An ant can increase and decrease its brain size, according to its reproductive status

Reversible plasticity in brain size, behaviour and physiology characterizes caste transitions in a socially flexible ant (Harpegnathos saltator). Clint A. Penick, Majid Ghaninia, Kevin L. Haight, Comzit Opachaloemphan, Hua Yan, Danny Reinberg and Jürgen Liebig. Proceedings of the Royal Society B, April 14 2021. https://doi.org/10.1098/rspb.2021.0141

Abstract: Phenotypic plasticity allows organisms to respond to changing environments throughout their lifetime, but these changes are rarely reversible. Exceptions occur in relatively long-lived vertebrate species that exhibit seasonal plasticity in brain size, although similar changes have not been identified in short-lived species, such as insects. Here, we investigate brain plasticity in reproductive workers of the ant Harpegnathos saltator. Unlike most ant species, workers of H. saltator are capable of sexual reproduction, and they compete in a dominance tournament to establish a group of reproductive workers, termed ‘gamergates'. We demonstrated that, compared to foragers, gamergates exhibited a 19% reduction in brain volume in addition to significant differences in behaviour, ovarian status, venom production, cuticular hydrocarbon profile, and expression profiles of related genes. In experimentally manipulated gamergates, 6–8 weeks after being reverted back to non-reproductive status their phenotypes shifted to the forager phenotype across all traits we measured, including brain volume, a trait in which changes were previously shown to be irreversible in honeybees and Drosophila. Brain plasticity in H. saltator is therefore more similar to that found in some long-lived vertebrates that display reversible changes in brain volume throughout their lifetimes.

4. Discussion

Workers of Harpegnathos saltator exhibited reversible changes in brain size similar to that found in relatively long-lived vertebrate species. Changes in brain volume observed in vertebrates generally track seasonal reproductive cycles and are triggered by reproductive hormone cascades [9]. Likewise, brain changes in H. saltator also track the reproductive status and are associated with changes in reproductive hormone levels [20,21] and the expression of key regulatory genes [42]. Changes in the vertebrate brain include the seasonal addition of new neurons [54], which we did not specifically measure here, but changes in total and region-specific brain volumes are comparable.

[Figure 6. Correlated plasticity in brain, behaviour, and physiology between reproductive and non-reproductive workers. (Online version in colour.)]

Task or experience-dependent plasticity of brain compartments has been demonstrated in various insects, including honeybees, ants, paper wasps, and moths (e.g. [11,34,55–57]). In H. saltator, gamergate brains were 19% smaller than the brains of foragers on average, which is in line with predictions that brain size should be reduced to divert metabolic resources to reproduction [12,13]. Even compared to comparatively younger inside workers, gamergate optic lobes were 24% smaller, suggesting they may not simply retain the brain size of young nurse workers, but most likely experience region-specific brain volume reduction. When gamergates of H. saltator were reverted back to non-reproductive status, their brains re-expanded and matched that of forager brains. Foraging requires the ability to orient towards the nest and to attack and retrieve live prey items, all of which requires higher cognitive processing. The observed reduction in central brain volume of gamergates and the subsequent expansion in reverted gamergates suggest it is used for the more demanding cognitive abilities of foraging [58]. Changes in the central brain of H. saltator, which includes the mushroom body, are consistent with results from other social insects [59]. In carpenter ants, foragers that perform cognitively demanding tasks exhibit an increase of more than 50% of mushroom body neuropile volume [11] and a similar pattern is found in the mushroom body of honeybees [60].

The pattern of size differences in the optic lobe of H. saltator suggests a programmed rather than experience-dependent change in brain volume. Gamergates displayed significantly smaller optic lobes than inside workers and foragers, both of which had equally large optic lobes. Gamergates were still exposed to light and thereby received visual stimulation from their nest-mates in our laboratory settings, so sensory deprivation is an unlikely cause for the size differences we observed. Given that gamergates do not rely on optic information under natural conditions, a reproduction-dependent size reduction seems most likely. The intermediate optic lobe size of reverted gamergates relative to gamergates and non-reproductive workers suggests a presumably slower reversion speed of the optic lobe compared to the central brain. However, the size reductions of the optic lobes and of the central brain compared to reverted gamergates both suggest this brain size reduction is an energy-saving mechanism as proposed previously [12,13].

The reversibility of changes in brain size in H. saltator contrasts with results in the honeybee and in Drosophila. Brain size in honeybees increases as nurse workers transition to foragers, but when foragers are reverted to nurse status, they do not show a decrease in brain volume [32]. Honeybee foragers in the study by Fahrbach et al. [32] were only reverted back to nurses for 5 days, while gamergates in the present study were reverted for 6–8 weeks, which may explain why brain changes were observed in our study but not in previous studies on honeybees. In addition, honeybee foragers typically only live for a matter of weeks, and there is no biological ‘reason' for why they should fully revert to nurse status—in a colony of 50 000 bees, foragers can easily be replaced by new workers. By contrast, H. saltator colonies are small (usually less than 100 individuals), and each worker represents a more valuable resource in terms of their relative contribution to colony productivity. Studies in Drosophila have looked at region-specific changes in brain size associated with adult age, and while the medulla of the optic lobe in D. melanogaster increases in size with age, sensory deprived medullae do not increase in size and this lack of growth seems to be irreversible later in life [33]. This difference in brain plasticity corroborates differences in the plasticity of the antennal lobe between ants and Drosophila. When the odorant receptor co-receptor (orco) was knocked out in ants, the antennal lobes showed a significant reduction in two ant species [61,62]. A similar morphological change was not present in Drosophila when orco was knocked out, which suggests a hardwired mode of olfactory glomeruli formation in the Drosophila antennal lobe [63] and potentially major differences in brain development between Drosophila and ants.

Along with reversion in brain size, we found behavioural and physiological reversions that include the ovarian activity, venom production, CHC profile, and expression of associated genes (figure 6). Combined changes in physiological traits and underlying gene expression levels demonstrate that the changes we observed in reverted gamergates were not random, but instead matched a clearly defined worker phenotype. If we had observed a mix of different physiological changes that were inconsistent with the worker phenotype, then we might have expected these changes to be driven by isolation stress alone. The effects of chronic stress are generally expected to increase the allostatic load and result in decreases in body mass and brain function [64], yet contrary to this prediction, we observed increases in brain size and venom production in reverted gamergates. Likewise, changes in gene expression of ELOV, which is involved in fatty acid elongation, and Vg, the vitellogenin egg yolk precursor, were consistent with downstream physiological responses of CHC profiles and ovarian activity.

The observed reversibility in phenotypic plasticity in H. saltator gamergates that transition back to non-reproductive workers is present despite the rarity of such events. Naturally, queens and gamergates reproduce until senescence and do not substantially contribute to foraging after the loss of status. Thus, there would appear to be little selective pressure to keep reproductive specialization reversible. However, reversibility of phenotypic plasticity could be maintained to allow workers the return to forager status after they have lost a reproductive tournament. Dominance tournaments last up to 40 days in H. saltator, and initially up to half the workforce of a colony may compete [41]. Physiological changes begin shortly after tournaments are initiated [21,42], and reversibility may allow workers to return to forager status without suffering long-term effects associated with the early transition period to reproductive status. Among social insects, lower termites offer another rare example of reversible plasticity, in which individuals develop regressively from nymphal instars to ‘worker' instars that lack wing buds [65]. The precise reason why regressive moults occur in lower termites is not understood, but it does have parallels to reversible plasticity in H. saltator. In both H. saltator and lower termites, reversible plasticity allows individuals to retain flexibility in shifting between non-reproductive and reproductive pathways.