A team of researchers at the National Cancer Institute’s Molecular Targets Development Program published a paper in the current issue of the Journal of Biomolecular Screening describing an assay for identifying inhibitors of the oncogenic transcription factor AP-1 (activator protein-1), which is implicated in tumor promotion and progression. The assay used Förster resonance energy transfer (FRET) to measure the expression of a beta-lactamase reporter — a relatively novel approach.
The MTDP, housed within the NCI’s Center for Cancer Research, helps discover compounds that could work as probes for functional genomics or other experimental studies, as well as leads or candidates for drug development. About half of the MTDP’s staff of 24 people work in assay design and screening, while the other half work in natural products chemistry.
Last week, Cell-Based Assay News caught up with Curtis Henrich, a senior scientist at NCI contractor SAIC and co-author on the paper, to discuss the results of the study and to find out a bit more about the role of MTDP within the NCI’s research goals.
Can you provide an overview of the activities of NCI’s Molecular Targets Development Program and the role of cell-based assays within the initiative?
The Molecular Targets Development Program developed from an effort a number of years ago looking for natural product modulators of molecular targets — targets involved in cancer and HIV. Currently we work with the Center for Cancer Research intramural NCI investigators who have specific targets that they’ve implicated through their basic science in particular cancers, and then the goal is to identify compounds that can modulate those targets. Those can range from receptors and protein kinases and phosphatases to, in the example here, transcription factors. We have others where we’ve looked at the ubiquitin pathway, where we’ve looked at reporter gene expression, and so forth. So they’re kind of all over the place.
They include both cell-based assays and biochemical assays, although in recent years most of our assays have been cell-based.
Typically what will happen is a researcher in the Center for Cancer Research at NCI will have an idea for some way to develop reagents or drugs that target their specific protein of interest or enzyme or whatever, and they will come to us with that idea, and then we’ll work with them to turn that into something we can use in a screening assay.
The goal is not necessarily drug development, although obviously if we come across something that’s worth developing as a drug, we will do that, we’ll put that into NCI’s system to develop that. But we also consider it a success if we come up with good reagents that will help advance the science related to the specific targets.
That said, what typically happens is that a CCR investigator will have kind of the skeletal outline of an idea of what they can do to look for modulators of their activity, and then we will take that and turn it into an assay that can work in a high-throughput environment.
The process is truly collaborative. We’re not a core facility or a service group, we really are a collaborative group, and we work with the PIs to develop the assays and make sure that, first, we’re in fact measuring what they hope we’re measuring, and second, that they’ve got downstream assays in place or we help them develop those so that when we get hits, they’ve got something to do with them. They can sort out specificity or whatever it is they’re interested in downstream.
We also stay involved in the process after the screening campaign is completed. So we will typically be involved in secondary assays — both development and running those assays, data analysis, [structure activity relationship] analysis, those kinds of things. But what probably makes us most unique is, one, the collaborative piece that I just mentioned, but the other is that we have in our group what I would consider really world-class natural products chemistry.
Probably the major focus area in screening is natural products. We have a repository of pure natural products that have been previously characterized, and those are available from NCI’s developmental therapeutics program. And we run those in all of our screens, so if we find something, we already know what it is. But we also have access to about 150,000 natural product extracts, so our goal with those is to find extracts that have activity in whatever assay we’re doing. And then our natural products group will go in and do fractionation to find the active compounds in those extracts.
That’s not completely unique, but it’s a little bit unusual, particularly with reference to, for example, big pharma. They’ve kind of moved away from natural products over the years, and that’s not terribly surprising, from their perspective. But we have an opportunity to really focus on that and find, again, reagents as well as potential drugs.
In terms of this study to identify inhibitors of AP-1, what was the motivation for NCI in studying this particular mechanism?
The AP-1 project was brought to us by Nancy Colburn, who’s in the Laboratory of Cancer Prevention here at NCI Frederick, and she’s basically built her lab around this particular target. AP-1 is involved in both tumor promotion and cancer progression. So if inhibitors could be found to AP-1 activity, they could be very valuable in cancer prevention, though that’s obviously way downstream from where we are right now. But in the meantime, any reagents we’re able to develop out of the project will certainly be useful in further identifying and understanding the role of AP-1 in tumor progression.
So that’s another example of how this has been a collaborative experience. Nancy brought us the project; we worked with her to develop that into a screening project, and then we’ve gone through the screening campaign and now we have a series of compounds … to kind of move downstream to look at other assays of AP-1 and to look at cross reactivity with other transcription factors and so forth.
One thing that’s a little bit unusual about this particular project is that the first author, Katie Ruocco, was a student in the Johns Hopkins University master’s program in biotechnology. They’ve got a concentration in molecular targets and drug-discovery technologies, and NCI has supported that by offering a couple of fellowships, and Katie was one of the first fellows in that program.
Can you tell me a little about this particular assay in the paper? Why did you use the FRET-based beta-lactamase approach rather than another reporter system?
Typically the approach that’s most commonly used for reporter genes is luciferase. Another alternative that’s used pretty often is green fluorescent protein and the various analogs of that. We’ve done some assays with luciferase and have been basically satisfied with the results that we’ve gotten with them. But we had an opportunity to try a different system, and that’s this beta-lactamase system from Invitrogen.
It’s a fluorescence assay, which potentially gives you pretty good sensitivity, but the fact that it’s a FRET-based system as opposed to just straight fluorescence gives you some advantages, particularly with regard to the sort of compounds that we work with. We have a lot of compounds and a lot of stuff in our extracts that tend to be fluorescent. So anything we can do that will allow us to take advantage of fluorescence capabilities without significant interference from fluorescent compounds in our screening libraries is potentially useful.
And we found this to be a pretty good assay for that application. When we applied it in the presence of extracts we saw very little interference in the assay itself. And again, these extracts are aqueous or organic extracts from plants, marine organisms, fungi, all kind of stuff. So they’re really kind of a soup of a large number of potential compounds, and, not surprisingly, a lot of those are fluorescent, so it can be a significant problem for us in various assays. It turned out not to be a problem with this particular assay.
Now, that said, at about the same time we were completing a luciferase screen, which is also referenced in the paper. That worked pretty well, as well, in terms of the relative advantages and disadvantages of the two assays, they’re roughly comparable — or at least were in our hands. The costs were pretty similar, the throughput was pretty similar.
One of the advantages of the beta-lactamase assay, though, over the luciferase was that we were able to do a cytoxicity assay in the same wells. So with the luciferase assay, we had to do a cytotoxicity assay in parallel, so that doubled the number of plates that we had to run. And that’s a significant issue for these sorts of assays, particularly with the kinds of libraries that we have. Up to 10 percent of the extracts can be cytotoxic, depending on the target cells and so forth. We obviously aren’t interested in following up on compounds that are just killing the cells. We’re more interested in things that are specifically affecting the target of interest, in this case the AP-1. So if we can do a cytotoxicity assay simply and quickly without having to do it in parallel, that was an advantage of this approach.
Were there any disadvantages to this approach that you would need to overcome in the future?
It took a while to optimize the assay and to really nail down the reproducibility. But it’s interesting because once we moved the assay from early development mode using multi-channel pipettors and so forth and moved it onto robotics instrumentation, most of those reproducibility issues went away. So it’s a little finicky in terms of volumes and control of additions and so forth. It’s light sensitive, so we had to light-protect the assay plates as we were running the assays. But that wasn’t too big a deal. It’s not the sort of thing where it needs to be in complete darkness, but we did have to do some light protection.
But probably the biggest downside was during the development process. Like I said, it’s a little bit finicky in terms of volumes and concentrations. But once we moved it onto a robotic system and had control over volumes, we didn’t have significant problems there.
Would you say that those problems have anything to do with why this FRET-based approach isn’t used more often?
Well, it’s relatively new, as well. For us, the cost was roughly comparable to luciferase. I don’t know if that would always be the case for everybody. There may be cost issues for some people. There’s also sometimes a barrier to entry to trying something new. If luciferase works, and you’re set up to use it, why change? So I think that’s an issue as well. But it worked out well for us. For future assays for us, a lot of that will probably depend on how comfortable the individuals working on the assay are with the two approaches.
What’s next? You mentioned that you have several assays planned to follow up on this AP-1 work.
Those will actually be occurring in Nancy Colburn’s lab. We’ll be involved in providing the materials and characterizing the compounds that we find. Their next assays are, first, another assay of AP-1 — just to confirm that we’re seeing what we think we’re seeing in the screening assay. And then, second, there are a couple of other transcription factors that they want to look at — SRE and NF-kappaB. So they have assays set up for those targets. Those aren’t high-throughput, so the fact that they’re working with dozens of compounds instead of hundreds of thousands is obviously a significant advantage there.
The other piece that we’ll be getting to eventually — we’re not there just yet — is mechanism of action. In one of the tables in the manuscript, in terms of the validation of the assay, we listed a number of the compounds that we found in the screening that are already known to affect AP-1 but have different mechanisms of action. Some of them affect protein kinase C, some of them affect Hsp90, and some of them affect other kinases or AP-1 complex formation.
We’re not terribly interested in finding more PKC inhibitors, so that will be a filtering device downstream as well to sort out what the actual targets are. AP-1 is at the end of a long signal transduction pathway, so anything that affects anything upstream could potentially hit AP-1.
So the plan is to systematically target each of those molecules along that path?
Right. And then the other piece that we’ll be moving ahead with is that we have a number of extracts that we found activity in, so we’ll be looking at fractionating those. Our natural product group will be doing purification of the active compounds from those. And the hope is that we would find some unique compounds that either haven’t been described before, or haven’t been described with this sort of activity.
Is there anything else that you’re working on that’s of particular interest right now?
Well, most of the assays that we’re developing at this point are in fact cell-based assays, and we have one that’s basically been completed in terms of screening and now we’re following up with a series of natural products that our group has identified on ABCG2, which is a multi-drug-resistance transporter. We’ve got some interesting compounds in that regard, some that have been identified in the past, and some that are unique. So that’s kind of a developing story right now. We found the compounds and we’re working on additional assays to try to understand specificity and potency and so forth of these compounds.
But that should turn out to be an interesting story because we have some, at least in a couple of cases, families of compounds that have come out of the natural products group, so we might be able to come up with some interesting structural features for those compounds.