Jordan Hale: Hey — so I'm going to start with a confession, which is that I thought I understood how cancer worked until about seventy-two hours ago.
Ryan Castillo: What changed seventy-two hours ago?
Jordan Hale: A paper dropped — October 2026, Nature Cell Biology, out of the University of Texas MD Anderson Cancer Center — and it describes something I genuinely did not have a mental model for. So let me try an analogy. Picture a factory. The blueprints tell every worker what to build. Most workers read the blueprints and build at normal speed. But there's this one enzyme — turns out it's called METTL13 — and what it does is modify the actual reading machinery inside the cell. Not the blueprint. The machinery that reads the blueprint. So certain products, the ones whose instructions happen to match the modified reader really well, get assembled way faster. Everything else? Same old pace.
Ryan Castillo: And the product that gets the speed boost is KRAS.
Jordan Hale: KRAS. One of the most commonly mutated oncogenes in human cancer — the protein that basically tells tumors to grow and survive. And Pawel Mazur, associate professor of Experimental Radiation Oncology at MD Anderson, co-led the team that found METTL13 is quietly doing this. No mutation in the KRAS gene. No extra copies of the gene. Just... the factory floor running a hidden fast lane for one product.
Ryan Castillo: So the gene is normal, the mRNA transcript is normal — it's only at the protein production step that something changes?
Jordan Hale: That's the click. Translation, not transcription. The copying stage is fine. It's the building stage where METTL13 has its thumb on the scale.
Ryan Castillo: That's a genuinely different layer than what most cancer research has been targeting.
Jordan Hale: And that's exactly what makes it a previously underappreciated layer — like, oncologists have spent decades hunting mutations, copy number changes, transcriptional upregulation. But this? The DNA is fine. The mRNA is fine. It's the ribosome-level reading where METTL13 has already tilted the game.
Ryan Castillo: Wait — so the DNA is just... clean? No mutation flagged?
Jordan Hale: Clean. That's the thing that broke my brain. Transcription — DNA to RNA — proceeds normally. It's the translation step, RNA to protein, where METTL13 intervenes. It modifies the machinery that reads the mRNA, and then only the mRNAs with the right codon sequences get the fast lane. KRAS mRNA happens to be codon-compatible with that modified machinery in a way most other mRNAs just... aren't.
Ryan Castillo: Hold on. So codon compatibility is the selector. The modified reader and the KRAS transcript are basically a lock and key.
Jordan Hale: That's the proposed mechanism, yeah. And I mean — it's elegant to the point where I wonder if we're maybe... I don't know, oversimplifying it? Like, is codon compatibility the whole story, or is it the tidy version of a messier reality?
Ryan Castillo: That's the load-bearing question for clinical development, actually. Because if the selectivity is real and robust — if METTL13 modification predictably fast-tracks KRAS and not your housekeeping proteins — the oncogenic protein production pathway you're disrupting is a viable target. If it's messier, you're flying blind on off-target effects.
Jordan Hale: Right — and the preclinical data from MD Anderson does show that blocking METTL13 slowed tumor growth and, crucially, improved sensitivity to existing KRAS-targeted drugs. That's not nothing. But you know, preclinical is preclinical — that gap between a cell model and a 65-year-old with metastatic cancer is basically the entire graveyard of oncology research.
Ryan Castillo: And that's precisely what makes the translation-versus-transcription distinction matter strategically — cancer amplified KRAS output without touching the gene, which means a whole regulatory layer was invisible to everything we were measuring. That's not a small miss.
Jordan Hale: And that regulatory layer being invisible is actually where the preclinical results get almost uncomfortable to look at. Because they didn't just show that blocking METTL13 slowed tumor growth — which, okay, you'd expect if KRAS production drops. The part that stopped me is that it also restored sensitivity to drugs like sotorasib and adagrasib. Drugs that already exist, already target KRAS, and already run into acquired resistance.
Ryan Castillo: So you're not replacing sotorasib. You're making sotorasib work again.
Jordan Hale: That's the combination story. METTL13 inhibition as an upstream master switch — not a standalone replacement. And I mean, think about what that means clinically. Like, imagine a patient, fifty-eight years old, lung cancer, initially responds to sotorasib. Tumors shrink, scans look better. And then, six months later, the resistance kicks in and the drug just... stops working. That's not rare — that's the expected trajectory for a lot of KRAS-mutant cancers right now.
Ryan Castillo: And the question the MD Anderson data raises is whether METTL13 inhibition could flip that resistance switch back off.
Jordan Hale: Exactly — and that's what makes it feel less like a discovery and more like a... you know, a key that was already missing from a lock we knew existed. We had the drug. We had the target. We just didn't understand what was quietly undermining it.
Ryan Castillo: The number that matters here is zero — as in zero human trial data. All of this is preclinical models. So how confident are we the combination effect actually transfers?
Jordan Hale: Not confident at all, and that's — wait, actually that's the thing I want to get into, because the gap between the preclinical result and a real patient is where this story gets genuinely complicated. And we haven't even touched on whether the selectivity mechanism holds up under scrutiny — which, spoiler, is not a settled question.
Ryan Castillo: Hold on — the codon compatibility explanation isn't closed?
Jordan Hale: That's what we need to dig into — because if the selectivity is messier than the MD Anderson paper in Nature Cell Biology makes it look, the whole therapeutic logic shifts under your feet.
Ryan Castillo: Right — but the part that doesn't fit is that codon compatibility being 'proposed but not fully resolved' isn't a footnote. If there are additional regulatory factors we haven't identified, METTL13 inhibition could hit proteins we're not tracking. Off-target effects in a patient with metastatic cancer aren't theoretical.
Jordan Hale: And that's — I mean, that's the honest version of this. The MD Anderson paper doesn't claim codon compatibility is the complete picture. It's the leading explanation. Which, you know, scientifically that's fine. That's how mechanism papers work. But therapeutically? That ambiguity is enormous.
Ryan Castillo: Therapeutic target identification is step one. Before drug development. Before Phase One trials. Before the FDA sees a single data packet. We're at step one.
Jordan Hale: Which is — no, I don't want to undersell that. Finding the target is genuinely hard. Pawel Mazur's team found a mechanism nobody had a model for. But I also don't want to oversell it to someone sitting in an oncologist's office right now.
Ryan Castillo: That's the exact tension. What does target identification actually mean for that person?
Jordan Hale: Think about a researcher — let's say she's a medicinal chemist, and she's just read this Nature Cell Biology paper at her desk at like eleven p.m. And she's excited, genuinely, because now she has a molecule to build against. METTL13. Concrete. Named. But between her excitement and a drug in a patient's bloodstream sits — I mean, literally years of preclinical validation, toxicology, Phase One dose-finding, Phase Two efficacy, Phase Three trials. The common cancer types modeled in these experiments aren't even specified beyond 'common.' We don't know if this holds across all KRAS variants or just G12C.
Ryan Castillo: The specific mutation question matters here. Because sotorasib targets G12C specifically. If METTL13 inhibition only restores sensitivity in G12C tumors, that's a narrow population — meaningful, but narrow.
Jordan Hale: And the paper doesn't answer that. October 2026 publication — zero human data, unresolved selectivity mechanism, unspecified cancer subtypes. Like, those three gaps together are not small.
Ryan Castillo: It's a map of a target. Not a drug. And the oncology graveyard is full of precise, elegant targets that just didn't survive contact with an actual human body.
Jordan Hale: You know what keeps landing for me, though? I started this thinking cancer was crude — it just screams and multiplies. And what MD Anderson found is that it's... not. It learned to hide inside the factory floor. The translation layer. The part nobody was watching as closely as the blueprints.
Ryan Castillo: What they found isn't just a target. It's proof the factory floor has its own set of locks. METTL13 is one of them. There are probably others.
Jordan Hale: And that's — I mean, that's the image I can't shake. One enzyme, one paper in Nature Cell Biology, and suddenly a whole hidden layer becomes visible for the first time. That's a different kind of discovery than finding another mutation.
Ryan Castillo: Good place to sit with that.
Jordan Hale: Yeah. Thanks for pushing on the gaps — genuinely made it sharper.
