Yamanaka Factors
And the practical road to age reversal
Well, in mice, they're using these Yamanaka factors to make the mice age the equivalent of like 250 years now. It's really incredible. And there are human clinical trial starting. So the Yamanaka factors, you guys will recall, are the four proteins that were identified that basically can turn any cell back into a stem cell.
Later, there was research done where they took those four Yamanaka factors and they applied a low dose of them to a cell. And rather than have the cell turn all the way back into a stem cell, that cell effectively became young again.
It started to repair and heal itself, repair its DNA, repair its gene expression networks, and the cell returned back to its original state. So the equivalent to think about this in a body is now you've got skin that loses its wrinkles, eye cells that start to see better, brain that starts to work better, muscles that start to work better, and so that is rejuvenation.
—David Friedberg, All In Podcast
What are Yamanaka Factors?
You may already know the headline idea. Four genes named Oct4, Sox2, Klf4, and c Myc can push an old cell back toward a youthful state. Scientists discovered this in 2006, then showed it in human cells in 2007. Turning these genes on fully wipes gene identity and creates stem cells.
The translational move is different. You turn a subset on, and only for short pulses, so the cell keeps its identity, yet regains youthful programs. That is called partial reprogramming.
Yamanaka factors are named after Shinya Yamanaka, who discovered them in 2006 in his lab in Kyoto. The four specific transcription factors he discovered are, Oct3/4, Sox2, Klf4, and c-Myc.
Fast primer for context
Think of the genome as hardware, and the epigenome as software that tells the hardware what to run. Aging shifts that software in predictable ways. DNA methylation changes, chromatin opens and closes in the wrong places, and gene expression drifts.
Yamanaka factors act like a software terminal. Used carefully, they do not replace the hardware, they debug damaged blocks of code toward a more youthful pattern so the code runs the way it used to.
What the factors actually do
Oct4, Sox2, and Klf4 are transcription factors. They bind DNA, recruit other proteins, and reprogram which genes are on or off. When you pulse OSK for short periods, several things tend to happen.
DNA methylation patterns move toward youthful baselines, often through TET enzymes that remove methyl marks.
Chromatin becomes more like the youthful state, which opens access to the right genes.
Whole gene programs involved in protein quality control, mitochondria, and stress responses shift in a youthful direction.
In some tissues, cells regain part of their ability to repair and regenerate.
The downside is c Myc speeds the full reprogramming process, but it also has an added risk of cancer. Considering this added risk, many partial protocols drop c Myc altogether and only use OSK.
Why partial reprogramming
Full reprogramming creates immature stem cells capable of giving rise to several different cell types within the body. These are both powerful and dangerous inside a living animal.
Partial reprogramming aims for a narrow window. You want the benefits of a youthful program without erasing cell identity. Researchers have been able to achieve this with short on/off schedules. A common starting pattern was two days on and five days off when it comes to expressing these factor genes.
That cadence is not magic, it is a starting guess that you tune by examining the tissue you want to affect and vector you are aiming for.
The Research So Far
Proof that the concept can work in a whole animal. A 2016 study used cyclic Yamanaka expression in a fast aging mouse model and improved several hallmarks of aging, including lifespan benefit.
Proof that you can restore real function in a normal tissue. A 2020 study showed that OSK in retinal ganglion cells reset methylation age, regrew optic nerve fibers, and improved vision after injury. Related work in nonhuman primates shows this same functional recovery.
Taken together, this says the method is not just a trick. You can get structure and function back in tissues that see degeneration, and that they care about in their day to day lives.
How delivery and control work in practice
Real translation is engineering.
Vector choice. AAV is the current workhorse. The eye often uses AAV2 for intravitreal delivery. Systemic work often uses AAV9.
Tight control. Most groups use a Tet On system so a simple drug like doxycycline can start and stop the factors. OSK coding sequences are big, so researchers compress them into a single cassette with 2A linkers, or split loads across two vectors.
Target where you want expression, remove where you do not. Tissue specific promoters focus expression. MicroRNA target sites in the message act as tags that cause degradation in unwanted tissues. For example, adding miR 122 sites helps prevent expression in the liver.
Safety features. Prefer OSK over OSKM. Add an inducible kill switch such as caspase 9. Consider degron tags that let you turn the proteins over quickly with a small molecule. None of this removes risk, but it does make the risk manageable.
Where small language models and protein folding models help today
You didn’t think you were going to get away with not reading about AI in this article, did you?
OpenAI, in collaboration with Retro Biosciences, developed a specialized miniature version of their GPT model called GPT-4b micro, designed specifically for protein engineering relevant to epigenetic reprogramming.
Using this model, they AI-assisted the redesign of key Yamanaka factors such as SOX2 and KLF4, achieving over a 50-fold increase in expression of stem cell reprogramming markers compared to the natural versions.
This AI-driven redesign not only enhanced the efficiency of generating induced pluripotent stem cells but also improved DNA damage repair, indicating higher rejuvenation potential.
OpenAI’s model distillation technology enabled creating this smaller, cost-efficient model by fine-tuning it with outputs from a larger, more capable GPT-4o model, allowing faster and cheaper iterations without sacrificing performance, accelerating breakthroughs in regenerative medicine and aging research.
This work was validated across multiple cell types and methodologies, showcasing a significant leap in applying AI to live cell rejuvenation efforts.
The Risks
Three risks dominate.
Oncogenic drift
c Myc raises risk, so most partial protocols remove it. Even with OSK, a small number of cells could slip into the wrong state. You counter this with short pulses, fast shutoff, and a kill switch.
Oncogenic drift is dangerous because the oncogene c-Myc increases cancer risk, so removing it from protocols and carefully controlling factor expression with short pulses and kill switches helps prevent cells from transitioning into harmful, cancerous states.
Immunogenicity and delivery
Many people have antibodies against common AAV serotypes. Redosing can be hard. Drug controlled expression helps, not perfectly.
Immunogenicity and delivery issues arise because many people have antibodies against common viral vectors used for gene delivery, making repeat dosing difficult and requiring careful drug-controlled gene expression, which is helpful but not completely reliable.
Measurement traps
Clocks can move in a good direction while cell composition changes hide a problem. Single cell identity checks and functional endpoints keep you honest.
Measurement traps are problematic because epigenetic clocks can indicate rejuvenation, even when harmful changes in cell composition occur, so verifying results with single-cell identity and functional tests is crucial to avoid misleading conclusions.
The precautionary principle here would indicate we don’t pursue this technology because of these risks. That is a mistake.
Aging is not a neutral baseline, it is a steady loss of function with total market penetration across every human person. The rational stance is to document known risks, build layered controls, start in tissues where you can read signal cleanly, and keep hard stop rules that you follow. This is how you pursue without being reckless. This is how we win.
Next Steps
Replication of the retinal result in an independent lab, with blinded outcome measures.
Single cell evidence that identity is preserved across a full time course, not just at one snapshot.
Durable functional gains in a naturally aged animal, not only in injury models.
A clean safety profile with predefined stop rules triggered rarely, and handled as designed.
Speculative but testable ideas
These are not ready for the clinic, they are promising directions with clear experiments.
Use CRISPR activation to turn on the body’s own Oct4, Sox2, and Klf4, which could improve control and reduce payload size.
Combine partial reprogramming with senolytics, so if a small subset of cells starts to drift you remove them.
Align pulse timing with circadian changes in chromatin accessibility, which might reduce the dose needed.
Use human organoids to tune pulse schedules before you move to animals.
At the End of the Day
Yamanaka factors using partial reprogramming is not a magic reset.
It is a controlled rewrite of epigenetic software that can restore youthful programs while keeping cell identity intact.
If we pair disciplined engineering with honest measurement, and move, then becoming a centenarian will become just another milestone in a storied life.
Additional Reading
Takahashi and Yamanaka 2006
https://www.cell.com/fulltext/S0092-8674%2806%2900976-7
Takahashi et al. 2007 human iPS
https://pubmed.ncbi.nlm.nih.gov/18035408/
Ocampo 2016 partial reprogramming in progeroid mice
https://www.cell.com/fulltext/S0092-8674%2816%2931664-6
Lu 2020 OSK restores youthful methylation and vision
https://www.nature.com/articles/s41586-020-2975-4
Macip 2024 AAV9 OSK lifespan and age reversal in aged mice
https://pmc.ncbi.nlm.nih.gov/articles/PMC10909732/
Life Biosciences nonhuman primate work and program plans
https://www.lifebiosciences.com/pipeline/
https://www.lifebiosciences.com/life-biosciences-presents-groundbreaking-data-at-arvo-demonstrating-restoration-of-visual-function-in-nonhuman-primates/
Muscle regeneration with partial reprogramming
https://www.nature.com/articles/s41467-021-23353-z
Reviews on partial reprogramming and aging clocks
https://www.nature.com/articles/s41467-024-46020-5
https://pubmed.ncbi.nlm.nih.gov/24138928/
https://www.nature.com/articles/s41576-018-0004-3
Detargeting and safety tools
https://pmc.ncbi.nlm.nih.gov/articles/PMC3686499/
https://www.cell.com/stem-cell-reports/fulltext/S2213-6711%2825%2900147-X
Protein language models for design
https://www.nature.com/articles/s41587-022-01618-2
https://www.science.org/doi/10.1126/science.ade2574
Context on AI and drug timelines
https://apnews.com/article/56f4d9e90591dfe7d9d840a8c8c9d553
https://www.wired.com/story/artificial-intelligence-drug-discovery/
It’s super interesting and super scary at the same time. I would think trying to fix vision and hearing issues would be a great first step. Then seeing if nerve damage can be repaired from spinal injuries.