NEW YORK (GenomeWeb) – A Bay Area company known for using its proprietary protein-engineering platform for industrial applications is turning its attention toward the in vitro diagnostics market.
Specifically, over the last year or so Redwood City, California-based Codexis has been using its platform, called CodeEvolver, to engineer high-performance enzymes for customers using next-generation sequencing and PCR in IVD applications.
Codexis has already developed an enzyme that plays a key role in library preparation steps for NGS. In internal testing, isolated from the rest of the workflow, the enzyme has provided a "multiple-fold" improvement in performance over comparable commercially available enzymes, company executives said.
The enzyme is now in the hands of beta testers who are evaluating it in their NGS workflows to see if these performance improvements can translate to an increase in analytical efficiency, processivity, and fidelity for DNA diagnostics.
"The data that we have … looks really good and we think we have something," said Chris Claeboe, director of custom manufacturing and product management at Codexis. "But, the ultimate determining factor is providing it to beta testers and saying, 'Hey this is what we think we've got; tell us how the results factor into your system. Assuming they're as good as we think they are, let's talk.'"
Codexis was founded 15 years ago specifically to engineer novel proteins that would perform as catalysts for more efficient chemical manufacturing, specifically for the pharmaceutical industry. The company's enzymes have since helped enhance the manufacture of a number of well-known pharmaceutical products such as Lipitor, Zocor, and Januvia, Codexis CEO John Nicols said.
"When we created the enzymatic catalyst for reducing the cost of manufacture of Lipitor, that was about 10 years ago, and that required about 20 people and two years," Nicols noted. Today, he added, developing a similar novel protein would require "a team of about four scientists about a half a year," and this is thanks in large part to the CodeEvolver platform.
"We have invested in the range of a half a billion dollars … throughout our 15-year history towards the use of and refinement of this protein-engineering toolkit," Nicols said. "That's led us to build proprietary software algorithms, increasingly artificial intelligence-based, to do predictive sequencing of amino acid protein sequences. That is proprietary, and it's enabled us to … create new, novel proteins that hit performance targets that heretofore could never be hit."
CodeEvolver is covered by some 250 patents and patent applications that cover software, the use of NGS in the protein engineering workflow, automation, enzyme isolation methods, and high-throughput testing, Nicols said.
A large proportion of the CodeEvolver workflow is bioinformatic in nature. The process begins with in silico screening, in which company scientists computationally test the docking of substrates into static and dynamic models of enzymes. They then employ proprietary data-mining methods to create protein sequence-activity relationship models from experimental testing of enzyme libraries, enabling the rapid assessment of thousands of mutations in context and accelerating protein improvement. The company also uses a computational methodology to determine the additive effect of individual mutations.
Business has been going well. Earlier this month Codexis announced full-year 2016 revenues of $48.8 million, a 17 percent increase over 2015 and the third consecutive year that the company has met or exceeded annual guidance targets to deliver double-digit revenue growth. The company also provided revenue guidance of $50 million to $53 million for 2017. Some of this revenue growth will be driven by licensing deals for CodeEvolver, but the company also expects its product sales to grow between 37 percent and 50 percent this year.
Beyond that, Codexis believes that its latest foray into the IVD market will be a multi-million dollar revenue contributor by next year, Nicols said in a conference call recapping earnings earlier this month.
"There is a large industrial enzyme world that's dominated by … companies like Novozymes and Dupont," Nicols said last week. "We've been reaching out and calling the customers of those great competitors and asking them: 'The enzymes that you're buying from those companies, where are they falling short of either current or future expected performance needs?' Those markets cover things like detergents, pulp and paper, animal feed, and diagnostic enzymes."
Nicols said that many potential customers indicated that enzyme performance is not keeping up with the desire and need of IVD developers to accurately detect increasingly challenging targets, one of the most obvious examples being trace DNA in liquid biopsy applications.
"We started to do research on what kind of enzyme properties might enable that sample and library prep workflow to get a higher-fidelity, lower-bias diagnostic result," Nicols said. The company remains tight-lipped on specifics for competitive purposes, but essentially it took a core enzyme that's frequently used in the NGS process now and improved upon it using CodeEvolver.
"Basically, we're taking that core enzyme, and we're changing amino acid positions along that backbone," Nicols said. For instance, the company might change 10, 20, or 30 positions along the backbone of amino acid chain lengths of 300 to 500.
"This particular enzyme is naturally available, and … our approach is to very agnostically and rapidly change … up to 10 percent of that backbone," Nicols said. "That's a lot of optionality, and that's what enables these performance benefits."
Because Codexis uses NGS in its own CodeEvolver workflow, it has been able to validate the performance of its enzyme and improve the efficiency of its own workflow. "Now we're targeting other people's workflows," Nicols said.
One beta tester is an executive at a leading liquid biopsy test developer who spoke on condition of anonymity due to competitive reasons. He said in an email that Codexis is a "leading company in enzyme optimization so an obvious partner for this type of exploration."
"We are always seeking ways to further optimize our NGS process and maximize the information we can gain from patients' precious samples," the executive said. "The enzyme in question is a key component of the process and so we are excited to explore if there are ways to improve efficiency."
The company is currently testing the enzyme in standardized assays and comparing its performance against existing versions, he said, adding that it is still too early to comment on whether it is increasing performance of the NGS workflow.
Codexis has supplied other beta testers with the enzyme to conduct similar trials, and it is already eyeing additional workflows that play a prominent role in the IVD space, such as qPCR.
This places Codexis squarely in competition with some well-established biological reagent suppliers such as New England Biolabs, Promega, and Takara Bio, and as such the company is currently building out its sales and marketing force to navigate this relatively uncharted territory.
"There's been a fair amount of consolidation in the space — Kapa Biosystems was acquired by Roche, and at least part of their offering was enzymes for the NGS workflow. Not too long before that, Enzymatics was acquired by Qiagen. So there seems to be fewer and fewer independent third-party marketers, so we see kind of a nice little demand pull for another potentially independent player in this channel," Nicols said.
Point-of-care testing is also an obvious market where better-performing enzymes for nucleic acid detection might make an impact. Essentially, the company is targeting any life science market where scientists are trying to accurately detect or measure low-quantity or low-quality DNA. This could eventually encompass areas like forensics and CRISPR-based gene editing, but for now Codexis is keeping its eye on the IVD ball.
Claeboe noted that the company first began putting its feelers out at the American Association for Clinical Chemistry meeting last summer, "and we were simply asking the question: Would engineered enzymes in your workflow be beneficial? Quite frankly, we found very few people that were actually thinking about it in that way. We found a lot of effort towards … platform adjustments; people trying to improve the technology [through] the actual machine."
However, he added, "we didn't find a lot of people working on improving the performance sensitivity, reducing bias of the enzymes, and I think that's where our core value resides. Based on some of the conversations we've had with beta testers, there seems to be a high level of interest."