Name: Peter Lichter
Title: Director, division of molecular genetics, German Cancer Research Center (DKFZ), Heidelberg, since 1992
Experience and Education:
Group leader, DKFZ, 1990-92
Postdoctoral fellow, department of human genetics, Yale University School of Medicine (working with David Ward), 1986-90
PhD, Max-Planck-Institute for medical research, Heidelberg 1986
Undergraduate degree in Biology, University of Heidelberg, 1983
Peter Lichter heads the PedBrain Tumor Research Project, one of three cancer sequencing projects in Germany that operate under the umbrella of the International Cancer Genome Consortium.
Funded in 2010 with €15 million from the German Federal Ministry of Education and Research and the charity German Cancer Aid over five years, PedBrain focuses on the two most common types of pediatric brain cancer, medulloblastoma and pilocytic astrocytoma.
As the director of the division of molecular genetics at the German Cancer Research Center, DKFZ, Lichter is also involved in other cancer genome projects, as well as several initiatives that aim to facilitate the transition of cancer genome sequencing into patient care. Clinical Sequencing News recently spoke with Lichter at his office in Heidelberg. Below is an edited transcript of the interview.
Can you provide an update on the PedBrain project and its aims?
We stick very much to the aims of the ICGC and follow its rules. Pediatric brain tumors are the tumors with the highest mortality in pediatric oncology, following the big success of treating leukemias in children, so [studying them] is very important for clinical applications as well.
We work on the two most frequent ones: pilocytic astrocytoma and medulloblastoma. For medulloblastoma in particular, mortality and overtreatment of children is a big issue. We plan to do 300 cases of each, analyzing the full genome, transcriptome, methylome, small RNAs, and chromosomal rearrangements.
Funding started in 2010, so we first had to establish [standard operating procedures], pipelines, and methodologies. For a good year now, data has been coming in. So far, we have done 50 cases of medulloblastoma, which is much more than we expected to do in the first one and a half years. We also have data for the first 25 cases of pilocytic astrocytoma, which are not analyzed yet.
How have you sequenced these samples? And where was the sequencing done?
The medulloblastoma [project] started on the [Illumina] GAII and is now done on the HiSeq. For the pilocytic astrocytoma [samples], we are using two platforms. The first series of samples were done on the Life Technologies SOLiD instruments, and we will also use HiSeq for this project in the future.
Several facilities were involved in the sequencing: [our own] facility at the DKFZ; [some] of the cases were outsourced to [sequencing service] companies; and the [European Molecular Biology Laboratory] has also contributed. GATC Biotech was very much involved in the medulloblastomas. We are also looking into other companies, but so far, GATC has been a very reliable and good partner to us.
As we are splitting the tasks into different work packages that are headed by different people, we have, for some of the profiles, places outside of Heidelberg [that will contribute]. For instance, the RNA sequencing is done by the group of Hans Lehrach [at the Max Planck Institute for Molecular Genetics] in Berlin; the small RNA is done by a group [at the University of] Düsseldorf, Pablo Landgraf and Arndt Borkhardt; and the chromosomal rearrangement by mate-pair mapping is done by Jan Korbel at the EMBL. We are doing the genome and methylome here in my group at the DKFZ, together with companies.
How are you equipped with sequencing platforms here at the DKFZ?
At the moment, there are three SOLiDs, two 5500, and four HiSeqs, and there are more HiSeqs on the way. We also have a 454, and we will get some MiSeq machines for validation experiments, and possibly an Ion Torrent, that is not clear yet.
Last year, Life Technologies and DKFZ announced a partnership to create a genome sequencing facility with 10 SOLiD hq systems (IS 6/8/2010). What has happened with that?
We are currently looking into the new generation of instruments that Life Technologies is offering, in particular the 5500 and the Ion Torrent.
As far as I understand, you conducted a pilot project for PedBrain, using a small number of medulloblastoma samples. Did you find anything interesting?
[The pilot project] started with one sample, and we have added a few more cases. In terms of full sequence information, it's less than five cases. We have also looked at an independent cohort of patients by specific methods, not total sequencing, [to confirm our results].
We found many new things, which are going to be published soon. This will be the first paper out of the PedBrain project. One aspect has to do with a new form of genomic instability that was reported for the first time earlier this year by Peter Campbell's group at the Wellcome Trust Sanger Center. That phenomenon is called chromothripsis, and we have some interesting links of sub-entities, molecular changes in this particular form of genomic instability.
We have a number of other interesting issues, like correlations of the number of mutations with the age of the patients, which seems to be true only for pediatric tumors, and of course we found a number of [previously] undetected genes that are mutated that we are currently validating. Before they are functionally validated, we will not be able to report on individual genes.
What have been the greatest challenges with the PedBrain project so far?
The greatest challenge, like in all [similar] projects, is that we get many new mutations and many genes that are mutated. As long as the genes are known to be pathogenic in tumor disease, that's pretty easy to follow up. But the real issue is, what are we doing with all the other real mutations, which are only in the tumor and not in the non-tumor cells? We have to functionally validate them to find out whether they really play a role or whether they are what is called these days 'passenger mutations.' The distinction of driver from passenger mutations, that's the biggest challenge.
And also as a community, we do not have sufficient technology to do this in a high-throughput fashion, so we have to work on the development of at least medium-throughput functional tests for many genes.
What kinds of tests could these be?
Obviously, one could start with cell cultures, but eventually, you have to do animal models, and it will not work if you do one gene per mouse, that's simply not possible. So we have to think about broader applications of more than one gene per mouse.
Do you have all the samples you need for PedBrain?
This project is planned for five years, and in the first one and a half years, we were able to collect more samples than we anticipated. We see that the pipelines are filling faster than we expected, so we will have sufficient samples.
What do you think will be the first results from PedBrain that will have an impact on patient treatment or diagnosis?
For example, we have important contributions to a further sub-classification of medulloblastoma. There are at least four subgroups, that's the current consensus of the world leaders in this particular tumor entity. But within the subgroups, we still have different clinical behaviors, and it seems that we will have very good markers that would further subdivide that.
For instance, certain children are currently treated, let's say, by radiotherapy, based on a particular marker that exists now. But we can tell that this marker is not sufficient and you would still overtreat a number of children. We now have better markers, and you should not do radiation [therapy] on [some of these children]. I think this is one of the immediate applications.
Are you reporting back results to PedBrain patients or their doctors if they could be clinically useful?
For PedBrain, the consent form of the patients clearly states that we will not return information back. We have been discussing this issue heavily in Heidelberg because PedBrain is a research project, but in addition, we have strong activities to get sequencing into clinical applications.
Because of the problem of incidental findings, we actually formed an interdisciplinary project [called Ethical and Legal Aspects of Whole Genome Sequencing, EURAT] that is funded by the Marsilius Kolleg, [a center at the University of Heidelberg that promotes interdisciplinary dialogue and research]. [EURAT] includes geneticists and informaticians, but also theologists, ethicists, and professors of law, economics, human genetics, and pediatrics.
We are a very active group trying to formulate rules and recommendations for how to handle these things. There are many issues involved: issues of previous consulting, and how much one can do that; issues of whether one should store the entire information or only a selected part of it, et cetera. This is an ongoing project, and we hope maybe in a year to come up with recommendations that could be used at least at the national level in Germany, and maybe even beyond.
What activities are you involved in that aim to bring sequencing into patient care?
We have an initiative at the DKFZ called HIPO [Heidelberg Initiative for Personalized Oncology] to bring sequencing into practical applications in the clinic. Christof von Kalle from the National Center for Tumor Diseases, NCT, Roland Eils at the DKFZ, and I are the main people driving that. At the moment, it is still a research project, but it is planned to transfer into common diagnostic settings.
We have two projects, one on breast and one on pancreatic cancer, which are just starting now, in collaboration with clinicians here in Heidelberg, Andreas Schneeweiss on the breast cancer side and Markus Büchler on the pancreatic cancer side.
The first projects started with exome and RNA sequencing, but it might turn into whole-genome sequencing — that's not been fully decided yet but it's on the agenda. We will first establish the pipelines, so there is a routine to get fresh tumor material directly after surgery and to work [with that].
The ethical issues will be accompanied by a project from the NCT to see how patients think about the consent and how we should develop that. And of course, like everyone else, we have to think about how to streamline the bioinformatic analysis and about the storage issue, which is not trivial.
Will the sequencing for those projects be done at the DKFZ?
At the moment, it will be done here at the DKFZ core facility. In the long run, if it turns from research into clinical applications, I would foresee that it's done within the NCT.
What are the chances that cancer exome or genome sequencing would be covered by health insurance in Germany?
There are different views on that. I think eventually, it will be the case, because it is actually much simpler and cheaper to look, for instance, at the whole exome than to sequence a larger number of individual genes.
At the moment, certain diagnostic gene analyses are covered by insurance, and it will develop from there. Obviously, one of the major issues before one would start large-scale sequencing as a diagnostic tool is, one would need certified methods for that, and at the moment, the methodology is not there, there are still too many false positives, false negatives. At the moment, this is sort of a pre-diagnostic procedure that would need validation in each single case by focused sequencing efforts with certified instruments and procedures.
Several US labs have recently started to offer sequencing-based cancer gene panels for patients. It looks like you are taking a different route?
At the moment, yes, because we are still on the research side. We will see how much we will have to turn that into disease-specific gene panels, or whether whole-exome [sequencing] would be the diagnostic procedure itself. That depends on the methodology.