Raw and processed GC-MS lipidomics and LC-MS metabolomics data presented in Ahsan, et al. Molecular Cell Volume 86 Issue 13 (2026), "Reductive death is averted by a conserved de novo lipogenic switch"
Description
Biguanides, including metformin, the world’s most prescribed oral hypoglycemic, extend health span and lifespan in vertebrates and invertebrates. Given the widespread use and apparent safety of metformin, it is assumed that its effects are not associated with toxicity, except when in marked excess. Here we determine that accumulation of damaging reducing equivalents is an unanticipated toxicity associated with biguanides, the defense against which requires post-transcriptional protection of de novo fatty acid biosynthesis. We demonstrate that biguanide treatment during impaired fatty acid biosynthesis drives NADPH toxicity, leading to catastrophic elevation of NADH/GSH reducing equivalents and accelerated death across metazoans. Multiple NADPH-generating interventions require fatty acid biosynthesis to prevent markedly shortened survival, indicating that this defense mechanism is broadly leveraged. We propose that fatty acid biosynthesis is a tunable rheostat which can minimize biguanide-induced reductive stress whilst maximizing its pro-longevity outcomes and serve as an exploitable vulnerability in reductive stress sensitive cancers.
Files
Steps to reproduce
Intracellular unlabeled liquid chromatography-mass spectrometry (LC-MS) metabolomics was performed as previously described. Briefly, exactly 1000 worms for each replicate of the indicated genotype, stage, or treatment conditions were counted, manually picked and harvested in M9 buffer, and washed three times to remove residual metabolite treatment or bacterial food source from the plates. Worms were then transferred into a prechilled microcentrifuge tube and metabolites were extracted using a solution containing HPLC-grade 40% (v/v) methanol, 40% (v/v) acetonitrile, and 20% (v/v) ultrapure distilled water. Samples were sonicated at 70% amplitude using a QSonica Q800R water batch sonicator at 4°C for 20 minutes using a 30 second on, 30 second off cycle. Samples were then spun down at 14,000 RPM at 4°C for 20 minutes to pellet cuticle and cellular debris. The supernatant was then transferred into a fresh microcentrifuge tube, and 5 µL of each sample was loaded onto a ZIC-pHILIC column (Millipore). Mass spectra were acquired on a Thermo Q-Exactive Plus run in polarity switching mode with a scan range of 70-1000 m/z and a resolving power of 70,000 at 200 m/z. For each run, the total flow rate was 0.15 mL/min and the samples were loaded at 80% B. The gradient was held at 80% B for 0.5 min, then ramped to 20% B over the next 20 min, held at 20% B for 0.8 min, ramped to 80% B over 0.2 min, then held at 80% B for 7.5 min for re-equilibration. Data were subsequently extracted using Xcalibur (v.4.1.31.9) and analyzed using TraceFinder (v.4.1) and Progenesis (v.2.3.6275.47961). Heatmaps of z-score normalized relative abundances of select metabolites were generated using R package pheatmap v1.0.12. Relative abundance or ratiometric metabolite values were otherwise plotted using Graphpad Prism 10. Biological replicate assessment and statistical tests used are as noted in the Figure Legends of the original manuscript.
Institutions
- Harvard Medical SchoolMA, Boston
- Massachusetts General HospitalMA, Boston