Neurons need their mitochondria to prune their own branches — in fruit flies

Edited by Samael A. Tejada, Founder & Editor-in-Chief

To rebuild itself for adulthood, a fly neuron must first destroy part of itself — precisely, and without dying. This study catches the moment the cutting machinery switches on, and finds the cell's power plants standing in the middle of the operation.

How can a neuron use death enzymes and live?

Caspases are best known as the executioners of apoptosis, but development deploys them sub-lethally for fate specification, migration and differentiation. The fly's dorsal body wall offers a side-by-side comparison in one intact animal: at metamorphosis, class III sensory neurons die by apoptosis while their neighbor, the class IV neuron ddaC, merely prunes its larval dendrites and regrows an adult arbor. With the SR4VH probe, caspase substrates were cleaved in ddaC within 20 minutes of puparium formation — before any branch was severed — at consistently lower levels than in its dying neighbors. Dose, it appears, is the difference between remodeling and death.

What roles do the RHG proteins and mitochondria play?

Single-cell H99 mutant clones showed the proapoptotic RHG proteins are cell-autonomously required for pruning — the signal must come from inside the pruning neuron itself. Although reaper, hid and grim are all transcribed in ddaC as remodeling begins, mutant combinations point to reaper and grim, rather than hid, cooperating to drive branch severing and clearance. Mitochondria sit upstream of the scissors: disrupting mitochondrial function or transport depleted dendritic mitochondria, prevented caspase activation, and blocked both pruning in ddaC and apoptosis in neighboring class III neurons. No mitochondria in the branches, no caspase signal, no cut.

Why does fly pruning matter beyond flies?

Carefully: this is a mechanism paper in an insect model, and it would be wrong to read a human therapy into it. Its value is conceptual and it is real. Developmental pruning is how mammalian brains — including human ones — also refine circuits, and failed or excessive pruning is implicated in neurodevelopmental research. Demonstrating in a living animal that mitochondria gate the caspase signal that severs branches gives that field a concrete, testable mechanism: mitochondrial position and transport inside a dendrite may decide where and whether the cut happens. That is a hypothesis for mammalian work now, not a finding about people.

Where the evidence stands

• Established: Caspases act sub-lethally in development as well as lethally in apoptosis; dendrite pruning is a normal developmental process in Drosophila metamorphosis; and mitochondria are central to caspase activation in cell-death pathways.
• Preliminary: In Drosophila sensory neurons in vivo (Scientific Reports, Nov 3, 2026; DOI 10.1038/s41598-026-77081-9), caspase cleavage began within 20 minutes of puparium formation at sub-lethal levels, Reaper and Grim were cell-autonomously required for pruning, and disrupting mitochondrial function or transport blocked caspase activation, pruning and neighboring apoptosis.
• Absent: Any mammalian or human data, any disease model, and any intervention — this is basic developmental mechanism research in fruit flies.