A cell that eats itself

The word autophagy, “self-eating”, was coined in the 1960s by Christian de Duve, who had discovered the lysosome, the cell’s digestive compartment. Under the electron microscope, cells could be seen wrapping parts of their own interior, including whole mitochondria, in a double membrane and delivering the package to the lysosome to be broken down. Nobody knew how the cell did it, or which genes were involved. The process was hard to study because it happened inside small, short-lived structures that were difficult to see.

Ohsumi’s experiment

Yoshinori Ohsumi, working in Tokyo in the late 1980s, chose baker’s yeast. Yeast has a large vacuole, its equivalent of the lysosome, which is easy to see in an ordinary light microscope. Ohsumi reasoned that if yeast cells carried out autophagy, the material they delivered to the vacuole would be digested too quickly to see. So he used mutant strains that lacked the vacuole’s main digestive enzymes and starved them.

The genes Ohsumi found in yeast turned out to have close relatives in animals and plants, and the machinery they encode is now known in detail: a complex that senses nutrients and starts the process, a system that builds the double membrane, and a protein, called Atg8 in yeast and LC3 in mammals, that marks the membrane and is used to measure autophagy in the laboratory (Mizushima and Komatsu 2011). Autophagy turned out to matter for far more than starvation: for clearing damaged mitochondria and protein aggregates, for immunity, and for development. Defects in autophagy genes are linked to neurodegenerative, inflammatory and other diseases (Levine and Kroemer 2019).

The leap from yeast to people

The popular version of the story runs: fasting switches on autophagy; autophagy cleans the cell; therefore fasting cleans the cells of the body, including the brain, and slows ageing. Each step is partly true, and the whole is not established in people. The weak points are the timing, the tissues, and the measurement.

mouse_kg, human_kg = 0.025, 70.0
# energy use per kg scales as body mass ** -1/4 (Kleiber)
scale = (human_kg / mouse_kg) ** 0.25
print(f"a mouse burns energy per kg about {scale:.1f} times faster than a human")
for mouse_h in (24, 48):
    days = mouse_h * scale / 24
    print(f"a {mouse_h} h fast in a mouse ~ {days:4.1f} days of a human's metabolic time")
# a mouse burns energy per kg about 7.3 times faster than a human
# a 24 h fast in a mouse ~  7.3 days of a human's metabolic time
# a 48 h fast in a mouse ~ 14.5 days of a human's metabolic time

What can reasonably be said

The honest summary of the Nobel story is therefore this: Ohsumi showed how cells digest themselves when starved, and in doing so opened a field that matters for medicine. He did not show that human fasting cleans the brain. That question is still open, and Chapter 9’s trial does not try to answer it; it asks something that can be measured in people, namely whether a fasting schedule changes attention.