LE Magazine December 2001
Page 3 of 4
L.E.: On a percentage basis, those 46 genes represent 0.9% of the known genes on the chip, around 1.1% of the active genes on the chip, and just 0.4% of the total genes on the chip. These age-related changes in liver appear to be considerably fewer than what was found in muscle. At the rate of 1.8% of the genes on the chip, one would expect a total of 198 total genes, or perhaps 109 known genes, to change in expression with age in the liver. However, you saw only 46, which is less than half of this predicted number. And yet you considered a change as small as a 1.7-fold change to be meaningful, whereas Weindruch and Prolla used a more restrictive detection threshold of over 2-fold (though they also reported some results with thresholds as low as 1.5-fold).
S.S.: I thought that we would see more transcription factors. I thought that we would see more fundamental regulators in gene expression changing. Aging, at least in the liver and in the muscle and in the brain, has a fairly subtle effect on gene expression.
L.E.: On the surface, everything seems to change with age, but at a deeper level, relatively few things change. As you said, aging is subtle.
S.S.: On the gene expression level it looks subtle. Now on the protein level it may be different because there's a lot of regulation that involves phosphorylation cascades and other kinds of modifications of proteins, and we know very little about that.
L.E.: On the other hand, calorie restriction may take care of a lot of that. In your study, for example, only 3 out of the normal 20 age-related increases in gene expression (15%) escaped correction by either long- or short-term calorie restriction! Two gene expression increases were reduced about 50%, and 15 increases were totally abolished by either long or short-term restriction. Of the 26 decreases in gene expression, 13 (50%) were blunted by long-term calorie restriction and 18 (69%) were blunted by short-term calorie restriction, leaving only 5 decreases (19%) unaffected by either long or short-term restriction. These results are amazing.
S.S.: Yes. I agree.
L.E.: Going further, it seems that a large fraction of the changes you saw with aging don't necessarily represent actual liver aging per se. Instead, they largely appear to represent changes that increase disease susceptibility. This means that the liver actually showed really very few true aging changes.
S.S.: We did see changes that look like the changes that you see in the development of age-related diseases. This has been found by other workers conducting micro-array gene expression studies in other tissues as well. I think our results are very consistent with theirs in showing that gene expression profiles in tissues begin to resemble profiles of tissues that have age related disease processes going in them. Our tissues looked healthy-we could slice them and look at them under the microscope and see no signs of liver fibrosis, for instance. But when we looked at gene expression in these tissues with age, we found changes that more and more resemble those that you see in diseased tissues. So, I think that's part of the development of age-related diseases-a drift towards gene expression that resembles the gene expression of diseased tissues. Calorie restriction reverses much of that, short and long-term.
L.E.: For example, a major fraction of the changes brought about by calorie restriction in your study had the effect of helping to prevent cancer from getting out of control.
S.S.: Most mice living under laboratory conditions die of cancer.
L.E.: So it seems that another possible practical spin-off of your work could be in the area of cancer prevention.
You point out in the paper several specific examples of how changes produced by calorie restriction might block cancer.
Might those insights allow for the development of anti-cancer drugs?
S.S.: I don't know if our agents will be anti-cancer drugs in the conventional sense of drugs for treating pre-existing cancers, but what I can see clearly is that we're going to be able to develop preventatives. Not just for cancer. But we can also attack other age-related diseases, because caloric restriction delays the onset and reduces the incidence and the severity of many age-related diseases.
L.E.: The beauty of CR is that it has such a broad effect against aging, hitting so many different pro-aging systems. But after you figure out what each specific CR gene effect is, you may find that you want to tailor a variety of drugs that are specifically targeted at those individual effects, rather than trying to replicate all CR effects at once.
S.S.: That's right. We don't really know yet if we'll find mimetics that will reproduce all of the effects of CR in all tissues. I think it's probably more likely that we're going to find treatments that reproduce some of the effects and are tissue-specific in doing that. I think we'll probably end up having to combine mimetics in order to achieve the full effects of caloric restriction.
L.E.: In your paper, you only reported changes in genes whose functions are known, except for maybe one case, which was the major urinary proteins, which were not really discussed.
S.S.: We didn't report what are called ESTs (expressed sequence tags).
L.E.: Did you see any EST changes?
S.S.: Oh, yes. But we just don't know what they mean. First of all the function of the ESTs is not known. Very often it is not known whether different ESTs even represent different genes. Some certainly represent the same gene.
Some ESTs are constant regions of antibodies.
L.E.: This points up the disadvantage of not having the complete Rosetta stone of biology-the complete genome-on a chip yet. In other words, there may be a lot of genes that are even more important than what's been observed, but they're just not on the chips yet.
S.S.: They're coming. It will only, I think, be a year or maybe two before we have a whole genome set on chips that will give you the gene expression levels of all of the genes and all of the splice variants. There may be 30,000 genes in the mouse and the human, but each one of those genes gives rise to, on average, 3.1 different messenger RNAs (mRNAs), the splice variants. So in many cases with the probes that are available now, we don't even know which splice variants, which of those three possibilities, we're assaying. But the commercial chips are getting better and better, and we're learning more and more about the genes as the Genome Project is continuing to yield information.
L.E.: Will the complete genome and variants chips be available for both humans and mice?
S.S.: For humans as well as mice. And then the field of proteomics (large scale protein surveying) is growing and expanding, and the technology is improving dramatically there. So I think the next ten years are going to give us a rich picture of what's going on during aging in different tissues and how that's affected by caloric restriction.
L.E.: Do you have in mind any hierarchies of the relative importance of different gene effects that you would use to distinguish a better calorie restriction mimetic intervention from a less good one?
S.S.: Yes. I think we have to look at each tissue and use what we know about the physiology of the tissue with aging and the effects of the change in gene expression that we see and compare that to the effects after long-term calorie restriction. In every tissue the effects are different, but there is a vast literature about the physiology of the tissues with aging and with caloric restriction. This can provide enormous help in interpreting the results that we get.
L.E.: Have you compiled all of that literature into some rules of thumb?
|We look at various times after the onset of caloric restriction: at the changes that happen early, at a middle time and late. Changes that appear early and disappear, and changes that appear early and are also present in the long-term state. I think that these kinds of studies are going to give us a better feeling for which may be primary changes and which may be secondary changes|
S.S.: Certainly we're working on that. It's a complex process and one that takes studies in a lot of tissues because every tissue is different so far. There are even differences between species that we've found, and that have been found by others. So, it's going to take some work to establish clear rules. But, yes, they're coming along. We have a good idea of what the effects are in liver, and now we're looking beyond that into other tissues as well.
L.E.: Your task would also be simpler if you could place the changes you see within some kind of hierarchy of cause and effect as well, so you would be able to say which age-related changes are primary and which ones are secondary. Do you have any feeling as to whether you can identify such a hierarchy of changes?
S.S.: This is a difficult question. One of the ways we're trying to get at the answer to that question is to look at various times after the onset of caloric restriction: at the changes that happen early, at a middle time and late. Changes that appear early and disappear, and changes that appear early and are also present in the long-term state. I think that these kinds of studies are going to give us a better feeling for which may be primary changes and which may be secondary changes. We can use bioinformatics to cluster the genes that change and start to ask which regulators are likely to be responsible for those changes. That will allow us then to do fundamental experiments that will ask more directly, do these specific genes affect health span and life span?
L.E.: That's very exciting. Another problem for interpretation of gene profiles might arise from the fact that very similar proteins can react in different ways to the same condition. For example, heat shock protein number 70 (HSP-70) doesn't seem to be affected by aging in your experiment, whereas HSP-86 and HSP-25 both go up with age.
S.S.: Proteins like these don't act alone, they act in groups. For HSP-70, where we didn't see an effect with age, it may be that there is a rather subtle effect. When we look at the livers of starved and fed animals, we find that HSP-70 responds strongly to feeding. After you eat, the level of messenger RNA for HSP-70 goes up about three-fold in a spike (peak), and then that decays after the meal. How many of those spikes you have and how high they are and how often they occur will affect the time-averaged level of HSP-70. If you fast an animal for a long time, the level of chaperone proteins tends to fall. So calorie restriction may have a subtle effect on some chaperones that we haven't picked up with these studies, where we looked at animals that had been fasted for 24 or 48 hours, because we missed those spikes that determined the averaged level of the protein over the long-term. So, these are very subtle and complex effects at work.
L.E.: It's a point we don't usually think about, but a lot of the things you might be measuring depend upon what time of day it is, how long it's been since you last ate, that sort of thing. The genome is very dynamic and if you look at different times, you'll get different results.
S.S.: That's so true.
L.E.: At least you're developing the tools to deal with the complexities.
S.S.: That's right.
L.E.: Another problem with using gene expression profiles to characterize aging and interventions could be that aging is a kind of mosaic. Different organs seem to age at different rates, and you die from whatever the weakest link is. So you could look wonderful based on many different markers and drop dead the next day because of one weak link that perhaps wasn't even measured.
S.S.: That's right. It's not clear to me whether we'll be able to get gene expression biomarkers from humans, for instance, from the cerebral cortex! But it is a promising technology.
L.E.: Have you found any contradictory effects of calorie restriction and aging, such as maybe having both an increase and a decrease in cell proliferation factors or both an increase and a decrease in the tendency for apoptosis or for inflammation? Or is everything consistent?
S.S.: We're not really looking at physiology. We're inferring changes in physiology from changes in gene expression. There is a lot more work to be done to see whether these changes in physiology actually occur as a result of changes in gene expression. Really, we're at the beginning of a very exciting era, and I think we need to use proteomics ultimately. We're going to need to know more about what's happening to the proteins in the cell.
Back to the Magazine Forum