Q&A with BLUEPRINT’s Henk Stunnenberg on the New Leukemia, Blood Epigenome Project

The European Union’s $53.5 million, 4-year BLUEPRINT project launched this week with the ambitious goal of cataloging the epigenomes of 100 different cell types, with a focus on hematopoietic cells and leukemia disease states.

The project is only the second such effort associated with the International Human Epigenome Consortium—the first being the U.S. National Institutes of Health’s Roadmap Epigenomics Mapping Consortium—and it’s composed of 41 research entities, including university labs, research institutes, and drug and diagnostics companies.

BLUEPRINT coordinator Henk Stunnenberg, a professor of molecular biology at the Radboud University Nijmegen in the Netherlands, says the project may publish early work as soon as a year from now, and that the epigenomes it uncovers have the potential to generate diagnostics and first-stage compounds for drug development. BLUEPRINT is also going to try out new technology, including different approaches to sequencing some of the more recently discovered epigenetic marks, such as formyl-C, as it records the patterns of methyl-C, non-coding RNAs, and other, more familiar epigenetic features in blood.

I had a brief chat with Dr. Stunnenberg this week to ask him about BLUEPRINT’s details, its focus, its technologies, and his hopes for its outcome. Here’s our conversation, lightly edited for the web.

What would you tell epigeneticists and non-epigeneticists about this project and its impact?

It’s an exciting time to be doing this type of work, and it will be giving us the blueprint of our regulation. The genome hasn’t told us everything yet. It’s given us the building blocks, and now we need to know how to use those building blocks to make buildings. That’s why the consortium is called BLUEPRINT.

It’s wonderful to be helping to develop this, and to help at the various levels—translationally, scientifically, and as a resource for the community.

Is the list of companies and organizations involved in BLUEPRINT set in stone?

No—it’s certainly not. We had two intentions. The first is that we will fund new groups to add to the activity, probably toward the second half of the project, once it becomes clear what we are delivering. But also when new technology comes up, so that we can incorporate that quickly and invite people to join. We have funds reserved for that, and we expect to have an increase [in consortium size] over the years.

So, [BLUEPRINT] will open up to the outside world once we’ve started. Not now, because it’s a big enough group of people, and we need to get it going. Once that’s achieved, then we can start looking at what else we need to complement, or what part of our activities started with good ideas that aren’t working out, so we can stop those activities. That can also happen.

To avoid having a sort of exclusive club, we will have associate members. They won’t get funding for their research, but they’ll be invited to meetings, they’ll have early access to data.

How long until new members can join?

We will probably have a call for that, and they’ll be approved and selected by our scientific advisory board, which will do a prioritization. We think we’ll get there by the end of the first year—we’ll have that out and we can start integrating in middle to end of the second year. The third year at the latest.

How long until the project can deliver results?

We have a number of functional pipelines already—we have one my lab, one in Barcelona for methylome sequencing—bisulfite based—there’s a group in Berlin that will do the RNA. So basically now [it involves] collecting the samples from the clinicians, from the blood banks, from the cohorts, and start.

We hope to have the first few epigenomes defined by, say, a half year for the work, the writing, and the analysis—about a year from now we think we’ll have the first couple out.

Why hematopoietic cells and leukemia disease states?

First, one could say that it’s low-hanging fruit because we have easy access. I think it’s also clear that a large number of diseases including cancer manifest in blood. We’ll be able to get normal primary cells from healthy individuals, as well as comparisons from a large number of [from individuals with] leukemias, lymphomas, but also diabetes, and so on.

There’s a lot of knowledge on cell-surface markers, so purifying and sorting is a good way to go, so that we can have relatively homogeneous populations—with the emphasis on “relative.”

Also, a large number of genome-wide association studies, linkage disequilibrium studies—everybody’s using blood. It’s the only quickly accessible material. Whether they’re always reflecting the disease in the analysis of blood cells—that’s a major question. We don’t know.

I think it’s very important that we define the “space” of blood from the hematopoietic stem cell down. It’s the system for developmental and differentiation control, because we can start from the hematopoietic stem cell and take the lymphoid myeloid lineage and go all the way down to the mature cells and define as many steps in-between—so it will give a very interesting picture of that process of differentiation at the epigenetic level. And that’s a system where we have many different decisions down the branching of the lineages, and we can study the process.

I think it’s very important as a proxy for a large number of studies. We need to define what the variation is between cells, or between individuals. Like when you look at GWAS or linkage disequilibrium-type analyses, you need to know what the variation is. And you need to know that at the epigenetic level, because that’s where we can say what function is.

In about the past year, it has become apparent that a large number of those loci that come out of GWAS analysis are not linked to genes, are not linked to promoters—and that’s often where it stops. We have a locus, we have a SNP, but genetically it’s in the middle of nowhere. Now it turns out that a very large proportion of those sites with a good association are turning out to be regulatory units at the epigenetic level.

By overlaying what we find at the epigenetic level with what’s known in the large number of cohorts—who’ve all used blood as a proxy—we can really start correlating there and start filtering out those loci that are informative from those that, at the time being, are not informative.

It seems like you’re mostly defining the lay of the land–what do you optimistically hope to come out of this? Possible therapies?

It’s a bit dangerous—we don’t want to raise hopes and say, “In five years, we’ll be able to do this and that.”

We have in the project a small drug-discovery line, which will be very early exploratory work involving the company Cellzome. They had a paper out just this Monday in Nature when we had the kick-off, in collaboration with big pharma and another academic group.

What we think we can do is define new compounds that prevent readers from interacting with their marks on chromatin. At the moment, and all of a sudden very popular, are bromodomains—it’s a big family that recognizes acetylated lysines. There are several “hits”—not drugs—that came out in the past half-year to a year, and big pharma is jumping on that.

That has also triggered analysis of many of the other readers, other domains that are interacting with and recognizing epigenetic marks. There’s a lot of activity within BLUEPRINT that will direct toward other proteins with domains that recognize methylated lysine or methylated arginine, for example.

Let’s be very optimistic and say that in 10 years there is a compound out there–but certainly new “hits,” new compounds will be identified—and that won’t take five years, they’ll be faster.

Are there diagnostic companies involved in the effort?

Yes, but mostly in the second phase, because we first need to define which are the potential biomarkers before we start spending money on companies without data or analysis to work with.

But yes, biomarkers will come in when the first series of experiments are done. But I’d say that’s more toward the second half.

Currently we need to get to the right technology, so we have to benchmark a number of the technologies that are out there, how reliable they are—and there’s a large number of ways to gauge DNA methylation, for example—that’s one of the activities we’ll start immediately so that we’re ready when the biomarkers are coming and we have a serious thing to work on.

So the BLUEPRINT project is going to involve some less-developed technologies in addition to standard technologies, such as bisulfite sequencing?

The bisulfite [sequencing] is still the golden standard. For the pipeline, for the genome resource, you can’t integrate technologies that are in very early phase development. They have to be validated procedures.

But there’s a project in there that looks at the new forms of DNA methylation. So hydroxymethylation, formyl-, carboxyl-. It’s a hot area at this point. The technology there is still very immature. There’s quite a big discordance among the data generated up to this point.

And clearly new technologies, like the next-next-generation sequencers are around the corner, and one of the companies is involved: Oxford Nanocore Technology—basically the European competitor of PacBio. And they promise to be able to read out modified DNA bases, the new ones, the existing ones that we have, at long reads. That would be very helpful for biomarker development and validation.

But certainly there’s quite a big project part that will focus from the exploratory down to [large numbers of specific] loci—that’s what you need for diagnostic prognosis, large numbers. So we have some ideas that are going in that direction.

[BLUEPRINT Hematopoiesis figure reproduced here with permission of Dr. Stunnenberg.]

And here’s that Cellzome paper:


Dawson, M., Prinjha, R., Dittman, A., Giotopoulos, G., Bantscheff, M., Chan, W., Robson, S., Chung, C., Hopf, C., Savitski, M., Huthmacher, C., Gudgin, E., Lugo, D., Beinke, S., Chapman, T., Roberts, E., Soden, P., Auger, K., Mirguet, O., Doehner, K., Delwel, R., Burnett, A., Jeffrey, P., Drewes, G., Lee, K., Huntly, B., & Kouzarides, T. (2011). Inhibition of BET recruitment to chromatin as an effective treatment for MLL-fusion leukaemia Nature DOI: 10.1038/nature10509