Skip to main content
PMC home page
eLife logoLink to eLife
. 2020 May 5;9:e55863. doi: 10.7554/eLife.55863

Whole-brain mapping of socially isolated zebrafish reveals that lonely fish are not loners

Hande Tunbak 1, Mireya Vazquez-Prada 1, Thomas Michael Ryan 1, Adam Raymond Kampff 2, Elena Dreosti 1,✉
Editors: Peggy Mason3, Catherine Dulac4
PMCID: PMC7282805  PMID: 32366356

Abstract

The zebrafish was used to assess the impact of social isolation on behaviour and brain function. As in humans and other social species, early social deprivation reduced social preference in juvenile zebrafish. Whole-brain functional maps of anti-social isolated (lonely) fish were distinct from anti-social (loner) fish found in the normal population. These isolation-induced activity changes revealed profound disruption of neural activity in brain areas linked to social behaviour, social cue processing, and anxiety/stress. Several of the affected regions are modulated by serotonin, and we found that social preference in isolated fish could be rescued by acutely reducing serotonin levels.

Research organism: Zebrafish

eLife digest

Socialising is good for people’s mental health and wellbeing. The connections and relationships that we form can make us more resilient and healthier. Researchers also know that prolonged periods of social isolation, and feeling lonely, can be detrimental to our health, especially in early childhood. The paradox is that loneliness often results in an even lower desire for social contact, leading to further isolation. But not everyone craves social contact. Some people prefer to be alone and feel more comfortable avoiding social interaction.

Zebrafish display the same social preferences. This, along with their transparent brains, makes them a useful model to study the links between social behaviour and brain activity. Like humans, zebrafish are social animals, with most fish taking a strong liking to social interactions by the time they are a few weeks old. A small number of ‘loner’ fish, however, prefer to avoid interacting with their siblings or tank mates. And so, if loneliness quells the desire for more social contact, the question becomes, does isolation turn otherwise social fish into loners?

Here, Tunbak et al. use zebrafish to study how social isolation changes brain activity and behaviour. Social fish were isolated from others in the tank for a few days. These so-called ‘lonely fish’ were then allowed back in contact with the other fish. This revealed that, after isolation, previously social fish did avoid interacting with others.

With this experimental set-up, Tunbak et al. also compared the brains of lonely and loner fish. When fish that prefer social interaction were deprived of social contact, they had increased activity in areas of the brain related to stress and anxiety. These lonely fish became anxious and very sensitive to stimuli; and their brain activity suggested that social interaction became overwhelming rather than rewarding. Positively, the lonely fish quickly recovered their normal, social behaviour when given a drug that reduces anxiety.

This work provides a glimpse into how human behaviour could be affected by lengthy periods in isolation. These results suggest that humans could feel anxious upon returning to normal life after spending a long time alone. Moreover, the findings show the impact that social interaction and isolation can have on the young, developing brain.

Introduction

Social preference behaviour, the drive for individuals to identify and approach members of their own species (Rogers-Carter et al., 2018; Winslow, 2003), is a fundamental component of all social behaviour. We previously found that most zebrafish develop a strong social preference by 2–3 weeks of age (Dreosti et al., 2015), yet we also found a small number (~10%) of ‘loner’ fish that were averse to social cues. A similar diversity of individual social preferences has been found in many species, including humans (Sloan Wilson et al., 1994). Loneliness, undesired isolation from social interaction, has been linked to a reduction in social preference (Engeszer et al., 2004; Shams et al., 2018). We therefore asked whether the socially-averse loner fish found in the normal population would show a similar behavioural phenotype and neuronal activity to socially-averse lonely fish raised in isolation. To answer this question, we compared the behavioural and functional responses of isolated fish to controls during viewing of conspecifics. This comparison found that isolation induces patterns of brain activity that are not present in the normal population. We then asked if we could rescue the aversive behaviour of isolated fish. Since some of the highly activated areas in isolated fish are serotoninergic, we used Buspirone, a 5HT1A receptor agonist. These findings will have important implications for how we understand and treat the impact of social isolation.

Prolonged periods of social isolation are particularly detrimental to humans during early development. However, even brief periods of social isolation have been shown to impact mental and physical health. We therefore tested two models of social isolation, Full (fish raised completely without social interaction) and Partial (fish isolated for 48 hr prior to behavioural testing). Each experiment comprised two sessions, 15 min of acclimation to the chamber followed by 15 min of exposure to two size matched sibling fish that were not isolated. To quantify social preference, we calculated a visual preference index (VPI) that compares the amount of time fish spend in the chamber nearest the conspecifics versus the opposite chamber where they are visually isolated from social cues (see Materials and methods). Full social isolation (Fi) caused a significant decrease in social preference relative to normally raised sibling controls (C) (Figure 1A, left and middle panel: C vs Fi, p=8.3e−8, Mann-Whitney). Specifically, there was an increase in the number of individuals that had a large negative VPI. We therefore decided to divide the fish into three sociality groups: a) anti-social (-S) fish with VPIs below −0.5; b) pro-social (+S) fish with VPIs above +0.5; c) non-social fish with −0.5 < VPI < +0.5. Fish that underwent Partial isolation (Pi), exhibited an intermediate, yet highly significant, change in social preference (Figure 1A, right panel: C vs Pi, p=2.5e−8, Mann-Whitney).

Figure 1. Isolation alters social preference behavior and swimming activity.

(A) Histograms of all the VPIs during the social cue period across different conditions: controls (C, left), full isolation (Fi, middle), and partial isolation (Pi, right). For visual clarity, red bars highlight strong pro-social fish (+S, VPIs > 0.5), blue bars anti-social fish(-S, VPIs < -0.5), and gray non-social fish (ns, -0.5 < VPI < +0.5). (B) Swarm plots comparing the activity levels of fish during the acclimation period expressed as percent time moving (C, n=380; Fi, n=47; Pi, n=157). Mean and standard errors are shown. (C) Swarm plots comparing the activity levels of anti-social (left) and social (fish) fish during visual social cue exposure for each rearing condition (C (-S), n=39; Fi (-S), n=21; Pi (-S), n=53) or (C (+S), n=193; Fi, n=11; Pi (+S), n=57). (D) Time projection through the video of a pro-social control, C(+S), and a fully isolated, Fi(+S), fish during social cue exposure. The dashed lines mark the division between the social cue side (SC) and the side without social cues (No SC) that was used to calculate VPI.

Figure 1.

Figure 1—figure supplement 1. Isolation alters social preference behaviour and swimming activity.

Figure 1—figure supplement 1.

(A) Swarm plots comparing the activity levels of fish during the social period expressed as percentage time moving for each rearing condition (C, n = 380; Fi, n = 47; Pi, n = 157). Mean and standard errors are shown. (B) Swarm plots comparing the activity levels of anti-social (-S) and pro-social (+S) fish during visual social cue exposure for each rearing condition (anti-social C n = 39; Fi n = 21; Pi n = 53) or social (C n = 193; Fi n = 11; Pi n = 57). (C) (Left) Swarm plots comparing freezes (movement pauses longer than 3 s) for the same groups in A during the acclimation period expressed as percentage time freezing for each rearing condition. (Right) Swarm plots comparing freezes for the separated anti-social (-S) and pro-social (+S) groups in B during the acclimation period. (D) Same presentation as C for the social cue period.
Open in a new tab

Video 1. Example of a control and a fully isolated +S fish video during social cue presentation.

Download video file (1.5MB, mp4)
Open in a new tab

Two minutes of behaviour is shown in 20 s (6x playback acceleration). The control fish shows a strong social preference for the social cue and has a stereotypical social phenotype (left). The test fish spends most of its time watching the social cue with a 45-degree angle and synchronizing its bout motion with the other two conspecifics. The fully isolated fish spends long periods of time as well on the side of the conspecifics. Its behaviour, however, its characterized by long pauses while watching the conspecifics (right).

As previously reported (Zellner et al., 2011), we found that fish raised in isolation were significantly less active than their normally raised siblings during the acclimation period (Figure 1B: C vs Fi, p=9.0e−6; C vs Pi, p=2.8e-9 Mann-Whitney) and during the social viewing session (Figure 1—figure supplement 1A: left C vs Fi, p=0.0001; C vs Pi, p=0.004 Mann-Whitney). We then divided fish into groups based on their social preference. Interestingly, anti-social fully and partially isolated fish showed very similar movement activity compared to anti-social controls during the acclimation (Figure 1C left: C (-S) vs Fi (-S), p=0.17 Mann-Whitney; C (-S) vs Pi (-S) p=0.23 Mann-Whitney) and during the social viewing session (Figure 1—figure supplement 1B left: C (-S) vs Fi (-S), p=0.48 Mann-Whitney; C (-S) vs Pi (-S) p=0.10 Mann-Whitney). The pro-social isolated fish, which also exhibited a reduction in activity relative to controls during the acclimation session (Figure 1C right: C (-S) vs Fi (-S), p=8.0e−5 Mann-Whitney; C (-S) vs Pi (-S) p=1.0e−7 Mann-Whitney), instead showed similar activity relative to controls during social viewing (Figure 1—figure supplement 1B right: C (-S) vs Fi (-S), p=0.02 Mann-Whitney; C (-S) vs Pi (-S) p=0.14 Mann-Whitney). In addition, we noticed that all isolated fish behaved qualitatively differently, exhibiting prolonged periods of quiescence (freezing) even when observing conspecifics (Figure 1D and Video 1).

Freezing is a hallmark of anxiety-like behaviour observed in many species, and reported in zebrafish exposed to stressors (Giacomini et al., 2015; Shams et al., 2018), including periods of social isolation (Egan et al., 2009; Shams et al., 2017). In order to quantify freezing behaviour, we measured the percentage of time spent in continuous periods (>3 s) without motion (Figure 1—figure supplement 1C-D). We found that both fully and partially isolated fish exhibited significantly more freezing than controls during the acclimation period (Figure 1—figure supplement 1C left: C vs Fi, p=3.4e−16 Mann-Whitney; C vs Pi, p=2.8e−5 Mann-Whitney), and that this increase relative to controls persisted for fully isolated fish during social viewing, but was reduced in partially isolated fish, perhaps representing some recovery during the 15 min of social interaction (Figure 1—figure supplement 1D left: C vs Fi, p=6.3e−13 Mann-Whitney; C vs Pi, p=0.03 Mann-Whitney). When we compared freezing behaviour of groups with similar social preference, we found, as expected, that anti-social fish exhibited increased freezing during social viewing regardless of rearing condition. However, pro-social fully isolated fish also showed increased freezing during social viewing, suggesting that they were not engaged in typical social interaction, but rather remained immobile on the side with the conspecifics (Figure 1—figure supplement 1D right).

The behavioural similarities between anti-social isolated (lonely) and anti-social control (loner) fish led us to hypothesize that isolation might simply predispose fish to the same anti-social state found in the normal population. If this is the case, neural activity of anti-social isolated and anti-social control fish should be similar when presented with social cues. To test this hypothesis, we performed whole-brain two-photon imaging of c-fos expression, an immediate early gene whose expression is associated with increased neural activity (Herrera and Robertson, 1996), in juvenile brains following testing in the social preference assay. Dissected brains were imaged with the dorsal surface down (bottom-up) to achieve clear views of the ventral brain structures that have been previously implicated in the social brain network (Figure 2A, also see Materials and methods). Volumes of 1.5 mm x 1.5 mm x 700 µm, with a voxel size of 1 × 1×3 µm, were acquired from 135 zebrafish brains across all experimental groups and registered to a reference brain (Marquart et al., 2017). These c-fos whole-brain functional maps were first normalised to a background intensity level (see Materials and methods) and then used to compare the neural activity patterns of different test groups. We compared the average activity map for each rearing/sociality condition with the average map acquired from similarly raised sibling fish that were placed in the behavioural assay for 30 min without any social cues (nsc, no social-cue). The resulting normalised difference stacks (e.g. (+S - nsc)/nsc) allowed us to identify changes in neural activity associated with exposure to a visual social cue (Figure 2A).

Figure 2. Functional maps of the social brain in normal and isolated fish.

Figure 2.

Open in a new tab

(A) Schematic of the custom-built two-photon microscope used for acquiring whole-brain volumes of dorsal-down mounted fish brains (top panel). Horizontal sections of pro-social control fish (C(+S)) responses at increasing imaging depth (lower panels). Images are average differences between (C(+S)) and siblings not presented with a social cue. Positive values (white) indicate increased cFos expression in socially preferring fish, while negative values (black) indicate decreased expression. Scale bar is 200μm. The intensity scale bar is shown in B, C(+S) row. (B) Region analysis of two different brain areas that have been implicated in social behavior: caudal hypothalamus and preoptic area. A schematic of the anatomical regions and corresponding DAPI staining is shown (top panel) with two sub-regions highlighted in green. Images showing changes in cFos activation in these areas for pro- (+S) and anti-social (-S) controls, fully isolated, and partially isolated fish are shown. Images are horizontal sections of the average difference between each test group and their corresponding sibling group not presented with a social cue. Scale bar is 100μm. Intensity scale bar is shown for each group. (C) Average image of TH1, TH2, Slc6a4b, and DAT expression in the same section of the caudal hypothalamus as 2B (n=3 each). Scale bar is 100μm. (D) Summary graphs showing the change in cFos activation for four different brain areas calculated by using the average difference images shown in (B) and using 3D masks (a single plane of each area of the masks is shown in green in B). Positive values indicate increases in cFos expression; asterisks mark significant changes relative to no social cue siblings. D=dorsal and V=ventral caudal hypothalamus; Pa=ventrolateral preoptic area, PM=dorsal preoptic area.

Several brain areas showed strong activation or inhibition in normally raised fish upon social cue exposure. We focused on areas that have been reported as social brain areas (O'Connell and Hofmann, 2011) and show differences between our experimental groups (Figure 2B: C (+S and -S)). The caudal hypothalamus was differentially activated in pro- vs. anti-social control fish. A dorsal sub-region was significantly activated in pro-social controls (Figure 2B and D: dHc - C (+S) vs C (nsc), p=0.007, Mann-Whitney), whereas it was inhibited, along with the adjacent ventral sub-region, in anti-social controls (Figure 2B and D: vHc - C (-S) vs C (nsc), p=0.003, Mann-Whitney). The caudal hypothalamus is known to express high levels of serotonin and dopamine, as well as glutamate and histamine (Filippi et al., 2010; Kaslin and Panula, 2001). Furthermore, a segregation into distinct dorsal and ventral areas of the caudal hypothalamus has already been shown for some of these markers, such as tyrosine hydroxylase 1 and 2, (Th1 and Th2) (Yamamoto et al., 2010) and we confirmed these previous results with immunostaining (Figure 2C left), as well as for the dopamine and serotonin transporters, DAT and slc6a4b (Figure 2C right) (Filippi et al., 2010; Lillesaar, 2011). Changes in serotonin and dopamine levels have been widely documented in response to social interaction (Scerbina et al., 2012), viewing social cues (Saif et al., 2013), and social isolation (Huang et al., 2015; Shams et al., 2018; Shams et al., 2015). While the serotoninergic system has been linked to stress and arousal (Backström and Winberg, 2017), the dopamine circuitry has been shown to regulate the reward system underlying social behaviour (Teles et al., 2013). Since the caudal hypothalamus expresses both of these neurotransmitters, and our data demonstrate a pattern of activation/inhibition that is distinct for pro- and anti-social fish, then this area could be crucial in regulating social preference.

The second social brain area we investigated was the preoptic area. Our data showed a similar activation pattern for anti-social and pro-social fish characterised by a small increase in the dorsal preoptic area (dPa) and a small decrease in the ventral preoptic area (vPa). However, only anti-social control fish showed a significant change in the ventral area (Figure 2B and D: C (-S) vs C (nsc), vPa p=0.003, Mann-Whitney). The activation of the preoptic area during social behaviour is consistent with previous literature in a number of species (O'Connell and Hofmann, 2011). This area has been shown to express several neuropeptides involved in social behaviour such as arginine/vasotocin and oxytocin (Heinrichs et al., 2009; Herget and Ryu, 2015). It was recently shown that oxytocin does not seem to be responsible for social interaction (Ribeiro et al., 2019) as mutants for oxytocin receptors shows no alteration in social preference, but rather reduced social recognition. Furthermore, injections of oxytocin do not have any effect on shoaling and interaction (Langen et al., 2015). The neuropeptide vasotocin, instead, has been shown to have a specific effect on reducing social interaction (Langen et al., 2015) and not shoaling behaviour. This neuropeptide has also been shown to be involved in aggression (Teles et al., 2016) and stress by stimulating cortisol release.

We then compared the brain activity maps of anti- and pro-social control fish with fully and partially isolated fish. As described previously, anti-social control (loner) fish showed a behavioural phenotype very similar to anti-social isolated (lonely) fish. Therefore, we investigated whether their brain activity maps were also similar following the presentation of a social cue. Contrary to our hypothesis, c-fos functional maps of anti-social fully isolated fish (Figure 2B: Fi (-S)) revealed a completely different activity profile than their anti-social sibling controls (Figure 2B: C (-S)). The ventral sub-region of the caudal hypothalamus (vHc) of Fi (-S) fish was not inactivated, while the preoptic area was strongly activated in both the dorsal (dPa) and the ventral (vPa) regions, but significantly only in the dorsal (Figure 2B and D: Fi (-S) vs Fi (nsc), p=0.006 dPa; p=0.07 vPa, Mann-Whitney). Furthermore, the pro-social fully isolated fish (Figure 2B: Fi (+S)), who exhibited an increase of freezes and reduced motility compared to control fish when viewing conspecifics, showed a similar activation to pro-social controls in the caudal hypothalamus, but increased activity in the dorsal preoptic area. Interestingly, the preoptic area was activated differently in pro-social and anti-social isolated fish, with only the dorsal preoptic area strongly activated in the pro-social group (Figure 2B and D: Fi (+S) vs Fi, p=0.04 vPa, p=0.002 dPa, Man-Whitney). These data suggest that long social isolation causes abnormal neural responses during viewing of social cues.

Furthermore, anti- and pro-social fish exposed to a brief isolation for only 48 hr prior to testing, showed similar functional activity changes to fully isolated fish, albeit less strong (Figure 2B and D: Pi (-S) vs Pi (nsc), p=0.18 dHc; p=0.28 vHc; p=0.04 vPa; p=0.04 dPa, Mann-Whitney; Figure 2B and D: Pi (+S) vs Pi (nsc), p=0.17 dHc; p=0.05 vHc; p=0.007 vPa; p=0.006 dPa). Together with the behavioural data, this finding supports the idea that short term isolation is enough to induce brain activity changes similar to those observed following complete isolation, and strikingly different than those observed in anti-social controls.

We were next interested in understanding why social isolation promotes social aversion instead of increasing the drive for social interaction. An important clue was found in the pattern of brain activity changes that were unique to isolated fish. When we directly compared the normalised c-fos functional brain maps of isolated and control fish that were not exposed to social cues during the assay (Figure 3A), we found a significant increase in two interesting areas; one associated with visual processing, the optic tectum, [McDowell et al., 2004]), and one involved in stress responses, the posterior tuberal nucleus (Ziv et al., 2013).

Figure 3. Changes in baseline brain activity following isolation.

Figure 3.

Open in a new tab

(A) Images of two areas that show strong c-fos activation in fully isolated fish independent of social stimuli (optic tectum and posterior tuberal nucleus (PTN)). Schematics of the horizontal sections and corresponding DAPI image are shown in the left panels. One plane of the 3D mask regions used for subsequent analysis is indicated (green). Images of Fully isolated fish c-fos neuronal activity, calculated as average differences between fully isolated (Fi) fish and normally raised fish without social cues (nsc) are shown in the right panels. Scale bar 200 μm. (B) Summary graphs showing the normalised c-fos expression in the optic tectum and PTN 3D masks for each experimental condition: non social cue (nsc), pro-social (+S) and anti-social (-S) for all the controls (C), fully isolated (Fi), and partially isolated (Pi) fish.

In pro-social control fish, viewing social cues resulted in a significant increase of neuronal activity in the optic tectum (Figure 3B top: C (+S) vs C (nsc), p=0.004 Mann-Whitney). However, in fully isolated fish, there was already increased neuronal activity in the optic tectum in the absence of social cues (Figure 3B top: Fi (nsc) vs C (nsc), p=0.0004, Mann-Whitney), suggesting that isolation increases visual sensitivity, as previously reported in humans (Cacioppo et al., 2015). This increased sensitivity of fully isolated fish not presented with social cues was weaker in partially isolated fish (Figure 3B top: Pi (nsc) vs C (nsc), p=0.03, Mann-Whitney). However, a much larger increase in tectal activity was observed when pro-social partially isolated fish viewed conspecifics, revealing that some visual sensitization had occurred (Figure 3B top: Pi (+S) vs C (+S), p=0.0002, Mann-Whitney). In addition, increased tectal activity was also present in both fully and partially isolated anti-social fish (Figure 3B top: Fi (-S) vs C (-S), p=0.048; Pi (-S) vs C (-S), p=0.005, Mann-Whitney), even though these fish largely avoided the chamber with visual access to conspecifics.

We also observed isolation-related activity increases in the posterior tuberal nucleus, an area associated with stress responses in zebrafish (Wee et al., 2019; Ziv et al., 2013). Full isolation caused a significant increase in posterior tuberal nucleus activity in the absence of social cues (Figure 3B bottom: Fi (nsc) vs C (nsc), p=0.015, Mann-Whitney) and in both anti-social and pro-social fish exposed to social cues (Figure 3B bottom: Fi (+S) vs C (+S), p=0.003; Fi (-S) vs C (-S), p=0.016, Mann-Whitney). Following partial isolation, posterior tuberal nucleus activity was not increased in the absence of social cues (Figure 3B bottom: Pi (nsc) vs C (nsc), p=0.29, Mann-Whitney), only slightly in pro-social fish (Figure 3B bottom: Pi (+S) vs C (+S), p=0.018), but significantly so in anti-social fish (Figure 3B bottom: Pi (-S) vs C (-S), p=0.0005).

Given these results from the optic tectum and posterior tuberal nucleus, we propose that isolation initially heightens sensitivity to social stimuli. However, when prolonged, this heightened sensitivity results in an increase of stress and anxiety levels during social viewing that leads to an aversion for social stimuli.

To test our hypothesis that reducing anxiety could reverse the anti-social behaviour observed in isolated zebrafish, we acutely treated control and partially isolated fish with Buspirone, an agonist of the auto-inhibitory 5HT1A receptor. The choice of isolation duration was motivated by the intermediate behavioural and functional phenotype of partial isolation relative to normal-rearing and full isolation, which would allow us to more easily detect both positive and negative impacts of treatment on sociality. The choice of Buspirone was supported by the changes in activity observed in the caudal hypothalamus of isolated fish, and by the fact that the caudal hypothalamus and the preoptic area strongly express Htr1ab receptors, one of the two orthologues of the 5HT1A receptor (Norton et al., 2008). Buspirone has been shown to reduce anxiety in humans, mice, and zebrafish (Bencan et al., 2009; Lalonde and Strazielle, 2010; Lau et al., 2011; Patel and Hillard, 2006). While it is not fully understood how Buspirone reduces anxiety, it has been shown to enhance social interaction of rats (File and Seth, 2003; Gould et al., 2011), sociability of zebrafish (Barba-Escobedo and Gould, 2012), and reduce social phobia in humans (Schneier et al., 1993; van Vliet et al., 1997). Its ability to counter the effects of social isolation in zebrafish has not been investigated.

We first tested the effects of acute exposure to Buspirone in control fish, and, as expected, we observed a small significant increase in social preference relative to untreated controls, however, a population of ~10% anti-social fish remained (Figure 4—figure supplement 1; C (no drug) vs C (30 µM), p=0.01, Mann-Whitney). We then treated partially isolated fish with 30 µM and 50 µM (Figure 4—figure supplement 1, n = 46, n = 72 fish) of Buspirone. Remarkably, the acute drug treatment was sufficient in both concentrations to reverse the anti-social phenotype caused by isolation (Figure 4A; Pi vs Pi (Buspirone 30 µM and 50 µM combined), p=2.56 e-05, Mann-Whitney).

Figure 4. Buspirone rescues social preference in isolated fish.

(A) Histogram of VPIs during the social cue period in partially isolated (Pi) fish treated with 30 μM and 50 μM of Buspirone (combined). For visual clarity, the bars are colored as in Figure 1. (B) VPI values calculated in one-minute time bins for controls (C, black line, n=380), partial isolated (Pi, blue line, n=157), and Pi treated with Buspirone (Pi+B, green line, n=118). Note how Buspirone treated fish recover normal social preference within the first 5 minutes. (C) Percentage of time moving calculated in one-minute bins for the same fish as B, thin lines indicate standard error.

Figure 4.

Figure 4—figure supplement 1. Buspirone rescues social preference in isolated fish.

Figure 4—figure supplement 1.

(A) Histograms of VPIs during the social cue period in control fish treated with 30 μM of Buspirone and in partially isolated fish treated with 30 μM and 50 μM of Buspirone. The bars are coloured as in Figure 1. (B) VPI values calculated in one-minute time bins for controls treated with 30 μM of Buspirone, Pi fish treated with 30 μM, and Partially isolated fish treated with 50 μ M of Buspirone, thin lines indicate standard error.
Open in a new tab

When we then compared the time course of this phenotype reversal by computing the VPIs for each minute throughout the 15 min of the behavioural experiment (Figure 4B). We found that the isolated fish treated with Buspirone, while initially anti-social, would rapidly recover normal social preference behaviour within the first 5 min of exposure to social cues (Figure 4B: C vs Pi (Buspirone), p=0.016, first minute; p=0.37, fourth minute, Mann-Whitney). In contrast, the VPIs of untreated isolated fish remained significantly lower than controls throughout the entire session. We next compared the time course of movement activity (Figure 4C), and found that it generally increased quickly throughout the first 5 min of the social viewing session. Notably, the activity of isolated fish treated with Buspirone was already at the level of controls from the start of the social viewing session (Figure 4B: C vs Pi (Buspirone), p=0.31, first minute, Mann-Whitney), which suggests that the recovery of normal movement activity, possibly as a result of reduced anxiety, precedes the recovery of normal social preference. Therefore, Buspirone’s impact on the rate of recovery of social preference indicates that it may do so by reducing anxiety, perhaps at the level of the preoptic and/or caudal hypothalamic area, allowing circuit plasticity to down-regulate the hypersensitivity to social stimuli acquired during the isolation period.

In summary, our results demonstrate that lonely fish, which have been isolated from social cues and show anti-social behaviour, have a completely different functional response to social stimuli than loner fish, anti-social fish found in the normal population. In addition, the functional changes caused by social deprivation are consistent with an increase in anxiety resulting from hyper-sensitization to social stimuli, similar to the effects of isolation on humans. We could reverse isolation’s effects in zebrafish with an existing anxiolytic drug that acts on the monoaminergic system. Zebrafish will thus provide a powerful new tool for studying the impact of loneliness (isolation) on brain function and exploring different strategies for reducing, or even reversing, its harm.

Materials and methods

Key resources table.

Reagent type
(species) or resource		 Designation Source or reference Identifiers Additional
information
Antibody Anti- digoxigenin-POD, sheep, polyclonal Fab fragments Sigma-Aldrich, Rouche Roche, Cat# 11207733910, RRID:AB_514500 1:3000
Sequence-based reagent cFos _F This paper PCR primers CCGATACACTGCAAGCTGAA
Sequence-based reagent cFos_R This paper PCR primers ATTGCAGGGCTATGGAAGTG
Peptide, recombinant protein Proteinase K Sigma-Aldrich Cat# P6556-10MG 2 mg/ml
Commercial assay TSA Plus Cyanine three system Sigma-Aldrich, Perkin Elmer Cat# NEL74401KT Dilution 1:50
Chemical compound, drug Buspirone hydrochloride Sigma-Aldrich Cat# B7148-1G 30 uM and 50 uM
Software, algorithm Anaconda, Spyder Anaconda (https://www.anaconda.com/) Spyder, RRID:SCR_017585 Version 4.0.1
Software ImageJ NIH
(http://imagej.nih.gov/ij/)		 RRID:SCR_003070
Software ANTs- Advanced Normalisation Tools http://stnava.github.io/ANTs/ RRID:SCR_004757 Version 2.1.0
Other DAPI staining Sigma-Aldrich Cat#
D9564-10MG		 1 mg/ml
Other Slc6a4b RNA probe Norton et al., 2008
Other DAT RNA probe Filippi et al., 2010
Other Th1 RNA Probe Filippi et al., 2010
Other Th2 RNA probe Filippi et al., 2010
Open in a new tab

Animals

AB strain zebrafish maintenance and breeding was performed at 28.5C with a 14 hr:10 hr light-dark cycle. Isolated fish were housed in custom chambers (length = 15 cm, width = 5 cm, height = 10 cm) made of opaque white acrylic with translucent lids, either from fertilization (full isolation) or for 48 hr prior to the behavioural experiment (partial isolation). All experiments were performed according to protocols approved by local ethical committee (AWERB Bloomsbury Campus UCL) and the UK Home Office.

Behavioural assay and analysis

Experimental details and image acquisition were performed as described previously (Dreosti et al., 2015). Fish were positioned in custom-built behavioural arenas (Figure 1D) made of white acrylic, and illuminated with visible light using a laser light projector (Microvision, ShowwX+, US). The videography system comprised a high-speed camera (Flea3, PointGrey, CA), an infrared light (Advanced Illumination, US, 880 nm), an IR filter (R70, Hoya, JP), and a vari-focal lens (Fujinon, JP). Experiments were recorded using custom written workflows in Bonsai (Langen et al., 2015). Test fish were positioned in the main C-shape compartment of the arena by pipetting, and left for 15 min to acclimate. A social cue, two fish of same age and similar size, was then introduced into one of the two adjacent chambers randomly. Test fish could see the social cue through a glass window. Each fish was run only once in the behavioural assay.

Images were analysed using custom written computer vision scripts in Python based on OpenCV (https://www.dreo-sci.com/resources/). Each frame was cropped, background subtracted, and thresholded. The centroid, position, orientation, and per frame motions of the test fish were identified, and stored in a CSV file. All videos have been saved with H.264 compression for subsequent offline analysis, and are available upon request.The source code can be downloaded at http://www.dreo-sci.com/resources/.

The visual preference index (VPI) was calculated by subtracting the number of frames in which the fish was located on the side of the arena nearest the social stimulus (Social cue (SC) side in Figure 1B) by the number of frames located on the opposite side of the arena (nsc (No SC) side). This difference was then divided by the total number of frames recorded [VPI = (SC – No SC)/Total frames]. The percent time moving was calculated by counting each frame with detectable changes in the fish image relative to the previous frame (i.e. motion), and dividing by the total number of frames. The percent time freezing was calculated by detecting contiguous sequences without motion longer than 3 s, counting all frames that are part of such sequences, and dividing by the total number of frames.

Whole mount in situ hybridisation

Fluorescent in situ hybridizations using digoxigenin-labelled c-fos were performed on dissected juvenile zebrafish with few modification to the original method (Brend and Holley, 2009). After overnight fixation in 4% PFA, protein K treatment (2 mg/ml 20 min of incubation), inactivation of endogenous peroxidase with H2O2 (22% v/v for 30 min at room temperature), additional fixation (30 min at room temperature) and 3 hr of incubation with the hybridisation buffer, fish were incubated with the c-fos probe (courtesy from Ricardo N. Silva (Forward CCGATACACTGCAAGCTGAA and Reverse ATTGCAGGGCTATGGAAGTG), or with dopamine transporter (DAT), tyrosine hydroxylase 1 (Th1), tyrosine hydroxylase (Th2) (Filippi et al., 2010), or the 5‐HT transporter, solute carrier family 6 member 4b (Slc6a4b) probes (Norton et al., 2008). C-fos, DAT and Slc6a4b probes were detected with anti-Digoxigenin-POD, Fab fragments (Roche, 1:3000) and TSA Plus Cyanine 3 System (Perkin Elmer, 1:50). Nuclear staining was obtained using DAPI (Sigma-Aldrich, 1: 500). Fish were then mounted for imaging in low melting point agarose (2.5% Agarose, 0.8% glycerol, PBS-Tween) and imaged.

Imaging and registration

A custom built two-photon microscope (INSS) was used for image acquisition of whole-brain in situs. Both DAPI and Cy3 Images were collected with a 10x objective (Olympus, W Plan-Apochromat 10x/0.5 M27 75 mm) using a ‘Chameleon’ titanium–sapphire laser tuned to 1030 nm (Coherent Inc, Santa Clara, CA, US) and controlled using custom written software in LabView. Registration of in-situ images was performed using ANTs (Advanced Normalisation Tools) version 2.1.0 running on the UCL Legion compute cluster. Images were down-sampled to 512*512 and parameters were slightly modified from Marquart et al. (2017) fixed registration:

antsRegistration -d 3 --float 1 -o [Registered_Image_, Registered_Image _warped.nii.gz] --interpolation WelchWindowedSinc --use-histogram-matching 0 r [reference_Image, Registered_Image,1] -t rigid[0.1] -m MI[reference_Image, Registered_Image _0.nii,1,32,Regular,0.25] -c [1000 × 500×250 × 100,1e-8,10] --shrink-factors 12 × 8×4 × 2 s 4 × 3×2 × 1 t Affine[0.1] -m MI[reference_Image, Registered_Image,1,32,Regular,0.25] -c [1000 × 500×250 × 100,1e-8,10] --shrink-factors 12 × 8×4 × 2 s 4 × 3×2 × 1 t SyN[0.1,6,0] -m CC[reference_Image, Registered_Image _0.nii,1,2] -c [1000 × 500×500x250 × 100,1e-7,10] --shrink-factors 12 × 8×4x2 × 1 s 4 × 3×2x1 × 0
antsApplyTransforms -d 3 v 0 --float -n WelchWindowedSinc -i Registered_Image _1.nii -r reference_Image -o Registered_Image _warped_red.nii.gz -t Registered_Image _1Warp.nii.gz -t Registered_Image _0GenericAffine.mat

Intensity normalisation

The registered image stacks were then normalised to adjust for intensity variations between imaging sessions caused by a variety of sources (staining efficiency, laser power fluctuations, light detector sensitivity, etc.). Normalisation was accomplished by computing an intensity histogram for each fish brain’s volume (with 10000 discrete intensity bins spanning the range −4000.0 to 70000.0) for all 512*512*273 voxels. The minimum value bin (with at least 100 voxels) was used as the bias offset, and subtracted from all voxel values. The mode value, minus the bias, provided a robust estimate of the background/baseline fluorescence and was thus used to normalise voxel values for the entire volume. Therefore, after normalisation, an intensity value of 1 reflected the background level while two indicates fluorescence level that is twice the background, and so on. Histogram normalisation was performed for each individual fish’s brain volume prior to any region or voxel-based analysis.

Figures 2B and 3A Reconstruction of cross section images were obtained by using the Fiji ‘Volume viewer’ plugin. Schematics of cross- and horizontal-section were obtained by using the ‘Neuroanatomy of the zebrafish brain’.

Figures 2D Percentages of c-fos activation were calculated for each of the six different areas highlighted in Figures 2B and 3A, using custom written Python functions, in the following way. A 3D mask for each area was generated by using the ‘Segmentation Editor’ plugin Fiji (https://imagej.net/Segmentation_Editor). C-fos percentage values for each condition (C (+S), C (-S), Fi (-S), Pi (-S)) were obtained by subtracting and then dividing each c-fos average value of the mask by the basal c-fos average value calculated in control fish No Social Cue.

Statistics

Statistical analysis was performed using Python scipy stats libraries. Since VPI, percent time moving/freezing, and c-fos activity distributions were generally not normally distributed, we used the non-parametric Mann-Whitney U-test of independent samples for hypothesis testing throughout the manuscript.

Drug treatment

Juvenile fish were treated with 30 µM or 50 µM Buspirone (Buspirone HCl, Sigma) for 10 min prior the experiment. After washing, fish were run through the behavioural assay. Each fish was used only once.

Data availability

All the images, video, protocols, analysis scripts, and data that support the findings of this study are available from this website (http://www.dreo-sci.com/resources/), or our GitHub repository (https://github.com/Dreosti-Lab/Lonely_Fish_2020; Dreosti, 2020; copy archived at https://github.com/elifesciences-publications/Lonely_Fish_2020), or from the corresponding author upon request.

Acknowledgements

The authors acknowledge Steve Wilson for providing lab resources. Ricardo Neto Silva for the c-fos probe. Jason Rihel for some of the reagents. Wolfgang Driver for the dopaminergic probes and William Norton for the serotoninergic probes. UCL fish facility team for fish care/husbandry. This work was supported by the Wellcome Trust Grant Ref 202465/Z/16/Z.

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Contributor Information

Elena Dreosti, Email: e.dreosti@ucl.ac.uk.

Peggy Mason, University of Chicago, United States.

Catherine Dulac, Harvard University, United States.

Funding Information

This paper was supported by the following grants:

  • Wellcome 202465/Z/16/Z to Elena Dreosti.

  • Gatsby Charitable Foundation 090843/F/09/Z to Adam Raymond Kampff.

Additional information

Competing interests

No competing interests declared.

Author contributions

Conceptualization, Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing - review and editing.

Data curation.

Software, Writing - review and editing.

Software, Formal analysis, Visualization, Writing - review and editing.

Conceptualization, Resources, Data curation, Software, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing - original draft, Project administration, Writing - review and editing.

Ethics

Animal experimentation: All experiments were performed according to protocols approved by local ethical committee (AWERB Bloomsbury Campus UCL) and the UK Home Office. PAE2ECA7E.

Additional files

Transparent reporting form

Data availability

All the images, video, protocols, analysis scripts, and data that support the findings of this study are available on our website (www.dreo-sci.com/resources) and GitHub repository (https://github/Dreosti-Lab/Social_Zebrafish), or from the corresponding author upon request.

References

  1. Backström T, Winberg S. Serotonin coordinates responses to social Stress-What we can learn from fish. Frontiers in Neuroscience. 2017;11:595. doi: 10.3389/fnins.2017.00595. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Barba-Escobedo PA, Gould GG. Visual social preferences of lone zebrafish in a novel environment: strain and anxiolytic effects. Genes, Brain and Behavior. 2012;11:366–373. doi: 10.1111/j.1601-183X.2012.00770.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bencan Z, Sledge D, Levin ED. Buspirone, chlordiazepoxide and diazepam effects in a zebrafish model of anxiety. Pharmacology Biochemistry and Behavior. 2009;94:75–80. doi: 10.1016/j.pbb.2009.07.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Brend T, Holley SA. Zebrafish whole mount High-Resolution double fluorescent in situ hybridization. Journal of Visualized Experiments. 2009;25:29. doi: 10.3791/1229. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Cacioppo JT, Cacioppo S, Capitanio JP, Cole SW. The neuroendocrinology of social isolation. Annual Review of Psychology. 2015;66:733–767. doi: 10.1146/annurev-psych-010814-015240. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Dreosti E, Lopes G, Kampff AR, Wilson SW. Development of social behavior in young zebrafish. Frontiers in Neural Circuits. 2015;9:39. doi: 10.3389/fncir.2015.00039. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Dreosti E. Analysis code used in Tunbak et al. 2020. d80c2f4GitHub. 2020 https://github.com/Dreosti-Lab/Lonely_Fish_2020
  8. Egan RJ, Bergner CL, Hart PC, Cachat JM, Canavello PR, Elegante MF, Elkhayat SI, Bartels BK, Tien AK, Tien DH, Mohnot S, Beeson E, Glasgow E, Amri H, Zukowska Z, Kalueff AV. Understanding behavioral and physiological phenotypes of stress and anxiety in zebrafish. Behavioural Brain Research. 2009;205:38–44. doi: 10.1016/j.bbr.2009.06.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Engeszer RE, Ryan MJ, Parichy DM. Learned social preference in zebrafish. Current Biology. 2004;14:881–884. doi: 10.1016/j.cub.2004.04.042. [DOI] [PubMed] [Google Scholar]
  10. File SE, Seth P. A review of 25 years of the social interaction test. European Journal of Pharmacology. 2003;463:35–53. doi: 10.1016/S0014-2999(03)01273-1. [DOI] [PubMed] [Google Scholar]
  11. Filippi A, Mahler J, Schweitzer J, Driever W. Expression of the paralogous tyrosine hydroxylase encoding genes th1 and th2 reveals the full complement of dopaminergic and noradrenergic neurons in zebrafish larval and juvenile brain. The Journal of Comparative Neurology. 2010;518:423–438. doi: 10.1002/cne.22213. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Giacomini AC, de Abreu MS, Koakoski G, Idalêncio R, Kalichak F, Oliveira TA, da Rosa JG, Gusso D, Piato AL, Barcellos LJ. My stress, our stress: blunted cortisol response to stress in isolated housed zebrafish. Physiology & Behavior. 2015;139:182–187. doi: 10.1016/j.physbeh.2014.11.035. [DOI] [PubMed] [Google Scholar]
  13. Gould GG, Hensler JG, Burke TF, Benno RH, Onaivi ES, Daws LC. Density and function of central serotonin (5-HT) transporters, 5-HT1A and 5-HT2A receptors, and effects of their targeting on BTBR T+tf/J mouse social behavior. Journal of Neurochemistry. 2011;116:291–303. doi: 10.1111/j.1471-4159.2010.07104.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Heinrichs M, von Dawans B, Domes G. Oxytocin, vasopressin, and human social behavior. Frontiers in Neuroendocrinology. 2009;30:548–557. doi: 10.1016/j.yfrne.2009.05.005. [DOI] [PubMed] [Google Scholar]
  15. Herget U, Ryu S. Coexpression analysis of nine neuropeptides in the neurosecretory preoptic area of larval zebrafish. Frontiers in Neuroanatomy. 2015;9:2. doi: 10.3389/fnana.2015.00002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Herrera DG, Robertson HA. Activation of c-fos in the brain. Progress in Neurobiology. 1996;50:83–107. doi: 10.1016/S0301-0082(96)00021-4. [DOI] [PubMed] [Google Scholar]
  17. Huang C-C, Lu R-B, Yen C-H, Yeh Y-W, Chou H-W, Kuo S-C, Chen C-Y, Chang C-C, Chang H-A, Ho P-S, Liang C-S, Cheng S, Shih M-C, Huang S-Y. Dopamine transporter gene may be associated with bipolar disorder and its personality traits. European Archives of Psychiatry and Clinical Neuroscience. 2015;265:281–290. doi: 10.1007/s00406-014-0570-0. [DOI] [PubMed] [Google Scholar]
  18. Kaslin J, Panula P. Comparative anatomy of the histaminergic and other aminergic systems in zebrafish (Danio rerio) The Journal of Comparative Neurology. 2001;440:342–377. doi: 10.1002/cne.1390. [DOI] [PubMed] [Google Scholar]
  19. Lalonde R, Strazielle C. Relations between open-field, elevated plus-maze, and emergence tests in C57BL/6J and BALB/c mice injected with GABA- and 5HT-anxiolytic agents. Fundamental & Clinical Pharmacology. 2010;24:365–376. doi: 10.1111/j.1472-8206.2009.00772.x. [DOI] [PubMed] [Google Scholar]
  20. Langen EMA, Lindeyer CM, Reader SM, Swaney WT. Nonapeptide influences on social behaviour: effects of vasotocin and isotocin on shoaling and interaction in zebrafish. Behaviour. 2015;152:897–915. doi: 10.1163/1568539X-00003261. [DOI] [Google Scholar]
  21. Lau BY, Mathur P, Gould GG, Guo S. Identification of a brain center whose activity discriminates a choice behavior in zebrafish. PNAS. 2011;108:2581–2586. doi: 10.1073/pnas.1018275108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Lillesaar C. The serotonergic system in fish. Journal of Chemical Neuroanatomy. 2011;41:294–308. doi: 10.1016/j.jchemneu.2011.05.009. [DOI] [PubMed] [Google Scholar]
  23. Marquart GD, Tabor KM, Horstick EJ, Brown M, Geoca AK, Polys NF, Nogare DD, Burgess HA. High-precision registration between zebrafish brain atlases using symmetric diffeomorphic normalization. GigaScience. 2017;6:1–15. doi: 10.1093/gigascience/gix056. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. McDowell AL, Dixon LJ, Houchins JD, Bilotta J. Visual processing of the zebrafish optic tectum before and after optic nerve damage. Visual Neuroscience. 2004;21:97–106. doi: 10.1017/S0952523804043019. [DOI] [PubMed] [Google Scholar]
  25. Norton WH, Folchert A, Bally-Cuif L. Comparative analysis of serotonin receptor (HTR1A/HTR1B families) and transporter (slc6a4a/b) gene expression in the zebrafish brain. The Journal of Comparative Neurology. 2008;511:521–542. doi: 10.1002/cne.21831. [DOI] [PubMed] [Google Scholar]
  26. O'Connell LA, Hofmann HA. The vertebrate mesolimbic reward system and social behavior network: a comparative synthesis. The Journal of Comparative Neurology. 2011;519:3599–3639. doi: 10.1002/cne.22735. [DOI] [PubMed] [Google Scholar]
  27. Patel S, Hillard CJ. Pharmacological evaluation of cannabinoid receptor ligands in a mouse model of anxiety: further evidence for an anxiolytic role for endogenous cannabinoid signaling. Journal of Pharmacology and Experimental Therapeutics. 2006;318:304–311. doi: 10.1124/jpet.106.101287. [DOI] [PubMed] [Google Scholar]
  28. Ribeiro J, Davids K, Araújo D, Guilherme J, Silva P, Garganta J. Exploiting Bi-Directional Self-Organizing tendencies in team sports: the role of the game model and tactical principles of play. Frontiers in Psychology. 2019;10:213. doi: 10.3389/fpsyg.2019.02213. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Rogers-Carter MM, Varela JA, Gribbons KB, Pierce AF, McGoey MT, Ritchey M, Christianson JP. Insular cortex mediates approach and avoidance responses to social affective stimuli. Nature Neuroscience. 2018;21:404–414. doi: 10.1038/s41593-018-0071-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Saif M, Chatterjee D, Buske C, Gerlai R. Sight of conspecific images induces changes in neurochemistry in zebrafish. Behavioural Brain Research. 2013;243:294–299. doi: 10.1016/j.bbr.2013.01.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Scerbina T, Chatterjee D, Gerlai R. Dopamine receptor antagonism disrupts social preference in zebrafish: a strain comparison study. Amino Acids. 2012;43:2059–2072. doi: 10.1007/s00726-012-1284-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Schneier FR, Saoud JB, Campeas R, Fallon BA, Hollander E, Coplan J, Liebowitz MR. Buspirone in social phobia. Journal of Clinical Psychopharmacology. 1993;13:251???256–256. doi: 10.1097/00004714-199308000-00004. [DOI] [PubMed] [Google Scholar]
  33. Shams S, Chatterjee D, Gerlai R. Chronic social isolation affects thigmotaxis and whole-brain serotonin levels in adult zebrafish. Behavioural Brain Research. 2015;292:283–287. doi: 10.1016/j.bbr.2015.05.061. [DOI] [PubMed] [Google Scholar]
  34. Shams S, Seguin D, Facciol A, Chatterjee D, Gerlai R. Effect of social isolation on anxiety-related behaviors, cortisol, and monoamines in adult zebrafish. Behavioral Neuroscience. 2017;131:492–504. doi: 10.1037/bne0000220. [DOI] [PubMed] [Google Scholar]
  35. Shams S, Amlani S, Buske C, Chatterjee D, Gerlai R. Developmental social isolation affects adult behavior, social interaction, and dopamine metabolite levels in zebrafish. Developmental Psychobiology. 2018;60:43–56. doi: 10.1002/dev.21581. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Sloan Wilson D, Clark AB, Coleman K, Dearstyne T. Shyness and boldness in humans and other animals. Trends in Ecology & Evolution. 1994;9:442–446. doi: 10.1016/0169-5347(94)90134-1. [DOI] [PubMed] [Google Scholar]
  37. Teles MC, Dahlbom SJ, Winberg S, Oliveira RF. Social modulation of brain monoamine levels in zebrafish. Behavioural Brain Research. 2013;253:17–24. doi: 10.1016/j.bbr.2013.07.012. [DOI] [PubMed] [Google Scholar]
  38. Teles MC, Gozdowska M, Kalamarz-Kubiak H, Kulczykowska E, Oliveira RF. Agonistic interactions elicit rapid changes in brain nonapeptide levels in zebrafish. Hormones and Behavior. 2016;84:57–63. doi: 10.1016/j.yhbeh.2016.05.020. [DOI] [PubMed] [Google Scholar]
  39. van Vliet IM, den Boer JA, Westenberg HG, Pian KL. Clinical effects of buspirone in social phobia: a double-blind placebo-controlled study. The Journal of Clinical Psychiatry. 1997;58:164–168. doi: 10.4088/jcp.v58n0405. [DOI] [PubMed] [Google Scholar]
  40. Wee CL, Nikitchenko M, Wang WC, Luks-Morgan SJ, Song E, Gagnon JA, Randlett O, Bianco IH, Lacoste AMB, Glushenkova E, Barrios JP, Schier AF, Kunes S, Engert F, Douglass AD. Zebrafish oxytocin neurons drive nocifensive behavior via brainstem premotor targets. Nature Neuroscience. 2019;22:1477–1492. doi: 10.1038/s41593-019-0452-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Winslow JT. Mouse social recognition and preference. Current Protocols in Neuroscience. 2003;22:8.16.1–8.16.8. doi: 10.1002/0471142301.ns0816s22. [DOI] [PubMed] [Google Scholar]
  42. Yamamoto K, Ruuskanen JO, Wullimann MF, Vernier P. Two tyrosine hydroxylase genes in vertebrates new dopaminergic territories revealed in the zebrafish brain. Molecular and Cellular Neurosciences. 2010;43:394–402. doi: 10.1016/j.mcn.2010.01.006. [DOI] [PubMed] [Google Scholar]
  43. Zellner D, Padnos B, Hunter DL, MacPhail RC, Padilla S. Rearing conditions differentially affect the locomotor behavior of larval zebrafish, but not their response to valproate-induced developmental neurotoxicity. Neurotoxicology and Teratology. 2011;33:674–679. doi: 10.1016/j.ntt.2011.06.007. [DOI] [PubMed] [Google Scholar]
  44. Ziv L, Muto A, Schoonheim PJ, Meijsing SH, Strasser D, Ingraham HA, Schaaf MJ, Yamamoto KR, Baier H. An affective disorder in zebrafish with mutation of the glucocorticoid receptor. Molecular Psychiatry. 2013;18:681–691. doi: 10.1038/mp.2012.64. [DOI] [PMC free article] [PubMed] [Google Scholar]

Decision letter

Editor: Peggy Mason1
Reviewed by: Peggy Mason2

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Acceptance summary:

I agree with you that in these times of widespread isolation, the questions addressed in this report, and the distinction between a self chosen loner mode and an extrinsically forced loneliness are ever so relevant.

Decision letter after peer review:

Thank you for submitting your article "Whole brain mapping reveals abnormal activity in socially isolated zebrafish" for consideration by eLife. Your article has been reviewed by four peer reviewers, including Peggy Mason as the Reviewing Editor and Reviewer #1, and the evaluation has been overseen by Catherine Dulac as the Senior Editor.

The reviewers discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

The most important point is that we would like for you to test whether the hypothalamic neurons are serotonergic using a method of your choice. As we discussed this manuscript just a week or two ago, we felt that the most important point was that we wanted you to test whether the hypothalamic neurons are serotonergic using a method of your choice. I continued with the explicit statement that "Acceptance will not be contingent on a specific answer – yes they are as you suspect or no they aren't. Either finding is fine. Obviously those two possibilities will require very different interpretations and discussions. It may also be that you choose to submit your manuscript, as is and without these added experiments, to a different journal. Or to resubmit with these experiments to eLife. Any of these courses of action are fine."

However, the world has changed since our editorial discussions took place. I am assuming that you, along with the bulk of scientists, do not have access to their research laboratories. I do not want to hold up your publication indefinitely as the world navigates the pandemic. I therefore offer you the option of proceeding with purely textual changes and explicitly withdraw the requirement of additional experiments. Of course, should you have those experiments in hand, we welcome their inclusion in the revised manuscript that we look forward to receiving.

A number of other points that stand out from the many suggestions that you will find below.

1) Introduction tightened up to foreshadow the key takeaway message as suggested by reviewer 4.

2) A clear indication of what the N counts, as reviewer 2 asks for. Also please explicate the full method of behavioural assay and analysis rather than referring to another paper.

3) It appears that immobility is used as a metric for "anxiety." If so, this should be quantified and immobility should take the place of anxiety in the results. Moreover. immobility can also signal fear- or depression-like affect. Thus, the rhetoric of the discussion should be toned down to recognize this ambiguity. It is noteworthy in this regard that buspirone has atypical antidepressant as well as anxiolytic effects.

4) Please explain why Busp was tested on partially isolated rather than total isolates.

One final point, there was disagreement among the reviewers regarding the use of the terms lonely and loner. I am going to take editor's prerogative and call this one on the side of keeping those terms and even extending their use. But to the dissenters' point, please add some justification and caveats to your use of these terms.

Reviewer #1:

This is a very interesting and compelling study showing that the lack of a preference for being near other fish that is typical is not present after social isolation. But the exciting piece is that the cfos activation is different in control raised animals that avoid other fish and in isolated fish that show the same avoidance. The authors suggest use the terms lonely and loner for isolated and control fish that fail to prefer other fish.

My only major concern is that the "anxiety" behavior is talked about but not shown or quantified. Figure 1D is referenced but it is not obvious to me how that figure of location shows me the amount of freezing. The solid black areas could be freezing or repetitive swimming in the same area – I can't tell which. And I would be happier if they called it freezing if that is in fact how they measured it.

Reviewer #2:

This paper utilizes zebrafish in order to study the effects of chronic social isolation on neural responses. The authors note that within a normal zebrafish population, there exists some variability in the social preference of the animals, such that while most animals display a strong tendency to shoal with a visually presented conspecific, there is a certain percentage that will not display this behavior (or display it to a very limited extent), a behavior that is reminiscent of what is observed in chronically isolated animals, which also tend to avoid social interaction. The authors then attempt to compare the neural response in particular the optic tectum and regions of the ventral midbrain (Preoptic nucleus of the hypothalamus, Posterior tuberal nucleus, and caudal hypothalamus) to visually presented social stimuli, and draw a distinction between these responses in chronically isolated fish, acutely isolated fish, and the response of a "normal" socially avoidant fish. They show that the neural responses in these brain areas differ as a function of isolation, and thereby conclude that the social avoidance behavior triggered by developmental isolation is not similar to the social avoidance found naturally within the population. Lastly, they rescue the anti-social phenotype by utilizing the anxiolytic drug Buspirone, both in partially isolated and non isolated fish, thus putatively showing that anti-social behavior is related to stress.

1) General comment: The authors report highly significant p-values for their findings (on the order of ^-8), this is to be expected given the reported group sizes. Are these group sizes actually from one experimental repeat (e.g n = 380 for non-isolated fish), or have they pooled together several repeats? If so then it would be more appropriate to analyze the repeated experiments by some other method (e.g. a 2-way-ANOVA or Generalized Linear Model with factors: "experiment number" and "treatment"), rather than pool them and utilize a single test for all of the data.

2) Third paragraph: "we found that fish raised in isolation were significantly less active than their normally raised siblings during the acclimation session (Figure 1B: C vs Fi, p=9.0e-6; C vs Pi, p=2.8e-9 Mann-Whitney). However, when divided into groups with similar social preferences, pro-social isolated fish were nearly as active as pro-social controls (Figure 1C right: C (+S) vs Fi (+S), p=0.02 Mann-Whitney; C (+S) vs Pi (+S) p=0.14 Mann-Whitney"

These statements are confusing/misleading. The authors did not compare the activity levels of pro-social, non-social, and anti-social fish in the same paradigm as they did in the first sentence (namely, the acclimation phase). They measure it during the social exposure phase (Figure 1C, legend), thus they actually show a differential response to the social cue in terms of overall motility, but not a change in baseline activity as a function of sociality levels per se. I agree that it would have been interesting to compare, post-hoc, the baseline activity levels of the pro\non\anti-social fish (i.e. classify the fish based on their score in the social assay, and find their baseline activity during acclimation, and compare between the groups), but this is not the data presented herein , and the phrasing ought to be changed to reflect that, or the correct comparison presented in the figure.

3)."…which were both not exposed to social cues during the assay (Figure 3A), we found a significant increase in areas associated with visual processing the optic tectum, (McDowell et al., 2004) and stress responses the posterior tuberal nucleus, PTN"

As far as I can tell, the authors did not attempt to assess the identity of the PTN activated neurons and therefore this statement may be a bit strong. For example, the PTN is also the major area of fish dopaminergic neurons associated with reward as well GABA-ergic neurons.

4) Paragraph four-six: What was the data normalized to in these experiments. In the preceding experiment (Figure 2) where the authors look at the response of the social areas, they compare the socially exposed to fish to socially nonexposed fish, but this cannot be the case here as the socially unexposed fish are part of the analysis (and their activity levels do not sum up to 1 on the normalized activity graph).

5) "…and social phobia in humans (Schneier et al., 1993; van Vliet et al., 1997)" should read: "reduces social phobia", not enhances.

6) "Buspirone's impact on the rate of recovery of social preference suggests it may be reducing anxiety by regulating circuit plasticity, perhaps by promoting down-regulation of the hypersensitivity acquired during the isolation period"

The authors previously quantified anxiety-like behaviors via measures of mobility (% time spent moving). Therefore, it would be appropriate to include a measure of time spent moving per minute in order to strengthen this claim. As it stands, this claim is an interesting hypothesis, but not quite shown by the data.

7) Three comments on the Buspirone rescue experiment:

a) Why did the authors only perform this experiment on partially isolated fish, and not fully isolated ones? Given the differences in behavior and brain activity between these two models it is not at all clear that the neural mechanism which drives their anti-social behavior is the same, and therefore full isolates may not be amenable to treatment by Buspirone. This would be an interesting and relevant addition to the paper.

b) Buspirone rescued anti-social behavior in Control (non-isolated) fish, and also in partially isolated fish. The authors main point in the paper is that the neural mechanisms which drive anti-social behavior following isolation are caused by increased anxiety, and are distinct from those that drive it in the general population. However, since Buspirone rescued anti-social behavior also in controls, does this not mean that anti-social controls and anti-social partially isolated animals share some neural phenotype?

c) If possible, it would be instructive to perform neural activity imaging on Buspirone treated animals, and in particular, show reduced activity in the optic tectum and PTN, as the authors claim that hyperactivity in these regions is anxiogenic and drives the anti-social behavior of isolates. Thus, if the authors hypothesis is correct, it would be expected that Buspirone treatment, which rescues the behavior, would also reduce activity in these areas.

Reviewer #3:

This study identifies brains areas that control social behavioural responses towards conspecifics in socially isolated and normally reared zebrafish. I am excited about this work because it opens up the possibility to understand the neurotransmitters and brain areas that control social interactions. The study is short and the data that are presented are clear. In particular, the differences in cfos expression between animals is very impressive.

This study would be improved significantly by directly measuring 5-HT levels in these animals (see below). In the absence of this, some of the conclusions are too strong. I also have some suggestions for improving the Discussion section and tidying up the writing.

1) This research could be improved by directly measuring 5-HT levels in the hypothalamus of isolated or socially-reared fish. This could be done by HPLC. The caudal hypothalamus contains a number of different neuron types, so IEG activity here could reflect activation of several neurotransmitter systems. In the absence of HPLC data, you need to consider other neurotransmitters that could also contribute to this phenotype.

2) The whole brain imaging of cfos is neat, but it loses any cellular resolution. The findings would be strengthened by making sections of the brain and taking higher magnification images. This could include co-labelling with specific markers for dopamine and 5-HT neurons.

3) The Abstract needs rewriting because the tense used changes throughout.

4) The second paragraph of the Introduction (describing isolation and the experimental setup) is not that clear, and can be rewritten to make it easier to follow.

5) There are a large number of acronyms for brain areas used. It might be better to replace all of these with the full names of the brain areas. This would also improve consistency – for example, you write preoptic (area) but not POA. It will also help to distinguish brain areas from treatment groups (nsc, Pi, Fi) etc.

6) Similarly please check name usage throughout – cf preoptic area, preoptic, pre-optic.

7) I missed the link to sertb expression in the hypothalamus. I would highlight this more clearly, particularly because you use buspirone to rescue behaviour in later experiments.

8) Some of the discussion points can be improved. There are already studies linking the POA to social behaviour in zebrafish, and these should be included here. It is possible that arginine vasopressin plays an important role in this behaviour. Linking the central hypothalamus to the control of social behaviour is also interesting; the function of sertb positive neurons is well studied. Again, this can be highlighted here.

9) It is not clear whether the caudal hypothalamus identified by cfos staining here is the same as the ventromedial hypothalamus described by O'Connell and Hoffmann.

10) Final paragraph of the Discussion I would remove the terms "lonely" and "loner". They are not necessary here and seem anthropomorphic.

Reviewer #4:

This manuscript describes a series of studies exploring the impacts of social isolation on zebrafish behavior and neural activity. In particular, the authors compare "loner" fish who by choice spend little time socializing with "lonely" fish who are experimentally isolated and discover that while behavior is relatively similar, neural activity is fundamentally different, the latter being rescued by reducing serotonin. In general, I like the paper a lot; it represents a significant effort by the authors and suggests that despite behavioral similarities, all isolation is not the same. This could profoundly shape how we think about sociality in general, and the impacts of isolation in particular. However, there are some aspects of the manuscript that need to be clarified.

Substantive concerns:

1) One of my big concerns is that the initial paragraph of manuscript does not do a good job of setting up the paper. The first paragraph needs to prepare the reader for the "lonely" vs "loner" fish comparison and explain why it matters. Also, the partial isolation fish are introduced in the second paragraph with no explanation of why they are important. I figured all of this out by the end of the manuscript, but the paper will be much stronger if the reader knows in the first paragraph what to expect. Specifically, I would add conceptual overview that sets the reader up for all of the comparisons and emphasizes why each is needed and important.

2) The only places I can find the sample size are in Figure 1 and the paragraph describing the Buspirone treatment. Please provide details on sample size and whether it was sufficiently powered, either throughout or in the Materials and methods section.

eLife. 2020 May 5;9:e55863. doi: 10.7554/eLife.55863.sa2

Author response

A number of other points that stand out from the many suggestions that you will find below.

1) Introduction tightened up to foreshadow the key takeaway message as suggested by reviewer 4.

We agree that the Introduction, and especially the first paragraph, did not clearly emphasise the main findings of the paper. We have now added a more general overview of the study’s goals and outcomes, introduced the concept of “lonely” and “loner”, and explained why we decided to use both partial and full isolation.

2) A clear indication of what the N counts, as reviewer 2 asks for. Also please explicate the full method of behavioural assay and analysis rather than referring to another paper.

Thank you for pointing this out. We now include N counts for every experiment reported, either in the figure directly or in the corresponding legend. As suggested, we have modified the Materials and methods section to include a complete explanation of the behavioural assay and analysis as opposed to referring to previous work. As requested by reviewer 2, we have also specified that each fish has been tested only once, i.e. 380 individual control fish were tested in the social assay, and therefore all the experiments are single trials and it is appropriate to use the Mann-Whitney statistical test as opposed to ANOVA.

3) It appears that immobility is used as a metric for "anxiety." If so, this should be quantified and immobility should take the place of anxiety in the results. Moreover. immobility can also signal fear- or depression-like affect. Thus, the rhetoric of the discussion should be toned down to recognize this ambiguity. It is noteworthy in this regard that buspirone has atypical antidepressant as well as anxiolytic effects.

Thanks for highlighting this ambiguity. We agree that the measure we used in the paper, percentage of time moving, could underlie other behavioural phenotypes apart from anxiety, such as depression or fear, and we have modified our Discussion accordingly. We have also now included a supplementary Figure (Figure 1—figure supplement 1C-D) where we quantify the percentage of time spent freezing during acclimation and social cue exposure. This parameter reports the percentage of time that fish remain motionless for 3 seconds and longer. The data show that fully and partially isolated fish show a significant increase in “freezes” during acclimation and social cue exposure, although fully isolated fish show a larger increase (Figure 1—figure supplement 1C-D, left). Furthermore, when we look at the anti-social fish during social cue exposure (Figure 1—figure supplement 1D right, blue dots), each experimental group (controls, fully isolated, and partially isolated), has a similar percentage of time freezing, supporting our claim that freezing is a hallmark of anti-social behaviour for both “loner” and “lonely” fish.

Finally, it is true that Buspirone has also been used as an atypical antidepressant, however, only as adjunctive treatment in patients with generalized anxiety disorders and depression. This is very interesting. We believe that there is still much to learn about the action of Buspirone and the interaction between anxiety and depression, and we feel that zebrafish will provide an excellent model for further study. We have therefore improved our discussion of Buspirone throughout the manuscript, both to motivate its use in our experiments and to suggest future directions.

4) Please explain why Busp was tested on partially isolated rather than total isolates.

We agree that this point needs to be better explained in the paper, thank you for highlighting this. There are several reasons why we opted for partially versus fully isolated fish to perform the drug experiments. First of all, partial isolation of only 48 hours instead of 3 weeks already induces a profound behavioural change in social preference behaviour (Figure 1A) similar to full isolation. Secondly, the functional activity observed in the presence (Figure 2B) or not (Figure 3A-B) of social cues is very similar between fully and partially isolated fish. However, in both behaviour and neural activity, the impact of partial isolation is less severe than full isolation. We therefore decided that partial isolation represented an “intermediate” phenotype that would allow us to detect both positive and negative impacts on sociality for a given treatment. This has now been explained in the manuscript.

One final point, there was disagreement among the reviewers regarding the use of the terms lonely and loner. I am going to take editor's prerogative and call this one on the side of keeping those terms and even extending their use. But to the dissenters' point, please add some justification and caveats to your use of these terms.

Thank you for this decision. We also believe that the terms are compelling and we did include them in the title of our first submission. Therefore, we have reintroduced the terms in the title, explained and justified there use in the Introduction, and used them throughout the manuscript where appropriate.

Reviewer #1:

This is a very interesting and compelling study showing that the lack of a preference for being near other fish that is typical is not present after social isolation. But the exciting piece is that the cfos activation is different in control raised animals that avoid other fish and in isolated fish that show the same avoidance. The authors suggest use the terms lonely and loner for isolated and control fish that fail to prefer other fish.

My only major concern is that the "anxiety" behavior is talked about but not shown or quantified. Figure 1D is referenced but it is not obvious to me how that figure of location shows me the amount of freezing. The solid black areas could be freezing or repetitive swimming in the same area – I can't tell which. And I would be happier if they called it freezing if that is in fact how they measured it.

We agree that the qualitative presentation of “freezing” behaviour in Figure 1D and the supplemental video were insufficient to demonstrate that anti-social fish exhibit anxiety-like behaviour. Therefore, we have now quantified the percentage of time spent freezing during acclimation and social cue period and included this analysis in Figure 1—figure supplement 1C-D. The data show that both fully and partially isolated fish have a significant increase in freezes during the acclimation period, suggesting a heightened level of general anxiety. This increased freezing persists during social cue exposure, but is largest in fully isolated fish. We believe this additional analysis greatly supports our claim that isolation increases anxiety-like freezing behaviour. Thank you very much for the suggestion.

Reviewer #2:

This paper utilizes zebrafish in order to study the effects of chronic social isolation on neural responses. The authors note that within a normal zebrafish population, there exists some variability in the social preference of the animals, such that while most animals display a strong tendency to shoal with a visually presented conspecific, there is a certain percentage that will not display this behavior (or display it to a very limited extent), a behavior that is reminiscent of what is observed in chronically isolated animals, which also tend to avoid social interaction. The authors then attempt to compare the neural response in particular the optic tectum and regions of the ventral midbrain (Preoptic nucleus of the hypothalamus, Posterior tuberal nucleus, and caudal hypothalamus) to visually presented social stimuli, and draw a distinction between these responses in chronically isolated fish, acutely isolated fish, and the response of a "normal" socially avoidant fish. They show that the neural responses in these brain areas differ as a function of isolation, and thereby conclude that the social avoidance behavior triggered by developmental isolation is not similar to the social avoidance found naturally within the population. Lastly, they rescue the anti-social phenotype by utilizing the anxiolytic drug Buspirone, both in partially isolated and non isolated fish, thus putatively showing that anti-social behavior is related to stress.

1) General comment: The authors report highly significant p-values for their findings (on the order of ^-8), this is to be expected given the reported group sizes. Are these group sizes actually from one experimental repeat (e.g n = 380 for non-isolated fish), or have they pooled together several repeats? If so then it would be more appropriate to analyze the repeated experiments by some other method (e.g. a 2-way-ANOVA or Generalized Linear Model with factors: "experiment number" and "treatment"), rather than pool them and utilize a single test for all of the data.

Thank you for pointing this out. The experiments are all single experimental repeats (i.e. 380 individual non-isolated fish were tested once in the social preference assay). We have now described this more clearly in the Materials and methods section.

2) Third paragraph: "we found that fish raised in isolation were significantly less active than their normally raised siblings during the acclimation session (Figure 1B: C vs Fi, p=9.0e-6; C vs Pi, p=2.8e-9 Mann-Whitney). However, when divided into groups with similar social preferences, pro-social isolated fish were nearly as active as pro-social controls (Figure 1C right: C (+S) vs Fi (+S), p=0.02 Mann-Whitney; C (+S) vs Pi (+S) p=0.14 Mann-Whitney"

These statements are confusing/misleading. The authors did not compare the activity levels of pro-social, non-social, and anti-social fish in the same paradigm as they did in the first sentence (namely, the acclimation phase). They measure it during the social exposure phase (Figure 1C, legend), thus they actually show a differential response to the social cue in terms of overall motility, but not a change in baseline activity as a function of sociality levels per se. I agree that it would have been interesting to compare, post-hoc, the baseline activity levels of the pro\non\anti-social fish (i.e. classify the fish based on their score in the social assay, and find their baseline activity during acclimation, and compare between the groups), but this is not the data presented herein , and the phrasing ought to be changed to reflect that, or the correct comparison presented in the figure.

This is a really good point. Thank you for highlighting our confusing presentation and description. We have now changed Figure 1C and replaced the swarm plots of anti-social and pro-social fish during the social exposure with the ones during the acclimation period (that were classified based on their subsequent performance in the social cue period). The new data support our argument that the baseline (i.e. without social cues) mobility of anti-social control, fully, and partially isolated fish is very similar during the acclimation period. This reduced mobility persists during the social cue phase, as we originally reported in the main figure. We have moved the swarm plots of mobility during the social cue period to the Figure 1—figure supplement 1.

3) "…which were both not exposed to social cues during the assay (Figure 3A), we found a significant increase in areas associated with visual processing the optic tectum, (McDowell et al., 2004) and stress responses the posterior tuberal nucleus, PTN"

As far as I can tell, the authors did not attempt to assess the identity of the PTN activated neurons and therefore this statement may be a bit strong. For example, the PTN is also the major area of fish dopaminergic neurons associated with reward as well GABA-ergic neurons.

Thank you for helping us to clarify this point. We have recently begun to perform in situs for dopaminergic, serotoninergic and acetylcholine markers in juvenile fish. These preliminary data show that the area activated is the PTN, however, the dopaminergic neurons are not the ones that we see activated. These dopaminergic neurons seem to surround the activated area shown in Figure 3A bottom. We do not know yet, what type of markers these neurons express. PTN neurons have been shown to express GABA (gad) and can also coexpress (vglut2) in addition to dopamine (Th1). We are planning to continue the characterization of the PTN neurons as soon as possible, but for now, we have improved our discussion of the PTN and other identified areas, throughout the manuscript to reflect these possible alternate roles.

4) Paragraph four-six: What was the data normalized to in these experiments. In the preceding experiment (Figure 2) where the authors look at the response of the social areas, they compare the socially exposed to fish to socially nonexposed fish, but this cannot be the case here as the socially unexposed fish are part of the analysis (and their activity levels do not sum up to 1 on the normalized activity graph).

Thank you for highlighting this confusion. All individual brain maps were first processed using an intensity histogram normalization, as described in the Materials and methods. As you point out, in Figure 2 we present the data as “difference images” to highlight relative changes in response to social cue exposure. In Figure 3, however, we compare between groups (i.e. controls vs. isolation) and identify c-fos activity changes that occur regardless of social cue presentation, and therefore we report the absolute (histogram normalized) c-fos expression in different regions. We have now emphasized the difference in presentation for Figure 2 and 3 in the manuscript text to highlight this for the reader.

5) "…and social phobia in humans (Schneier et al., 1993; van Vliet et al., 1997)" should read: "reduces social phobia", not enhances.

Thank you for finding this error. We have changed the text accordingly.

6) "Buspirone's impact on the rate of recovery of social preference suggests it may be reducing anxiety by regulating circuit plasticity, perhaps by promoting down-regulation of the hypersensitivity acquired during the isolation period"

The authors previously quantified anxiety-like behaviors via measures of mobility (% time spent moving). Therefore, it would be appropriate to include a measure of time spent moving per minute in order to strengthen this claim. As it stands, this claim is an interesting hypothesis, but not quite shown by the data.

Thank you for suggesting this important analysis. We have now quantified the percentage of time moving in one minute bins for each treatment group presented in Figure 4 A-B and included a new panel in Figure 4 (Figure 4C). The results show that Buspirone treatment increases the mobility of partially isolated fish, which reach control levels within the first minute of the social assay. Therefore, these new data, together with the VPI results (Figure 4B), supports the argument that Buspirone reduces anxiety/increases mobility and that this happens prior to a recovery of social preference. Very interesting and a great addition to the study. Thanks again for the suggestion.

7) Three comments on the Buspirone rescue experiment:

a) Why did the authors only perform this experiment on partially isolated fish, and not fully isolated ones? Given the differences in behavior and brain activity between these two models it is not at all clear that the neural mechanism which drives their anti-social behavior is the same, and therefore full isolates may not be amenable to treatment by Buspirone. This would be an interesting and relevant addition to the paper.

Thanks for allowing us to clarify this decision. In addition to the reply to the Editor’s comment above, we would like to say that we are (or we were prior to lockdown) currently pursuing experiments testing Buspirone treatment on fully-isolated fish. However, given that we have a limited number of isolation tanks, and that these tanks are entirely used by one fish for the full three week period of isolation, then these full isolation experiments are much lower throughput. Therefore, given that brief periods of isolation are likely much more common than complete lack of social interaction from birth, we decided to start with partial isolation fish to test possible treatments/rescues. However, we agree, that full (or at least longer) isolation tests need to follow for any identified treatment candidates.

b) Buspirone rescued anti-social behavior in Control (non-isolated) fish, and also in partially isolated fish. The authors main point in the paper is that the neural mechanisms which drive anti-social behavior following isolation are caused by increased anxiety, and are distinct from those that drive it in the general population. However, since Buspirone rescued anti-social behavior also in controls, does this not mean that anti-social controls and anti-social partially isolated animals share some neural phenotype?

Thank you for pointing out this potential confusion. It is not clear that Buspirone has altered the social preference of anti-social control fish, as we still see ~10% anti-social control fish following drug treatment (Figure 4—figure supplement 1A left panel), similar to the proportion found in an untreated control population. However, given that anti-social fish are rare in normally-raised populations, we do not have enough data to conclude that Buspirone is only effective at reducing anxiety in isolated fish.

c) If possible, it would be instructive to perform neural activity imaging on Buspirone treated animals, and in particular, show reduced activity in the optic tectum and PTN, as the authors claim that hyperactivity in these regions is anxiogenic and drives the antisocial behavior of isolates. Thus, if the authors hypothesis is correct, it would be expected that Buspirone treatment, which rescues the behavior, would also reduce activity in these areas.

We fully agree that these are interesting experiments that could shed light on the mechanism of action of Buspirone. We have indeed kept and fixed all the Buspirone treated fish so that we can explore their functional activity by using c-fosin situs, however, this will be part of a future study where explore the impact of Buspirone treatment in more detail. Thank you for all the suggestions.

Reviewer #3:

This study identifies brains areas that control social behavioural responses towards conspecifics in socially isolated and normally reared zebrafish. I am excited about this work because it opens up the possibility to understand the neurotransmitters and brain areas that control social interactions. The study is short and the data that are presented are clear. In particular, the differences in cfos expression between animals is very impressive.

This study would be improved significantly by directly measuring 5-HT levels in these animals (see below). In the absence of this, some of the conclusions are too strong. I also have some suggestions for improving the Discussion section and tidying up the writing.

1) This research could be improved by directly measuring 5-HT levels in the hypothalamus of isolated or socially-reared fish. This could be done by HPLC. The caudal hypothalamus contains a number of different neuron types, so IEG activity here could reflect activation of several neurotransmitter systems. In the absence of HPLC data, you need to consider other neurotransmitters that could also contribute to this phenotype.

Thanks for raising this point. We agree that in order to fully understand the impact of social isolation and the mechanism of action of Buspirone, we would need to characterise the levels of 5-HT via HPLC. We are not currently able to do this, but will pursue these experiments for a future publication. Thank you for the suggestion.

We are also planning to measure the expression levels of different hypothalamic neurotransmitters following isolation and after drug treatment. It is already known, indeed, that several of these markers are increased or reduced after isolation, such as Dopamine (Shams et al., 2018) and Serotonin (Shams et al., 2015). These data have been obtained by looking at the level of the expression in different brain areas, but not at single cell resolution. Our behavioural and imaging assay should allow us to achieve this resolution, and we have already started to perform double in situs of c-fos and candidate markers. These data will also be included in a future publication.

2) The whole brain imaging of cfos is neat, but it loses any cellular resolution. The findings would be strengthened by making sections of the brain and taking higher magnification images. This could include co-labelling with specific markers for dopamine and 5-HT neurons.

We fully agree that it would be very informative to know where and what type of cells are activated during social viewing, after isolation and Buspirone treatment. These data, together with functional imaging, could shed more light on how social isolation affects the social circuit, and how Buspirone is able to rescue the anti-social phenotype. One advantage of using juvenile fish is that their brain is still relatively small, and, therefore, we do not need to do brain sections to achieve cellular resolution. We have indeed started to do double in situs for c-fos and putative markers, and these data will be included in another publication.

3) The Abstract needs rewriting because the tense used changes throughout.

Thank you for pointing this out, we have corrected the Abstract.

4) The second paragraph of the Introduction (describing isolation and the experimental setup) is not that clear, and can be rewritten to make it easier to follow.

Thank you for your suggestion. Other reviewers also made this point and we have now changed the Introduction to improve clarity and better introduce our main findings.

5) There are a large number of acronyms for brain areas used. It might be better to replace all of these with the full names of the brain areas. This would also improve consistency – for example, you write preoptic (area) but not POA. It will also help to distinguish brain areas from treatment groups (nsc, Pi, Fi) etc.

This is a great suggestion. The text now contains the names of the areas in full, and we have slightly modified the acronyms on the figures to follow the most common conventions found in the literature.

6) Similarly please check name usage throughout – cf preoptic area, preoptic, pre-optic.

Thanks for spotting these inconsistencies. We have changed the text accordingly.

7) I missed the link to sertb expression in the hypothalamus. I would highlight this more clearly, particularly because you use buspirone to rescue behaviour in later experiments.

We agree that this is an important point of the paper, and we have changed the text to emphasise the presence of serotoninergic neurons in the caudal hypothalamus.

8) Some of the discussion points can be improved. There are already studies linking the POA to social behaviour in zebrafish, and these should be included here. It is possible that arginine vasopressin plays an important role in this behaviour. Linking the central hypothalamus to the control of social behaviour is also interesting; the function of sertb positive neurons is well studied. Again, this can be highlighted here.

Thank you for this excellent observation. We have expanded our Discussion and included your main points in the section describing the areas differentially activated in pro- and anti-social fish.

9) It is not clear whether the caudal hypothalamus identified by cfos staining here is the same as the ventromedial hypothalamus described by O'Connell and Hoffmann.

Thanks for highlighting this. Yes, the VMH is the region we see activated in our c-fos functional maps. In O’Connell and Hoffmann, 2011, there are two hypothalamic areas already known to be involved in social behaviour: the anterior hypothalamus and the ventromedial hypothalamus (VMH). The VMH is the part of the hypothalamus that is more caudal and comprises the area we see activated in our experiments.

10) Final paragraph of the Discussion I would remove the terms "lonely" and "loner". They are not necessary here and seem anthropomorphic.

We understand that these two terms may not be ideal. However, in accordance with the decision of the Editors, we will keep them, albeit with a more complete explanation of how they apply to our zebrafish experiment.

Reviewer #4:

This manuscript describes a series of studies exploring the impacts of social isolation on zebrafish behavior and neural activity. In particular, the authors compare "loner" fish who by choice spend little time socializing with "lonely" fish who are experimentally isolated and discover that while behavior is relatively similar, neural activity is fundamentally different, the latter being rescued by reducing serotonin. In general, I like the paper a lot; it represents a significant effort by the authors and suggests that despite behavioral similarities, all isolation is not the same. This could profoundly shape how we think about sociality in general, and the impacts of isolation in particular. However, there are some aspects of the manuscript that need to be clarified.

Substantive concerns:

1) One of my big concerns is that the initial paragraph of manuscript does not do a good job of setting up the paper. The first paragraph needs to prepare the reader for the "lonely" vs "loner" fish comparison and explain why it matters. Also, the partial isolation fish are introduced in the second paragraph with no explanation of why they are important. I figured all of this out by the end of the manuscript, but the paper will be much stronger if the reader knows in the first paragraph what to expect. Specifically, I would add conceptual overview that sets the reader up for all of the comparisons and emphasizes why each is needed and important.

Thank you for your great suggestions on how to improve the text. We have now changed the first and second paragraphs of the Introduction to incorporate your comments, please see our additional response to the Editors’ comments.

2) The only places I can find the sample size are in Figure 1 and the paragraph describing the Buspirone treatment. Please provide details on sample size and whether it was sufficiently powered, either throughout or in the Materials and methods section.

Thanks for finding this omission in the paper. We have now added the sample sizes in either the figure panel or legend for all experiments throughout. Our power analysis was based on variance measurements obtained during a previous work investigating the impact of small effect drug treatments on sociality (Dreosti et al., 2015). Given the effect size of Buspirone that we observed, our sample sizes ended up being larger than necessary. This is now explained in the Materials and methods section.

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    Transparent reporting form
    elife-55863-transrepform.pdf (237.6KB, pdf)

    Data Availability Statement

    All the images, video, protocols, analysis scripts, and data that support the findings of this study are available from this website (http://www.dreo-sci.com/resources/), or our GitHub repository (https://github.com/Dreosti-Lab/Lonely_Fish_2020; Dreosti, 2020; copy archived at https://github.com/elifesciences-publications/Lonely_Fish_2020), or from the corresponding author upon request.

    All the images, video, protocols, analysis scripts, and data that support the findings of this study are available on our website (www.dreo-sci.com/resources) and GitHub repository (https://github/Dreosti-Lab/Social_Zebrafish), or from the corresponding author upon request.

    Articles from eLife are provided here courtesy of eLife Sciences Publications, Ltd

    RESOURCES