At some point in their careers, most bench biologists will have to order antibodies for their experiments, and Glenna Burmer has a warning for them. “The world of commercial antibodies is the Wild West,” she says. Consumers cannot simply take the manufacturer’s word that a reagent will work as marketed.
This has been made clear — in 2016, the International Working Group on Antibody Validation offered recommendations on how to confirm the quality of antibodies. But Burmer believes some of these suggestions are impractical for the average researcher, requiring expertise in specific technologies such as mass spectrometry, or the costly and labor-intensive process of producing transgenic mice.
Burmer is a molecular pathologist, and a founder of antibody provider Lifespan Biosciences. In the 25 years since the company’s launch, she estimates that LSBio has performed IHC testing on more than 80,000 antibodies and she has personally evaluated more than 25,000 different monoclonal and polyclonal antibodies. Drawing on this expertise, she offers suggestions for researchers looking to make the most of their antibody-based experiments:
Diversify your portfolio. Rather than picking a single antibody for your target, select several and compare them. Humans have around 20,000 genes coding proteins, which can then come in alternative isoforms, with post-translational modifications, and carrying disease mutations. However, Burmer estimates that the number of highly validated antibodies (such as those used in diagnostics) is to fewer than 1,000 protein targets. Researchers should seek out multiple reagents and test them in the same assay. “If you compare multiple antibodies to different epitopes of a given target, you can very rapidly conclude what’s real staining and what’s not,” says Burmer.
Research your reagents. If a commercial antibody is available, there should also be a published track record of immunoassays for its target. By rooting through the literature, researchers can learn which antibodies have succeeded or failed, and gain insights into their tissue specificity and target epitope. There are a number of useful public sources of expression information, such as the Human Protein Atlas and Genecards® Human Gene Database, but no one source is comprehensive for all species or genes. Don’t despair if there are no good protein data — RNA also offers a good guidepost. “A 20-year-old paper that shows a northern blot will still tell you what the positive control tissues are,” says Burmer. Reverse transcription PCR assays can also uncover RNA signatures that reveal where a protein ought to be found.
Pick the right test for quality control. Much of Burmer’s experience is in the realm of immunohistochemistry (IHC), and her company uses tissue arrays to assess antibody performance in such experiments. As a simpler alternative, she recommends generating transiently transfected cell lines that express the antibody target. This requires some investigation to identify cells that do not already express the protein of interest, but Burmer notes that “you can look up RNA-seq data for a lot of these cell lines”. She adds that newly transfected cells tend to express high levels of exogenous protein, making them preferable to stably transfected cells, where expression may taper over time.
Double-check your antibody. A positive result with transfected cells is encouraging, but Burmer strongly suggests performing a western blot as an independent secondary assay of target specificity. Results between the two need not align perfectly. “We have many experiences with multi-band westerns, which some people would reject,” she says, “but when you get the antibody under a microscope, the whole thing cleans up and works perfectly well.” Likewise, a negative western result is less concerning if transfected cells and their negative controls show the right result in IHC. A western blot presents the protein in a different shape than is present in cells, so different results between IHC and western blot is not a conclusive reason to declare the antibody non-specific. However, a strong positive result in both assays is an excellent sign the researcher is on the right track.
Recognize the restrictions of your experiment. In pathology, samples are typically treated with fixatives and other chemicals that can alter the structure and integrity of the protein target. “Antibodies that work in formalin don’t always work in paraformaldehyde, and antibodies that work in frozen specimens don’t always work as well in formalin,” says Burmer. She also notes that paraffin embedding can weaken the signal in an IHC assay, such that detection of lesser-expressed proteins can become difficult. On the plus side, the antibodies that perform well in IHC generally also perform well in other assays that bind the native protein, such as flow cytometry.
Click for more information on IHC validated antibodies at Lifespan Biosciences (LSBio).