Scientists identified a self-perpetuating loop in Alzheimer's disease: tau damages mitochondria, mitochondria produce signals that make tau worse, tau damages mitochondria more. A Stanford-led study published in Neuron reveals the exact mechanism driving this cycle and how to break it. Tau is a protein that's been heavily implicated in Alzheimer's disease and other neurodegenerative disorders. The standard view has been that tau forms tangled clumps called neurofibrillary tangles inside nerve cells, disrupting their structure. But mitochondrial dysfunction consistently appears alongside tau abnormalities across all tauopathies. The connection between these two hallmarks has remained unclear. Mitochondria are your cells' energy factories. They generate ATP, the molecule that powers nearly every cellular process, by shuttling electrons through a series of protein complexes called the electron transport chain. Electrons normally flow forward through these complexes, from Complex I to Complex IV, producing energy efficiently. Reverse electron transport, or RET, is when electrons flow backward through Complex I. Instead of generating useful energy, this reverse flow produces excess reactive oxygen species, or ROS. These are damaging molecules that oxidize proteins, lipids, and DNA. RET also depletes NAD+, a molecule critical for hundreds of metabolic reactions. The new study demonstrates that tau directly regulates RET. When tau gets phosphorylated, meaning enzymes attach phosphate groups to specific sites on the protein, it enters mitochondria and binds to a Complex I subunit called NDUFS3. This interaction activates reverse electron transport. Here's what the researchers found: • Tau enters mitochondria and physically interacts with the NDUFS3 subunit of Complex I • Phosphorylated tau enhances RET activation in a dose-dependent manner • RET produces excess ROS and reduces the NAD+/NADH ratio • The ROS and metabolic changes caused by RET drive further tau phosphorylation • This creates a feed-forward loop where tau-induced RET makes tau more pathological • Removing tau eliminates stress-induced RET entirely The team tested this mechanism across three different model systems: fruit flies, mice, and human neurons derived from induced pluripotent stem cells. In all three systems, depleting tau prevented stress-induced RET and conferred significant protection against cellular stress. When they blocked tau from entering mitochondria or disrupted the tau-NDUFS3 interaction, RET decreased. When they inhibited RET either genetically or pharmacologically, they protected against tau-induced neurodegeneration across all species tested. The phosphorylation sites that enhanced RET activity corresponded to the same modifications found in human Alzheimer's disease brain tissue. This suggests the mechanism operates in actual disease conditions, not just experimental models. What makes this discovery particularly important is that it reveals a normal function of tau that becomes pathological. Tau appears to regulate RET under physiological conditions, possibly as part of stress response signaling. During aging or chronic stress, this regulatory function spirals out of control. The vicious cycle works like this: stress or aging triggers tau phosphorylation. Phosphorylated tau enters mitochondria and activates RET. RET produces oxidative stress and metabolic dysfunction. These changes drive more tau phosphorylation. More phosphorylated tau activates more RET. The loop reinforces itself. This explains why mitochondrial dysfunction appears so consistently in tauopathies. It's not a separate problem running parallel to tau abnormalities. The two processes are mechanistically linked through RET, with each driving the other. The therapeutic implications are direct. Blocking RET, preventing tau entry into mitochondria, or disrupting the tau-NDUFS3 interaction could all interrupt the pathological cycle. Several compounds that inhibit RET already exist, including metformin and rotenone at specific doses, though their therapeutic windows need careful definition. The study also reframes how we think about neurofibrillary tangles. While tangles remain a pathological hallmark, the mitochondrial dysfunction driven by tau-RET interaction may cause neurodegeneration before tangles even form. Early intervention targeting this mechanism could prevent damage at stages where neurons remain salvageable. Previous work showed that reducing tau levels protects against various neurological insults, from seizures to traumatic brain injury. This study provides a molecular explanation: without tau, cells can't activate pathological RET, maintaining mitochondrial function under stress. The research identifies RET as a convergence point where tau's physiological function, its pathological modifications, and mitochondrial dysfunction all intersect. Understanding this intersection opens intervention strategies that target the mechanism driving neurodegeneration rather than just clearing protein aggregates after damage occurs.

Bender, Scannell, Shaywitz and colleagues in NRDD: a decade of AI in drug discovery, clinically relevant impact still limited. Their fix is right. Stop benchmarking model accuracy; benchmark whether the tool changes a real Go/No-Go. We proved translation is possible with Rentosertib. nature.com/articles/s41573-0…

Introducing PRIME, Cultivarium’s mitochondrial engineering program, supported by ARIA. We’re developing tools for nucleic acid delivery, gene expression, and genome maintenance in mammalian mitochondria. The non-model biology within us: blog.cultivarium.org/p/the-n…

The article from Daphne is valuable; this plot is truly absurd. - each "cell" from scRNAseq is 50K reads which is not 3GB, off by many magnitudes. - we are a preseed company and work with 1PB of data- no causal map of biology is occurring LOL - The amount of public sequencing data on NCBI well exceeds 100PB. still no causal map of biology. - the idea that "more data" will solve biology is an absurd mental framework. it just makes AI+bio folks sound like a joke.

We're making psychedelics*! We design nature-inspired therapeutics and work to get them functioning inside human biology. With huge challenges ahead, and particularly in this automation + LLM era, we need human brains that are creative, connected, and on fire. This is us putting that into practice. *said our Ebru (paper marbling) instructor.

Spotted in the wild: Powerhouse Biology's @rachel8407 hits our very first #JPM running at @cglifeagency's Sweat Equity 5K. We are building a new therapeutic platform centered on mitochondrial control, arguably one of the most underexploited levers in age-related disease. Find us tonight at @statnews @ #JPM26 to learn how we’re opening a new frontier in drug development.

What happens when you gather hundreds of builders and researchers at the frontiers of science and tech for three days of bouncing ideas, parties, and lab tours? This weekend, we hosted our flagship event Vision Weekend! With a strong track record of matching scientists for collaborations, entrepreneurs with their first investors, and startups with early employees, we love seeing how these gatherings help turn big ideas into reality faster. Thank you for building a better future with science and tech! Recordings of the talks will be available on our YouTube channel in a few weeks. And a special thanks to our partners: @AsteraInstitute @netholabs @protocollabs @ARIA_research @RetroBio_ and Poseidon Research

Towards genome editing for mitochondrial diseases nature.com/articles/d41586-0…

We have been awarded a prize in the Catalyst Phase of the Healthy Longevity Global Competition of @theNAMedicine! Particularly proud because it’s a meaningful science-first signal for our bold new endeavor to enhance cellular resilience and durability. NAM selected 15 awardees from 662 U.S. submissions. Announcement: buff.ly/70WSH9i

It's a true pleasure to work with Christina Jensen, an exceptional scientist and person. Christina joined Powerhouse Biology at the start of September - our first Experimental Biologist actively pushing forward what was a wildly ambitious mission in a pitch deck in February. Christina brings pedigree from stellar labs: deep expertise in hiPSC model development, plus broad experience applying frontier technologies in molecular biology to design and validate synthetic systems with real biological impact. And yes, her former lab’s tongue-in-cheek motto “anything can be high-throughput if you work hard enough” is currently making the rounds at MBC BioLabs in San Carlos.

On Powerhouses and Foodstuff: in this delightful primer in Life Sciences Insights, a @CALifeSciences magazine, co-founder and CSO @rachel8407 explains why we do what we do at Powerhouse Biology. Have a read on how we use high content data to resolve multidimensional, mitochondrial behaviors that drive age-related diseases. "Enhance, enhance!" ca-ls.org/lsi-ai-powerhouse

First day on site! @arubikscube joined Powerhouse Biology as a Computational Scientist. He’s our trailblazer on multi-modal network models and ML analysis, and he’s set to grow and shape our computational team. We were instant fans: his creative spark and bold move (yep, he packed up and left Boston for the Bay!) are pretty telling. Welcome to sunny California, Aru - super excited to have you!

Few things solidify a reputation better than a sticky nickname. 68 years ago this month, Philip Siekevitz called the mitochondrion “a small body which appears to play a central role in the oxidation of foodstuff,” and popularized its now iconic title: the powerhouse of the cell. - Vol. 197 No. 1 (July 1957), p. 131 @sciam scientificamerican.com/artic…

Life is beautiful. We want to give people more of it.

We can't wrap our head around how she does it all - our champion CSO @rachel8407 just published "The AI drug revolution needs a revolution" in npj Drug Discovery. Rachel co-founded Powerhouse Biology with a vision to implement AI-guided multi-modal analysis of mitochondrial biology for therapies by grounding algorithms in physiologically relevant, real-world data that capture true human variability. Curious how that mindset rewires drug development? Read her Comment here: nature.com/articles/s44386-0…

First data in!! We've transferred into the brand-new office and laboratory space at MBC BioLabs in San Carlos! Huge thanks to the MBC team for moving mountains to get us (and many others) to work, and to @LeicaMicro for helping us hit the ground running with core equipment and support. It’s awesome to share space with fellow founders and tech optimists - they make the good days exponentially better, and on the tough days, they help solve the seemingly impossible. (The sun terrace and well-stocked fridge don't hurt either.) Curious how we’re rebooting the human powerplant? Check out powerhouse.bio, or hop in for coffee!

Thanks to everyone who joined our AI x Bio workshop this weekend! 180 scientists, entrepreneurs, engineers, and funders gathered to explore how AI can accelerate longevity research. The workshop included presentations on the current state of the field and its biggest bottlenecks, with insights from experts like @bettslacroix @DrMorganLevine @SGRodriques @jrkelly, as well as working groups to develop promising projects at the intersection between AI and biology.

Happy #NewMemberMonday! 👩‍🔬 @Powerhouse_Bio makes precision peptide therapeutics to combat mitochondrial dysfunction 💡 @serna_bio leverages AI to discover RNA-targeting small molecules 👨‍🔬 @TelixPharma drives radiopharmaceutical innovation from its base in Australia

Technically challenging and historically siloed from broader genomics and therapeutic innovation, mitochondrial biology is now at the center of converging tools, data, and demand. We are incredibly excited to see @virushunter and @ARIA_research opening up an opportunity space in bioenergetic engineering. Check it out: aria.org.uk/opportunity-spac… There's so much important work ahead: designing, synthesizing, and assembling human mitochondrial DNA; reconfiguring coding sequences currently encoded in the mitochondrial genome for expression in the nucleus; and ultimately extending healthy human lifespan by enhancing mitochondrial biology with superior, mitochondria-centric traits. Let's build!

Here we go!! We’re hiring a Computational Scientist and an Experimental Biologist to join our fast-paced platform therapeutics startup! We need scientists and engineers who own their work, adapt quickly without losing precision, and stay laser-focused on what truly matters. We are looking for exceptionally skilled pioneers with a translational mindset - people who don’t just explore problems but drive solutions. We do things thoroughly, well, and at speed, because that’s what it takes to make an impact. If you thrive in a high-impact, no-nonsense environment* where we innovate together to drive results, we want you on our team! *much like the mitochondria in a cell! Learn more and apply here: tally.so/r/w5M6Ob