s results] completely because of the lack of enough specific detail,” says McGill University ophthalmologist Leonard Levin. “Don’t just say, ‘This is what we did.’” Instead, “say ‘This is how we do it.’”
Ashley P. Taylor is a freelance writer and science journalist in Brooklyn, New York. She has written for Yale Medicine, the Yale School of Medicine alumni magazine, since 2014 and Medicine@Yale, the medical school’s newsletter, since 2017.
March 16, 2018
I would add that it is important to make the rules of engagement public. How are the IP rights split? What information will be proprietary and what information will be published? Will the results of all trials be published, or only the ones that show effectiveness? What is the company paying for?
Making terms public from the outset helps shield the academic scientist from suspicion of being a company shill. That suspicion will hurt all the parties.