Batsios et al. · bioRxiv (preprint) · 2025
Lactylation fuels nucleotide biosynthesis and facilitates deuterium metabolic imaging of tumor proliferation in H3K27M-mutant gliomas
Batsios G, Taglang C, Udutha S, Gillespie AM, Robinson SP, Phoenix T, Mueller S, Venneti S, Koschmann C, Viswanath P
The study
What was asked, and what was found
Batsios et al., bioRxiv (preprint), 2025 asked what lactate does for a tumour, rather than treating it as waste. Working in H3K27M-mutant diffuse midline glioma, a paediatric brain tumour, the authors traced labelled glucose through five patient-derived models and found that the mutation raises phosphoglycerate kinase 1 and drives lactate production, and that lactate is needed for nucleotide biosynthesis. Lactate modifies the nucleoside diphosphate kinase NME1 at lysine 49, and that modification is what keeps the supply of nucleoside triphosphates up. Blocking it arrested cells in S phase and triggered replication stress and DNA damage, and the same modification was present in patient biopsies carrying the mutation and absent from wild-type tumours and normal brain.
Every one of those comparisons needs cells with the mutation and matched cells without it. Two models, BT245 and DIPG-13, came from CRISPR-Cas9 editing. For the other three, SF8628, DIPG-6 and QCTB-R059, the authors used AUMsilence sdASOs against the mutant allele. The oligonucleotide is 21 nucleotides long, spans the mutation site in exon 2 of H3F3A and is complementary to the mutant allele, and it was run against a scrambled control. It was added to the culture medium at 5 μM with no transfection reagent, and loss of the mutant protein was confirmed at 72 hours.
The silencing reproduced what the gene editing did: the repressive H3K27me3 mark rose, H3K27 acetylation fell, and the flow of labelled glucose into glycolysis, the tricarboxylic acid cycle, the pentose phosphate pathway and nucleotide biosynthesis dropped in every model. It also cut production of deuterium-labelled lactate, which matters because the second half of the paper builds a non-invasive imaging readout on exactly that signal. Deuterium metabolic imaging localised the lactate signal to the tumour in mice and rats at 9.4T and at the clinical field strength of 3T, and the signal fell within a week of treatment, before tumour volume changed.
Key findings
- AUMsilence sdASOs against the mutant H3.3K27M allele, added to the medium at 5 μM with no transfection reagent, were the tool used to silence the driver mutation in three of the five patient-derived glioma models in this study.(Materials and Methods, Gene silencing and overexpression, page 19)
- Silencing the mutation with the oligonucleotides reproduced what gene editing did in the other two models, raising the repressive H3K27me3 mark and lowering H3K27 acetylation across every model.(Results, page 7)
- Loss of H3K27M expression cut the flow of labelled glucose into glycolysis, the tricarboxylic acid cycle, the pentose phosphate pathway and nucleotide biosynthesis in every model.(Results, page 7)
- It also lowered phosphoglycerate kinase 1, the enzyme the authors identify as the link between the mutation and lactate production.(Results, page 8)
- Silencing the mutation cut the deuterium-labelled lactate signal that the study proposes as a non-invasive readout, in every patient-derived model, which is the step that ties the imaging signal to the mutation itself.(Results, page 12)
- The mechanism the study lands on is that lactate modifies the nucleoside diphosphate kinase NME1 at lysine 49, and that this modification is what supports nucleotide synthesis and proliferation.(Results, page 11)
For research use only. Not for use in diagnostic or therapeutic procedures.