by Simon Glerup, Kathrin Weyer & Ib Nielsen Labtimes 03/2014
An astrocyte stained using antibodies against the sorting receptor SorLA (green) and the cytoskeletal protein GFAP (red). The visualisation of specific proteins in cells and tissue is of critical importance in life science. The end result can be truly aesthetic but the road to get there can be tortuous and expensive. Photo: Simon Glerup
Help to identify all the antibodies that didn’t work… as well as those that do and how.
The human genome contains approximately 30,000 individual genes that can undergo alternative splicing and post-translational modifications, creating a proteome of potentially hundreds of thousands of different protein molecules. The same goes for the numerous model organisms commonly used in life science, rendering an enormous number of individual protein targets of interest to scientists.
The unprecedented tool for visualisation of individual gene products in complex mixtures, such as body fluids, cells and tissues is antibodies. Generation of a specific antibody for research is a laborious and time-consuming task. Fortunately, numerous companies have specialised in providing researchers with thousands of commercial research antibodies – however, they do not always perform as promised in the data sheets.
Our group at the Department of Biomedicine at the Aarhus University in Denmark studies the signalling cascades induced by brain-derived neurotrophic factor (BDNF), specifically in subtypes of GABA interneurons of the hippocampus. Finding the best-performing antibodies for our work turned out to be a daunting task, illustrating the complexity of selecting research antibodies and, further, the difficulties in reproducing published results.
There are 21 different subtypes of GABAergic interneurons in the mammalian hippocampus, identified on their expression of different markers and their electrophysiological properties. Ideally, all can be marked using an antibody directed against the inhibitory neurotransmitter GABA but parvalbumin, calbindin and calretinin are also commonly used as markers to narrow down the specific interneuron subtype. According to Linscott’s directory, there are several hundreds of different antibodies on the market directed against these four markers. Searches in the literature databases allowed us to narrow it down to a subset of antibodies most commonly used in publications.
In our hands, however, the most-cited ones did not work for immunofluorescent staining of tissue sections and we had to test our way through fifteen different commercial antibodies, before finally finding a suitable antibody for each of the four markers. But this was only the beginning; we still had to find the tools for visualising TrkB and the phosphorylated components of its signalling cascade. We needed to test fourteen different commercial antibodies against TrkB, before finding one that worked convincingly for immunohistochemistry! We further had to test a total of eight different antibodies against phospho-AKT and six different phospho-MAPK antibodies, before being able to detect BDNF signalling by immunofluorescence.
Thinking about the costs of this work in materials and working hours is not pleasant, when you are running a lab on a limited budget. It is very expensive and a lot of work, testing your way through all the commercial antibodies, before finding the one that really works. We concluded that there is an immediate need for transparency in the use of research antibodies and propose to solve this by micropublication of unbiased user experiences in a publicly available database, by and for the research community.
For this purpose, we have generated the website pabmabs.com using an application-specific star rating system based on a Bayesian algorithm. Everyone is welcome to contribute. User comments appear in blog string, which enables quick researches to get an overview of the best-performing antibody in a given application category. We encourage life scientists to publish both good and bad antibody experiences.
Several antibody suppliers, including Everest Biotech, St John’s Laboratory, and Immuquest, have agreed to give something back to the researchers that contribute. This has enabled us to organise antibody review contests, where we give out a cash prize to a random review. Last month’s winner was postdoc Stine Klinger from NIH, Washington. The next prize of €500 will be given out on May 20, 2014 and is sponsored by St John’s Laboratory.
Last Changed: 08.05.2014