Among the first observations in humans were that prenatal famine exposure and lower birth weight are both associated with cardiovascular and metabolic diseases, as well as cognitive alterations and mental health disorders (Talge et al., 2007; Van den Bergh et al., 2017). This phenomenon can be modelled in zebrafish by exposing adult females to a starvation stress prior to breeding. A relatively short duration of starvation may not affect body size or gross morphology, but can nonetheless have profound effects on the developing brain and behaviour, including hypoactivity (Fan et al., 2019), and suppressed cell proliferation in the forebrain of the early larva (Higuchi, 2020). The mechanisms via which maternal starvation influences offspring development and function are unclear, but given that maternal cortisol levels are increased in zebrafish embryos from starved mothers (Higuchi, 2020) and that cortisol exposure can alter early forebrain neurogenesis in zebrafish larvae (Best et al., 2017) in a similar manner to that seen in the early offspring of starved females, it seems likely that cortisol is a mediator.

Behavioural studies following maternal cortisol exposure in zebrafish have reported a variety of phenotypes including increased boldness (Best et al., 2017) and altered anxiety-like behaviour (Van Den Bos et al., 2019) of offspring. The study by Best et al. (2017) shows that zebrafish larvae exposed to increased maternal cortisol display hyperactivity during the light period of a light-dark assay, and reduced thigmotaxis, or wall-hugging behaviour, in an open field assay. These behaviours, suggestive of increased exploration and reduced anxiety respectively, together indicate a tendency towards increased boldness. Further work has suggested that strain can influence the effects of maternal stress on behaviour in the offspring (Van Den Bos et al., 2019).

The mechanisms underlying behavioural alterations in zebrafish larvae following exposure to increased maternal cortisol are only beginning to be investigated. Initial studies have uncovered alterations in the developing brain, which correlate with the behavioural abnormalities. Best et al. (2017) detected an increase in primary neurogenesis in the preoptic region and pallium, as well as increased brain expression of proneural gene neurod4 in zebrafish embryos following exposure to maternal cortisol, suggesting that maternal cortisol may alter brain development in a region-specific manner. Additionally, D’Agostino et al. (2019) reported increased basal expression of cfos in prenatally stressed zebrafish larvae, as well as altered reactivity of cfos expression level to an acute stressor. The cfos gene is often used as a marker of neuronal activity, and as such, these data suggest that prenatal exposure to cortisol might affect neuronal function in the basal state, as well as during stimulus processing. Similarly in humans, the offspring of mothers who experienced high levels of anxiety during pregnancy exhibited altered patterns of brain activation during cognitive control tasks (Mennes et al., 2019).

The zebrafish studies described here suggest that the zebrafish system can be used to model the effects of maternal stress and that many of the effects of maternal stress in humans and other animal models can be recapitulated in zebrafish. The zebrafish system offers some advantages over other animal models, including the ability to mimic increased deposition of maternal cortisol under highly controlled conditions, as well as ease of studying the concurrent effects of stress on the developing brain and behaviour in a high-throughput manner. In addition to maternal stress, studies in rodents have demonstrated that paternal stress can also influence behaviour and physiology in the offspring (Yeshurun and Hannan, 2019). This sperm-driven inheritance is thought to be mediated at least in part, by non-coding RNAs (Rodgers et al., 2015). Whilst this area has not yet been well studied in zebrafish, a recent study demonstrated that male zebrafish that were exposed to stress during adulthood had altered levels of several small non-coding RNAs in the spermatozoa, and that their offspring subsequently displayed weakened endocrine and behavioural responses to acute stress (Ord et al., 2020).

In addition to transmitting excess cortisol to the developing foetus, human mothers can also expose the foetus to alcohol via the placental blood supply. Animal models, more recently including zebrafish, have been used to model Foetal Alcohol Spectrum Disorder (FASD). FASD is caused by prenatal alcohol consumption, which affects the developing foetus, and has a prevalence of 2–5% in the human population (May et al., 2009). In humans, severe exposure is often associated with gross skeletal and craniofacial abnormalities, and severe central nervous system (CNS) dysfunction, whilst moderate exposure is associated with more subtle cognitive and behavioural defects including depression, vulnerability to stress, impulsivity and inattention and antisocial behaviour. These more moderate cases of FASD are estimated to make up 98% of all cases, and are often left undiagnosed or misdiagnosed (Clarren et al., 2015).

Given the utility of the zebrafish as a shoaling animal, attention has focused on the effects of lower doses of ethanol given during early life on social behaviour. Despite variation in exposure level, developmental timing of exposure, timing and experimental set-up of behavioural assays, several zebrafish studies have reported that ethanol exposure during early development results in impaired social behaviour in adulthood (Fernandes and Gerlai, 2009; Buske and Gerlai, 2011; Parker et al., 2014; Fernandes et al., 2015, 2019a, b). The observed defects in social behaviour were apparent in various contexts, including when an individual fish was presented with the visual cue of a live shoal (Parker et al., 2014), or a computer image (Fernandes et al., 2015), and when fish were observed in a group of conspecifics (Parker et al., 2014). The impairments in social behaviour in zebrafish models of FASD fit well with the reduction of social skills seen in children with FASD (Stevens et al., 2015) and social impairments in mammalian models (Varlinskaya and Mooney, 2014). This suggests that zebrafish models of FASD may be able to provide crucial mechanistic insights into the aetiology of abnormal social behaviour resulting from developmental ethanol exposure.

Although studies have only just begun to investigate the mechanisms underlying impaired social behaviour resulting from developmental ethanol exposure in zebrafish, the results suggest that the dopaminergic neurotransmitter system plays a role. A number of studies have detected a reduction in levels of dopamine and its metabolite, 3,4-Dihydroxyphenylacetic acid (DOPAC), in the brains of adult fish exposed to ethanol during early development, compared with controls (Buske and Gerlai, 2011; Mahabir et al., 2014; Fernandes et al., 2015). Data from Fernandes et al. (2015) suggests that this represents not a reduction in baseline levels, but rather a reduction in the dopaminergic response to a stimulus. If specific to the shoaling stimulus, this could suggest that ethanol-exposed fish find the conspecific cue less rewarding, perhaps due to impaired dopaminergic signalling, leaving exposed fish with a reduced motivation to socialise.

In addition to the defects in neurotransmitter systems following developmental exposure to ethanol, further alterations in brain development exist. These include altered morphology of the midbrain-hindbrain boundary (MHB), reduced expression of pax6a in the eye (Zhang et al., 2017), forebrain, and hindbrain, as well as reduced expression of epha4a in rhombomere 1 of the hindbrain, which point towards altered brain development, including cerebellar development resulting from ethanol exposure. Further, early ethanol exposure has been shown to contribute to altered neurogenesis during early zebrafish development. Specifically, exposure corresponded with a clear reduction in differentiated neurons in the hindbrain and spinal cord (Joya et al., 2014). The defect was found to relate to populations of sensory neurons, where decreased cell proliferation and increased apoptosis was observed. Whilst these observations are interesting, it is not clear to what extent such alterations might contribute towards behavioural phenotypes and further investigation of the effects of developmental ethanol exposure on brain development are warranted.

One seemingly important factor in modulating brain development and behaviour in zebrafish is movement during early life. During early development, restraint stress is associated with reduced progenitor cell proliferation in the zebrafish forebrain (Hall and Tropepe, 2018). In restrained animals, neuronal differentiation is increased at the expense of self-renewal, thereby depleting the progenitor pool. This, and to a lesser extent increased apoptosis, lead to smaller forebrain size in exposed larvae. In contrast, forced swimming during early life is associated with increased progenitor cell proliferation in the larval zebrafish pallium, and reduced neuronal differentiation (Hall and Tropepe, 2018). This finding suggests the opposite observation to that of restrained larvae: maintenance of an expanded progenitor pool at the expense of neuronal differentiation. In another study, an involuntary swim paradigm resulted in dysregulation of the HPA axis (Castillo-Ramírez et al., 2019). Here, in a later larval stage, animals that had experienced involuntary swimming during early development reconfigure their cortisol response to homotypic and heterotypic stressors. These results show how early life stress exposure can be linked to altered glucocorticoid function later in life in zebrafish.

Hall and Tropepe (2018) determined that it is physical cues associated with movement in early life that alter neurogenesis in the forebrain and these cues are conveyed by dorsal root ganglia. The mechanism by which movement affects neurogenesis in the zebrafish forebrain is not clear, although changes in cell cycling, cell fate and apoptosis may all play a part. Although glucocorticoids are known to modulate neurogenesis, it is not clear whether cortisol plays a direct role in movement-mediated alteration of forebrain neurogenesis or behaviour. As both restraint and forced swimming are expected to increase cortisol level, their differing effects on neurogenesis do not support a direct cortisol effect on neurogenesis in these paradigms. In another study, cortisol administration has been associated with increased progenitor cell proliferation in the early zebrafish forebrain and with behavioural alterations (Best et al., 2017). Increased forebrain proliferation is also observed after forced swimming (Hall and Tropepe, 2018), suggesting a potential role for cortisol in mediating movement-mediated alteration of forebrain neurogenesis. The difference in neurogenesis phenotypes following early life restraint or forced swim could reflect the different ways in which these paradigms modulate the HPA axis.

The effects of social isolation during development on the brain and behaviour can be readily studied in zebrafish, since the lack of maternal care means that, unlike in mammals, the effects of social deprivation can be separated from other variables connected with maternal care, such as nutrition. The locomotor startle response is influenced by rearing environment in zebrafish: larvae raised in isolation exhibit a blunted startle response to a dark pulse in comparison to group-reared larvae, and this is thought to be attributable to a lack of tactile stimulation (Steenbergen et al., 2011). It has also been observed that socially isolated larvae exhibit enhanced social avoidance of conspecifics during social interactions, an effect mediated by lateral line inputs (Groneberg et al., 2020). Similarly, another study found that developmental isolation reduced social preference in juvenile fish and these fish were more anxious (Tunbak et al., 2020). In this study, isolation also lead to increased visual sensitivity and increased activity in the posterior tuberal nucleus of the brain, a region associated with the stress response, suggesting that in isolated fish heightened visual sensitivity leads to increased stress during social viewing, which may lead to the observed social aversion. On the other hand, another study found that unlike fish that were chronically isolated, developmentally isolated fish did not exhibit any changes in anxiety or social behaviours in adulthood (Shams et al., 2018). These data suggest that timing and duration of isolation may affect the behavioural outcomes.

A separate study also suggested that long-term isolation may affect the stress response, since isolation-reared zebrafish were found to have lower levels of dopamine and serotonin’s metabolite, 5-Hydroxyindoleacetic acid (5-HIAA), following exposure to an unpredictable chronic mild stress, when compared to group-reared fish (Fulcher et al., 2017), however, it was also observed that developmental isolation had no significant effects on baseline or stress-induced cortisol levels (Shams et al., 2020). Interestingly, in another study, developmentally isolated fish (isolated 0–7 dpf) were observed to have significantly reduced levels of DOPAC, but not dopamine, serotonin or 5-HIAA, when compared with social-reared fish in adulthood (Shams et al., 2018). Further, Varga et al. (2020) observed that social isolation of fish during a critical period of behavioural metamorphosis that occurs in the juvenile stage affects later behavioural responsiveness and serotonergic signalling in the dorsal pallium. Blockade of 5-hydroxytryptamine (5-HT) signalling was sufficient to prevent the serotonin-associated brain changes and behavioural deficits, suggesting that a serotonergic mechanism, sensitive to environmental manipulation during a specific time window, modulates context-dependent behaviours in later life.

Long-term isolation of zebrafish, such as from 6 dpf until 6 months, can have profound effects on brain development in a region-specific manner. Lindsey and Tropepe (2014) showed that isolated fish exhibited a significant reduction in proliferation in sensory niches of the brain, but not in the telencephalon and that these changes could not be consistently rescued by exposure to social novelty. An interesting recent study shows that social isolation reduces the expression of neuropeptide pth2, and remarkably a brief exposure of isolated fish to a social group can rescue this (Anneser et al., 2020).

One route by which activation of the HPA axis may occur is by exposure to chemicals. A large number of studies have tested the effects of exposure to chemical substances in early development on the brain and behaviour in zebrafish. These toxicological studies have relevance both as models of the adverse effects of chemical exposure in humans and as studies of the effects of such chemicals on fish, since human actions can cause certain chemicals to contaminate aquatic environments. This vast amount of literature is outside of the scope of this review and the readers are referred elsewhere (Hill et al., 2005; Dai et al., 2014; Lee et al., 2015). Instead, here we will mention two exemplary studies testing the effects of chemicals that alter the HPI axis.

Bisphenol A (BPA) is commonly used in the production of plastics and other consumer products and in rodent models prenatal exposure to BPA is associated with the development of behavioural challenges in later life (Kinch et al., 2015). Treatment of embryonic zebrafish with BPA results in precocious neurogenesis in the developing hypothalamus, as well as hyperactivity during the larval stage (Kinch et al., 2015). This work identified an unexplored pathway through which BPA exerts its effects on neurogenesis and behaviour: via androgen-receptor mediated up-regulation of aromatase (Kinch et al., 2015). It is not clear, however, how altered aromatase activity might affect neurogenesis and behaviour. In other animals, BPA exposure is known to cause dysregulation of the HPA axis (Panagiotidou et al., 2014). Given that cortisol is known to affect neurogenesis and behaviour in the larval zebrafish (Best et al., 2017), and that cortisol is known to interact with sex steroids, this may represent a downstream effect of BPA exposure.

Another example that has been studied in zebrafish is exposure to fluoxetine, a selective serotonin reuptake inhibitor (SSRI), which is widely prescribed to treat depression in humans. Fluoxetine alleviates symptoms of depression by inhibiting 5-HT reuptake transporters on presynaptic neurons, thereby enhancing serotonergic neurotransmission (Stahl, 1998). However, concerns have been raised regarding the effects of exposure to fluoxetine during developmentally sensitive periods (Suri et al., 2015), such as exposure to the foetus when taken by pregnant women, and exposure to children and adolescents, to whom fluoxetine may also be prescribed (Jane Garland et al., 2016). Fluoxetine is widely prescribed globally and following excretion can enter aquatic environments where variable concentrations have been quantified in different water bodies (Parolini et al., 2019).

Exposure to fluoxetine during early life in zebrafish is associated with modulation of stress-associated genes and key neurotransmitter transporters in larvae (Parolini et al., 2019). These changes correlate with a reduced locomotor response to a touch stimulus. Early life exposure to fluoxetine also has long-term effects on the HPA axis and behaviour in zebrafish and these effects depend on the dose and timing of exposure, and are also sex-specific. Exposure during embryonic – early larval, larval, or juvenile stages lead to a blunted stress response in adult males. Additionally, the reduced cortisol levels resulted in reduced locomotion and exploratory behaviour following embryonic-early larval stage and larval fluoxetine exposures, but not following juvenile exposure (Vera-Chang et al., 2018). Meanwhile fluoxetine exposure during juvenile stages lead to reduced basal cortisol in females but not males (Vera-Chang et al., 2019). Further, in adult females, exposure during the juvenile, but not larval stage resulted in increased locomotion and exploratory behaviour.

The observed hyporeactivity of the stress response following early life fluoxetine exposure in zebrafish is consistent with observations in children prenatally exposed to SSRIs (Oberlander et al., 2008) and chronic blunting of basal cortisol levels has been linked with a number of psychological disorders in humans (Wichmann et al., 2017). One of the most intriguing findings from the studies of fluoxetine exposure, is that the suppression of cortisol levels and altered behaviour persists for multiple generations in unexposed fish (Vera-Chang et al., 2018). The HPA axis and behavioural defects may well be mediated by epigenetic alterations that in some cases are stably transmitted across generations, as has been previously observed following stress (Babenko et al., 2015). Indeed, 5-HT is known to alter GR expression via epigenetic modifications. The zebrafish data suggests that cortisol level maybe responsible for the observed behavioural alterations (Vera-Chang et al., 2018). Whether fluoxetine exposure alters behaviour in zebrafish via epigenetic modification of GR remains unknown. The analyses of the effects of fluoxetine exposure on the HPA axis and behaviour in zebrafish seem to be consistent with findings in rodent and human studies. Given the rapid external development of the early zebrafish and lack of maternal care, it is an advantageous system with which to model the effects of prenatal fluoxetine exposure in humans.

Brief exposure to anoxia during embryogenesis has also been shown to have profound and long-term effects on zebrafish behaviour. Anoxia-exposed zebrafish develop to become more dominant, exhibit more aggressive behaviours and have increased testosterone levels (Ivy et al., 2017). Interestingly, exposure to mild stress during early life may confer seemingly advantageous behavioural changes. In one study, zebrafish that were exposed to a combination of environmental enrichment and daily net chasing exhibited enhanced learning and reduced anxiety when tested as juveniles, and the reduction in anxiety behaviour was apparent even when fish were tested in adulthood after returning to standard housing for several months (De Pasquale et al., 2016).

In addition to acute and chronic exposures to individual stressors, paradigms are also established that use exposure to a combination of different stressful stimuli to induce ELS, such as chronic unpredictable early life stress (CUELS) (Fontana et al., 2020). Larvae exposed to a 7 day CUELS protocol were observed to have reduced anxiety levels as adult fish, meanwhile a 14-day exposure had the opposite effect (Fontana et al., 2021). The CUELS paradigm did not affect memory and cognition or social behaviour. On the other hand, exposure of juvenile fish to combined stress for 3 days had a positive effect on working memory in adulthood, while having no effect on shoal cohesion or anxiety behaviour. Interestingly, zebrafish exposed to a chronic unpredictable stress paradigm during the larval stage exhibited anxiety-like behaviours in the days following stressor exposure, but these effects were transient (Golla et al., 2020). These studies indicate that the timing and duration of exposure, as well as timing of sampling will influence the effects of ELS on behavioural outcomes.

The genetic tractability of the zebrafish has made it an attractive model organism for some time and the availability of CRISPR-Cas9 technology in recent years has extended this further (Albadri et al., 2017). Zebrafish lines with stable mutations that disrupt individual components of the HPA axis have been created and have begun to shed light on the development of the HPA axis and its role in brain development and behaviour, across the life-course.

One of the most widely studied HPA axis molecules is the GR, one of two receptors via which cortisol exerts its actions including the negative feedback on the HPA axis. A number of different mutants for nr3c1, the gene encoding the GR, have been created, and used to study a variety of processes, including development of the HPA axis and behaviour. Loss of GR function means that negative feedback to the HPA axis is compromised and thus leads to high basal cortisol levels in both larvae (Griffiths et al., 2012; Facchinello et al., 2017; Faught and Vijayan, 2018) and adult fish (Ziv et al., 2013) and the HPA axis is not responsive to an acute stressor (Ziv et al., 2013; Facchinello et al., 2017; Faught and Vijayan, 2018).

A number of studies have investigated the role of nr3c1 in behaviour, using various assays and the results of the different studies have sometimes conflicted. Reduced general locomotion has been widely reported in both larvae (Griffiths et al., 2012; Sireeni et al., 2020) and adult (Ziv et al., 2013; Sireeni et al., 2020) GR mutant fish in some studies, but not all (Faught and Vijayan, 2018). The first reports on a GR mutant showed that in larval stages mutants displayed a heightened startle response, in that they would not habituate to repeated stimulus exposure (Griffiths et al., 2012), and that as adults, they displayed freezing behaviour and an avoidance of the walls of a novel environment (Ziv et al., 2013), which the authors likened to depressive-like behaviour.

However, subsequent analysis of the same mutant revealed that mutant larvae exhibited a similar avoidance of the walls and confirmed the freezing behaviour in adulthood, however, this was accompanied by increased thigmotaxis in adulthood (Sireeni et al., 2020), the opposite effect to that observed previously. In contrast to the study of Ziv et al. (2013), the authors suggested that these behaviours were indicative of increased anxiety in fish lacking GR activity. An analysis of anxiety-like behaviour in larvae, however, revealed no significant effects, suggesting that in GR mutants, anxiety behaviour may only become apparent in adulthood. Further to the effects of loss of GR function on general locomotion and anxiety-like behaviour, a lack of feeding entrainment has also been reported in juvenile and adult GR knockout zebrafish (Morbiato et al., 2019). Whilst the authors focused on the requirement for GR in circadian rhythm, they concluded that clock gene expression was not responsible for the altered behaviour. This leaves the interesting possibility that other alterations in brain development or function such as alterations of the motivation system might be responsible for this behavioural change.

These studies of loss of GR function have highlighted the role of GR in anxiety- and depressive-like behaviours across the life course. In some cases, the research has also shown that the increased cortisol level resulting from loss of GR function is not directly responsible for the observed behavioural abnormalities: while behavioural defects can be normalised via treatment with fluoxetine, this is not accompanied by a reduction in cortisol level (Griffiths et al., 2012; Ziv et al., 2013). This suggests that a defect in serotonergic function in the brain of GR mutants, perhaps downstream of cortisol, is responsible for the observed behavioural defects. Indeed, glucocorticoids are known to influence the serotonergic system and both are implicated in the aetiology of affective disorders (Lanfumey et al., 2008). Despite this initial insight into the mechanism underlying behavioural defects in the GR mutant zebrafish, a comprehensive analysis into the effect of loss of GR function on the brain, including the serotonergic system, is lacking.

In contrast to GR, much less is known about the effect of MR manipulation in zebrafish. However, a recent analysis of MR^–/– in zebrafish reveals that MR is involved in stress-related behaviour, as knockout animals fail to exhibit glucocorticoid-mediated larval hyperactivity to a light stimulus (Faught and Vijayan, 2018). Meanwhile in another study, MR^–/– larvae exhibited no difference in their rapid locomotor response to an acute stressor, although the authors questioned whether they had achieved a complete loss-of-function in their MR knockout line (Lee et al., 2019). Further work is needed to establish whether MR plays a role in modulation of additional behaviours in zebrafish, such as anxiety-like behaviour, as is the case in mice (Rozeboom et al., 2007).

The HPA axis can also be manipulated via direct targeting of pharmacological agents. The most commonly used are synthetic cortisol and dexamethasone, both agonists of the GR. Stimulation of GR in early life initiates negative feedback to the HPA axis, leading to a reduction in basal cortisol levels (Wilson et al., 2013, 2016). In contrast, fish that were exposed to GR agonists in early life that were subsequently raised under normal conditions, exhibit normal basal cortisol levels but show a heightened HPI axis response to an acute stressor (Wilson et al., 2016).

Dysregulation of the HPI axis using GR agonist treatment in early life is also associated with behaviour changes in zebrafish. Whilst dexamethasone exposure in early life has not been associated with any effects on basal locomotion in zebrafish larvae (Steenbergen et al., 2011; Wilson et al., 2013, 2016), cortisol treatment has often been shown to cause hyperactivity (Best and Vijayan, 2017; Faught and Vijayan, 2018). Additional behavioural phenotypes in larvae include altered thigmotaxis in an open field test (Wilson et al., 2013; Faught and Vijayan, 2018), and a reduced locomotor response to an acute stressor (Steenbergen et al., 2011). Whilst few studies have investigated the long-term consequences of early life exposure to GR agonists such as dexamethasone in zebrafish, the results of initial studies are confusing. Whilst some reported increased anxiety-like behaviour, as measured in the novel tank diving test (Khor et al., 2013; Kumari et al., 2019), others, using the same assay, reported the opposite (Wilson et al., 2016). This conflict may be caused by the different dosing regime, which would suggest that varying the timing and duration of exposure could have profoundly different effects on behaviour. However, this notion requires further testing.

These studies highlight the utility of the zebrafish to uncover behavioural alterations following dysregulation of the HPI axis in early life. Initial studies suggest that cortisol signalling via MR, rather than GR plays an important role in regulating cortisol-induced hyperactivity following early life exposure (Best and Vijayan, 2017; Faught and Vijayan, 2018). Beyond this, the mechanisms underlying behavioural alterations following HPA axis dysregulation in early life remain elusive and the effects of such treatments on the brain have not yet been investigated.

Several studies also examined the effect of GR antagonist, mifepristone treatment in zebrafish. Early life exposure to mifepristone was observed to reduce basal locomotor activity in exposed larvae (Wilson et al., 2013). An additional study observed that, as in an nr3c1 mutant, mifepristone significantly decreased the locomotor response to an abrupt challenge in larvae (Lee et al., 2019). Further studies are required to establish whether early life mifepristone treatment has long-term effects on the brain and behaviour in zebrafish.

Optogenetics has great potential within the zebrafish field, due to the transparent nature of the larval body, rendering its tissues accessible to light. Tissues that are too small to be amenable to surgical manipulation can be non-invasively manipulated using cell type specific promoters, and this manipulation is aided by the genetic tractability of the zebrafish. Using optogenetic techniques to selectively modify the level of individual HPI axis components with a high degree of temporal precision has recently been demonstrated using the zebrafish larva (De Marco et al., 2016, 2013; Gutierrez-Triana et al., 2015).

De Marco et al. (2013) first demonstrated the technique in zebrafish by manipulating pituitary signalling. Using a cell-type specific promotor, beggiatoa photo-activated adenylate cyclase (bPAC) was targeted to corticotroph cells of the anterior pituitary that produce ACTH (Figure 1A; De Marco et al., 2013). Exposure of these transgenic larvae to blue light, or white light with a blue component, leads to increased cAMP levels specifically within bPAC-positive pituitary corticotroph cells. In pituitary corticotroph cells, cAMP increase causes release of ACTH, and thus in bPAC-positive larvae light exposure leads to ACTH-induced secretion of cortisol (Figure 1C). Using a similar technique, transgenic zebrafish lines have also been created in which bPAC is targeted to Steroidogenic acute regulatory protein (StAR)-expressing interrenal cells (Gutierrez-Triana et al., 2015; Figure 1B). In these larvae, blue light causes an increase in cAMP in steroidogenic interrenal cells, which stimulates cortisol synthesis (Figure 1C).

These studies demonstrated that such transgenic zebrafish lines can be used to study rapid non-genomic effects of hormone signalling on behaviour in free-swimming larvae (De Marco et al., 2016). Additionally, these transgenic lines can be used to study the delayed and long-term effects of early life stress, since exposure to blue light could induce hypercortisolaemia in transgenic larvae. The ability to specifically and precisely manipulate the level, duration and timing of HPA axis activation will allow critical questions in the ELS field to be addressed, such as how the exact GC level, exposure duration and developmental timing of exposure affects the brain and behaviour. This technique could also be combined with imaging and with high-throughput behaviour assays in zebrafish larvae to identify the effects of ELS on brain structure and function. However, it should be noted that the use of these cell-type specific promotors does not necessarily mean specific manipulation of individual HPA axis hormones. Indeed other molecules that co-localise to the same cell types manipulated in these studies may also be altered via light exposure and contribute to the phenotypic outcome.