Antibody screening has enjoyed considerable success in finding antibodies that can grasp disease-associated targets and contribute to therapeutic applications. For example, since the mid-1980s, slightly over 100 monoclonal antibodies have been approved for treating various maladies, such as cancer, metabolic disorders, and autoimmune conditions, and for preventing transplant rejection. Also, many additional antibody-derived drugs are in development. But continued success may require that antibody screening expand its reach beyond the low-hanging fruit of relatively simple targets. Consequently, antibody screening is eyeing complex targets such as ion channels and G protein–coupled receptors. To ensure that discovery efforts focused on such highly sought-after molecules won’t turn sour, researchers are getting a boost from better technology and more sensitive assays. For example, researchers are deriving more value from fluorescence-activated cell sorting (FACS).
Combining FACS with a cell–cell interaction system
“FACS is an important technology in the set of tools used to find agonist antibodies,” says Richard A. Lerner, MD, professor of immunochemistry and institute professor at Scripps Research. Several FACS approaches have been developed to generate antibodies against specific proteins. However, these approaches have had limited success in identifying antibodies that can bind to difficult targets such as membrane proteins.
“The biggest limitation is that membranes are very hydrophobic,” advises Lerner. “The resulting extensive nonspecific stickiness makes it difficult to visualize the real signal in signal transduction pathways.”
To bypass the need to purify membrane proteins, Lerner and colleagues developed a whole-cell screening platform based on a yeast cell–mammalian cell interaction system that allows antibodies to be selected against native transmembrane proteins. In this platform, which combines yeast display technology with FACS, the yeast cell and the mammalian antigen-bearing target cell are labeled with two different fluorophores, and yeast cell–mammalian cell complexes are identified by FACs.
Using this approach, Lerner and colleagues were able to direct affinity maturation of an antagonist antibody specific for the human acid-sensing proton-gated ion channel ASIC1a. “A key therapeutic field that will benefit is the identification of antibodies targeting opioid receptors,” Lerner notes.
Investigators in Lerner’s laboratory paired yeast cells expressing a large naïve human library with mammalian cells overexpressing the human GPCR µ opioid receptor, and they identified agonist antibodies to the human µ opioid receptor that bound with nanomolar affinities. These antibodies may avoid the adverse effects that develop after opioid use.
“Cell sorting has become an essential tool, in that if used on a population of candidate antibodies, it allows to one to select the antibodies where the transduction pathway is activated,” Lerner asserts. “Then, PCR can help identify what antibody it is.”
Microfluidics pairs with yeast display and molecular genomics
“The idea behind our approach is that we can run a sample of cells through a microfluidic device that we designed and go very deep into the antibody repertoires,” says David S. Johnson, PhD, CEO and co-founder of GigaGen. The approach developed by GigaGen combines the use of emulsion droplet microfluidics with yeast single-chain variable fragment (scFv) display and molecular genomics. “The challenge,” Johnson explains, “is keeping the heavy and light chains of the antibodies intact at a single-cell level to maintain the functionality of the antibodies.”
Microfluidics is critical for maintaining the endogenous pairing of the heavy and light chains of immunoglobulins, which would otherwise be lost after the bulk lysis of large numbers of B cells. “The antibody that we express is tethered to the yeast cell surface, so one can mix the yeast library with the target of interest, which may be a virus or a cancer cell. The antigen binds to the yeast cells. Flow cytometry can sort the binders, and perhaps those can become candidates,” explains Johnson.
After interrogating the antibody repertoire of unique scFv binders against the human programmed cell death protein 1, PD-1, Johnson and colleagues used FACS to achieve an average 800-fold enrichment. Most antibodies that were part of the repertoires of immunized mice were eliminated by this step, indicating that they are not strong binders to the immunogen. Of the candidates that were expressed as full-length antibodies, nine emerged in cell-based checkpoint blockade assays as potential candidates.
“We can run a few million cells per hour, which allows us to go very deep into the cell repertoire,” asserts Johnson. This inexpensive and time-efficient strategy emerges as a viable alternative to hybridoma-based discovery. “Our method,” he continues, “provides a lot of diversity very quickly, and this constitutes an important part of our discovery efforts, where we are looking for rare antibodies that do interesting things.”
Investigators at GigaGen are using this strategy for some of their infectious disease programs. “We are especially interested in human patients who recovered from a particular infectious disease, as they tend to have some antibodies that are neutralizing against the pathogen,” notes Johnson. “And our strategy provides an inexpensive way to do it.”
One of the advantages in the case of infectious diseases is that target identification is more obvious. However, for cancer research, which is another effort at GigaGen, target identification is more challenging because cancer cells continually undergo profound changes in vivo.
Intricacies of synaptic transmission
“Our strategy is called immunoprecipitation detected by flow cytometry (IP-FCM),” says Stephen E.P. Smith, PhD, assistant professor of pediatrics at the University of Washington and principal investigator at the Center for Integrative Brain Research. “It typically focuses on immunoprecipitating proteins, but it can just as easily be used to screen antibodies.”
One of the advantages of IP-FCM is that it requires quantitatively less protein than comparable experimental strategies. In a conventional western blot experiment, for example, a lot of protein is needed if the protein is to be detectable, that is, if it is to manifest as an observable band. “IP-FCM,” notes Smith, “requires that only a couple hundred of five-micron beads be coated with enough protein to obtain antibody binding.”
Investigators in Smith’s lab use this technology to immunoprecipitate protein complexes and subsequently probe with fluorescently conjugated antibodies for different members of the protein complex. “We are able to interrogate protein interaction networks and see how protein interactions change as a unit, which we think is relevant for cell signaling questions,” says Smith.
Researchers in Smith’s lab focus on signaling at the glutamate synapse to understand how the protein interaction networks downstream of the synaptic input respond to different types of synaptic input that may be disrupted in autism. “This is a different way of looking at signal transduction networks as compared to what has been traditionally done,” declares Smith.
Several hundreds of genes have been implicated in autism, but understanding how they are collectively linked to a single pathway that is dysregulated has been a daunting task. “I like to use the analogy with cancer,” says Smith. Hundreds of mutations were individually linked to many cancer types, and while each patient’s cancer is unique, all cancer mutations feed into a shared cell growth pathway. “Similarly, every patient with autism may represent an ultra-rare disease,” he continues, “but the question is how these different mutations feed into a cell signaling pathway downstream of a synaptic input.”
In unpublished work, Smith and colleagues conducted two versions of the same experiment, one using IP-FCM and one using mass spectrometry. The scientists then compared how well these approaches captured the dynamics of protein interaction networks. “While the two approaches agreed with each other, our technique was able to see 20–30 interactions changing, whereas mass spectrometry captured only 5 interactions,” says Smith.
A proof-of-concept study that validated the sensitivity of IP-FCM measured the amount of huntingtin protein in the cerebrospinal fluid of mice. Although quantitating huntingtin protein is valuable both as a marker of disease progression and as a way to monitor gene silencing in clinical trials—for example, during the use of antisense oligonucleotides—traditional protein quantitation techniques were not successful in detecting the protein in cerebrospinal fluid. The alternative, which would involve the use of brain tissue samples, is not feasible in humans.
By combining microbead-based immunoprecipitation with IP-FCM, Smith and colleagues developed a highly sensitive assay that enabled the accurate detection of mutant huntingtin protein in the cerebrospinal fluid of Huntington’s disease patients and mouse models, raising the possibility that this technology could be applied in research and as a clinical biomarker.
“We rely on antibodies—and they are the best and the most terrible of the reagents that we can have,” says Smith. One of the concerns is whether antibodies are sufficiently specific, and this involves extensive quality control steps. “But every antibody is unique,” Smith points out, “and when a commercial antibody supplier decides to discontinue an antibody, that can be pretty devastating for a research program.”
Antibodies for tricky antigens
“When an antigen is hard to work with, a small fraction of a yeast-display antibody analogue library can allow one to identify molecules that bind to that target,” says Pete Heinzelman, PhD, associate scientist in the University of Wisconsin-Madison laboratory of Philip A. Romero, PhD, professor of biochemistry. A recent effort by Heinzelman and colleagues focused on identifying binders to whole Zika virions. The Zika virus represents a major medical and public health challenge, considering the severity of the infection, the multitude of complications, and the lack of a vaccine or effective therapies.
Traditionally, generating antibodies or other binders to a virus involves vaccinating an animal, a considerable undertaking that is not always doable in many labs. Moreover, obtaining acceptable levels of antibody or antibody analogue specificity and affinity often require an in vitro engineering platform. Finally, labs face technical and financial limitations when expressing immunoglobulins.
“What we wanted to do is take a small amount of virus and then—without the need for an animal, with just a naïve yeast-displayed library—identify the molecules that bind to virions,” recalls Heinzelman. While this was accomplished previously with phage display, the approach is somewhat more difficult with yeast display due to the requirement for large amounts of virus isolated via laborious purification procedures.
The approach used by Heinzelman and colleagues involved performing FACS on clones obtained from a naïve yeast surface-displayed antibody analogue, that is, human fibronectin domain, against whole Zika virions. “We did not need to go through the labor-intensive virion purification process,” asserts Heinzelman. “It was easy to make mutant progeny of the clones and find higher affinity binders to virions.”
FACS screening of libraries with site-specific antigen-binding loop mutations helped isolate progeny clones with enhanced Zika virion binding, and ELISA assays showed that their affinity is in the low double- or single-digit nanomolar range. “Going forward,” anticipates Heinzelman, “we will use this strategy for additional target binder isolation and engineering applications that would be very difficult or impossible to tackle with any other approaches.”
Solutions for antibody discovery
“Agilent offers a full analytical solution portfolio that is beyond flow cytometry and can help enable personalized diagnostics, drug development, and disease monitoring,” says Geziel Aguilar, PhD, global product manager, Reagent Partnership Division, Diagnostics and Genomics Group, Agilent Technologies. Flow cytometry instruments and reagents developed by Agilent help biopharmaceutical partners identify novel therapeutic antibodies.
“The use of the flow cytometer enables rapid screening and/or validation of these new antibodies for their specificity to bind their targets on cells,” adds Jeff S. Xue, PhD, marketing head, Cell Analysis Division, xCELLigence and NovoCyte Platforms, Agilent. xCELLigence, an impedance-based system that allows the label-free and real-time monitoring of cell proliferation, morphology, and viability, has contributed to research that has generated more than 1400 publications to date. The NovoCyte® high-performance benchtop flow cytometers have one to three different laser options and a customizable optical platform.
“With the advancement of analytical technologies, there is great potential for the screening of new drugs and therapies to become increasingly personalized to specific patient profiles,” suggests Aguilar.
“One difficulty is the identification of markers that can offer the sensitivity to assess effectiveness of candidate drugs,” notes Xue. Detecting antibody targets that are at a low abundance on cell surfaces requires highly sensitive instruments, and antibodies that target rare cells in a population require instruments that can process many cells concomitantly.
“Agilent flow cytometers are capable of both,” asserts Xue. “In addition, the NovoCyte Flow Cytometry system’s automation provides antibody screening 24/7 with walkaway convenience.”