Science Signaling Podcast for 2 August 2016: Patient-specific protein complexes

See allHide authors and affiliations

Science Signaling  02 Aug 2016:
Vol. 9, Issue 439, pp. pc17
DOI: 10.1126/scisignal.aah5912


This Podcast features an interview with Adam Schrum and Steven Neier, authors of a Research Article that appears in the 2 August 2016 issue of Science Signaling, about a method for identifying protein-protein interactions in patient tissue samples. The authors used this method to compare signaling complexes downstream of the T cell receptor in T cells from healthy skin with those in T cells from the skin of patients with the autoimmune disease alopecia areata. The study revealed differences in the relative abundance of some protein complexes between T cells from the control and patient groups. This technique could be adapted for use as a diagnostic tool to stratify patients by molecular phenotype and predict the therapeutic strategy that is likely to work best for each patient.

Listen to Podcast

Listen to Podcast | Podcast RSS Feed | Featured Research Article

Length: 16 min; file size: 13 MB; file format: mp3


Host – Annalisa VanHookWelcome to the Science Signaling Podcast for August 2nd, 2016. I’m Annalisa VanHook, and today I'll be talking with Steven Neier and Adam Schrum about a method for identifying protein-protein interactions that are specific to individual patients (1).

A lot of diseases are caused by changes in signaling pathways. It’s fairly straightforward to recognize gross inactivation of a signaling pathway or aberrant activation of a signaling pathway in diseased cells versus healthy cells. But more subtle changes in signaling pathways can also be physiologically relevant. Signaling networks are finely tuned, so even changes in the relative abundance of the protein complexes that transduce signaling can affect cell biology, and therefore also affect the entire organism. Adam Schrum, Steven Neier, and their colleagues have developed a method for comparing the protein complexes that form in cells from diseased patients with those that form in cells from healthy subjects. Schrum and Neier spoke to me from the Mayo Clinic in Rochester, Minnesota, about how this technique works and how they’ve already used it to identify a molecular signature associated with the skin disease alopecia areata. They hope that this method could be used to develop personalized therapies for complex diseases.

Interviewer – Annalisa VanHookHi, Dr. Neier and Dr. Schrum. Welcome to the Science Signaling Podcast.

Interviewee – Steven NeierHello.

Interviewee – Adam SchrumHi, thanks very much for having us.

Interviewer – Annalisa VanHookAdam, can you explain the logic of your approach? What did you measure and why?

Interviewee – Adam SchrumWell, what we're really trying to get at is, what is the signaling language that the cell perceives in order to do something? And when one thinks of signal transduction pathways being driven by all these myriad protein-protein interactions, one can think of that as kind of a language. So you have… just like the alphabet has 26 letters, but how many messages can you send with it? Well, you can form different words and put those letters together in different arrangements to send different messages. And we're thinking of proteins in that way. So, what we are trying to look at in our approach is what are the different combinations in which proteins actually join together in shared signaling complexes in order to convey a message to the cell, instruct it to do something, so that cell can do it.

And in the case of T cells, which we focused on, we looked at the T cell antigen receptor, which is a receptor known to be able to transmit different functional signals. It can cause a T cell to, you know, reject a tissue or eradicate an infection, or it can cause a T cell to protect tissue and stop other cells from destroying it in a tolerance kind of response. So, our approach was to try and look at the different combinations that signaling proteins can adopt in order to send these functionally distinct messages to this cell. And what we measured was a sort of matrix, which we'll tell you more about, a matrix approach to seeing the combinations that these proteins are able to form together in a network kind of analysis.

Interviewer – Annalisa VanHookSteven, how does the method work?

Interviewee – Steven NeierSo, the method is a fluorescence-based assay. It uses and leverages colored beads that we're able to use to pull down specific native complexes. And then we're able to come in and detect protein co-associations in kind of a pairwise manner using fluorescent probes. For those who are familiar, it's kind of a like a multiplex ELISA. And in this way, we're able to measure pairwise associations and thereby generate this matrix format where we're able to combine the different interactions into a network format to look at the network signaling pathway.

Interviewer – Annalisa VanHookHow have you applied the method?

Interviewee – Steven NeierSo, we first started by applying the method to cell lines just kind of as a proof of principle application. We then wanted to show some specificity with the technology, and to do that we used a physiologic inhibition of a signaling pathway, specifically of the costimulation pathway that is correlated and important for T cell function and signaling. We then took the technology and wanted to apply it to patient samples, and specifically we want to do it with situations where we had limited biomaterial. So—we applied our technology to look at the few T cells that we're able to isolate from these very small 4 mm punch biopsies that we're able to collect from either control patients or alopecia areata patients. And with those samples, we're able to measure some basal reactivity and a dysregulated signaling module.

Interviewer – Annalisa VanHookYou mentioned that you use skin punch samples from the patients. What specifically is alopecia areata, and how is the skin affected? Why is the skin important for that disease?

Interviewee – Adam SchrumSo, alopecia areata is a T cell–mediated autoimmune disorder. T cells aren't the only cells involved, but they are thought of as kind of one of the main players and master regulators. And it's basically autoimmune baldness. So what happens to these patients, it can strike at any age and with various levels or degrees of severity. What happens to people is that, you know, they're suddenly combing their hair, and lo and behold their hair is falling out in their hands, in their own comb. And it's very disturbing. They come into the clinic, and they basically say, “What's going on? What's happening?” And it ends up being T cells attacking the hair follicle.

One of the interesting things about this disease—I would call it an interesting thing from the point of view of trying to think of treatments—is that it is reversible. So, when patients have been put on immunosuppressants, the hair has grown back. And there are various immunosuppressive treatments that are in trials right now to try and contain the immunosuppressive nature of such medication, but with the goal of inhibiting T cells such that the hair grows back. So that's what alopecia areata is and what the patients go through.

What these punch biopsies are, are small, you know, almost the size of a hole punch with a little bit of thickness. And these donors who gave us their tissue for this, allowed this size hole to basically be punched in their scalp where active lesions are occurring. And it's that little punch biopsy they call it that was brought to the lab, and it's that that we got the T cells out and put in our analysis to look for these network patterns.

Interviewer – Annalisa VanHookSo you essentially compared the protein complexes that are formed in T cells from unaffected patients to those in T cells from affected patients. Which signaling pathways were affected in the alopecia patients?

Interviewee – Adam SchrumSo, this was very interesting. And the first thing to say is actually there were a lot of signaling protein complexes that were made that were common to both the alopecia areata patients and also our control patients. But that makes all the more interesting the answer to your question: What was different in the alopecia areata patients? And the main pattern we found takes the form of something we call a module. So, there was a lot of network style data to be gone through. And what a module is in the context that I'm stating it is a set of proteins that joined shared complexes in a manner that correlated with each member of this module in a statistically significant manner.

And so, the alopecia areata patients had a different balance of two proteins. One is called GRB2, the other is called GADS. And these two are related to each other, but they also compete for binding sites on a very important signaling protein called linker for activated T cells or LAT. That's a long protein that's got a lot of different binding sites on it; it's a signaling protein hub is what it is. These two proteins—GRB2 and GADS—compete for some binding sites on LAT. And the alopecia areata patients, their balance between those two favored more GADS over GRB2. That which we observed in the control patients favored more GRB2 over GADS at the signaling hub, LAT. The significance—or potential significance—of that is that these two proteins link up in terms of signal transduction to some different types of effector functions. So, GADS, which was the one favored in alopecia areata, has a predilection for interacting with SLP-76 and some other proteins in a specific pathway; one is called ADAP/SLAP130.

But the functional consequences is that that signaling pathway can link to the cytoskeleton and to integrins, can make cells stickier, and can help a T cell that's trying to reject some tissue—it actually strengthens that function. GRB2, which was the one favored slightly more in the control patients, has a constitutive interaction with SOS1; it tends to push towards the RAS-MAP kinase pathway, which has a whole series of functional consequences there. So what I would summarize as the interesting difference is that there is a different balance in these two proteins; each of them leads to specific effector functions with these cells, and it makes it look like the T cells in the diseased patients are sending a different message than the cells of control patients, at least at this particular molecular interface that we're talking about.

Interviewer – Annalisa VanHookDid all of the alopecia patients have similar changes in their signaling, or was there some variability between patients?

Interviewee – Steven NeierThat's really a great question. What we noticed across both our patients and controls was quite a bit of heterogeneity. And that's really interesting to us, but because of the small sample size that we had, we didn't extrapolate and make many conclusions on an individual level. And so, for our publication, we primarily focused on the patterns and the network signaling that were identified across groups and trends that we're able to distinguish between our patient and control groups. However, it's something that we're really interested in looking at further, because this heterogeneity might define specific subgroups of alopecia areata patients that might benefit from one treatment strategy over another. And so, distinguishing how this heterogeneity might have functional consequences in the clinic is something that we're really excited about transitioning forward as we increase our sample size.

Interviewer – Annalisa VanHookDid you use the findings from this study to tailor treatments for these patients?

Interviewee – Adam SchrumNo, we haven't been able to do that yet because the study isn't quite at the point where we would be able to do that. So, we studied T cells from autoimmune lesions—or control lesions—that had not undergone treatment yet. But the question that you're asking goes right in the direction that we would like to take this. So, treatment should be one of the next variables we look at. Patients that have undergone treatment X, what is the response of their T cells? But also, before they've undergone treatment, can we look at the kinds of patterns—such as the ones that we're reporting in this publication, but also others we might notice when we increase our n and do more of this—can we identify patterns that we could later associate to responsiveness to treatment X versus treatment Y? So we think there's a lot of potential for this network protein interaction business to get into the realm of different treatment regimens and efficacy.

Interviewer – Annalisa VanHookTo put you on the road to some personalized medicine.

Interviewee – Adam SchrumYes, we think so.

Interviewer – Annalisa VanHookAnd not necessarily just for alopecia—this is just the test case that you've used—but potentially for other diseases as well.

Interviewee – Adam SchrumQuite right. We don't see any reason that this should be limited to alopecia areata. It's been a place for us to start.

Interviewer – Annalisa VanHookHow feasible would it be to use your method clinically? I mean, could it be adapted to be fast enough and affordable enough to use it to create treatment plans for individual patients?

Interviewee – Steven NeierWell, we definitely think this strategy is directly applicable and consistent with other technologies that are currently used in the clinic. It definitely has the benefits of the speed and the cost that would be consistent with current clinical assays being performed. Additionally, we think that the biosignatures and other network signatures that we're generating with this technology can be made more concise and more targeted, which will enhance this ability to translate this technology into the clinic.

Interviewee – Adam SchrumAnd so, if I might follow up on that, just on that last point, which is, sure, to do what we did we looked at a matrix of 20 proteins in a 20 by 20 matrix. Once one finds what the key signatures are, there may be four proteins in that. And then, you would create your clinical diagnostic or your clinical assay just on what you had learned. You don’t have to do all those proteins anymore necessarily, you just look at the relevant subset. So, it's got some real potential in the clinical arena, we think.

Interviewer – Annalisa VanHookAnd that relevant subset of proteins that you look at would depend upon the particular disease to which you wanted to apply it and the signaling pathways that were implicated in the pathology of that disease.

Interviewee – Adam SchrumExactly right. And, the discovery of those key proteins that create that signature involves looking at many more proteins than the ultimate key signature proteins would be. At least, if we're to take this lesson learned from what we've done so far, that would be it: that it ends up being a subgroup. And once we identify that, then you would target that for further clinical workup.

Interviewee – Steven NeierAnd just to add one more thing to that, even if the network signature depends on a much bigger network, the system and the format is directly compatible and applicable to generate kind of higher order network signatures that might be beneficial for subtyping and looking at some nuances between treatment strategies. So even if we were to expand to a broader network, we think that this would have interesting implications in the clinic.

Interviewer – Annalisa VanHookAdam and Steven, thanks for speaking with me.

Interviewee – Steven NeierThank you so much.

Interviewee – Adam SchrumThanks very much; it's been a pleasure.

Host – Annalisa VanHookThat was Steven Neier and Adam Schrum discussing a paper from the August 2nd issue of Science Signaling by Smith and colleagues (1). You can read that article online at stke.sciencemag.org.


The Science Signaling Podcast is a production of Science Signaling and the American Association for the Advancement of Science—Advancing Science, Serving Society. If you have any comments or questions, you can write to us at sciencesignalingeditors{at}aaas.org. I'm Annalisa VanHook, and on behalf of Science Signaling and AAAS, thanks for listening.

Educational Details

Learning Resource Type: Audio

Context: High school upper division 11-12, undergraduate lower division 13-14, undergraduate upper division 15-16, graduate, professional, general public and informal education

Intended Users: Teacher, learner

Intended Educational Use: Learn, teach

Discipline: Biochemistry; bioinformatics, genomics and proteomics; biotechnology, human biology, immunology

Keywords: Science Signaling, alopecia areata, autoimmune disease, baldness, diagnostic tool, GADS, GRB2, LAT, linker of activated T cells, personal medicine, protein-protein complex, T cell receptor, TCR signalosome


View Abstract

Stay Connected to Science Signaling

Navigate This Article