Antibodies. Your typical biology research can’t live without ‘em—they label proteins for view under the microscope, pull them out of complex solutions for study and identify their presence and sizes on Western blots. But lately, researchers are realizing that there are challenges to living with ‘em, too. Antibodies may recognize the wrong protein in some assays, or fail altogether, and this accounts for a big part of the billions of research dollars wasted on irreproducible results [1].
It happened to Anthony Couvillon, scientific marketing project manager for Cell Signaling Technology, when he was in graduate school. He wasted two years working with an antibody before he figured out it detected not the protein he cared about but rather another of the same size. “Every single person I talk to has a horror story of some kind,” he shares.
In recent years, those antibody-induced horrors have started to attract attention [2,3]. Those with a stake in the issue—including researchers, vendors, journals and funding agencies—have come together in key discussions, including the 2016 Antibody Validation conference [Editor’s note: Mike Okimoto, Biocompare’s chief content officer, participated in a panel discussion at the conference.]
Possible solutions are starting to emerge, including guidelines for antibody validation, unique identifiers to distinguish antibodies and requirements to perform such validation for funding or publication.
According to Alejandra Solache, head of reagents, innovation and manufacturing at Abcam, “Abcam has been raising antibody standards for years, most recently with our knockout (KO) validation initiative, started in September 2015, which now numbers over 650 KO-validated antibodies.”
Some reagents are straightforward: If you order sodium chloride, you can rely on receiving salt, whenever and wherever it’s purchased. Not so with antibodies, which are protein reagents produced by living organisms. As such, they create certain challenges. For example, depending on the epitope an antibody binds, it might work great with the unfolded proteins on a Western blot but be blind to the natively folded versions under the microscope—or vice versa. According to Jason Li, CEO of Proteintech Group, the question of whether “the antibody [is] recognizing what it is supposed to recognize” poses a key challenge within the antibody market for both tool providers and researchers. Polyclonal antibodies are raised in animals, which have differing immune responses, causing lot-to-lot variation. How one treats a sample, such as the number of freeze-thaw cycles, can affect experimental results. And antibody products can change names or be shuffled among vendors, making it difficult to keep track of a favorite version, or to tell if the antibody provided by one provider differs from that of another.
And some antibodies just aren’t very good, as Couvillon discovered with his cross-reactivity issue. “It’s very hard to get truly monospecific antibodies,” says Michael Snyder, director of the Center for Genomics and Personalized Medicine at Stanford University. His group once tested 11 antibodies against protein arrays of 5,000 different proteins and found significant cross-reaction [4]. “You’d be amazed at how often they’ll react with things they’re not raised to,” says Snyder.
In fact, one-third of researchers say they often try three or four antibodies to a given target before they find one they like.
Yet many scientists don’t understand why it’s important to test each lot of an antibody in positive and negative control situations.
Less than half of junior researchers regularly validate the antibodies they use, according to one survey [5].
Recombinant antibodies are an option to traditionally generated antibodies and for some researchers represent an alternative with a great validation potential. According to Kalpana Singh, product manager flow cytometry at Miltenyi Biotec GmbH,” Recombinant antibodies are clearly the future of antibodies as they are highly defined in nature and generated under standard conditions to achieve high consistency.” In terms of vendor responsibilities, Singh notes that “investing in recombinant antibodies can be a costly affair for vendors, but it is certainly an investment to ensure that in long term only consistent antibodies are available in the market.”
In Biocompare’s 2015 antibody market survey, 68% of respondents (n=884) shared that when testing a new antibody for the first time, they would only purchase from a trusted vendor—one whose products had contributed to successful results in their previous experiments. In addition, 66% of the survey’s participants commented on using technical data and citations provided by the vendor in making a first-time purchasing selection.
Solutions
Snyder and co-authors recently published guidelines for such researchers to consider, with five different methods, or “pillars,” for verifying antibodies [3]. They laid out methods to test antibody reactivity—such as trying it in a cell line missing the target protein, or comparing it to another antibody that binds the same protein in a different spot—and matched those methods to different applications, such as Western blotting or immunohistochemistry. The “gold standard” is to test an antibody in a knockout cell line, says Snyder, but that isn’t always possible.
The pillars formed the starting point for discussions at the Asilomar conference. Meeting attendees have now divided into working groups to expand those guidelines. Then, a committee of vendors will weigh in on what’s reasonable. For example, it would surely be nice to test every antibody for sale via extensive CRISPR-based knockout lines, but that would make the products quite expensive, notes Karen Padgett, vice president of the antibody business unit at Bio-Techne.
Vendors also are trying to improve their documentation about where an antibody comes from, and what testing it has undergone. “It is important for transparency that the nature of the immunogen, how an antibody was screened during development and specific details pertaining to lot-specific validation are communicated,” says Carl Ascoli, chief science officer for Rockland Immunochemicals. That way, users can double-check the manufacturer’s claims and use the same experimental conditions to obtain good results in their own experiments.
Li adds that the five pillars of validation are a good start, but it will take time to implement. Right now, he suggests “transparency on the origin of a product.” This provides greater visibility to end users as to the product they are working with and ultimately translates into the prevention of wasted time, money or the risk of a research paper retraction.
Nomenclature is also a key issue. “None of these pillars matter if you actually aren’t identifying the antibodies properly,” notes Anita Bandrowski of the University of California, San Diego, a co-author on the pillars paper. Publications should include the company name, catalog number and lot for each antibody, she says.
But that information could be less than helpful, Bandrowski says, if an antibody has passed from one manufacturer to another, or it now goes by a different moniker. That’s why she and others are trying to give a unique number to every antibody out there, through the Antibody Registry. So far, the database contains about 2 million antibodies, she says, and its ID numbers have been cited in more than 1,700 papers.
Solache shares, “At Abcam, we name our products based on the protein name as defined by UniProt. Although, if we find out through citations that there is a more commonly referred name for the target, then we’ll use that name instead.” We also include host species, clonality, clone number and if appropriate isotype. However, continued efforts are needed to standardize nomenclature of antibodies.
It takes a village
Those projects are just a start. “A community effort is needed to improve the way antibodies are produced, validated and used,” says Ascoli. Vendors can contribute by ensuring that they sell valid products and by clearly telling buyers how to cite each antibody. Solache also adds “The Life Sciences industry as a whole needs improved training for new researchers in order to make them more aware of the value of validation and good reporting. At Abcam we have many resources for students starting out in the lab, and we share our expertise at Antibody Validation Workshops which aim to improve the education of new researchers.”
Journals, in turn, can ensure that the papers they publish provide sufficient detail on the antibodies used. Cell Press, in particular, has made strides with its new STAR Methods program, supporting logical, complete methods sections with no size limit. Funders can play a similar role; for example, in 2016 the National Institutes of Health began requiring grant applications to include reagent authentication.
Investors are another player in the antibody world, according to Li. His advice to investors is to “do their due diligences to understand the antibody industry better and use their influence for quality, not quantity.”
Singh, notes that a first but major step in terms of the responsibilities of funding agencies has been the reporting guidelines established by the NIH. Singh states, “the NIH is just one of the major funding agencies worldwide” and a global standard of guidelines should be established, and enforced by all funding agencies around the world, describing which antibody-related information must be disclosed to ensure a successful grant application.”
And researchers must protect their own research by applying the pillar guidelines and by sharing information about the antibody testing they perform. Resources such as Biocompare and other online product sites offer user reviews and citations for antibodies, and some company sites also post customer reviews. Keep in mind that the most-published antibody isn’t necessarily the best one out there, advises Padgett.
Tackling the antibody crisis is a big job, but smaller efforts indicate it can be done. The ENCODE project has set clear standards for antibody characterization, and it has tested more than 400 so far. The gaps and challenges with antibody validation will be explored in greater detail by Biocompare in an upcoming documentary in May 2017 titled “Reproducibility Issues in The Lab: Crisis or Opportunity?”
In the end, proper testing of antibodies and sharing of that information will save researchers time and money, Snyder says. Continuing discussions amongst all vested parties will be instrumental in identifying solutions. Stay tune as this discussion advances and resolutions are made.
References
[1] Friedman, LP, “Irreproducibility: A $28B/year problem with some tangible solutions,” Genetic Engineering News, June 30, 2015.
[2] Baker, M, “Blame it on the antibodies,” Nature, 521:274-276, 2015. [PMID: 25993940]
[3] Uhlen, M, et al., “A proposal for validation of antibodies,” Nature Methods, 13:823-827, 2016. [PMID: 27595404]
[4] Michaud, GA, et al., “Analyzing antibody specificity with whole proteome microarrays,” Nature Biotechnology, 21:1509-1512, 2003. [PMID: 14608365]
[5] Friedman, LP, et al., “The need for improved education and training in research antibody usage and validation practices,” BioTechniques, 61:16-18, 2016. [PMID: 27401669]
Image: Shutterstock Images