angptl4 Gene Expression as a Marker of Adaptive Homeostatic Response to Social Isolation Across the Lifespan in Zebrafish

Social isolation has detrimental health effects, but the underlying mechanisms are unclear. This study used zebrafish to investigate the impact of two weeks of isolation on behaviour and gene expression in the central nervous system (CNS) at different life stages. Results showed that socially deprived young adult zebrafish experienced increased anxiety, accompanied by changes in gene expression. Most gene expression patterns returned to normal within 24 hours of reintroduction to a social environment, except angptl4 , which remained upregulated, suggesting an adaptive mechanism. Similarly, aging zebrafish displayed heightened anxiety and increased CNS expression of angptl4 during isolation, but effects were reversed upon reintroduction to a social group. The findings imply that angptl4 plays a homeostatic role in response to social isolation, which varies across the lifespan. The study emphasises the importance of social interactions for psychological well-being and highlights the negative consequences of isolation, especially in older individuals.

factor for coronary artery disease and stroke (Cené et al., 2022;Valtorta et al., 2016;Xia and Li, 2018) as well as type 2 diabetes (Brinkhues et al., 2017;Hackett et al., 2020). However, the molecular mechanisms by which social isolation increases risk for coronary artery disease, stroke and metabolic disorders such as type 2 diabetes require further investigation (Jeste et al., 2020).
Zebrafish (Danio rerio) is a social species, which actively forms small shoals in nature (Engeszer et al., 2007) and in the laboratory where it exhibits a higher degree of social cohesion compared to other traditional laboratory animals, such as the mouse or the rat (Saverino and Gerlai, 2008). They are, therefore, a popular model for studying social behaviour (Larson et al., 2006;Pagnussat et al., 2013) with isolation-induced behavioural effects being well studied in adult zebrafish (Daniel and Bhat, 2022;Forsatkar et al., 2017;Giacomini et al., 2015;Larson et al., 2006;Lindsey and Tropepe, 2014;Marchetto et al., 2021;Onarheim et al., 2022;Shams et al., 2018Shams et al., , 2015. Zebrafish share the major neurotransmitter systems with humans such as the dopaminergic and serotonergic pathways (Horzmann and Freeman, 2016;Rico et al., 2011) with many studies demonstrating the involvement of these pathways in the shoaling behaviour Saif et al., 2012;Scerbina et al., 2012;Tunbak et al., 2020). Moreover, it has been previously shown that zebrafish may display reduced interest or engagement in social interactions when subjected to social isolation (Tunbak et al., 2020) and this has also been demonstrated in humans (Vitale and Smith, 2022) suggesting similar behavioural effects for isolation in both species.
Here, we aimed to study the mechanisms by which social isolation impacts upon psychological health, and how this may change during ageing. In order to identify potential mechanisms underpinning this, we initially examined effects on CNS gene expression and behaviour in young adult zebrafish (6 months post fertilization).
J o u r n a l P r e -p r o o f

Animals and husbandry
Wildtype zebrafish (AB strain; ~ 50:50 male: female ratio at 6-month [young adult] and 24-month [ageing] post-fertilization) were bred in-house and raised in the University of Portsmouth Fish Facility. Animals were maintained on a recirculating aquatic housing system (Aquaneering, USA) in groups of n = 10 (in 2.8L tanks). Water temperature and pH were maintained at 28.5°C (±1 °C) and 8.4 (±0.4), respectively. Fish were kept in a 14-hour light: 10-hour dark cycle (lights on at 9:00 am) and fed on a daily diet containing live brine shrimp and dried flake flood three times a day (and once a day at weekends). All experiments were approved by the University of Portsmouth Animal Welfare and Ethical Review Board, and under license from the UK Home Office (Animals (Scientific Procedures) Act, 1986, Licence number PP8708123).

Social isolation and experimental groups
Young adult (6-month-old) and ageing (24-month-old) zebrafish were randomly assigned to two groups: the control (social) group and the experimental (socially isolated) group. Group 'Social' fish were kept in groups of n = 10 (5 male, 5 female) per 2.8 L tanks, and Group 'Socially Isolated' fish were placed, individually, in 1.4 L tanks, during which they had no access to olfactory or visual cues of other fish (we placed white opaque baking paper between tanks; all experiments were carried out on a recirculating rack). Photographs of the experimental setup are shown in Figure 1.
For initial profiling of the effects of isolation on changes in gene expression and behaviour, we isolated fish for 14 days. After 14 days, the behavioural profile of both groups was evaluated for anxiety using the Novel Tank diving test (Blaser and Rosemberg, 2012;Egan et al., 2010;Parker et al., 2012). We carried out two fully independent replicates of the initial 14-day isolation study. To understand the effects of shorter isolation periods on gene expression in 6 month-old fish, we also isolated fish for 2 days and 6 days. The 2-day isolation period was chosen to examine acute effects (and to allow sufficient time for changes in gene expression). We had initially planned to study two additional interim time points (6 days isolation and 10 days), but as there were already strong effects (similar to 14 days) after the 6-day isolation period, we chose to omit the 10 days for 3Rs reasons (i.e., to save unnecessary use of animals). J o u r n a l P r e -p r o o f [FIGURE 1 HERE]

Novel Tank diving test
Novel Tank diving tests were carried out in a fully randomised order between 10:00 am and 16:00 pm using two independent batches (n = 17-18 per batch/per group).
Animals were transported individually to a black test tank (L 20 cm × H 17.5 cm × W 5 cm). Tanks were filled with 1 litre of aquarium water and were divided into three equal virtual horizontal zones (bottom, middle and top). Zebrafish swimming behaviour was recorded using the fully automated Zantiks AD behavioural testing system for adult zebrafish (Zantiks Ltd., Cambridge, UK) for 6 minutes. The system was fully controlled via a web-enabled device, which allowed the monitoring of fish during behavioural testing. The time spent in the top zone of the tank was compared between the isolated and social groups to evaluate differences in anxiety. The time spent in the top zone data was analysed using an analysis of covariance with the total distance travelled data being the covariate values. Results were expressed as estimated marginal (EM) means ± SEM.

Total RNA extraction and RNA sequencing
After behavioural testing, group-housed and socially isolated animals (n = 6 in each group, three males and three females) were euthanized by rapid cooling (immersion in 2ºC water) followed by immediate excision of the whole brain tissue for RNA extraction. Six of the isolated fish (three males and three females) were group-housed for 24 hours, and then whole brain tissue was removed. Brain tissue was snap-frozen in liquid nitrogen and kept at -80ºC. RNA extraction was performed using the GeneJET RNA Purification Kit (Thermo Scientific) as described in the manufacturer's instructions. Genomic DNA was removed using the RapidOut DNA Removal Kit (Thermo Scientific). RNA concentration was determined using the Bioanalyzer 2100 (Agilent Technologies) using an RNA 6000 Nano kit. RNA quality was evaluated using the NanoDrop ND-1000 spectrophotometer (Thermo Scientific) and the Bioanalyzer 2100 (Agilent Technologies). A260/A280 and A230/A260 ratios greater than 1.8 and an RIN greater than 8 were considered acceptable. RNA samples were stored at -80º C until they were shipped to BGI Tech Solutions (Copenhagen, Denmark) to conduct a polyA mRNA enriched non-stranded RNA sequencing with 100-bp paired-end reads on the DNBseq platform.
J o u r n a l P r e -p r o o f

Bioinformatics
Base calling quality across all raw reads was assessed using fastQC V0.11.8 (Andrews, 2010), contamination was assessed using FastqScreen V0.13.0 (Wingett and Andrews, 2018), and results were combined using multiQC V1.8 (Ewels et al., 2016). Reads were trimmed using trim-galore (Krueger, 2012) to remove poor quality reads and adapters (parameters '--illumina -q 20 --stringency 5 -e 0.1 --length 20 -trim-n'). Trimmed reads were mapped to the GRCz11 Danio rerio consensus genome sequence from Ensembl using the STAR universal RNA-seq aligner V2.7.10a (Dobin et al., 2013) (parameters '-outSAMmultNmax 300'). Mapped reads were filtered to remove non-mapping reads and reads mapping with a quality score less than 20 using samtools V1.6 (Danecek et al., 2021). Gene models for GRCz11 were taken from Ensembl V107, and read counts over genes was quantified using the summarizeOverlaps function in the GenomicAlignments package (Lawrence et al., 2013) in R (R Core Team, 2021) (parameters 'mode = "Union", singleEnd = FALSE, ignore.strand = TRUE, fragments = FALSE'). Differential gene expression analysis between groups was performed using the DESeq2 package (Love et al., 2014) in R. P values were adjusted based on the Benjamini and Hochberg False Discovery Rate (FDR) multiple testing correction (Benjamini and Hochberg, 1995). Significant differential expression was identified based on fold change greater than two-fold (upor down-regulated) with an adjusted p-value less than 0.05. Genes with a mean Fragments Per Kilobase Mapped (FPKM) less than 1 in both groups being compared were removed to avoid the impact of low abundance transcripts. Gene ontology enrichment analysis was performed using the clusterProfiler package (Wu et al., 2021;Yu et al., 2012) in R.

Identification of Homo sapiens orthologs
Human orthologs of the zebrafish genes were identified using the Alliance of Genome  (Hu et al., 2011). The tool assigns a score based on the number of methodologies that call a specific ortholog. The J o u r n a l P r e -p r o o f descriptions of human orthologs were derived from GeneCards (https://www.genecards.org).

Identification of gene-disease associations
We used the Harmonizome database (Rouillard et al., 2016) to investigate whether there is an association between the identified DEGs in the experimental groups and anxiety and depressive and cardiovascular phenotypes. We chose these because social isolation is linked to anxiety and depression (Santini et al., 2020;Umberson and Karas Montez, 2010;Wilkialis et al., 2021b), and to cardiovascular disease (Cené et al., 2022;Valtorta et al., 2016;Xia and Li, 2018). The gene sets "Depressive Disorder" and "Anxiety Disorders" from the CTD (Comparative Toxicogenomics Database) Gene-Disease Associations dataset (Davis et al., 2023), which contain 5242 and 2774 genes/proteins, respectively, were used. The gene set "Cardiovascular Diseases", which contains 4608 genes/proteins from the CTD dataset was also used to examine the association between the DEGs and cardiovascular diseases. The CTD dataset contains curated and inferred gene-disease relationships. Curated relationships are derived from the literature or the OMIM database. Inferred relationships are established via CTD-curated chemical-gene interactions (Davis et al., 2023).

cDNA synthesis and Real-Time Quantitative PCR (RT-qPCR) analysis
Total RNA samples from the social, isolated, and reintroduced fish groups (n = 5 in each group) were converted to cDNA using the Applied Biosystems High-Capacity seconds followed by a melt curve analysis to confirm product specificity. Negative reverse transcription (RT) controls were used to confirm the absence of genomic DNA in RNA samples. Relative changes in gene expression were calculated using the deltadelta CT method after normalisation to the reference gene elf1a (elongation factor 1 J o u r n a l P r e -p r o o f alpha) . Changes in expression were presented as mean ± SD of fold change to the control social group. Data was tested for normality using a Shapiro-Wilk test and homogeneity of variances using Levene's test. Normally distributed data was analysed using an unpaired t-test, or an ordinary one-way ANOVA and Tukey's multiple comparisons test (when groups had equal variances), while non-normally distributed data was analysed using a Mann Whitney test or Kruskal Wallis and Dunn's multiple comparisons tests.

Social isolation induces changes in gene expression in 6-month-old fish in a time-dependent manner, and are rescued by reintroduction to the social group
Zebrafish (Danio rerio) is a social species, which actively forms small shoals in its natural environment (Engeszer et al., 2007). In the laboratory, zebrafish exhibit a high degree of social cohesion compared to other laboratory animals such as rodents (Saverino and Gerlai, 2008), making them a popular translational model for studying the basis of human social behaviour (Larson et al., 2006;Pagnussat et al., 2013). To establish the impact of social isolation on behaviour, we first singly housed 6-monthold fish for 2 weeks, then reintroduced them to their original housing group. We then carried out RNAseq on the brain of isolated (after 2 weeks), reintroduced (24 hours after reintroduction), and social control fish. The majority of the differentially expressed genes (DEGs) in the isolated fish group were upregulated (n = 46) in comparison to the control group, while four genes were downregulated (Log2 (Fold Change) ≥1 and FDR-adjusted p value < 0.05) ( Figure 2). [FIGURE 2 HERE] Gene symbols for the DEGs are listed in Table 1, in addition to Log Fold Change (LogFC) and adjusted p-values. We used the Alliance of Genome Resources website to identify human orthologs of the DEGs listed in Table 1. Ten of the genes had no orthology data available, and two genes were not found in the database. Table 1 lists the identified human orthologs, their description, as well as a score value, which is the number of methodologies used by the Alliance website that call a specific ortholog.
When more than one human ortholog was identified for a zebrafish gene, only the J o u r n a l P r e -p r o o f orthologs with the highest scores were reported. For four of the genes, more than one ortholog with the same score were identified. We found several genes of interest that were downregulated in the isolated fish.
Previous studies have found that social isolation causes a downregulation of pth2, and have identified a role for pth2 in the social behaviour of zebrafish (Anneser et al., 2022). We found strong support for this, and replicated the finding that expression of this gene was downregulated in socially isolated fish, and fully rescued after reintroduction (Anneser et al., 2020). The Biological Process Gene Ontology (GO) terms as well as the Cellular Component GO terms enriched in the upregulated genes are displayed in Figure 3. In terms of the biological process GO terms, we found several enriched clusters relating to muscle contraction and organism movement.
Similarly, in terms of the cellular components, the most enriched clusters were related to muscle function and activity (e.g., actin cytoskeleton, sarcomeres, myofibrils, and contractile fibres): these are all components of muscle tissue that are involved in generating force and movement.

[FIGURE 3 HERE]
We next used the Harmonizome database to identify whether there is an association between the identified DEGs and three common disorders/diseases linked to social isolation: anxiety, depressive disorders and cardiovascular diseases. Twenty-four of the DEGs were found in the gene sets "Depressive Disorder" and "Anxiety Disorders" from the CTD Gene-Disease Associations dataset, and twenty of the DEGs were found in the "Cardiovascular Diseases" gene set. Table 2 shows the relative strength of association between these DEGs and anxiety and depressive disorders as well as cardiovascular diseases expressed in standardized values, which are related to empirical p-value as follows |standardized value| = −log10(p-value). We found high overlap between several of the DEGs and the genes associated with all three diseases J o u r n a l P r e -p r o o f in the Harmonizome database, suggesting a strong potential relevance of the DEGs to all three diseases.

J o u r n a l P r e -p r o o f
The differences in gene expression in the isolated fish compared to controls largely returned to normal after the reintroduction to age-matched counterparts. Interestingly, a single gene angptl4 was differentially expressed (upregulated) in the reintroduced fish vs controls, but there was no difference in the expression of this gene between isolated and social fish. This suggested that this change in expression was directly related to the process of reintroduction. We therefore repeated the isolation and conducted an independent RT-qPCR assay, which confirmed the differential expression of angptl4 in the reintroduced fish (Figure 4.A), where we identified a statistically significant difference in angptl4 expression between controls, 2-week isolated, and reintroduced adult fish ((χ 2 (2) = 8.660, p = 0.0055) (Figure 4.A). angptl4 expression was higher in the reintroduced fish compared to controls (p = 0.0400) as well as the 2-week isolated fish group (p = 0.0267) (Figure 4.A). As can been seen from Table 2, angptl4 is related to anxiety disorders and depression, so we hypothesized that there may be changes in expression of this gene associated with different durations of social isolation. To determine the time course of changes in gene expression, we examined the effects on angptl4 expression between controls and isolated fish after 48-hour and 6-day isolation. We found a statistically significant difference in angptl4 expression between 48-hour and 6-day isolated young adult fish and their age-matched social controls ((χ 2 (2) = 6.752, p = 0.0245) (Figure 4.B). The 6month-old fish, which were isolated for 6 days showed an increase in angptl4 expression compared to social controls (p = 0.0483), while an isolation period of 48 hours did not cause a change in angptl4 expression (p > 0.9999) (Figure 4.B).

[FIGURE 4 HERE]
To further explore the role of anxiety in social isolation we examined if socially isolated fish showed differences in responses in a validated test of anxiety in zebrafish, the novel tank test. We found no sex differences, so these data were not included in further analyses. We found no significant effect of the 48-hour isolation on the time that angptl4 expression and anxiety, with a correlation after a short isolation, but this normalised later (after 2 weeks). This suggested a homeostatic role of angptl4 expression following isolation. This was further supported by the increase in angptl4 expression following reintroduction to the social group (Figure 4.A).

Social isolation increases anxiety and affects angptl4 in 24-month-old fish
Social isolation is particularly problematic in older adults: in humans, social isolation in the elderly is a significant public health challenge and is related to physical and mental health decline and poor prognosis for both acute and long-term conditions  (Brock et al., 2017;Cleal et al., 2021;Gerhard, 2007;Keller and Murtha, 2004;Kishi et al., 2009;Yu et al., 2006).

[FIGURE 5 HERE]
To test the hypothesis that expression of angptl4 was important in social isolation in the ageing fish, we next compared levels of angptl4 gene expression between young adult and ageing social controls. We found that angptl4 expression is significantly lower in ageing (24 months) than young adult (6 months) fish (U = 0, p = 0.0079) ( Figure 5.A). Next, we examined the effect of social isolation on angptl4 gene expression in ageing fish. We found a statistically significant difference in angptl4 expression between controls, isolated, and reintroduced ageing fish (χ 2 (2) = 9.420, p = 0.0024), but interestingly, the effect was different in the ageing fish than in the young adults, with angptl4 expression higher during social isolation (vs social controls) in ageing fish (p = 0.0175) but this normalised again following reintroduction to their social group (p > 0.9999) (Figure 5.B). angptl4 expression was also significantly lower in ageing reintroduced fish compared to the isolated group (p = 0.0327) (Figure 5.B). Altered anxiety responses were seen in both young adult and ageing populations after 2 weeks of isolation. Previous work in several species has shown that social deprivation elicits anxiogenic effects (Amiri et al., 2017;Cacioppo et al., 2012;Dos Santos et al., 2010;Lukkes et al., 2009;Santini et al., 2020;Yasuda et al., 2016). We also observed an isolation time-dependent increase in anxiety in young adult zebrafish, as we found that 6-day isolated fish experience increased anxiety compared to their social counterparts, while those isolated for 48 hours have no change in anxiety compared to controls. Our observation that 48-hour isolation has no effect upon anxiety responses is not consistent with a recent study which showed that 48 hours of isolation induce an anxiogenic response in young adult zebrafish (Daniel & Bhat, 2022). However, the study protocol was different to ours with different degrees of isolation involved, which may explain the different anxiety response observed.
Furthermore, other studies have even shown anxiolytic responses following long-term isolation (Parker et al., 2012;Giacomini et al 2015;Shams et al 2015).
We found several DEGs associated with 2-weeks isolation of 6-month-old zebrafish, the majority of which were completely rescued by reintroduction. We, and others (Anneser et al., 2022(Anneser et al., , 2020 have found that pth2 is significantly downregulated in socially isolated fish, and is completely rescued by resocialisation. This gene, J o u r n a l P r e -p r o o f however, has not been linked to anxiety, depression or cardiovascular/metabolic diseases in humans, suggesting that it may not be translationally relevant as a marker of social isolation. Human orthologs of several of the DEGs following social isolation are associated with anxiety, depressive disorders, and cardiovascular disease. By cross checking our DEGs against the disorders linked to social isolation in the Harmonizome database, we identified several DEGs that were found to have high overlap with anxiety, depression and cardiovascular diseases. This included several genes that were related to two or more of these disorders, including actc1, ckm, col1a1, col1a2, gapdh, myh6, nme2, and pkm. There is some suggestion that the glycolytic protein glyceraldehyde-3-phosphate dehydrogenase (GAPDH) may be involved in the development of depressive symptoms in animal models  and in human GWAS of treatment-resistant depression (Fabbri et al., 2019).
However, the other genes have not, to our knowledge, been linked to social isolation and psychiatric or physiological disorder in the past. While the specific biological mechanisms underlying these conditions are not yet fully understood, there is evidence to suggest that there may be shared genetic and biological pathways involved in their development and progression. Thus, the identification of genes that are associated with all three conditions in the Harmonizome database suggests that this gene may be involved in shared pathways or processes that contribute to the development of these disorders. However, it is important to note that the overlap of these DEGs with the disorders does not necessarily imply a direct causative relationship. Anxiety, depression, and cardiovascular disease are complex disorders with multifactorial aetiologies. Further research will be needed to investigate the specific mechanisms by which the identified genes contribute to the pathogenesis of these disorders, and the potential therapeutic implications of targeting these genes.
Finally, because reintroduction of the fish to their social group rescued the majority of these DEGs, these effects are transient, meaning that the negative effects of isolation are likely to be reversible.
In both the biological process and cellular GO terms, we found several enriched terms related to muscle physiology, including muscle contraction and general organism movement factors (biological processes) and actin cytoskeleton, sarcomeres, myofibrils, and contractile fibres. Collectively, this suggests that social isolation had a significant impact on muscle structure and function. When an individual is socially J o u r n a l P r e -p r o o f isolated, they experience increased levels of stress, anxiety, and even depression (Ahmadi Sarbarzeh et al., 2020;Mautong et al., 2021). We found that the isolated zebrafish show increased anxiety in the novel tank test, supporting the translational relevance of this effect. Chronic anxiety can lead to a decrease in physical activity, which can then lead to a reduction in muscle strength and endurance. In addition, social isolation can also affect an individual's motivation to engage in physical activity (i.e., 'depression-like' phenotypes), which can further exacerbate the negative effects on muscle function. Furthermore, social isolation can also impact an individual's movement patterns: social interaction provides opportunities for movement and physical activity, such as swimming around with shoal-mates. Without these social opportunities, the fish may become more sedentary, which can lead to a decrease in movement variety and coordination. Interestingly, the DEGs were transient, and therefore appear to be adaptive changes, as they were completely rescued by reintroduction to the social group.
Despite the majority of the DEGs being rescued following 24 hours of reintroduction to the social group, there was one gene, angptl4, in which the expression was unchanged after social isolation of young adult fish but was significantly upregulated after resocialisation. angptl4 expression was significantly lower in (non-isolated) ageing fish compared to non-isolated adult fish, suggesting that expression decreases with age in social populations. However, in ageing fish, social isolation induced upregulation of angptl4 which was rescued by resocialisation. These diverse transcriptomic changes suggest a functional role for angptl4 in response to isolation, perhaps with upregulation following reintroduction in adults reflecting a protective homeostatic response against the effects of (potential) future isolation. It may be that younger adults habituate to being isolated, and assuming that angpl4 is playing an adaptive homeostatic role, it would be predicted that there would be an increase in expression on re-introductioni.e., it is acting as a protective mechanism against future isolation. This hypothesis was supported by the findings that adult fish isolated for a shorter period of 6 days exhibit increased anxiety responses in the Novel Tank test that are associated with an increase in angptl4 expression compared to social controls. For the ageing fish, howeverwho have had a lifetime of no isolationadaptive homeostasis will decrease with age (Pomatto and Davies, 2017). Therefore, older adults show sustained levels of anxiety, even after 2 weeks, and sustained J o u r n a l P r e -p r o o f increases in angpl4. The acute increase in angpl4 in the older animals during isolation suggests that at this stage, the change in expression represents an acute response, that decreases once fish are resocialised.
angptl4 encodes a protein that is involved in the regulation of lipid metabolism by acting as an inhibitor of lipoprotein lipase (LPL) activity (Dijk and Kersten, 2014;Köster et al., 2005;Vienberg et al., 2015) and thereby suppressing the clearance of triglyceride-transporting lipoproteins and leading to a rise in circulating triglyceride levels (Kersten, 2014;Wang and Eckel, 2009), which is a key feature of metabolic disorders such as obesity and cardiovascular disease (Fernández-Hernando and Suárez, 2020;Kumari et al., 2021;Young and Zechner, 2013). It has been suggested that angptl4 can be a potential target for the treatment of metabolic disorders, and individuals with a genetic loss of angptl4 function have improved glucose homeostasis, reduced risk of type 2 diabetes (Abid et al., 2016;Gusarova et al., 2018) and protection against cardiovascular disease (Abid et al., 2016;Dewey et al., 2016;Romeo et al., 2007). The metabolic role of angptl4 in zebrafish remains to elucidated; however, it is believed to be well conserved (Anderson et al., 2013). This is evidenced by both germfree zebrafish and mice introduced with microbiota displaying markedly reduced intestinal angptl4 expression (Bäckhed et al., 2004;Camp et al., 2012)). Therefore, here we postulate that the upregulated angptl4 expression will result in an inhibition of LPL activity and reduction in triglyceride clearance with decreased energy release and fat storage (Camp et al., 2012).
Ageing-related alterations in the genomic regulation of gene expression have been well documented in different human tissues (Frenk and Houseley, 2018;Glass et al., 2013;Lu et al., 2004;Rodwell et al., 2004), and ageing pathology has been found to be conserved across eukaryotes (Bou Sleiman et al., 2020;Frenk and Houseley, 2018). angptl4 upregulation in isolated ageing fish was accompanied by increased anxiety-related behaviour. Another study has found reduced anxiety and a decrease in angptl4 expression in whole brain tissue of adult zebrafish after exposure to clozapine, a schizophrenia medication (Viana et al., 2020). This suggests that a link might exist between anxiety, angptl4 expression, and that isolation may exacerbate this, particularly in older animals. However, resocialisation rescues this effect, so it J o u r n a l P r e -p r o o f seems likely that if older people are isolated, re-socialisation may be sufficient to protect against negative effects.
Population-based studies have demonstrated that social isolation is correlated with a higher risk of cardiovascular disease (Cené et al., 2022;Valtorta et al., 2016;Xia and Li, 2018), diabetes (Brinkhues et al., 2017;Hackett et al., 2020), and metabolic syndrome (Henriksen et al., 2019). Metabolic syndrome is also higher in depressed compared to non-depressed individuals (Moradi et al., 2021), and there is a significant positive association between anxiety and metabolic syndrome (Tang et al., 2017).
Interestingly, metabolic syndrome has recently been linked to ANGPTL4, with ANGPTL4 gene SNP rs1044250 being identified as an independent risk factor for metabolic syndrome in the Shandong Han population (Tong et al., 2021). Given that different disorders that are associated with metabolic syndrome have been studied using zebrafish, such as hyperglycaemia, obesity, diabetes, hypertension and cardiovascular disease (Benchoula et al., 2019), zebrafish could be a very useful model to study the mechanisms linking social isolation and metabolic conditions; and angptl4 expression could play a key role in this process.

Limitations
We acknowledge several limitations. First, although zebrafish exhibit shoaling and exhibit basic social interactions, their social structure and behaviours are not as intricate as those seen in humans. This limits the direct applicability of zebrafish studies to understanding complex aspects of human social isolation. Second, zebrafish lack the cognitive and emotional complexity of humans. Social isolation in humans can have profound effects on mental health and cognition, whereas the emotional and cognitive experiences of zebrafish may not be directly comparable.
Third, our work here was carried out in controlled laboratory environments, which may not fully capture the complexity of human social interactions and the influence of social context. Human social isolation will be influenced by factors such as cultural, societal, and environmental aspects, which cannot ever be accurately modelled in zebrafish (or any other model species). Fourth, although zebrafish share a significant amount of genetic similarity with humans (~84% of psychiatric disease-related genes; (Howe et al., 2013), there are still notable differences in their biological and physiological systems. This may limit the translational relevance of findings from zebrafish studies to human conditions, so extrapolating these findings to human social isolation requires J o u r n a l P r e -p r o o f caution. Finally, although we measured the time course of the effects of social isolation in younger adult fish, we did not explore this in older fish. We also did not evaluate the potential age-dependent interactions linked to social states on gene expression and anxiety responses. To understand fully the role of angptl4 in ageing, these experiments could prove useful. They could be accompanied by detailed study on lipid metabolism following social isolation in order to understand the mechanisms.

Conclusions
Social interactions are essential for social species, and social deprivation can result in many negative effects including cognitive deficits and increased anxiety and depression (Pantell et al., 2013;Umberson and Karas Montez, 2010;Wilkialis et al., 2021b). Older people are more at risk of depression and anxiety, and negative cardiovascular effects that result from social isolation (Vrach and Tomar, 2020). Here, we found that social isolation of ageing zebrafish for only two weeks can have a negative impact on their anxiety responses. This was combined with a robust change in expression of angptl4. However, reintroduction to a social group rescued this change in gene expression, suggesting that the effects are reversible on resocialisation. These data suggest an adaptive homeostatic function of angpl4 in response to social isolation that changes across the lifespan. This may explain why older people are at heightened risk of negative physical effects of social isolation, such as cardiovascular problems, and why younger people are more resilient. Future research may examine the mechanisms by which social isolation affects angptl4 expression, and functional effects of this gene during development and the ageing process. In addition, future work may examine the time course of changes in gene expression to assess longer-term effects.