NEW YORK – In late January, James Crowe, an antibody researcher at Vanderbilt University, was gearing up for a test run to develop potential antibody treatments for a virus outbreak.
The goal: do it in just sixty days, a mark set by set by a Defense Advanced Research Projects Agency (DARPA) program that has funded Crowe and three other labs around the country.
Crowe's lab had come close in a test scenario last year, where they were able to generate antibodies against Zika virus and show they were effective and safe in primates in just 78 days.
Like their 2019 run-through, the 2020 one was supposed to be a mock exercise. But as the SARS-CoV-2 virus started showing up in the US — the first case, from Seattle, was confirmed with an RT-PCR test on Jan. 20 — Crowe got word from his DARPA program officer that this was no longer just a drill.
Crowe said his lab received a sample from US patient number one just over a week after they were confirmed for COVID-19. And while it has been more than 60 days since then, the Vanderbilt lab worked with astonishing speed to find antibody therapy candidates that are already being developed in partnership with two companies, IDBiologics and AstraZeneca. Ology Bioservices, a Florida-based Department of Defense contractor, has also received $14 million to develop and manufacture antibodies discovered by Crowe's lab and other investigators. Two other companies are investigating licensing antibody sequences from Vanderbilt, but Crowe declined to disclose them.
"Microfluidic single-cell sorting and high-throughput NGS. Those are really the two enabling technologies that have allowed this," said Amy Jenkins, the DARPA official overseeing the Pandemic Prevention Platform (P3) program. Other antibody discovery methods, including hybridomas, could take a year or more to generate a drug candidate, she said. Crowe added that concepts from synthetic biology and protein engineering have been crucial to his lab's success. "All this stuff is coming together all at once. It enables us to go really fast and work at scale," he said. "Because it's a stochastic process, the more searching you do the better [antibody] clones you get."
Crowe credited several specific genomics and synthetic biology firms and technologies with enabling his lab to complete its work so quickly. 10x Genomics and cell sorting firm Berkeley Lights both went out of their way to make sure Vanderbilt had the necessary instrumentation to isolate antibody-producing B cells, he said. And next-generation sequencing with technology from Illumina and Pacific Biosciences helped determine the sequences needed to recreate the genes coding for those antibodies, which the lab accomplished using Twist Bioscience's synthetic DNA platform. The lab has also incorporated automated liquid handling from Beckman Coulter and data analysis using Illumina's BaseSpace Sequence Hub.
"For COVID-19 work, we did 10 or 12 different workflows in parallel," Crowe said. "We did many different things. Some of them worked and some of them didn't work. It's sort of like improvisation."
The basic process involves winnowing all the cells available in a blood sample down to only the antibody-producing memory B cells that are of interest, which comprise about 5 percent of white blood cells. Only one in 10,000 B cells generate antibodies for a virus of interest.
Because single-cell analysis is still expensive, the lab often pre-enriches the B cell suspension with flow cytometry. The lab uses either 10x's Chromium single-cell analyzer with the Immune Profiling kits or Berkeley Lights' Beacon instrument to isolate cells for NGS analysis. For the rapid response platform, sequencing was done on Illumina's platform, although Pacific Biosciences' single-molecule, long-read platform is also used at certain stages. With antibody-generating sequences in hand, the lab uses Twist Bioscience's synthetic DNA platform to generate thousands of constructs to then express in cells that generate the antibody proteins. "At some point, we mix antibody with virus protein and that's when we really know what's going on," Crowe said.
To respond to the coronavirus, the Vanderbilt lab worked with both 10x Genomics and Berkeley Lights to get instruments they needed. When they started, they only had one Chromium instrument, which was situated in a biosafety level-2 lab. To work with the virus, they needed to move that instrument to a BSL-3 lab and get another one for the lower-risk work.
Since last year, 10x has been prevented from selling Chromium instruments that work with older GEM microfluidic chips due to a federal court injunction issued as part of a successful patent infringement suit brought by Bio-Rad Laboratories. The Vanderbilt researchers had validated their workflows on these older chips.
"I said, 'Oh my gosh, we do not have time to rework everything,'" Crowe said. By mid-February, 10x had obtained permission from both Bio-Rad and the court to provide a legacy Chromium instrument.
Similarly, Berkeley Lights worked closely with the lab to get their Beacon instrument into their workflows. "We had been planning to collaborate to develop virus inhibition assays, but we didn't own a Beacon instrument," Crowe said, "and it was not possible to move our research to California." Berkeley Lights brought a loaner instrument to Nashville and sent application specialists to stay on site to work with the Vanderbilt team.
That platform can put live cells into "pens" and observe them for a short time. Crowe's team was able to watch as cells secreted antibodies that subsequently blocked SARS-COV-2 spike protein that was added to the pen. "That was very novel to be able to do that," he said. "We had to work that out on the fly right in the middle of the pandemic."
Most of the NGS analysis is done with Illumina, especially when looking at only the variable region of an antibody sequence. But sometimes the lab wants to see polymorphisms in the genes and antibody allotypes, which require PacBio's long read capabilities. That platform is also useful for "antibodies we know we like, to find the authentic sequence we want to manufacture and use forever," Crowe said.
Crowe also noted that Twist's DNA synthesis was "an essential component" of the workflow. (Twist has its own COVID-19 antibody discovery program and announced earlier this month that it has found competitive antibodies to SARS-CoV-2 as well as antibodies that bind to the human ACE2 receptor that the virus primary uses to infect cells.)
Though Vanderbilt's discovery run was successful, there is still room for improvement. Single-cell technologies return anywhere from 10 to 60 percent of the antibody genes. And while not a technological limitation, because the body sometimes takes several months to develop a robust immune response to a virus, the most promising antibodies came from patient samples that the lab didn't receive until mid-March, Crowe said. But that also means it took just five or six weeks before the lab was able to pass the baton to clinical trial researchers. Now, his lab is supporting their partners bringing the antibodies through clinical trials, although they may do another round of discovery for COVID-19 soon.
Vanderbilt wasn't the only success story, P3 program manager Amy Jenkins said. "Some may have been a little bit ahead of the others, but they all generally discovered antibodies in approximately the same time frame." That includes AstraZeneca, which also has its own discovery platform; Vancouver-based antibody therapy company AbCellera, which has partnered with Eli Lilly to develop COVID-19 antibodies; and the Duke University School of Medicine's Human Vaccine Institute. The US Government retains rights to use the technologies funded by DARPA, but any therapies that come out of P3 will be available to both civilians and the DOD, Jenkins said. Moreover, the DOD contract with Ology Bioservices is "a way for us to ensure we have access to some of these products," she said.
Ultimately, Jenkins would like to see genomics play a role in delivering the antibodies into the body, via genetic constructs. "That's incredibly difficult, maybe more difficult than finding them," she said. "Hopeful we'll succeed."