Effect of ultraviolet radiation (UVR) on the life stages of fish

Current levels of Ultraviolet Radiation (UVR) represent a significant threat to many fish species. The first studies on the effects of UVR on organisms were performed on fish at the beginning of the twentieth century, and the topic has been progressing continuously until the present. Here, we review the reported harmful effects of ultraviolet B (UVB) and A (UVA) radiations in fish at different lifecycle stages, including embryo, larvae, juveniles and adults. The most evident negative effects during the early development stages are an increase in mortality and incidence in developmental malformations, with the skin and gills the most affected tissues in larvae. Growth reduction, a loss in body condition, and behavioral, physiological and metabolic changes in juveniles/adults occur under short- or long-term UVB exposure. The skin in juveniles/adults undergoes profound morphological and functional changes, even after acute exposure to UVR. Impairment of molecular and cellular processes was evidenced in all development stages by increasing the levels of DNA damage, apoptosis and changing tissues’ antioxidant status. The different photo-protective mechanisms to cope with excessive UVR exposure are also revised. Currently, stratospheric ozone dynamics and climate change interact strongly, enhancing the potential exposure of fish to UVR under water. Due to these environmental changes, fish are exposed to new and complex interactions between UVR and environmental stressors, which potentially affects fish growth and survival. Understanding the ability of fish to cope and adapt to these environmental changes will be essential to evaluate the potential impact in fisheries and mitigate ecological problems.


Introduction
Solar energy reaching the earth's surface includes ultraviolet radiation (UVR) that can be divided into three spectral bands: ultraviolet C, highly harmful (UVC, mostly absorbed by stratospheric ozone and oxygen; does not reach the earth's surface); ultraviolet B, highly energetic and moderately harmful (UVB,; and ultraviolet A, mildly energetic and less harmful (UVA, 320-400 nm) (Madronich et al. 1995;McKenzie et al. 2007).
In the aquatic environment, both UVA and UVB radiation bands can penetrate the water column, showing variable attenuation across saltwater and freshwater ecosystems both seasonally and geographically. Dissolved organic matter and suspended particles are the major components that contribute to the attenuation of light under water. UVR is largely absorbed by chromophoric dissolved organic matter, which consequently reduces the exposure of aquatic organisms to UVR (reviewed by Häder et al. 2007;Williamson et al. 1996;Zagarese and Williamson 2001). Short UVR wavelengths are strongly absorbed under water, and UVB radiation is highly attenuated, penetrating from only a few centimeters below the surface in turbid lakes to more than 20 meters in transparent oceanic waters (Huovinen and Goldman 2000;Huovinen et al. 2003;Michael et al. 2012;Tedetti and Sempere 2006). UVA wavelengths are less attenuated than UVB, penetrating deeper into the water column, reaching depths greater than 70 m (Schlichter et al. 1986;Tedetti and Sempere 2006). Nonetheless, the significant depletion of stratospheric ozone due to anthropogenic emissions of atmospheric pollutants has enhanced the UVB radiation that reaches the biosphere (Barnes et al. 2019;Crutzen and Arnold 1986;Molina and Rowland 1974;Rowland 2006;Rowland and Molina 1975), causing detrimental effects to aquatic organisms and ecosystems (Häder et al. 1998;Helbling et al. 2003;Agustí 2006, 2010;Llabrés et al. 2013).
Current levels of UVA and UVB radiation in aquatic ecosystems can cause damage at different levels to a broad range of organisms, from bacteria to higher vertebrates (reviewed by Häder et al. 2007Häder et al. , 2011Häder et al. , 2015Llabrés et al. 2013;Peng et al. 2017;Williamson et al. 2019;Xiao et al. 2015). UVR is mutagenic and is considered a strong evolutionary selective force in organisms (Rothschild 1999;Rozema et al. 2002). A meta-analysis study observed that aquatic organisms from the Northern Hemisphere tend to be more susceptible to the effects of UVB than those from the Southern Hemisphere, due to strong stratospheric ozone asymmetries between the hemispheres .
In the early 1930s, the harmful effects of exposure to UVR during embryonic development were reported for the first time in fish (Hinrichs and Genther 1931). A considerable number of Fundulus heteroclitus fertilized eggs and early embryos exposed to UVR exhibited severe degrees of axial duplication and showed several abnormalities including poor eye development (Hinrichs and Genther 1931;Hinrichs 1938). A few years later, Bell and Hoar (1950) observed high mortality in sockeye salmon (Oncorhynchus nerka) fertilized eggs in the later stages of development and larvae that had been exposed to UVR. These authors also noticed several skin lesions in the sockeye salmon larvae after UV exposure including the displacement between the epidermis and the basement membrane, loss of scales and disruption of the mucous producing cells (Bell and Hoar 1950).
The early development stages appear to be the lifecycle stages that are most prone to damage (Dahms and Lee 2010); however, the tolerance of juveniles and adult fish to UVR exposure has also been studied, and many species appear to be highly sensitive to both UVA and UVB radiation at later development stages in their lifecycle (García-Huidobro et al. 2017;Jokinen et al. 2008;Kazerouni et al. 2017;Rick et al. 2014;. Reduction in growth, impaired development, changes in behavior, development of skin and eye lesions, suppression of the immune system, reduction on diseases resistance, DNA damage and a series of metabolic and physiological stress changes are some of the described effects of UVR exposure in fish (Browman et al. 2003;Hunter et al. 1981;Salo et al. 2000a;Sandrini et al. 2009;Sharma et al. 2005). Many fertilized eggs and larvae, as well as visual predators, herbivores and farmed fish obligated to live at the photic surface layer, are potentially exposed to significant UVR radiation. From an economic point of view, several cases of sunburn due to overexposure to high natural solar radiation resulted in numerous losses in aquaculture fish farms during the 1980-1990s, particularly in those where the fish were grown in outdoor tanks (Bullock 1982(Bullock , 1984(Bullock , 1988Bullock and Coutts 1985;Lowe and GoodmanLowe 1996).
Although fish species can develop several strategies to cope with the harmful effects of UVR (e.g., the avoidance of UV, production of UV-absorbing compounds and DNA damage repairing mechanisms), recent meta-analyses continue to corroborate the negative effects of UVR on aquatic organisms (Braun et al. 2016;Williamson et al. 2019). During the past decade, many reviews have addressed the effects of UVR on aquatic ecosystems, mainly on primary producers, zooplankton and invertebrates. To the best of our knowledge, there are still few literature reviews, specifically compiling the effects of UVR effects. The effects of UVR on fish has been reviewed briefly as part of broader general reviews on the topic (Barnes et al. 2019;Häder et al. 2007Häder et al. , 2011Häder et al. , 2015, or has been reviewed more specifically such was done by Zagarese and Williamson (2001). Recently, Lawrence et al. (2019) revised the impact of UVR exposure in the fish immune system and mentioned that UVR exposure could have a negative effect on the immune control of infection by some fish species, especially in the case of fish produced by aquaculture. These authors also described some of the photoprotective mechanisms used by some, but not all fish species, which help to mitigate the negative impacts of UVR exposure. However, more recent reviews considering other adverse effects and analyzing results collected over the last two decades do not yet exist. Our goal here is to review the literature and recent studies describing the harmful effects of UVR on both marine and freshwater fish species. We also review the effect of different exposure periods (acute, short-term, or long-term exposure) and the mechanisms developed by fish organisms to cope with the effects of UVR. Lastly, we revise recent studies that address the interaction between the harmful effects of UVR and environmental stressors, like climate change or pollutants.

Detrimental effects of UVR on fish
Early development stages: embryos and larvae Increase of mortality, developmental abnormalities, behavioral and metabolic changes During early development, both fresh and seawater fish are sensitive to UVR. The most evident effects of UVR (mainly UVB) exposure are the reduction of survival rates and the increase in the number and types of developmental malformations in both embryos and larvae, when exposed to an acute dose, or for a shortand long-term exposure period. Subsequently, these developmental abnormalities have been associated with high mortality after UVR exposure (e.g., Dong et al. 2007;Lesser et al. 2001;Mahmoud et al. 2009;Vásquez et al. 2016) (Tables 1 and 2, see summary in Fig. 1).
4 h post-fertilization (hpf) zebrafish (Danio rerio) embryos exposed for 2.4 h (UVB, 295 nm cutoff) showed a reduction of more than 50% in their survival rate after 6 days, with a high incidence of developmental abnormalities, including caudal (posterior) notochord torsion and bending (Nuñez et al. 2012). In the same species, embryos during the midgastrula stage of development (6-7 hpf) exposed to 31.1 kJ m -2 UVB radiation had mortality rates higher than 70% (Dong et al. 2007). Lower hatching rates and several embryonic malformations, such as enlarged pericardial sacs, spinal deformities and minor spinal bending, also occurred after UVB exposure (Dong et al. 2007). Woundfin (Plagopterus argentissimus) embryos are sensitive even to low levels of UVB radiation (0.15 W m -2 ) when exposed for longer periods (14.5 h). These UVB levels correspond to 25% of the ambient irradiance observed in the bubbling ponds of some fish hatcheries in Arizona, where no embryo survival was measured after UVB exposure (Holmquist et al. 2014). The penetration of UVR in the water column depends on several variables such as the incident irradiance, optical properties of the water itself, phytoplankton, concentration of dissolved organic matter, and density of suspended particles. Noteworthy, the most significant factors modulating the UVR attenuation in the water column are the chlorophyll a and the chromophoric dissolved organic matter (reviewed by Häder et al. 2007Häder et al. , 2011. The amount of dissolved organic matter together with the nest location/depth choice can have an important role in the spawning success and embryotic survival of some fish species, such as the case of the bluegill sunfish (Lepomis macrochirus) in the lakes Tahoe (California-Nevada border, USA) and Giles (Pennsylvania, USA), (Olson et al. 2006Tucker et al. 2010). In surface waters of the Lake Tahoe with low dissolved organic carbon concentrations (high UVB transparency, 22.65 kJ m -2 ), almost 90% of the bluegill larvae died after 4 days of UVR exposure. Still, only 15% of the larvae died in the surface waters showing high dissolved organic carbon concentrations (low UVB transparency, 0.60 kJ m -2 ), (Tucker et al. 2010). Little information is available on the effects of UVA radiation on freshwater embryos. For example, Japanese medaka (Oryzias latipes) fertilized eggs (4 cell stage) showed high resistance to different UVA radiation levels. Nevertheless, the number of resulting deformed embryos increased and the hatching time was prolonged with an increase in UVA dose (Sayed and Mitani 2017). An increased hatching time was also noticed in zebrafish embryos when exposed to UVB radiation (Dong et al. 2007). Survival rates in the early stages of the seawater fish species, Atlantic cod (Gadus morhua), red seabream (Pagrus major), dab (Limanda limanda), North Sea plaice (Pleuronectes platessa), and northern anchovy (Engraulius mordax) are also affected by UVB radiation (Beland et al. 1999;Dethlefsen et al. 2001;Hunter et al. 1979Hunter et al. , 1981Steeger et al. 2001), ( Table 2). The UVB sensitivity of North Sea plaice depends on the timing of radiation exposure during embryonic development. No embryo survival was observed if the exposure to UVB radiation occurred at the early embryonic stage Ib, whereas no differences in survival were observed between non-irradiated and UV exposed gastrulation stage II embryos (Steeger et al. 1999). In Atlantic cod, more than 50% of the fertilized eggs died after 32 h of exposure to UVB (4.04 W m -2 ) under ozone layer depletion (around 20%) conditions (Kouwenberg et al. 1999). Dethlefsen et al. (2001) demonstrated that increasing levels of UVB as a consequence of ozone depletion (reduction to 270 Dobson units, DU), resulted in high embryo mortality, a decrease in hatching rate and a loss of buoyancy in dab and North Sea plaice 24 hpf exposed embryos. The embryonic development of North Sea plaice during spring spawning is not endangered by the actual UVB levels, or in the case of a reduction of 180DU. In Chile, actual levels of UVB observed in the Gulf of Arauco and Conception Bay are considered harmful to the planktonic anchoveta (Engraulis ringens) and common sardine (Strangomera Fig. 1 Summary of the current knowledge regarding the adverse effects of UVR in fish during early development (embryo and larvae). a Effects on survival, growth and development of body malformations and b tissues lesions, physiological, immunological and metabolic changes, including impairment of molecular and cellular processes. The color intensity represents the number of studies reporting each detrimental effect, in which the stronger the color/bar size the higher the number of references for each detrimental effect bentincki). Embryos from both species at stages I (without embryo) and II (early embryo: embryo covers half of the chorion) were irradiated over 4 days, whereas embryos from stage III (late embryo: the embryo covers more than half of the chorion) were exposed over 3 days. Both species showed a decrease in hatching success, changes in buoyancy and the development of several malformations. Embryo abnormalities included blisters on the yolk at the earliest stages, twisting of the notochord at different levels of intensity in the most advanced stages, and the presence of dead tissue in all stages of development (Vásquez et al. 2016).
High variability on UVB tolerance during larval stages have been shown among fish species (Fukunishi et al. 2012;Mitchell et al. 2008;Sucré et al. 2012;Vehniäinen et al. 2012). Caspian Sea salmon (Salmo trutta caspius) larvae exposed to 0.5-1.3 W m -2 experienced 100% mortality after 9 days of exposure to UVR (Kazerouni and Khodabandeh 2011). Similar mortality levels were observed in woundfin larvae exposed to 0.15-0.60 W m -2 for 15 h. Susceptibility of woundfin larvae to UVB radiation depends on the larvae development stage (Holmquist et al. 2014). Northern pike (Esox lucius) larvae subjected to daily doses of 1.8 and 2.7 kJ m -2 of UVB for 2 days, showed an increase in mortality by 10-20% (Häkkinen et al. 2004). In two sparidae species, black seabream (Spondyliosoma cantharus) larvae showed significantly higher survival rates than the red seabream larvae after exposing them to the same UVB conditions, suggesting that black seabream is probably better adapted to habitats with high UVB radiation than red seabream (Fukunishi et al. 2006). Short-term exposure in yellow perch showed that larval survival was inversely related to the UVR intensity dose, and similar results were obtained for exposure for 7 days for both UVA/UVB and UVA only (Boily et al. 2011). UVB negatively impacted the antipredator escape performance of Atlantic cod larvae exposed to a dose rate of UVB radiation (2.9 kJ m -2 h -1 ) for 15 h, which led to higher predation mortality (Fukunishi et al. 2012). UVB induced atrophy in the European seabass (Dicentrarchus labrax) fin fold after 2 days of exposure resulting in the loss of its normal swimming capability. Such atrophy may have been caused by apoptosis or necrosis processes (Sucré et al. 2012). In addition, the orientation behavior of red seabream larvae changed after exposure to UVB, 1.41 W m -2 .
The authors noticed that this species can develop UVB tolerance during ontogenic development (Sharma et al. 2007). After 4 days of exposure, only 50% of northern anchovy larvae survived and they showed retarded growth and development, several lesions in the eyes and in the brain, and evident dispersion of pigment within melanophores (Hunter et al. 1979). A decrease in the specific growth rate of Roho labeo (Labeo rohita) was observed after 40 days of exposure to 0.80 W m -2 (Singh et al. 2013).

Tissue lesions, physiological changes and immune system modulation
During the early development stages, several tissues are exposed to the harmful effects of UVR and physiological changes and immune system modulation have been reported (Tables 1 and 2, see summary in Fig. 1). The occurrence of lesions in the skin and gills of larvae exposed to UVR have been reported in recent years. Microscopic examination of native Lahontan redside minnow (Richardsonius egregius) and non-native warm-water bluegill sunfish skin exposed to UVR revealed that the native species is more adapted to high UV conditions that are characteristic of the environment of the near shore Lake Tahoe (Gevertz and Oris 2014). The skin from bluegill sunfish exhibited greater damage in both epidermis (more irregular and thinner) and dermis (formation of extracellular space) and potential DNA damage and impairment of cellular respiratory processes (Gevertz and Oris 2014;Gevertz et al. 2012). In sole (Solea solea), larval skin is slightly pigmented, and five days of exposure to UVB (2.15 kJ m -2 d -1 ) resulted in the appearance of the characteristic sunburn because of damaged cells and a reduction in the size of the mucous producing cells. In contrast, the highly pigmented skin of the turbot (Scophthalmus maximus) larvae was not affected after exposure to the same UVB conditions (McFadzen et al. 2000). UVR led to the appearance of sunburn cells and a reduction in the epidermis thickness, as well as the number of mucous producing cells in Caspean Sea salmon. The disappearance of pavement cell microridges and a lifting of the epidermis from the basal membrane were observed in the UVR exposed larvae (Kazerouni and Khodabandeh 2010). In addition, UVR exposure resulted in ionocytes deformation as well as a reduction in their number and cell size. These mitochondrial-rich cells are important for osmoregulation, respiration and excretion functions in the skin, especially during early development, and the damage observed in these cells may have contributed to the high mortality observed after UV exposure (Kazerouni and Khodabandeh 2011). The loss of osmoregulatory capacity in skin integument was also observed in European seabass larvae after 2 days of UVB (0.8 W m -2 ) exposure. Ionocytes were less abundant, and a decrease in the fluorescent immunostaining of two important osmoeffectors, Na ? /K ? -ATPase and the Na ? /K ? /2Clcotransporter, was observed in the UVB exposed larvae (Sucré et al. 2012). The gill filaments and lamellae in Indian major carp (Catla catla) were damaged by UVB radiation. Larvae exposed for 54 days (1.45 W m -2 , 15 min each day) showed damage in gill epithelium, and scanning electron microscopy revealed a decrease in the number of microridges. In addition, the pavement cells were severely affected (Sharma and Chakrabarti 2006). The same conditions of UVB radiation may also have had negative impacts on the digestive physiology and immune system of Indian major carp, contributing to poor growth and survival. After 55 days of exposure, the activity of the digestive enzymes amylase, trypsin and chymotrypsin decreased in carp larvae, suggesting an impairment of carbohydrate metabolism and protein digestion. Lower levels of the lysozyme, an important innate immune parameter, were observed in UVB treated fish, being an indicator of immune system suppression in carp larvae (Sharma et al. 2010). The tissue damage caused by UVB exposure in this species can be explained by the higher levels of glutamate oxaloacetate transaminase and glutamate pyruvate transaminase observed in UVB exposed larvae (Sharma et al. 2010). Further evidence of immune system suppression caused by UVB radiation during early development stages was suggested by Singh et al. (2013). They observed a decrease in leucocytes and in the myeloperoxidase activity in exposed larvae of roha labeo.

Impairment of molecular and cellular processes
At the molecular and cellular level, direct and indirect photochemical pathways characterize the toxic effects because of UVR exposure (Vincent and Neale 2000). Certain macromolecules, mainly nucleic acids and proteins, are directly targeted biologically by UVR due to the absorption of specific wavelengths by these molecules, followed by the dissipation of the absorbed energy under photochemical reactions Setlow and Setlow 1962;Wilson et al. 1995). Photochemical transformation or degradation of these macromolecules can lead to impairment or even loss of their specific biological functions. Photooxidative breakdown of proteins and cross-linking of amino acids can occur due to the excess excitation energy that results from the absorption by specific aromatic amino acids including tyrosine, tryptophan, and phenylalanine of specific UV wavelengths (Gerhardt et al. 1999;Wilson and Greenberg 1993;Wilson et al. 1995).
Among the UVR absorbing macromolecules, nucleic acids are the most sensitive to UVR exposure (Buma et al. 2003;Vincent and Neale 2000). The wavelength of maximum absorbance of nucleic acids is around 260 nm, but also extends into the UVB spectral region, and can result in the photo-destruction of nucleotides, essentially the pyrimidines, thymine, and cytosine, generating different photoproducts (Görner 1994;Setlow 1974). The degree and type of DNA damage depends on the intensity and the specific wavelength of the exposure. Three types of photoproducts can be generated as a result of UVR exposure: cyclobutane pyrimidine dimers (CPDs), pyrimidine [6-4] pyrimidone photoproduct (6-4 PP) and photohydrates (Görner 1994). Moreover, the DNA damage caused by these photoproducts can induce the impairment of essential cellular processes, such as blocking DNA replication and transcription that can result in severe deleterious consequences, including mutagenesis, apoptosis, and carcinogenesis (Hart and Setlow 1974;Mitchell et al. 1993Mitchell et al. , 2001Setlow et al. 1989Setlow et al. , 1993. Pyrimidine dimers, CPDs (mainly the thymine TT dimers), are the predominant photoproducts generated as a result of UVB exposure (Buma et al. 2003;Thoma 1999) and CPDs have been described as a potential inhibitor of embryonic and larval development in fish (Lesser et al. 2001;Vehniäinen et al. 2012;Vetter et al. 1999). In blackfin icefish (Chaenocephalus aceratus) fertilized eggs, biological weighting functions (BWF) and exposure-response curves showed that CPDs formation is significantly correlated with the cumulative daily dose of UVB radiation. It was calculated that damage to DNA of approximately 35 CPD/Mb was induced in icefish eggs (unhatched, late-somitic stages) due to ozone layer depletion (Malloy et al. 1997). Likewise, due to the abundance, buoyancy and transparency of icefish eggs, the authors propose that this species could be a potential biological indicator of the DNA-damaging effects of UVB in zooplanktonic communities confined to Antarctic surface waters (Malloy et al. 1997). In another study, 10% of the estimated Atlantic cod embryo mortality was caused by an increased load of 10 CPD/Mb (megabase) to DNA after exposure over 1 h to 150 kJ m -2 of UVB. The CPDs loads were generally lower in eggs than in larvae, and only wavelengths shorter than 360 nm were shown to have a strong effect on CPDs formation. The eggs' characteristics, including the chorion (membrane) and the fluid-filled perivitelline space, can provide the embryos with some protection from UVB induced DNA damage (Browman et al. 2003).
This direct relationship between the number of generated pyrimidine dimers and the increase in mortality was also observed in fathead minnow (Pimephales promelas) embryos and rainbow trout (Oncorhynchus mykiss) larvae, (Applegate and Ley 1988;Mitchell et al. 2008). UVB induced DNA damage in the form of CPDs was also observed in different strains of Japanese medaka larvae exposed to different UVB conditions (2.7-5.9 mW m -2 ) and a positive correlation was observed between the UVB intensity and the generated CPDs number (Armstrong et al. 2002). Similar findings were reported for northern pike exposed for 2 days to UVB (0.24-1.7 W m -2 , 3 h d -1 ), (Vehniäinen et al. 2012). The CPDs localization was investigated using immunohistochemistry in the northern pike tissues. At the highest UVB irradiances (0.97-1.7 W m -2 ), CPDs were found not only in the epidermal cells but also in the brain, eye and muscle. Such DNA damage in the eyes and brain in this species is associated with the severe behavioral disorders and mortality that were observed in this study (Vehniäinen et al. 2012). To the best of our knowledge, information on the occurrence of 6-4 PP photoproducts during early fish development is scarce. Although 6-4 PP induced lesions are less frequent under exposure to UVR, these photoproducts can have more damaging effects by blocking replication and transcription (Mitchell and Nairn 1989).
Several indirect photochemical mechanisms also mediate the damaged caused by UVR exposure; they generally involve the absorption of some photosensitizing agents, generating reactive oxygen species (ROS), such as superoxide radicals (O 2 -Á ), hydrogen peroxide (H 2 O 2 ), and hydroxyl radical ( Á OH), (Kieber et al. 2003). These highly energetic oxidative species can diffuse and react rapidly with several cellular components, resulting in damaged sites that can differ from the site of photoproduction (Vincent and Neale 2000). Cha et al. (2011) observed in zebrafish embryos a significant increase in intracellular ROS after exposure to 0.5 kJ m -2 UVB. The detection of ROS in the embryos was analyzed using the oxidationsensitive fluorescent probe dye, 2 0 ,7 0 -dichlorofluorescein di-acetate (DCF-DA), (Cha et al. 2011). Another study in the same species revealed more ROS in 120 hpf larvae when the embryos were exposed to UVB, 0.25 kJ m -2 (Hurem et al. 2018). Despite the involvement of UVA in the mechanisms of repairing DNA damage, exposure to UVA in North African catfish (Clarias gariepinus) resulted in a significant increase in DNA damage, confirmed by the comet assay technique, (Mekkawy et al. 2010). As the levels of DNA damage were positively correlated with the lipid peroxidation results, the authors emphasized an oxidative nature of DNA damage in that study (Mekkawy et al. 2010). The DNA damage triggered by UVA is generally an indirect mechanism, and involves the formation of reactive chemical intermediates, including O 2 -Á and Á OH radicals and their interaction with the DNA. Such interaction can result in the DNA strand breaks, DNA-protein cross-links, and alkali labile sites (Thoma 1999). In addition, lipid peroxidation (LPO) and consequent oxidative damage was observed in zebrafish larvae when the embryos were exposed to UVA doses C 374 kJ m -2 , and afterwards resulted in behavioral changes including a reduction in larval movement (e.g., changes in the resting heart rate, less time spent swimming), (Hurem et al. 2018).
Exposure to UVR during the early development stages can induce transcriptional changes. In zebrafish embryos, the expression of osteonectin (osn) increased after exposure to 4.9 W m -2 of UVB for 150 min, and it was accompanied by an increase in mortality and developmental abnormalities (Nuñez et al. 2012). The increase in osn expression may be one of the plausible molecular mechanisms of UV radiation-induced phenotypic developmental abnormalities (Nuñez et al. 2012). In mammals, osteonectin has been described as a facilitator in the development of skin tumors in response to UVR exposure (Aycock et al. 2004).
Moreover, the expression of p53 was also upregulated in response to UVR (UVB and shorter UVA wavelength). This gene is generally associated with the DNA repair system, and low expression levels during embryogenesis are associated with normal development (Nuñez et al. 2012). Additionally, UVB radiation induces p53 activation in the brain, suggesting that neural molecular changes can be associated with behavioral changes observed in pike larvae (Vehniäinen et al. 2012). DNA damage and subsequent activation of p53 was observed in Atlantic cod embryos exposed to UVB. The activation of p53 is generally associated with DNA damage after exposure to UVB, which results in delays in cell division while DNA repair is taking place (Lesser et al. 2001). However, when the embryonic cells are unable to repair the DNA damage due to UVB exposure, the apoptosis pathway can be triggered. The activation of apoptotic pathways in response to UVR was reported in Japanese flounder (Paralichthys olivaceus) embryos, (Yabu et al. 2003). Using a caspase-3-like activity assay and terminal deoxynucleotidyl transferase-mediated dUTP nick end-labeling staining, the authors observed an induced extensive apoptosis in embryos after exposure to UVR (0.2-1 kJ m -2 ). Flounder embryos exposed to UVR revealed apoptotic cells distributed throughout the body, particularly in the head, spinal cord, yolk sac, heart and larval fin (Yabu et al. 2003).
An up-regulation in cellular defense (superoxide dismutase sod1, catalase cat1, heat shock protein 70 hsp70) and immune (interleukin-1 beta il-1ß, tumor necrosis factor alpha tnfa) related genes was observed in 96 hpf zebrafish larvae after exposure to UVB (0.11 W m -2 ) over 4 days (Aksakal and Ciltas 2018). The authors suggest that the antioxidant defense response and the chaperoning mechanisms involved in larvae development were induced after UVB exposure. Furthermore, UVB seems to modulate the innate immune system in zebrafish larvae, as evidenced by the increase in gene expression of proinflammatory cytokines il-1b and tnfa (Aksakal and Ciltas 2018). Similar results were observed by Banerjee and Leptin (2014) where the up-regulation of il-1b was proportional to the increase in UVB dose. The authors suggested that such increased levels of this pro-inflammatory cytokine counteract the lethal effect of high doses of UV in zebrafish.

Juveniles and adults
Growth reduction, behavioral and metabolic changes UVB radiation is known to affect fish growth and their body condition (Tables 3 and 4, see summary in Fig. 2). In Atlantic salmon (Salmo salar) juveniles, a reduction in growth and loss in body condition was observed after exposure in cages for 8 weeks to sunlight supplemented with UVB radiation (at depth of 1 cm-31.0 kJ m -2 d -1 , at bottom of the cage-3.50 kJ m -2 d -1 ). The results were compared with fish kept under UVB depleted natural sunlight (at depth of 1 cm-0.45 kJ m -2 d -1 , at bottom of the cage-0.00 kJ m -2 d -1 ), (Jokinen et al. 2008). Subadult three-spined stickleback (Gasterosteus aculeatus) showed a similar reduction in growth and loss in body condition after exposure over 10 weeks to UV-enhanced conditions (0.33 W m -2 , daily dose-6.48 kJ m -2 ), when compared with fish exposed to natural UVB conditions (0.21 W m -2 , daily dose-3.89 kJ m -2 ), (Vitt et al. 2017). A potential reduction in the investment to the adaptive immunity due to a lower splenosomatic index was observed in the UVBenhanced treatment (Vitt et al. 2017). As suggested by the authors mentioned above, under these chronic stress conditions, fish possibly allocate energy for repairing UVB-induced DNA damage mechanisms (nucleotide excision repair) instead of using the energy for digestion (Jokinen et al. 2008;Vitt et al. 2017). Changes in feeding behavior, appetite reduction and catabolism stimulation in fish can be triggered by stress during UVB exposure (Arts et al. 2010;Holtby and Bothwell 2008;Vitt et al. 2017). Such suppression in the energy allocated for digestion was also suggested by the reduced growth observed in sea chub (Graus nigra) juveniles after short-term exposure (7 days) to a total UVB dose of 22.68 kJ m -2 .
The mechanisms underlying how fish perceive UVR and the changes in behavior that are triggered by UVR are poorly understood. Nevertheless, exposure to UVB generally results in an increase in oxygen consumption. This increase can be related to restless behavior and to an increase in swimming activity (Alemanni et al. 2003;García-Huidobro et al. 2017). Such an increase in oxygen consumption seems to be positively correlated with an increase in UVB doses in rainbow trout juveniles (Alemanni et al. 2003).  Kaweewat and Hofer (1997)    Nevertheless, zebrafish showed reduced swimming performance when exposed to UVB for two weeks (3.3 W m -2 , 1.19 kJ m -2 d -1 ). This reduction in muscular activity can be part of a mechanism to minimize intrinsic reactive oxygen species (ROS) production after exposure to UVB . The predatory performance of the reef fish Patagonothen cornucla, a species inhabiting the intertidal areas of the Patagonian coast, was affected after UVB exposure (Valinas and Helbling 2016). The prey capture time increased in the juveniles exposed to UVB for 9 days (Valinas and Helbling 2016). Longterm exposure to UVA radiation (30 days, daily dose of 55.6 kJ m -2 ) in three-spined stickleback adults affects reproductive performance by decreasing sperm velocity and by inducing changes in the breeding coloration (Rick et al. 2014).

Tissue lesions and physiological changes
Exposure to UVR results in severe lesions in organs and tissues, including those that are directly exposed such as the skin and eyes (Tables 3 and 4, see summary in Fig. 2). The skin constitutes a primary barrier and acts as the interface between the fish and its external environment (Esteban 2012). Unlike humans, the inexistence of a keratinized outer layer in fish skin results in higher vulnerability to UVR exposure. In the Lahontan cutthroat trout (Oncorhynchus clarki henshawi) and rainbow trout, signals of sunburn appeared after 3 days (UVB daily dose, 34.2 kJ m -2 ) of exposure. The darkening of the skin may have resulted from melanosome dispersion, a characteristic symptom observed after sunburn. The largest area of sunburn occurred just after the head and relatively close to the dorsal fin. After 6 days of exposure, both species had a significant fungal infection. No sunburn or fungal infection was observed in the apache trout (Oncorhynchus apache) or razorback suckers (Xyrauchen texanus) under the same UVB conditions (Fabacher and Little 1995). Appearance of sunburn cells, epidermis necrosis and edema, disruption between the epidermis and the basement membrane, changes in the skin thickness, and hyperplasia of superficial mucous cells are some of the microscopic changes observed in the cutthroat trout skin after UVB exposure. Below the stratum compactum, the melanocyte layer was thickened and more disorganized in some areas in UVB exposed fish (Blazer et al. 1997).
Other evidence of the harmful effects of UVB in the skin are changes observed in the metabolically active mucous tissue. The mucous layer corresponds to the outer extrinsic barrier representing the first line of defense against pathogenic microorganisms in the surrounding water (Esteban and Cerezuela 2015; Gomez et al. 2013). The number of mucous producing cells, also called goblet cells, decreased significantly in the dorsal skin of juvenile Arctic char (Salvelinus alpinus), European minnow (Phoxinus phoxinus), Danube bleak (Alburnus chalcoides) and rainbow trout after 5-7 days of UVB exposure (7.5 kJ m -2 d -1 ). No effects of UVA were observed in the number of mucous producing cells. A decrease in the number of mucous producing cells can affect the innate immune response because of less mucous production (Kaweewat and Hofer 1997). Similar findings were observed in adult fathead minnow after acute exposure to UVA/UVB for 8 h (irradiance levels 43 W m -2 ). In addition, a reduction of 50% in the number of epidermal club cells (alarm cells) was observed in the UVR treated fish, but without significant changes in the epidermis thickness (Manek et al. 2012). Both mucous and alarm cells in the dorsal skin of North African catfish were severely damaged after 3 days exposure to UVA (total dose-10.8 kJ m -2 ). In the ventral skin, an increase in the epidermis thickness as well as the appearance of pyknotic cells and signs inflammation were visible in the UVA exposed fish (Sayed et al. 2007). During the last 3 decades, several platyfish interspecies genetic hybrid models were developed to investigate the role of UVR on the induction of cutaneous malignant melanoma (Ahmed and Setlow 1993;Mitchell et al. 1993Mitchell et al. , 2001Setlow et al. 1989Setlow et al. , 1993. A pioneer study by Setlow et al. (1989) demonstrated that UVR has a potential role in the etiology of melanoma. The authors observed that backcross hybrids irradiated with UVB showed a higher frequency of induced melanoma compared to the spontaneous melanoma levels. Later, a study performed by Mitchell et al. (2010) on the UVR action spectrum for melanoma induction, using the Xiphophorus couchianus model, showed that UVB could induce melanoma in this animal model; however, UVA did not. These authors also suggested that the direct DNA damage (formation of CPDs and 6-4 PP) associated with UVB could play a major role in the initiation of melanomas. Moreover, the efficient capacity of fish to repair UVB-induced DNA damage by the photoenzymatic repair system seems to decrease the occurrence of melanoma in the Xiphophorus melanoma model (Mitchell et al. 1993(Mitchell et al. , 2001.
Corneal epithelium/stroma damage and persistent cataractous changes in the anterior part of the crystalline lens occurred in the eyes of rainbow trout exposed to UVB 0.75-15 kJ m -2 (Cullen and Monteithmcmaster 1993;Doughty et al. 1997). Furthermore, cataractous changes in rainbow trout under long-term exposure to UVB (205 days, dose 5.12 kJ m -2 d -1 ) include 'doughnut' opacities, discrete anterior subcapsular and peri-nuclear haze in the fish eye lenses, supporting the cataractogenic role of UVR (Cullen et al. 1994). In zebrafish, UVB exposure resulted in a reduction in corneal thickness but the lens capsule diameter and fiber thickness were not affected (Marlow 2010). UVB-exposure over 6 days resulted in an increase of Diplostomum spathaceum metacercariae parasites present in the rainbow trout eye lenses compared with the control group. This increase in the number of parasites was accompanied by a thinner mucus layer in the dorsal skin (Markkula et al. 2007). An ultrastructural study demonstrated that acute exposure to UVB 1.4 W m -2 (10.08 kJ m -2 ) for 2 h changed the outer surfaces of the eyes and skin in ayu (Plecoglossus altivelis). Damage to the specialized microridges, which are generally related to stress adaptation, was observed in the skin and eyes of UVB exposed fish (Sharma et al. 2005). UVR-induced changes and damage have also been noted in organs that were not directly exposed to radiation, such as, in the red blood cells and liver (Table 3). In Japanese medaka, UVA exposure for 3 days promoted apoptosis and induced several morphological malformations in red blood cells, such as acanthocytes, hemolyzed cells, sickle cells, swollen cells and cell membrane lysis. Moreover, several nuclear abnormalities including deformed nuclei, eccentric nuclei, nuclear budding and bilobed nuclei were identified in the UVA exposed fish (Sayed 2018;Sayed et al. 2013. Hepatic tissue was also damaged by UVA in Japanese medaka. Dilated thick-walled blood vessels in association with inflammatory lymphocytic infiltration, diffused hepatic pigments, irregular shape of some nuclei and loss of cellular integrity, fatty degeneration (lipidosis) associated with lipid accumulation and cytoplasmic vacuolation were evident in UVA-treated fish (Sayed et al. 2007(Sayed et al. , 2013. Several physiological responses to UVR exposure have been reported in juvenile and adult fish (Tables 3  and 4, see summary in Fig. 2). Plasma cortisol levels increased after exposure to UVB radiation in rainbow trout, roach (Rutilus rutilus) and fathead minnow. Such increases in cortisol levels, a primary endocrine response of fish to stressful conditions, generally induce physiological, immunological and metabolic changes, thus affecting fish growth (Manek et al. 2012;Markkula et al. 2006;Salo et al. 2000a, b). Moreover, the role of UVB as an endocrine disruptor was investigated in the platyfish Xiphophorus coachianus melanoma model. The exposure to a sublethal dose of UVB for 8 min in adult males resulted in a remarkable increase in the circulating levels of both sex steroid hormones and gene expression levels in its associated hormone receptor, ARa. The authors suggested that this photoendocrine response could be a key factor in male tumorigenesis due to its potential involvement in the initiation of UV-induced melanoma (Mitchell et al. 2014). Short-and long-term UVB exposure decreases both hematocrit and plasma total protein levels in Atlantic salmon, roach, rainbow trout and common carp (Cyprinus carpio). A decrease in both indicators revealed poor nutritional and health status in the fish exposed to UVB (Jokinen et al. 2008(Jokinen et al. , 2011Markkula et al. 2007;Salo et al. 2000b). Nevertheless, the total protein levels in the plasma increased in North African catfish after 3 days of UVA exposure (3 h d -1 ). UVAtreated fish revealed changes in several biochemical, hematological and metabolic parameters. For example, the levels of creatinine, alanine amino transferase and aspartic amino transferase increased after UVA treatment, but most of the parameters, such as the levels of glucose, cholesterol, alkaline phosphatase, hematocrit, hemoglobin, red blood cells, and platelets, decreased in UVA exposed catfish, suggesting that UVA can also induce a metabolic and physiologic disturbance in adult fish (Sayed et al. 2007).

Immune system modulation
The reported detrimental effects of UVR in innate and acquired immune systems in fish showed variations between species and the duration of exposure. Impacts of UVA and UVB include negative effects on both cellular and humoral components of the fish immune A database of physiological, cellular and molecular UV damages in seawater species was generated using literature searches on experiments of acute, short-and long-term exposure to UVA and UVB. UV harmful effects occur at different levels: growth, behaviour and abnormalities development; tissue lesions; physiological, metabolic and immunological responses, and impairment of both molecular and cellular processes.  Fig. 2 Summary of current knowledge regarding the adverse effects of UVR in juveniles and adult fish. a Effects on growth and changes in behavior and metabolism; b changes in physiology status and impairment of molecular and cellular processes; c changes in the immune system. The color intensity represents the number of studies reporting each detrimental effect, in which the stronger the color/bar size the higher the number of references for each detrimental effect. RBA respiratory burst activity, NCC non-specific cytotoxic cells system (Tables 3 and 4, see summary in Fig. 2). One of the most common negative effects of UV exposure is the change in the proportion of the different leucocytes in peripheral blood. In Rutilus rutilus, which generally inhabits lakes with low UVR, exposure to 4.3 kJ m -2 UVB resulted in a significant increase in the percentage of granulocytes (control-4-6%, UVB exposed fish-35-37%) and a significant decrease in the percentage of lymphocytes (control-56-60%, UVB exposed fish-30-33%) on day 1 of post-irradiation. These changes were not observed under UVA exposure for the same species. However, on day 14 after the exposure, the percentage of these leucocytes returned to similar levels to those observed in the control. No significant changes were observed in the percentage of monocytes or thrombocytes (Jokinen et al. 2000;Salo et al. 2000a, b). Under the same UVB dose, similar results were obtained in the common carp but not in the rainbow trout. The increase in the UVB dose (11 kJ m -2 ) in the rainbow trout resulted in a decrease in the lymphocytes percentage, but the granulocytes percentage remained unchanged. The number of thrombocytes increased 20% in the UVB exposed trout. In the common carp, these changes were dependent on the UVB dose (Markkula et al. 2006). The high levels of cortisol observed in the UVB exposed fish suggest that the occurrence of lymphopenia/granulocytosis in blood can be a stress response indicator to UVB exposure (Jokinen et al. 2000;Markkula et al. 2006;Salo et al. 2000a). In addition, repeated exposure (short-and long-term) to UVB also induced peripheral blood lymphopenia and granulocytosis. UVB exposure during six consecutive days (total dose 9 kJ m -2 ) improved the immune function of uninfected rainbow trout by increasing the number of circulating granulocytes in the blood (Markkula et al. 2007). Long-term exposure to UVB (28 days, total dose 28.8 kJ m -2 ) decreased the proportion of lymphocytes in the blood of common carp. Fish, showing lesions and infection in the dorsal skin during the third week of exposure, also presented an increase in the proportion of granulocytes and monocytes (Markkula et al. 2005). These results suggested that lymphocytes are sensitive to UVB radiation, and the increase of granulocyte-to-lymphocyte ratio in the UVB exposed fish can result in a higher contribution of the innate immune system compared to the adaptive immune system (Markkula et al. 2005;Vitt et al. 2017).
Production of ROS during phagocytosis, also known as respiratory burst activity (RBA), is essential for the effective destruction of pathogens (Uribe et al. 2011). This cellular component of the innate immune system is generally suppressed in the head kidney phagocytes after UVB exposure. In the roach, RBA of both macrophages and granulocytes decreased during the first two days after the UVB exposure (4-5 kJ m -2 ), but the capacity of both leucocyte types to produce ROS was restored on day 7 post irradiation (Jokinen et al. 2000;Salo et al. 1998Salo et al. , 2000b. Similar findings in the common carp (0.5-5 kJ m -2 ) and rainbow trout (10 kJ m -2 ) granulocytes were described on day 1 after irradiation (Markkula et al. 2006). In the common carp, granulocytes RBA was significantly decreased in fish irradiated with 2.4 kJ m -2 d -1 UVB for 1 week, whereas long-term exposure to UVB radiation decreased in fish irradiated with both 1.2 and 2.4 kJ m -2 d -1 UVB doses (Markkula et al. 2005). In contrast, leucocytes respiratory burst activity in peripheral blood generally increased in fish after UVB exposure. This probably occurred because of the granulocytosis observed in peripheral blood (Markkula et al. 2006;Salo et al. 2000a, b). In rainbow trout, repeated UVB exposure can increase or decrease the peripheral blood granulocytes RBA (Markkula et al. 2007(Markkula et al. , 2009. Another important cellular component of the innate immune system affected by UV exposure is the activity of non-specific cytotoxic cells, which generally shows the same pattern as RBA in peripheral blood and head kidney (Jokinen et al. 2000;Markkula et al. 2005Markkula et al. , 2006Salo et al. 1998). NCC in teleosts are the equivalent of the mammalian natural killer (NK) cells, being part of the natural defense against parasites, virus or even neoplastic diseases (Uribe et al. 2011); therefore, this protective mechanism can be affected by exposure to UVR.
Humoral components of the immune system of fish can also be suppressed by the exposure to UVR. In rainbow trout, four hours of exposure to UVB (10 kJ m -2 ) resulted in a significant decrease in lysozyme, which provides an essential defense against Gram-positive bacteria and activates both phagocytes and complements activity (Markkula et al. 2006;Saurabh and Sahoo 2008). Long-term exposure to UVB (54 days, dose at the bottom of the cage-0.12 kJ m -2 d -1 ) in Atlantic salmon decreased both complement bacteriolytic activity and immunoglobulin M (IgM) levels in the plasma (Jokinen et al. 2008(Jokinen et al. , 2011. Salo et al. (2000a) demonstrated that UVA could also be a modulator of the immune systems in roach, by decreasing the plasma immunoglobulin levels. However, short-term exposure to UVB had no clear negative effects on the roach plasma immunoglobulin levels (Jokinen et al. 2001). Plasma IgM concentration decreased significantly in common carp juveniles exposed to the highest dose of UVB (2.4 kJ m -2 , total dose 28.8 kJ m -2 ) after 4 weeks but not with other doses. Fish showing skin lesions and fungal infections also revealed significantly lower IgM levels compared to healthy fish (Markkula et al. 2005). This decrease in the IgM levels may be an indicator of the impairment in lymphocytes function due to UVB exposure (Jokinen et al. 2008(Jokinen et al. , 2011. In addition, IgM levels may be indicative of the overall function of the acquired immune system, as the synthesis of IgM requires the activity of several cell populations, such as T-/B-lymphocytes and cytokines (Jokinen et al. 2008). The potential role of UVB to modulate the functioning of lymphocytes, and subsequently compromise the immune status of the fish, was also demonstrated through the suppression of the spleen lymphoproliferative (lymphocytes mitogen-activated proliferation) responses after exposure of R. rutilus to moderate UVB levels (Jokinen et al. 2000). In fact, the immune system modulation/suppression that resulted from the exposure to UVB can subsequently decrease the resistance of the fish to diseases. A high incidence of fungal infection on the skin of rainbow trout and lahontan cutthroat trout was observed after 6 days of exposure to UVB (1.9 W m -2 , ambient mid-latitude summer irradiance (Fabacher and Little 1995). Moreover, long-term UVB exposure in guppy (Poecilia reticulate) increased the fungal infection rates by white spot in their offspring, suggesting that UVB has an immunosuppressive effect across generations (Kazerouni et al. 2017). The negative effects of UVB on disease resistance were observed in rainbow trout, where the juveniles exposed to UVB showed a suppression in the resistance against trematode parasites (Diplostomum spathaceum) and bacteria (Yersinia ruckeri), (Markkula et al. 2007).

Impairment of molecular and cellular processes
DNA damage in the form of CPDs increased significantly on the skin of moon wrasse (Thalassoma lunare) after UVR treatment (1 h, UVA-21.9 kJ m -2 , UVB-48.2 kJ m -2 ). The level of DNA damage in the juveniles' wrasse skin was negatively correlated with the integrated mucus absorbance, represented by the amount of mycosporine-like amino acids (MAAs) present in the external skin mucus (Braun et al. 2016). UVR-induced DNA damage and its repair in the skin of platyfish species were described for the first time by Ahmed and Setlow (1993). Induced CPDs are up to 10 times more frequent than (6-4) PPs, and the efficiency of photoproduct generation seems to be tissue dependent in platyfish species. CPDs and (6-4) PPs can be efficiently repaired by dark-and light-dependent repair processes, being highly diverse in the Xiphophorus genus. This nucleotide excision repair showed high efficiency for the (6-4) PPs compared with the CPDs in most platyfish species (Mitchell et al. 1993(Mitchell et al. , 2001. The authors concluded that differences in vulnerability of the several platyfish species/hybrids to the damaging effects of UVB to DNA could be a result of differences in the skin structure (e.g. morphology and composition of scales) and pigmentation. For example, in the absence of scales, the photoproducts generation was similar between skin and fins (Mitchell et al. 2001). DNA damage was also observed in the liver of G. laevifrons after 3-5 h of exposure to UVB (Carrasco-Malio et al. 2014). UVA-induced DNA damage and repair in erythrocytes was investigated in Japanese medaka and North African catfish. Exposure of adult medaka during 3 days to UVA radiation resulted in high frequencies of c-H2AX foci (marker for DSBs) and tail moment score (neutral comet assay) immediately after exposure, and then decreased within 24 h (Sayed and Mitani 2017). A replication-independent formation of UVA-induced double-strand breaks through the generation of ROS and oxidative damage was suggested by the authors. In North African catfish, DNA damage and repair was a function of UVA dose and recovery after exposure. Comet assay results demonstrated that fish exposed to UVA had a significant increase in DNA damage compared with the controls (Sayed 2018). Based on the morphological changes, DNA damage and apoptosis results, the data suggest that Japanese medaka might be more sensitive to UVA exposure than the North African catfish. The effects of UVR on oxidative stress have been reported in some fish species (Carrasco-Malio et al. 2014;Kazerouni et al. 2017). In adult zebrafish, changes in muscle and skin antioxidant status were observed after ambient UVB (4 h, 1.72 W m -2 ). Total glutathione concentration and glutathione peroxidase activity decreased after 6 and 12 h of exposure. In contrast, superoxide dismutase and catalase activity peaked at 6 and 12 h, respectively. In addition, lipid peroxidation was revealed by the increase in ROS in UVB-treated fish compared to the control fish. These results showed significant free radical-mediated lipid membrane damage by an increase in cytosolic hydrogen peroxide (H 2 O 2 ) and superoxide anion radicals (O 2 -) due to UVB exposure (Charron et al. 2000). An increase in the ROS-induced damage to proteins and membranes was also observed in adult zebrafish when exposed to UVB (daily dose 1.19 kJ m -2 ) during 15 days . These authors suggested that in the presence of UVB, fish could minimize the ROS production by reducing their swimming activity. In Girella laevifrons, differences in the antioxidant capacity were observed between liver and muscle. Catalase activity in the liver increased with the duration of UVB exposure (0.58 W m -2 ), but the superoxide dismutase activity peaked at 2 h, followed by a decrease down to basal levels after 5 h of radiation. In the muscle, lipid peroxidation increased during 5 h of exposure to UVB (Carrasco-Malio et al. 2014).

Photoprotective mechanisms
Fish have developed different protective strategies to cope with the harmful effects of UVR exposure. These include: (1) behavioral changes to avoid UVB exposure; (2) the presence of physical barriers such as scales; (3) acquisition of UV-absorbing compounds; (4) elimination of toxic compounds resulting from the exposure to UVR; and/or (5) UV-induced damage repair mechanisms in proteins and DNA (Dahms and Lee 2010;Zagarese and Williamson 2001;Zamzow 2003Zamzow , 2004Zamzow , 2007. The changes in behavior to avoid UVR exposure and photoprotective mechanisms by UV-absorbing compounds are very important components in overall UVR protection (Dahms and Lee 2010). Avoidance of UVR exposure has been reported in several fish species. For example, coho salmon (Oncorhynchus kisutch) juveniles reared in outdoor flumes can perceive UVR and avoid overexposure by exhibiting a shade-seeking behavior. Agonistic interactions between individuals and natural feeding behavior were suppressed in the fish exposed to UVR (Holtby and Bothwell 2008). During a UVR avoidance experiment, black seabream (but not red seabream) showed the capacity to avoid the highest UVB levels (20 min of exposure to 1.1 W m -2 ) by hiding in the half of the tank that was covered with UV-blocking film. In fact, black seabream during their early life cycle stages generally live in shallower water with high UVB radiation levels (Fukunishi et al. 2006). Bluegill sunfish locate their nests in higher depths characterized by high UVR attenuation and in areas protected by overhanging trees or other structures, therefore increasing larvae survival success (Olson et al. 2006. One of the strategies of both intertidal fish species, Girella laevifrons and Patagonotothen cornucola to avoid UVR is living under rocks and/or areas abundant in algae (Carrasco-Malio et al. 2014;Pulgar et al. 2015Pulgar et al. , 2017Valinas and Helbling 2016).
UV-absorbing compounds (UVACs) play a protective role during UVR exposure in aquatic organisms and include carotenoids, mycosporins, melanin, scytonemin and the well-described mycosporine-like amino acids (MAAs). UVACs are common UVsunscreens that are synthesized by photosynthetic organisms and can be acquired by animals through their diet (Carefoot et al. 1998(Carefoot et al. , 2000Mason et al. 1998;Riemer et al. 2007;Zamzow 2004). MAAs are small photostable water-soluble compounds that can absorb UVB and UVA wavelengths (between 309 and 360 nm). Besides their role in UV absorbance, these intracellular compounds can also be responsible for ROS scavenging (Oren and Gunde-Cimerman 2007). During early development, MAAs including gadusol were detected in eggs and larvae of several marine fish species (Lesser et al. 2001;Plack et al. 1981;Zagarese and Williamson 2001). Several MAAs were identified in the eye tissue of 52 reef fish species, including palythene (k max-360 nm), palythinol (k max-332 nm), asterina-330 (k max-330 nm) and palythine (k max-320 nm), (Dunlap et al. 1989). Most studies found that UVACs including MAAs in fish were found in the epidermal mucus. Fabacher and Little (1995) identified a UVB-absorbing compound in skin extracts of razorback suckers (Xyrauchen texanus) and apache trout (Oncorhynchus apache). This compound was more abundant in these two species when compared to the amount observed in rainbow trout and Lahontan cutthroat trout. These last two species showed high signals of fungal infection and skin sunburn after exposure to UVR (Fabacher and Little 1995). UVA and UVB absorbing compounds were detected in the epidermal mucous of more than 120 tropical reef fish species. For example, in the Hawaiian saddle wrasse (Thalassoma duperrey), UVR exposure induced changes in the absorbance of its mucous (Zamzow and Losey 2002). The type of diet and UVR exposure affected the UV-absorbing compound content in the mucous. Furthermore, differences in UV absorbance by T. duperrey epidermal mucous were found between male and females. Mucous in females seems to absorb less UV and additionally, females showed higher levels of skin damage than males (Zamzow 2004). The use of MAAs as a strategy of protection against UVR exposure in sea chub was suggested by Carrasco-Malio et al. (2014). In the Australian reef fish Pomacentrus amboinensis and Thalassoma lunare, MMAs act as a first line of defense against UVR-induced DNA damage (UVA-6.1 W m -2 , UVB-13.4 W m -2 ). In P. amboinensis, the epidermal mucous absorbance was significantly higher in the UVB exposed juveniles when compared with the control (Braun et al. 2016). The mechanisms underlying the transport of MAAs into the mucous layer and regulation of MAAs content in mucous are poorly understood. Braun et al. (2016) suggested two plausible explanations for the increase in mucus absorbance observed in P. amboinensis. First, MMAs can be stored in tissues such as the gastrointestinal tract and gonads and then transferred by specific transporters, and accumulated in the mucous producing globet cells. On the other hand, MAAs originating from bacteria with a functioning shikimate pathway can be potentially transferred into fish via the epidermal fish mucous (Braun et al. 2016). Juveniles of the reef fish Patagonotothen cornucola fed with a rich-UVAC diet showed a significant lower respiration rate than those fed with a poor-UVAC diet (Valinas and Helbling 2016). Cha et al. (2011) demonstrated how phlorotannins in zebrafish can have a significant photoprotective role during UVB exposure. Zebrafish embryos pretreated with these polyphenols, synthesized by algae, exhibited reduced hyperpigmentation, as well as UV-B-induced reactive oxygen species and nitric oxide levels, thus offering protection against UV-B-induced cell death (Cha et al. 2011).

Interaction between UVR, environmental factors, and pollutants
The anthropogenic impact of climate change (e.g., changing temperature and atmospheric CO 2 , precipitation, and ice melting) and other stressors on ecosystems are resulting in an increasingly difficult habitat for organisms. The health status of aquatic animals generally depends on a combination of several environmental factors, and where a disturbance in their tolerance limits occurs, they can become stressors (Schulte 2014). Indeed, recent evidence indicates that the deleterious effects of UVR may be enhanced by interactions with some environmental factors (e.g. increased temperature and hypoxia due to climate change) or pollution (Häder et al. 2015;Häder and Barnes 2019). The interaction between UVR and other stressors can increase the negative impact compared to UVR exposure alone, and can result in antagonistic, synergistic or additive effects that affect fish physiological status, growth or survival .
The interactive effects of UVB radiation and temperature have been reported in a few studies. In zebrafish, UVB radiation and temperature (low 24°C and high 30°C) can disrupt embryonic metabolism, modulate immune system response and impair its embryonic development (Aksakal and Ciltas 2018). Low and high temperatures led to an increase in hatching time, development of malformations and mortality. Additionally, for both temperatures, an upregulation in oxidative stress (superoxide dismutase 1, catalase 1), heat shock protein 70 and immune-related (interleukin-1 beta, tumor necrosis factor alpha) genes was observed. Similar results were obtained in the presence of UVB at the control temperature (28°C). The combination of both stressors revealed that low and high temperatures have additive effects on top of the damaging effects of UVB during early development in zebrafish (Aksakal and Ciltas 2018). The authors suggested that the low temperature strengthens the sensitivity of zebrafish larvae to UVB exposure for two possible reasons: (1) a reduction in the enzyme-mediated DNA damage repair mechanism that can occur due to a retardation of overall biochemical reactions at low temperatures; and (2) delayed embryo/larvae development resulting from a prolonged exposure to UVB radiation (Aksakal and Ciltas 2018). The impact of both global warming and the increase in UVB radiation levels due to stratospheric ozone depletion was evaluated in Atlantic salmon juveniles. After 8 weeks of exposure, the effects of temperature and UVB were mainly additive, and the lowest complement-dependent bacteriolytic activity, hematocrit and plasma protein levels were observed when the fish were exposed simultaneously to both stressors, suggesting an innate immune system suppression (Jokinen et al. 2011). The combination of temperature and UVB can interact synergistically to suppress metabolism and increase the susceptibility to pathogens in mosquitofish (Gambusia holbrooki). The highest infection levels by the ciliated protozoan Ichtyhophthirius multifiliis were observed in fish exposed for 10 days to the highest UVB and temperature treatments (mean infection abundance-about 55 parasites per host), compared to the the infection levels of fish exposed to both stressors separately (high UVB-about 20 parasites per host; high temperature-about 20 parasites per host; and control-about 10 parasites per host), (Cramp et al. 2014).
The role of UVR interactions with other stressors in species survival and successful habitat selection and adaptation is still poorly studied. For example, a reduction in water transparency and/or an increase in water temperature can permit the establishment of the warm-water species largemouth bass (Micropterus salmoides) in the highly transparent cold waters of Lake Tahoe, California/Nevada. Such habitat invasion results in the higher tolerance of larvae to UVR and can reduce the population size of the native redside minnow (Richardsonius egregius) through predation or competition (Tucker and Williamson 2014). The interaction between temperature and UVB can determine the movement and the microhabitat selection of zebrafish. In an open field arena with a thermal gradient (20-30°C) and under-exposure to UVB (daily dose of 1.19 kJ m -2 ) during 3 weeks, fish avoided the temperature extremes compared to the control . Groff et al. (2010) showed that DNA damage caused by UV exposure in tambaqui (Colossoma macropomum) erythrocytes can be enhanced by co-exposure to hypoxia conditions, most probably due to ROS generation under low dissolved oxygen levels in the water. Using a comet assay, the authors observed a significant increase in the damage index and damage frequency in tambaqui under normoxia and as a function of UVR exposure doses, compared to unexposed fish. However, a higher damage index and frequency in the erythrocytes was observed when the fish were co-exposed to UVR (0.504 W cm -2 UVA, 1.080 W cm -2 UVB) and hypoxia than when exposed only to the UVR (Groff et al. 2010).
UVR can enhance the toxicity of polycyclic aromatic hydrocarbons (PAHs) (Bridges et al. 2018). Photo-induced toxicity can increase the generation of ROS, and the subsequent increase of oxidative stress has been suggested as one of the mechanisms of PAH photo-toxicity in fish (Weinstein and Oris 1999). For example, the co-exposure of UVR (UVA: 1.31 W m -2 ; UVB: 0.11 W m -2 ) and anthracene (ANT) in bluegill sunfish liver microsomes resulted in oxidative stress through the increase in lipid peroxidation levels and superoxide anion production (Choi and Oris 2000). The authors measured malondialdehyde (MDA) nmoles produced in the liver microsomes exposed during 60 min to the following treatments: PAR, PAR ? ANT, UVR, and UVR ? ANT. No significant differences were observed in the produced MDA nmoles between the control (PAR) and the anthracene (PAR ? ANT) treatments. Two times more MDA nmoles were produced in the presence of UVR (approx. 500 MDA nmol) when compared to PAR, but the highest number of MDA nmoles were observed in the presence of UVR and ANT (approx. 800 MDA nmol), (Choi and Oris 2000). The hatching rate success was reduced in mahi-mahi (Coryphaena hippurus) after 7 h of embryo exposure to both natural solar radiation (UVR) and a mixture of PAHs (tPAH 50 , defined as the sum of the concentrations of 50 PAHs analytes present in the mixture), (Alloy et al. 2016). Similar hatching success percentages ([ 80%) were observed between the control (\ 10% of natural solar UVR, 0.4 lg L -1 tPAH 50 ), UVR (100% of natural solar UVR, 0.4 lg L -1 tPAH 50 ) and PAHs (\ 10% of natural solar UVR, [ 2.7 lg L -1 tPAH 50 ) treatments. Nonetheless, a decrease in more than 50% in the hatching rate success was observed in the UVR and PAHs treatment (100% of natural solar UVR, [ 2.7 lg L -1 tPAH 50 ). This decrease resulted in the delayed development in the co-exposed embryos, which may affect survival and later recruitment compared with the control embryos (Alloy et al. 2016).
The photo-induced toxicity of a mixture of PAHs was also evaluated in yellowtail kingfish (Seriola lalandi) during early development. Embryos coexposed to UVR and a mixture of PAHs showed an evident decrease in the hatching rate success when compared to those exposed to UVR or the mixture of PAHs alone. In the absence of UVR, no significant changes were observed in the hatching rate success when exposed to different concentrations of the mixture of PAHs (3.9-172 ng L -1 tPAHs), (Sweet et al. 2018). The cardiac function in the yellowtail kingfish embryos was affected by exposure to PAHs, where the exposed embryos showed an increase in the pericardial area and a higher incidence of cardiac arrhythmias and edema. However, the co-exposure of PAHs and UVR only increased the incidence of cardiac arrhythmias (Sweet et al. 2018). Synergistic effects between UVB and retene (7-isopropyl-1methylphenanthrene) were observed during early development in the whitefish (Coregonus lavaretus), as suggested by Häkkinen et al. (2003). In whitefish larvae exposed either to UVB radiation for two days (2.8 and 5.4 kJ m -2 d -1 ), or separately to different retene concentrations (10, 32 and 100 lg L -1 ), no significant mortality (max. 4%) was observed. In the control group without exposure to any UVB or retene, the larvae survival was 100%. However, when the larvae were exposed to UVB (2.8 and 5.4 kJ m -2 d -1 ) together with retene (32 and 100 lg L -1 ), more than 90% of the larvae died after exposure. Furthermore, signs of hypoxia and behavioral changes (uncontrolled spiral swimming, fish remained at the bottom of the bowl) were observed in the larvae exposed simultaneously to both stressors. No changes in behavior occurred in larvae exposed to UVB or retene alone . Severe lesions in the skin and liver fish were exhibited by the larvae co-exposed to UVB and retene. The histopathological changes observed on the skin comprised loss of membrane integrity, uplifting of the epidermis due to sloughing and vacuolization, and the appearance of necrotic cells in the epidermis showing shrunken nuclei. Interestingly, the number of neutral and acidic mucous producing cells increased by around 60% in the epidermis of the whitefish larvae exposed to UV-B and retene (32 mg L -1 ). Such an increase may be part of a protective mechanism against the retene toxicity caused by UVB. In the liver, whitefish larvae exposed to UVB and 10 or 32 mg L -1 retene showed hepatocytes containing necrotic nuclei . Gevertz et al. (2012) demonstrated that the nonnative bluegill sunfish is more susceptible than the Lake Tahoe native Lahontan redside minnow to the harmful effects of combined exposure to UVB radiation and fluoranthene (FLU). When exposed only to UVB, the native redside minnow displayed more tolerance (LD 50 , 28.0 W cm -2 h -1 UVB) to the radiation than the non-native bluegill sunfish (LD 50 , 4.6 W cm -2 h -1 UVB). Co-exposure to UVB and FLU reduced the LD 50 significantly in the native Lahontan redside minnow (LD 50 , 15.4 W cm -2 h -1 UVB). Furthermore, damage to the skin that resulted from the combination of UVR and FLU (50 ng L -1 ) was more pronounced in the non-native species as shown by the transmission electron microscopy-ultrastructural tissue examination (Gevertz and Oris 2014).
Manufactured nano-scale titanium dioxide particles (nano-TiO 2 ) are present in a broad range of products, and are usually found in personal care products including cosmetics and sunscreens. The phototoxicity of the nanomaterials in Japanese medaka was shown by exposure of the larvae to both nano-TiO 2 particles and UVA radiation. High mortality was observed in the larvae co-exposed to both stressors. Under the simultaneous exposure to both nano-TiO 2 and UVA, the photo-toxicity of nano-TiO2 increased by two powers of magnitude in the medaka larvae .

Conclusions
Exposure to ultraviolet radiation (UVA and mainly UVB) is harmful during all stages of a fish life cycle, from egg fertilization to the adult phase (Fig. 3). This current review summarizes results reported in numerous studies on fresh and seawater fish species.
Short and long-term exposure to UVR can induce damage to fish at molecular, cellular and/or tissue levels. During early development (eggs, embryos and larvae), an increase in mortality and a high incidence of developmental abnormalities are the most reported negative effects. The majority of these abnormalities include spinal/notochord deformities, enlarged pericardial sacs and the presence of blisters in the yolk. A remarkable decrease in the hatching success and a more prolonged hatching time has also been documented. Loss of normal swimming capacity and low escape performance from predators in fish larvae are the most evident behavioral changes. The skin and gills seem to be the most affected tissues in larvae exposed to UVR (Fig. 3a). Lesions include structural and functional changes in the following tissues: sunburn on the skin, changes in epidermis thickness, a decrease in number and size of mucous cells, deformation in the ionocytes structure, and damage in the epithelium tissue of the gills. In short, under UVR exposure, the following biological functions can be compromised: (1) the role of the skin as the first line of defense against pathogenic microorganisms in the surrounding water; (2) the skin and gills osmoregulatory capacity; (3) the digestive physiology and (4) the immune response.
In juveniles and adults, growth reduction and loss of body condition were described after short-or longterm UVB exposure, probably influenced by physiological and metabolic changes incurred by exposure (Fig. 3b). Behavioral changes following UVR exposure were evident in these life cycle stages and included feeding behavior, loss of appetite, restless behavior, changes in swimming activity and low predatory performance. Nevertheless, the mechanisms underlying how fish perceive UVR and how these mechanisms are triggered are poorly understood and should be further explored in the future. Several tissues/organs are negatively affected by both UVA and UVB exposure. These effects include: sunburn on skin, hyperpigmentation, a decrease in mucous production and club cells, inflammation, appearance of necrotic tissue in the epidermis, damage in corneal epithelium of the eyes, cataractous changes, higher incidence in number of parasites in the eye lenses, loss of cell integrity in the liver, lipidosis, cytoplasmic vacuolation, inflammatory lymphocytic infiltration, and high incidence of morphological malformations in red blood cells. A potential risk from UVR exposure on fish health is evidenced by the disruption in both innate and acquired immune systems, which can decrease the resistance of fish to diseases. Impairment of molecular and cellular processes was evidenced in all development stages and in different tissues (Fig. 3). Such damage is characterized by an increase in the extent of DNA damage and apoptosis and changes in the antioxidant status. Information on the overall transcriptional changes in these tissues/organs, including those involved in the immune system, is scarce. A key challenge in the future will be to establish the overall molecular mechanisms involved in the different responses of these tissues/organs to the detrimental effects of UVB and UVA exposure. Large-scale transcriptome analysis in these tissues/ organs under different cumulative doses of UVR will give new insights about the direct and indirect photochemical pathways that characterize UVB-induced damage in fish. Moreover, there is limited knowledge on the strategies used by fish to reduce the impact of UVR, and analysis of mucous proteome from different species could provide new insights on the photo-protective mechanisms.
Current evidence suggests that the destruction of stratospheric ozone, climate change and interaction with other environmental and anthropogenic stressors can lead to significant changes in underwater UVR levels. These changes may lead to more damaging effects on fish species in inland and ocean waters, which may have an impact on the fisheries and aquaculture sectors. One of the challenges in the near future will be to predict how the fish will cope with these changes, and if they will be able to adapt to future levels of UVR. There are still few studies describing how UV-detrimental effects are enhanced by interactions between UVR and temperature or pollutants; more effort should be made in light of future climate-change scenarios and the presence of both persistent and emerging contaminants. A better understanding on the harmful effects of UVR and how to reduce the impact of UVR on fish is important to mitigate ecological problems, such as predicting invasive species and their impact on native species populations. Such knowledge can also be used to improve fish aquaculture production by adjusting the fish rearing conditions in the offshore cages to minimize the damage caused by UVR, as well as formulating feed with UVAs.