Socialization causes long-lasting behavioral changes

Abstract

In modern human societies, social isolation acts as a negative factor for health and life quality. On the other hand, social interaction also has profound effects on animal and human, impacting aggressiveness, feeding and sleep, among many other behaviors. Here, we observe that in the fly Drosophila melanogaster these behavioral changes long-last even after social interaction has ceased, suggesting that the socialization experience triggers behavioral plasticity. These modified behaviors maintain similar levels for 24 h and persist up to 72 h, although showing a progressive decay. We also find that impairing long-term memory mechanisms either genetically or by anesthesia abolishes the expected behavioral changes in response to social interaction. Furthermore, we show that socialization increases CREB-dependent neuronal activity and synaptic plasticity in the mushroom body, the main insect memory center analogous to mammalian hippocampus. We propose that social interaction triggers socialization awareness, understood as long-lasting changes in behavior caused by experience with mechanistic similarities to long-term memory formation.

Keywords: Aggression; CREB; Drosophila; Feeding behavior; Sleep; Social interaction; Socialization awareness; Synaptic plasticity.

Fig. 1

Long-lasting behavioral changes induced by socialization depends on memory-related genes. (A) Scheme of the socialization protocol: recently eclosed animals were either grouped or isolated for five days, and subsequently tested. (B) Quantification of food consumption of wt, rut and dnc mutant flies in socialized and isolated conditions (single fly CAFE assay) in the 0–24 h time window (Kruskal-Wallis chi-squared = 75.905, df = 5, p-value = 6.022e-15; post hoc Dunn comparisons: wt^social|wt^isolatedp = 6.24e-13, rut^social|rut^isolatedp = 1.00, dnc^social|dnc^isolated = 1.00). (C) Scheme of the modified socialization protocol: 5-day grouped or isolated animals were isolated for 24 h before testes. (D) Quantification of food consumption of wt, rut and dnc mutant flies in socialized and isolated conditions (sCAFE) in the 24–48 h h time window (Kruskal-Wallis chi-squared = 32.698, df = 5, p-value = 4.32e-06; post hoc Dunn comparisons: wt^social|wt^isolatedp = 1.04e-03, rut^social|rut^isolatedp = 1.00, and, in fig S1, dnc^social|dnc^isolated = 1.00). (E) Sleep profile and (F) sleep quantification of the 24–28 h time window for wt and rut mutant background (Kruskal-Wallis chi-squared = 94.165, df = 3, p < 2.2e-16; post hoc Dunn comparisons: wt^social|wt^isolatedp = 1.07e-02, rut^social|rut^isolatedp = 0.438). (G) Sleep quantification of the 24–28 h time window for wt and dnc mutant background (Kruskal-Wallis chi-squared = 36.476, df = 3, p-value = 5.94e-08; post hoc Dunn comparisons: wt^social|wt^isolatedp = 3.43e-02, dnc^social|dnc^isolatedp = 3.48e-03).

Fig. 2

Time course of socialization effects after isolation. (A) Scheme for the time course of the aggression protocol: either isolated or grouped animals were isolated for 1,4, 8–24 h, then grouped with other male and their aggression quantified. (B) Quantification of proportion of time expended lunging after different times of re-isolation. Flies either wt or in a rut mutant background were grouped or isolated for 5 days and then socialized flies were tested after 1, 4, 8–24 h after isolation (Kruskal-Wallis chi-squared = 139.99, df = 9, p-value < 2.2e-16, post hoc Dunn comparisons: wt^{24h_after_social}|wt^isolatedp = 1.08e-07, wt^{8h_after_social}|wt^isolatedp = 1.61e-10, wt^{4h_after_social}|wt^isolatedp = 1.88e-05, wt^{1h_after_social}|wt^isolatedp = 2.24e-09, rut^{24h_after_social}|rut^isolatedp = 1.00, rut^{8h_after_social}|rut^isolatedp = 0.815, rut^{4h_after_social}|rut^isolatedp = 0.598, rut^{1h_after_social}|rut^isolatedp = 7.52e-10).(C) Scheme for the time course of the sleep protocol: flies were either isolated or grouped after eclosion for 7, 6 or 4 days and subsequently isolated for 0, 1 or 3 days (named as socialized, 6 + 1, 4 + 3 and constant isolation); after introducing them in ethoscopes, sleep behavior was recorded for 3 days. (D) Sleep profile of animals isolated for 1 to 4 days, using isolated flies as control. Total number of days in isolation for E-I is depicted in the panel. (E) Quantification of sleep from ZT0 to ZT4 for day 1–4 and flies under constant isolation (CI); Kruskal-Wallis chi-squared = 44.32, df = 7, p-value = 1.85e-07; post hoc Dunn comparisons: wt^1d_isolation|wt^CIp = 4.20e-05, wt^2d_isolation|wt^CIp = 5.51e-04, wt^3d_isolation|wt^CIp = 5.65e-03, wt^4d_isolation|wt^CIp = 0.399. (F-H) Analysis of bout length (H), total number of bouts (G) and latency to first bout (H) from ZT0 to ZT12 for day 1–4 and animals under CI. (F) Kruskal-Wallis chi-squared = 18.47, df = 7, p-value = 0.01, post hoc Dunn comparisons: wt^1d_isolation|wt^CIp = 0.409, wt^2d_isolation|wt^CIp = 0.084, wt^3d_isolation|wt^CIp = 0.443, wt^4d_isolation|wt^CIp = 0.414. (G) Kruskal-Wallis chi-squared = 35.44, df = 7, p-value = 9.23e-06, post hoc Dunn comparisons: wt^1d_isolation|wt^CIp = 2.63e-05, wt^2d_isolation|wt^CIp = 8.37e-04, wt^3d_isolation|wt^CIp = 2.50e-02, wt^4d_isolation|wt^CIp = 0.164. (H) Kruskal-Wallis chi-squared = 36.16, df = 7, p-value = 6.75e-06, post hoc Dunn comparisons: wt^1d_isolation|wt^CIp = 1.93e-04, wt^2d_isolation|wt^CIp = 3.88-03, wt^3d_isolation|wt^CIp = 0.70e-04, wt^4d_isolation|wt^CIp = 0.207.

Fig. 3

Anesthesia abolishes socialization effects on sleep and food consumption. (A) Scheme of the cold-shock protocol (twice per day). (B) Quantification of food consumption using sCAFE (Kruskal-Wallis chi-squared = 15.954, df = 3, p-value = 1.16e-3; post hoc Dunn comparisons: non-shocked^social|non-shocked^isolatedp = 5.26e-3, shocked^social|shocked^isolatedp = 1.00). (C) sleep profile and (D) sleep quantification of the 24–28 h time window (Kruskal-Wallis chi-squared = 31.184, df = 3, p-value = 7.78e-07; post hoc Dunn comparisons: non-shocked^social|non-shocked^isolatedp = 3.05e-06, shocked^social|shocked^isolatedp = 0.116) of cold-shocked socialized and isolated wt flies, together with non-shocked control wt flies.

Fig. 4

Socialization correlates with cellular and synaptic plasticity. (A) Representative confocal images of CAMEL tool for wt and rut mutant MB, either socialized or isolated. Only one representative MB is shown. (B) Number of CREB GFP-positive cells in the MB of socialized or isolated wt and rut mutant animals after 5 days of socialization. Kruskal-Wallis chi-squared = 33.735, df = 3 p-value = 1.93e-05; post hoc Dunn key comparisons: 5 days: wt^social|wt^isolatedp = 6.01e-05, rut^social|rut^isolatedp = 1.00. (C) Example of CAMEL tool (MB cells marked by GFP) combined with the pre-sinaptic marker brp-cherry after 5 days of socialization for wt and rut mutant animals, either socialized of isolated. (D) Quantification of the number of synapses after either isolation or socialization in both wt and rut mutant flies (see fig S4 for a detail on the quantification). Kruskal-Wallis chi-squared = 9.7691, df = 3, p-value = 0.021; post hoc Dunn key comparisons: wt^social|wt^isolatedp = 2.51 e-02, rut^social|rut^isolatedp = 1.00.