Faulty Antibodies Undermine Widespread Research
Faulty Antibodies Undermine Widespread Research

Faulty Antibodies Undermine Widespread Research

Two papers reveal that many commonly used research antibodies don’t bind as believed, highlighting the need to validate these reagents before use.

Ruth Williams
Ruth Williams
Jan 30, 2020

ABOVE: ISTOCK, ANUSORN NAKDEE

The ability of research antibodies to bind to their proteins of interest, or to the peptide tags attached to such proteins, can vary depending on the particular amino acids surrounding the target binding sites or modifications to those sites, according to reports published Tuesday (January 28) in Science Signaling. In some cases, such antibodies even cross-react with other proteins, the authors warn.

“These papers, like others, clearly indicate that even commercially available antibodies need thorough evaluations for the respective purposes for which they will be used,” pharmacologist Thomas Wieland of Heidelberg University who was not involved with the papers writes in an email to The Scientist.

Antibodies, proteins that bind with high affinity and specificity to target molecules, are one of the most commonly used tools for studying protein biology. Some have been produced to recognize particular proteins of interest, while others recognize short peptide chains (tags) that are fused to proteins to enable detection or isolation.

But these antibodies don’t always bind as expected, as a number of recent articles have highlighted. Still, the issue “keeps popping up its head because people don’t follow directions for validation,” says pathologist David Rimm of Yale School of Medicine who was not part of the research team. “[Researchers] don’t take it seriously enough,” he says, explaining that there is still data from experiments with unvalidated antibodies being published in the literature.

People need to know the danger attached to antibody use, and if you do not want to be tricked by your antibodies, then you need to validate them.

—Egon Ogris, Medical University of Vienna

Egon Ogris of the Medical University of Vienna who led the two new studies, is similarly frustrated, and after he got concerned about a specific commercial antibody that is commonly used, he felt compelled to examine the issue in more depth, as a “service to the community,” he says.

Ogris studies the protein phosphatase 2 (PP2A)—a ubiquitously expressed multi-subunit enzyme involved in numerous cellular functions and implicated in diseases including cancer and neurodegeneration.  To examine one particular subunit of the enzyme, he and a coworker appended to the subunit the commonly used Myc peptide tag. But, because they happened to use different enzymes to insert the tags, they got two versions of the tagged subunit that differed by just a few amino acids at the region linking the tag to the protein, explains Ogris.

When the researchers used a commercially available Myc-targeting antibody called 9E10 to detect the two subtly different tagged PP2A subunits, one version of the protein produced a nice clear signal, says Ogris, “and the other was hardly visible.”

“The only difference was these four or five amino acids,” Ogris says, suggesting this was the cause of 9E10’s failure to recognize the tag. To investigate, his team produced their own Myc antibody—the now commercially available 4A6—and showed that, unlike 9E10, this reagent could bind equally well to both tags, and produce similar strength signals in both samples. The team went on to examine four other commercially available Myc antibodies, finding that all but one exhibited variable binding affinity for the Myc tag depending on the surrounding amino acid sequences.

The findings indicate that how and where a researcher attaches a Myc tag to a protein could influence detection. The antibody may indicate “high expression or low expression, though the protein may actually be expressed at equal levels,” says Ogris, “which is actually a catastrophe if you think about it.”

In their second study, Ogris and colleagues examined antibodies that directly target the catalytic subunit of PP2A. A number of such antibodies exist, but there had been concern that methylation of the subunit—a necessary modification for the enzyme to carry out its function of removing phosphate groups from certain molecules—interfered with the antibodies’ binding abilities. Out of four antibodies tested by the team, all but one did indeed show dramatically reduced affinity for the methylated version of the protein. More alarmingly, the same three antibodies showed weak affinity for the related yet entirely separate phosphatase PP4.

All of the tested antibodies are commercially available and one (1D6) forms part of a widely used PP2A activity assay kit. Because of his findings, Ogris advises other PP2A researchers not to use the kit.

Rimm expressed concerns about scientists expecting the companies selling the antibodies to do the validation for them. Ultimately, “the motivation of the vendor [is to] . . . increase shareholder value.” So it is not always in their interests to do vigorous validations, he says. By contrast, “the motivation of the scientist is to do science, and if we want to do good science we have to be sure that all our reagents are giving us the expected values and are reproducible.”

In short, people “need to know the danger attached to antibody use,” says Ogris, and “if you do not want to be tricked [by your antibodies] . . . then you need to validate [them].”

For a guide on how to validate your own antibodies, commercial or otherwise, see here.

S. Schuchner et al., “The Myc tag monoclonal antibody 9E10 displays highly variable epitope recognition dependent on neighboring sequence context,” Sci. Signal., 13, eaax9730, 2020.

I.E. Frohner et al., “Antibodies recognizing the C terminus of PP2A catalytic subunit are unsuitable for evaluating PP2A activity and holoenzyme composition,” Sci. Signal., 13, eaax6490, 2020.

Ruth Williams is a freelance journalist based in Connecticut. Email her at ruth@wordsbyruth.com or find her on Twitter @rooph.