Brain gene expression syncs between bonded prairie voles

Just as romantic partners exhibit more similar brain waves than do strangers when, say, drawing on an Etch A Sketch toy together, animal pairs also show neural synchrony during social interactions and cooperation tasks.

“Neural synchrony is something that happens in these minute-to-minute engagements that you have with another individual,” says Zoe Donaldson, associate professor of behavioral neuroscience at the University of Colorado Boulder. But over time, too, pairs in a relationship learn to infer what their partner is going to do, she adds.

In prairie voles, at least, that learning process may unfold at the molecular level in the form of “transcriptional synchrony,” according to a preprint Donaldson and her colleagues posted on bioRxiv in November. Prairie voles are socially monogamous, and after two of them bond, gene-expression patterns in their nucleus accumbens—a forebrain region linked to reward and social interaction—start to align.

It remains unclear whether this transcriptional synchrony causes pair bonding or only correlates with it, she adds, but in the meantime, it offers researchers a new place to hunt for the basis of these strong social ties.

This new study “pushes the limits of what’s possible” technically, says Robert Froemke, professor in New York University’s Neuroscience Institute and otolaryngology department, who was not involved in the study. Though the existence of neural synchrony logically suggests that there may also be shared patterns of gene expression, “it’s still remarkable to actually have it documented,” he says.

T

he new preprint offers the first evidence of transcriptional synchrony in prairie voles, Donaldson says, but a 2020 study revealed that fighting pairs of Betta splendens fish show a strong correlation of gene expression after 60 minutes of fighting, and only a weak correlation after 20 minutes.

To look for the same phenomenon among prairie voles, Donaldson and her lab matched 11 opposite-sex and 9 same-sex pairs of the animals, representing something like “romantic relationships” and “friendships” in human-centric terms, she says. They let each pair live together in one cage for two weeks; “in vole land, that’s a long time,” she adds.

RNA sequencing of single nuclei from cells in the nucleus accumbens was successful in 39 out of the 40 animals, resulting in a dataset of more than 142,000 nuclei. A unique feature of this study is that the team ran these samples individually and did not pool them, says Kristen Berendzen, assistant professor in residence in psychiatry at the University of California San Francisco, who was not involved in the study. Although more onerous and expensive, Donaldson says, this sequencing approach provides a more detailed look at the differences and similarities in specific cell populations between individual animals.

To test the hypothesis that transcription would be similar between partners but not between non-partners, Donaldson’s team plugged the sequencing data into a support vector machine, a supervised machine-learning algorithm adept at finding differences among inputs and grouping similar ones together.

The team employed a “leave-one-out” approach to cross-validate the model and found that it was more likely to match cells from the vole excluded from any one iteration to the animal’s partner than to other voles.

The model correctly identified a vole’s partner 27.3 to 48.5 percent of the time—a “pretty good” result, well above the 2.7 percent accuracy expected to occur by random chance as reported in the study, Froemke says.

Additional analysis revealed that genetic similarity between animals also correlated with transcriptional similarities, but “this [effect] is exponentially smaller than the effect of who you were partnered with,” Donaldson says.

T

he genes most synchronized between bonded voles were expressed in oligodendrocyte progenitor cells and interneurons, further analysis suggests.

“Oligodendrocyte precursor cells would not be the first cell type where I would expect transcriptional synchrony to be evident,” Donaldson says, mainly because no one has definitively tied these cells to social interaction before. But how these cells might be involved is an “open question” she says she wants to investigate. Myelination is responsible for encoding complex memories in mice, previous evidence suggests, which might offer clues, she says. “We really do think of social bond formation as a form of complex social learning.”

The stronger the correlation between bonded voles’ oligodendrocyte gene expression, the more the two voles freely interacted over the course of three hours, a follow-up experiment revealed. These two variables were moderately correlated, Donaldson says. The interneuron genes, by contrast, did not predict pair behavior.

The use of this free-association assay is important, says Froemke, because the classic partner preference test limits an animal to a choice between their partner and one “stranger” vole, both of which are restrained and so themselves have no choice. “Science also advances when we realize that maybe it’s time to expand our assays and really just start looking again at the kind of spontaneous behaviors these animals enact.”

This work “opens up a lot of really cool pathways to go down,” such as how long it takes for gene expression to synchronize over the course of the relationship, Berendzen says. Also of interest, this transcriptional synchrony doesn’t appear to differ between same-sex and opposite-sex pairs, she says. “There is a social environment effect that is the most robust.”

This work leads to many more questions, Froemke agrees. Perhaps paramount among the concepts to explore, he notes, is that a change at the level of RNA-seq doesn’t necessarily mean something is functionally changing with the oligodendrocyte precursor cells in the brain, for example. The functional consequence of this transcriptional synchronization on neurons is a “complex, interesting question that needs to be worked out.”