2026-01-08 23:00:00
The treatment converted the liver into an immune cell “nursery” that pumped out greater numbers of healthy T cells.
Our immune system is a fierce brigade. Roaming immune cells scan for bacteria, viruses, and other invaders. They also communicate with tissues to catch early signs of cancer. After detecting a threat, the immune system kickstarts formidable defenses to snuff it out.
But our immunity loses power as we age. Immune cells dwindle, and those that remain struggle to perform their usual roles. As a result, immune defenses weaken, increasing the chances of infection and cancer. This also makes vaccines less effective in older adults.
Now, a new treatment using mRNA technology similar to that in Covid vaccines rejuvenated the immune systems of old mice with twice-weekly shots. The injections transformed the liver into a temporary nursery to boost the numbers and health of key immune cells.
Treated mice, aged the human-equivalent of their early 60s, saw a rapid rise in multiple T cell types after vaccination. They also rallied against tumors with a popular cancer immunotherapy.
Resetting immunity isn’t just about defense. The immune system is intricately tied to the health of other organs. Chronic inflammation steadily rises as we age, wreaking havoc on memory, cognition, and metabolism. It also stiffens tissues in multiple organs, increasing the chances of heart attacks and kidney failure.
“If we can restore something essential like the immune system, hopefully we can help people stay free of disease for a longer span of their life,” study author Feng Zhang at MIT said in a press release.
Multiple immune cell types protect our bodies, but T cells are one of the most prominent.
Some T cells seek and destroy virus-infected cells and cancer. Others coordinate immune responses and balance the attack to prevent autoimmune problems or runaway inflammation. Still more “remember” prior threats to trigger a faster immune response when re-exposed.
Despite their wide range, all T cells are born in the bone marrow. Baby T cells then journey to the thymus, a small organ sitting at the top of the heart, where they mature and diversify. In this nursery, the cells learn friend from foe, ensuring they’ll only attack legitimate threats while leaving healthy cells alone. The process is mostly coordinated by cocktails of proteins and other signaling molecules, which direct the fate of immature cells and help them survive.
The aging process gradually degrades the nursery. The thymus shrinks, and much of its working tissue is replaced by fat, leading to a drop in newly minted T cells.
“As we get older, the immune system begins to decline. We wanted to think about how can we maintain this kind of immune protection for a longer period of time, and that’s what led us to think about what we can do to boost immunity,” said study author Mirco Friedrich.
For years scientists have tried to revive the organ. Hormones and immune-related proteins have struggled to bring it back to health. More exotic approaches, such as infusing the blood of young animals, transplanting stem cells, or directly tinkering with blood stem cells have shown some promise but are hard to turn into clinical treatments.
“Much has already been attempted to halt or reverse the age-related involution of the thymus,” said Friedrich. “Unfortunately, without much success so far.”
Rather than reviving the struggling organ, the team built a new T cell nursery in another part of the body.
They began by comprehensively mapping genetic changes in infant and elderly mice and deciphering how shifts in gene expression influenced T cell production.
The screen surfaced three genes that play a critical role in T cell maturation. The proteins those genes produce fall with age, correlating with lower T cell numbers. Refreshing the proteins could, in theory, reboot immune cell production.
This “is more of a synthetic approach,” said Zhang. “We’re engineering the body to mimic thymic factor secretion.”
They decided on the liver as a temporary nursery for several reasons. The organ faithfully synthesizes proteins even into old age, and it’s a relatively easy target for mRNA treatments.
The team packaged mRNA encoding the three nurturing proteins into fatty nanoparticles and injected them into mice’s blood twice weekly for a month, beginning when the mice were aged the rough equivalent of people in their 60s. While far from elderly, T cell defects are noticeable around this age, and the cells could benefit from early intervention.
Compared to untreated peers, those given the shots produced more, healthier T cells. The treatment also boosted the critters’ immunity. In one test, mice vaccinated against ovalbumin, a major protein in egg whites, had a far stronger immune response against the protein compared to peers without the mRNA treatment.
The shots also helped the mice’s laggy immune systems better coordinate with checkpoint inhibitors, a common cancer medication. Mice with cancer given both treatments survived longer and at higher rates than those given only the inhibitors. More tests found all three protein-encoding mRNA sequences were needed to rejuvenate the immune system.
To be clear, this isn’t a one-and-done shot. The effects wane after treatment ends. While it seems like an inconvenience, the flexibility allows scientists to further tinker with dosage and treatment schedule and minimize side effects. More broadly, the study shows restoring the thymus isn’t necessary for turning back the clock on the immune system. Mimicking its signals in other parts of the body could also help T cells thrive, even in old age.
These are early results, and more tests are needed before bringing the therapy to people. The team plans to study the mRNA trio in other animals and hunt down more proteins that nurture T cells. They’re also looking to expand the strategy to other immune cell types, like the B cells that pump out antibodies.
“The immune system ages, but it does not irreversibly lose its abilities. If we provide it with the missing signals again, it can once more perform amazing feats,” said Friedrich.
The post Aging Weakens Immunity. An mRNA Shot Turned Back the Clock in Mice. appeared first on SingularityHub.
2026-01-06 23:00:00
This year saw the meteoric rise of a promising new therapy for brain health.
Microglia are the silent guardians of the brain. They hunt down pathogens, clean up toxic protein clumps, and even shape the brain’s wiring. They’re also robust. Neurons can’t divide to generate new copies of themselves. But microglia can renew, especially during inflammation, stroke, or diseases that erode cognition.
And yet this regenerative ability has a limit, especially when the cells harbor genetic mutations. One solution? Replace diseased or injured cells with a fresh supply.
This year saw a meteoric rise in microglia replacement therapy, with clinical trials highlighting its brain-protecting potential. Refreshing microglia could, in theory, boost their beneficial effects.
Tinkering with the brain’s complex immune system isn’t straightforward, but “microglia replacement has emerged as a groundbreaking paradigm,” wrote Bo Peng and colleagues at Fudan University. The therapy could tackle a range of conditions from rare genetic diseases to more familiar foes such as Alzheimer’s.
Microglia are odd ducks. Like other immune cells that patrol the body, they usually start out as blood stem cells in bone marrow before migrating to the brain. Once settled, they stay at their post, exclusively protecting the brain.
The cells are usually shaped like shrubs in need of a haircut. But once activated, they shrink into puff balls and recruit other brain cells to fight off invaders and prevent brain damage.
Microglia also reconfigure the brain’s wiring. They prune extra synapses—connection points that allow neurons to talk to each other—and pump out nutritious molecules to support established neural networks and encourage baby neurons to grow.
It’s no wonder that when microglia go awry so does the brain. This happens in Alzheimer’s, other neurodegenerative diseases, and even just as we age. But more commonly, it’s because of genetic mutations in the cells.
Gene therapy is seemingly the best way to fix these problems. But microglia are notoriously terrible candidates. A gene therapy is usually shuttled into cells within safe viral carriers or tiny bubbles of fat. Few of these can enter the brain’s immune cells. Microglia-specific carriers exist, but they need to be injected directly into the brain. Complications from surgery aside, injected cells only reach a small area—hardly enough to make a notable difference.
Microglia replacement gets around this roadblock. Replacing mutated or aged cells with a healthy supply could correct genetic problems and “replenish populations lost to degeneration, inflammation, or developmental failure,” wrote Peng and colleagues.
Transplanting healthy donor microglia directly into the brain is nearly impossible because existing microglia often turn against the new arrivals. But because microglia start life as blood stem cells, a bone marrow transplant from a healthy, matching donor is a viable alternative. Once mature, the cells journey to the brain, where they divide and thrive.
The first and most taxing step of a bone marrow transplant is making space for the new cells. This requires extensive radiation or chemotherapy, but often without direct treatment to the head. The step also destroys the recipient’s immune system, leaving them vulnerable to infections and at higher risk for cancer.
Unfortunately, the standard treatment doesn’t work for microglia replacement, largely because diseased microglia still living in the brain leave little room for healthy new cells to settle.
But in 2020, Peng’s team developed a drug that depleted microglia in the brains of mice, making room for healthy cells. Then this July, Peng and colleagues successfully used a bone marrow transplant to treat a fatal brain disease called CAMP (CSF1R-associated microgliopathy). Here, mutations in a gene critical to microglia survival destroys the cells’ health, causing the brain’s wiring to physically disintegrate over time. Within a few years, people with the condition struggle with everyday reasoning, motor skills, and often fall into depression.
In mice and eight people in a small clinical trial with the disease, the treatment halted their decline for at least two years without notable side effects.
Researchers have also seen early success in other conditions.
Sandhoff disease is one that stands out. People with this inherited condition can’t break down certain fats, which leads to neuron death. The disease is partly caused by miscommunication between microglia and neurons. Normally, microglia shuttle an enzyme to neurons that helps recycle the fatty molecules. Mutated microglia can’t do this. In mice, bone marrow transplants of cells without the mutation improved the mice’s mobility, survival, and brain health.
Another study tackling Sandhoff disease used a different, more daring method. The team isolated the young cells that eventually become microglia and grew them in petri dishes.
After radiation therapy in mice, targeted to their heads, the team infused the healthy lab-grown microglia into the mice’s brains. The cells made themselves at home and worked as normal. The treatment avoided full-body radiation and damage to other organs but the approach could also kill off stem cells that generate new neurons in the brain and so may be limited in its efficacy.
Immune rejection also poses a major stumbling block. But induced pluripotent stem cells (iPSCs), where a person’s skin cells are reprogrammed into other cell types, may reduce the risk. In a proof of concept also in mice, microglia made from iPSCs replaced damaged microglia and slowed neurodegeneration by gobbling up toxic proteins related to Alzheimer’s.
Physicians will need to study the long-term consequences of head-only radiation, and test microglia replacement in a wider range of diseases. If all goes well though, the versatile cells could be used to even ferry medications into the brain like Trojan horses.
In just five years, microglia replacement has gone from animal studies to the first clinical treatment. Once a niche moonshot, it’s now “a topic of great interest in neuroscience and cell therapy,” wrote the team. While there’s plenty more work to do, the therapy could “mature from early breakthroughs into a generalizable platform across neurological diseases.”
The post Refreshing the Brain’s Immune Cells Could Treat a Host of Diseases appeared first on SingularityHub.
2026-01-05 23:00:00
US nuclear capacity is forecast to rise 63 percent in the coming decades thanks largely to data-center demand.
Nuclear energy has had a tough few decades, bedeviled by high costs and waning public support. But AI’s appetite for electricity could be a shot in the arm for the beleaguered industry.
AI’s energy demands are rising quickly, with global data center electricity use expected to double by the end of the decade. And nuclear power’s ability to provide large amounts of emission-free baseload power is hugely attractive for AI firms trying to balance their energy needs against climate commitments.
Google, Amazon, Meta, and major data center operators are signing power-purchase agreements with existing reactors, investing in the development of advanced small-modular reactors, and even helping restart shuttered nuclear plants.
This is a significant turnaround for a sector that has long been struggling to compete with cheap natural gas and rapidly falling renewable energy prices. But if the AI industry’s energy demands continue to grow as expected, the nuclear energy industry could be one of the big winners.
The most immediate impact of this trend could be to extend the lives of existing plants. In June, Meta inked a long-term contract with the utility Constellation Energy to keep its Clinton Clean Energy Center in Illinois operating for a further 20 years, after the plant faced closure due to the upcoming expiry of a credit program for low-emission energy producers.
Constellation says more deals could soon be coming. “We’re definitely having conversations with other clients, not just in Illinois, but really across the country, to step in and do what Meta has done, which is essentially give us a backstop so that we could make the investments needed to re-license these assets and keep them operating,” CEO Joe Dominguez told Reuters.
But demand for nuclear power is so acute that technology companies are also looking to bring already shuttered plants back online. Constellation closed a reactor at its Three Mile Island site in 2019 for economic reasons, but Microsoft has since stepped in to bring it back to life. Last September, the company agreed to a 20-year power purchase agreement to fuel its data centers, giving Constellation the certainty required to restart the reactor.
And Google appears to be following suit. In October, the company announced it was partnering with the utility NextEra Energy to bring back to life the Duane Arnold Energy Center, which closed in 2020. The company has committed to buying power from the facility for the next 25 years, and it could be back up and running by 2029.
But perhaps the biggest impact of Silicon Valley’s new love of nuclear could be a boom in investment in fresh nuclear capacity. Given how long it takes to build and commission nuclear plants, it may be a while before that impact is felt, but this could boost long-term confidence in the sector.
Last December, Meta announced it was seeking proposals from nuclear developers to help meet its energy demands. The company said that it was looking for 1 to 4 gigawatts of new capacity starting in the early 2030s, and that it was open to proposals to build either regular nuclear reactors or small modular reactors—an emerging class of advanced reactors that have yet to be commercialized.
These small reactors have caught the attention of technology giants due to their potential for lower costs and fast deployment. And they typically produce less than a third of the output of a regular nuclear reactor, which makes them suitable for powering smaller facilities. But their modular design means they can also be combined to create higher capacity plants.
Google has agreed to purchase power from Kairos Power, which is developing a fluoride-salt-cooled small modular reactors, becoming the first company to sign a commercial contract with the startup. The agreement covers six or seven reactors, with the first unit targeted for 2030 and the rest by 2035, supplying Google data centers with up to 500 megawatts of nuclear power.
In a similar vein, Amazon has agreed to buy electricity from four small modular reactor modules under development by X-Energy in Washington State, with the option to buy up to eight additional modules once they’re built. The data center operator Equinix has also placed a preorder for 20 transportable microreactors from California-based Radiant Nuclear.
A recent Bloomberg Intelligence report forecasts that US nuclear capacity could rise 63 percent by 2050 thanks in large part to demand from data centers. This would represent a net gain of 61 gigawatts in generation, most of which would come after 2035 when small modular reactors are expected to transition from demonstration projects to scalable commercial deployment.
Whether this comes to fruition will depend largely on whether big tech’s energy demands continue to balloon. There is mounting concern the industry is in an AI bubble primed to burst at any minute, which could put a major dampener on the nuclear resurgence.
But for the time being at least, the industry’s future is looking considerably rosier than it was a decade ago.
The post Your ChatGPT Habit Could Depend on Nuclear Power appeared first on SingularityHub.
2026-01-03 23:00:00
The passage of time is inextricably tied to how humans perceive our own experiences. We confuse our perspective on reality with reality itself.
“Time flies,” “time waits for no one,” “as time goes on”: The way we speak about time tends to strongly imply that the passage of time is some sort of real process that happens out there in the world. We inhabit the present moment and move through time, even as events come and go, fading into the past.
But go ahead and try to actually verbalize just what is meant by the flow or passage of time. A flow of what? Rivers flow because water is in motion. What does it mean to say that time flows?
Events are more like happenings than things, yet we talk as though they have ever-changing locations in the future, present, or past. But if some events are future, and moving toward you, and some past, moving away, then where are they? The future and past don’t seem to have any physical location.
Human beings have been thinking about time for as long as we have records of humans thinking about anything at all. The concept of time inescapably permeates every single thought you have about yourself and the world around you. That’s why, as a philosopher, philosophical and scientific developments in our understanding of time have always seemed especially important to me.
Ancient philosophers were very suspicious about the whole idea of time and change. Parmenides of Elea was a Greek philosopher of the sixth to fifth centuries BCE. Parmenides wondered, if the future is not yet and the past is not anymore, how could events pass from future to present to past?
He reasoned that, if the future is real, then it is real now; and, if what is real now is only what is present, the future is not real. So, if the future is not real, then the occurrence of any present event is a case of something inexplicably coming from nothing.
Parmenides wasn’t the only skeptic about time. Similar reasoning regarding contradictions inherent in the way we talk about time appears in Aristotle, in the ancient Hindu school known as the Advaita Vedanta, and in the work of Augustine of Hippo, also known as St. Augustine, just to name a few.
The early modern physicist Isaac Newton had presumed an unperceived yet real flow of time. To Newton, time is a dynamic physical phenomenon that exists in the background, a regular, ticking universe-clock in terms of which one can objectively describe all motions and accelerations.
Then, Albert Einstein came along.
In 1905 and 1915, Einstein proposed his special and general theories of relativity, respectively. These theories validated all those long-running suspicions about the very concept of time and change.
Relativity rejects Newton’s notion about time as a universal physical phenomenon.
By Einstein’s era, researchers had shown that the speed of light is a constant, regardless of the velocity of the source. To take this fact seriously, he argued, is to take all object velocities to be relative.
Nothing is ever really at rest or really in motion; it all depends on your “frame of reference.” A frame of reference determines the spatial and temporal coordinates a given observer will assign to objects and events, on the assumption that he or she is at rest relative to everything else.
Someone floating in space sees a spaceship going by to the right. But the universe itself is completely neutral on whether the observer is at rest and the ship is moving to the right, or if the ship is at rest with the observer moving to the left.
This notion affects our understanding of what clocks actually do. Because the speed of light is a constant, two observers moving relative to each other will assign different times to different events.
In a famous example, two equidistant lightning strikes occur simultaneously for an observer at a train station who can see both at once. An observer on the train, moving toward one lightning strike and away from the other, will assign different times to the strikes. This is because one observer is moving away from the light coming from one strike and toward the light coming from the other. The other observer is stationary relative to the lightning strikes, so the respective light from each reaches him at the same time. Neither is right or wrong.
How much time elapses between events, and what time something happens, depends on the observer’s frame of reference. Observers moving relative to each other will, at any given moment, disagree on what events are happening now; events that are happening now according to one observer’s reckoning at any given moment will lie in the future for another observer, and so on.
Under relativity, all times are equally real. Everything that has ever happened or ever will happen is happening now for a hypothetical observer. There are no events that are either merely potential or a mere memory. There is no single, absolute, universal present, and thus there is no flow of time as events supposedly “become” present.
Change just means that the situation is different at different times. At any moment, I remember certain things. At later moments, I remember more. That’s all there is to the passage of time. This doctrine, widely accepted today among both physicists and philosophers, is known as “eternalism.”
This brings us to a pivotal question: If there is no such thing as the passage of time, why does everyone seem to think that there is?
One common option has been to suggest that the passage of time is an “illusion”—exactly as Einstein famously described it at one point.
Calling the passage of time “illusory” misleadingly suggests that our belief in the passage of time is a result of misperception, as though it were some sort of optical illusion. But I think it’s more accurate to think of this belief as resulting from misconception.
As I propose in my book A Brief History of the Philosophy of Time, our sense of the passage of time is an example of psychological projection—a type of cognitive error that involves misconceiving the nature of your own experience.
The classic example is color. A red rose is not really red, per se. Rather, the rose reflects light at a certain wavelength, and a visual experience of this wavelength may give rise to a feeling of redness. My point is that the rose is neither really red nor does it convey the illusion of redness.
The red visual experience is just a matter of how we process objectively true facts about the rose. It’s not a mistake to identify a rose by its redness; the rose enthusiast isn’t making a deep claim about the nature of color itself.
Similarly, my research suggests that the passage of time is neither real nor an illusion: It’s a projection based on how people make sense of the world. I can’t really describe the world without the passage of time any more than I can describe my visual experience of the world without referencing the color of objects.
I can say that my GPS “thinks” I took a wrong turn without really committing myself to my GPS being a conscious, thinking being. My GPS has no mind, and thus no mental map of the world, yet I am not wrong in understanding its output as a valid representation of my location and my destination.
Similarly, even though physics leaves no room for the dynamic passage of time, time is effectively dynamic to me as far as my experience of the world is concerned.
The passage of time is inextricably bound up with how humans represent our own experiences. Our picture of the world is inseparable from the conditions under which we, as perceivers and thinkers, experience and understand the world. Any description of reality we come up with will unavoidably be infused with our perspective. The error lies in confusing our perspective on reality with reality itself.
This article is republished from The Conversation under a Creative Commons license. Read the original article.
The post Time Doesn’t Really Flow—Your Brain Just Makes You Think It Does appeared first on SingularityHub.
2026-01-02 23:00:00
The time to build safeguards is before something goes wrong, not after.
Generative AI is biology’s new playground. The technology powering popular chatbots can also dream up new, entirely novel versions of life’s most basic molecules, from DNA to proteins.
Once the domain of highly trained specialists, relative novices can now design synthetic molecules using open source AI software. But ease of access is a double-edged sword. While lower barriers to entry might spur creativity or even yield new medicines, the technology could also be used for nefarious purposes, such as designing novel toxins.
In 2024, two experts wrote an essay highlighting the need for biosecurity in the field. One of them, David Baker at the University of Washington, earned a Nobel Prize for RoseTTAFold, an AI that predicts protein structures from their amino acid building blocks. The other, Harvard’s George Church, has long been at the forefront of genetic engineering and synthetic biology.
They argued we should embed a barcode into each new designer protein’s genetic sequence to form an audit trail that scientists can trace back to the protein’s origins.
But a genetic tracer alone isn’t enough. A Microsoft study found AI-designed genetic sequences often escape the biosecurity screening software used by companies synthesizing designer DNA. AI-generated proteins with alien DNA sequences confuse these programs. Anything with genetic bits previously labeled “safe” flies under the radar, even if it encodes a dangerous final product.
These early studies are raising awareness. They’re not meant to stymie progress or enthusiasm—scientists welcome ideas for self-regulation. But for AI-powered designer biology to grow responsibly and be used for good, argue Church and other experts in a new preprint, the right time to build comprehensive biosecurity is before something goes wrong, not after.
From individual proteins to DNA, RNA, and even entire cells and tissues, AI is now learning the language of biology and designing new building blocks from scratch.
These powerful AI systems don’t simply recognize patterns. They eventually generalize those learnings across biology to analyze and dream up hordes of molecules at a prompt. RFdiffusion2 and PocketGen, for example, can design proteins at the atomic level with specific health-altering purposes, like sparking biological reactions or binding to drugs.
Generative AI is also beginning to read and write RNA. Like DNA, RNA is composed of four genetic letters, but RNA treatments don’t mess with the genetic blueprint. This makes them an exciting way to tackle disease. Unfortunately, they’re hard to design. RNA folds into intricate 3D shapes that are often difficult to predict using older software.
“Generative AI models are uniquely suited” for the job of capturing these intricacies, which could bolster the field of RNA therapeutics, wrote the team.
But the same AI galvanizing the field can also be used to create dangerous biological material. A person intent on jailbreaking an algorithm can, for example, repeatedly write prompts a generative AI system would normally refuse but is tricked into answering through repetition.
The dangers aren’t theoretical. A recent study compiled a dataset of toxic and disease-causing proteins and challenged multiple popular AI protein design models to create new variants. Many of the generated proteins retained their toxicity and evaded biosecurity software. In another case, scientists developed a method to test algorithmic security called SafeProtein. They managed to jailbreak advanced protein-design models 70 percent of the time.
Beyond proteins, researchers developing a framework called GeneBreaker found carefully tailored prompts can coax AI to spit out DNA or RNA sequences resembling viruses, such as HIV. Another team produced 16 viable genomes for bacteria that infect viruses, known as bacteriophages. Some of the resulting phages outcompeted their natural peers.
Even drug discovery tools can be flipped to the dark side. In one case, researchers easily reconfigured an AI model trained to find antiviral molecules. Within hours the AI suggested a known nerve toxin as a potential drug candidate.
“This demonstrates how even well-intentioned AI models can be rapidly misused to design toxins, especially when safety constraints are absent,” wrote the team.
To address these risks, the authors argue we need rigorous frameworks and regulations at every step of the process.
Scientists are leading the charge, and governments are on board. Last year, the UK released guidance for gene synthesis screening that urges providers of DNA and RNA molecules to vet their customers and increase screening for potentially dangerous sequences. The US launched similar rules and included biosecurity in its AI Action Plan.
Meanwhile, the tech giants behind AI models in biology are echoing calls for broader oversight. Some have pledged to exclude all viral sequences that are potentially dangerous to humans from their training databases. Others have committed to rigorous screening for new designs.
These safeguards, although welcome, are fragmented.
To gain a broader picture of the biosecurity landscape, the new study interviewed 130 experts across industry, government, academia, and policy. They agreed on several themes. Most think AI misuse is an urgent concern in biology and advocate for clearer regulatory standards. Roughly half were highly skeptical of current screening systems, and a majority supported upgrades.
The authors wrote that securing generative AI for biology isn’t about “finding a single solution.”
“Instead, it requires building a fortress with multiple layers of defense, each designed to anticipate, withstand, and adapt to threats.”
They designed a roadmap based on that principle. The strategy’s primary defenses target three stages in the AI life cycle. The first step is about controlling who can access training data and different AI versions. The next would add moral training that fine-tunes AI output. And finally, “live fire drills” to stress test models could reveal ways the AI could go sideways.
For example, algorithms trained on viral genomes are useful for drug or vaccine development. But they would be restricted. Users would have to apply for access and log usage. This is similar to how scientists must record the use of controlled narcotics in research. A tiered access system would allow others to use a version of the tool trained on data without dangerous content.
Meanwhile, strategies used to ensure chatbots (mostly) behave could also keep biology-focused AI in check. Moral training would guide a model’s output such that it aims to match public health and biosecurity standards. Stress testing to pinpoint a model’s vulnerabilities, known as red-teaming, would simulate misuse scenarios and inform countermeasures. Finally, biosecurity systems won’t work in a vacuum. Increasingly sophisticated AI could benefit from greater biological or general context, in turn improving its ability to detect and raise red flags.
“An effective biosafety system is not a firewall, it is a living guardian,” wrote the team.
Awareness is only the first part of the story. Action is the next. Although a unified vision of AI biosecurity doesn’t yet exist, the team calls on the field to collectively stitch one together.
The post AI Can Now Design Proteins and DNA. Scientists Warn We Need Biosecurity Rules Before It’s Too Late. appeared first on SingularityHub.
2026-01-01 23:00:00
Once only available for children under two, a one-and-done treatment is now approved for older kids too.
Waking up, hopping out of the bed, and stumbling to the kitchen for a cup of coffee: It’s an everyday routine most people don’t think twice about.
But for children with spinal muscular atrophy, simply propping themselves up in bed is an everyday struggle. The inherited disease is caused by mutations in the SMN1 gene. Without a working copy of the gene, motor neurons—cells that control muscles—rapidly wither.
Symptoms occur early in life. In the most severe cases, six-month-old babies can’t sit up without help. Others struggle to crawl or walk. The disease doesn’t affect learning and other cognitive abilities. Babies with the condition soak in their surroundings, and their brains develop normally. All the while, the disease cruelly destroys their bodies.
Left untreated, muscle weakness expands to the lungs, potentially causing deadly breathing problems. If there’s a silver lining, it’s that the disease has a clear genetic foe to target. Thanks to gene therapy, three treatments, approved by the FDA, can halt the disease in its tracks—if a patient is under two years old.
There’s a reason for the age limit. After two, the disease has already damaged motor neurons to such a degree that the therapy is no longer helpful.
Not so fast, two international teams of physicians and scientists wrote in December.
The teams published highly promising results from separate trials testing an experimental gene therapy, called Itvisma, in kids between 2 and 18 years of age. The new therapy is based on a previously approved version made by the drug company Novartis. Both have the same gene-correcting ingredient but are administered differently. The original relies on a shot into the bloodstream. Itvisma is delivered directly into the spinal cord.
The two recent trials brought significant improvement in participants’ ability to move over the course of a year. From not being able to walk, treated kids were able to roll into a sitting position from lying down and climb stairs, compared to children who did not receive treatment.
The results “demonstrate clinical benefits across a broad…population with a wide range of ages and baseline motor functions,” wrote Richard Finkel at St. Jude Children’s Research Hospital and team, on behalf of a broader STEER Study Group that conducted one of the trials.
The FDA agreed. In late November, the agency approved Itvisma for the disease, making it the only gene replacement therapy for people two years and older on the market.
“This achievement is not only a significant step forward for SMA [spinal muscular atrophy]–it also signals new possibilities for the broader field of neurological disorders and genetic medicine,” said John Day at the Stanford University School of Medicine in a Novartis press release.
Like its predecessor, Itvisma uses a harmless virus to carry a healthy version of the SMN1 gene into the body. The virus shuttles its cargo into cells but doesn’t tunnel into the genome. This makes it relatively safe, as it doesn’t raise the risk of unintended vandalism to the cell’s native DNA.
The previous therapy was a one-and-done shot into the bloodstream. The virus hitched a ride to motor neurons and restored their connection to muscle fibers. The liver and heart also received an unintentional dose, which could potentially cause side effects. Researchers carefully monitored children given the therapy for liver problems. These were relatively mild and easily treated.
The results were dramatic. Most treated infants were able to sit up, roll around in their cribs, and some could even crawl. But the treatment was only approved for children aged two years or younger.
Two problems hampered its broader use. One was timing: The disease rapidly eats away at motor neurons, causing long-term damage that’s difficult to restore. The other was safety. Gene therapies injected into blood are tailored to the recipient’s body weight—the higher the weight, the larger the required dose. Higher doses raise the risk of dangerous side effects, potentially causing the immune system to hyperactivate or cause damage to the liver.
For a toddler or teenager, the risk-benefit calculation didn’t work in the gene therapy’s favor.
Itvisma took an audaciously different approach by injecting the gene therapy directly into the fluid surrounding the spinal cord.
The procedure is much more invasive than a standard shot, but has a unique edge. Gene therapies delivered in this way don’t depend on body weight. Rather, their effectiveness can be carefully calibrated in a single off-the-shelf dose for anyone with the disease—toddlers, teenagers, or even adults. And because the therapy mostly circulates in liquids surrounding the spinal cord and brain, it rarely reaches other organs to cause unexpected mayhem.
Two clinical trials validated the daring new strategy.
One trial, STRENGTH, recruited 27 participants with the disease between the ages of 2 and nearly 18. The main goal was to test the treatment’s safety. The trial was single-armed, meaning that all participants received the gene therapy without a control group.
Overall, Itvisma was found to be safe. Some participants experienced cold-like symptoms, such as a runny nose and a sore throat. Others reported temporary headaches and stomach discomfort. A few suffered more severe problems, like a temporary spike in liver toxicity, fever, and motor neuron problems, which eventually went away.
Giving all participants a working treatment can lead to placebo effects. So, a second trial, STEER, followed the “gold standard” of clinical trials: double-blind, randomized, and placebo-controlled. The trial recruited 126 participants from 14 countries but separated them into two groups. One received the gene therapy; the other went through the same injection procedure but without the treatment. Neither the patients, their families, nor their doctors knew who got an active dose.
A year later, patients given the gene therapy could stand up from sitting on the couch, and some climbed stairs without support. Those who didn’t receive the treatment fared far worse. Once the trial was unblinded—in that both patients and doctors knew who received what treatment—the control participants also got a dose of the gene therapy.
Results from both studies prompted the FDA to approve Itvisma for people older than two.
The “approval shows the power of gene therapies and offers treatment to patients across the…disease spectrum” including various ages, symptoms, and motor function levels, said Vinay Prasad, the FDA’s chief medical and scientific officer in an announcement.
Itvisma is the latest in a burgeoning field of one-and-done gene therapies this year. From tackling a devastating genetic disease that torpedoes normal metabolism to broadening gene editors for rare inherited diseases and slashing cholesterol to protect heart health, gene therapy is finally tackling diseases once deemed unsolvable. The momentum is only building.
The post Kids With Spinal Muscular Atrophy Show Dramatic Improvement With FDA-Approved Gene Therapy appeared first on SingularityHub.
