Strange Squeaks: A Uniquely Human Gene Alters Mouse “Songs”

Many animals have surprisingly advanced forms of communication: Birds use different alarm calls to signal how dangerous a predator is, dolphins have signature whistles that are thought to be similar to human names, and chimpanzees, like the famous Nim Chimpsky, can learn more than a hundred sign language words.^1–3 But human language is unique in its complexity. Endless combinations of nouns, verbs, and adjectives allow for the transmission of crucial information about the past, present, and future, and contribute to the enormous success of our species. Because language itself does not fossilize, however, scientists have long wrestled with the question of when and how this key adaptation arose.

Now, detailed investigations into the genomes of diverse modern human populations, along with Neanderthals and Denisovans, are providing new clues about the origins of spoken language. In a recent study published in Nature Communications, neuroscientists Robert Darnell and Erich Jarvis of The Rockefeller University present evidence suggesting that a specific variant of the gene NOVA1 could be a key factor in language development.⁴ They determined that this variant is possessed by nearly every person on Earth and is also unique to modern humans, separating us from both living primates and our closest, now extinct, relatives. Furthermore, mice engineered with this human variant displayed subtle changes in the splicing of some genes related to language along with changes in ultrasonic vocalizations.

“I was really excited by [this study],” said Andreas Pfenning, a neurogenomics researcher at Carnegie Mellon University who was not involved in the study. “So much of the field of human speech evolution has been dominated by FOXP2.” Back in 2002, researchers proposed that FOXP2 was important for the development of spoken language, as the human variant was not found in other extant primates and mutations in this gene were associated with language difficulties in humans.⁵ However, the gene was deposed sixteen years later when larger and more diverse human genome datasets failed to show evolutionarily recent positive selection of this gene.⁶ “It's really encouraging to see the repertoire expand—of molecular correlates of human speech and vocal behavior more broadly,” noted Pfenning.

Darnell has long been fascinated by NOVA1, but his initial interest had nothing to do with language. When Darnell was a neurology resident at Weill Cornell Medical Center in the late 1980s, he became intrigued by a set of rare conditions called paraneoplastic neurologic disorders, in which patients with cancer suddenly developed problems with balance, vision, or memory. While the symptoms were clearly neurological in origin, the patients’ brains were cancer-free. When Darnell and his colleagues analyzed the serum of an individual with this condition, they found autoantibodies specific for proteins found in some types of tumors, but also in the brain, triggering an immune response against both the cancer cells and the neurons.⁷ They subsequently identified the culprit: neuro-oncological ventral antigen, or NOVA.⁸

Over the following decades, Darnell explored the crucial functions of this RNA-binding protein, demonstrating its role in brain development, synaptic function, and glioblastoma cell fitness.^9–11 Darnell was inspired to explore NOVA1’s evolutionary history in part by Max Planck Institute geneticist Svante Pääbo’s 2012 genome sequencing of a Denisovan, a type of archaic human.¹² Pääbo identified only a handful of genes involved in brain development that were highly conserved in other primates but different in modern humans: NOVA1 was among them.

In collaboration with Cold Spring Harbor Laboratory computational biologist Adam Siepel, Darnell and his team explored the prevalence of this particular NOVA1 variant using human genome databases. Out of more than 650,000 human sequences, only six lacked this particular variant. “[Essentially] every other human on the planet has this human-specific variant,” said Darnell. “But no Denisovan does, no Neanderthal does, no chimpanzee does… Our interpretation is that as Homo sapiens emerged in Africa—before they emigrated north to populate Asia and Europe—there was an evolutionary sweep of a lot of things.”

Darnell said the switch to the new NOVA1 variant likely occurred between 40,000 and 200,000 years ago, altering RNA regulation and potentially changing human brain development in a way that enabled the emergence of spoken language.

To explore the functional implications of human NOVA1, researchers used CRISPR gene editing to create mice that expressed the uniquely human version of this protein. Within the cell, NOVA1’s job is to bind to RNA to regulate alternative splicing—a process by which exons from the same gene are assembled in different ways to produce different RNA products, thus influencing the proteins that are subsequently translated. The researchers assessed NOVA1-bound transcripts throughout the brain in wildtype and humanized mice and found them to be quite similar; both showed an enrichment of transcripts related to synapse function.

However, when they compared alternative splicing events in the periaqueductal gray, a midbrain region which is a key part of the vocal motor pathway in mammals, the researchers found subtle, but potentially important differences between the two groups. Of the handful of differently spliced genes, four were related to vocalization. In humans, mutations in these genes have been associated with neurodevelopmental disorders, hearing loss, and epilepsy.

Finally, the researchers wanted to determine whether the human NOVA1 impacted mouse behavior. They recorded the ultrasonic distress calls of mouse pups separated from their mothers and found subtle variations in the “syllables” the pups produced, although there were no appreciable changes in the mothers’ responses. Darnell and his team also identified differences in the ultrasonic courtship “songs” produced by adult males. Collectively, these data indicate that this human-specific NOVA1 variant induces alternative splicing of certain vocalization-related genes as well as altering the quality of vocalizations themselves.

“What I would want to know next is exactly how [these findings] relate to the neural circuits for vocalization and for speech,” said Pfenning. “There's evidence at the molecular level and there's evidence at the behavioral level, but the two are still disconnected. It's not yet clear how the molecular differences relate to a difference in how the brain regions connect to each other, which would allow for the difference in behavior.”

References