Scientist Stories: Elsa Logarinho, Chromosomal Instability and Aging

Good afternoon. I would like to start by acknowledging the organisers for the invitation and congratulations for setting up such an impressive programme. Since 2015, my group, Ageing and Aneuploidy, at i3S in Porto has been studying chromosomal instability in the context of ageing. Today's seminar is entitled Inhibition of Age-Associated Chromosomal Instability: An Immersion Strategy to Delay Cell Senescence and Aging. Aging represents a major risk factor for many chronic diseases, including cardiovascular diseases, diabetes, cancer, arthritis, and frailty. Aging occurs at the molecular, cellular, and organismal level. And a host of age-associated features has been identified, with a subset being potential drivers of the aging process, the so-called aging hallmarks. At the molecular level, the aging hallmarks comprise DNA damage, telomere attrition, epigenetic alterations, and protein aggregation. At the cellular level, aging features include accumulation of a pure mitochondrion lysosomes, deregulated nutrient sensing, and cellular senescence. Senescent cells are cells that stop dividing in response to stress or damage from exogenous or endogenous sources.. And these cells, they exhibit distinct biomarkers and a secretory phenotype causing local chronic inflammation with an impact in healthy surrounding cells, as you can see here. So, the senescent cells are represented in pink and they accumulate during aging. Having a paracrine signaling in the neighbor cells. At the organismal level, the chronic low-grade inflammation associated also with cell senescence leads to stem cell exhaustion and to altered intercellular communication. In the lab, we have been focused on chromatin aging, and how this leads to cellular senescence. Our rationale is that increased chromosomal instability and consequently DNA damage leads to a progressive loss of chromatin architecture and epigenetic remodeling, which in turn compromise the the cell type-specific gene expression that determines cell identity and function. Although an intervention is still missing to correct chromosomal instability and DNA damage or improve the DNA repair mechanisms, there are other mainstream strategies that are available that either improve the epigenome or targets the senescent cells. So aging, once considered irreversible, is now seen as remarkably malleable. Diet regimens such as caloric restriction and fasting, even more if coupled to circadian cycles, they are able to preserve the epigenome. More dramatically, Cellular reprogramming by the Yamanaka factors acts by resecting the epigenome and has been validated in animal models. But such a pluripotency triggering prevents its clinical translation thus far. So, an intervention later on based on the ablation or clearance of dysfunctional senescent cells has been increasingly popular Senescent cells, they are responsible for the age-associated pathologies due to their secretory phenotype known as the SASP. Many drugs targeting senescence or the SASP, they are under clinical trials. One thing that caught our attention as we reviewed in this paper was the fact that the reprogramming and the senolytic therapies, they all end up by improving the cell proliferative fitness. So we ask whether by somehow modulating the age-associated regulatory features to proliferation could actually work to either reprogram epigenetics or senescence. To this end, we start by analyzing the cell cycle dynamics of aged cells. We use human dermal fibroblasts retrieved from Caucasian males with age ranging from neonatal to octogenarian, and these are movies of early cell culture passages recorded under phase contrast microscopy for several days. When comparing a neonatal with a middle-aged sample, you can easily appreciate that the striking difference is the loss of proliferative capacity during aging. Even though it is widely accepted that this loss of proliferative capacity acts to prevent damaged cells from proliferating, the exact regulatory mechanisms have remained elusive. Also, we realized that analysis of the mitotic process in aging models and diseases was scarce, and the comprehensive analysis of cell division efficiency in naturally aged cells was largely missing. To address these questions, we establish a human cellular model of natural aging consisting of human dermal fibroblasts retrieved from skin biopsies of healthy Caucasian males with ages ranging from neonatal to octogenarian. Since inter-individual differences exist in the rate at which a person ages, the so-called biological age, we included 2 donors of each age in a total of 5 chronological ages to increase the robustness of any correlation found with aging. In addition, we use dermal fibroblasts from an 80-year-old child with the Hutchinson-Gilford progeria syndrome as a model of premature cellular aging. Considering that during a normal postnatal lifespan, cells in vivo will hardly reach the exhaustion number of replications observed in culture, and to limit any artifacts and clonal selection of in vitro culturing, only early cell culture passages way below replicative lifespan exhaustion were used. To validate the suitability of our cellular models as models of normative and premature aging, respectively, we measure the percentage of cells exhibiting senescent markers. Chronic accumulation of macromolecular damage during natural aging induces a cellular stress response known as senescence, and accumulation of senescent cells has been widely reported for chronological aging and age-related disorders. Senescent cells exhibit distinct biomarkers such as increased beta-galactosidase activity due to the high accumulation of lysosomal mass, heterochromatin foci, DNA damage markers, and also an increased expression of cell cycle inhibitors. In addition, these cells, they exhibit a secretory phenotype causing local inflammation in surrounding cells. Accordingly, we found that there are significantly higher levels of senescence-associated biomarkers measuring under strict quantitative parameters by microscopy analysis and in the proliferating fibroblast cultures from older individuals and AGPS patients, thus validating the suitability as models of normative and premature aging. So you see here that not only there is an increased percentage of cells staining positive for DSA beta-gal assay, as well as cells become increasingly positive for the double immunostaining for a DNA damage marker, 53BP1, and the cell cycle inhibitor P21. And also the cell size, which is also a typical feature of the senescent cells, tends to increase in the elderly samples and age-epis. To investigate the cell cycle behavior of fibroblasts from the different age individuals, we perform long-term phase contrast live cell imaging. Interestingly, we found that not only the number of dividing cells decreases with advancing age, but also the interval between nuclear envelope breakdown and anaphase onset, which we consider the measure for mitotic duration. And this also increased steadily with advancing age, and the quantification you can see here. So, there is a gradual increase of mitotic duration from nuclear envelope breakdown to anaphase onset with advancing age. To investigate the chromosome or the spindle defects that are contributing to the increased mitotic duration in the older mitotic cells, we performed high-resolution spinning disc confocal microscopy in cells that were transduced with lentivirus for the expression of a bicistron, H2B-GFP to label the DNA and the chromosomes, and mCherry tubulin to follow the microtubules and the spindle. We found an increase in several mitotic phenotypes, significantly at the middle age and old age samples. And these include a delay in chromosome alignment at the metaphase plate, as well as chromosome missegregation, as you can see here, with an event of an anaphase lagging chromosome, a chromosome that is left behind in the cell division, which later on will generate a micronucleus. We also found that because of these mitotic defects, there is an increase in aneuploidy from middle age on. And we measure this by performing fluorescence in situ hybridization for using centromeric probes against 3 chromosome pairs. And you see that the aneuploidy index of these 3 chromosomes increases from middle age on. But also, it is even more striking if you look at aneuploidy in post-mitotic cells, cells that we arrest in telophase. Using the cytokinesin D inhibitor for cytokinesis. So, aneuploidy increases during normative aging due to an age-associated loss of mitotic fidelity. To identify the molecular mechanisms behind this complex aging-associated mitotic phenotype, we perform RNA sequencing, gene expression profiling of cells captured in mitosis, accordingly to the experimental layout shown here. So basically fibroblast cultures from neonatal and octogenarian donors, they were treated with a mitotic block drug, the STLC, to enrich for mitotic index,. And the mitotic cells were collected by mechanical detachments. Since cell synchronization, it was limited in elderly versus the neonatal cultures, we use of course a higher number of cell culture flasks with octogenarian fibroblasts to collect equivalent number of cell— of mitotic cells in both elderly and neonatal cultures. We confirmed that this procedure yielded an enrichment of around 95% of mitotic cells, so this is a very pure mitotic population, which was the population that we used for the RNA sequencing analysis. This was important because using this methodology, we uncouple any changes that we found that we could find in mitotic gene expression from the sample differences in mitotic index and overall proliferation capacity. The RNA-seq revealed that the abundance of 337 300 gene transcripts was significantly altered in octogenarian mitotic fibroblasts compared to neonatal mitotic fibroblasts. And in agreement with elderly cell defects in mitosis, the top 10 most altered gene ontology terms included, uh, sick cell cycle and mitosis-related gene ontologies. And very interestingly, the 71 mitosis gene ontology genes that were significantly altered in the elderly samples, they were mostly downregulated. So this means that the elderly mitotic cells, they are entering mitosis with a very low level of mitotic transcription. Interestingly, these genes are targets of the BMI-FOXM1 transcription complex, which drives the expression of G2/M cell cycle genes containing the homology region CHR motif in their promoters. Therefore, we ask whether the loss of mitotic proficiency during normative aging was due to FOXM1 downregulation. Indeed, there is a decrease in the levels of both FOXM1 transcripts and protein levels with advancing age, as you can see in this graph. So this is in agreement with the gradual reduction that we saw we saw in mitotic. So, this is in agreement with increased duration of mitotic duration with advancing age. So, we concluded that an aging-associated FOXM1 repression accounts for an extensive transcriptional downregulation of mitotic genes and the phenotypes that we observed in the early senescent dividing Moreover, a comprehensive list of 55 genes that were identified as comprising the senescent core signature was interrogated in the RNA-seq dataset. Although the transcriptome of senescent cells is highly heterogeneous and depends on the cell type and the senescence stimulus. There are 55 genes that were identified as altered in different types of senescent cells in response to different stimuli. And some are downregulated, whereas others are upregulated. Interestingly, we found that 19 genes out of these signature, they were differentially regulated in the RNA-seq dataset of mitotic octogenarian dermal fibroblasts, with 15 of them behaving as reported. Thus, an unforeseen SAS phenotype evolves in healthily dividing cells alongside a global transcriptional shutdown of mitotic genes accounting therefore for the aging-associated mild aneuploidy levels. To test the correlation between FOXM1 repression and age-associated phenotypes further, we depleted FOXM1 in fibroblasts from a 10-year-old donor using RNA interference. And again, to uncouple gene expression analysis from the reduction in the mitotic index following the RNAi, we performed RNA-seq profiling of cells capturing mitosis accordingly to the experimental layout that I have shown before. And the RNA-seq revealed that the expression, that FOXM1 repression altered the abundance of 5,000 gene transcripts, which were again enriched for cell cycle and mitosis gene ontologies. The top 10 most altered GO terms included mitosis, as I told you, and also the cell cycle genes controlled by this important transcription factor FOXM1. Moreover, we found that 50% of the genes that were altered in old age fibroblasts, they were also altered in the FOXM1-depleted cells. FOXM1-depleted young fibroblasts, they displayed again a mitotic delay and increased aneuploidy similar to old age fibroblasts. The percentage of cells with senescence markers, they also increased. And again, we found that there is an altered expression of the SASP genes. And so the results show that the young cells with low levels of FOXM1 equivalent to those that we found in octogenarian cells, they recapitulate all the aging-associated mitotic defects and neuploidy and senescent phenotypes. As FOXM1 depletion reduced mitotic fidelity, we hypothesized that FOXM1 overexpression in the early cells should counteract the aging-associated mitotic decline. So, we did the expression of constitutively active truncated form of FOXM1. It is a truncated form because we need to remove this N-terminal domain, which does an autoinhibitory loop in the transactivation domain. And this N-terminal inhibition, it's only removed if there is CDK activity, which we know is low in healthy cells. And so using this truncated form, we can generate a form that is constitutively active and independent of CDK activity levels. And we transduce cells with lentivirus for the expression of this truncated form.. And we found that the mitotic genes that were depleted in the aged cells and also by RNAi, they were being recovered upon the overexpression of this active FOXM1. And we also found that the mitotic delay and the mitotic defects that we were seeing before in the octogenarian cells, they were also rescued. So now cells, they are doing mitosis much faster and under fidelity. Because mitosis is more efficient, we also saw, of course, that there are decreased aneuploidy levels. In these octogenarian cells under FOXM1 overexpression. Furthermore, the percentage of cells with the senescent biomarkers was also decreased consistently with the amelioration of SASP and senescence core transcriptional signatures. So those senescence-associated genes that were down, they come up, and those that were up, they go down. So, overall, these data demonstrate that the modulation of mitotic efficiency through FOXM1 induction in elderly cells prevents aneuploidy and delays cellular senescence. To directly demonstrate that chromosomal segregation acts as the ultimate trigger to the establishment of a full senescent state, we perform the experimental layout shown here. Octogenarian fibroblasts expressing H2B-GFP. They were followed under long-term time-lapse microscopy. We use a plate with coordinates so that at the end of, of the movie, we can fix the plate and identify the cells that are daughter cells from cells that divided during the first day. And so, basically, we record the cells for 3 days, we end the imaging, we do the immunostaining procedure for senescent markers, we acquire these cells, we see which ones are senescent or not, and then using the coordinates of these cells, we can go back into the live cell imaging analysis and see which mother cell gave the— or generated the daughter cells that we are examining. So, we found that using this technique, the daughter sulfate of mitotic cells that do not have any chromosome segregation defects, they normally keep on cycling with a low percentage of cells that arrest in the cell cycle during the timeline of our experiment, and only very few becoming positive for senescence markers. However, we noticed that cells from mitosis with missegregation events such as the one you can see here. So, there is a daughter cell that has a micronuclei that is generated due to an anaphase lagging event. So, following this cell and staining for the senescence markers at the end of the movie, you can appreciate that this cell has beta-gal positive signaling, which are these green dots that you see in the cell. It also has a positive signal for the DNA damage marker 53BP1 and the p21 cell cycle inhibitor. So, when quantifying all these cells that we saw misaggregation events, we noticed that in this case the situation is that all the daughter cells will arrest in the cell cycle, and 85% of the cells, they actually become positive for the senescence markers. So, in this way, we could show directly that cells that missegregate chromosomes, they will generate an ultimate trigger to become permanently arrested in the cell cycle and generate the senescent phenotypes. So altogether, our data supports a model in which proliferating naturally aged cells with FOXM1 insufficiency and with a senescence-associated transcriptional priming, early senescence, they lose mitotic fidelity, that, if ending in chromosomal segregation, ultimately triggers the accumulation of full senescent phenotypes in the aneuploidy progeny. By reinstating FOXA1 transcriptional activity in the elderly cells, we could ameliorate the cell autonomous and also the cell non-autonomous feedback effects between mitotic fidelity and senescence, suggesting that this mechanism actually evolved as a positive feedback loop between cellular aging and aneuploidy. So, if chromosomal missegregation accounts for full senescence, this means that chromosomal instability inhibitions should significantly delay senescence. Chromosome missegregation is often due to lagging chromosomes as the one used occurring right now in the movie of this H-cell. And lagging chromosomes, they form whenever overstable erroneous interactions between kinetochores and microtubules are left uncorrected into anaphase. The stability of these kinetochore-microtubule interactions has to be tightly regulated within levels that allow enough stability for the formation of the attachment, but that also provide enough instability for any correction. So, if the microtubule interactions are very hyper-stable,, you tend to have chromosomal instability. If these interactions are hypostable, then you generate a mitotic arrest. So, the levels need to be here. And this stability of the interactions is safeguarded by microtubule destabilizing kinasins the KIF to be an M-Cog, and thus whenever the error correction machinery is compromised, improper kinetochore-microtubule interactions will originate aneuploid daughter cells. So, we ask if microtubule-kinetochore interactions are somehow affected in old cells. And indeed, suggesting that that is the case. We found that the MCAC levels, so the levels of this important kinase protein that controls the stability of microtubule interactions, this kinase is downregulated in the octogenarian cells as measured here by Western blot analysis and immunostaining analysis. And in line with this finding, we actually confirmed that there is an increased K-Fiber intensity in the metaphase cells of elderly samples. And you can appreciate that in the graph. So the interactions are hyperstable. And can we correct this defect. We tested, we took advantage of an agonist that was available for this depolymerizing kinase, MCAC. It's an inhibitor known as OMK-57 in elderly cells. So we treated cells with this compound, which is known to improve the activity of MCAC and destabilize the kinetochore-microtubule interactions. We measure the optimal concentration of WNK to be used in the elderly cells by measuring the mitotic duration. And you can see here that at 1 micromolar concentration there is a rescue of the mitotic delay. And importantly, the duration of mitosis in neonatal fibroblasts is not affected by this drug concentration. So using UMK57 agonists of MCAC activity at 1 micromolar, we found that the metaphase of elderly cells no longer exhibited overstable K-fibers. And so we were rescuing the microtubule dynamics in elderly cells. And we found that this UMK57 treatment was also able to reduce chromosomal instability. Importantly, we found that by preventing chromosomal instability, we could also decrease the number of micronuclei that were staining positive for C-GAS and negative for retinoblastoma, which is a signal that is very important to generate the senescence phenotype. And in line with this, we also found that 2MK57 was able to delay senescence because we found a reduced number of cells staining positive for the senescent markers. When this agonist was used and published for the first time in cancer cell lines, the chromosomal instability was rescued, but only transiently. So, there was this issue whether the treatment would be able to be performed in a long-term scheme. So, we tested to discard this adaptive resistance to a prolonged WMK57 exposure in the aged cells, we measure all the parameters of being rescued by the treatment after 96 hours and after 4 weeks. And you can appreciate here, I'm showing you the 96-hour treatment, but it was the same after 1 month. So you can still rescue the K-fiber intensity in the elderly cells after 3 days. You also see a rescue in the anneosomy index. The same for the C-GAST positive micronuclei. And the same for the percentage of cells with senescent markers.. So we demonstrated that an adaptation that was observed before in cancer cells does not occur in our aging model, since upon long exposure to the drug, there were benefits on mitotic duration, chromosome segregation, and senescence markers that were still being rescued. So, in conclusion, we disclose the first small molecule inhibition of SIN that delays senescence. The disturbed kinetochore microtubule dynamics is a major driver of chromosomal instability in elderly cells, and this contributes to senescence. The aged cells, they exhibit defective kinetochore microtubule depolymerising activity of Mcac, and this compromises the error correction machinery. So, we found that by overexpressing Mcac, even though I did not show this data, but also by treating with the Mcac agonist UMK57, we could rescue these overstable attachments, we could prevent aneuploidy events, and we were also rescuing the senescence, the accumulation of senescent cells, which is mainly due to the activation of the Sigalskin pathway pro-inflammatory signaling. We are now addressing the impact of both FOXM1 genetic modulation and UMK57 pharmacological epigenetic biological intervention in other cellular aging hallmarks such as the epigenome and the 3D chromatin architecture, as well as we are investigating for the benefits in organismal aging. I hope very much to disclose these findings in a near opportunity. So now I finish by acknowledging the funding agencies and members in the lab, particularly Rui Ribeiro, Joanna Macedo, and Monica Villar, whose results were described here today. I also like to acknowledge my collaborators, Duane Compton, and lab members at Geisel Medical School in the US, and for their collaboration on the MK57 Compound work, and also Floris Ploeger at Teriba, Netherlands, and Hanne Medema at the NKI, Netherlands, for their contribution in our FOXM1 study. So under the healthy aging goal for the future, we like to use this parallel with our popular vintage Porto wine that has been improving for decades in the bottle. Thank you so much for your attention, and I'm pleased to take any questions.