NEW YORK (GenomeWeb) – Antibody firm Abcam has developed a new next-generation sequencing-based workflow that it said could allow it to produce reagents to a variety of difficult to target proteins.
The method uses NGS to more thoroughly explore the antibody repertoire produced by rabbits than is possible in conventional rabbit hybridoma workflows, allowing the company to isolate and analyze molecules in what John Baker, Abcam's vice president, diagnostics and licensing, termed "the long tail of the immune response."
The hope, he noted, is that in being able to explore the full diversity of the animal's immune repertoire, researchers will be able to better identify antibodies with the precise properties they are looking for.
Antibodies are traditionally made by injecting a rabbit with an antigen of interest and then harvesting antibodies produced by the animal's immune response to the target antigen. These antibodies are then screened against the target to identify effective reagents that can then be further refined.
A limitation of traditional hybridoma or B cell cloning approaches, Baker said, is that they are biased towards the most abundant B cell clones produced in response to the introduction of the target antigen. And while effective antibodies can often be found among these populations of abundant clones, they aren't necessarily the best reagents to a particular target.
The rabbit immune response is notable for its breadth, Baker noted, but, he added, "that breadth is only useful if you can mine it. The challenge we set ourselves is, wouldn't it be great if we could pull out every single responding B cell and be able to analyze that full repertoire, to make sure we're pulling out the right binder."
To do that, the company has combined single-cell microfluidics with NGS to generate a pool of candidate antibodies that includes not just the more abundant clones but all the antigen-responding B cells produced by the rabbit.
First, antigen-responding B cells are pulled out and enriched using the target antigen. This typically leaves a population of around 5 million to 20 million cells, which are then put through a single-cell separation and emulsion followed by NGS to sequence the heavy and light chain variable regions of each cell.
"The initial output from that is hundreds of thousands to millions of paired heavy and light chain sequences," Baker said, adding that Abcam has found that, as anticipated, the method picks up "sequences from right across the abundance range of the B cells in the immune response."
"It's picking up a lot from the very abundant, highly proliferating B cells that you might expect to catch with a fusion-driven, hybridoma-driven process," he said. "But it's also pulling out sequences of B cells that were only there in one or two copies in that input cell pool – and that gives us a candidate pool that is typically one to three orders of magnitude higher than any other platform we've seen."
This, though, raises the question of how to analyze this expanded pool, Baker noted. "The breadth is only useful if you can screen it."
To do so, Abcam researchers developed what he said is "basically a phylogenetic analysis" of the B cells, which produces "a classic looking phylogenetic tree that really maps the way the affinity maturation and the hypermutation in the rabbit has progressed."
Branches of this phylogenetic tree correspond to portions of the target antigen. "So, we see similarities in physical behavior between clones in the same arms," Baker said. And this, he added, allows Abcam or its customers to start their screen by looking at a few clones from all of the different arms, and then follow up with more in-depth screens of clones from the arms that looked promising in the initial round.
"Once you've done that primary screen, then you can say, 'Let's go broad in that arm and let's explore the 10, 20, 50, 100 clones most closely related as we look to pick out the one that has exactly the performance we need,'" Baker said.
Looking more deeply into the immune response has, as hoped, allowed the company to generate antibodies to proteins it has been unable to target in the past.
"We have a number of instances where we have been able to rescue programs that have been through multiple rounds of multiple different platforms, and we have been able to find that rare clone that sticks, where previously there was no clone [that worked]," Baker said, adding that thus far the system has had a success rate of around 70 percent to 80 percent for generating antibodies against targets that have failed in other systems.
The ability to have a more comprehensive picture of the antibody repertoire for a given target also has business implications, he said, particularly in the case of therapeutic antibodies, where having the full set of antibodies that could bind a target "is potentially extremely powerful from an IP perspective."
Abcam is still working to scale its new discovery system, but it is up and running with several initial commercial partners, primarily for therapeutics work. Baker said that given the complexity of the process, it is unlikely it will ever displace traditional rabbit hybridoma workflows for basic research antibodies, but the system could open up a number of targets that have, to date, been largely inaccessible.
The process costs roughly the same as a traditional rabbit hybridoma discovery project, and takes around nine months to a year, which, Baker said, is roughly the same as a traditional workflow, but longer than the accelerated hybridoma process Abcam has developed, which can produce an antibody in around six weeks after the immunization process is complete.
The company is also working to integrate the new NGS-based workflow with technology it acquired through its 2015 purchase of AxioMx.
AxioMx is building a sequencing- and microfluidics-based workflow that it hopes will allow the company to screen massive libraries of antibodies and antigens against each other and significantly speed the development process.