Research antibodies: do not use them to stain your reputation

in a recent roundtable organized by the Federation of American Societies For Experimental Biology (FASEB) on data reproducibility and antibodies (8), we learned how the sale and use of antibodies place both basic and preclinical research at risk of losing support from the public at large, from funding agencies, and from Congress. Hundreds of millions of dollars are wasted annually on poor data collected with nonvalidated antibodies (3). The National Institutes of Health (NIH) is increasingly concerned about data reproducibility and will require applicants to include a section addressing scientific rigor, transparency, and validation of reagents in grant applications. This new mandate might start as early as January 2016. This gives us an opportunity to revisit the expectations of AJP-Cell Physiology regarding the use of research antibodies.

There are two major factors that contribute to the problem on the vendor side: one is transparency, the other is lack of adequate validation. There are over 2 million antibodies on the market (1), but estimations are that only 100,000-200,000 can be considered unique. Indeed, vendors buy and sell antibodies from each other in order to fill their catalogs and to offer the widest variety to their customers. Worse is the fact that many of the vendors repackage these antibodies and change their names without any consideration for transparency regarding their origin or identity. Unique identification of research resources is an essential component for the reproducibility of science (9). How many of us have on occasion purchased several antibodies just to increase the chance of finding a good one? The untold truth is that, in such situations, the chance of purchasing the same antibody twice is rather high. This problem is compounded by the fact that, in contrast to antibodies that are FDA approved for clinical use, very little effort goes towards the rigorous validation of research antibodies (1). If you think that because a company sells an antibody, it will do what it is supposed to do, be prepared for this shocker: there are some companies that reportedly sell antibodies with up to 100% failure rates (2). Now, this is obviously not the majority of cases and numerous reputable companies try their best to provide quality products. The sheer number of antibodies that are available on the market and the number of unique uses and applications that investigators come up with are also contributing factors to this sad state of affairs. It is clear that significant improvements could be made by commercial antibody producers and distributors to increase transparency regarding the identity/origin of antibodies and for initial validation. It is, however, the final responsibility of investigators to use these reagents in the most rigorous way. Reagents should be optimized for the very specific conditions in which they will be used. Therefore, it is only good science to characterize them rigorously and to conduct a full set of controls in your experiments. Published guidelines might help AJP-Cell authors design proper controls for their immunohistochemical assays (5, 6).

What do AJP-Cell Editors expect? Aside some basic mandatory requirements, the journal also requires high-quality and rigorous science to ensure that a reported observation is reproducible and therefore will withstand scrutiny and the test of time. Let's consider these two expectations separately with regard to the use of antibodies.

1) The mandatory requirements identified in the Journal's instructions to authors are imposed to ensure that basic reporting standards exist. AJP-Cell receives many manuscripts that fail to provide key basic information on reagents or that fail to include proper labels in Figures. While this is typically caught during review, effort should be made to include this important information at first submission. Editors and reviewers evaluate manuscripts differently if they are rigorously put together.

• Authors are expected to provide full antibody information in the Materials and Methods section. This includes target, host species, polyclonal vs. monoclonal (clone if monoclonal), vendor, catalog number, and lot numbers(s). Note that while lot numbers are typically not disclosed in publications, because of lot to lot variations, and possible drift of monoclonal clones over years and decades, it is good practice to include this information in publications (7). When antibodies are received and aliquoted, the lot number should be recorded in laboratory notebooks for future reference and disclosure. Additional information regarding antibody dilutions (or concentrations—if known) is required. It is also important, especially for “in house” antibodies, to provide details of the purification method(s).

• Authors are encouraged to describe existing evidence of antibody validation (e.g., from the literature). The description should be in a few sentences with reference(s) rather than simple citations. AJP journals have no page limitation and therefore space does not constrain an author in providing this essential information. For example, a sentence like “Smith and colleagues demonstrated that the antibody recognized” protein X “using Western blot analysis experiments involving increased amount of cRNA injected in Xenopus laevis oocytes (Ref.). In addition, they established antibody specificity using a knockout cell line (Ref.).” While this information is typically omitted in manuscripts, as mentioned above, it will soon become a NIH grant requirement to provide a detailed assessment of reagent validation.

• According to APS publication guidelines regarding figures, each gel or blot should contain a molecular-weight size marker. Each panel should retain space above and below the band of interest from the original image. It is not appropriate to crop the panel right on the band itself. The methodology for signal capture (X-ray film, phosphoimager, etc.) and for data analysis also needs to be included in the Methods section. For microscopy panels, it is good practice to include a negative control, for example an irrelevant antibody of the same immunoglobulin class, as well as the test images. All micrographs should include a scale bar.

2) Good science implies that observations reported in publications withstand scrutiny and are reproducible. It is, however, not appropriate for journals to prescribe specific solutions. The number of times an experiment is replicated, for instance, might indicate how reproducible an observation is. We would not want, however, to mandate that all experiments be systematically replicated. Let's suppose that a single experiment demonstrates a drug effect on a physiological measure (e.g., compound A inhibits Ca²⁺ influx in a cell line). It would be perfectly appropriate for an author to place that experiment, performed only once, in a Figure panel, provided that other panels expand this original observation (for instance, another panel could depict an experiment reporting an 11-point dose response of the drug with a determined IC₅₀). Reproducibility concerns the basic scientific observation and not the replicability of any single experiment. The same rigor and common sense apply to the use of antibodies. It is good science to design an experiment properly with both positive and negative controls, but precise details of the controls should be left to the authors. Manuscript reviewers must assess the strength of the controls and the validity/reproducibility of the observations.

Our responsibility as researchers is to always question the validity of our data and to always build the strongest case by using the most rigorous approaches. At the level of publications, the more details we provide regarding the source and identity of antibody reagents, the better our observations can be scrutinized and our studies reproduced. A similar message was recently conveyed to AJP-Gastrointestinal and Liver Physiology authors (4).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author. The author was representing the American Physiological Society at the FASEB roundtable.

AUTHOR CONTRIBUTIONS

E.D. wrote and approved the final version of the manuscript.