Skip to main content
Premium Trial:

Request an Annual Quote

How Does Merck Use siRNA Screens to ID Drug-Enhancing Genes?

Name: Steven Bartz
Position: Research fellow, cancer biology/molecular profiling, Merck Research Laboratories
Background: Senior research biologist, Merck Research Laboratories — 2001-2004
Postdoc, Fred Hutchinson Cancer Research Center — 1995-2001
PhD, medical microbiology and immunology, University of Wisconsin, Madison — 1995
BSc, microbiology, University of Wisconsin, La Crosse — 1989

At Merck Research Laboratories’ Rosetta Inpharmatics site, Steven Bartz performs siRNA screens to find genes and gene pathways that impact disease and drug efficacy, primarily in oncology.
Last month, Bartz and other researchers from Merck published data in Molecular Cell Biology showing how an siRNA screen can identify genes that enhance the toxicity of the chemotherapeutic agent cisplatin when knocked down.
This week, RNAi News spoke with Bartz about his research and the findings published in Molecular Cell Biology.
Let’s start with what you do at Merck.
My group does siRNA screening basically. I have a group of about eight people and we do a lot of focused screens, as well as subgenomic screens. We [also] take advantage of sending those screens to our … screening facility in North Wales, [Penn.], that has a lot more of the higher-throughput robotics [compared with] what we have. We can easily screen 2,000 to 4,000 genes here, but genome-scale screens we tend to send to them. What we end up doing is covering a lot more … assays in terms of approaches or pathways. The ones that look like they have a higher probability of success, [we] send off for genome-scale screening.
Do you primarily run your screens in areas of interest for Merck from the pharmaceutical side of things to get more clarity on drug targets?
Yeah. It’s therapeutically relevant target areas. There are a few other groups around Merck that will do other therapeutic areas, but we tend to focus on oncology, although it has branched out recently.
Branched out into where?
We work with an immunology group, and there is a diabetes group involved. Also bone biology. There are pathways that kind of cut across oncology, bone biology, immunology, and diabetes. They may be interested in genes that up-regulate a particular pathway, while oncology [groups] may be interested in genes that down-regulate it. But there is interest in the pathway from different therapeutic areas, so we tend to focus on those.
You just had a paper come out in Molecular Cell Biology detailing one of these screens. Can you talk about what you did and what you found?
To preface this, the other type of screens that we do are drug-enhancer screens. That is relevant from two perspectives: one is that you can identify genes that enhance the activity of a drug, and those may become targets themselves, such as the ones in the paper for cisplatin — if you can enhance the activity of a chemotherapeutic that is used heavily in the clinic, particularly in just the tumor cells, then [the gene involved] is potentially a target itself.

The other way to use [drug-enhancer screens], as the paper suggested, [is to use the] genes that enhance activity … in predicting who will respond and who will not respond [to a particular treatment.] The example we gave in the paper was that for cisplatin, the BRCA family was one of the enhancers, and the literature suggests that patients with BRCA mutations have better responses to cisplatin.

To get more into the paper, what we did was to do drug-enhancer screens from the perspective of one, trying to find new targets and two, trying to find new targets that were selective in particular tumor contexts. The first question was, would they work, because in yeast when they’ve done drug-enhancer screens, it’s been over multiple generations — they typically go 20 to 30 generations in order to find the drug enhancers. Obviously, we’re not going to be able to do that with an RNAi screen so we typically do 72- to 96-hour assays with three or four doublings at the most. So one question was whether the window would be big enough to see anything. The window is there, but it’s quite small. … You can do them, but it means that you’re just dealing with a bit of noise in the assay.
The nice thing was one, that we were able to do [the screens] and two, when we looked at different drugs the enhancers were specific for each of the drugs and could be linked to the drug’s mechanism of action. For example, with cisplatin, we get out pathways involved in homologous recombination and translation, [and] DNA repair. So we were able to pull out pathways that were involved rather than just individual genes. For gemcitabine, we got out genes that regulate nucleotide metabolism, and for [paclitaxel] we got out genes that are involved in mitotic checkpoints, for example.
The other nice thing was that, at least for cisplatin, we pulled out genes that have been found in model organisms such as yeast or Drosophila. … So pathways that are evolutionarily conserved we were able to identify in the siRNA screen, as well.
Once we had validated that the windows were sufficient, at least on the small scale, to get hits out and [determined] that they were relevant, we published on the genome screen we did for cisplatin [in the Molecular Cell Biology paper]. By going to the genome-scale screen, we were able to pull out more of the pathway members that we didn’t have in smaller-scale screens. Also, the genome library that we have [now] contains a lot more of the novel genes.
In the paper, we tended to focus on validating hits that fell into known pathways because we were primarily interested in finding targets that enhance cisplatin. For the novel genes, there are certainly things we think fall into some of the pathways we’d identified. We’re working on validating those, but one of the most time-consuming steps is the validation, and validation of novel genes requires a lot more time than things that have some information associated with them. So for the novel genes, we’re still working on them.
For the pathways that came out, one of the questions we had was, “Were they selected in a p53-negative background?” because many tumors are p53 deficient. The reason that they might be was that p53 is heavily involved in regulating the integrity of the DNA, so when the DNA is disrupted then you get a p53 response. So we suspected that cells being p53 deficient may be more susceptible to cisplatin if these genes were targeted. In particular, we followed up on the BRCA pathway. There are previous suggestions that BRCA disruption would increase the sensitivity of a p53-deficient background, but they hadn’t looked in p53-plus and –minus cells simultaneously, and they only looked at BRCA1. So we made isogenic pairs using an shRNA targeting p53, validated those both phenotypically and using gene-expression technology, and then assessed the effect of knockdown of the BRCA pathway on cisplatin sensitivity. We found that cells that were deficient in p53 were more sensitive to cisplatin than p53-intact cells.
There is a lot of interest in ubiquitin ligases right now. The BRCA/BARD1 complexes ubiquitin ligase, so this suggests that you could increase cisplatin sensitivity by having small molecules directed at the BRCA1/BARD1 complex.
The other nice thing, relating back to the model organisms, is that we did identify the RAD6/RAD18 pathway also as enhancing cisplatin activity, and this is one that is well-characterized in yeast as being able to do that as well. So we got out two major pathways — the RAD6/RAD18 and the BRCA1/BRCA2 pathway. But at least in yeast screens you wouldn’t be able to pull out the BRCA pathway since it’s not present there. So there’s nice overlap between the model organisms and the model screens, and it shows the power of doing this kinds of screens directly in human cells.
You mentioned small molecules, but could RNAi molecules be used to do the inhibition?
Yeah. With Merck’s acquisition of Sirna [Therapeutics], obviously everything can be considered druggable using RNAi approaches. Certainly Merck is interested in fully taking advantage of that technology from a drug-target perspective.
So what’s the next step?
The next step is moving promising targets into drug-development programs. The nice thing about having done these screens internally is that you’ve established a lot of the cell-based assays that can be used to help drive the medicinal chemistry effort should some of these targets be picked up for development. The only thing that is not existing is the actual biochemical assay, but you’ve already got a lot of the infrastructure in place with the cell-based step, which wouldn’t be the case if you didn’t do the target discover internally.
The two components that come out of that are the value of using siRNA screens for target discovery and the value of doing drug-enhancer screens to identify pathways, as well as helping to identify pathways that may regulate the drug response in terms of moving towards personalized medicine. You can imagine using siRNA screens to help identify what patients may or may not respond [to particular drugs].
I suppose that would be useful not only once a drug is on the market but when you’re designing clinical trials.
Exactly. Both.

The Scan

Myotonic Dystrophy Repeat Detected in Family Genome Sequencing Analysis

While sequencing individuals from a multi-generation family, researchers identified a myotonic dystrophy type 2-related short tandem repeat in the European Journal of Human Genetics.

TB Resistance Insights Gleaned From Genome Sequence, Antimicrobial Response Assays

Researchers in PLOS Biology explore M. tuberculosis resistance with a combination of sequencing and assays looking at the minimum inhibitory concentrations of 13 drugs.

Mendelian Disease Genes Prioritized Using Tissue-Specific Expression Clues

Mendelian gene candidates could be flagged for further functional analyses based on tissue-specific transcriptome and proteome profiles, a new Journal of Human Genetics paper says.

Single-Cell Sequencing Points to Embryo Mosaicism

Mosaicism may affect preimplantation genetic tests for aneuploidy, a single-cell sequencing-based analysis of almost three dozen embryos in PLOS Genetics finds.