How a Childhood Brain Tumor Hijacks Lactate to Build DNA (And How We Can Watch It Happen)

Why This Matters

Diffuse midline gliomas (DMGs) are among the cruelest cancers we know. They strike children, grow in midline brain structures where surgical resection is typically not feasible, and carry a devastatingly poor prognosis.[1] The H3K27M mutation has been the recognized oncogenic driver, but we've struggled to translate that knowledge into treatments that actually work. Meanwhile, researchers have been wrestling with a fundamental question: why do these tumors gorge on glucose and pump out lactate at such extreme rates? Is lactate just metabolic exhaust—cellular garbage—or is it doing something?

This paper from Batsios and colleagues answers that question definitively: lactate is fuel. More precisely, it's a signaling molecule that directly activates the machinery for building DNA. And because this process is so hyperactive in H3K27M-mutant DMGs, it creates an opportunity to watch these tumors in real-time using a clever imaging trick involving heavy hydrogen.[1]

This matters because DMG research has been stuck. We've known the mutation. We've known it's deadly. But knowing and treating are different games entirely. This work connects mechanism to medicine—and gives us a way to see if our treatments are working within days, not months.

Following Carbon Atoms Through Tumor Cells

The team started with patient-derived DMG cells—the gold standard for studying these tumors because they maintain the genetic and metabolic features of actual patient disease. They fed these cells glucose labeled with carbon-13 (a heavy, trackable version of carbon) and watched where those atoms ended up.[1]

The result: lactate production was dramatically elevated. But here's the key—this wasn't just generic cancer metabolism. The H3K27M mutation itself was driving this metabolic rewiring, as demonstrated through stable isotope tracing and loss-of-function studies.[1]

Think of it this way: cancer cells are often described as "rewiring" their metabolism. But that makes it sound random, like someone cut wires and reconnected them haphazardly. This is more precise. The H3K27M mutation flips specific genetic switches, turning up production of particular enzymes, creating a coordinated metabolic program.

From Mutation to Metabolic Reprogramming

How does a histone mutation—a change in the proteins that package DNA—crank up lactate production? The researchers traced the path:[1]

The H3K27M mutation alters the epigenetic landscape of the cell. Think of epigenetics as the system that controls which genes are easy or hard to read—like bold versus faint text on a page. One gene that gets epigenetically upregulated: PGK1, phosphoglycerate kinase 1.[1]

PGK1 is described as a rate-limiting glycolytic enzyme.[1] That term—rate-limiting—has specific meaning in metabolism. It's not just any enzyme in the pathway. It's the bottleneck. The step that controls how fast the whole process runs. More PGK1 means the entire glucose-to-lactate assembly line speeds up.

But the real surprise came when they looked at what lactate does next.

Lactylation: When Lactate Becomes a Signal

Lactate doesn't just sit around. It can attach to proteins through a modification called lactylation—basically, lactate molecules get stuck onto lysine residues on proteins, changing how those proteins behave.[2]

The team discovered that lactate lactylates a protein called NME1 (nucleoside diphosphate kinase). NME1's job is to convert nucleoside diphosphates (NDPs) into nucleoside triphosphates (NTPs)—the building blocks needed to make new DNA. When NME1 gets lactylated, it becomes more active.[1]

This is your cell's assembly line for replication. DNA is made of nucleotides. Those nucleotides need to be in their triphosphate form (ATP, GTP, CTP, TTP) to be incorporated into growing DNA strands. NME1 is the enzyme that phosphorylates nucleoside diphosphates into triphosphates. Lactate makes it work faster.

Here's the chain reaction the paper identifies as the H3K27M-lactate-NME1 axis:[1]

H3K27M mutation → epigenetic upregulation of PGK1 → more lactate → lactylated NME1 → more NTPs → faster DNA synthesis → tumor proliferation.

Lactate facilitates the synthesis of nucleoside triphosphates, which are essential for DNA replication and tumor proliferation.[1] It's not just a byproduct. It's a signal that tells the tumor: build more DNA, divide faster.

Proving It Matters

To confirm this wasn't just correlation, the researchers did loss-of-function studies.[1] When they disrupted this pathway, tumor proliferation slowed. The mechanism was validated using clinically relevant patient-derived DMG models—actual cells from patients, carrying the real mutations, behaving like real tumors.[1]

This is important because cancer biology is littered with findings that work beautifully in artificial cell lines and then collapse when tested in more realistic systems. Patient-derived models aren't perfect (we'll get to limitations), but they're far closer to the real thing.

The Imaging Innovation: Watching Tumors Grow in Real-Time

Understanding the biology is one thing. Being able to see it in living patients is another.

The team realized that because H3K27M-mutant DMGs produce so much lactate, they could use this as an imaging biomarker. But instead of carbon-13 (which requires specialized equipment and doesn't show up well on clinical scanners), they used deuterium—heavy hydrogen.[1]

They gave mice with DMG tumors glucose labeled with deuterium at specific positions: [6,6-²H]-glucose.[1] As the tumor cells metabolized this glucose, they produced deuterium-labeled lactate. Using deuterium metabolic imaging (DMI) at clinical field strength (3 Tesla)—the same field strength as standard clinical MRI machines—they could create spatial maps of lactate production.[1]

The imaging allowed visualization of the metabolically active tumor lesion.[1] This isn't just detecting that there's a tumor. It's spatially mapping where the tumor is metabolically active—where it's actually making lactate, which means where it's actually proliferating.

Testing Therapies: An Early Warning System

The real power of this imaging approach emerged when they tested it during treatment. The team treated DMG-bearing mice with standard of care and targeted therapy.[1]

The deuterium-lactate signal provided an early readout of treatment response that preceded extended survival.[1] This is crucial: the imaging changed before the mice lived longer. It was predictive, not just correlative. The imaging also reflected pharmacodynamic alterations in tumor tissues—actual drug effects on tumor biology, not just tumor size.[1]

This matters enormously for clinical trials. Right now, we wait months to see if a treatment works, using tumor size as the endpoint. Tumor size is a lagging indicator. It tells you what happened weeks ago. With metabolic imaging, we could potentially know within days whether a therapy is hitting its target. For children with rapidly progressive tumors, that speed could be lifesaving.

What Makes This Different

Metabolic imaging isn't new—PET scans have been tracking glucose uptake in tumors for decades. But this approach has several advantages:

Specificity: By tracking the conversion of glucose to lactate (not just glucose uptake), this method specifically interrogates the metabolic pathway that's hyperactive in H3K27M-mutant tumors. It's not looking at all metabolic activity. It's looking at the specific metabolic vulnerability created by this specific mutation.

Clinical feasibility: Deuterium imaging can be performed at clinical field strength (3T), without requiring radioactive tracers or PET scanners.[1] This is a big deal. PET requires a cyclotron nearby to make the radioactive glucose. Deuterium MRI requires specialized RF coils and pulse sequences, but the base infrastructure is already in most hospitals.

Mechanistic grounding: Unlike imaging approaches developed empirically ("let's try this and see if it works"), this one is built on a deep understanding of how H3K27M mutations drive metabolism. We're not just seeing that something changes—we understand why.

Limitations and Caveats

This is preclinical work in mouse models.[1] All the imaging was done in mice, not humans. The approach needs validation in human patients before we know if it works as well in children with DMGs.

The studies used patient-derived cells, which is excellent, but they were still grown in mice, not humans.[1] The tumor microenvironment matters. Mouse brains aren't perfect replicas of pediatric brains. Immune responses differ. Blood-brain barrier properties differ.

The mechanistic work focused on the H3K27M-lactate-NME1 axis, but lactate likely has other targets in these cells. This is one important pathway, not necessarily the only one. Lactylation is a relatively new post-translational modification—we're still discovering all the proteins it affects and all the things it does.[2]

The imaging can detect metabolically active tumor, but distinguishing tumor from inflammation or treatment-related changes will require careful validation. Lactate production can increase in inflamed tissue too.

Finally, these findings are specific to H3K27M-mutant gliomas. DMGs without this mutation may have different metabolic profiles and may not be as amenable to this imaging approach.

The Path Forward

This paper does something rare: it connects mechanism to medicine. By understanding how lactate drives DMG proliferation through nucleotide biosynthesis, the researchers created a tool to measure that proliferation non-invasively.

For DMG research, this opens several doors:

Better clinical trials: Metabolic imaging could serve as an early endpoint, allowing faster assessment of whether experimental therapies are working. Instead of waiting months to see if tumors shrink, we could potentially see within a week whether the drug is hitting its metabolic target.

Therapeutic targets: The H3K27M-lactate-NME1 axis itself might be druggable. Blocking lactylation or targeting downstream effects could slow tumor growth. If lactate is feeding nucleotide biosynthesis, starving that pathway could starve the tumor.

Patient stratification: If lactate metabolism varies between patients, imaging could identify which tumors are most dependent on this pathway and most likely to respond to metabolism-targeted therapies. Not all H3K27M tumors are identical. Imaging could help us personalize treatment.

The H3K27M mutation has taunted researchers for years—a clear driver mutation in a devastating disease, but one that's proven difficult to target directly. You can't easily drug a histone mutation. But this work suggests an alternative strategy: don't target the mutation itself, target the metabolic vulnerabilities it creates. And now we can watch whether it's working, in real-time, in living subjects.

For the families facing DMG diagnoses, that's not just scientific progress. It's hope with a mechanism attached. It's the difference between "we're trying things" and "we understand what's broken and we can see if we're fixing it."

The next step is obvious: test this in humans. The technology is ready. The biology is understood. The need is urgent. Let's see if what works in mice works in children.

---

References

[1] Batsios G, Taglang C, Udutha S, et al. Lactylation fuels nucleotide biosynthesis and facilitates deuterium metabolic imaging of tumor proliferation in preclinical models of H3K27M-mutant gliomas. Science Translational Medicine (2026). doi:10.1126/scitranslmed.adw0834

[2] Lactylation is a post-translational modification discovered relatively recently (2019) where lactate-derived lactyl groups are added to lysine residues on proteins, particularly histones, altering their function. This paper extends that concept to show lactylation of metabolic enzymes like NME1, not just structural proteins.