A lab space of one’s own
The QB3 Garage: An incubator for innovation
California Historical Landmark No. 976 is a mythical place for entrepreneurs. Located at 367 Addison Avenue, Palo Alto, CA, it is home to the garage in which Bill Hewlett and David Packard developed HP’s first product, the Model 200A audio oscillator. The garage is not only the birthplace of Silicon Valley, but also a symbol for innovation and the Californian entrepreneurial spirit.
For entrepreneurs in the life sciences, developing ideas into innovative products requires more than what can typically be found in a backyard garage – they need lab space that’s suitable for performing experiments, and in compliance with environmental health and safety regulations. For emerging companies with limited financial means, this space is hard to find in commercial real estate: the minimum unit of lab space that can be rented is around 2,500 square feet, too expensive for most beginning entrepreneurs to put on their credit cards.
When Regis Kelly and Douglas Crawford joined the California Institute for Quantitative Biosciences, or QB3, they identified the space issue as one of the main barriers between great scientific discoveries and innovative products that reach the marketplace. They decided to start a tiny incubator of ~2,500 square feet at UCSF, which they called the Garage to commemorate HP’s humble place of origin. The smallest unit that can be rented at the Garage is a single lab bench, equivalent to ~120 square feet, which in many cases is sufficient for carrying out proof-of-principle experiments to get the company off the ground. The idea was initially met with skepticism by venture capitalists: “Some of them said don’t bother, this is a recipe for mediocrity, an intensive care unit for small companies that will not amount to anything,” recalls Crawford. “But we proceeded because it was QB3’s strategic goal to promote great science and to help enrich our society. We believe that basic research will lead to economic growth, but if we don’t help it move through the final mile, to get the discovery to the marketplace, we are not meeting our social contract.”
The success story of the QB3 Garage’s first tenant, Fluxion Biosciences, proves Crawford right: founded in 2006, the company moved to South San Francisco in 2008 and now has 30 employees. “When you go there, it’s exactly what people hope for from the science in our universities. Now there’s a small factory in South City, hiring high school graduates to manufacture microfluidics devices. It’s the full impact – it’s jobs, it’s cool research tools that will drive future discoveries, and it is the realization of the potential of laboratory research. It showed us that it is possible to start with very little, a tiny amount of space, and produce a company of great value,” says Crawford.
From postdoc to entrepreneur
The idea for Fluxion Biosciences was born in Luke Lee’s lab in the Department of Bioengineering at UC Berkeley when Cristian Ionescu-Zanetti, a postdoc at the time, became interested in working outside of academia. He enjoyed his research, but felt that in the academic environment he was “taking things maybe a fifth of the way towards something that really works, a product that’s better than the status quo.” Together with a graduate student in the lab, he applied for a Small Business Innovation Research (SBIR) grant, entered business-plan competitions, and eventually became the first company to move into the Garage at UCSF. “They came and knocked on our door when we were still planning; we had dedicated the space, but we hadn’t even started to get the approvals from the university,” Crawford says. “In the end, their inquiry for space precipitated it all; it was a nice synergy between us and Fluxion.” While starting the company at the Garage, Ionescu-Zanetti continued working as a postdoc half time, but soon devoted all his efforts to the company.
Fluxion currently markets two products, called IonFlux and BioFlux. “Technology-wise our focus has always been to take labor-intensive processes, such as drug screening, and parallelize and automate them, to make them faster, better, and cheaper,” explains Ionescu-Zanetti. The IonFlux automatically records the flux of ions through membrane channels without the need for intermediate user intervention. The machine uses standard-format multi-well plates commonly used in high-throughput screening, and can be readily adapted to existing screening platforms. Pharma companies and academic labs use the IonFlux for screening the effect of drugs on membrane channels, as well as for characterizing the consequences of mutations on ion channel currents. Conceptually, the IonFlux was Fluxion’s first product, but after talking to potential customers, the company soon started developing its second product, the Bioflux, which allows researchers to perform live-cell assays under shear flow. For a variety of applications, such as studies of platelet adhesion that naturally occurs in the blood stream, the BioFlux mimics the physiological environment much better than traditional assays. The BioFlux instrument is used by scientists in both academia and industry for a number of cell-based assays, ranging from wound-healing research to studies of bacterial microfilms. “The rewarding moments came after our first instruments went into pharma companies, and their people came back and said ‘This is much better than what we were doing before.’ They were really excited and even wanted to publish papers with us,” says Ionescu-Zanetti.
Once Fluxion had moved into the Garage, the remaining space filled up quickly. “It’s been full ever since it opened, and we get one to four inquiries a week from nascent companies looking for space,” says Crawford. The increasing demand for space prompted Crawford to expand capacity, which led to the creation of the QB3 Garage/Innovation network that now comprises four incubators: the original Garage at UCSF, the Garage at the UC Berkeley campus in Stanley Hall, the QB3 Mission Bay Innovation Center, and finally the QB3 East Bay Innovation Center, which opened in July 2011 – and there are already plans for adding the next incubator to the network.
Getting funding in tough times
Allopartis is one of the companies that started out in the QB3 Garage at UCSF and then moved into the Mission Bay Innovation Center, which is now home to more than 20 start-ups. It was cofounded by three former students from Richard Mathies’ lab in the Department of Chemistry at UC Berkeley: Robert Blazej and Nick Toriello, who graduated from the joint UCSF/UC Berkeley Bioengineering graduate program, and Charlie Emrich, a biophysics graduate.. One of the hardest things about starting the company was getting funding. “We were founder-financed at the beginning for about 8 months, which meant that all three of us went almost totally broke before we got it funded,” says Emrich. Meeting venture capitalists could sometimes be a surreal experience for someone who had just gotten out of graduate school: “Here we were in our beaten down cars, driving down to Menlo Park where all the venture capitalists work, parking in between a Maserati and a brand new BMW,” says Emrich. Allopartis eventually got funded right around the time when the market crashed, which forced them to be creative with the resources they had. They used the money to prove their core technology, the AlloScreen, and have attracted further investments, including grants from the Department of Energy, ever since.
The AlloScreen employs the principles of directed evolution and a unique selection system to generate enzymes with optimized properties, such as activity or stability. While natural evolution happens on a timescale of millions of years and relies on spontaneous mutations to create proteins with altered properties, directed evolution in the laboratory accelerates this process by artificially creating a library of many variants of the original DNA gene that encodes the enzyme. In the AlloScreen, each of the DNA variants is then attached to a substrate particle, and emulsified with the contents of a cell-free expression system in order to produce the many different protein variants that are encoded in each DNA variant. If a particular protein variant is active, it will be able to digest its substrate particle and release its coding DNA, which can subsequently be separated by centrifugation from all the inactive variants, which will stay attached to the bead. The information obtained by sequencing the released DNA molecules is the basis for the further characterization of the altered proteins they encode. Emrich and his colleagues have used the AlloScreen to improve the activity of cellulases, enzymes that digest cellulose. “Cellulose is the most abundant biopolymer on earth,” explains Emrich, “it is a linear chain of glucose molecules, and these chains are magically very crystalline, not very soluble, and very recalcitrant, so they do not break down easily.” Because of these properties, cellulosic enzymes are the holy grail in the making of biofuels. While glucose can relatively easily be fermented into ethanol, breaking down cellulose into glucose is the rate-limiting step. With improved cellulases, abundant and renewable resources such as agricultural waste and non-food crops could be used for the production of low-emission biofuels that could substitute fossil fuels and lower greenhouse gas emissions. Although getting there will be a long journey, scientists at Allopartis have already created cellulase variants with improved activity and are optimistic. “We’re now getting some commercial traction for those variants,” says Emrich. In collaboration with Louise Glass in the Department of Plant and Microbial Biology at UC Berkeley, they are now working on co-evolving cellulases with engineered strains of the cellulolytic fungus Neurospora crassa to better understand the activity profiles of different types of cellulases.
Crossing the valley of death
Allopartis has successfully crossed what is known among entrepreneurs as the first “valley of death” – a gap in funding opportunities for projects that go beyond the academic research that the NIH will fund, but are not yet at a stage of maturity where they can attract commercial funding. Once a start-up has obtained minimal funding to bridge this gap, the next big challenge usually lies in finding affordable equipment. Omniox, a current start-up at the Garage at UCSF, was lucky in that the economic crisis worked in their favor. “Many biotech companies were going out of business at the time and sold their equipment, often at bargain prices. We had the deal of a century—we easily got a million worth of equipment for $30,000,” recalls Stephen Cary, co-founder of Omniox.
The company is developing a molecular carrier that will deliver oxygen to hypoxic tissues, areas of the body that are starved of oxygen such as tumors. The first protein that comes to mind for this purpose is hemoglobin, the protein that transports oxygen from the lungs to all other tissues in the body. However, government agencies and companies have tried for decades to develop hemoglobin into an oxygen transporter that could be used as a blood substitute, with no success: taken out of red blood cells, hemoglobin scavenges nitric oxide, with devastating effects for the body.
It was during his final days as a graduate student in Michael Marletta’s lab in the Department of Molecular and Cell Biology at UC Berkeley when Cary had an idea that seemed too good to be true: “I was reading up on all the efforts around developing hemoglobin as an oxygen delivery therapeutic, reading paper after paper about how the FDA was rejecting it for trauma and surgery because of the toxicities, when I suddenly remembered a group meeting from a few weeks before where Elizabeth Boon, a postdoc in the lab, had presented on a protein that didn’t have very much nitric oxide reactivity, and I thought wait a minute, maybe this is a much better platform, because it’s a stable gas sensor, rather than being a gas reactor.” The protein is part of the heme nitric oxide/oxygen binding family, or H-NOX. Like hemoglobin, it uses heme as a cofactor, but subtle differences in the coordination geometry result in very different oxygen-binding properties. Cary presented his idea to Jonathan Winger, a postdoc in the lab, and together with Marletta they successfully applied for a Rogers Bridging-the-Gap Award for translational research at QB3. To develop their idea further, they founded Omniox and Emily Weinert, a postdoc in the Marletta lab, created and characterized more variants of the protein with a range of oxygen binding affinities. Under Cary’s leadership and supported by a National Cancer Institute SBIR grant, the company then moved into the Garage at UCSF.
The H-NOX protein could potentially be used as a therapeutic in many different diseases that are associated with hypoxia. Potential applications include treating stroke, managing sickle cell pain, and wound healing. Although Cary considers branching out, the company currently focuses on overcoming hypoxia in tumors. “Hypoxia is a huge driver of tumorigenesis and metastasis,” explains Cary. If the growth of blood vessels can’t keep up with the growth of the tumor, large regions are starved of oxygen. Those regions are hard to target using conventional therapies such as radiation, which relies on the damaging effects of reactive oxygen species. In addition, cells in the hypoxic regions usually become more aggressive as a consequence of being starved of nutrients, energy, and oxygen, and tend to form metastases in other parts of the body. The goal of Omniox is to improve existing cancer therapies by bringing oxygen to the tumor. Initial studies using a mouse model show that the protein very efficiently travels from blood vessels into the tumor tissue to deliver oxygen to previously hypoxic areas – a phenomenal success that was rewarded with a $3 million SBIR Phase II Award from the National Cancer Institute. Omniox is now eligible to apply for a SBIR Bridge Award given they can secure matching funds from private investors, which would add another $6 million to their budget. These funds will pay they costs of optimizing a lead candidate that will then be used for pre-clinical studies to determine its efficacy and its toxicity profile in animals. If all goes well, H-NOX will be ready for clinical studies in about two to three years. The company will then face what Crawford calls “the second valley of death.” QB3 has worked hard towards bridging the first valley of death by providing mentoring, funding, and lab space to start-ups, but bridging the second valley of death, the gap between pre-clinical and clinical studies, poses further difficulties: “The enterprise back at the discovery end is expensive, but the cost in the clinic dwarfs that,” explains Crawford. Cary estimates the costs for phase I and II studies at around $15 million; getting H-NOX to the market through phase III studies will add another $50 to 100 million. But, given Omniox’s latest results, it seems likely that they will find investors willing to pitch in to help Omniox in helping cancer patients.
From bench to business
One of Omniox’s neighbors at the Garage at UCSF is Refactored Materials, a start-up that works towards the synthetic production of spider silk. Spider silk is a material of phenomenal strength, lightness and flexibility that outperforms all man-made materials and could potentially be used in applications ranging from lightweight and durable clothing to biomedical applications such as artificial tendons. A big challenge for commercialization is the production of spider silk: “Spiders can’t be farmed, they’re territorial, they will attack each other, eat each other, and no one has been able to make spider silk recombinantly on a commercial scale,” explains David Breslauer, a graduate from the joint UCSF/UC Berkeley Bioengineering graduate program. Together with two other graduates from the same program, Dan Widmaier and Ethan Mirsky, Breslauer co-founded Refactored Materials and has been a tenant at the Garage at UCSF since May 2010. They decided to use yeast cells for the recombinant production of the large silk proteins, and have already produced enough silk protein to try to make fibers. “Fibers are generally either melt-spun, meaning that you melt a polymer, extrude, and cool it, or wet-spun, meaning that you dissolve a polymer and extrude it into a non-solvent that coagulates it,” explains Breslauer.
In contrast to many other emerging companies, Refactored Materials was funded from the beginning: they got their first grant right when Breslauer graduated. Although the company is now minimally funded through federal and state grants for several years, they’re still looking for additional sources of money: “We’re moving faster than those grants can support,” explains Breslauer. “It’s nice to have them, but it’s not necessarily something to rely on.”
While the first six months at the Garage felt very similar to working in an academic environment for Breslauer, the mindset changed once they started to work more on developing the business aspect. “You suddenly stop caring as much about publications, you’re just trying to make something that really works, rather than understand every little detail about it,” Breslauer remembers.
Strategic partners to finance growth
Silicon BioDevices was the second company to move into the Garage at Berkeley after it opened its doors in summer 2010. The company is developing diagnostic devices that are based on digital microchips and can detect tiny amounts of specific proteins in a liquid sample such as blood. The ease of use combined with high sensitivity and low costs – the single-use device will be available for $1.50 – sets them apart from the bulky and expensive analyzers that are currently available on the diagnostics market.
Once a drop of whole blood is applied, a membrane at the top of the device separates red blood cells from the plasma. The plasma then solubilizes antibody-coated magnetic particles on the back of the membrane, allowing them to bind the protein to be detected and to bind a secondary antibody. Unspecifically bound particles are removed by magnetic forces, and the remaining particles are detected by the chip. After the signal is read out and processed, the test result can be sent directly to the physician’s cell phone by the wireless transceiver that’s integrated into the device. A significant advantage with regard to safety is the self-testing capability of the device: “The sensor can control the assay by making sure it’s run correctly, in a timely way, and you can disable it if it has been compromised in any way,” explains Silicon BioDevices’ co-founder Octavian Florescu, a graduate from Bernhard Boser’s lab in the Department of Electrical Engineering and Computer Science at UC Berkeley.
The user-friendly design might eventually allow for diagnostics at home, but for the near future Florescu hopes that the device will find its way into emergency departments and physician’s offices, rendering time-consuming laboratory testing obsolete. “The number one reason doctors don’t perform in-office testing is because it requires extra time and extra staff,” says Florescu. The device would be the first highly sensitive diagnostic tool that could be integrated seamlessly into the physician’s workflow.
Although initial results are promising, it might still be another three years until you encounter one of Florescu’s chips at your local doctor’s office. Developing the final prototype and making it manufacturable will take approximately two years before the device is ready for approval by the FDA, which might then take another year.
The company has raised money from business plan competitions, but is financed out of Florescu’s pocket for the main part. In order to finance further development and production, they’re now approaching life science investors and strategic partners. “It’s a very slow process; even if you have a great technology, you have to add another 12 to 18 months if you are to strike a good deal,” explains Florescu.
Judging from the history of its predecessors at the Garage, it seems likely that Silicon BioDevices will be able to close a deal: out of the first six companies that started at the Garage at UCSF, four closed venture financing rounds and a fifth was acquired by Affymetrix for $25 million. “There’s now one very wealthy 28 or 29-year old after starting a biotech company at the Garage,” says Crawford, and he adds: “We’re not promising that for everyone, but it’s nice to know that there is at least that possibility.”
Starting a start-up
Contrary to what one might think, the high success rate is not based on an evaluation of the commercial potential of the companies by QB3. “We don’t want to be rigorous in the evaluation of the market opportunity of what they brought to us, we want to be rigorous in our evaluation of the people. Most real innovations are diamonds in the rough: over and over again, we don’t see it when it comes. If you have good people, you get to the right conclusions most of the time, and we want to help them grow as quickly as possible.”
Cary’s advice for nascent entrepreneurs? Just do it! “You can be a postdoc with an idea, and you can start a company. For $1,200 you incorporate your company, then you take your science idea and submit it as a six page SBIR grant, and eventually you get half a million dollars.” Although the process might not always be so smooth, starting a company is a rewarding experience, says Crawford: “I do not know of a single case where an individual regretted their decision. All admit that it’s the hardest thing they’ve ever done, stressful, but satisfying in a way that then exceeds their expectations.” So what are you waiting for?