Brain Tissue Grown in a Dish Kept Track of Its Own Age for Five Years

A human cortical neuron is in no hurry. It can take years to finish wiring itself, and in the front of the cortex the pruning of connections runs into the third decade of life. That patience is a problem for anyone trying to watch human brain development in a laboratory, because the lab grown clumps of brain tissue known as organoids have mostly been studied for a few months at a time, reaching something like a newborn’s brain after roughly 250 to 300 days. A team in Paola Arlotta‘s laboratory at Harvard University wanted to know what came after that, so they kept the same cultures going, tended in an isolated room, for more than five years.

Keeping tissue alive is not the same as letting it grow up. Telling those two apart was the real question, and it needed a measurement that did not simply count how many cells were still breathing.

So the team read the chemical tags that cells stick onto their own DNA. Certain spots in the genome gain or lose these methyl marks on a schedule regular enough that the pattern works as an age gauge, and the best known version, built from about 350 such spots, can place most human tissues within a few years of their true age. Reprogramming an adult cell back into a stem cell wipes that gauge nearly to zero, which is where every organoid in this study began. Sampling the tissue at nine points between three months and five years, the researchers found the predicted age climbing in step with time in the incubator, and a version of the gauge tuned specifically to human cortex agreed.

The Tissue Passed Birth Without a Body

By about a year in, the tissue had crossed a line that organoids had not crossed before. Cells started carrying molecular signatures that in a real brain show up only after birth, and the different cell types had arrived in roughly the order they do in a developing person. “In the human brain, these epigenetic changes accumulate according to a characteristic developmental pattern,” says Noelia Antón-Bolaños, an assistant professor at University Medical Center Utrecht who co-led the work with Irene Faravelli. “We observed the same pattern in the brain organoids.”

The neurons were the weak link. Their share of the cells shrank over the years, until the five year samples turned up a neuronal cluster of only about a hundred cells, though part of that decline is an artifact of the method: neurons are fragile, and breaking tissue into single cells for sequencing tends to destroy them. Staining intact slices told a kinder story, turning up neurons at two, three, four and nearly six years, but that method grew unreliable after the first year as background signal crept up.

The fix was a change of liquid. Standard organoid medium does not support the spontaneous electrical chatter that keeps young neurons alive, so from about ten weeks onward the team moved some cultures into a formulation designed to permit it. The difference showed in the wiring: at one year, roughly half the connections in the treated organoids sat on dendritic spines, the small knobs that mark a settled synapse, against about a quarter in the standard cultures. It showed more starkly on an electrode array. At the one year mark, all eight conventionally grown organoids produced no coordinated network bursts whatsoever, while all nine grown in the new liquid burst reliably.

That gap matters for how the result gets described. The organoids that reached five years were the conventionally grown ones, the same cultures in which neurons were thinning out, and the activity permitting recipe has so far been shown to hold neurons and network bursts for about two years, not five. The paper says plainly that these conditions would benefit from further optimization before neurons can be counted on across the whole span.

Old Cells Skipped the Steps They Had Already Taken

The strangest result came from taking organoids apart. The researchers broke up organoids that had been growing for nine to twelve months, then let the loose cells clump back together into fresh spheres. Left to themselves, the old cells picked up where they had left off and made the cell types appropriate to their age, mostly support cells rather than neurons.

Then the team mixed old cells with cells from organoids only two weeks old, using different donors for each group so they could tell afterward which cell came from where. Within two weeks the old cells were making neurons again, but not the early born ones their young neighbors were making. They jumped straight to a late arriving type that normally takes months to appear, and it accounted for nearly half of everything the old cells produced, against roughly one percent in an ordinary nine month organoid. “When we dissociated an older organoid and allowed the cells to grow again, they produced the cell types associated with a late developmental stage,” Antón-Bolaños says. The cells were not merely old; they carried a record of what they had already done.

What the Dish Still Cannot Do

None of this makes an organoid a brain. These are millimeter scale balls of tissue with no blood supply, and without vessels the interior runs short of oxygen and nutrients, which stresses and kills cells and caps how large the tissue can grow. They also lack the layered organization of a real cortex, and they have no body, no senses, and nothing to be about. Ethicists who have worked through the sentience question in detail argue that current cortical organoids fall well short of the structure most theories of consciousness assume, and that the measures used to gauge awareness in patients with brain injuries have no validated meaning when pointed at a dish.

What the tissue does have is time. Two year old organoids in the activity permitting liquid were bursting in two distinct groups, each with its own rhythm, which is not what younger tissue does. Whether that keeps sorting itself out at year five is unknown, because those cultures are only partway there.

FAQ