Evolving views of ionic, osmotic and acid–base regulation in aquatic animals

Journal of Experimental Biology (JEB) has been communicating studies on comparative physiology since its very origin (Crew et al., 1923). Maintaining ionic, osmotic and acid–base (IOAB) homeostasis is essential for all organisms, and aquatic animals (see Glossary) have evolved diverse strategies to regulate the IOAB status of their biological fluids in response to metabolic and environmental challenges. This makes IOAB regulation by aquatic animals exceptionally well suited for comparative physiology studies.

All biochemical reactions are affected to some degree by ionic strength and pH. Ionic composition and gradients determine the transmembrane and transepithelial transport of molecules, water fluxes can lead to disruptive cell swelling or shrinking, and pH affects the function of proteins and enzymes. ATP hydrolysis produces H⁺ and thus cellular work is the most significant (and a continuous) challenge to acid–base (A–B) homeostasis. The majority of the H⁺ produced is consumed by mitochondrial oxidative phosphorylation (Hochachka and Mommsen, 1983), which establishes additional links between oxygen availability and A–B regulation (Tresguerres et al., 2020). However, living surrounded by water imposes interlinked advantages and restrictions on aquatic animals that are not experienced by air breathers. Chiefly, aquatic animals must maintain permeability with the surrounding water to sustain gas exchange, waste excretion and nutrient acquisition, which makes them prone to diffusive fluxes of gases, ions and water that can challenge their IOAB homeostasis. Additionally, the lower content of O₂ and the higher solubility of CO₂ in water relative to air increase the energetic cost of O₂ uptake, decrease the HCO₃⁻ buffering capacity of biological fluids and virtually rule out the ventilatory regulation of CO₂ excretion as an A–B regulatory strategy in aquatic animals. Instead, they largely compensate for an acidosis by actively excreting H⁺/NH₄⁺ in exchange for external Na⁺ and for an alkalosis by excreting HCO₃⁻ in exchange for external Cl⁻. These processes take place across the gills and skin and are mechanistically similar to IOAB regulation in the mammalian nephron. However, the IOAB conditions in aquatic environments can be highly diverse and variable, which drove the evolution of correspondingly diverse and dynamic physiological mechanisms to maintain IOAB homeostasis. For example, Na⁺ and Cl⁻ concentrations can range from ∼500 mmol l⁻¹ in seawater down to <10 μmol l⁻¹ in freshwater, which imposes thermodynamic constraints on the IOAB regulatory mechanisms used by aquatic animals in each water body and an intrinsic link between systemic A–B and ionic regulation in some freshwater species. Estuarine animals face changes in salinity resulting from tides and changes in freshwater input, and animals that migrate between seawater and freshwater experience even more pronounced changes. The levels of O₂, CO₂, pH and ammonia can dramatically change owing to the photosynthetic and respiratory activity of organisms, organic matter decomposition, upwelling and algal blooms. This induces hypoxia, hyperoxia, hypercapnia (see Glossary) and high environmental ammonia for periods of time ranging from minutes to days or weeks, or can be cyclical. Furthermore, aquatic animals cannot escape anthropogenic impacts on water bodies that affect IOAB homeostasis, including changes in salinity regimes owing to landscape disruption and ice melting, CO₂-induced acidification, heavy metal pollution­, hypoxic zones and other disturbances caused by agriculture run-off (Fig. 1).

The main goal of this Review is to convey the progression of knowledge about IOAB regulation in aquatic animals. This synopsis is divided into five arbitrarily defined and overlapping eras (Fig. 2), beginning with a brief summary of knowledge before JEB (Box 1) and highlighting some of the outstanding work published in JEB throughout its centennial history (with our apologies for the many omissions). For practical reasons, the focus is on NaCl transport and H⁺ and HCO₃^– excretion in adult aquatic animals; additional details can be found in the exhaustive reviews cited throughout this article and in other recent reviews about NaCl uptake in freshwater animals (Lee et al., 2022), sessile crustaceans (Sundell et al., 2019), cephalopods (Hu et al., 2015) and freshwater insects (Silver and Donini, 2021), and about the pH/temperature α-stat hypothesis (Wang and Jackson, 2016) and the osmorespiratory compromise (Wood and Eom, 2021).

The first JEB article to discuss aquatic IOAB regulation was a review on the factors that affect the migration of anadromous fishes, including salinity, pH and CO₂ (Chidester, 1924). This article posed the question of whether, by disrupting O₂ transport, the pH/CO₂ of the water could affect fish movement and distribution to a greater extent than salinity, and recommended that ‘true field ecologists’ included a pH-measuring apparatus in their toolkit. However, another view argued that environmentally relevant pH variations were too small to have a significant effect on aquatic animals (Saunders, 1926). Other JEB articles continued to document the effects of salinity changes on fluid composition (Bateman, 1932, 1933; Pantin, 1931) and on the respiratory rate of aquatic invertebrates (Beadle, 1931; Löwenstein, 1935), and the ionic and osmotic properties of hagfish blood (Bond et al., 1932) and trout eggs (Gray, 1932). Experiments on excised hearts showed the importance of urea for shark cardiac function (Simpson and Ogden, 1932), and back-to-back articles provided a comprehensive comparison of the mechanisms of urine formation in aquatic crustaceans and molluscs (Picken, 1936, 1937). These studies greatly contributed to establishing the osmoregulatory patterns described in modern textbooks: marine invertebrates and hagfish are iso-osmotic (see Glossary) with seawater, marine fishes actively excrete NaCl to remain hypo-ionic (see Glossary) to seawater, but there are osmoregulatory differences between bony and cartilaginous fishes. Moreover, freshwater animals actively take up NaCl and excrete a dilute urine to remain hyper-ionic and hyperosmotic (see Glossary).

Tonometer experiments described that, as observed in vertebrates, CO₂/H⁺ decreased the affinity of squid hemocyanin for O₂; however, the opposite effect was seen for horseshoe crab and whelk hemocyanins (a ‘reversed Bohr effect’; Redfield et al., 1926), and a JEB article reported a biphasic response for crab hemocyanin (Hogben, 1926). A few years later, a seminal study described the effects of CO₂/pH on fish blood–O₂ binding in unprecedented detail. It confirmed differences between active and sluggish species, concluded that the effect of CO₂/pH on fish blood O₂ binding was more complicated than in mammalian blood and proposed a molecular mechanism whereby elevated CO₂ sequentially inactivated each of the heme groups that bind O₂ (Root, 1931). The diverse effects of A–B on blood–O₂ transport are a reminder that, although mammalian physiology can provide valuable guidance, gross extrapolations to aquatic animals are not advisable owing to the unique physiological mechanisms that have evolved in the many distinct aquatic taxa and environments.

A new wave of research established the sites and mechanisms for ion absorption and excretion using whole-animal experiments combined with dissection, perfusion, ligation and divided-chamber (see Glossary) techniques. It was deduced that the osmoregulatory strategy of marine bony fishes involved seawater drinking, intestinal water and NaCl absorption, branchial NaCl excretion and renal handling of divalent ions (Smith, 1930). It was noted that freshwater animals absorb NaCl from dilute solutions and proposed that fish and crabs took Na⁺ in exchange for NH₄⁺ and Cl^– in exchange for HCO₃^– across their gills (Krogh, 1938). Influential JEB articles described the ion-transporting properties of frog skin (Francis and Pumphrey, 1933) and of different portions of the insect larval gut (Koch, 1938; Wigglesworth, 1933a,b, 1938). Studies at the cellular level were pioneered by the histological identification of the fish gill ‘chloride cell’ (Keys and Willmer, 1932) and the histochemical identification of Cl^–-absorbing cells in anal papillae of freshwater larval insects and gills of crustaceans (Koch, 1934). These fundamental concepts were compiled in August Krogh's classic work ‘Osmotic regulation in aquatic animals’ (Krogh, 1938).

The overwhelming majority of JEB articles during the period originated from laboratories in the UK. Like many other normal aspects of society, JEB's publication output was severely disrupted by World War II (WWII). The few articles on aquatic IOAB regulation during the war years continued to document the effects of hypotonic environments on the fluid composition of invertebrate species (Beadle and Cragg, 1940; Ellis, 1939; Panikkar, 1941; Robertson, 1939), invertebrate muscular and ciliary function (Wells and Ledingham, 1940; Wells et al., 1940) and salmon sperm activation (Ellis and Jones, 1939). These types of studies continued in the post-war era, in some cases using more accurate methods to measure ions in minute volumes of fluid (Ramsay, 1949, 1950; Robertson, 1949). In the USA, an iconic textbook for medical students synthetized the basics of A–B physiology (Davenport, 1947); it included detailed explanations of metabolic and respiratory challenges and adjustments using the ‘pH-bicarbonate diagram’ (today known as the Davenport diagram).

As the devastating effects of WWII began to settle, there was a vigorous ramping up of technology, industry and research together with an explosion of new tools. Radioisotopes, such as ²²Na⁺ and ³⁶Cl⁻ (generated as byproducts in nuclear fission reactors), were utilized to measure ion fluxes with high precision and accuracy. Radiotracer fluxes coupled with electrophysiological measurements in the Ussing chamber allowed for the identification and measurement of ionic currents across the skin of amphibians (Ussing, 1949). Later on, the short-circuit current would be used as a proxy for net transepithelial ion transport (reviewed in Kirschner, 2004) across other flat ion-transporting epithelia, including the opercular membrane of bony fishes, split gill lamellae of decapod crabs, cultured epithelial cells growing on permeable supports, segments of the fish gastrointestinal tract and the clam mantle.

Many paradigms, including the DNA double-helix structure and mitochondrial function, originated during the 1950s and 1960s and revolutionized biological research. A large number of enzymes were biochemically characterized, including carbonic anhydrase (CA) (Meldrum and Roughton, 1933) and Na⁺/K⁺-ATPase (NKA) (Skou, 1957), and the drugs acetazolamide (Miller et al., 1950) and ouabain (Schatzmann and Räss, 1965) were developed to inhibit their respective activities in vitro and in vivo. Pharmaceutical companies developed the loop diuretics thiazide, furosemide and bumetanide, and the K⁺-sparing diuretic amiloride; these inhibit renal NaCl absorption and were used to treat high blood pressure (Jucker, 1976). Although the target proteins were not known at the time, these drugs were quickly adopted for investigating NaCl transport mechanisms in aquatic animals. Decades later, thiazide, bumetanide and amiloride would help identify Na⁺/Cl⁻ cotransporters (NCCs), Na⁺/K⁺/2Cl⁻ cotransporters (NKCCs) and Na⁺/H⁺ exchangers (NHEs). The stilbene compounds SITS (4-acetamido-4'-isothiocyano-stilbene-2,2'-disulfonic acid) and DIDS (4,4′-di-isothiocyano-2,2′-disulfonic stilbene) were used to evaluate Cl⁻-transporting mechanisms; however, their poor specificity made it difficult to discriminate between anion exchangers, Cl⁻ channels and other off-target effects.

The production of commercial transmission electron and scanning electron microscopes (TEM and SEM, respectively), coupled with the optimization of biological sample processing, enabled the visualization of the cellular ultrastructure during the 1960s and 1970s. Antibodies, fluorochromes and epifluorescence microscopy began to be used to detect proteins in tissues and cells (reviewed in Brandtzaeg, 1998); however, it would take until the 1990s for these techniques to be used for studying aquatic IOAB regulation.

The synergistic use of these novel concepts and tools triggered a rapid leap in knowledge. Robust CA and NKA activities were found to be a nearly universal characteristic of ion-transporting epithelia. Electrophysiology and radiotracer experiments coupled with drugs identified ion-transport mechanisms at the apical and basolateral membranes. TEM revealed that the ‘chloride cell’ of fish gills contained numerous large mitochondria and a tubular system (Doyle and Gorecki, 1961) that was later shown to be connected with the basolateral membrane (Philpott, 1966). Because these cells transport other ions in addition to Cl^–, they were eventually renamed ‘mitochondrion-rich’ cells (MRCs). ‘Denser’ and ‘lighter’ MRC subtypes were noticed, the former being more abundant in seawater and having an apical pit (Doyle and Gorecki, 1961). Research during the subsequent decades (e.g. Laurent and Dunel, 1980) found a variety of MRC subtypes and other ion-transporting cells (collectively referred to as ‘ionocytes’; see Glossary) in different aquatic animals and environments.

During this era, JEB published hundreds of articles about aquatic IOAB regulation, especially on NaCl uptake by freshwater insects, crustaceans and fishes. This is reflected in the analysis of more frequently used words in JEB article titles during the 1960s and 1970s, which prominently feature ‘sodium’, ‘chloride’, ‘ionic’, ‘transport’, ‘pH’, ‘gills’, ‘osmotic’, ‘trout’, ‘goldfish’, ‘crayfish’, ‘crab’ and ‘malpighian’ (Franklin, 2023). Typical experimental approaches entailed measuring ionic fluxes, water–blood potential difference and composition of physiological fluids, sometimes in conjunction with inhibitors or exposure to environmental or metabolic challenges. These studies helped solidify the general concept of independent Na⁺/H⁺ and Cl⁻/HCO₃ exchange in insect larvae, crustaceans and fish, as well as the link between ion uptake and A–B regulation in freshwater.

The presence of NKA enzymatic activity in gill extracts together with the inhibition of Na⁺ uptake by ouabain and lack of inhibition by external K⁺ indicated that basolateral NKA was the driving force for Na⁺ uptake (Maetz, 1971). This would be confirmed using TEM observations of [³H]ouabain binding in the basolateral membrane of freshwater gill MR-ionocytes (Karnaky et al., 1976). The pathway for Na⁺ entry from freshwater into fish and crab gill cells was inhibited by amiloride (Kirschner et al., 1973); however, the NHEs (slc9 family) were yet to be characterized at a biochemical (Murer et al., 1976) and molecular level (Sardet et al., 1989). Whether Na⁺ was taken up in exchange for metabolic H⁺ or NH₄⁺ was the subject of a healthy scientific debate that began to have some closure in the 2000s with the identification of ammonia-transporting Rhesus (Rh) channels (Wright and Wood, 2009; Box 2).

Inhibition of Na⁺ uptake upon exposure to acidic pH (Maetz, 1973; Shaw, 1960) led to the realization of the potential damaging effects of acid release from coal mining and of acid rain on IOAB regulation (see Ultsch et al., 1981). A few years later, another link between anthropogenic pollution and aquatic IOAB regulation would be established, as scientists discovered the effects of metals on gill apical transporters involved in Na⁺ and Ca²⁺ uptake, paracellular NaCl efflux and NKA activity (reviewed in Paquin et al., 2002).

All aquatic animals take up Cl⁻ in exchange for internal HCO₃⁻ through a mechanism that is generally inhibited by SITS (Perry et al., 1981). However, Cl⁻ uptake demonstrated high inter-individual variability and species-specific sensitivity to acetazolamide (Kerstetter and Kirschner, 1972), indicating a large influence of the internal milieu and diverse internal HCO₃⁻ sources. Inhibition of active Cl⁻ uptake by thiocyanate and anion-stimulated ATPase activity in gill plasma membranes were interpreted to reflect an apically located Cl⁻/HCO₃⁻ ATPase that energized Cl⁻ uptake (de Renzis, 1975). However, the anion-stimulated ATPase activity was later explained by mitochondrial contamination (Siebers et al., 1990), and the mechanisms that energize Cl^– uptake across the gills of freshwater teleosts remain unidentified.

In seawater fish gills, NKA was originally thought to be in the apical pit of the chloride cells and to directly pump Na⁺ into seawater (Maetz, 1971). However, TEM and [³H]ouabain binding assays determined that, just like in freshwater fishes, NKA was in the basolateral membrane (Karnaky et al., 1976). The enigma was solved through a series of functional studies using radiotracer fluxes, intracellular ion measurements, electrophysiology, and ouabain and furosemide treatment of live eels (Silva et al., 1977a), Ussing chamber-mounted killifish opercular membranes (Karnaky et al., 1977) and isolated and perfused shark rectal glands (Silva et al., 1977b). The model is among the most elegant in epithelial physiology: NKA drives electroneutral Na⁺ and Cl⁻ cotransport from the blood into the cell, and because it also kicks the Na⁺ back out, it acts as a chloride pump that actively loads up the cell with Cl⁻. The negative intracellular electrical potential extrudes Cl⁻ into seawater or the rectal gland lumen, which in turn allows Na⁺ to be extruded paracellularly down its electrochemical gradient (Silva et al., 1977b). The vibrating-probe technique would confirm that this mechanism takes place in MR-ionocytes (Foskett and Scheffey, 1982), and the model would be further perfected with the molecular, functional and localization characterization of the NKCC, the cystic fibrosis transmembrane regulator (CFTR) Cl⁻ channel and K⁺ channels (reviewed in Evans, 2010; Box 3).

In the area of A–B regulation, experiments on rainbow trout identified the inverse relationship between plasma HCO₃⁻ accumulation and Cl⁻ loss during exposure to hypercapnia (Lloyd and White, 1967). A JEB article revealed that exposure to hypercapnia resulted in an equivalent increase in blood CO₂ that maintained the gradient for CO₂ excretion, instead of engaging the mechanisms to blow off excess CO₂ observed in air-breathing animals (Cameron and Randall, 1972).

The teleost opercular membrane and the shark rectal gland enabled studies on the effects of hormones, neuropeptides and signaling molecules on ion transport without confounding effects from the rest of the animal (Epstein et al., 1983; Foskett et al., 1983). Similar studies on marine fish intestines identified the furosemide-sensitive co-transport of Na⁺, K⁺ and Cl⁻ across the apical membrane (Musch et al., 1982), the differential effects of the second messengers cAMP and cGMP on ion absorption (Rao et al., 1984), and the links between HCO₃⁻ excretion and NaCl and water absorption (Ando and Subramanyam, 1990). The flounder urinary bladder revealed thiazide-sensitive apical electroneutral Na⁺,Cl⁻ co-transport (Stokes, 1984), which would eventually lead to the initial identification and cloning of an NCC (Gamba et al., 1993). It was clear that IOAB regulation in aquatic animals was achieved by the concerted action of multiple ion-transporting proteins with species- and tissue-specific differential expression and regulation.

Attempts to study ion transport across fish gills independently from the rest of the animal relied on perfused head and isolated arch preparations (Perry et al., 1984a). However, this approach fell out of use midway through the 1980s, probably owing to technical limitations that affected reproducibility between studies. Nonetheless, it generated interesting conclusions, such as the potential partitioning of Cl⁻/HCO₃⁻ and Na⁺/H⁺ exchanges between gill filaments and lamellae (Avella et al., 1987) and the rapid differential adrenergic regulation of ion uptake (Perry et al., 1984b). Other studies used SEM coupled with NaCl uptake kinetic analysis to correlate ultrastructural changes in apical ionocyte surface area with blood IOAB regulation in response to environmental and metabolic challenges. This approach identified a ‘morphometric regulation’, whereby ion transport rates were modulated by differential changes in the apical surface area of the ionocyte subtypes through the development of microvilli and the retraction of adjacent pavement cells (see Glossary), resulting in increased ionocyte exposed apical membrane surface (Cameron and Iwama, 1987; Goss et al., 1992, 1994; Perry and Laurent, 1989).

The biochemical discovery of the V-type H⁺-ATPase (VHA) in the fungus Neurospora crassa (Bowman and Bowman, 1982) led to another re-evaluation of the mechanisms for IOAB regulation. The physiological roles that were ascribed to VHA during the 10 years following its discovery have been documented in an iconic JEB Special Issue (Harvey and Nelson, 1992). This early research focused on eukaryotic endocytic vacuoles and organelles, and also investigated the plasma membrane of specialized cells in the mammalian kidney, the frog skin and the gut of insects. Subsequently, the VHA proved to be a critical driving force for IOAB regulation in freshwater, for HCO₃⁻ excretion in seawater and for many other A–B-related cellular processes in a variety of aquatic animals.

The need to visualize larger tissue areas and to establish the cellular localization of ion-transporting proteins led to the adoption of immunohistochemistry. To our knowledge, its first use in aquatic IOAB regulation research was to study the localization of VHA in trout gills; this research was published in JEB (Lin et al., 1994). Other landmarks include the original immunolocalization of NKA in fish gills using the a5 monoclonal antibody (Witters et al., 1996), the co-immunolocalization of multiple proteins in tilapia, trout and mudskipper (Wilson et al., 2000a,b), and the visualization of distinct NKA- and VHA-rich ionocytes in freshwater and seawater Atlantic stingray (Piermarini and Evans, 2000).

The idea that VHA may play a role in IOAB regulation in freshwater fishes (Avella and Bornancin, 1989) and the first evidence supporting it (Fenwick et al., 1999; Lin and Randall, 1991; Lin et al., 1994; Sullivan et al., 1995) were published in JEB during the 1990s. These studies concluded that VHA in the apical membrane of the pavement cells in gill lamella excreted H⁺ and energized Na⁺ uptake via an apical Na⁺ channel, similar to the role of MR-ionocytes in frog skin and toad urinary bladder and to mammalian renal intercalated cells. However, pharmacological VHA inhibition also depressed Cl⁻ uptake (Fenwick et al., 1999), and VHA was additionally present throughout gill filament cells (Lin et al., 1994). Furthermore, freshwater Atlantic stingray gill cells had VHA in their basolateral membrane, which suggested a role in H⁺ absorption and energizing HCO₃⁻ excretion to seawater (Piermarini and Evans, 2001; Box 4).

Other emerging experimental approaches were to isolate and culture gill cells for histological, TEM and biochemical studies (e.g. Hootman and Philpott, 1978; Perry and Walsh, 1989) and to measure intracellular pH regulation in live cells using ratiometric dyes (Pärt and Wood, 1996). To address the issue of cell polarity loss and to study transepithelial NaCl uptake, pavement cell primary cultures were grown on semi-permeable supports (Wood and Pärt, 1997) and ‘double seeded’ with MR-ionocytes (Fletcher et al., 2000). Although these preparations have not yet been able to adequately mimic the active transepithelial ion-transporting mechanisms of gills, they have proven helpful for studying ammonia excretion, paracellular permeability, hormonal regulation and the effects of pollutants (reviewed in Bury et al., 2014).

A flurry of functional and molecular studies published in JEB during the 1990s and early 2000s greatly increased our understanding of ion transport across the posterior gills of decapod crustaceans. These gills are specialized for ion transport, can be isolated and perfused with relative ease, and their hemilamellae can be mounted in Ussing chambers. The cellular mechanisms underlying branchial ion transport in crustaceans have been extensively reviewed (Freire et al., 2008; Henry, 2001; McNamara and Faria, 2012). As environmental salinity decreases, gill epithelial tightness, NaCl uptake capacity and CA activity increase in association with a switch in ion-transporting mechanisms that provides clues about the evolution of hyperosmotic regulation during the invasion of freshwater environments. The gills of osmoconforming marine crabs show apical Na⁺/H⁺ and Cl/HCO₃⁻ exchange utilizing H⁺ and HCO₃⁻ generated by cytosolic CA as substrates and powered by basolateral NKA. Similarities with hagfish (Evans, 1984), hemichordates and cephalochordates (Sackville et al., 2022) support the hypothesis that the original function of gill ion transport was A–B regulation and that this was later co-opted for hyperosmotic regulation. Weak hyperosmoregulators (e.g. Carcinus maenas) that live in brackish water (10–35 psu) absorb most NaCl through apical NKCC and basolateral NKA and an associated paracellular Na⁺ flux (Riestenpatt et al., 1996). Moderate hyperosmoregulators (e.g. Neohelice granulata) use similar mechanisms in brackish water (Onken et al., 2003), and may additionally engage apical VHA to energize NaCl uptake in more diluted environments (1–2 psu) (Genovese et al., 2005). However, an inability to reduce paracellular permeability and the ensuing leak of NaCl precludes these animals from living in freshwater permanently. Strong hyperosmotic regulators (e.g. Eriocheir sinensis) have tight gill epithelia and can live in freshwater with <0.5 psu. They use apical VHA coupled with cytosolic CA activity to energize Na⁺ uptake across a channel, or an electrogenic 2Na⁺/H⁺ exchanger and Cl⁻ uptake via an anion exchanger. At the basolateral membrane, NKA energizes the movement of both Na⁺ and Cl⁻ from the ionocyte into the hemolymph; under open-circuit conditions, the Na⁺ and Cl⁻ currents short-circuit each other and allow for transepithelial NaCl uptake (Onken, 1999; Onken and Putzenlechner, 1995). The red freshwater crab Dilocarcinus pagei uses a similar mechanism to E. sinensis; however, Na⁺ and Cl⁻ absorption take place in opposite hemilamellae (Onken and McNamara, 2002), which implies a regional partitioning of ionocyte subtypes that deserves further attention.

Some crab species are able to excrete NaCl across their gills to hypo-osmoregulate (Luquet et al., 2002; Martinez et al., 1998) through an unknown NaCl excretion mechanism, although gene upregulation during acclimation to concentrated seawater suggests the involvement of NKA, NKCC and VHA (Luquet et al., 2005). Interestingly, the outside positive transepithelial potential difference (Luquet et al., 2002) and the lack of CFTR in invertebrates indicate a mechanism that is fundamentally different from that of vertebrates. Clearly, IOAB regulation in crustacean gills would benefit from future studies investigating the cellular and subcellular localization of ion-transporting proteins and potential ionocyte subtypes associated with IOAB regulation.

Ion transport across isolated crab gill preparations is rapidly regulated by eyestalk extract, hormones, neuropeptides and second messengers, and by changes in hemolymph osmolarity, pH and [HCO₃⁻] (Onken, 1996; Riestenpatt et al., 1994; Siebers et al., 1985; Tresguerres et al., 2003, 2008) (Table S1). This type of regulation, which is too fast to involve changes in gene expression, is likely to play an essential role for the crab (and other animals) in vivo during sudden salinity changes encountered during tides, in estuaries and in tide and rain pools.

The comparative focus fostered by JEB catalyzed a vigorous exchange of ideas among subdisciplines that helped identify mechanistic links between IOAB regulation and multiple other physiological processes in aquatic animals. Among these, biomineralization stands out because of its critical importance for global carbon cycling and an urgency to understand the impacts of climate change on aquatic life beyond those predicted by geochemical models that tend to dismiss biological regulation. Research published in JEB characterized links between A–B regulation, H⁺ excretion and the formation of the crab carapace and fish otoliths (Cameron, 1985; Cameron and Wood, 1985; Payan et al., 1997). These studies, together with research on corals (Furla et al., 2000; Tambutté et al., 1996), indicated that a significant portion of the dissolved inorganic carbon in biomineralized structures was sourced from metabolically generated CO₂, whereas the rest comes from environmental HCO₃⁻. Although not surprising (after all, terrestrial vertebrates build their bone using CO₂ and HCO₃⁻ derived from mitochondrial respiration and blood plasma), these mechanisms go against the mainstream biogeochemical view that ocean acidification impairs biomineralization owing to the associated decrease in environmental CO₃²⁻ availability (NOAA PMEL Carbon Program, 2023). Importantly, our ability to forecast potential effects of ocean acidification on biomineralization is strictly linked to the generation of further knowledge about the underlying cellular and physiological mechanisms, both general and species-specific. This is an urgent topic that is uniquely suited for comparative IOAB experimental physiologists.

Research during the 1980s and 1990s identified a novel source of biomineralization resulting from intestinal IOAB regulation in marine bony fishes that was proposed to have geological significance (Walsh et al., 1991). This prediction would be confirmed 20 years later, and icthyocarbonates would be estimated to contribute ∼15% of oceanic carbonate production (Wilson et al., 2009) and >70% of mud carbonates in mangroves (Perry et al., 2011). This biological process is the result of cellular IOAB regulatory processes that allow bony fishes to stay hydrated in seawater (reviewed in Grosell and Oehlert, 2023), and also deserves more consideration by oceanographers and marine biogeochemists.

The new century saw an explosion in the utilization of molecular techniques to clone cDNAs encoding ion-transporting proteins, determine their tissue expression and quantify their mRNA abundance in response to salinity or A–B challenges. Initially, these tasks relied on degenerate reverse transcriptase PCR (RT-PCR), northern blots and semi-quantitative RT-PCR, but they later transitioned to expressed sequence tag libraries, mRNA microarrays and real-time PCR (qPCR). Although these techniques were time consuming and expensive, they had the benefit of a requirement for a priori hypotheses that guided experimental design and interpretation of results. Early in the 2000s, a new experimental trend started whereby animals were exposed to an IOAB challenge and changes in mRNA abundance were interpreted as evidence for the involvement of the encoded proteins in the underlying homeostatic mechanisms. Among the first novel findings, this approach identified the existence of multiple NKA α-subunit isoforms in rainbow trout gills and a switch in isoform expression during salinity transfer (Richards et al., 2003b). This JEB paper stimulated research in multiple other fish species, including studies on Atlantic salmon that improved mRNA quantification accuracy with the use of qPCR and used mRNA in situ hybridization (Madsen et al., 2009), western blotting and immunohistochemistry (McCormick et al., 2009) to establish the cellular localization of the NKA isoforms.

The presence of only two ionocyte subtypes in elasmobranch gills coupled with an exclusive role of the gills in A–B regulation in seawater greatly facilitates studies at the whole-animal and cellular levels. Furthermore, elasmobranchs are amenable to comparative studies on ionic and osmotic regulation owing to the presence of stenohaline and euryhaline species (see Glossary; Box 4); they undergo a metabolic blood acidosis upon exhaustive exercise (Richards et al., 2003a), and they naturally experience a pronounced postprandial blood alkaline tide (Wood et al., 2005; Box 5). The discovery that pendrin (slc26a4) is an HCO₃⁻-excreting anion exchanger in renal β-intercalated cells that express basolateral VHA (Royaux et al., 2001) initiated a series of studies that would resolve many questions about IOAB regulation in elasmobranch gills. First, pendrin was identified in base-excreting (B)-ionocytes from both marine and freshwater Atlantic stingrays (Piermarini et al., 2002). The apical localization of pendrin and basolateral localization of VHA in freshwater was a clear indication of their role in Cl⁻ uptake and HCO₃⁻ excretion for hyper-ionic regulation; however, both pendrin and VHA demonstrated a diffuse cytoplasmic localization in marine stingrays. A few years later, the induction of metabolic alkalosis in dogfish was found to cause the translocation of VHA to the basolateral membrane of B-ionocytes (Tresguerres et al., 2005), a mechanism that was dependent on functional microtubules (Tresguerres et al., 2006a) and CA (Tresguerres et al., 2007), and was essential for compensation of the blood alkalosis. Later, soluble adenylyl cyclase (sAC) was identified as the enzyme that sensed the blood alkalosis and triggered the translocation of VHA to the basolateral membrane (Tresguerres et al., 2010) and pendrin to the apical membrane (Roa et al., 2014) during a natural postprandial blood alkalosis. More recently, experiments on isolated round ray gill cells demonstrated that alkalosis and sAC-dependent activation of B-cells does not require central hormonal or neurotransmitter input (Roa and Tresguerres, 2016). The widespread presence of CA, sAC, anion exchangers and VHA in base-excreting epithelial cells suggests an evolutionarily conserved mechanism for sensing and counteracting a blood alkalosis (Tresguerres, 2014). The gills of both freshwater and marine elasmobranchs also have acid-excreting (A)-ionocytes that co-express NKA and NHE3 (Choe et al., 2005, 2007; Reilly et al., 2011) and take up Na⁺ in exchange for H⁺ in a manner analogous to mammalian renal proximal tubule epithelial cells. These cells become inactivated during the alkaline tide (Roa and Tresguerres, 2019), and NKA and NHE3 abundances increase during transfer from seawater to freshwater, but not during short-term (hours) hypercapnia (Choe and Evans, 2003; Piermarini and Evans, 2000). Interestingly, both A- and B-cells become activated upon migration of euryhaline elasmobranchs from seawater to freshwater to mediate NaCl uptake for blood hyperionic regulation. Could the evolutionary invasion of freshwater environments have been facilitated by an activation of A- and B-ionocytes cells owing to more frequent bursts of active swimming, exposure to hypercapnia and postprandial alkaline tides?

Teleost fishes have more ionocyte types than elasmobranchs, and the mechanisms for IOAB regulation are accordingly more complex. This article briefly reviews the state of knowledge about freshwater rainbow trout and zebrafish; further details about these and other species can be found in a number of recent exhaustive reviews (e.g. Hiroi and McCormick, 2012; Hwang and Chou, 2013; Kwong et al., 2014; Shih et al., 2023).

Early in the 2000s and inspired by rabbit renal research, PNA⁺ and PNA⁻ ionocyte subtypes were identified in freshwater trout gills based on their ability to bind peanut lectin agglutinin (PNA) (Galvez et al., 2002; Goss et al., 2001). These ionocytes were proposed to correspond to the denser and lighter chloride cell subtypes discovered in the 1960s and to the chloride MR and pavement MR cells observed in the 1980s. The idea that PNA⁺ ionocytes mediated Cl⁻/HCO₃⁻ exchange and PNA⁻ ionocytes mediated Na⁺/H⁺ exchange was very attractive; however, the picture turned out to be much more complicated than that. For starters, PNA binding by fish ionocytes is restricted to rainbow trout, and even then, the PNA⁺ ionocyte localization pattern throughout the gill epithelium is quite variable among studies (Brannen and Gilmour, 2018; Ivanis et al., 2008; Tresguerres et al., 2006b). Moreover, PNA⁺ ionocytes express NHE2 and NHE3 (Ivanis et al., 2008), which implies a role in Na⁺/H⁺ exchange that was originally proposed for the PNA⁻ cells. Some of these discrepancies could be explained by the presence of PNA⁺ ionocyte subtypes; for example, one may express basolateral NKCC1 and apical NHE3 and play a role in Na⁺/H⁺ exchange and ammonia excretion (Dymowska et al., 2014; Hiroi and McCormick, 2012), another may take up Cl⁻ through apical NCCs (Brannen and Gilmour, 2018), and yet another may express still elusive apical Cl⁻/CO₃⁻ exchangers in a metabolon with cytosolic CA and basolateral VHA (Tresguerres et al., 2006b). This hypothesis could be tested by immunolocalizing NCC and slc26a6, which is the most promising candidate for a Cl⁻/HCO₃⁻ exchanger in trout (Boyle et al., 2015; Leguen et al., 2015). The PNA⁻ ionocytes are proposed to take up Na⁺ through apical acid-sensing ion channels (ASICs; Dymowska et al., 2014) energized by apical VHAs. However, the frog skin and the mammalian renal collecting duct exemplify that the Na⁺ channel and VHA do not necessarily have to be in the same cell (Larsen et al., 2014). Could apical VHA be present in the pavement cells, as originally proposed (Sullivan et al., 1995; Wilson et al., 2000b), and still energize Na⁺ uptake by the PNA⁻ ionocyte cells?

Fast generation times, year-round breeding and small size contributed to the emergence of zebrafish as a model system for vertebrate development during the 1990s, the sequencing of its genome in the 2000s (Grunwald and Eisen, 2002) and the adoption of zebrafish for studies IOAB regulation. The zebrafish embryo became an extremely useful model owing to the presence of abundant ionocytes in the yolk sac. These cells can be conveniently stained, quantified, probed with the scanning ion-selective technique and analysed for the effects of morpholino and CRISPR/Cas9 genetic manipulations before major complications are manifested. Current models propose at least four ionocyte subtypes in larval zebrafish skin: (1) VHA-rich (HR) cells with apical VHA, Rhcg1, NHE3 and a putative Na⁺ channel and basolateral NKA (isoform α5); this cell type takes up Na⁺ and H⁺ and excretes NH₄⁺; (2) NCC cells with apical NCC and basolateral NKA (isoform α1) that take up NaCl; (3) NaR cells with apical Ca⁺ channels and basolateral NKA (isoform α5) that take up Ca²⁺; and (4) KS cells with apical K⁺ channels and basolateral NKA (isoform α4) that excrete K⁺ for some unknown function (reviewed in Hwang and Chou, 2013). Similar to all other teleosts, the mechanisms for Cl⁻/HCO₃⁻ exchange remain elusive; however, one of these four ionocyte subtypes presumably expresses slc26a3 and is responsible for Cl⁻ uptake and HCO₃⁻ excretion. In addition, slc26a3 was also reported in non-NKA-rich cells (which make a fifth ionocyte subtype), and mRNA data suggested the presence of slc26a4 and slc26a6 in unidentified zebrafish larvae skin cells (Bayaa et al., 2009). It would be interesting to reassess the subcellular localization of these and some of the other transporters using multi-protein immunolocalization and confocal 3D optical reconstructions (e.g. Hiroi et al., 2008).

The IOAB regulatory mechanisms in zebrafish gills have not been explored to the same extent as those in larval skin. Zebrafish gills have HR cells that mediate H⁺ excretion (Shih et al., 2022) and express slc26s (Perry et al., 2009), ASICs (Dymowska et al., 2015) and NHE3b (Yan et al., 2007; Zimmer et al., 2019). However, the presence of these proteins in specific ionocyte subtypes and their subcellular localizations are not clear. Recently, adult zebrafish were reported to activate a novel Na⁺ uptake mechanism during short-term (2–8 h) exposure to low pH that is coupled to K⁺ efflux and does not involve apical NHEs, Na⁺ channels, NCCs or NH₄⁺-trapping (Clifford et al., 2022a). The identity of the apical transporter remains unknown, but mRNA evidence points to K⁺-dependent Na⁺–Ca²⁺ exchangers (NCKXs; slc24a1-6). It was proposed that this K⁺-dependent Na⁺ uptake mechanism may gain ecophysiological relevance in the natural environment of zebrafish, which includes densely populated acidic shallow streams and stagnant ponds. Future studies will certainly examine whether other species inhabiting acidic waters use this mechanism, which may also inspire research on renal physiology and hypertension (Mrowka, 2023). Indeed, this mechanism has been deemed ‘among the most significant advances in the field of comparative osmoregulation over the past several decades’ (Perry and Wang, 2022).

The need to better understand the responses of marine fishes to environmental stressors (Fig. 1) sparked research to understand the molecular mechanisms for branchial H⁺ and HCO₃⁻ excretion. The generalized pattern is that these NKA-rich ionocytes also express NHE3 and are responsible for excreting H⁺ (Christensen et al., 2012; Hiroi and McCormick, 2012; Hsu et al., 2014; Seo et al., 2013), a process that is favoured by the pronounced inwardly directed Na⁺ gradient caused by the high concentration of Na⁺ in seawater coupled with the activity of NKA. Upon exposure to hypercapnia, many seawater species are able to rapidly activate H⁺ excretion to compensate for the ensuing blood acidosis. At least in European sea bass, the mechanism involves the widening of the ionocyte apical membrane with an associated increased availability of NHE3s exposed to seawater (Montgomery et al., 2022). Such a mechanism resembles the morphometric regulation of ion transport described in freshwater fishes during the 1980s, albeit at much faster time frames.

In addition, there is some evidence for the presence of separate VHA-rich ionocytes in gills from longhorn sculpin (Catches et al., 2006) and of VHA in NKA-rich ionocytes in red drum gills (Allmon and Esbaugh, 2017). VHA seemed to be present in the basolateral tubular system or in cytoplasmic vesicles even during exposure to extreme hypercapnia (Allmon and Esbaugh, 2017), which rules out a role in H⁺ excretion across the apical membrane. Our immunofluorescence confocal microscopy images confirm these patterns in sardine (Fig. 3A) and rockfish (Fig. 3B,C). One type of ionocyte lacks NKA and has VHA in its cusp-like shape basolateral membrane (Fig. 3Ci–iii), whereas the other is the traditional ionocyte that has NKA in its heavily infolded basolateral membrane and has VHA also in the basolateral membrane or in cytoplasmic vesicles (Fig. 3Cii,iii). We tentatively propose that VHA in one or both ionocytes becomes activated during blood alkalosis in an analogous manner to VHA in the B-cells of elasmobranch gills. Confirming this hypothesis would require, once again, identifying the presumed apical Cl⁻/HCO₃⁻ exchanger and establishing its cellular localization followed by functional studies following a blood alkalosis.

Why is there so much diversity in the gill ionocytes of teleosts compared with elasmobranchs and other aquatic animals? One reason may be related to the evolutionary history that includes an invasion of freshwater environments, a subsequent return to the sea for some species followed by multiple additional freshwater invasions (Halstead, 1985) that could have resulted in the selection of different branchial cellular mechanisms for IOAB regulation. Another potential reason is the great ecophysiological diversity of teleost fishes, which have disparate sizes, occupy a large range of freshwater and semi-terrestrial environments, and have diets with varied organic and ionic content and impacts on A–B physiology. Indeed, some teleosts experience a postprandial alkaline tide like elasmobranchs (Bucking et al., 2009; Cooper and Wilson, 2008), but others have an acidic tide (Wood et al., 2010; Box 5). In seawater, ionocyte function may have gained additional diversification owing to emerging interactions between IOAB transport both within each ionocyte and between gills and the gut (e.g. Taylor and Grosell, 2009). Another potential factor affecting ionocyte diversity might be differences in ammoniagenesis, gluconeogenesis and glycolysis in gill ionocytes that could favour apical H⁺ excretion and Na⁺ uptake via NHEs or VHA-Na⁺ channels (Tseng et al., 2020). Moreover, there is also diversity in branchial and skin Ca²⁺ transport, which is not discussed in this Review but is an essential process for bony fishes. When we consider the potential combinations of all of these factors, the observed ionocyte diversity is less surprising.

Next-generation sequencing has revolutionized research by enabling the relatively affordable analysis of whole genomes and extensive transcriptomes. However, even such powerful methods have limitations. Aquatic animals have unique physiologies enabled by genes that do not have a counterpart in model species and thus lack functional annotation (Melzner et al., 2022). The mRNA levels do not always reflect protein levels (Gerbin et al., 2021), and whole-tissue transcriptomics analysis does not provide information about differential mRNA localization in cell types or about the subcellular localization of the encoded proteins. However, genomics and transcriptomics are superb for studying processes that depend on sequential switches of gene expression patterns (i.e. development), for processes for which the underlying physiological mechanism are already known, or for hypothesis generation followed by functional testing. Some recent examples include using mRNA network analysis to examine the effects of environmental stressors during freshwater–seawater transitions in salmonids (Monette and Velotta, 2022), single-cell transcriptomics to elucidate the molecular mechanisms of biomineralization in sea urchin larvae (Chang et al., 2021), and genomic analyses to investigate the molecular evolution of ionic regulation in natural populations spanning a salinity gradient (Velotta et al., 2022).

The growing interest in understanding the responses of aquatic animals to their dynamic environments and to anthropogenic stressors is leading to an expansion of experimental approaches used to study IOAB regulation in aquatic animals (Blewett et al., 2022; Zimmer et al., 2021). Many studies expose animals to extreme conditions to maximize compensatory responses and facilitate their detection and quantification; however, conclusions from such studies do not always apply to natural environments, nor are the compensatory responses necessarily proportional to the magnitude of the stressors. Additionally, many environmentally relevant physiological adjustments can occur very rapidly and do not rely on transcriptomic adjustments (Table S1). Thus, extrapolations from extreme experiments to the ‘real world’ should be performed with extreme caution.

Modern studies on IOAB regulation will require multidisciplinary collaborations with biogeochemists to measure and model environmental salinity, CO₂/pH, ammonia, temperature and O₂ levels with sufficient temporal and spatial resolution (e.g. Cyronak et al., 2020; Frieder et al., 2012). The expertise of aquatic IOAB physiologists will be essential to assess the extent to which environmental variability challenges animal homeostasis and to evaluate compensatory responses. But although it is currently possible to efficiently measure blood ionic composition and osmolarity in wild aquatic animals (e.g. Dwyer et al., 2020), measuring other physiological variables will require the development of implantable sensors of blood A–B parameters and P_O2, and of non-invasive biopsies to assess cellular physiology based on biochemical and cellular biomarkers. This endeavor will require collaborations with bioengineers to miniaturize sensors, amplifiers, processors and data loggers.

In addition, there will always be the need for experimental laboratory research on aquatic IOAB regulation. Much of our current knowledge has been generated from model species or from species with economic and ecological interest in developed countries, most times following Krogh's principle: ‘For many problems there is an animal on which it can be most conveniently studied’. However, relatively little is known about aquatic animals from developing countries throughout the world. Perhaps we could learn something from plant biologists who, concerned about a growing trend for studies on the model species Arabidopsis thaliana, formulated ‘No single organism (or technique) exists that can provide easy access to the diversity of hidden mechanisms that underlie all interesting and important physiological and biochemical problems’ (Wayne and Staves, 1996).

This Review has attempted to convey that scientific research is a convoluted process. As time passes by, there is the risk of dismissing the nuances – much like a broken telephone game – potentially leading to oversimplification. The readers, and especially junior colleagues, are encouraged to exhaustively study the original literature, follow the trail of knowledge progression and to not automatically dismiss results that do not conform to current knowledge. Such a strategy is not only useful to better understand how the generation of scientific knowledge works, but can also help identify unresolved topics worth investigating through the lens of current paradigms and technologies (Romero, 2022).

Aquatic environments are extremely diverse both spatially and temporally. Moreover, the ocean and other aquatic ecosystems remain relatively unexplored and are increasingly threatened by anthropogenic activities that can affect IOAB homeostasis. Discovering physiological mechanisms underlying IOAB regulation and characterizing organismal responses and trade-offs will continue to be essential for satisfying human curiosity about how aquatic animals work, understanding and predicting how they will fare in rapidly changing environments, and identifying ‘winners and losers’ and informing conservation and policymaking agencies. Thus, IOAB research on aquatic animals should only gain more prominence during the upcoming decades. To fulfill this role, the aquatic IOAB physiology community should continue and increase our efforts to train the next generation of scientists in fundamental concepts as well as in emerging research areas and techniques.

We thank Drs Joshua Lonthair and Nicholas Wegner (NOAA Southwest Fisheries Science Center, La Jolla, CA, USA) for the Pacific sardine gill samples.