Oldfield mice and deer mice differ in their parental care, most dramatically in the behaviour of fathers. A study reveals the genetic and neuronal contributions to variation in parental care. See Article p.434
One of the most pressing questions facing biologists is how variation in the genome is translated into variation in complex traits. Perhaps no trait is more complex or more interesting than social behaviour. How do genomic differences influence aggression, courtship or bonding? Few studies have asked. On page 434, Bendesky et al.1 combine genetic crosses, detailed behavioural assays and the latest neuroscience tools to understand species differences in parental care in mice of the genus Peromyscus.
Peromyscus is a diverse North American genus, whose habitats range from arid deserts to montane cloud forests2. Along with these tremendously disparate habitats come comparably variable behaviours. For example, sharing of parental care and social monogamy are rare traits in mammals, but they have evolved at least twice in Peromyscus3. Oldfield mice (P. polionotus), for instance, live in sandy habitats at low population densities, and seem to have adapted to their sparse environments by forming pair bonds, with both sexes providing ample parental care (Fig. 1). By contrast, the deer mouse (P. maniculatus) is tremendously widespread and has a promiscuous mating system similar to that found in many other rodents3. Deer mice provide less parental care than do oldfield mice, with this difference being particularly pronounced in fathers.
Despite their differences, these closely related species can be crossed to produce fertile offspring, making them excellent subjects for exploring the genetic basis of behavioural variation. Bendesky et al. crossed the two species for two generations, producing hundreds of grandchildren that varied in the genomic segments they inherited from each species. The group found that some regions of the genome predicted care in general — influencing how much nest building took place as well as how readily pups were retrieved when they were moved away from their parents, for example. Others were more specific, influencing only nest building. Surprisingly, the relationship between parental care and specific DNA sequences often differed between males and females, indicating that many of the mechanisms underlying species differences are specific to either maternal or paternal care.
To investigate which genes might shape species differences in parental care, the authors focused on a region of the genome associated with male nest-building behaviour alone. The segment contained 498 genes; to shorten this list, the authors examined gene expression in first-generation hybrid mice, asking whether the expression of copies derived from each species differed when the two copies were present in the same cells. The authors focused on the hypothalamus, a brain region extensively implicated in parental care. Nine genes showed species-specific expression levels in this region, of which the gene that encodes vasopressin — a hormone involved in male social behaviour4 — showed the biggest interspecies differences. In particular, the gene copy derived from the oldfield mouse was expressed at lower levels than the copy from the deer mouse.
Do high vasopressin levels create bad dads? Bendesky et al. showed that injecting vasopressin into the brains of oldfield mice did in fact reduce nest building, but had no effect on other dimensions of parental care — a finding consistent with the authors' genetic data. However, the injected vasopressin might be acting in many brain areas other than the hypothalamus, or in any of several brain regions within the hypothalamus3,4,5.
Among the possible sources of vasopressin, a region of the hypothalamus known as the paraventricular nucleus (PVN) was a particularly likely candidate. The PVN is an important source of vasopressin and, through its projections to another hypothalamic region called the medial preoptic area, is a major regulator of parental care4,5,6. Bendesky and colleagues investigated whether the vasopressin-expressing neurons in the PVN might inhibit parental care. Switching to laboratory mice, the group manipulated vasopressin neurons to express a gene that alters neuronal firing rate in response to a drug7. Increasing the activity of these PVN neurons inhibited nest building, and decreasing vasopressin activity had the opposite effect. Thus, it seems that paraventricular vasopressin contributes to differences in male parenting behaviours between species.
In many respects, this work dovetails nicely with the existing literature. Studies of the prairie vole, another socially monogamous rodent that shares parenting, have suggested that similar parenting behaviours between sexes might arise from different neurobiological mechanisms8. Even human imaging suggests that paternal care may involve neuronal mechanisms that are distinct from those of maternal care9. Bendesky and colleagues' systematic look at the genome suggests that this may prove to be a general pattern.
Although the current study is consistent with much of the existing literature, it does provide some surprises. Hypothalamic vasopressin has been implicated in parental care in several rodents, but the hormone promotes, rather than inhibits, care in these species8,10,11. It is not clear why this disparity might have arisen. One difference is that previous studies examined vasopressin released in the medial preoptic area; the current study is the first to directly analyse the effects of paraventricular vasopressin neurons on parental care. Perhaps more crucially, the vast majority of previous work on parental care has focused on pup-directed behaviours such as retrieving, grooming and crouching over pups. Bendesky and colleagues' demonstration that the genetic variation in nest building can be dissociated from pup-directed behaviours suggests that these aspects of care rely on distinct mechanisms.
Overall, the current experiments demonstrate that species differences in a complex behaviour can be traced back to differences in specific genes and circuits. This work serves as a harbinger of exciting science to come. As non-model organisms become increasingly amenable to genetic and neuronal analysis, researchers can explore how brain and behavioural variation emerge from the complex interaction of the genome and the environment. Such work offers a more integrated view of biology than was previously achievable, and allows us to explore questions that have long been intractable.Footnote 1
Bendesky, A. et al. Nature 544, 434–439 (2017).
Bedford, N. L. & Hoekstra, H. E. eLife 4, e06813 (2015).
Turner, L. M. et al. Mol. Biol. Evol. 27, 1269–1278 (2010).
Caldwell, H. K., Lee, H.-J., Macbeth, A. H. & Young, W. S. Prog. Neurobiol. 84, 1–24 (2008).
Bester-Meredith, J. K., Young, L. J. & Marler, C. A. Horm. Behav. 36, 25–38 (1999).
Numan, M. & Young, L. J. Horm. Behav. 77, 98–112 (2016).
Roth, B. L. Neuron 89, 683–694 (2016).
De Vries, G. J. & Panzica, G. C. Neuroscience 138, 947–955 (2006).
Atzil, S., Hendler, T., Zagoory-Sharon, O., Winetraub, Y. & Feldman, R. J. Am. Acad. Child Adolesc. Psychiatry 51, 798–811 (2012).
Bosch, O. J. & Neumann, I. D. Proc. Natl Acad. Sci. USA 105, 17139–17144 (2008).
Bales, K. L. & Saltzman, W. Horm. Behav. 77, 249–259 (2016).