ENERGY SCIENCE
Full-Text Link to Primary Source
Most of us remember fairly well two main things from high school biology: that we inherit stable sets of chromosomes from mom and dad, and that mitochondria are the powerhouse of the cell.
New research shows that neither are quite right. Rapidly jumping genes from mitochondria to nuclear chromosomes show how dynamic our cell’s genomes are, which might have implications for how long we live.
Reconsidering How We Think About Biology
A popular conception is that our cells work like little machines. The outdated analogy goes as follows: genes control cellular functions by making RNA, which in turn make proteinaceous cogs and wheels that mechanically churn out substrates into products, bind to one another, and make up cellular structures and organelles.
Somehow, the interactions of the molecular cogs give cells the ability to sense, produce appropriate responses, and faithfully replicate to produce two little clones of themselves. Repeat a few million times, and you have an organism—a breathing, thinking, feeling, conscious organism.
“But you see, this isn’t quite how life works,” Philip Ball tells us in his book How Life Works.
One core concept getting overturned by new biology is the concept of stability. Machines are stable, predictable, and immutable things—until they break, that is. Cells and organisms, on the other hand, are dynamic, changeable, and adaptive systems.
The dynamic dance of genes gives rise to “the music of life.”
The musical score cryptically inscribed in the four-letter code in the genome is interpreted by the flux of energy and information, fueling a collective that is unified in a sort of social contract. Interconnected by ever-changing bioelectrical signals, hormones, and other forms of communication, the cell and mitochondrial collective operates second by second, through an ever changing world, as a dynamic process.
“But” you may say, “the genome is fixed and stable, right?” Not quite. Several dynamic changes happen in the human genome.
In particular, virus-like genes called retrotransposons “replicate” and are “transposed” in new places across our inherited chromosomes. This happens more frequently as cells get older, triggering inflammation and possibly contributing to the aging process itself. Mutations, deletions, and other types of dynamic changes slowly take place in the nuclear genome of our cells all the time.
New Findings on the Prevalence and Timescale of Mitochondrial DNA Insertions
A paper from our lab published in Plos Biology in August 2024 showed that a process presumed to happen at an evolutionary scale only, in fact, happens across the human lifespan. And it happens in just weeks in cultured human cells.
The process involves the ancient “bacterial” genome of mitochondria—the mitochondrial DNA (mtDNA). Mitochondria and their genomes live in the cytoplasm (the gel-like substance within cells), surrounding the nucleus.
In some instances, the mtDNA “jumps” into one of the chromosomes inside the nucleus. Just like a DNA virus that cuts the genome and pastes a copy of itself into it, or like retrotransposons, the mtDNA becomes integrated within the sequence of the human genome. These nuclear mtDNA insertions are called Numts (pronounced “new-mites”).
If an incoming piece of mtDNA gets inserted inside a coding gene or in a regulatory gene region, the new Numt can disrupt important genomic elements. That is, it can break existing genes, or dysregulate their expression.
The field of genomics has known for a couple of decades that, on a few occasions over evolutionary timescales, pieces of mtDNA, or even entire mtDNA genomes, have somehow become integrated inside the human genome and can be inherited by future generations.
This was confirmed in 2023 in a catalogue of human Numts from >60,000 people. The study found that numtogenesis (i.e., the creation of new Numts) occurs in about one in 4,000 new births. If you count the number of generations between you and the mother of all mothers 250,000 years ago, that makes for a lot of mitochondria-derived Numts in our genomes.
As a result, all humans walk around with hundreds of these vestigial mtDNA segments in our maternally- and paternally-inherited chromosomes.
Links Between Mitochondrial DNA Insertions and Health
The process of mtDNA insertion into nuclear genomes is sped up in cancers where it may contribute to the transformation of normal cells into cancer cells. As a result, as Keshav Singh found in 2018, tumors have more Numts than non-cancerous tissue from the same person.
If new Numts happen during our lifespan in tissues like the brain, heart, gut, and kidneys, it could have important, but as yet unknown, functional consequences on human health or lifespan.
Kalpita Karan, a postdoctoral fellow in my laboratory, reached out to Numts expert Ryan Mills at the University of Michigan.
Working with Hans Klein at Columbia, we got access to data from Rush University Medical Center and the team led by David Bennett. Ryan and Arthur were able to hunt for new Numts in more than 1,000 brain DNA sequences.
What Ryan and research scientist Arthur Zhou in his group found was stunning: human brains, particularly the prefrontal cortex (the fore-most part of the brain known for its roles in attention, emotion, complex learning, and other executive functions), had many Numts!
The brain has substantially more Numts than blood. And it has even more than the cerebellum—a small integrative brain structure not thought to be directly involved in working memory and “high-level” thinking like the prefrontal cortex.
There was only one rational conclusion: nuclear mitochondrial DNA insertions happen in the human brain—likely several times over—across a person’s lifespan.
Strikingly, people with more Numts in their prefrontal cortex died earlier than individuals with fewer Numts. In individuals with normal cognition or without dementia, there were about 5 years of life lost for each abnormal mtDNA insertion.
If numtogenesis was inconsequential to the life of a cell and organism—or, in genetic terms, if Numts were “silent”—then neither neurons nor organisms would care. And so, there should be no association between Numts number and lifespan. Or perhaps you’d get more as you age, so the older you were when you died, the more you would have. But we found the opposite.
On the contrary, the significant association between having more Numts and dying earlier suggests for the first time that brain Numts may have functional consequences. They may shape the human lifespan.
Investigating the Origins of Numts
The reason the scientific community had missed this stunning fact for decades is because all previous studies were performed using DNA sequences from blood. But immune cells in the blood are known to undergo quality control—the bad ones are outcompeted and only the best cells survive to be sequenced.
There is selection bias when looking at DNA sequences from circulating blood cells.
But how fast can new Numts arise? To address this question, Kalpita in our group used the Cellular Lifespan Study dataset created a few years earlier by Gabriel Sturm, where cells from different individuals are followed longitudinally as they age and their telomeres gradually erode.
This confirmed what we saw in the brain. In aging human cells in vitro, numtogenesis happened dynamically. And surprisingly quickly! Healthy cell populations, Kalpita found, accumulate one new Numt on average every 13 days. That’s fast evolution.
But like the brains of living people, do some cells experience more numtogenesis than others? Does stress accelerate this process?
Yes—we found that cells of patients with mitochondrial disease (in this case, a genetic mutation that causes a deficiency in the mitochondrial respiratory chain) accumulated Numts up to 4.7 times more rapidly.
Other stressors that were tested only increased Numts moderately and non-significantly. But all cells accumulated Numts over a period of days to weeks. This is in line with the notion that stress costs excess energy (which we’ll cover in subsequent posts). And that stress can compromise the integrity of the genome, and subsequently affect mitochondrial biology.
The study showed a new way in which stress can affect the biology of our cells—by making mitochondria more likely to release pieces of mtDNA that “infect” and change the sequence of the nuclear genome. Mitochondria can dynamically change the genetic sequence in the nucleus!
Overall, by revealing the existence of numtogenesis in the human brain and showing that it takes only a few days to happen in cultured human cells, the new study demonstrates that this “evolutionary” process takes place over a much shorter timeframe than previously believed: days or weeks, not millennia.
Moreover, given that people whose brains have more Numts die earlier, for the aging field, this discovery adds numtogenesis to the list of age-related genome instability mechanisms that may contribute to why we get old, why we experience functional decline, and why we eventually die when we do.
A Research Area Ripe With Unanswered Questions
“How does the mitochondrial DNA get into the nucleus?” you might ask. “Isn’t the mtDNA supposed to be inside a mitochondrion?” If so, how could the mitochondrial and nuclear genomes ever interact to produce these potentially damaging insertions?
It turns out that mitochondria have many ways to release their DNA into the cytoplasm. Once in the cytoplasm, mtDNA fragments can make their way into the nucleus through nuclear pores, or they can seep in while the nuclear envelope dissolves and reassembles during cell division.
More work needs to be done to better understand this phenomenon mechanistically, but the release of mtDNA appears to be a controlled, regulated process from mitochondria.
The electron micrograph of a human cell below show three mitochondria in close proximity to the cell nucleus, separated by the cytoplasmic space. Red arrows show a transfer path for signals and DNA from mitochondria to the inside of the nucleus, through a nuclear pore—a stable opening in the nuclear membrane.
This image is adapted from Mitochondrial Signal Transduction paper, published in Cell Metabolism in 2022, where we describe the Mitochondrial Information Processing System (MIPS)
Nuclear mitochondrial DNA insertions (Numts) involve the transfer of mtDNA segments to the nucleus, where they are genomically inserted, similar to viral DNA from retroviruses.
Another big question is, once in the nucleus, how do Numts “decide” where to insert? If there are Numt hot spots within the genome, why are these sites more vulnerable or receptive to mtDNA insertions?
Further research is needed to compare the rate of numtogenesis between tissues within individuals to get a more holistic sense of whether certain cells and tissues are more susceptible than others. In our experiments on the impacts of stress, the rates of numtogenesis in cells may be impacted by slight increases in cell death, which could preferentially eliminate cells with harmful new Numts, so sequencing dead cells and debris could be informative.
In addition, we still don’t definitively know if Numts directly contribute to disease, aging, and the embedding of psychosocial stress or other forms of stress. We also don’t yet have a sense of whether numtogenesis happens more quickly in some people than others.
If so, could increased genomic disruption by Numts drive accelerated aging?
What these Findings Mean in the Context of Mitochondrial Biology and Our Health
Beyond altering the long-held belief that our chromosomes are largely static, our new study about numtogenesis adds one more way in which mitochondria shape our health beyond energy production—by directly changing the sequence of our genome.
Mitochondria also move about in the cell, fuse and fragment among a dynamic network, sense and integrate information, switch from anabolism (making stuff) to catabolism (breaking down stuff), and they have a life cycle—old ones die and new ones are born.
Machines don’t do these kinds of things, and they are generally unifunctional.
A powerhouse transforms energy from one kind into another, period. The 1957 powerhouse of the cell analogy is expired.
Mitochondria are multifunctional. They behave as dynamic networks and act as intracellular processors—an intracellular brain within each cell.
By speaking the metabolic language of the epigenome, mitochondria even regulate which genes are on or off. And they have a veto on cell life and death, executed by releasing proteins that chew up nuclear chromosomes. With the new findings that they send Numts to be integrated inside the nuclear chromosomes, we now know that mitochondria can effect the most radical identity transformation, by changing the nuclear DNA sequence of a cell. And this has the potential to impact our health and lifespan.
Every second, we breathe to bring oxygen to our mitochondria, feeding life-giving processes within our cells and organs. They are not like machines. Mitochondria are more like little candle flames: combinations of dynamic processes move us. We see this movement everywhere—in our metabolism, our beating hearts, and in our minds.
Now we know that even at the smallest level, in a dialogue between mitochondria and the nucleus, mitochondria rewrite pieces of our genetic code. Like the candle light, we are an ever-evolving, transforming flame of life.
Curious about mitochondria and how they move, connect, and give rise to life and mind? Visit MitoLife to learn more about the beautiful diversity of mitochondria.
This is a fascinating and important reframing!
For decades we treated the nuclear genome as the “master script” and mitochondrial DNA (mtDNA) as a small, mostly static side note. But accumulating data suggest mtDNA variation and signaling aren’t passive; they can influence nuclear gene expression, stress responses, metabolism, and even inflammatory tone through retrograde signaling.
A few implications that make this exciting from a systems biology perspective:
1. Genome is not just sequence, but it’s cross-talk. Nuclear and mitochondrial genomes are in constant bidirectional communication. mtDNA variants can shift metabolic flux, redox balance, and epigenetic regulation in ways that alter nuclear transcriptional programs.
2. Energy state as information. Mitochondria aren’t just ATP factories; they’re sensors. Changes in mitochondrial function can reshape chromatin states and stress signaling pathways, effectively modulating how the nuclear genome is “read.”
3. Heteroplasmy matters. The proportion of different mtDNA variants within cells can influence phenotype, and that distribution can drift with age, stress, and tissue context, introducing another dynamic layer to genetic expression over the lifespan.
4. Disease and aging connections. If mtDNA can reprogram nuclear responses, it may help explain variability in aging trajectories, stress resilience, and susceptibility to metabolic or neurodegenerative disease.
What I appreciate most is that this shifts the narrative from “genes are destiny” to “genes are part of a responsive bioenergetic network.” It opens up a more dynamic view of biology, where environment, stress, and metabolism interface directly with genetic expression.
Really compelling synthesis!
It is wild to think of our brains as a living diary where stress leaves actual genetic footprints. Thanks for the great article.