This story originally ran on Aug. 24.
By Tony Fong
Can proteomics make you pretty?
Earlier this month California-based CellWorks Group announced a deal to apply its computational proteomics technology to cosmetic compounds from International Specialty Products, bringing the field into a space it has rarely gone before.
The deal, for an undisclosed amount, is for up to 18 months and will be directed at the evaluation of protein functions in compounds for potential use in anti-aging cosmetics, Pradeep Fernandes, president of CellWorks, told ProteoMonitor last week.
Formed in 2005, CellWorks has proof-of-concept deals with other firms investigating disease-associated molecules. The deal with ISP is the first forged for work in the cosmetics market, though Fernandes said that it could open up a substantial new track of business for the company.
ISP, headquartered in Wayne, NJ, supplies specialty chemicals to a wide range of clients including beauty firms Lancôme and L'Oréal, and companies such as Proctor & Gamble and Unilever, which manufacture and sell beauty and personal care products.
Globally, the cosmetics industry generates about $250 billion annually in sales, but to date, proteomics and the protein sciences have largely been left out of the beauty picture.
Proteome Systems, now called Tyrian Diagnostics, inherited a class of compounds used in anti-aging products by Estée Lauder and L'Oreal when it purchased biopharma Eukarion in 2005, though it's unclear if Tyrian is continuing to develop those compounds. And New Jersey-based firm Proteoderm, a subsidiary of Proteonomix, uses protein-based technology to develop anti-aging skin products.
While these examples have been the exception to date, Fernandes said that proteomics may be poised to make a bigger push into the cosmetics market, particularly as the industry looks for alternatives to animal-based testing.
"This whole cosmetics angle is pretty big," he said. "That industry [is becoming] very sensitive to the whole animal testing [controversy] because of [People for the Ethical Treatment of Animals] … so the more that you can do virtually, the better for everybody."
Before signing the current deal with ISP, CellWorks carried out proof-of-concept work using its technology, broadly called Virtual Cells, to predict the effects of two ISP biofunctional ingredients on skin pigmentation under specific environmental conditions. According to CellWorks, in that part of the project it was able to assay hundreds of biomolecules and examine different application rates, allowing ISP to predict the effects of the ingredients on different skin types. ISP validated CellWorks' results in its own laboratories.
In this next phase, the focus will be on predicting the biofunctions of chemicals for potential use in anti-aging products.
In a statement, Claude Dal Farra, vice president of R&D for ISP, said the Virtual Cells technology could potentially and "dramatically" accelerate the discovery and development of biofunctional ingredients and improve their efficacy.
The ability of the platform "to perform complex experiments with easy setup and to analyze hundreds of biomolecules and skin attributes gives us unprecedented insight into the workings of biofunctionals we have under development," Del Farra said. "Working with their virtual skin platform in order to model the interaction of the melanocyte, keratinocyte, and fibroblast is a tremendous advantage as the co-culture equivalent would require primary cell lines that last only 48 hours at best, imposing high costs and extreme time pressures."
Modeling Molecules of Beauty
CellWorks' Virtual Cells technology is borrowed from technology used in the electronics field where Fernandes once worked. Before starting up CellWorks with CEO Taher Abbasi and Chief Scientific Officer Shireen Vali, he founded a company that developed software to model how electronic chips work.
Building a chip can costs millions of dollars and consume considerable time, so before building the chip, hundreds or even thousands of experiments are conducted virtually to ensure the systems behave correctly. Using the same concept, he and his CellWorks co-founders now apply that technology to the life sciences.
The Virtual Cells platform, Fernandes said, "leverages" the experimental data created by other researchers to determine "the mathematical relationships between the proteins that are involved in those phenotypes in those cell types. And then we [get the] computational engine to [predict] out how these reactions will change due to some other stimuli, or response, or inhibition over expression — how a drug would react with it potentially."
[ pagebreak ]
"So we're making virtual diseases by putting together virtual cells and those virtual cells contain functional interactions of proteins, and all this is done mathematically," he said.
Using virtual models in the life sciences, however, has proven to be "a lot harder" than doing the same for the electronics field "because Nature has built all the systems … and we're kind of piecing together the puzzle," he said. "Most of our efforts have been in trying to model things to a degree [where] we can have a predictable system.
"The tough thing about life sciences is that everything is so empirical, but if you could do a lot more virtual and theoretical analyses about drugs, or even diseases, then you can focus and hone in [on] what experiments you want do physically, or understand why something is working the way it is, which is pretty much how people do it in engineering."
Clients such as ISP provide CellWorks targets they are interested in and CellWorks runs the targets through its model. The company is able to use its models to provide data because every protein it has modeled is assayable, Fernandes said, adding that CellWorks has about 6,000 assayable proteins.
"In proteomics, one of the pretty tough things is every protein that you want to assay costs you money, and you need [significant sample size] to extract that protein, whereas for us there is no cost to assay anything," he said.
Because Virtual Cells incorporates other people's data, CellWorks continually needs to update the model. As it does so, Fernandes said, the company is, in effect, validating the results of these new experiments.
For example, two similar experiments may result in completely contradictory data. When that happens, "we put both sets of data in our system and we see which one … actually produces an accurate, plausible phenotype because we're checking it in the context of a much bigger picture," he said. "So in some sense, we're actually validating some papers."
At the same time, though, its technology also needs to be validated in order to become commercially viable, and in research conducted with other partners leading up to the ISP announcement, the focus has been on proving the utility of Virtual Cells.
Commercial partners on such work include a top-10 pharma and a mid-sized drug firm. Fernandes declined to identify either one. CellWorks is also collaborating with academic institutions such as the University of California, Davis; the Dana-Farber Cancer Institute; and the Indian National Institute of Mental Health and Neuro Sciences.
"We're doing these deals just to build credibility because in some sense what we're trying to do is very ambitious, so we're just doing this to get them to validate our system," he said.
Now, that the company feels that it has made some progress, the next step is to go after partnerships for novel discoveries and the design of new therapeutics. CellWorks' business model is to work with other firms on a classic partnership basis in which it will share any profits realized as well as receive fees.
The company, Fernandes said, wants to avoid offering the technology on a fee-for-service basis. He and his partners also currently do not plan to commercialize Virtual Cells as a stand-alone product for sale.
CellWorks has approximately 80 employees, most in Bangalore, India, where its R&D operations are based. The molecular biological work is conducted there while building and reconfiguring the computational engine and its business operations are based out of the Saratoga, Calif., headquarters.
In the true spirit of many start-ups, Fernandes, Abbasi, and Vali self-funded the company to get it up and running, then got additional financing from friends and family. About three months ago, CellWorks closed on its first venture capital round. Fernandes declined to say how much money was raised in the round or the name of the VC firm.
The money will be used accelerate internal programs — "to validate some ideas that we have," Fernandes said — and fund research on the Virtual Skin platform. In addition to the anti-aging work for ISP, CellWorks is targeting partnerships in cancer, especially those directed at research into lung cancer and colorectal cancer, and inflammation, in particular rheumatoid arthritis, Fernandes said.