Scientist Stories: Jan Hoeijmakers, The Aging Process

Thank you very much. And before I start, let me say that I have no disclosures. I have recently also started in two other laboratories, one in Utrecht on children's oncology and one in Cologne on aging of the skin. But most of my activities that I will report today refer to the Molecular Genetics Department at the Erasmus Medical Center in Rotterdam. So what I would like to do is to explain to you what we currently think is the cause of aging and transcription stress as one of the consequences, and how it relates to many aspects of aging. So, and I promise it will be amazing in a number of regards because there are a number of very unexpected developments in the research. So DNA is damaged all the time by factors outside, UV radiation, chemicals, also our own reactive oxygen species, alkylation products, numerous natural products damage DNA. Even water reacts with DNA. Every day, every cell, every nucleus loses 10,000 purines because of hydrolysis. So DNA damage is everywhere, every day. And DNA, of course, when damaged, will change metabolism. One will be that there are mistakes during replication or that chromosomal segregation leads to aberrant chromosomes, and that leads to mutations, and that will trigger cancer. So cancer, the main cause of cancer starts with damage in your DNA. But a very different consequence is the other one. So if you have a break or an interstrand crosslink, a bulky adduct, an intrastrand crosslink, they will pose a problem for transcription. So transcription will be blocked or replication will be arrested. If it takes too long for the cell to repair, it will delay cell cycle progression. It will maybe, if the damage is irreparable, arrest permanently. It may functionally decline, and it may trigger cell death. And those, we think, are the main causes of aging. And I'll explain to you later why. Constantly, there are repair systems that try to cope with this problem and to counteract and to remove all of these damages before either cancer or cell death and functional decline occur, and cellular senescence, and you age. So repair systems are constantly protecting us from aging. And every day, every cell may experience up to 100,000 lesions per day. So it's a massive problem. And before I continue a little bit, let me also reiterate Some damages are not recognized by any repair system, so they will accumulate in time. Some lesions are irreparable. You cannot repair them anymore. They are too severe. Sometimes repair is too late, and sometimes repair is not error-free, so you can make mistakes. So despite efficient repair, still DNA lesions accumulate in time, and that leads to cancer and aging as we experience it in our lifetime. So repair mutants, mutants in DNA repair, will accelerate all of these effects because they cannot repair the lesions that normal people do. So since there are many different types of damages, there is not one repair solution, but there are multiple pathways, each dealing with part of the problem. So for instance, nucleotide excision repair and transcription-coupled repair, they deal with a problem that damage can distort helical structure, can disrupt base pairing, and can block transcription. So that's a very wide spectrum repair pathway. So base excision repair deals with subtle lesions. Homologous recombination and end joining are two of the multiple solutions for double-strand breaks, which are very difficult for cells to heal. Cross-links are very difficult to remove. So that's a dedicated pathway, and there are multiple pathways there as well, in the strand cross-link repair. So I'll highlight nucleotide and transcription coupled repair as examples, but remember, this is an example for the whole basic problem of DNA damage. So this is a summary of the repair reaction. I'll not go through it. I only can point out there is a global genome NER system where damages everywhere in the genome that disrupt base pairing are detected and removed. So that is why we call it global genome nucleotide excision repair. And there is another pathway which is called transcription-coupled repair. So when transcription is blocked by a damage, this pathway will try to repair the damage. And the damage is excised and new DNA is synthesized, and that is how the system works. So many proteins involved. All of the proteins indicated in pink, there are human children that are born with a defect in the genes for those repair proteins. And these diseases are abbreviated over here, XPCS and TTD. I'll explain them in a minute. And there are more of them, but I'll only highlight those three. And next to the patients, we have also generated in the laboratory many mutants, all of them indicated in M. We have made mouse models for these genes. And in many cases, we made even the same point mutation of a patient in the mouse to have an accurate mouse model for the human disease. So let me show you the diseases first. So all of these diseases are characterized by pronounced sensitivity to UV light. UV causes damage in your skin. DNA repair cannot repair it. As a consequence, all of these syndromes are characterized by pronounced UV sensitivity. The first disease is XP, very rare, autosomal recessive, 1 in a million, 7 genes involved, XPA to G. The hallmark of this disease is skin cancer. You can see pigmentation abnormalities in sun-exposed parts over here and over here, but what is the most important problem is the beginning of skin cancer. And when children are not exquisitely protected from sunlight, they will develop skin tumors in sun-exposed parts very early on. This is a mild case. There are children 10 times as worse. The whole face full of tumors. So skin cancer is the prominent feature. There is also elevated other cancers, and this is due to a defect in the global genome, so where you repair damage everywhere in the DNA. Now, another disease which is called Cockayne syndrome, very rare as well, 1 in a million, also autosomal recessive. But the phenotype is almost the opposite of the previous one. In this case, no cancer. These children have no cancer. But they are worse off because they have very severe neurological developmental abnormalities and a life expectancy on average of only 12 years. We have no medication for them. This is Baptiste, a child with Cockayne syndrome, 3 years old. 7 years old, sorry, 7 years old, you see very small, needs glasses, needs hearing aids, cannot speak, cannot walk, has a cacti appearance. And this child, Baptiste, became only 10 years. And for comparison, his sister, 1 year older. So dramatic growth retardation, very severe phenotype, No solution. And this is due to transcription-coupled repair. So the other branch, where when the transcription is blocked by damage, this system would be able to help you. Now, another disease, trichothiodystrophy, extremely rare as well, 1 in a million, TTD. The same, almost the same, not identical, the same, very many of the same features of Cockayne, but what makes this disease very standing out is that these children have typical brittle hair, brittle nails, scaly skin. So, and this disease is always caused by a defect where both of these pathways are partly affected and also a transcription function is affected because the complex involved in these mutants is called TF2H. It's opening DNA for transcription. Repair, but it's also opening every promoter for transcription. So when transcription is affected, then you get brittle hair, brittle nails, scaly skin. I can explain it. We have recently found 4 other genes for this disease, and they all link to gene expression. So this is typical. If you have a problem in gene expression in whatever stage it is, then you get brittle hair, brittle nails, and scaly skin. In, but that's a trivial problem and I will not explain it further. So we have one pathway, nucleotide excision repair, but so many different diseases. And even you can have a mutation in the XPD gene. In some patients with an XPD mutation, it's XP. In others, it is Cockayne and XP. In yet others, it's TTD. And again, others have TTD and XP. And again, another disease which is called COFS, cerebrooculofascioskeletal syndrome, very severe. Patients don't get older than 1 year and a half, all caused by mutations in the XPD gene. But all of them are point mutations. And one way to try and find out what is going on is to try to make mouse models, what we have done. I'll show you the TTD mouse. As an example, we have made many mice, so just an example before I show you that we first explain XPD is a helicase. This is essential. You need it for transcription of all your genes. So only point mutations are viable, but every point mutation is linked with one clinical outcome. So this point mutation, R683W, all balls, all patients, but all XP. So this is another one, R722W, all TTD. So all point mutations are linked with one clinical outcome. So what we have done is we have mimicked this point mutation in the mouse XPD gene, made XP, XPCS, and TTD mice. And this is the TTD mouse, this is the patient, and this is the mouse. So the mouse has also the typical brittle hair features of the children. And the mouse is a very accurate model for the human disease, there are striking parallels. I can say you that we found clinical phenotypes in the mouse first, asked clinicians to look in the patients, and upon inspection, they turned out to be there. So in this case, the mouse and the human patient are extremely similar. But an eye-opening observation was that these mice became gray. We had never expected TTD to be linked with, with aging. But now we started to look very closely in these mice, and these mice indeed get kyphosis, and they show osteoporosis. That has been very thoroughly investigated by one of the persons in the audience, Karen Diederich, in our laboratory. So these mice show many features of aging, and they reveal, they revealed that TTD actually is a progeriatric syndrome. So if you look at the mice, their lifespan is reduced, although not as severe as patients. Patients, all of these patients with that mutation died before the age of 5. In the mice, they can become 1.5, almost 2. They have cachexia, sarcopenia, early infertility, cutaneous symptoms, skeletal abnormalities, osteoporosis for instance, progressive neurological abnormalities, lymphoid depletion, liver, and kidney show all features of aging. But it is segmental. Not all features, not all organs do show it. Actually, cataract in the islands is delayed. Less cataract, but overall more aging. And if you look in the patients, some clinicians had noted, for instance, osteoporosis. But of course, these children are so rare Nobody could get a complete picture, but the mouse provided it. And now in retrospect, there is not the slightest doubt that these children also show premature aging. So mice and humans are not the same, but the TTD mouse is a very accurate model for the human disease. And TTD is not primarily a developmental, but we think it is early aging that compromises development. That's why It looks like a developmental disease, but it is simply early aging that interferes with development. So what would happen if you now take this mutant, which has a partial defect in global and transcription coupled repair, and cross it with an XPA mouse, which is a very severe cancer-prone mouse model? So that's what we did. So the XPA mouse by itself, also generated in our laboratory as a conditional mutant, shows cancer. Has a global repair defect, completely deficient in some aspects of transcription-coupled repair. This mouse, the TTD mouse, ages in 1 year and a half, shows hair loss already, but has no cancer. Cancer is in fact reduced in this mouse. The double mutant, DTD/XPA, lives only 3 to 5 weeks. You see it gets kyphosis in that short period of time. And it shows aging. Now, the aging that took 1 year and a half in the single mutant is so much accelerated that it now takes only 3 to 5 weeks. And it took a long time for the aging community to really realize that this is really true. So the severity of the NER defect correlates with the acceleration of a segmental aging phenotype. There is a dose-response relationship, and dose-response relationships point to causality. There must be the repair, must be the reason. And I show you another mouse. We have made many mice, so we have made Cockayne A, Cockayne B, XPA, and we have also XPC mice generated elsewhere. If you take a CSB mouse, cross it with an XPA mouse, you have no hair problems because the hair is typical for DTD, but now the mouse shows also kyphosis. And has a lifespan of only 3 to 6 weeks. So this is a very reproducible and robust result. So this was a double mutant. Now let's go to 4 repair defects. If you now make a mouse which has 4 repair defects, well, we made also another mouse which is called the ERCC1 mouse. And ERCC1 is part of a complex with XPF, and that is involved in 4, at least 4 pathways of repair. So if you take that one gene, you have a problem in four pathways. So now that mouse is really the champion of accelerated aging in mammals because it shows aging both in the post-mitotic compartment that is linked with transcription coupled repair, and in the proliferative compartment of the body. The bone marrow is aging in this mouse also. So now this mouse shows vision loss, hearing loss, neurodegeneration, very extensive, I'll show you later. Hypertension, vascular stiffness, cardiomyopathy, kyphosis, osteoporosis, liver and kidney aging, muscles, the fat tissue, the skin. They get cachexia, sarcopenia, urinary incontinence. The stem cells get exhausted both in the bone marrow and the satellite cells in the muscle. So, and overall frailty. Virtual normal embryonal development, but then progressive accelerated aging. And this mouse, because it's a hypomorph, it's not a full knockout, lives 4 to 6 months. So that's a very optimal time frame for studying aging. So widespread aging, widespread aging is the same as multimorbidity that we all experience when we all age, because we never age in one organ or tissue at the same time. So this actually shows unequivocally what an endogenous— this is spontaneous DNA damage— can do when you cannot properly repair it. So all of these phenotypes, all of them, they can be derived and attributed to DNA damage that is not properly repaired. So in fact, this mouse shows 3 and maybe 4 human diseases at the same time. It shows the transcription coupled repair defect from Cockayne. It shows global repair defect from xeroderma pigmentosum. It shows a cross-linked repair defect from Fanconi anemia. And it shows also a single-strand annealing defect for double-strand breaks that persist. So this mouse shows these three diseases together and the transcription coupled repair defect, which is aggravated when you have On top of that global repair defect that shows cell death and functional decline, mainly in the post-mitotic tissues. So the neurons, the heart, the liver, and the kidney, they are the ones that suffer from aging, but not so much the proliferative compartment of the body. Because if you proliferate, if your cells divide, you dilute damage. All new DNA is not damaged, so you dilute constantly damage, whereas a cell that does not divide will only accumulate damage. And that leads to transcription problems I'll show you later. The cross-link repair defect and single-strand annealing defects of the YACC1 mouse, they cause cell death and chromosomal instability and thereby aging and cancer in the proliferative compartment. So the bone marrow, like in Fanconi anemia where you have pancytopenia, and leukemia, both together, as the consequence of cross-link repair defect. So that explains why aging in ESCC1 mice is so all-round. So let me focus on the brain, because the brain has— there, in this mouse, there is tremendous neurodegeneration. I'll summarize before I show you the data, and in the interest of time. So there is brain shrinkage. There is ventricular dilatation. There is loss of neuronal plasticity. —when you measure with long-term potentiation physiology, neurons, memory, motor performance, hearing, vision, they all get gradually lost. And cells die, neurons, or get p53 active, positive in time in 4 weeks and more so in 16. So I cannot show you all of the data. I will focus on the gene expression profiles that we generated from these mice. There was something when we studied all the activity of all the genes in the brain in the cortex hippocampus, there was something very unexpected that we found. There were very strong parallels with Alzheimer's disease. Alzheimer's. All of these genes that are critical for beta amyloid plaque formation, they are deregulated in the very same way in the profile of these mice. So, that— indicated maybe ESCC1 mice show features of Alzheimer's. I didn't tell you, ESCC1 mice we made before we found later the patients. So there are 3 patients that I know in the world with an ESCC1 mutation, so it's 1 in a billion. Yet the phenotype of the brain points to Alzheimer's, which occurs in 50 million people in the world. So I don't have the time to summarize. I only summarize the data. I don't have the time to show them. There are not only key alterations in the same genes as Alzheimer's, but if you take the overall profile, the overall profile shows a much higher overlap with Alzheimer's than the best Alzheimer's models that have been used in vain for Alzheimer's research. And if you compare two human Alzheimer's profiles and compare their overlap, the ERCC1 overlap is in the same range. So, and if you follow the progression of that profile in time, it follows the progression of Braak stage 1 to 7. So it's the same parallel development. So that, in fact, makes a lot of sense, because the most important risk factor for Alzheimer's is not plaques and tangles, but is your age. How old are you? So aging is the most important risk factor. And if these mice reflect aging in the brain pronouncedly, then they must also reflect Alzheimer's. And I'll show you later, we think now we can explain all the dementias and all the proteinopathies on the basis of this mouse. I'll show you later. So ESCC1 mice may be a valuable additional model for Alzheimer's disease, addressing a tremendous unmet medical need. So we can manipulate aging, so we can only target part of the brains with the repair defect. I'll show you now a mouse in which only the hippocampus and the cortex has a repair defect. The rest of the body completely normal. And now you see that if you stain for GFAP, that this is the hippocampus and this is the cortex compared to the wild type, and there is dramatic shrinkage of the tissue. And now I'll show you a comparison between two mutants. So this is the ESCC1 where the hippocampus is shrinking half a year. In one year, there is hardly any hippocampus left anymore. If you now take an XPG mouse, that's another mouse mutant that we also generated, one, at least one repair pathway less. So that one has 3 instead of 4 defects. And as a consequence, the degeneration takes longer. So in 1 year, XPG is almost equivalent to half of a year in ESCC1. So again, those proteins response relationships that points to causality. If you take the brain of an Alzheimer's, then this is the normal hippocampus. This is what's left from the hippocampus in an Alzheimer's patient. So also there, you see that there is tremendous loss of neurons in those tissues. Again, now a mouse in which only the heart is repaired defective, and only the heart shows aging. So this is the echocardiogram of a normal mouse. A small heart that vigorously pumps. This is the dilated heart of an ESCC1 cardiomyocyte-specific knockout. You see the heart volume is much larger. The pumping function is much weaker. And I can show you on two movies. So this is the normal beat of a normal mouse, very vigorous beat. And this is the beat of a mouse in which only the ESCC1 heart is accelerated in aging. So we can target any organ. For accelerated aging. OK, now what we also found, and that is very important. So we found that all the mice that show accelerated aging, when you study their expression profile in the liver, in the kidney, in the lungs, in the brain even, you see that these mice suppress the growth hormone IGF axis. They don't want to grow. They don't want to spend their energy in growing. They invest their energy in the maintenance, in the protection system of the body, in stress resistance. So we think that these mice, they invest, they want to protect themselves from DNA damage, and in that way to try to extend their lifespan. And that explains not only why Cockayne syndrome children stay always small, because they invest also their energy into maintenance, they don't want to invest their energy in growth, But we think this is also very important because dietary restriction, when you give too little food to a healthy organism—yeast, worms, flies, and primates—then the energy is not going into growing, but the energy is going into trying to survive the period of starvation, to make you strong. Because if you don't survive, you then don't get offspring. So then priority is to to favor your maintenance pathways, and that delays aging. Dietary restriction, uniform, delays aging in all organisms tested in laboratory conditions. So we wondered, so these mice, they suppress growth, invest in maintenance, what would happen when we actually apply dietary restriction? Normally we give these mice food as much as they want. So, but what would happen if we now would these little mice, put them on a diet. So that is done here. So this is ESCC1 females. The first female dies at libitum conditions at about 13 weeks, and the last one 27. And now the other half of the females on diet. Week 7, 10% less food. Week 8, 20%. Week 9, 30%. And then continue, 30% less food. And then see what happens. And this happened. Lifespan extended threefold. Median as well as maximum lifespan. So this is really an extreme lifespan extension. But of course, what happens to health? Are they healthier or is there more misery? Well, there is no, no, not the slightest doubt. All organs and tissues have very much delayed aging. In the liver, the kidney, the immune system, the retina, the neuronal system, vascular system, they all are tremendously benefiting. But most neurodegeneration. Neurodegeneration is the most severe problem of cocaine children. They die from neurodegeneration. If you now look at these movies, these are two mice, little mates, born same day. The mouse over here, this mouse, Food all the time. This one from week 7 on diet. Now they are 16 weeks old. And you'll see, if I start the movie, this mouse has a problem in walking because it's neurologically already very deteriorating. It cannot control its muscles. It has tremors, very uncontrolled spastic movements you will see in a minute, and also the mouse cannot keep proper balance. So they, like the patients. So, but if you take the same mouse and from week 7 on diet, this mouse is running around because it expects, aha, I get food, I'm very happy. But normally the mouse is very quiet, but neurologically it's in a much better state compared to the other one. So now, if the next slide shows a rotorod experiment. You put mice on a rotating rod and accelerate rotation, some point they drop off. Measure to how the mice can control their muscles. 4 mice. ESCC1 on dietary restriction. ESCC1 ad libitum, the mouse that didn't, couldn't walk in the previous movie. This is the wild type on dietary restriction. This is the wild type, the fat mouse, ad libitum. I'll start the movie. Already in the first round, this drops because it couldn't walk. You saw it on the previous movie. The thick mouse, after 3 minutes, has fallen down. This is the wild type, and this is the ESCC1. So the only difference is the food. And in this movie, the wild type drops first, but normally not. So if you measure the onset of tremors, very interesting. So if you take the ad libitum animals, the first mouse that develops tremors by week 7.5, 8. By week 10, all mice have tremors. Now we start At week 6, with dietary restriction. One week before you have tremors. The first tremors, sorry, the first tremors, ah, the first tremors that appear 44 weeks, and not even all mice died with tremors. Some of them died without. If you now give them for 6 weeks dietary restriction, and then stop, within one week after stopping the dietary restriction, they get tremors. So it's a very prominent and a very powerful protection from neurodegeneration. And if you start counting neurons in the neocortex, 30% more neurons in the brains of the dietary restriction animals, as you see. So with ad libitum food, 3 million, and with dietary restriction, about 5 million neurons in the neocortex. So if you look at p53, it's exactly the other way around. In ad libitum, there are many neurons that have p53-positive responses, and if you apply dietary restriction, less. And that suggests p53 responds to DNA damage. So maybe dietary restriction leads to less damage in your genome. So we wondered what happens to the DNA damage. We used gamma-H2AX as a damage marker. As a marker for DNA damage in the Purkinje cells of the EACC1 mouse upon dietary restriction. And it turns out indeed that the number of neurons that are gamma H2H positive in the dietary restriction is reduced by 50%. So dietary restriction somehow reduces DNA damage and preserves your genome. And since these mice are very sensitive to damage, they overrespond to dietary restriction. Dietary restriction normally will not triple your lifespan. It gives you maybe 35% lifespan extension, but not triple. These mice show tripling of lifespan, and we think it's because they are sensitive to DNA damage, and that's why when you have less damage, they show also a very strong response to that. So the ERCC1 mice have lost one-third of neurons because of limited food. And dietary restriction has preserved 50% more neurons, which explains why they are neurologically much better. And the neuronal system, we think, benefits disproportionately from dietary restriction. And if you think in evolution, back when there would be starvation, the first thing you need is your brains. So when dietary restriction conditions want to allow you to survive starvation, then it should give priorities to your brains. So that may be why with dietary restriction we see neuroprotection. And dietary restriction works better than any medication would be able to accomplish. Instead of 12 years, if the same would happen to patients that we find in mice, and there's no reason a priori not to believe so,, then these patients might become 36 and would have much better health. What's the mechanism? So one thing which was striking is that transcription-coupled repair, that is always linked with pronounced neurodegeneration, but also liver and kidney, the post-mitotic tissues. So that suggests maybe in these tissues there is something going on with transcription. That DNA damage interferes with transcription and that that may be a cause of aging. So we looked what happens to transcription. We developed a new method for transcription in vivo, in living animals. I give you only the outcome without showing you the data. It's difficult to see on the slide anyway. If you measure transcription in the liver of ERCC1 mice at 4 weeks, the transcription is already decreased. Decreased in this way, about 40%. In 16 weeks, transcription is below 50%. In 20 weeks, the mice be— that's already the age when a number of mice start to die. Transcription, overall transcription of the whole nucleus of all hepatocytes has decreased to about 25%. So a dramatic reduction in the transcriptional output of the whole genome. So we wondered— this we found first in our repair mice— so we wondered what happens with natural aging. So we took now an old mouse, a normal mouse, 102 weeks old, and measured transcription in the liver of these mice. And lo and behold, also in normal aging, transcription declines very, very significantly. So the transcription machinery also with natural aging is declining in post-mitotic tissues. So you can make a prediction. If DNA damage causes this transcriptional decline, then when a gene is twice as long, it has twice as much chance to have a damage. So a longer gene has a more high likelihood to be suppressed compared to a short one. So we looked in the expression profiles of ERCC1 mice and wild types, what is happening if you compare gene length in transcriptional output. And then it turns out, and I can show you, it takes some time for me to explain and I have not so much time. So what I can show you is that if you look at the expression profile of a wild-type liver of a mouse on diet— a wild-type liver on dietary restriction. Compare that to the profile of a wild-type ad libitum. Look which genes are up and which genes are down, and then correlate the up genes and the down genes in terms of their length. So this is a length scale. This is a log scale. 1 kilobase, 10 kilobase, 100. These are the very big genes. And the black line indicates the size distribution of the up genes in wild type upon dietary restriction with wild type ad libitum. The most genes are between 10 and 100 kilobase. That's the average length of the most genes. And the down genes have a similar distribution over here. Now look at the profile of ERCC1, the animals at ad libitum. Compare those. So that's the blue line. So the down genes, most down genes, the top of the down genes is between 100 and 1,000 kilobase. So the very long genes are selectively suppressed. And if you look at the up genes in the ESCC1 profile, that profile is shifted to the shorter lengths. So there is a very discordant between now gene length determines how genes are silenced. So we wondered, is this something special for our ESCC1 mice? So we took rat liver and rat liver's normal aged they show a similar distribution, so the down genes are longer, and this is hundreds of genes, so it's selected. It's highly significant compared to the up genes. And then we looked in the human hippocampus. So old people compared to young human hippocampus, look which genes are up in young and down in young compared to the old. And then it turns out that the down genes in old also are much longer on average. Than the up genes in old. So also in a natural hippocampus, we see this transcriptional bias. So, and it was also observed in Alzheimer's, which I showed you before. So we call this transcription stress. And we can do nascent RNA synthesis of the whole genome, map all the nascent transcripts. This one is to the IGF-1 receptor gene. So if you look for all the nascent RNA synthesis in the liver of IGF-1 receptor gene, it's a long gene. You see that nascent RNA synthesis of a wild-type liver 4 weeks old, this is what is in the beginning of the gene. It gradually a little bit declines, and this is the end of the gene. Now, if you look at a 16-week-old liver, you see there is more in the beginning, compared to 4, and in 100 weeks it's even much more because there are more transcripts that start, but there are fewer that succeed in reaching the end. So ESCC1 at 4 weeks, young ESCC1, show already a profile that is much worse than 16 weeks wild type. So we can also measure all genes. So now, And this is an experiment where we took all nascent RNA from all genes, and in this case, 4,100 genes which had sufficient reads for the whole gene length. And then we divided all these genes in 3 equal parts. So the first 1/3, the second 1/3, and the third 1/3. And these genes are arranged by length. So these are the very shortest genes, And this one, the 4,100, these are the very long genes. And now look, where are the nascent RNA synthesis coming from? And then you see bluer is declining. So if you take the first one third, that is the long genes, these are the short genes, these are the long genes. There is some increase over here. But if you then go through the second third, And if you go to the third third, you see there is a— the longer the gene, the more suppressed, the less nascent RNA. So the longer the gene, the less nascent RNA in the most distal part of every gene. And longer genes are longer away, so that's why it is more pronounced in longer genes. Now we looked, where are the RNA polymerases in these genes? And it then turns out The longer, sorry, the longer the gene, the more RNA polymerases are in the last part. So there are more RNA polymerases where there is less nascent RNA. So that must mean more polymerases further down in the gene which are largely nonproductive. They must be stuck. They must be stuck by DNA damage. So we can Yes. May I ask you to— Oh, wow. What I'll skip, what I summarize, and then I make another summary of a part that I was supposed to present also, but I apologize. I have always very optimistic. So age-dependent decline of transcriptional output, we first discovered in repair mice, and later we found it in natural aging. So that means that our mice are very valuable for natural aging. And it is mainly in organs with low or no cell renewal, which accumulates DNA damage, and where damage is not diluted by replication. Dietary restriction strongly reduces transcriptional stress in these models. I didn't show you, but there was another curve where you could see that the bias to long genes that are being suppressed was very much reduced by dietary restriction, consistent with dietary restriction reducing DNA damage. DNA damage associated with declining transcriptional output is a new aging-related process, which must have major functional consequences because transcription is vital. Without transcription, you do not function well. And it strengthens the parallel between natural and accelerated aging and highlights the values of our repair-deficient mice. The last part of my presentation was about protein aggregation. So I have no time to show you the data and to explain it. What I can tell you as a take-home message is when you have transcriptional stress affecting long genes more than short genes and stochastic random damage, then the prediction is that in post-mitotic tissues this will lead to protein aggregation. And that explains the aging contribution to all proteinopathies. So I'll skip this. So the general concept, so in a wild-type mouse and a wild-type animal, wild-type organism, there is normal repair. And so DNA damage that accumulates takes a long time, and transcriptional decline takes also a long time, and that's an organ dysfunction as well, and that is why you age normally. In the ERCC1 mutant, under normal ad libitum conditions, the repair capacity is strongly reduced. So there is a repair defect, 4 pathways affected. As a consequence, there is much faster accumulation of DNA damage, most of which will block transcription quickly, and that leads to the dramatic accelerated aging in 3 to 5 weeks in a double mutant. Now, upon dietary restriction, there is less damage. Of course, dietary restriction cannot compensate for the repair defect, but it can reduce your DNA damage. And as a consequence, it takes longer for transcriptional output to decline, and as a consequence, aging, the rapid aging, is declined. Okay, this is our laboratory, the people, the mice, very important, and the names of people that have contributed, all of them very important, and I have no time to go through all of the names. We have also collaborations with many people outside, within and outside of our country and our institution, and also these are people that have left the laboratory and continued funding organizations. I'm grateful for support from many different sources. And sorry for taking so much time. Maybe I was not clear in all regards. So if you have a question, maybe there is still time for a little bit or not. We still have time for one or two short questions, so please. It's not a question.