At A Glance
Name: Erich Nigg
Position: Director, Department of Cell Biology, Max Planck Institute for Biochemistry, since 1999; honorary professor of cell biology, Ludwig Maximilians University, since 2000
Background: Postdoc, department of biology, University of California, San Diego — 1980-1982; PhD, biochemistry, Swiss Federal Institute of Technology — 1977-1980
Erich Nigg of the Max Planck Institute for Biochemistry in Munich is one of the world’s foremost experts in the mechanisms that control chromosomal segregation during cell division. At last week’s American Society for Biochemistry and Molecular Biology meeting in Boston, Nigg delivered a presentation about abnormalities in chromosome segregation, their links to disease, and their promise as therapeutic targets. Inside Bioassays caught up with Nigg at the conference to hear some additional thoughts about this work.
How did you become interested in this field of study?
I worked on kinases for a number of years, and the cell cycle is one of the places where kinases play a prominent role. The first decade or so of [my work] in this area was devoted to cyclin-dependent kinases which had been discovered by other people. But then we eventually moved into other kinases that are more related to mitosis, per se. And that’s where we work now — other types of kinases and the cyclin-dependent kinases.
You gave a lecture at the ASBMB conference about some recent major findings regarding aneuploidy? Can you give an overview of these?
The fact that tumor cells are aneuploid has been known for decades. And it’s also been known for a long time that there’s a general correlation — the more scrambled the karyotype of a tumor, the more aggressive the tumor is. Yet people thought for a long time that this is basically just a nuisance — that the genetic aberrations you see are simply hiding a few critical changes in tumor suppressor genes and oncogenes, and that you have to pinpoint those and then target the tumor to it.
But now, in the last few years, more and more people have started believing that this plasticity of the genome that you see is not just a nuisance or a bystander phenomenon, but it’s actually important for the biology of the tumor. And I think you can make three plausible postulates. One is that this mutation of this tumor and its ability to rearrange its genome very quickly will favor the odds of having the necessary types of mutations come together in a single cell. So if you think of a human tumor and the epidemiology, people have been saying for years that it takes four, five, six, or seven mutations — something on that order — for a human tumor cell to evolve, and to transform a normal cell into a tumor cell. Now what are the odds for getting these mutations together in a single cell? That’s hotly debated, but basically if you have a genome that is undergoing mutations at an increased rate, you will increase the chances of getting the mutations together in a single cell. That’s number one.
Number two, which I think is even more important, is the fact that once you have a population that has a plastic genome, then as the conditions in the tumor change — because the tumor’s growing, and now there’s hypoxia, and there’s nutritional limitations and so forth — but now you can select subpopulations of cells that have a particular genetic constitution that allows them to thrive better in this new environment. So you have a chance to select out for new clones that have best adapted to the environment.
And finally, the same applies if you’re conducting treatment. As you throw drugs at it, again, if you have a genome that is rigid, then all the cells will respond in the same way. But if you have a genome that is plastic … then again, you will have tumor cells that will resist whatever treatment you try to use against them. This is what I think — I don’t think this hard fact at the moment, this type of thinking, to me, is enough to ask the question: Where does this chromosomal aberration come from? Now if you look at it you basically see two kinds, and they’re often mixed in the same tumor. You see structural instability — the chromosomes are often fused together in bits and pieces. And you see numerical instability — so entire chromosomes get mis-segregated. You have an odd number from the normal diploid genome.
So which process is your lab interested in?
My own lab is not so interested in the structural aberrations, but the numerical aberrations. Structural aberrations, though, I think will come about as a result of telomere erosion … Numerical aberrations, I think, are most likely due to a deregulation of synchrony between chromosome segregation and cytokinesis. If you cleave the cell at the wrong time, then the chromosomes have not yet separated properly, then of course you will end up with uneven or unbalanced configurations. Or, [you could] have problems at the spindle sentry checkpoint, which is a checkpoint that makes sure all the chromosomes get patched into the spindle apparatus, or, [you might] have problems with the centrosome cycle. The centrosome … in mitosis … basically duplicates, and it forms two poles. So you have two poles that you segregate into two cells. But if you have multiple poles, then inevitably you will [miss] chromosomes.
Which of these possible causes do you think have potential as possible drug targets?
I sort of skipped a bit over what findings we have contributed — I only told you what the background is, so let me just complete that. So I’ve been working on those three areas. Here’s where the kinases play a major role. They play a major role in the centrosome cycle, and there, one thing that we recently found is that we think the fact that many people see excessive numbers of centrosomes in tumors can come about through multiple mechanisms. It is an important parameter in the tumor’s biology, because once you have multiple centrosomes, you will mis-segregate chromosomes — it’s almost inevitable. But we think that often times these multiple centrosomes arise through division failure, rather than through [deregulated] duplication. That means that the cell that has these multiple centrosomes will actually have a tetraploid genome — that is to say it will have double the normal chromosome number. I think the chances of having viable and potentially harmful cells out of divisions where you start with a tetraploid genome are substantially increased over the chances if you start with a diploid genome.
So about your question of how to approach this as a drug target, I honestly think it’s very early in this research — we’re still conducting basic research on this. But what is clear is that the mitotic kinases like the [polar] kinase and the aurora kinase — at the moment they are considered in many companies as attractive drug targets. In the case of aurora kinase, it is overexpressed in many tumors — that is clear. And there is also evidence that this is causally linked to tumorogenesis in the sense that there is evidence for aurora kinase being a tumor susceptibility gene. So there are good reasons you could say: OK, if you target aurora kinase with a drug, you may preferentially block tumor cell proliferation. In the case of other kinases in mitosis, the evidence for deregulation in cancer is less strong, although there is anecdotal evidence in favor of that view. But it’s just another way to block proliferating cells. And to compare to drugs that have been successful in clinical applications, like Taxol — that targets microtubules, and has side effects in the brain and in other areas. The idea of blocking mitotic kinases that are essential for cell division is plausible. I mean, it’s not going be a wonder cure in the sense that these will not have side effects ... because these are required for all cell proliferation. And so the therapeutic window has to come, to some extent, through serendipity. I don’t think it’s a rational argument to say: This will be a perfect drug that only hits tumor cells.
The spindle checkpoint, I think, is also a very attractive issue — this checkpoint that makes sure all chromosomes get attached to the spindle before they get segregated. There is evidence, although not terribly strong, that tumor cells … may have a weakened checkpoint. So you could argue that if you weakened that checkpoint even further, that you could cause such an extensive chromosome mis-segregation that no viable cells would come out of mitotic division. So you would kill cells not so much by apoptosis, but by what is called mitotic catastrophe, or a lethal mitosis. And I think this is a wide-open field that will gain prominence in the years to come.
Have you or anyone in your lab screened any compounds against these possible targets?
No, we haven’t. We are very basic scientists, and we want to figure out how things work, and we leave it to the pharma industry to make the drugs.
What type of cell biology tools and instrumentation do you use to investigate these pathways?
We do a lot of imaging, obviously, fluorescence microscopy, and RNAi nowadays. And we do a lot of biochemical assays that look for protein interactions, but not through any high-throughput or hybrid screens, but really targeted coimmunoprecipitation experiments from endogenous or in vivo situations. And then we go through mass spectrometry to identify proteins.
It seems like in this type of work it is almost essential to conduct cell-based assays. What role does this play in your research?
Definitely, there is one experiment that we did a while ago which was to inhibit kinesin-related motor proteins with antibody injection, and that gave such clear-cut mitotic arrests that later on people went on and developed drugs on cell-based assays that used the exact same phenotypes, and those drugs turned out to inhibit the exact same modes that we had inhibited with antibodies. There are certainly biotech and pharma companies that are using this approach. I think the major problem that people are discovering in this type of approach is to go from a drug to actually then define a target. It’s not too difficult to find drugs that give you very interesting-looking arrests and phenotypes. But then to go from there and actually fish out which proteins these drugs are blocking turns out to be more difficult than maybe people thought it would be.
So that would insinuate that you have to have to look at doing both biochemical and cell-based assays?
I think there are arguments on both sides. If you start out with a preconceived target, let’s say a kinase or a motor protein, and you go and screen for drugs and inhibit that target, then of course you’re going to face downstream problems that the drug may not be acting as well in vivo, or it might not be as specific as it could be. So you have problems on that side. If you start with a cell-based approach, you right away see that you have a nice, tight block, but you don’t really know what the target is. The more knowledge you have about the process you’re cell-based assay is interfering with, the better off you are at making guesses as to what the target might be. If you have no clue about what’s going on, you’ll never find your target. That’s why you should have a basic understanding of the process, because you can then make predictions about what you might hit in your assay.
And that’s what you’re in the business of doing?