Uppsala University, genetics and pathology department
Who: Ola Soderberg
Position: Researcher, Uppsala University, genetics and pathology department, 2001 to present
Background: PhD in pathology, Uppsala University, 1997; post-doc work, Instituto de Patologia e Imunologia Molecular da Universidade do Porto University of Porto (IPATIMUP), Portugal, 1998 to 1999; post-doc work, University of Linköping, biomedicine and surgery department, 1999 to 2001.
The proximity ligation assay is a novel technology in proteomics that some say offers a more sensitive and specific way to measure protein-protein interactions. Proximity probes containing oligonucleotide-labeled antibodies are designed to bind to proteins. Because the protein targets are represented as DNA targets, they can be amplified and measured by common DNA-analysis methods such as PCR.
Last year, Ola Söderberg and his colleagues at Uppsala University developed an in situ application for PLAs. An Uppsala-based startup, Olink, is commercializing the technology under the name of Duolink.
Last month, Söderberg and his fellow researchers published a paper in Molecular & Cellular Proteomics describing further development of the technology. The improved method is a “generalized” version of the approach that “preserves the sensitivity and selectivity of the in situ PLA method while permitting the use of general, species-specific antibodies as proximity probes with any suitable primary antibody pair,” according to the paper.
Below is an edited version of a conversation ProteoMonitor had with Söderberg last week about his work.
You published a paper on the in situ proximity ligation assay technology in Nature Methods last year. How have you improved on the technology since then?
One problem with the in situ PLAs was that you have to conjugate the oligo to the antibody, and then you must have pure antibodies without BSA or any other carrier protein. So we thought it would be good to have secondary antibodies instead so then people can use whatever antibodies they have at home. So this is just a way to make it easier for other people to use it for their special antibodies.
Of course, if you buy secondary antibodies, you can buy a large amount for a small cost, and then you conjugate the oligo to this and then you can purify it with HPLC, for example, and you’ll have very pure conjugates. And then you can use it for any combinations of antibodies from different species.
Can you explain what you mean by secondary antibodies?
If you stain against two different proteins – if you have one mouse protein that is protein x and a rabbit protein that is protein y, then you have a secondary antibody that binds rabbit [immunoglobulin] and mouse [immunoglobulin]. So then it would be like two antibodies sitting on top of each other. Then you don’t measure the interaction of the protein molecules, but you measure the amount of primary antibodies that bind to these proteins or modifications.
So it’s a way to make it easier. We have a lot of other projects ongoing where we’re trying to modify the technique to multiplex it and so on, but we don’t have data yet to publish that.
What is the current throughput of the approach without multiplexing?
You can measure two or three proteins binding together, or modifications, so you have two or three antibodies at the same time. You can study as many proteins as you want, but you have to take one set at a time.
So scaling that up to the tens or hundreds would probably be very useful.
Yes, but if you have an antibody pair for anything you want to study, you can do it. We have tried a lot of other proteins as well, or interactions or modifications. I think there have been 10 or 20 different proteins studied so far. A lot of people from around the world have contacted us and want reagents, so they’ve started to use this, though we don’t now exactly what they’re studying.
Also now, there is a small company that has started here in Uppsala selling these reagents as well, so people can buy a kit now and do the staining.
It’s going well. It’s a very small company. I’m not involved at all in the company. I just do work for the university, and they are commercializing this aspect, but I don’t have any funding or money from them. It’s a quite nice situation. Otherwise, you always have this problem where you can become a bit biased. I’m trying to keep it separated.
It appears from what Olink is doing in this agreement with ABI that this method has a lot of potential applications. In this instance, they licensed out the rights for use in the research market to ABI, but they seem to think there is also diagnostic potential. What are your thoughts in terms of where and how this technology can be applied — especially in terms of diagnostics or biomarkers?
The standard technique you use nowadays is you study how much protein is expressed in a cell, and then you try to correlate that to prognostic markers. You correlate [that] to survival or response to chemotherapy for cancer treatment, for example.
But with this method, with the in situ PLA, you can also study interactions between proteins, modifications of proteins, [which is an improvement because] even if you have the same amount of proteins in a cell, they can interact with each other or not.
If you compare it to soccer — you have 11 players on each team. If you just measure how many players you have, it doesn’t tell you much about the game, but if you can see how the players interact with each other, that is a completely different level. So I think we have the potential to see a lot of new things. People will probably have to study a lot of different protein interactions to see which ones are good candidates for diagnostic purposes, but I think it has tremendous potential.
It seems like there is a lot of work being done on new methods for measuring protein interactions. How does in situ PLA compare to some of the standard and emerging protein-interaction methods?
The advantage of this method is that we can study clinical samples. If you have antibodies that can bind to your protein interactions with clinical samples, and if you’re using yeast two-hybrid or FRET or bimolecular fluorescence complementation, you have to transfect cells with vectors with fluorescent-labeled proteins. So this is a way to study patient material.
I think there’s also the possibility to multiplex it further, so you can do many things at the same time in many different colors. And we also have single-molecule resolution. All the small dots you see are single molecules.
In situ PLA results in an amplification process where you get a thousand fluorophores for each molecule, so it’s a very bright signal. It becomes larger than the actual interaction, but it’s possible to count how many events you have in the cell and not just say you have more in this cell than that. So I think that’s an additional advantage.
However, I think the major advantage is that you can study patient material.
Are you doing that now? Are you applying this in the clinical setting at all?
We have started a study to look at breast cancer. We have a few [projects] we are starting now. Our group is mainly focused on developing new techniques, so now we have developed this and we want to go on and make it better, so if people want to collaborate, then we try to help them do some studies.
So your collaborators are doing the breast cancer work?
Yes. But when this is available to everyone in the world, I hope that a lot of people will use their material to study whether protein interactions or modifications can be a diagnostic marker.
What are your next steps? It sounds like there are a lot of different directions you could go in.
I think in the short term we would like to multiplex this further, so if five proteins interact with five other proteins, you want to see all these combinations simultaneously in the cell, and they have to be color coded. If one color is a certain interaction, then you can follow signaling cascades in cells. Usually one protein can interact with many different partners, and they can have different effects depending on what partners they interact with. So we want to see how the balance is between the different signaling pathways in cells and see if that can be used as a diagnostic marker.
Then we also want to study protein and nucleic acid interactions. It’s a bit complicated, but we have some promising results.
In terms of the multiplexing, is that a question of developing better reagents, or are there platform or instrumentation aspects to getting that up to scale?
It’s different oligo designs. It would be different designs for the technique.