Nature Podcast 12 April 2007
Introduction
This is a transcript of the 12th April edition of the weekly Nature Podcast. Audio files for the current show and archive episodes can be accessed from the Nature Podcast index page http://www.nature.com/nature/podcast, which also contains details on how to subscribe to the Nature Podcast for FREE, and has troubleshooting top-tips. Send us your feedback to mailto:podcast@nature.com
Advertisement
The Nature Podcast from Nature, the international weekly journal of science brought to you in association with Bio-Rad. Find your perfect match among the full line of Bio-Rad thermocyclers and real-time PCR instruments — on the web at http://www.bio-rad.com/genomics
End Advertisement
Ben Valsler: This week, are these the genes which could make cancer cells more susceptible to drugs whilst keeping normal cells safe?
Michael A. White: This gene comes on in 70% of all ovarian tumours, but is not expressed in normal ovarian tissue at all. In fact, it is a male-specific gene.
Ben Valsler: Michael White tells us more, later on in this show. Also, John Trauger tells us how he can see distant planets that we could never see before.
John T. Trauger: So, we should be able to see Jupiter's and Saturn's orbiting nearby stars if we could simply take our laboratory experiment and launch it.
Ben Valsler: And we discover a possible mechanism behind intermediate-depth earthquakes.
Peter B. Kelemen: These rocks will heat up all the way to their melting temperature, and as they do so they become very, very weak and release a huge amount of stored stress, or energy, in a very short time.
Ben Valsler: Hello, I am Ben Valsler. Welcome to the Nature Podcast. First up this week, Chris Smith spoke to Michael White, from the UT Southwestern Medical Center in Dallas, about identifying those genes in cancer cells, which could enhance the effectiveness of drugs whilst keeping normal, non-cancer cells safe. Nature 446, 815–819 (12 April 2007) .
Michael A. White: The purpose was really two-fold for us. One was trying to help tackle this problem of how one eliminates cancer cells without having serious side effects on normal cells, and the other was to take advantage of this convergence of the knowledge of the entire human genome together with this new technology, this capacity to be able to inactivate any gene at will to be able to interrogate the function of that gene in any biological system of interest. This is a fabulously powerful way to approach the biology behind various problems having to do with human disease.
Chris Smith: So, in a nutshell then, the idea is to try to discover genes that make cancers nasty so that we've got novel targets to hit cancer with?
Michael A. White: Absolutely. The whole idea is the conditional dependency, things that cancer cells rely on to be able to survive, but may be less important to the normal cell.
Chris Smith: So, how did you do it?
Michael A. White: The context that we started from was with non-small-cell lung carcinomas and the problem that we have there is that one of the most commonly used drugs— Paclitaxel— has a very variable consequence in different patients. Often people who do respond have a relapse where that drug is completely ineffective and that is kind of a plantable hocus. So, what one would like to be able to do is identify mechanisms to make a drug that we already know is effective even more effective.
Chris Smith: So, how do you actually manage to flush out the Achilles' heel for these cancers and make these drugs more effective in these contexts then?
Michael A. White: The way we decided to do that was to use a cell line that was derived from a cancer that we know Paclitaxel can have an effect on, and then inactivate one by one every gene in the human genome to find those that when we take them away will selectively kill these cells, but only in the presence of the drug. This is known as a synthetic lethal screen. It's something that has been applied in classical genetic model organisms for a long time now, but now perceived genome-wide RNA libraries will have the capacity to do this in human cancer cells.
Chris Smith: And when you do this, do you find many hotspots?
Michael A. White: We actually found significantly fewer than one might expect. We went through over 21,000 genes and at the end of the day wound up with about 87, which had a very strong impact on the capacity of these cells to survive in the presence of that drug.
Chris Smith: And do you anticipate if you looked not at lung cancer, but to, say, breast cancer or ovarian cancer, or something similar, you would see a similar spectrum of genes?
Michael A. White: Well, that is a great question and I think that the answer is both. I think that there is going to be some genes that are very specific for one tumour type versus another. There is going to be some genes that are very specific for subsets even within a tumour type, but something I am really excited about is one of the most potent genes that we identified as really provoking profound chemosensitivity in these lung cancers is a gene that is normally not even expressed in lung cells, but only comes on in the tumours; and another tumour that this gene has been associated with is ovarian cancer. This gene comes on in 70% of all ovarian tumours, but is not expressed in normal ovarian tissue at all. In fact, it is a male-specific gene. And if we can specifically target that protein, this may have very little toxic consequences because it is simply not expressed in women.
Chris Smith: That was going to be the next question I asked you, which was —it is all very well identifying these targets, but if they render other cells in the body, which you do not actually want to be harmed by the chemotherapy, even more vulnerable to the chemotherapy, then obviously you would magnify the side effects wouldn't you?
Michael A. White: Yeah, that is a great point and that is one reason why we try very hard to examine the consequences of knocking down these genes in normal cells as well as in tumour cells; and a variety of the targets that we found turned out to have absolutely no negative impact in cell cultures from normal lung cells, whereas they had a strong impact in cell cultures from tumour cells.
Chris Smith: What do these genes actually do?
Michael A. White: That is another interesting question. They fall into a few classes. One class of genes that we identified has known to support mitotic progression in cells, that is the ability of a cell to be able to replicate its DNA and distribute it effectively in the daughter cells so that they can continue to proliferate themselves. Another class we found is important to turn over proteins that need to be removed in cells after they served their useful function. This is a machine called the proteosome, but a large class of these things that we identified have completely unknown functions in cells, and that is going to generate a lot of investigation from our labs and many others to really figure out what these proteins do in normal cells.
Chris Smith: And I presume that now you have got this identified in the dish that probably you are itching to get into in vivo work to try and work out whether you can exploit these loopholes, and see whether you can sensitize cancers to things like Paclitaxel again?
Michael A. White: Yeah, so you really hit on a fabulously important point, and that is that everything that we have done so far has been in the context of tissue-culture cells in a plate, and we really need to find out how important these things are in the actual tumour and move into model systems where we can address those questions.
Ben Valsler: Michael White discussing how knowledge of which genes are active only in cancer cells will provide specific targets for anticancer drugs. But most people who die from cancer are often not killed by the primary tumour itself, but by secondary spread into other organs, known as metastasis. Intriguingly, some cancers show preferences for spreading into certain specific organs, which suggests there might be genetic mechanism behind it. Identifying it, of course, could lead to novel ways to block the spread of certain types of tumour. To find out, Joan Massagué injected mice with cells from human tumours and tracked where the cells spread to and which genes were critical for making it happen. Encouragingly, they have also now found combinations of drugs which can block these genes and potentially hope the spread of certain cancers. Nature 446, 765–770 (12 April 2007) .
Joan Massagué: We want to understand better how it is that tumours that start in one particular organ or tissue in our bodies gain the capacity to spread to other tissues and other organs. What are the tools that the tumour cells employ to gain access to different organs?
Chris Smith: So, how did you try and get to the bottom of that question?
Joan Massagué: Well, we took cells from the breast-cancer patient, who has many metastases in her bones and in her lungs and in her brain. And so we said, look, if cells have managed to colonize these various organs, they must have found the ways or tools by misusing different genes to invade those particular organs. So, let us take this mixed pack of metastatic cells and see if, when we inoculate them into a mouse, the different cells will in turn invade the different organs.
Chris Smith: So, in other words, you are going to see invasion in the mouse recapitulating exactly what happened in the human?
Joan Massagué: Indeed! We asked would the mouse do us the favour of allowing its organs to be invaded the way these cells invaded the organs in the patient, and that worked. So now we could compare the cells to each other and ask, well, these cells come from the patient, from the same breast tumour originally, but one group of cells they have managed to find a way to go to the lung, one goes to the brain, let us see what is different between them. Looking for these differences at the level of how these cells use and, in fact, misuse certain genes, we found a group of genes that allow some of these cells to colonize.
Chris Smith: And what genes are they? What do they do, those genes?
Joan Massagué: Most of these genes encode proteins that the cell is using to deal with its environment, its neighbours, and that make a lot of sense because colonizing a particular organ, it is all about gaining access to forbidden land. A breast cell of course has no business invading and colonizing the lung or the bone marrow. So, it makes sense that the genes that the cells are misusing in order to insinuate themselves, penetrate and colonize, are genes whose products create the conditions for that cell not to be eliminated by the new environment. So, we find that four of these metastasis genes collaborate by making products that allow the mammary gland to form new blood vessels to feed the growing tumour.
Chris Smith: What about if you look at all the cancers, because other cancers spread of course, do they have a different set of genes, or can you see similar genes doing similar jobs in different cancers?
Joan Massagué: Both, we have in fact the same cancer using different genes to go to a different organ. We find different cancers using some of the same genes to go to the lung. Our sense is that this is basically a toolbox of genes that different tumours use at their convenience. Whatever gives them an advantage, they will use it.
Chris Smith: And now you have got those four genes, you know what the gene products are, can you neutralize their effects with, say, drugs, so that if you had a patient with this kind of breast cancer you could give them that cocktail of agents, and therefore abolish the prospect of spread to at least one organ, the lung for example?
Joan Massagué: Yes, indeed. This is possible, at least in the mouse. We became aware that there were a number of drugs that were designed to block the action of the products of these genes, and these are drugs that are already in clinical use for other indications. Now that we know that these genes are used in combination by these tumour cells, it became immediately obvious to us that we might use these drugs in combination to attack the problem and, at least in the mouse when we tested the drugs in combination, we saw a very, very marked reduction in the rate of tumour growth in the lung, and so that provides a great basis for clinicians to begin to think about how to test this combination of drugs in patients.
Ben Valsler: Joan Massagué, from the Memorial Sloan-Kettering Cancer Center.
Jingle
Nature's podcast, bringing the world of nature to life.
End Jingle
Ben Valsler: And now to the search in deep space for planets like our own. A major problem with spotting these so-called exoplanets is that they reflect only a tiny amount of light, which is very difficult to see amidst the glare from the parent star, which is millions of times brighter. To solve this problem, John Trauger from the Jet Propulsion Lab at the California Institute of Technology has designed a system which can be mounted on to a space telescope, which should make it possible to see planets the size of Jupiter. It uses a coronagraph to blot out the star itself and then a deformable mirror which can be tuned to within one ten-thousandth of a wavelength of light to achieve unprecedented resolution. Nature 446, 771–773 (12 April 2007) .
John T. Trauger: What we are looking for is a method to image nearby planetary systems that would be orbiting the stars in our galaxy from space. So, we are extending what we have learnt how to do with the Hubble Space Telescope and we have to do something that Hubble can never do, which is to improve the optics to a very high degree so that we can reduce the glare that is normally surrounding a star.
Chris Smith: So, if you were to compare how bright the planets you are trying to spot are with how bright the parent star is, what is the differential there?
John T. Trauger: Well, the differential is huge. If we were to look back at our Solar System from a nearby star, we would see that Jupiter is about a billion times fainter than our star and the Earth is about 10-billion times fainter than our star.
Chris Smith: So, you reckon that you can surmount that problem?
John T. Trauger: Well, that is what we are demonstrating with this experiment that we have reported. It primarily relies on being able to correct the optics with what is called active correction, a deformable mirror which we can very precisely adjust, and also something called a coronagraph, which is a pair of masks designed to eliminate diffracted light. The first mask is inserted where the image of the star first appears. That blocks most of the star light, but it is not much bigger than the star image and a lot of the light goes around it. So, there is the second mask called the pupil, where that scattered light is scrapped away.
Chris Smith: Where does that glare come from, that speckle that you get from the optics? Is that just, say, deformities or errors or lack of precision in making the instruments in the first place?
John T. Trauger: Well, there are two main sources of this light that we need to get rid of. One is the diffraction from the boundaries of the telescope mirrors that spread light into rings and spikes. That diffracted light is removed with a coronagraph. If we have done that, we still have the problems of imperfections in the objects that may be as small as a ten-thousandth of a wavelength of light, a fraction of the diameter of an atom.
Chris Smith: So, how do you get rid of those?
John T. Trauger: That is what our precision deformable mirror is for. It can be adjusted to that kind of accuracy, a ten-thousandth of a wavelength of light, to clear out little scattered light all around the star at a very low level. So, we should be able to see Jupiter's and Saturn's orbiting nearby stars if we could simply take our laboratory experiment and launch it.
Chris Smith: Can you get any smaller than Jupiter? Because obviously we know planets like Jupiter, we can spot those already at a pinch, but they are not terribly hospitable places. People are interested in finding potentially habitable planets like our own. Can you spot those?
John T. Trauger: Sure. Let me go back one step. We have not yet imaged a Jupiter that is anything like our Jupiter, a mature planet. So, it is still really worthwhile to get a sense for what a planetary system looks like, but, yes, the idea is to go for planets as faint as the Earth and that in fact is what we have demonstrated.
Chris Smith: How have you actually demonstrated it? Because you were saying at the moment you felt this as obviously a prototype, you have not been able to see anything yet in terms of actually dissecting planets away from stars. So, how are you showing that it is going to work?
John T. Trauger: Well, in the laboratory we can simulate the conditions, but primarily we are applying a very bright simulated star and creating a dark field around it, and we can measure how bright that dark field is. We know that if a planet much fainter — which would otherwise look just like a star — were to appear in our dark field, our dark field is faint enough that we will see it.
Ben Valsler: John Trauger, with a new system capable of spotting Jupiter-sized planets orbiting other stars, and hopefully we will see that technology launched and imaging new planets for us in the near future. For now though, we look back inside cells as Michael Resnick and Francesca Storici explain how they have discovered the pieces of RNA can be used to repair damaged DNA. Nature advance online publication (11 April 2007) .
Michael A. Resnick: The question we have been interested in is how DNA damage can be repaired, and a particularly important type of damage is damage to both strands of a chromosome called double-strand break, and it was known in the past that this could be repaired by interactions with another DNA molecule — a chromosome or sister chromatid. What we have speculated was, because RNA can pair with DNA, maybe at a break site beside the break it could interact with complementary RNA. So, we have searched for RNA that can actually comes in and repairs the double-strand break. Now, the RNA that we have used is RNA that is added from outside the cell, comes into the cell and it repairs the break. So, right now it is a proof of principle of what might take place inside the cell.
Chris Smith: This also depends on you knowing the sequence for the area where the damage or the break has occurred, doesn't it, so that you know what sequence of RNA you put in there?
Michael A. Resnick: That's right, and so in the design of the experiment we actually used a defined double-strand break at a precise position, so we know the sequences on both sides of the break and therefore we can design complementary sequence in RNA to add to that, and so we can ask that question if we do have complementary RNA to both sides of the break, will it repair the double-strand break?
Chris Smith: So, what did you actually do, could you talk us through the nuts and bolts of how you added the RNA and then how you think it goes about repairing this double-stranded break in order to link the two bits of broken DNA back together again?
Michael A. Resnick: Yes, we have a system-based on yeast, which has been ideal for settling many aspects of repairing-chromosome metabolism, in which at will you can induce a double-strand break at a precise site and then, we add the RNA which is complementary to both sides of the break, and what we perceive is that they actually come in, pair with sides of the break, and then there was synthesis upon the RNA across the break, so that you can actually get an end of the space between the two sides of the break using the information from the RNA.
Chris Smith: How do you know, for example, that it is the RNA that links the two broken bits of DNA first and, say, the RNA is not first of all converted into a DNA copy, a cDNA, and it is that that's doing the repairs?
Michael A. Resnick: Well, that is a really interesting question, which we have addressed in several ways. In some cases, we ask a simple question — if you have the RNA internal to two stretches of DNA, can you get a copy of the RNA, and the answer is yes, and that would not be the type of molecule that would be subject to this reverse transcriptase that you are referring to; and the other thing that we have used is various mutants that we know will get rid of the reverse transcriptase activity in the cell. So I think, both in the genetic sense and in the molecular sense, we have pretty much excluded the possibility of a reverse transcriptase.
Chris Smith: So, if I could now come to Francesca Storici, who is the first author on this paper. Francesca, why do you think we have never seen, despite hell of lot work, this effect before?
Francesca Storici: Well, the earlier studies did not investigate the possibility of mode-driven repair of chromosomal breaks. There were reports that RNA can repair double-strand breaks, but these were actually involving a reverse transcription of the RNA, so was an indirect role of RNA in the repair of the break, and it is we are sure that for the first time RNA can be directly used as a template in repair of the damage and there is no requirement for reverse transcription.
Chris Smith: Now, you did this in yeast— do you think this is a phenomenon specific just to yeast cells, or is yeast therefore viewed as a good model for fairly advanced cell types?
Francesca Storici: We started with yeast because recombination is known to be very efficient in yeast, but definitely this could happen in other organisms as well, like higher carriers.
Chris Smith: So, given that case, do you think this has potential to, say, correct, to find abnormalities of the genome where there are breaks or translocations that characterize certain diseases and certain cancers or tumour types, for instance, could you put right damage there?
Francesca Storici: Well, we stipulated one application could be to improve a new way for gene targeting, considering that RNA can be amplified in the cells so it could be generating in very high number of copies, and if RNA would have the capacity to transfer information directly to the chromosome, then this could be a new way of how to approach gene targeting and then this could lead to gene therapy in a model sense.
Ben Valsler: Michael Resnick and Francesca Storici, of the National Institute of Environmental Health Sciences, North Carolina, explaining how they have observed for the first time how small segments of RNA can be used as a template to repair damage to DNA.
Ben Valsler: Now, the process of photosynthesis is fairly well understood— that can be found in most high-school textbooks; however, researchers at the University of California at Berkeley have now observed how all light was captured so efficiently in the first place. Here is Gregory Engel. Nature 446, 782–786 (12 April 2007) .
Gregory S. Engel: We are trying to understand how photosynthesis, this exquisitely tuned process of evolution where light is absorbed and then eventually transforms to chemical energy, to understand how this process is designed, what the underlying principles are behind it, and how ultimately we may be able to use some of those principles in artificial systems.
Chris Smith: I thought people have worked out relatively well how photosynthesis works because we know it is effectively electrons tumbling from a high potential to a low potential and driving a chemical reaction. So, what is different here?
Gregory S. Engel: That is very, very true. What we are looking at actually is what happens before that. In photosynthetic systems light is absorbed through a vast array of antenna molecules and is transferred extremely efficiently to the reaction centre where you first see this charge-separation event, and subsequent to that you have the electron transfer that you mentioned, and that is well understood and well studied. We are looking at what happens prior to that event.
Chris Smith: So, how do these molecules actually soak up sunlight then?
Gregory S. Engel: So, the molecules absorb the sunlight and then, by being in close proximity to one another, they can transfer the energy between them. The energy flows extremely efficiently to the reaction centre: that is the critical stuff of photosynthesis, where we move from simply absorbing light to actually being able to use it for chemical energy.
Chris Smith: So, how did you study it Greg? Because it is one thing to see this happening in leaves, but it is another to try and get a handle on this under the controlled environment of a lab.
Gregory S. Engel: In the lab, we use extremely fast laser pulses to excite the sample and then a second series of laser pulses to try to understand what happened to the energy after we excited it. These processes of energy transfer are extraordinarily fast — on the order of hundreds of a tenth of a second. So if you take a billionth of a second, you divide it into10,000 pieces, it would be about that long. So, we have to follow these extremely fast processes with our lasers and we watched the energy actually move through these protein systems with a wave-like nature, and that is the real discovery of this paper.
Chris Smith: How are you actually seeing energy though, because as far as I know energy is not something that you can usually see, is it?
Gregory S. Engel: Oh, absolutely not. We excite the system and then we use a second set of laser pulses to get the system to then emit energy and that is called the 'photon echo' as we use it in this experiment. This echo that is generated by the system is what we use to probe where the excitation is in the system and how it is moving through the system.
Chris Smith: So, once you have got a handle on how things like chlorophyll actually work, how are you going to exploit this?
Gregory S. Engel: At this point in our studies, we are really trying to understand the basic design principles behind the system. Ultimately, we hope that researchers find this useful and can copy some of these design principles in artificial systems. The first industry that I think would be most interested in this sort of discovery would be photovoltaic solar cells, where they try to capture energy from the sun and efficiently create electric energy, and I think that they may take an interest in this work, but our work is fundamentally studying the photosynthetic system itself and trying to understand what nature has done so that we may be able to copy it.
Ben Valsler: Greg Engel explaining how they used a type of two-dimensional electronic spectroscopy to observe what happens less than a billionth of a second after light hits a photosynthetic system. Finally this week, Peter Kelemen of Columbia University is shaking up the world of seismology with a new model to explain how certain forms of earthquakes occur. Nature 446, 787–790 (12 April 2007) .
Peter B. Kelemen: We were interested in understanding intermediate-depth earthquakes, which prior to this time were difficult to understand because at a depth of 50 kilometres or more in the Earth the pressures on faults deep beneath the surfaces are so large that it is not clear how they can undergo brittle fractures.
Chris Smith: So, when you say intermediate depth, I suppose if it is a lay person, most people would be forgiven for thinking that an earthquake just originates all from the same place. So, are you saying that earthquakes of different types come from different depths in the ground?
Peter B. Kelemen: Very much so, and particularly in the so-called ring of fire around the Pacific, where oceanic crust and upper mantle are being pushed beneath the surrounding continental plates. There are earthquakes from a whole range of depths, from very near the surface to several hundred kilometres. It has been theoretically difficult to understand how the typical earthquake mechanism that involves rocks actually breaking could be operating at depths of 50 kilometres or more.
Chris Smith: So, what do you think is going on?
Peter B. Kelemen: Well, we found that there could be a viscous earthquake. We have done a mathematical model of this using standard laws for the deformation of mantle materials under stress, and we found that under the circumstances of very localized deformation in pre-existing fine-grained shear zones, these rocks will heat up all the way to their melting temperature and, as they do so, they become very, very weak. They act almost like cracks and release huge amount of stored stress or energy in a very short time. So, we predict displacements of tens of metres in a few seconds.
Chris Smith: What is actually responsible for making them heat up quickly enough in order to melt like that?
Peter B. Kelemen: Well, it is something like friction, it is called viscous shear heating, and it is simply related to the amount of force that is being released and the velocity at which it is being released. So, it is the stress times the strain rate.
Chris Smith: You mentioned that you have done this by modelling. Are there any signs or are there any particular things you could look for, which would prove this is definitely what is going on?
Peter B. Kelemen: Well, very much so, and so this project started out from our observational work on shear zones and exposed outcrops of the mantle that have been thrust up onmto land for various reasons, and we are now going back to look at those exposures and see if we can find evidence for transient periods of very, very high temperature. There is one prior study from a mantle outcrop in the Italian Alps where they found glass formed by very, very rapid melting of mantle materials under high stress and strain rates.
Chris Smith: So, what would be the implications if what you have described here turns out to be true, which we have no reason to think it wouldn't, what do you think the implications are for the field in this respect?
Peter B. Kelemen: Well, the most exciting thing for me is thinking about what happens beneath plate boundaries all over the Earth at mantle depth. For example, in California where we have the San Andreas Fault, it is locked in the upper crust near around San Francisco and we are all expecting a great earthquake, what happens in the deeper part of the plate beneath that fault, and I think what we found is that there is not necessarily steady deformation in the deeper parts of the tectonic plates, but there may be periodic shear heating instabilities, and those in turn might have feedback into the upper crust and control the period and nature of earthquake failure where people live, and so it is that implication that I would like to explore now.
Chris Smith: Presumably, with the idea of being able to predict where the hotspots might be in future in mind?
Peter B. Kelemen: Yes, indeed. Many people may have a toy where you have two pendulums that are attached to one another and, although each pendulum individually is very predictable, when you attach one to the bottom of the other the motion of the two becomes very, very difficult to predict. In this case, we can imagine a very large pendulum swinging with a very regular period and that might be the sort of thing that we have predicted for the mantle, and then the crustal response would be a very small pendulum attached to that. And the idea here is that if we understood the mantle process, we would be much better off in trying to predict what the little crustal pendulum is going to do in response.
Ben Valsler: Peter Kelemen explaining how earthquakes starting at depths of 50 kilometres in the Earth may occur because of the rapid melting and viscous creep in the rock, and hopefully that work will take us a step closer to being able to predict when earthquakes are about to happen.Well, that is all we have for this week, but I will be back next time with a new way of tracking how our continents took shape, and you can send your thoughts, opinions or any other feedback to mailto:podcast@nature.com. This week's Nature Podcast was produced and presented by me, Ben Valsler, Anna Lacey and Chris Smith. From us until next week, good-bye.
Advertisement
The Nature Podcast is sponsored by Bio-Rad, at the centre of scientific discovery for over 50 years, and on the web at http://www.discover.bio-rad.com.
End Advertisement
