A New Genetic Clue to How Humans Got Such Big Brains

Three genes that appeared during our early evolution probably increased the number of neurons in our heads—but at a cost.

A woman sits besides a real human brain, displayed in Bristol, England.
Matt Cardy / Getty

It started with some blobs of brain-like tissue, growing in a dish.

Frank Jacobs, then at the University of California at Santa Cruz, had taken stem cells from humans and monkeys, and coaxed them into forming small balls of neurons. These “organoids” mirror the early stages of brain development. By studying them, Jacobs could look for genes that are switched on more strongly in the growing brains of humans than in those of monkeys. And when he presented his data to his colleagues at a lab meeting, one gene grabbed everyone’s attention.

“There was a gene called NOTCH2NL that was screaming in humans and off in [the monkeys],” says Sofie Salama, who co-directs the Santa Cruz team with David Haussler. “What the hell is NOTCH2NL? None of us had ever heard of it.”

The team ultimately learned that NOTCH2NL appears to be inactive in monkeys because it doesn’t exist in monkeys. It’s unique to humans, and it likely controls the number of neurons we make as embryos. It’s one of a growing list of human-only genes that could help explain why our brains are so much bigger than those of other apes.

These human-only genes are typically created when bits of DNA accidentally get duplicated. Duplication creates backup copies of existing genes, which are then free to mutate with impunity and take on new roles. In this way, duplication events provide fresh fuel for evolution.

They also cause headaches for researchers who are trying to understand our genome. Scientists sequence genomes by breaking long stretches of DNA into more manageable fragments. They then decipher each piece separately, before assembling the pieces into a whole. But when genes are duplicated, fragments of the copies are almost indistinguishable from fragments of the originals, which causes confusion. It’s like trying to assemble several jigsaw puzzles that are only slightly different: When their pieces are jumbled up, it looks like they all come from a single puzzle.

That was the case for NOTCH2NL. In earlier drafts of the human genome, it looked like one gene. But when the latest (and 20th) draft was released in December 2013, Jacobs and his colleagues realized that this mysterious gene was actually three genes. They’re known as NOTCH2NLA, NOTCH2NLB, and NOTCH2NLC. They’re 99.7 percent identical to each other. And they have a convoluted history.

In the common ancestor of all great apes, there was just one gene: NOTCH2. At some point, it was duplicated, but only partially. Its doppelganger, the first NOTCH2NL gene, was missing important sections, and so didn’t work properly. It was useless, an instruction manual with random chapters torn out. To this date, chimps and gorillas still have these dead versions of NOTCH2NL.

But between 3 and 4 million years ago, in the ancestors of humans, something special happened. The original NOTCH2 gene partly overwrote its broken duplicate. This process, known as gene conversion, revived NOTCH2NL, allowing it to play an active role in its owners’ biology. And having been resurrected, it duplicated twice more, creating the A, B, and C genes that we have today.

While Jacobs’s team was learning all of this, Ikuo Suzuki and colleagues from KU Leuven, a university in Belgium, were homing in on the NOTCH2NL genes through a different route. They started by identifying genes that have three characteristics: They arose from duplication events, are strongly active in the developing brain, and are unique to humans. Suzuki and his team came up with a shortlist of 35 genes, and introduced several of these into the brains of embryonic mice to see what would happen.

One gene—NOTCH2NLB—had a particularly interesting effect on the radial glia, the cells responsible for building a brain. The radial glia are like factories that manufacture two products: neurons and more factories. Both Suzuki and Jacobs found that NOTCH2NL genes nudge the glia toward the latter: They make more of themselves. As their numbers swell, they collectively churn out more neurons and build bigger brains. By influencing the radial glia, the NOTCH2NL genes might have contributed to the evolution of our large brains and vaunted intellects.

These changes could have come at a cost. The NOTCH2NL genes are so similar that even our cells can get them confused. As a result, the stretch of DNA where these genes reside is very unstable. Sometimes, it gets duplicated. Sometimes, it’s deleted. Sometimes, the A gene might overwrite the B one, or vice versa. These genetic upheavals can lead to developmental disorders.

In extreme cases, the duplication of the NOTCH2NL genes can lead to macrocephaly, where people grow up with unusually large brains and heads. Conversely, the wholesale loss of these genes can lead to microcephaly—a condition of small brains and heads. Other changes in this region have been linked to autism, schizophrenia, and intellectual disorders. “It’s fascinating to think that the same mechanism that helped enable a bigger brain might also make us susceptible to these disorders,” Salama says. “We’re paying the price for the gain we got in our evolution.”

For now, it’s hard to accurately tell how much NOTCH2NL genes vary between people, and how specific variations influence either brain size or risk of disease. That will likely change, as new “long-read” technology allows scientists to sequence large continuous stretches of DNA without having to first break it into pieces. “As we sequence more human genomes using long-read methods, we will get a more complete picture of NOTCH2NL’s role in disease and human traits,” says Megan Dennis, from UC Davis, who wasn’t involved in either study.

The NOTCH2NL genes are far from the only ones related to brain size. Others like them have been recently identified, with equally tortuous names like SRGAP2C and ARHGAP11B. And don’t forget that Suzuki focused on NOTCH2NLB after first identifying a shortlist of 35.

“We hit the jackpot here but there might be many more jackpots,” says Pierre Vanderhaeghen, who led Suzuki’s study.