Finding. Mature neurons still make DNA to repair themselves, and polymerase kappa is one of the enzymes on that patch. Goyal and Paul show POLK next to EdU in mouse and human neurons. Take POLK down and EdU falls, arbors simplify, nuclei swell and wrinkle, and a senescence-like program comes on, including less mitochondrial-associated staining. In APOE e4/e4 neurons the same loss raises cytoplasmic double-stranded DNA and phospho-tau. Put POLK back and EdU rises while 53BP1, gamma-H2AX, p38, cytoplasmic dsDNA, 6E10 amyloid stain, and phospho-tau fall.
Why this paper matters
Repair synthesis in postmitotic neurons needed a named polymerase. POLK is now a candidate, with an Alzheimer-relevant overlay on APOE4. The mitochondrial reason to keep it is modest and real: staining drops, and cytoplasmic dsDNA appears, which is how cGAS papers start. This week's TFAM-cGAS melanoma brief is the tumor version of a DNA leak. Here the leak is in a neuron.
What they actually measured
EdU, morphology, senescence panel, staining, IP-MS/PLA, AD marks, depletion and augmentation. No Seahorse.
How to read the score
Around 80. Nuclear repair first, mitochondria as a stress node. Confidence is medium for the organelle claim.
Caveats
Staining is not function. dsDNA origin unknown. No mouse.
What to do with it
If you EdU-label neurons, co-stain POLK. If you work APOE4 iPSC neurons, add cytoplasmic dsDNA and a mitochondrial stain to the POLK dose. Do not call POLK a mitochondrial polymerase from this brief.