🎮 French Scientists Turn Mouse Brains Into Nintendo Controllers (Dementia Optional)

Remember when the most advanced brain technology was hitting yourself on the head to fix a stuck thought? Well, science just invented the biological equivalent of a TV remote, except instead of controlling your Netflix addiction, it controls cellular powerhouses and reverses memory loss. This week, we're diving into mouse brain remote controls that cure dementia, an antidote for carbon monoxide that only took humanity several millennia to figure out, AI that can actually design antibiotics (shocking, we know), and DNA synthesis technology that finally graduated from the Miami Vice era.

Science is getting so good at impossible things that we're starting to suspect the universe is just showing off at this point. Pour yourself something caffeinated and prepare to question whether we're living in a simulation designed by overachieving grad students.

Remember when your TV remote died and you had to actually get up to change the channel? Well, French scientists just invented the biological equivalent, except instead of controlling your Netflix binge, it controls brain cell powerhouses and reverses memory loss. Published in Nature Neuroscience, this is the first time anyone has proven that dead mitochondria actually cause dementia, not just hang around the crime scene looking suspicious.

The team at Bordeaux created something called mitoDREADD-Gs, which sounds like a rejected Star Wars droid but is actually a genetic remote control for mitochondria. Inject a designer drug called CNO, and boom, your brain batteries start working like they just chugged a Red Bull. They tested it on stoned mice (yes, THC-induced memory problems), mice with tau tangles, and mice with Alzheimer's plaques. Every single time, the memory problems vanished faster than your motivation on a Monday morning.

Here's the clever bit: instead of trying to clean up all the toxic protein gunk in the brain like every failed Alzheimer's drug for the past 30 years, they went straight to the power source. It's like realizing your computer isn't broken, it just needs to be plugged in. The mitochondria get a molecular espresso shot that makes them produce ATP like they're 20 years younger.

Of course, before you start planning grandma's brain remote control surgery, there are some minor issues. The activating drug converts to clozapine, an antipsychotic with side effects that make the cure potentially worse than the disease. Plus, mouse studies translate to humans about as reliably as Google Translate handles poetry. But hey, at least we finally know what's actually breaking in broken brains.

Read more from here

After millennia of carbon monoxide poisoning being basically a death sentence and sixty years of oxygen masks being our best shot, scientists at the University of Maryland said "screw it" and stole superpowers from dirt bacteria. Published in PNAS, they created RcoM-HBD-CCC (they really need a marketing department), which sucks up CO faster than a Roomba on cocaine.

The protein binds carbon monoxide 50 times better than your own hemoglobin and clears it from the blood in under one minute. For context, current oxygen therapy takes hours and often leaves you with permanent brain damage as a parting gift. This new antidote basically kidnaps the CO molecules and escorts them out through your pee before they can kill you.

The origin story makes sense. Soil bacteria naturally make proteins to detect trace CO for their metabolism. Scientists looked at this and thought, "What if we turned that sensor into a vacuum cleaner?" Dr. Jason Rose and Dr. Mark Gladwin have been working on this since 2016 through their company Globin Solutions (finally, a biotech name that makes sense).

Carbon monoxide kills 1,500 Americans annually and sends 50,000-100,000 to the ER, making it the most common poisoning worldwide. Current treatment involves sitting in a hyperbaric oxygen chamber for hours, hoping your brain cells don't die while waiting. The new antidote could be given by paramedics right at the scene, turning a potentially fatal poisoning into a really expensive ambulance ride.

The catch? It only works in mice so far, and we all know how that usually goes. Plus, they need to figure out dosing, manufacturing, and whether humans will have some exciting new side effect like glowing in the dark. But after thousands of years of fire-related deaths, we'll take "promising mouse studies" over "definitely dead."

In a field where "AI discovers new drug!" usually means "computer suggests molecule that can't exist in our dimension," MIT researchers actually delivered. Published in Cell, they used generative AI to design completely new antibiotics that kill drug-resistant bacteria. Not tweaked versions of old drugs, not AI-optimized known compounds, but brand new molecular structures that no human would ever think to create.

James Collins' team generated 36 million potential compounds (because go big or go home), identified two that might work, and both actually killed superbugs in mice. One targets gonorrhea by attacking a protein that no current antibiotic touches. The other cleared MRSA skin infections in four hours, working as well as our best drugs, but through a completely different mechanism.

Here's why this matters: antibiotic resistance killed 1.27 million people in 2019 and is projected to kill 10 million by 2050. We've discovered exactly two new antibiotic classes in 50 years, while bacteria have been speedrunning evolution like it's their job. The AI can imagine compounds no chemist would dream up, processing millions of possibilities while you read this sentence.

The reality check is pretty brutal, though. From 80 AI-designed gonorrhea drugs, only 2 could actually be made in real life, and only 1 worked. It's like having a friend who gives amazing restaurant recommendations for places that exist only in parallel universes. Ukraine-based Enamine has become the go-to manufacturer for turning these AI hallucinations into actual molecules, apparently being the only company that speaks both "computer" and "chemistry."

Despite billions poured into AI drug discovery since 2010, zero AI-designed drugs have reached patients. Collins admits it'll be "several years" before these antibiotics help humans, assuming the mouse results translate, which historically happens about as often as your New Year's resolutions survive past January.

While we can read entire genomes for the price of a nice dinner, writing DNA has been stuck using technology from when Miami Vice was cool. Ansa Biotechnologies just changed that with engineered enzymes that can synthesize DNA fragments over 750 base pairs, crushing the old limit of 200-300.

Their trick? Attach an enzyme to each DNA letter, let the enzyme add it to the growing chain, then chemically remove the enzyme so the next letter can join. It's like building with Legos, where each block comes with its own tiny construction worker. No harsh chemicals, no sequence restrictions, and they can make DNA that other companies literally can't.

With $80+ million in funding, Ansa offers 400-750 bp DNA delivered in 6 business days with an "On-Time Guarantee" where your entire order is free if any part is late. That's more confidence than a pizza delivery promise, except instead of lukewarm pepperoni, you get custom genetic sequences.

But wait, there's more. Michigan Tech's lab, led by Shiyue Fang, just made phosphoramidite synthesis look like a typewriter in the smartphone era. Published in Chemical Science, they synthesized a 1,728-nucleotide DNA polymerase gene directly on smooth glass surfaces. Not assembled from pieces, not stitched together, but chemically synthesized as one continuous molecule. They also knocked out an 800-mer green fluorescent protein gene for good measure.

The secret sauce? Ditching the porous glass beads everyone's been using since the Reagan administration. Smooth surfaces eliminate the steric hindrance that blocks long DNA assembly, and somehow this also "drastically reduced synthesis errors". They use something called "catching-by-polymerization" to fish out the perfect full-length sequences from the synthesis soup. It can even handle nightmare sequences like long repeats, bizarre secondary structures, and site-specific modifications that make regular synthesis methods cry.

Why should you care? We're living in a bizarre world where reading DNA costs less than a fancy coffee but writing a single gene costs more than a car. If enzymatic synthesis scales up, we could finally close this "read-write gap" and make designing biology as easy as typing code. Microsoft's already storing data in DNA because apparently regular hard drives aren't weird enough for them.

We’re reminded again that reality keeps outpacing our ability to process it. We've got brain remote controls fixing dementia in mice, CO antidotes stolen from dirt bacteria, AI designing actual working antibiotics, and DNA synthesis that finally entered the 21st century. At this point, science fiction writers should just give up and become documentarians.

Did the brain remote control make you rethink your relationship with your mitochondria? Ready to carry around a carbon monoxide antidote just in case? Planning to ask AI to design your next prescription? Hit reply and share your thoughts – we read every email between checking if our brain batteries need a molecular espresso shot!

Share this with someone who thinks medical breakthroughs happen slowly. We're still cruising with way over 70% open rates, proving that you're as addicted to impossible science as we are!

Stay curious, stay skeptical, and maybe invest in a good carbon monoxide detector,
Prateek & Jere

P.S. If mouse studies translated to humans as reliably as we pretend they do, we'd all have been cured of everything by now. But hey, at least the mice are living their best cyborg lives! 🧠💊✨

And if you’re still not convinced, no worries! Just unsubscribe here. You can always check back on us later.