4. Discussion
As the nervous system forms the fundamental link between animals and the environments they inhabit (Kelley et al., 2018; O’Donnell, 2018), understanding the neurobiological impacts of environmental change is key to predicting how and why animals will respond to anthropogenic climate change. In this study, we sought to understand how projected end-of-century CO2 levels alter the nervous system at a molecular level, and how such changes may affect behaviour of the whole animal. To do this, we investigated the transcriptomic response to ocean acidification (OA) of the central and peripheral nervous system of a marine invertebrate with a complex nervous system, the two-toned pygmy squid Idiosepius pygmaeus. We then correlated patterns of gene expression with CO2 treatment levels and OA-affected behaviours in the same individuals. The central nervous system (CNS) and eyes of I. pygmaeus responded to elevated COwith significant differential expression (DE) of a small number of genes, and widespread small, coordinated changes of genes belonging to important functional categories between CO2conditions. Furthermore, we identified 169 genes in the CNS, six genes in the eyes and ten genes in both tissues that were correlated with CO2 treatment and one or more behaviours affected by OA, indicating these genes potentially contribute to OA-induced behavioural changes.
The GABA hypothesis is the predominant mechanistic explanation for OA-induced behavioural changes in fish (Nilsson et al., 2012; Tresguerres & Hamilton, 2017) and may also apply to marine invertebrates (Thomas et al., 2020). Pharmacological work has supported the GABA hypothesis in marine molluscs (Clements et al., 2017; Watson et al., 2014), including in I. pygmaeus (Thomas et al., 2021). In the whole-body of a pteropod mollusc, a GABAA receptor transcript was upregulated after OA exposure (Moya et al., 2016). In fish nervous tissue, OA exposure has variable effects on GABAA R subunit transcript expression, causing upregulation in some species (Cohen-Rengifo et al., 2022; Schunter et al., 2018) but not another (Williamset al., 2019). Furthermore, differences in OA exposure duration (Kang et al., 2022; Schunter et al., 2018; Schunter et al., 2016) and magnitude (Toy et al., 2022) are associated with variable effects on the expression of genes coding for GABAA R subunits within species. In I. pygmaeus, we found small, coordinated downregulation in the CNS and upregulation in the eyes of genes for ion-channel receptors, which included GABAA receptor subunit transcripts. In the CNS, there was also upregulation of syvn1-b (implicated in GABAAα1 receptor subunit degradation (Crider et al., 2014; Jiao et al., 2017)), regulators of GABAergic neurotransmission were positively correlated with CO2treatment and behaviours (phf24 , rac1), and aldh5a1(involved in the final degradation step of GABA (Kim et al., 2009)) was negatively correlated with CO2 treatment and behaviours. Together, this data suggests an effect of OA on GABAergic signalling may be widespread, occurring not only in fish but also marine molluscs, however, effects may be species and tissue-specific.
Recent research suggests that various other types of neurotransmission may also be affected by OA and potentially contribute to consequent behavioural responses. Pharmacological research has identified altered function of a range of different types of ligand-gated Cl- channels in I. pygmaeus (Thomas et al., 2021) and the dopamine D1 receptor in a damselfish (Hamilton et al., 2023) contributing to OA-induced behavioural changes. Elevated CO2 upregulated glycinergic, cholinergic and glutamatergic transcripts in the whole body of a pteropod mollusc (Moya et al., 2016), upregulated glutamatergic transcripts in the non-nervous tissue of oysters (Ertl et al., 2016; Wang et al., 2020), and altered acetylcholine receptor transcript expression in the whole body of another pteropod mollusc (Johnson & Hofmann, 2017). In fish nervous tissue, exposure to OA conditions caused upregulation of genes coding for glutamatergic and cholinergic neurotransmission of some species (Cohen-Rengifo et al., 2022; Schunter et al., 2018; Williams et al., 2019), but downregulation in others (Kanget al., 2022; Porteus et al., 2018). In I. pygmaeus, we identified small, coordinated downregulation in the CNS and upregulation in the eyes of genes involved in neurotransmission mediated by ligand-gated ion channels, including transcripts for subunits of ionotropic glutamate, glycine and acetylcholine receptors. There was also small, coordinated downregulation in the CNS of G protein-coupled receptor (GPCR)-mediated neurotransmission, including genes for subunits of metabotropic glutamate, serotonin, dopamine and GABA (GABAB) receptors. Key genes for glutamatergic and monoaminergic signalling were DE, and a subunit of nicotinic acetylcholine receptors (chrna10) was correlated with CO2 treatment and behaviours in both tissues. Overall, this suggests that OA not only impacts GABAA R function, but various different types of neurotransmission mediated by ligand-gated ion channels and GPCRs. However, as with GABAergic signalling, OA effects on other types of neurotransmission may vary by species and tissue type, and possibly other factors such as the magnitude and duration of OA exposure. Further experimentation is needed to understand how OA-induced transcriptomic responses translate to neurotransmission function and behaviour.
Genes involved in the general processes required for synaptic neurotransmission were attenuated in the CNS of I. pygmaeus after OA exposure, including Ca2+, K+ and Na+ ion channels required for maintenance of membrane potential, action potential generation and neurotransmitter release. There was also a negative correlation between the expression of genes regulating synaptic neurotransmission with CO2 treatment and activity traits. Genes for Ca2+ and K+ transporters and the regulation of neurotransmitter release were also downregulated in the brain of spiny damselfish collected from CO2 seeps (Kang et al., 2022), but upregulated after acute and developmental OA exposure (Schunter et al., 2018), and in fish olfactory tissue after short-term (Williams et al., 2019) and transgenerational OA exposure (Cohen-Rengifo et al., 2022). These transcriptomic signatures suggest an even more widespread effect of OA on neurotransmission, potentially altering the general processes required for synaptic neurotransmission to occur. However, experimentation is required to determine functional effects.
Neuroplasticity is the ability of the nervous system to change. Both neurogenesis (the process by which new neurons are generated and integrated into existing neural circuits) and synaptic plasticity (changing of synaptic strength over time) contribute to neuroplasticity (Costandi, 2016). In the CNS of I. pygmaeus we found small, coordinated upregulation and positive correlation with CO treatment and activity traits of genes involved in all the stages required for neurogenesis (re-entering and exiting the cell cycle, cell proliferation and differentiation to form new neurons, and neural wiring involving cell migration and adhesion for new neurons to be incorporated into existing circuits). Widespread upregulation and positive correlations with CO2 treatment and behaviours of genes in the CNS of I. pygmaeus involved in transcription, RNA processing, and protein processing could potentially be a response to deal with the changed protein demand required due to increased neuroplasticity. Notably, cdk10, which plays an important role in neurogenesis (Yeh et al., 2013) was identified as correlated with the OA-induced increase in exploratory interactions in both the CNS and eyes. Genes involved in synaptogenesis (the formation of new synapses) and synaptic plasticity were also positively correlated with CO2 treatment and OA-affected behaviours in the CNS of I. pygmaeus .
Despite not assessing the nervous tissue specifically, transcripts involved in neuronal cell adhesion, neuronal differentiation and survival and synaptic plasticity were also upregulated in the whole body of a pteropod mollusc after OA exposure (Moya et al., 2016). In fish nervous tissue, genes involved in neurogenesis were upregulated in some species but not others (Lai et al., 2017), and genes involved in synaptic plasticity were upregulated in some species (Cohen-Rengifo et al., 2022; Schunter et al., 2018), but downregulated in another species (Porteus et al., 2018). Thus, OA-induced transcriptomic responses related to neuroplasticity may be widespread, occurring in fish and marine molluscs. However, this response may be taxa-specific and/or could be affected by differences in CO2 exposure duration and magnitude that have differed between studies.
Elevated CO2 alters the molluscan immune response, with most research focusing on bivalves (Bibby et al., 2008; Liet al., 2015; Liu et al., 2016; Su et al., 2018; Wu et al., 2016)), though the immune response of an octopus was also affected by elevated CO2 (Culler-Juarez & Onthank, 2021). Here, we found DE of genes that regulate immune signal transduction pathways and which are also implicated in the molluscan immune response (map4k5/3, syvn1-b, psenen , cbs) (Canesi et al., 2006; De Zoysa et al., 2010; Goodson et al., 2005; Salazar et al., 2015). There were also changes in expression of a range of genes that code for immune effectors, including iron sequestration (tf and cbs), autophagy (map1l3ca/b), controlling the pool of available nucleoside triphosphates (nme6), and phagocytosis (‘cell adhesion’ and multiple cytoskeleton functional categories). Previous research has also indicated altered phagocytosis in molluscs at elevated CO2; adhesion capacity of haemocytes was decreased in a clam and expression of integrin (involved in cell adhesion for phagocytosis) was decreased in an oyster species (Ivanina et al., 2014) and increased in another oyster (Ertl et al., 2016). Furthermore, the phagocytic rate and cytoskeleton component abundance was decreased, and the expression of cytoskeleton genes was upregulated, in a clam at elevated CO2 (Su et al., 2018). Our results show an effect of OA on the transcriptional profile of genes implicated in the immune response suggesting OA-induced alterations in immune function may also occur in molluscan nervous tissue, though further research directly measuring immune function within the nervous system is required.
Cross-talk between the neuroendocrine and immune systems coordinates appropriate physiological and behavioural responses to environmental change (Demas et al., 2011). In molluscs, neuronal release of norepinephrine regulates immune responses through a neuroendocrine-immune axis-like pathway (Liu et al., 2017). Specifically, changes in the expression and activity of maoa (upregulated in I. pygmaeus eyes here) plays a key role in immune functioning via norepinephrine in molluscs (Liu et al., 2018; Sunet al., 2021; Zhou et al., 2011). Immune-derived factors can also feedback to alter the nervous system and behaviour (Adamo, 2006; Dantzer & Kelley, 2007). Indeed, tf, which is a key component of the molluscan innate immune response (Lambert et al., 2005; Ong et al., 2006; Herath et al., 2015; Salazar et al., 2015; Li et al., 2019) was upregulated and positively correlated with COtreatment and activity traits in the CNS of I. pygmaeus. We also found the expression of genes coding for integrins itga4 and itga9, cell adhesion molecules playing a key role in invertebrate immune responses (Johansson, 1999; Terahara et al., 2006), were positively correlated in the CNS with CO2 treatment and activity traits. Therefore, it is possible that OA-induced changes in neurotransmission could have consequences on immune function, and changes in immune function could also feedback on the nervous system to alter behaviours at elevated CO2. However, the potential links between OA, neurotransmission, immune function and behaviour remain to be experimentally tested.
Oxidative stress occurs when there is an imbalance between the production of reactive oxygen species (ROS) and protection by antioxidant mechanisms (Halliwell & Gutteridge, 2015). Elevated CO2 induces oxidative stress in molluscs, increasing ROS and altering antioxidant defences inducing DNA damage, lipid peroxidation and apoptosis (Cao, Liu, et al., 2018; Cao, Wang, et al. , 2018; Kim et al., 2023; Tomanek et al., 2011; Wang et al., 2016; Zhang et al., 2021). In I. pygmaeus , we found DE of genes implicated in the production of antioxidants, including ascorbate and glutathione. We also identified upregulation in the CNS of genes involved in DNA damage and repair, protein damage and endoplasmic reticulum stress, and cellular stress-induced apoptosis. In molluscs, OA exposure has previously been shown to result in DNA damage (Cao, Liu, et al., 2018; Nardi et al., 2018) and increased apoptosis (Cao, Liu, et al., 2018; Zhang et al., 2021).
The nervous system is particularly vulnerable to oxidative stress (Halliwell, 2006; Valko et al., 2007) and oxidative stress-induced damage within the nervous system can disrupt neurotransmission and neuronal function (Bouayed et al., 2009; Halliwell, 2006; Halliwell & Gutteridge, 2015; Lebel & Bondy, 1991). In mammals, a link between oxidative stress in the nervous system and changes in behaviour has been demonstrated (Bhatt et al., 2020; Bouayed, 2011; Rammal et al., 2010). In the CNS of I. pygmaeus , we identified a positive correlation between the expression of genes implicated in oxidative stress and COtreatment and OA-affected behaviours. In the eyes, two genes (crb , zranb1) potentially correlated with OA-induced behavioural alterations of I. pygmaeus are implicated in oxidative-stress induced retinal degeneration (Chartier et al., 2012; Wang et al., 2018). In particular, crb prevents photoreceptor degeneration by limiting the production of ROS and the resultant oxidative damage (Chartier et al., 2012). A recent study in a cuttlefish found the behavioural effects of OA were associated with an altered retinal structure and an increase in apoptotic cells within the eyes (Xie et al., 2023). Thus, it’s possible that OA-induced oxidative stress could contribute to behavioural alterations at elevated CO2, potentially through central and peripheral mechanisms, but further electrophysiological and whole-animal behavioural experimentation is required.
When interpreting our results, there are a few important things to consider. Firstly, despite the reasonable assumption that changes in gene expression driving behavioural responses occur prior to behavioural production (Fischer et al., 2021), we measured gene expression immediately after the OA-induced behavioural responses in I. pygmaeus due to the necessity of terminal sampling to obtain nervous tissue. Furthermore, the process of transcribing genes is far too slow to mediate rapid behavioural responses, which are instead mediated by fast electrical signals passed along and between neurons (Fischeret al., 2021). Thus, our transcriptomic results do not describe the neuronal mechanisms driving the immediate behavioural responses to a stimulus, but rather those that likely contribute to longer-term changes in behaviour (Clayton et al., 2020; Fischer et al., 2021) in OA conditions. Secondly, organism responses to OA can be sex-specific (Ellis et al., 2017), including marine invertebrate behavioural responses (Marčeta et al., 2020; Richardson et al., 2021). We used males only in this study. Future research could consider using both sexes to determine whether behavioural responses to OA are sex-specific, and if so whether differing transcriptional profiles underly these sex-specific responses.