Can Memories Be Inherited? How RNA Is Breaking the Barrier on Transgenerational Inheritance

In an episode of the Huberman Lab Essentials podcast, Stanford neurobiologist Andrew Huberman sat down with geneticist Oded Rechavi to explore a question that challenges conventional wisdom: can experiences acquired during a lifetime be passed down to the next generation? The answer turns out to be far more complex than the textbook "no."

Genes, RNA, and Proteins: Starting with an IKEA Manual

To understand the boundaries of inheritance, we first need to understand how genes work. Rechavi offered a clever analogy: DNA is like an IKEA assembly manual that every cell carries — a complete set of instructions for building every "piece of furniture" in the body. That's the genome.

Every cell in the body has the same manual, but different rooms need different furniture — the kitchen needs kitchen pieces, the bathroom needs bathroom pieces. So a cell pulls out just one page of instructions (corresponding to a specific gene), transcribes it into RNA, and uses that to "assemble" the protein — the finished piece of furniture.

One striking figure: the human genome is enormous, yet less than 2% of it encodes messenger RNA (mRNA). The vast majority is transcribed into other types of RNA that serve regulatory functions — many of which remain poorly understood. This sets the stage for the "new channels of inheritance" we'll explore later.

Two Seemingly Insurmountable Barriers

Why can eye color be inherited, but not knowledge of architecture? The key lies in the distinction between somatic cells and germline cells.

Only germline cells (sperm and eggs) can contribute to the next generation. Changes that occur in somatic cells — like the brain or muscles — theoretically cannot be transmitted to germline cells. The muscles you build at the gym are muscles your children will have to build themselves. That's intuitive common sense.

Rechavi identifies two core barriers that block the inheritance of acquired traits:

The first is the Weismann barrier, proposed by August Weismann in the 19th century: the separation of somatic cells from the germline. This has been called biology's "second law" (natural selection being the first) — a measure of how fundamental it is.

The second is epigenetic reprogramming. Cells accumulate various chemical modifications over a lifetime, but at the transition between generations, roughly 90% of those modifications are erased in sperm, eggs, and early embryos — allowing the new generation to develop from a "blank slate" following the original genetic instructions.

Using the IKEA analogy: it's like erasing all the pencil marks a previous owner made in the manual, restoring its full potential to build any piece of furniture, rather than being limited to the few items used in one particular room.

---

Epigenetics literally means "above the gene" — it refers to heritable chemical modifications that affect gene expression without changing the DNA sequence itself. The two most common types are DNA methylation (adding a methyl group to specific positions on DNA, typically suppressing gene expression) and histone modification (altering the protein spools around which DNA is wound, thereby controlling gene accessibility). These modifications can dynamically switch specific genes on or off in response to environmental signals — stress, diet, toxins, and more. Epigenetic reprogramming is the process by which, during fertilization and early embryonic development, specialized enzyme systems clear these modifications at scale, preventing a parent's "life annotations" from interfering with the new individual's full developmental potential. The roughly 10% of modifications that escape erasure represent the biological window through which acquired traits might cross generations — and form the core of this article's debate.

---

The Ghost of Lamarck and the Rise of RNA

The idea of inherited acquired characteristics traces back to Lamarck. The classic example is the giraffe: Lamarck believed giraffes stretched their necks to reach high leaves and passed those longer necks to their offspring, while Darwin's natural selection held that individuals born with longer necks happened to survive and reproduce, gradually replacing others.

Interestingly, even physicist Schrödinger, in his 1944 book, flatly stated that the inheritance of acquired characteristics doesn't hold — and lamented that this was discouraging, since it meant nothing you do in your lifetime can biologically influence the next generation.

But Rechavi emphasizes that beyond DNA, there exist mechanisms for transmitting information across generations via RNA. In recent years, even in mammalian research, RNA has increasingly emerged as a carrier of transgenerational information. This is the frontier of the field.

---

Lamarckism was long treated as a byword for pseudoscience in the 20th century, partly because Soviet-era Lysenko politicized it — leading to the persecution of thousands of Soviet scientists who followed Mendelian genetics and contributing to catastrophic agricultural failures. This history saddled "inheritance of acquired characteristics" with enormous stigma in the scientific community, such that any related research demands an exceptionally high evidentiary bar. Rechavi's pointed insistence in the interview that the evidence in C. elegans is "rock solid" is a deliberate distancing from that history. His work isn't a revival of Lamarck — it's the molecular-level unveiling of an entirely new RNA-mediated information transmission mechanism, one whose logic is completely different from the intuition of "use and disuse": what's transmitted isn't the trait itself (like the muscle you built), but a specific molecular state triggered by the information that "a particular threat was encountered."

---

C. elegans: The Model Organism Rewriting the Rules

Verifying these mechanisms requires model organisms. Rechavi chose Caenorhabditis elegans (C. elegans).

The advantages of this nematode are remarkable: researchers worldwide have numbered and named every one of its neurons, allowing different labs to study "the same neuron"; the worm's body is transparent, so neural firing can be observed directly; it was the first animal to have its genome sequenced, even before humans; each mother produces roughly 250 offspring with nearly identical genomes, and the generational cycle is just three days — enough to run hundreds of generations within a single PhD.

More critically, in C. elegans, the transgenerational inheritance of acquired traits has "very clear, rock-solid evidence" — so much so that almost no one in the epigenetics field disputes it anymore.

From RNA Interference to Transgenerational Viral Resistance

The scientific foundation of this story is the 2006 Nobel Prize awarded to Andrew Fire and Craig Mello for RNA interference (RNAi), based on their 1998 paper. They discovered that injecting double-stranded RNA into C. elegans silences genes with matching sequences, with small RNAs destroying the corresponding messenger RNA.

Even more striking: the silencing effect doesn't stay at the injection site — it spreads throughout the body, including germline cells, and is passed to offspring. Even simply feeding worms bacteria that produce double-stranded RNA causes the effect to spread from the gut through the whole body and into descendants. This has become routine procedure in C. elegans labs worldwide.

Rechavi's lab then ran an elegant experiment using fluorescent viruses: worms glow green if the virus successfully replicates, and stay dark if it's destroyed. They let worms become infected, clear the virus, and then knocked out the machinery for producing small RNAs in the offspring — meaning those offspring couldn't produce small RNAs themselves. Despite lacking the relevant genes, these offspring still stayed dark when exposed to the virus, successfully silencing it — and the effect persisted for multiple generations.

Through RNA sequencing, researchers confirmed that small RNAs matching the viral genome found in these offspring genuinely originated from infected parents. In other words, the parents wrote a "head start against a specific threat" into their offspring's heritable instructions.

---

RNA interference (RNAi) fundamentally changed molecular biology's understanding of gene regulation. The core logic: cells contain a defense and regulation system guided by small RNA molecules. When double-stranded RNA (dsRNA) enters a cell, an enzyme called Dicer cuts it into fragments roughly 21–23 base pairs long — small interfering RNAs (siRNAs). These fragments are loaded into the RISC (RNA-induced silencing complex), which uses complementary base pairing to target and degrade the corresponding messenger RNA, effectively "switching off" a specific gene with precision. This mechanism is evolutionarily ancient and likely originated as a defense against viruses — viral replication generates double-stranded RNA, which triggers the host's silencing response. Fire and Mello's Nobel-winning work revealed its remarkable power in C. elegans, and Rechavi's lab connected it to transgenerational inheritance — elevating RNAi from an intracellular regulatory tool to a potential carrier of immune memory across generations.

---

Can the Brain "Write Letters" to the Next Generation?

The most paradigm-shifting question is: can brain activity be transmitted to offspring? Rechavi repeatedly stresses that in mammals, "we don't yet know" — but in C. elegans, important discoveries have already been made.

The central challenge is a translation problem between two "languages": the brain stores information in the three-dimensional architecture of synaptic connections, while heritable information must pass through the single-cell bottleneck of the fertilized egg. How to translate synaptic structure into molecular form is the core challenge.

Rechavi's team's 2019 paper published in Cell offered a path that sidesteps this translation: simply altering the production of endogenous small RNAs in the worm's brain — without touching the offspring's brain — changed their descendants' ability to find food, with effects lasting three generations. The mechanism works by influencing the expression of a gene called saga-2 in the germline.

They demonstrated this is a genuine epigenetic effect because it spans multiple generations and depends on proteins that physically carry RNA between generations — remove that protein, and the effect disappears.

From Science Fiction to Diagnostic Possibilities

Looking ahead to applications, Rechavi outlined several directions. Rodent experiments have already shown that overfeeding causes metabolic problems in offspring — but having the animals exercise can correct this parentally transmitted effect.

A more distant vision involves diagnostic breakthroughs. In Israel, most couples planning a pregnancy undergo DNA-based genetic disease screening, but no one currently tests RNA. The beauty of DNA is its malleability — DNA is fixed, but RNA is variable. If its role in humans were one day confirmed, perhaps "spending some time on the treadmill" could alter one's heritable RNA profile, used during IVF to select healthier combinations.

Rechavi repeatedly and carefully caveats all of this: in humans, this is "currently completely unknown" and still belongs to science fiction. But that's precisely what makes it so fascinating — an almost entirely unexplored territory with enormous potential, waiting to be mapped.