Epigenome Reprogramming

A review of methods used to rejuvenate organisms

19 min readApr 3, 2021

In the past years (2012–2021), multiple exciting discoveries in the world of genomics have been made. From CRISPR, that allowed us to edit more efficiently, to revelations about the relationship between genome and epigenome, we are now more equipped than ever to rejuvenate organisms efficiently by modifying their genes.

Epigenome reprogramming is one particularly impactful way of achieving this. Itcan be done multiple ways. In this review, I explore the differents methods used to achieve it and their effects based on recent experiments. For each method, I will expose exactly what can been modified, why the chosen target of modification has been chosen, what the modification does to the organisms’ cells and what this might mean about the efficiency of the method for epigenetic reprogramming.

***A little disclaimer before we start: in this review article, I will not be writing about the impact of interventions of lifestyle on epigenetic age. Only interventions related to gene editing will be discussed. This means that if you’re interested in calorie restriction, or the impact of certain medications, this article is not the place to go. However, if you’re curious about the impact of tweaking sirtuins or the Yamanaka factors, read on.***

***A second disclaimer: this piece is not meant for beginners. That is not to say I am an expert. Just a biotech enthusiast.***

Outline of Paper

Method 1: Modifying Sirtuin Activity:

1.1 What are sirtuins?

1.2 SIRT6

1.3 SIRT1

2. Method 2: Editing Yamanaka Factors:

2.1 What are the Yamanaka factors?

2.2 OSK

2.3 SKM

3. Conclusion

4. Sources

1. Method 1: Modifying Sirtuin Activity

1.1 What are sirtuins?

The first thing to know about sirtuins is that they are part of the deacylase class, which is defined as “any hydrolase that removes an acyl group by hydrolysis.” To break it down further, an hydrolase is a class of enzyme that uses water to catalyze a biochemical reaction that breaks a chemical bond. In the case of sirtuins, that chemical bond is the one that links an acyl group to a histone (histones are the main proteins that make up chromosomes, thus being crucial for DNA stability).

Why is this important? Because acyl groups relate to DNA acetylation. And DNA acetylation relates to epigenetic programming. You see, DNA acetylation, as well as DNA methylation, regulates how DNA is expressed. Acetylation or methylation of a particular locus will make the protein encoded by the gene in that particular locus produced in lower amounts, and less acetylation (or methylation, I won’t say it every time) will have the reverse effect. As we age, the DNA acetylation pattern changes, some loci because hyperacetylated, while other become hypomethylated, causing changes in our phenotype and health status. This is because some proteins become over-expressed, while others become under-expressed, causing inadequate amounts to keep our bodies functionning properly over time.

To give you an example, take melanin. Melanin is not a protein in itself, but it is derived from the amino acid tyrosine. The amino acids the cell produces are regulated by genes, because amino acids make up proteins. Thus, when the locus for the gene involved in melanin production become hyperacetylated, our melanin production decreases, and this is why our hairs turns white as we age.

It should then be understood that sirtuins, by playing a key role in programming the epigenome, are responsible for many essential cell function that aim to keep us young. Such functions are, for example, enhancing metabolic efficiency, upregulating the mitochondrial oxydative mechanism (which aids resistance to oxydative stress), increasing antioxydant pathways and facilitating DNA repair, notably by de-acetylating loci for repair proteins.

This is why many researchers over the past few years thought that by increasing sirtuin activity, they could, to some extent, reprogram the epigenome of organisms to expand their longevity. Let’s see some of the sirtuins that have been experimented with, and the accompanying effects on longevity.

1.2 SIRT6

The first modification we’re going to look at is the overexpression of the sirtuin SIRT6. It is one of the seven sirtuins found in mammals, simply named SIRT1 to 7. SIRT6 was a strategic choice for the researchers to target, because it regulates so many of the phenotypic manifestations of aging. It is involved, for example, in glucose and lipid metabolism, inflammation management, genomic instability mitigation and thus, cancer prevention as well.

Another interesting thing about SIRT6 is that it is a mono-ADP-ribosyltransferase of DNA polymerase 1. This means that it adds an ADP-ribose group to the protein DNA polymerase 1 after it has been synthetized by a ribosome, allowing it to correctly accomplish its function. As you may know, the function of DNA polymerase 1 is to copy DNA before cell division. It then ensue that its proper functionning is crucial to correctly translate the genome from the old cell to the new. Lower amounts of SIRT6 are thus hypothesized to make DNA replication more error-prone, driving aging.

This is not just theoritical; there have been experiments where mice overexpressing SIRT6 have exhibited a younger phenotype. And by phenotype, I do not only mean what is directly visible, but also what is behaves differently (proteins, cells, etc). Those effects are the ones I will mostly focus on now and throughout this paper. Please bear in mind that it is impossible for me to mention every single observed effect, and so I will focus on the most relevant ones. If, however, you are interested in knowing more about any experiment I will talk about, their respective papers will be sourced at the bottom.

The first effect the overexpression of SIRT6 had is to modify the production of the p16 protein, a tumor suppressor. Due to genomic instability, one of the hallmarks of aging, cells become more cancer-prone when aging. Thus, the body must work harder to prevents its cells from becoming cancerous and that is done through tumor-suppressing proteins such as p16, p21 and p53. It follows, then, that older cells express higher amounts of such proteins. The amount of p16 mRNA was measured in mice modified to express higher levels of SIRT6 (MOSES mice) in five different tissues: liver, spleen, thymus, kidney and white adipose tissue. The finding was surprising: higher levels of SIRT6 seemed to downregulate the expression of p16 in some tissues (spleen and thymus), to change nothing for other tissues (liver and white adipose tissue) and to upregulate it in the kidneys. It is not fully clear why SIRT6 did not only downregulate the expression of p16. However, due to other observed effects, the research still concluded that the overexpression of SIRT6 causes a younger phenotype. Let’s see why.

Possibly the most important thing that overexpression of SIRT6 caused is differences in gene transcription. What I mean by that is that 2385 genes were expressed differently in young wild-type mice and old wild-type mice, while only 175 genes were expressed differently between young MOSES mice and old MOSES mice. This is very significant when our main focus is epigenetic reprogramming. Indeed, this finding means that the epigenome of old MOSES mice ressembled the epigenome of young MOSES mice more closely than their wild-type counterparts. In other words, their epigenome hadn’t changed that much with aging. Granted, it hasn’t been “reprogrammed” per se, but this overexpression of SIRT6 has close to eliminated the need for such reprogramming by preventing harful programming in the first place.

So, what implications might this bear for the reprogramming of the epigenome? Clearly, a genotype allowing higher levels of expression of SIRT6 is a good thing to be born with, as it prevents much of the degradation of the epigenome. However, we do not know if inducing higher production of SIRT6 in wild-type organisms with gene-editing will bear the same effects, as it has not been tested out. The answer is likely to be yes, since lifestyle interventions that upregulate sirtuin activity, such as calories restriction, have been proven to improve many hallmarks of aging. Still, this is a gap in our knowledge that is worth mentionning.

Additionally, you may have noticed that the observed effects of overexpression of SIRT6 indicate a younger epigenome, but that not actual epigenetic age has been determined for the MOSES mice cells, and that no comparison has been made with their chronological age. It would have been possible to do exactly that, since this is what the Horvath clock, and many other epigenetic clocks, allow for, and the Horvath clock specifically has been invented in 2013. The experiment with the MOSES mice was conducted in 2017. We then have to remember that while the data collected in this experiment about the mice’s phenotype is strong, it is not as strong as it possibly could have been, and that a data point that would have been extremly compelling if collected has been left out.

1.3 SIRT1

The second modification we’re going to review has to do with SIRT1, the most widely studied of sirtuins. It is also somewhat controversial. Early studies have showed that Sir2 and its mammallian orthologs (which SIRT1 is part of) are important for mediating hallmarks of aging in multiple organisms, such as yeast, worms, and flies, through environmental stressors as a trigger. However, SIRT1 specifically has been found to be useless in the case of full-body overexpression. That being said, brain-specific increase in SIRT1 expression has been shown to reduce the effects of aging in mice.

SIRT1 is known for being important to the regulation of metabolic responses in many tissues such as the liver, skeletal muscle, adipose tissue and brain. This regulation is done through its deacetylase activity. We talked about deacylases and their role; deacetylases are very similar, but their particularity is that they cleave an acyl group, specifically, from the N terminal of a histone.

Why, then, does SIRT1 have no effect on longevity and aging phenotype when overexpressed in the whole body, but is beneficial when overexpressed in the brain? We do not know the answer to the first part of that question, but we can explain the second part. For that, we have to learn some background about brain physiology. The part of the brain that particularly interests us in this case is the hypothalamus. It communicates with many other tissues through the medium of hormonal and neural networks, coordinating metabolic and behavioral responses to environmental stimuli. The specific part of the hypothalamaus that is important to explain the importance of SIRT1 is the neurons that produce growth-hormone (GH)-releasing hormone and somatostatin. The former stimulates GH release in the anterior pituary gland, while the latter inhibits it. A particularly interesting GH released by the anterior pituary gland is the insulin-like growth factor (IGF-1). Its link to longevity in mammals is well-established.

Graphs illustrating longevity in BRASTO and control mice, and death causes. https://www.semanticscholar.org/paper/Sirt1-extends-life-span-and-delays-aging-in-mice-of-Satoh-Brace/5658512b7a6f9c2f1896ef797646af849c6cbf05

It may be because it this, then, that mice modified to overexpress SIRT1 in the brain (BRASTO mice) show multiple signs of decreased aging phenotype. The first and most straighforward one is the simple lenght of their lives, measured in days. Overall, BRASTO mice showed a 11% expension of median lifespan control compared to the control group, who lived 926 days compared to 835. On the left side, you can see multiple graphs representing this finding. At the bottom are graphs illustrating causes of death, and we can see a clear trend: for any studied cause, BRASTO mice die at a slower rate for longer.

A second hallmark of aging that was reduced in BRASTO mice is mitochondrial disfunction in the muscle cells. As you can see on the picture at the bottom, the muscle cell mitochondrias of wild-type mice were significantly more absent, fused or swollen, all indicating abnormality. The proper functionning of mitochondrias is essential for organismal health, as they produce the cell’s energy source. When a motichondria becomes dysfunctionnal, the cell can be lead towards senescence, another hallmark of aging.

Picture indicating abnormal mitochondria in wild-type control mice (white arrows), and comparison to BRASTO mice. https://www.semanticscholar.org/paper/Sirt1-extends-life-span-and-delays-aging-in-mice-of-Satoh-Brace/5658512b7a6f9c2f1896ef797646af849c6cbf05

But why did BRASTO mice exhibit younger muscles? To explain it, we have to understand another part of the brain. We previously discussed the hypothalamus; well, the hypothalamus activates the sympathetic nervous system when needed. One of the sympathetic nervous system’s roles is to provide stimulation to the skeletal muscles, causing an increase in the expression of the β2 adrenergic receptor (Adrb2). What is this, and why is it important? Let’s explain. Adrb2 is a receptor present on all the surface of the cell membrane whose role is to bind adrenaline. Through its signaling, adrenaline regulates physiologic responses such as smooth muscle relaxation. Smooth muscle relaxation occurs when inhibitors of the Rho kinase, such as fasudil and Y-27632, block its action by taking up space on the binding sites of the kinase, that would normally have been occupied by ATP, the cell’s energy source. The Rho kinase is a protein regulating cell movement by acting on the cytoskeletton. We can then hypothesize that since muscle relaxation was promoted though Adrb2, the muscle cells used less ATP, downregulating mitochondiral activity, thus delaying mitochondrial dysfunction.

SIRT1 overexpression in the brain of BRASTO had many other fascinating effects. For the sake of conciseness, no more of these effects will be exposed here. However, from the findings of the study, we can conclude one thing: if we want to use the upregulation of SIRT1 to enhance longevity, the data we currently have indicates the modification should target the brain only.

2. Method 2: Editing Yamanaka Factors

2.1 What are the Yamanaka factors?

The Yamanaka factors are a set of genes highly expressed in embryonic stem cells (ES). Those genes code for transcription factors, which are relevant to the epigenome. Indeed, certain transcription factors are referred to as master transcription factors. Those factors directly bind to DNA (not to histones like sirtuins) and regulate the expression of the genes they bind to. The Yamanaka factors are master transcription factors, and as we will see, they could be considered the most important ones for longevity.

When expressed, the Yamanaka factors regulate the methylation of DNA. In ES, they cause the cells to adopt a methylation profile that is both dedifferentiated and youthful. The idea when overexpressing the Yamanaka factors in an organism is to reprogram the epigenome for it to adopt back a youthful state, but not a dedifferentiated one.

There are four Yamanaka factors: Oct-3/4, Sox2, Klf4 and c-Myc. I will refer to them as OSKM when talking about the four of them, and leave out some letters when talking about only some of them (obviously leaving out the letter corresponding to the left out factor). The role of each individual factor in inducing a youthful genome is not fully understood yet, although studies have attempted to shed light on this knowledge gap. One of them has particularly helped to better understand the role of c-Myc, but that of other factors remains obscure.

Indeed, a study conducted in 2009 found that the genes bound by c-Myc ,with or without additional bonding with OS or K, during epigenome reprogrammation were significantly involved in regulating metabolic processes, such as the control of translation, RNA splicing, cell cycle and energy production. On the other hand, genes bound by factors other than c-Myc were mainly involved in translational control of development.

2.2 OSKM

Reprogramming the epigenome has been done using many different combinaisons of the Yamanaka factors in the past. In 2016, a study on mice aiming to induce overexpression of all the four Yamanaka factors, and by doing so, rejuvenate the aforementionned mice, successfully reached its goal, although not without casualties in the way.

When writting about sirtuins, I could explain the rationale behind experimenting with the sirtuin of interest. When it comes to the Yamanaka factors, as I mentionned before, this is not possible, as the specific roles of each of the factors are not clearly known. I will then, in this section and the two next, skip the part where I would normally explain why a specific target of interest has been chosen, and skip straight to the results.

Here, there are two groups to observe. Normally, I would only analyse the experimental group, and compare it with the control group if needed. In this case, the control group is a little bit different. You see, the goal of the study that made mice overexpress OSKM was to see if intermittent overexpression led to rejuvenation. It then had to be tested against continuous overexpression. And so it was, and the mice that were made to continually overexpress OSKM ended up losing a significant amount of weight and, after 4 days, died in high proportions. This is explained by the dedifferentiation that likely occurred in these mice’s vital organs, due to too high a dose. This is why uncontrolled dedifferentiation in-vivo is so dangerous: it turns specialized cells, made to accomplish a specific function, into pluripotent cells, able to create any kind of cell, but not very helpful on their own.

The experimental group’s story is a bit more cheerful: the experiment was a near-complete success for these mice. They showed improvement in all measured phenotypic manifestations of aging, except weight loss, which was not prevented. Some things that were prevented, though, are genomic instability, hetorochromatin degradation, senescent cells, mitochondrial damage, and nuclear envelope degradation.

First, for genomic instability, the treatment significantly decreased the amount, intensity and volume of the spaces where the histone γ-H2AX can develop. This means that fewer DNA double-strand breaks occured. Why? Because the histone γ-H2AX repairs those breaks. It follows, then, that when fewer spaces for this histone develop, fewer double-strand breaks have occured. In addition, the amounts of the p53 binding protein 1 (p53PB1) in the mice’s cells also decreased. This means the organism had to work less to suppress imminent tumors, because there were fewer of those. The link from the p53PB1 to this conclusion is that one of p53PB1’s roles is to mitigate DNA damage, which isn’t surprising when we know that the consequence of a mutation in the p53 gene often results in cancer. Furthermore, overexpressing the Yamanaka factors in an intermittent fashion downregulated the expression of age-related stress response genes in the p53 tumor suppressor pathway. Examples of such genes are p16INK4a, p21CIP1, Atf3, and Gadd45B. Again, less work required by the body to prevent cancer formation.

Second, improvements were noticed in the cells’ heterochromatin. The heterochromatin is responsible for supporting and protecting the chromosomes, thus preventing genetic material damage. The treatment upregulated the activity of H3K9me3, which is normally downregulated with aging. Plus, it downregulated the activity of H4K20me3, which is upregulated during aging. Both H3K9me3 and H4K20me3’s normal activity are essential for proper heterochromatin maintenance.

Third, the treatment also improved the mice’s health relating to senescent cells. Indeed, the senescence-associated metalloprotease MMP13 and interleukin-6 was downregulated, which means fewer senescent cells were created. Senescent cells are generated as a means of the organism to protect itself. When a cell’s DNA is damaged, it can die, become cancerous, or become senescent. Ideally, the body uses autophagy to kill those damaged and potentially dangerous cells, but as it ages, it progressively looses its ability to do so. Thus, it makes them senescent, which is less dangerous than letting them become cancerous, at least in the short-term. Another senescence-related observed effect was the reduction of senescence-associated β-galactosidase activity.

Fourth, the overexpression of OSKM decreased the production of oxygen reactive species (ROS) by the mitochondria. ROS are produced when damage has occurred in cells, as I explain further in this article about the mitochondria. This means possible lessened mitochondrial damage, as well as lessened overall stress in the cell.

Lastly, nuclear enveloppe damage was lessened. This is remarquable since the mice used for this experiment were progeroid, which means afflicted by progeria. This disease causes, among others, the accumulation of progerin in the nuclear enveloppe, thus degrading it. This is a driver of premature aging in organisms that suffer from the Hutchinson-Gilford Progeria Syndrome, as well as in healthy organisms, for which it happens over a longer period of time.

The implications of the results of this study are significant. Since c-Myc is an oncogene, it is always feared that its overexpression will lead to tumor growth. However, no tumor formation was observed. This is explained by the researchers by hypothesizing that that this intermittent method of expression could potentially be responsible for the prevention of cancer as an unwanted side-effect of this treatment.

**All being said, if you want to know more about this particular experiment, I wrote another article specifically about it, where I go more in-depth about the methods used.***

2.3 SKM

Another possible combinaison to use to reprogram the epigenome of cells is comprised of SKM. This was a surprising finding, because until an experiment proved otherwise, it was thought that Oct-4 was the most important of Yamanaka factors. Not only did a 2019 study prove that reprogrammation only using SKM was possible, but it also yielded induced pluripotent stem cells (iPSCs) that were of greater developmental potential.

Now, you might think this is not so related to aging, since clealy the aim here was to generate iPSCs, whereas for rejuvenation, we want to avoid creating iPSCs. Still, as we saw in the above experiment, rejuvenation lies just a step before pluripotency. The epigenome is reprogrammed in either case, it is just reprogrammed further back with iPSCs. The experiment that supports my claims here is still very relevant to the topic, as it is highly likely that its finding are transferable.

You may be asking yourself how the SKM-induced iPSCs are superior to their SKM-induced counterparts. First, you have to know that OSKM-induced iPSCs suffer from loss of imprinting (LOI). LOI happens when a monoallelic gene is no longer regulated by the parent-of-origin DNA methylation. This is a common feature of human tumors. Second, they present karyotypic instability, which translates to a high amount of DNA mutations. Those all have the potential to lead to cancer. Third, they aberrantly express some genes (overexpressing or underexpressing them), making them somewhat unreliable. And fourth, they are prone to mitochondrial defects. While SKM-induced iPSCs still bear many of these imperfections, they are not prone to LOI.

This is not the only way in which SKM-induced iPSCs are superior. To determine developmental potential, the researchers subjected the cells to the most precise test of pluripotency to exist: tetraploid complementation assay (4NCA). This is a technique in which a two-cell embryo gets both its cells fused by an electric current to form one teraploid cell. Such cells continue to divide and form a blastocyst that can implant in the uterine wall and form extraembryonic issues such as the placenta, but rarely form a fetus. This is when the iPSCs come into play. The 4N cells are combined to iPSCs from another organism. If the iPSCs have good developmental potential, a normal fetus will form out of the iPSCs and the 4N cells will compose the extraembryonnic tissues. The more developmental potential the iPSCs have, the more fetuses will survive and the more orgaisms will normaly develop into adulthood.

5 iPSCs lines for both SKM and OSKM-induced iPSCs were tested. All the SKM-induced iPSCs developped into pups, while only 3 out of the OSKM-induced iPSCs also did. 44.1% ± 9.1% of SKM-induced iPSCs generated fully developed pups, in contrast to only 2.3% ± 1.4% for OSKM-induced iPSCs. Additionally, none of the OSKM-induced pups lived through foster nursing, while 34% of the SKM-induced pups survived until adulthood. Of those, 100% were healthy and fertile.

However, this does not mean that there is no use for O in the reprogramming cocktail. The absence of this factor makes more doxycycline use necessary to achieve pluripotency and it is thus hypothesized that it makes establishment and maintenance of pluripotency more difficult. Hence, this may indicate that O overexpression towards the end of the process of becoming pluripotent would be beneficial and could possibly combine the best of both worlds, although this has never been tested. Another way to get to reap all the possible benefits of different factor combinaisions could be to lower OSKM overexpression, and/or to use a rapidly disappearing vector, and/or to use a certain O:SKM ratio. To reason this is thought is that one out of 3 OSKM and episomal-generated iPSC lines developped, but without the defects usually associated to OSKM-induced iPSCs.

Those results are transferable to the field of aging, as it also seeks to generate high quality reprogrammed cells, although less reprogrammed than iPSCs. Apart from that, the two discipline are strikingly similar, at least when we turn to the Yamanaka factors as our fountain of youth.

Conclusion

To conclude, the methods we can use to reprogram the epigenome are diverse. In all methods, it is necessary to modify the expression of a cellular component that plays a role in histone modification, either acetylation, deacetylation, methylation or demethylation. Sirtuins are a widely studied group of deacylases and multiple past experiments have proven that the modification of their activity can extand healthspan and lifespan. The Yamanaka factors are a group of transcription factors known to be highly expressed in embryonnic stem cells and, as this indicates, involved in induction of pluripotency and, most importantly in our case, epigenome reprogramming. To give a cell back its youthful phenotype, it is necessary to overexpress the Yamanaka factors at a lower level than to give it back its pluripotency. The concepts of youth and pluripotency are hence closely related.

Many questions are left to be answered. In the case of sirtuins, it would be helpful to the field of longevity to determine if overexpressing SIRT6 in organisms born as wild-type has the same effect as for transgenic mice. In addition, it would be interesting to use the Horvath clock to discover the epigenetic age of SIRT6 transgenic mice cells. As for SIRT1, it remains a mystery why its overexpression is only beneficial if induced exclusively in the brain. In the case of the Yamanaka factors, discovering the precise role of each of OSKM could make us discover new reprogramming combinaision with unprecedented potential. Additionally, it would be interesting to test whether a lower level of overexpression of an OSKM cocktail would yield youthful cells devoid of important defects, and/or if a faster disappearing vector helps in that case as well. We could also experiment with different O:SKM ratios.

Sirtuins and the Yamanaka factors aren’t the only components we can modify in order to reprogram an epigenome back to its youthful state, at least partially. Any component that acts on histones could be worth investigating, as some have abeen. By researching those, we could discover new important groups of components for the epigenome that could potentially give rise to therapies in the future.

Although sirtuins are promising, the Yamanaka factors seem more so. Indeed, it has been proven that when overexpressed in an organism that previously wasn’t overexpressing them (the mice in the OSKM section), they induce back a youthful phenotype. On the other hand, sirtuins have only been proven to maintain a young phenotype.

On a final note, it seems we are very close to achieving the breakthrough that will finally give rise to the first healthspan extension therapies. Thanks to these advances in the research that give us significant insight into the effects of various gene-editing interventions on the epigenome, the research can now become more and more targeted towards not only understanding the cells, but finding the best way to rejuvenate them. Lasting youth might be closer than we think.