Interview with an IF Guru: Joshua Sanes

http://www.hms.harvard.edu/dms/neuroscience/fac/sanes.php

Fascinated with the brain from an early age, Joshua Sanes became a premier scientist in the field of neurobiology. Currently, Sanes is a professor of Molecular and Cellular Biology at Harvard University in Massachusetts.

Sanes areas of interest include the molecules and structures which regulate synapse formation including intercellular interactions. In addition, Sanes is also known for his role in developing a new method for visualizing synapses formation and synaptic circuits, published as “Brainbow” in Nature.

We had a talk with Dr. Sanes about the method of Immunofluorescence, the difficulties, the utility, and how antibodies play a role. Keep reading for our exclusive interview!

What is your sense of the success rate [for antibodies in the application of Immunofluorescence]?

Sanes: There isn’t a single answer. When we started making monoclonal antibodies in my lab, many years ago, we screened them by immunofluorescence, so by definition the success rate was 100%.  With polyclonal antibodies that we’ve purchased, there has been a huge variation bases on the type of antigen. Antibodies to membrane proteins, for example, seem to have a much lower success rate than antibodies to cytoplasmic proteins.

Aviva:  Can you give us a sense, 10, 20 percent [success rate]?

Sanes: No, I don’t want leave you with any number because its, such a motley assortment of antibodies of all sorts and sources. For example, when we get antibodies from collaborators or colleagues its because they’ve already shown they work and so you know maybe 50% of what they tell us work, work in our hands, but I don’t know whether they threw away ten before they got the one.

Aviva: So maybe you’re a bit more critical than some researchers when it comes to the utility of antibodies for IF?

Sanes: I think they are totally essential, but there are just a lot that don’t work. I can tell you three types of failure we’ve seen. One, is we get antibodies that if we transfect cells in culture we can stain them perfectly well, so it isn’t that there’s something about the antibody that makes IF impossible, but then we stain tissue and see no specific signal. One possibility is that levels are just way lower in tissue; another is that the background is way higher in tissue. Yet another is that the antigen in tissue is somehow  complexed with other proteins or masked. Second, a lot of antibodies just don’t work period in our hands, even if they are perfectly satisfactory for immunoblotting. Presumably, they react well with denatured protein but not with native protein. Finally, a third situation is that some antibodies are just poor -no specific reaction that we can find, or such high nonspecific reactivity that they are useless.

And just because you see staining, even patterns that look reasonable, they don’t necessarily represent specific staining. We’ve had stunning examples of antibodies either commercial or ones we’ve made ourselves, both, where we kind of know what we want to see from various sources of evidence.  We believe we know where the antigen is going to be localized and we immunostain the tissue, it looks more or less conforming to our expectations and we’re really happy. Then we stained tissue from knockout mice that we had made or obtained from colleagues -and the staining was still there! You know I honestly believe there’s a substantial fraction of reports in literature using antibodies that just aren’t true.

Aviva:  You know that western blot is obviously a denatured situation, but are there other applications that you look for from antibody companies that would facilitate your native proteins, immunohistochemistry, or frozen tissues perhaps?

Sanes: My guess is that antibodies that work for western and not IHC recognize denatured proteins, so in principle antigen retrieval or on-slide denaturation could work. But, much of our work is on delicate neural tissue -often in embryos and sometimes in whole mounts- so we are restricted in what methods we can use.

Aviva: From your experience, what companies are making the largest impact for your research and for your Immunofluorescence?

Sanes: We’ve used so many antibodies. For many years, as you know, there weren’t very many commercially available antibodies so we’d make our own, and we still do that sometimes. We buy a lot of commercial antibodies these days but because of the problems I’m mentioned getting good signals in IF we have to be very careful or we’d rapidly go bankrupt.

Aviva: How many different kinds of antigen retrieval methods would you go through for a typical antibody, or would it depend on how interested you were in the antibody?

Sanes: We may not have done as a good a job as we should have in implementing antigen retrieval methods. Maybe that’s because when we have done it, we don’t feel like its worked all that well, and then people in the lab aren’t that motivated to try a whole panel. Part of the problem, as I mentioned, is that we use developing central nervous tissue from mammals, which is pretty delicate.  You can do a lot of antigen retrieval on sections of tumor or heart that really tear up developing brain. An alternative we haven’t explored is to use paraffin sections, which are well-suited for antigen retrieval. Also, we do a lot of double and triple labeling, so antigen retrieval that might work for one antibody can destroy another antigen.

What we do typically is try unfixed tissue, PFA fixed tissue, acetone fixed, methanol fixed, and a few commercial retrieval buffers, such as Monet and Renoir. These proprietary solutions probably vary in pH and salt and urea. When we were working on extracellular matrix, we used urea denaturation and low pH pretreatment. But these treatments tend to do too much damage to the very fragile tissue we’re using now.

Aviva:  As far as antibody companies addressing the need, you mentioned before there are many more reagents available out there, would that be predominately the message that IF might be growing in the research area just because there are now antibodies that are readily available? It is being used much more in the field from what we can see.

Sanes: It may be also that there are more people using it. IF has been a mainstay for my lab the whole time I’ve had one, which is 30 years now. I would not say we use it more than we used to, we’ve always used it a lot. I think the shift is that we’ve moved from making our own antibodies which were good but difficult to make -and obviously you can’t make very many. So we’ve moved in the past years towards commercial antibodies.

Aviva: What are the resources you might suggest other researchers doing this type of histology staining? Are there any good sources?

Sanes: No, I don’t know of any good ones.

Aviva:  It’s a very complicated place to be a company and it’s a very complicated place for a researcher and one of Aviva’s goals is to provide the best resources that we can to help facilitate more research. That’s the purpose of this phone call, to get assistance there.

Sanes: Well I think the sort of thing that you’re doing is terrific.  Where you can partner with people or get feedback from people and put applications on the website. You know I wish there was some better technology for making antibodies but I have not really worked in that area of methods development.

Aviva:  Aviva’s one of the largest polyclonal manufacturers in the world, if not the largest, and we’re challenged with how do we, from a business standpoint produce quality reagents that validate well in multiple applications

Sanes: Well I think the key is what you’re already doing. Obviously you want to make good reagents, but to just accept that a lot of them are not going to work, at least for some applications. I think by being up front about it, putting it out there on your website, being open to it, you’ll get people’s confidence and you will grow.

Aviva: The sampling program is intended for that. It’s intended to really get the reagents out there and see how they work. Companies like us are challenged with the opportunity of how else can we screen and test our antibodies so they’re more useful. We have 30,000 of them. That’s really what some of our challenges are, how we can test these, obviously we’re high throughput, and we use numbers to our advantage. We’re going to make these antibodies and we know a fraction of them won’t work, and we know a fraction of them will work. What else can we do from our methods to screen, you mentioned transfecting cells, that is not a practice we take on because we feel that provides a little differentiation for us, that we have an endogenous method of validation.

Sanes: I very much wish I had something wise to say, we’ve never gotten beyond the sort of trial and error, and look and look again method that we’ve been talking about, whether it is with our own antibodies, antibodies that we buy, antibodies we get from collaborators. It is a problem because it’s the quintessential low-throughput method. We just don’t have anything better at this point.

Aviva:  I thank you again for your time Dr. Sanes, that was a good conversation, we will review it and see what we can take from it to help other researchers, as you can imagine, companies like ours receive many inquiries for optimizing applications you know, we do our best to find solutions.

This entry was posted in Blog, News & Announcements and tagged , , , , , , , , , , , . Bookmark the permalink.

Leave a Reply