Exploring the link between ultraviolet B radiation and immune function in amphibians: implications for emerging infectious diseases

Levels of ultraviolet B radiation (UVBR), a potent immunosuppressant, have increased in parts of the world due to anthropogenic thinning of the Earth’s protective ozone layer. However, little is known about how UVBR influences amphibian immune function, or if increased UVBR exposure can contribute to disease emergence in susceptible species.


Introduction
Amphibians are currently the vertebrate taxon most threatened with extinction (Houlahan et al., 2000;Hof et al., 2011;IUCN, 2016). Global amphibian numbers have undergone substantial declines over the last four decades, with between 40 and 50% of examined species listed as at least 'Near Threatened' by the International Union for the Conservation of Nature (IUCN) Red List of Threatened Species (IUCN, 2016). A major contributor to recent amphibian declines was determined following the discovery of the novel fungal pathogens Batrachochytrium dendrobatidis (Bd) in 1998 (Berger et al., 1998) and Batrachochytrium salamandrivorans (BSal) (Martel et al., 2013). Both pathogens cause the deadly amphibian disease chytridiomycosis (Berger et al., 1998;Longcore et al., 1999). Since its identification almost 20 years ago, Bd 1 has been recorded in over 500 amphibian species worldwide and has been implicated in the decline or extinction of several hundred of these (Berger et al., 1998;Lips et al., 2006;Cheng et al., 2011;Van Rooij et al., 2015). In addition to Bd and Bsal, several other diseases caused by amphibian pathogens (Ranaviruses and trematode parasites) have recently increased in prevalence. Ranaviruses have now been detected in more than 70 amphibian species in 20 countries (Miller et al., 2011;Gray and Chinchar, 2015), making them the second most common infectious agent of amphibians after Bd (Chen and Robert, 2011). Similarly, trematode infections caused by the parasite Riberioa ondatrae have increased in North American amphibians to such an extent that they have contributed significantly to recent local population declines (reviewed by Rohr et al., 2009).
An apparent rise in the number of disease-related amphibian declines has prompted questions as to why such a trend should be occurring now (Hof et al., 2011;Blaustein et al., 2012;James et al., 2015;Kolby and Daszak, 2016). Increased contact between wildlife, humans and domestic animals has undoubtedly amplified the exposure of wildlife to novel, potentially pathogenic organisms (Daszak et al., 2000). However, the consequences of a novel host-pathogen interaction depend not only on the specific physical/physiological characteristics of the host and pathogen, but also on how these are shaped by their interactions with the surrounding environment (Daszak et al., 2001;Blaustein et al., 2012). Environmental conditions not only affect pathogen growth, transmissibility and pathogenicity but also influence host behaviour, immune function and pathogen exposure regimes (Fisman, 2007). Host immune function is sensitive to a range of biotic and abiotic environmental variables like temperature (Engelsma et al., 2003;Raffel et al., 2006;Ndong et al., 2007;Barber et al., 2016), habitat quality (Cary et al., 2014;Katzenback et al., 2014;Krynak et al., 2015;Makrinos and Bowden, 2016), nutritional status (Venesky et al., 2012), competition (Groner et al., 2014) and importantly, solar UVB radiation (Kripke et al., 1992;Lahnsteiner et al., 2011;Debecker et al., 2015;Abu Bakar et al., 2016).
Ultraviolet B radiation radiation (UVBR) forms a part of the solar electromagnetic spectrum (wavelength range 280-320 nm). Although the majority of solar UVBR reaching the outer atmosphere is absorbed by stratospheric ozone, a small amount does reach the Earth's surface (van der Leun, 2004). Spatial and temporal variations in ozone thickness influence UVBR levels at the Earth's surface. Other factors, such as solar angle and proximity, cloud cover, surface reflectance (albedo) and altitude, also influence terrestrial UVBR levels (Xenopoulos and Schindler, 2001). While UVBR can penetrate into aquatic environments, surface waves, reflectivity and levels of dissolved organic carbon (DOC) significantly affect the transmission of UVBR and the depth to which it can influence aquatic systems (Xenopoulos and Schindler, 2001).
UVBR is a powerful natural stressor because it can interact with a range of biological molecules and is capable of causing extensive cellular and molecular (DNA and protein) damage (Diffey, 1991). At the organismal level, UVBR exposure can adversely impact survival, growth rates, developmental trajectories, locomotion and predator avoidance. In amphibians, embryonic and larval stages are particularly sensitive to UVBR. These life stages are at increased risk from UVBR as they are often exposed to direct sunlight, are most commonly found during spring and summer (when UV levels are highest) and/or have a limited capacity to avoid UVBR exposure . UVBR can have a variety of effects on embryonic and larval amphibians including increased mortality (e.g. Crump et al., 1999;Belden et al., 2003;Alton et al., 2010), decreased growth and impaired development (e.g. Belden and Blaustein, 2002;Calfee et al., 2006), developmental abnormalities (e.g. van Uitregt et al., 2007Romansic et al., 2009) and reduced locomotor performance and altered behaviour (e.g. Kats et al., 2000;van Uitregt et al., 2007). UVBR can also modulate the effects of other stressors (both biotic and abiotic) on amphibians (reviewed by Alton and Franklin, 2017). However, not all species respond to UVBR in the same way (Licht and Grant, 1997). Responses to UVBR can vary with geographic and seasonal exposure patterns (Peterson et al., 2002), as well as with the capacity to avoid and/or repair DNA damage (e.g. Hansen et al., 2002;Blaustein et al., 2004;Palen et al., 2005;Thurman et al., 2014).
Studies on non-amphibian taxa show UVBR is both genotoxic and a powerful modulator of immune function, with studies of fish and mammal species showing increased incidence of, and susceptibility to, disease when exposed to sublethal doses de Gruijl, 2008;Blount and Pike, 2012;Siroski et al., 2012;Ullrich and Byrne, 2012). In comparison with other vertebrate taxa, the impacts of UVBR on amphibian immune function and disease are poorly documented and the role of UVBR in disease-related mortality and decline of amphibian populations remains conjectural. There are, however, a number of lines of evidence to suggest that disease-related amphibian declines in some areas may be linked to increased UVBR exposure. For example, many Bdrelated amphibian declines have occurred in montane environments (Lips, 1998;Middleton et al., 2001;Bosch et al., 2007;Kriger and Hero, 2008;Rohr and Raffel, 2010;Walker et al., 2010) where ambient UVBR levels are significantly higher than at lower altitudes because there is less atmospheric filtration of UVBR (UV levels increase by 10-12% for every 1000 m of elevation) (Madronich et al., 1998). In addition, many declines attributable to disease (e.g. declines in eastern Australia and South America during the late 1970s and 1980s) coincide spatially and/or temporally with increases in UVBR associated with stratospheric ozone depletion (Berger et al., 1998;Lips, 1998;Lips et al., 2006). UVBR-induced immunosuppression was suggested as a potential factor contributing to amphibian declines as far  (Carey, 1993). However, the impacts of UVBR exposure on amphibian immune function and disease susceptibility have been little studied since, and much remains unknown about whether changes in global UVBR levels have contributed to recent disease-associated amphibian declines.
In the relatively small number of studies that have attempted to address this issue, results have been conflicting and/or inconclusive. For example, Garcia et al. (2006) found that exposure of juvenile (post-metamorphic) frogs of three species (Rana cascadae, Bufo boreas and Hyla regilla) to low levels of UVBR for 3 days did not increase susceptibility to Bd or subsequent mortality. Likewise, Searle et al. (2010) found no increase in mortality rates or in the infectiousness of Bd when Rana cascade larvae were exposed to UVBR and Bd simultaneously. UVBR exposure has also been linked to reduced Bd susceptibility in larvae of Bufo bufo (Ortiz-Santaliestra et al., 2011) but increased Bd susceptibility in green tree frog (Litoria caerulea) larvae (Cramp and Franklin, preliminary unpublished observations). Differences in the source of the UVBR (natural or artificial), UVBR dose, the timeframe of exposure, life history stage at the time of exposure and species influence amphibian UVBR exposure outcomes (Licht and Grant, 1997;Alton and Franklin, 2017). Experimental differences across existing studies may also contribute to the lack of consensus on the role that UVBR might play in Bd susceptibility. Although Bd epidemiology is undeniably complex, the links between chytridiomycosis and elevation/altitude, with increased prevalence/ severity of chytridiomycosis at higher altitude sites, are clear. With UVBR levels increasing with altitude (Madronich et al., 1998), it is conceivable that the greater prevalence of this and other diseases at high altitude may be related to increased UVBR exposure. Currently, our capacity to understand the relationship between UVBR and disease outcomes is significantly hampered by the lack of understanding of the fundamental way(s) in which UVBR can interact with, or modulate, amphibian immune function.
Pathways to impact: mechanisms through which UVBR can influence amphibian immune function The vertebrate immune system is a suite of complex, interrelated morphological, physiological, cellular and chemical facets that animals use to determine self from non-self (Schulenburg et al., 2009). Immune defences can be classified as innate or adaptive; innate immune defences are considered the 'first line of defence' and provide rapid, non-specific protection against a variety of potential pathogens. The innate immune system includes physical barriers like the skin which impedes the movement of microbes into the body, mucus which often contains a variety of antimicrobial substances and the hosts' own commensal microbiome with which a potential pathogen must successfully outcompete in order to become an established infection. The innate immune system also includes cell-mediated defences in the form of white blood cells and chemical cascades like complement and lysozyme. The acquired or adaptive immune system is a set of immune responses highly specific to the pathogen that induced them and includes the production of antibodies against foreign antigens and the retention of immunological memory. The functions of the vertebrate immune system can be affected by exposure to a range of environmental factors and through the effects of other (competing) physiological processes. The amphibian immune system is similar to that of most other vertebrates in that it incorporates both innate and adaptive immune pathways (Carey et al., 1999). In amphibians, as in other vertebrate taxa, the immune defences of early life (embryonic and larval) stages are less well developed than those of adult frogs. Some embryonic immune defences are, nevertheless, established very early (Rollins-Smith, 1998) with egg membranes forming a direct barrier against pathogen entry (Robert and Ohta, 2009). There is also some evidence that maternal antibodies are passed into the eggs during vitellogeneis providing additional protection during embryonic development (Poorten and Kuhn, 2009). Initial adaptive immunological competence is achieved approximately 2 weeks post-fertilization in Xenopus laevis larvae with the development of the principle lymphoid tissues, the spleen and thymus (Robert and Ohta, 2009). Larval adaptive immune function is comparatively less robust than that of adult amphibians, with larval amphibians having fewer lymphocytes and reduced antibody diversity than adults (Rollins-Smith, 2017). During metamorphosis, hormonal changes result in a sharp reduction in larval lymphocyte abundance (immunosuppression) as the immune system is reorganized from the larval type to the adult type to prevent the larval immune system from attacking the newly formed adult tissues (Rollins-Smith, 1998;Robert and Ohta, 2009). Following metamorphosis, immune function is gradually restored, however, mature immune function may not be achieved for up to a year post-metamorphosis (Rollins-Smith, 1998, 2017. Taken together, the larval and juvenile life history stages are comparatively most at risk from disease or infection because of the immaturity of the immune system. It is not surprising then that many amphibian diseases are most problematic for embryonic, larval and juvenile life stages (Gray et al., 2009;Miller et al., 2011;Langhammer et al., 2014).
Embryonic and larval amphibians are more likely to be exposed to UVBR than adult stages, although several declining anuran and urodele species (i.e. Taudactylus eungellensis and Taudactylus acutirostris, several Atelopus spp. and a number of aquatic salamander species) also have diurnal juvenile and adult stages for which UVBR exposure may be harmful. In the subsequent sections of this review, we will explore how UVBR exposure influences immune defences early in development (embryonic, larval and juvenile). We will show that UVBR can act directly on components of the innate and adaptive immune systems, as well as indirectly influencing immune function through impacts on other physiological systems or by influencing the capacity of other environmental stressors to affect immune function ( Figure 1). In doing so, we review not only studies examining the impacts of UVBR on amphibian species directly but also relevant studies on other taxa where data for amphibians is lacking. In reviewing the evidence for potential effects of UVBR on amphibian immune function, we also identify gaps in our understanding of how UVBR, through its impact on immune function, may contribute to disease processes in amphibians.

Direct modulation of immune system function by UVBR Skin integrity
The barrier function of epithelia is an integral component of the innate immune system of all animals and serves as the 'first line of defence' against potential microbial invasion. Damage to the integrity of the skin compromises the barrier role of the skin and increases the capacity for pathogens to enter the tissues and cause infection. In most vertebrates, UVBR is absorbed by the epidermis and often does not penetrate into the deeper skin layers (Biniek et al., 2012). However, larval amphibian skin is potentially highly sensitive to UVBR, since their protective pigment (melanin) layers lie largely beneath the epidermis within the dermis (Fox, 1986). In most taxa, including amphibians, acute responses to elevated UVBR can include sunburn (Blazer et al., 1997;Little and Fabacher, 2003), melanogenesis (skin darkening), lesions on the dorsal skin and eyes (McFadzen et al., 2000;Zamzow, 2004;Kazerouni and Khodabandeh, 2010), necrosis and sloughing of the skin (Noceda et al., 1997;Abedi et al., 2015), a reduction in epidermal strength (Biniek et al., 2012), and epidermal hyperplasia and oedema (Matsumura and Ananthaswamy, 2004). Larvae of two frog species, H. regilla and Rana aurora, experienced both skin damage and lens opacity following exposure to elevated levels of artificial UVBR (Flamarique et al., 2000). Subsequent fungal infections in areas of UV-related skin damage have been reported (Flamarique et al., 2000;Little and Fabacher, 2003), indicating that UVBR exposure can compromise the immunological barrier function of amphibian skin. Whether these results apply to natural exposure regimes is presently unclear as the aforementioned studies were undertaken under laboratory conditions using artificial light sources.
While exposure to elevated UVBR can sometimes lead to obvious outward signs of skin damage, sub-erythemal exposures can also lead to levels of barrier dysfunction that can  Figure 1: Schematic representation of the potential direct and indirect pathways through which UVBR can influence immune function in amphibian early life stages. UVBR can directly kill and damage cells in the outer skin layers, disrupting the physical integrity of the skin, compromising the function of cutaneous dendritic cells (D) and leucocytes (L), and triggering exposed keratinocytes (K) to release a cascade of immunosuppressive molecules that inhibit innate and adaptive immune functions of the systemic immune system. UVBR can disrupt the innate immune function of the cutaneous 'secretome' (mucus, antimicrobial peptides, complement, lysozyme, etc.) and influence the composition of the host microbiome. Indirectly, UVBR may influence immune function via its impact on interrelated physiological systems. For example, UVBR can affect immune function by disrupting energy production and/or distribution pathways, by influencing neuroendocrine signalling pathways controlling immune system maturation or inducing a physiological stress response (involving glucocorticoids). UVRB-associated damage to DNA and other biomolecules may also have a lasting impact on immune function by influencing gene expression patterns in subsequent life stages. Image: Cameron Baker©.

4
precede infection by secondary fungal pathogens (Nowak, 1999). Since sub-erythemal skin damage is difficult to determine macroscopically, it may go unnoticed in wild amphibian studies where levels of embryonic and larval mortality are naturally very high anyway. If, like in other taxa, suberythemal skin damage increases the risk of pathogen infections in early amphibian life stages, further work is needed to better understand how UVBR exposure regimes influence infection outcomes. Moreover, exposure 'thresholds' are likely to vary across species and populations with the capacity of animals to behaviourally and/or morphologically limit their exposure to, or tolerate, UVBR in their environment. Similarly, since UVBR levels vary temporally and spatially, exposure regimes (fluctuating vs acute exposures) may also affect skin responses within and across life history stages. We, therefore, need to examine the possibility that early UVBR-associated skin damage may have long-term consequences for cutaneous immune function in later life history stages and appreciate that the acute and long-term effects of UVBR on skin function are likely to vary both between and within species.

Skin secretions
Skin secretions form a vital component of the innate immune system of all animals. Secretions function as both a physical barrier to pathogens, by trapping and preventing their establishment on the skin surface, and as a chemical barrier, containing a variety of antimicrobial proteins and peptides like proteases, lectins and lysozyme, which actively kill microbes on contact (Ellis, 2001;Zasloff, 2002;Magnadottir, 2006). Embryonic and larval amphibians have several types mucusand peptide-secreting cells in the integument (Altig and McDiarmid, 1999) that produce substances with well-defined immunoprotective functions (Woodhams et al., 2016). Damage to this layer can directly increase the risk of pathogen infection (Kanno et al., 1989;Magarinos et al., 1995;Zamzow and Siebeck, 2006). In some reef fish, cutaneous mucus also (indirectly) influences the innate immune defences (i.e. the barrier function of skin) by reducing the penetrance of UVBR through to the more UVBR-sensitive skin layers (Zamzow and Losey, 2002;Zamzow, 2004;Zamzow and Siebeck, 2006;Eckes et al., 2008); whether larval amphibian mucus has a similar photoprotective function is largely unknown. Notably, UVBR exposure can reduce the abundance and distribution of mucus-secreting cells in fish larvae (Kaweewat and Hofer, 1997;McFadzen et al., 2000;Kazerouni and Khodabandeh, 2010;Abedi et al., 2015). It is currently uncertain if UVBR exposure can influence mucus production in amphibians. The loss of mucus-producing cells may be particularly problematic for embryonic and larval animals whose adaptive immune systems are underdeveloped relative to adult frogs (Rollins-Smith, 1998).
Non-mucus skin secretions are an important component of the innate immune system of larval, juvenile and adult amphibians (Woodhams et al., 2007(Woodhams et al., , 2016. Antimicrobial peptides (AMPs) have been shown to be a particularly important constituent of non-mucus skin secretions, since many are effective against cutaneous pathogens like Bd (Rollins-Smith, 2009). Indeed, the diversity and abundance of cutaneous AMPs is thought to be a major factor contributing to the differential susceptibility of frog species to Bd (Rollins-Smith, 2009). Abiotic stressors such as pollutants and low environmental temperatures can differentially influence the production, composition and release of AMPs from skin glands (Davidson et al., 2007). However, little is known about the capacity for UVBR exposure to directly influence the production and/or composition of cutaneous AMP in larval amphibians or for early UVBR exposures to affect AMP production or composition in subsequent life history stages.
As far as we could determine, no studies have explored the hypothesis that early UVBR exposure can disrupt the physiochemical properties of embryonic and larval amphibian skin. Laboratory studies are required to establish a cause-and-effect relationship between UVBR exposure and skin secretion properties. If such a relationship can be established, it will be important to ascertain how the timing, dose and intensity of UVBR exposure might affect the subsequent composition of the cutaneous 'secretome' or influence the subsequent development of cutaneous mucus and glandular tissues.

The cutaneous microbiome
The microbiome of the amphibian integument plays an important role in host immunity and disease progression Rebollar et al., 2016). As with all organisms, the 'external' surfaces of amphibians are colonized by a large and diverse array of microbiota (bacteria, fungi, viruses and mites), most of which are commensal or transient and cause no harm at all to the host. A subset of these microbes produce metabolites that actively discourage establishment of potential pathogens, thereby contributing to disease resistance (Patra et al., 2016). The host's microbiome also provides a highly competitive microenvironment in which a potential pathogen must outcompete the commensal microbiota in order to become established. The composition of the host microbiome can determine the outcome of infection by pathogens, with recent studies showing variation in host susceptibility to Bd is linked to the presence/absence of certain bacteria on the host's integument (Briggs et al., 2010;Jani et al., 2017).
Given the links between Bd infection outcomes and microbiome composition, there is considerable interest in understanding those factors that may regulate or disturb amphibian microbiomes. Like all ecosystems, host-associated microbiomes are sensitive to environmental disturbances. Aquatic pH, salinity and temperature have been identified as primary regulators of host-and non-host-associated microbial communities (Fierer and Jackson, 2006;Lozupone and Knight, 2007;Costello et al., 2009;Kueneman et al., 2014). Solar UVBR has well-documented antimicrobial properties, capable of affecting both commensal and pathogenic organisms (Faergemann and Larko, 1987;Dotterud et al., 2008;Wang et al., 2012). Recent work in humans suggests that UVB radiation can modulate aspects of the host microbiomes more broadly, and in doing so, influence disease development and outcomes (Patra et al., 2016). However, the role that UVBR may play in the regulation of innate immune function in amphibians, through its impact on host microbiome composition remains largely unknown. In the only study to date that has addressed this question in amphibians, pond shading had no impact on larval microbiome composition in Rana catesbeiana (Krynak et al., 2015), suggesting that UVBR may not influence the structure of the microbiome in some aquatic amphibians. However, since UVBR penetrance into aquatic systems is dependent on the amount of dissolved organic matter in the water, pond shading levels and solar angle (Xenopoulos and Schindler, 2001), further work is needed to determine if aquatic microbiomes can actually be influenced by natural levels of UVBR in other environs where UVBR penetrance is higher (e.g. at high altitude, clear stream sites).

Cutaneous and systemic immune signalling
The cutaneous immune system comprises interconnected innate and adaptive cellular networks involved in localized and systemic immune responses to potential pathogens (Mann et al., 2012). In vertebrate skin, keratinocytes in the epidermis produce a complex array of AMPs, pro-inflammatory cytokines and chemokines in response to the activation of pathogen recognition receptors on their cells' surfaces (Mann et al., 2012). In addition to these cells, antigen-presenting dendritic cells (Langerhans cells) actively patrol the integument for would-be pathogens. These cells perform a vital role bridging the cutaneous innate and adaptive immune systems: they can recognize a diverse array of pathogen-and self-associated molecules and can initiate the appropriate immune or tolerogenic responses (Sleijffers et al., 2004;Mann et al., 2012).
The effects of UVBR exposure on cutaneous immune signalling in amphibians is largely unexplored; however, data from mammalian models show that UVBR exposure can induce local and systemic immunosuppression by reducing cutaneous dentritic cell abundance, and by reducing their capacity to recognize antigens and induce an adaptive immune response (Kripke, 2013). In addition, UVBR exposure induces the secretion of immunosuppressive cytokines (including interleukin-10) by keratinocytes, which suppresses other local and systemic immune responses (Elmets et al., 1983;Beissert and Schwarz, 1999) and impedes DNA damage repair mechanisms (Sreevidya et al., 2008). While data relating to non-mammalian taxa are limited, studies on fish suggest a common action of UVBR on elements of the systemic immune system. In some fish species, UVBR exposure correlates with a reduction in the abundance of circulating peripheral lymphocytes (Jokinen et al., , 2001Markkula et al., 2005Markkula et al., , 2007Markkula et al., , 2009, with a reduced functional capacity of white blood cells to respond to stimulation (Salo et al., 1998;Markkula et al., 2009) and with lower blood immunoglobulin levels (Jokinen et al., 2001(Jokinen et al., , 2008(Jokinen et al., , 2011Markkula et al., 2005).
Importantly, local and systemic immunomodulation by UVBR exposure has been directly linked to a wide range of mammalian pathogen infection outcomes (reviewed by Sleijffers et al., 2004). Similarly in fish, UVBR can negatively affect pathogen infection rates even when UVBR doses are not sufficient to cause overt skin damage (Markkula et al., 2007;Cramp et al., 2014). Moreover, the effects of UVBR on infection susceptibility can manifest after just a single exposure and can persist for several weeks (Jeevan and Kripke, 1989). The effects of UVBR on cutaneous immune function and systemic signalling in amphibians (larval or adult) are largely unknown. However, given the similarly of the amphibian immune system to that of other vertebrates (Carey et al., 1999), it is not unreasonable to expect that similar responses to UVBR might occur.
In amphibians, thyroid hormones (THs), along with glucocorticoids, play a major role in the timing and control of tissue development and metamorphosis (Denver, 2009;Croteau et al., 2010). Changes in the levels of these hormones in response to environmental stressors can, therefore, influence growth rates, tissue development and metamorphic timing (Denver, 2009). UVBR exposure has been shown to impede larval development and reduce deiodinase 2 expression in some tissues of Rana pipiens tadpoles and (Croteau et al., 2009) suggesting that UVBR exposure may impair development via its influence on the thyroid axis. Environmental factors that influence neurohormonal pathways are likely to influence a number of physiological systems, including the transition from the larval-to adult-type immune system (Rollins-Smith and Smits, 2004). Indeed, maturation of the larval immune system is highly dependent on both TH and glucocorticoid levels (Rollins-Smith and Blair, 1990;Rollins-Smith, 1998). Conceivably then, larval UVBR exposure may disrupt both the glucocorticoid-and TH-signalling pathways to impair immune system development. While this question is yet to be explicitly explored, Ceccato et al. (2016) showed that juvenile frogs reared as larvae under moderate levels of UVBR, had lower levels of circulating leucocytes and reduced antigen swelling responses suggesting a delayed transition from larval to adult phenotype. Recently metamorphosed frogs are at greater risk from pathogens because of the relative immaturity of the immune system (Rollins-Smith, 2017), so there is potential for early UVBR exposure to compound this risk by delaying the development of the adult-type immune system. The role that the glucocorticoid or TH axes played in these responses remains to be determined.

Energy production and supply to the immune system
Immune responses and resistance to pathogens and parasites are physiologically demanding (Demas et al., 1997;Martin et al., 2003;Hawley and Altizer, 2011). Energetic costs associated with UVBR exposure (including UVBR damage repair or avoidance mechanisms) may result in trade-offs, reducing an animal's capacity to mount an effective immune response. Consistent with this view, studies of amphibian larvae have shown that energetic costs associated with repair or avoidance of UVBR damage may be significant, compromising larval growth and foraging opportunities (van de Mortel and Buttemer, 1998;Alton et al., 2010Alton et al., , 2011. Direct UVBRinduced damage to biomolecules within mitochondria (e.g. ROS damage) may also compromise energy production (Cramp et al., 2014). Studies of amphibian larvae have also shown that increased cutaneous melanin production in response to UVBR (Bancroft, 2007) may limit the amount of energy available for growth and, potentially, immune function as well-a view supported by studies of damselflies showing a trade-off between melanin production and immune function (Debecker et al., 2015). Whether this trade-off in energy investment influences the outcome of a pathogen challenge, however, remains to be determined.

Carryover effects from early life experiences
The effects of UVBR on amphibian immune function may not manifest immediately, which may mask the potential risk from this stressor. Juvenile amphibian immune function can be significantly influenced by environmental stressors experienced during the larval period, including pond drying and dietary stress (Gervasi and Foufopoulos, 2008;Venesky et al., 2012;Krynak et al., 2015); however, relatively little is known about how early UVBR exposure might influence long-term immune function in amphibians. UVBR has been linked to long-term effects on immune function though direct DNA damage (mutations) as well as through epigenetic processes in mammals (e.g. Gronniger et al., 2010;Shen et al., 2017). For instance, the development of skin cancer in humans follows UVBR-induced immunosuppression that can occur many decades after the causative exposure (Elmets et al., 2014). Early development is a particularly vulnerable stage for mutagenic and epigenetic modifications, since it is a period of rapid cell replication (Perera and Herbstman, 2011). Given that early life stages are most likely to experience UVBR, our group has been exploring the capacity for early UVBR exposure to have a delayed effect on immune function in amphibians. We have found that juvenile frogs, exposed to elevated UVBR as larvae, have reduced antigen responses, reduced white blood cell counts (Ceccato et al., 2016) and are more susceptible to Bd than those exposed to low/no UVBR as larvae. These findings indicate that early developmental exposure to UVBR has long-term ramifications for amphibian immune function and disease susceptibility. Further work is required to understand the mechanistic basis for the latent effects of UVBR exposure on amphibians, and how UVBR dose, intensity and the timing of exposure can influence latent immunosuppression.

Conclusions and future directions
In this review, we have explored the potential for one changing environmental variable, UVBR, to influence amphibian fitness through its capacity to exert direct and indirect effects on immune function. While relatively little is known about how UVBR can influence amphibian immune function, a large body of literature from studies of other taxa strongly supports the idea that UVBR exposure may increase the susceptibility of amphibians to disease. The potential immunosuppressive effects of UVBR, however, have been largely overlooked in studies investigating the cause(s) of diseaserelated amphibian population declines. Understanding the role that elevated UVBR levels may have played and may continue to play, in the emergence or exacerbation of amphibian diseases is, therefore, important.
Anthropogenic increases in solar UVBR exposure correlate positively with patterns of enigmatic amphibian decline in many parts of the world (Kiesecker et al., 2001;Middleton et al., 2001). While a novel pathogen (Bd) was subsequently shown to be the proximate cause of many of these declines, the possibility that increased UVBR may have contributed to the emergence and epidemiology of the disease has remained largely unexplored. Elevated UVBR has been largely neglected from major environmental models of Bd-related declines because of a lack of data on natural UVBR levels in the amphibian-pathogen microenvironment. UVBR levels are difficult to accurately quantify and can vary enormously, over both space and time, as a consequence of solar distance and angle, vegetation cover, cloud thickness and levels of dissolved organic matter in water (Xenopoulos and Schindler, 2001). UVBR levels also co-vary with other changing environmental factors such as temperature, cloud cover and rainfall patterns. Temperature and rainfall significantly influence amphibian behaviour as well as incident UVBR levels, so disentangling the potential negative effects of increased UVBR exposure from the effects of other, covarying environmental factors, has been, and remains, a complex challenge.
Satellite-based monitoring of global ozone levels and incident UVBR provide a relatively coarse measure of exposure risk and can be used to monitor changes in UVBR levels. Satellite-based UVR data have been used in a small number of studies to model Bd infection patterns over a relatively broad spatial scale (the Iberian Peninsula) (Walker et al., 2010;Ortiz-Santaliestra et al., 2011). While the data indicated that Bd presence in the study populations was not predicted by any one environmental factor, the conditional prevalence of infection was weakly, negatively correlated with solar radiation levels (Walker et al., 2010). However, the authors stress caution in the interpretation of these findings, mainly because of the incongruity of scale between the coarse climatic statistics used in these studies and the actual microenvironments occupied by both the hosts and the pathogen (Walker et al., 2010). In order to understand if/ how UVBR may impact amphibian health and disease processes and to mitigate potential risks for highly threatened species or insurance populations, we need a better understanding of how amphibians interact with UVBR in their environment. Future studies need to consider realistic measures of UVBR in the immediate environment where animals are most likely to be exposed and consider both spatial and temporal variation in UVBR exposure patterns. Likewise, further work is needed to understand how amphibians behaviourally influence individual exposure histories through their choice of oviposition site (e.g. Palen et al., 2005;Thurman et al., 2014) and/or through avoidance or thermoregulatory activities.
Following on from this, the question of 'how much UVBR is too much' is key to assessing whether anthropogenic increases in UVBR have contributed to some diseaseassociated amphibian declines. Amphibians, as a group, occupy a diverse array of environments and may experience a wide range of UVBR levels: how much UVBR is too much is likely to be highly species, population and life-stage specific. For instance, species or populations naturally living in high altitude, open canopy and/or low DOC environments typically experience higher levels of UVBR than fossorial, or fully nocturnal species. Moreover, exposure to UVBR only becomes problematic when an animal's capacity to respond to, or repair, induced damage is exceeded. While amphibians employ a range of defences against UVBR, their capacity to invoke these can differ enormously across species, populations and life history stages (van de Mortel and Buttemer, 1998;Belden et al., 2003;Palen et al., 2005)

8
avoidance strategies are ineffective is likely to be highly context (species, life history stage, population, etc.) specific. Future work is needed to assess whether UVBR exposure thresholds exist for immunosuppression. Thresholds for UVBR-associated immunosuppression are likely to be different for embryos vs larvae and adults and differ across species and/or populations. Understanding the factors that influence the performance of morphological and molecular defence mechanisms, and the plasticity within repair and defence mechanisms to contend with diverse and variable UVBR levels is needed. Likewise, the capacity for larvae to detect and behaviourally modulate their exposure to UVBR needs to be more widely explored.
Ecoimmunological studies often measure immune system responses to environmental stimuli as a proxy for individual fitness (Downs and Stewart, 2014). However, increasingly, studies that employ single immune assays, like white blood cell counts or lectin-induced swelling assays, have been criticized for being overly simplistic or overinterpreted due to the challenge of linking immune measures to actual pathogen susceptibility, resistance or recovery (Hawley and Altizer, 2011). There is also often a lack of correlation among immune assays within single individuals or species (Hawley and Altizer, 2011). Attempts to link host susceptibility to pathogens with responses to non-pathogenic immune assays are hampered by a lack of understanding about which immune parameters will best predict pathogen susceptibility. Moreover, different pathogens may differentially interact with the immune system, and so markers that may predict outcomes for one pathogen, may bear no relevance to infection outcomes for another pathogen. Consequently, it is important that future studies of UVBR on amphibian immune function examine a suite of traits that reflect the diverse ways that the immune system may to respond to a challenge, both within and across life stages, including by directly measuring pathogen susceptibility, infection intensity and immune responses to antigenic challenges.
Immune function is a key individual-level trait that influences populations because it directly affects the survival outcome of a pathogen challenge (Downs and Stewart, 2014). However, within a community, variability in host immune function can promote selection and the evolution of resistance to infection (Lazzaro and Little, 2009). Variability in the responses of immune system components to the environment can promote polymorphism within a population; environmental heterogeneity can then maintain genetic variation in immunity by promoting alternative phenotypes over space and time (Lazzaro and Little, 2009). Given that UVBR exposure history likely differs significantly within and between populations, UVBR may differentially influence selection pressure on amphibian immune function genes, which in turn may shape differences in disease prevalence and infection outcome. Relatively fast-paced changes in UVBR levels may have placed additional pressure on populations, as the rate and magnitude of change could have exceeded the capacity for evolutionary mechanisms to maintain genetic heterogeneity in immune traits. This may have rendered some amphibian populations or species more susceptible to novel pathogens and contributed to the catastrophic declines that followed. Analysis of the way that UVBR can shape genetic heterogeneity within and between species and populations would inform our understanding of how UVBR may influence the evolution of disease resistant or susceptible phenotypes.
Historically, the primary role of the immune system has been regarded as the detection and elimination of pathogens (i.e. 'resistance'). While in this review we have focused on the potential for UVBR to disrupt amphibian resistance pathways, tolerance of pathogens is increasingly recognised as an important immune strategy for managing host-pathogen interactions (Medzhitov et al., 2012). Tolerance can be defined as responses that limit the negative effects of the infection on the host without influencing the pathogen burden. Resistance-based immune responses are energetically costly and often damaging to the host; tolerance of a pathogen may allow hosts to maximize their fitness by preventing costly immunopathological responses (Medzhitov et al., 2012). The concept of tolerance is relatively new to ecoimmunology and the mechanisms underpinning it are poorly understood. Conceivably though, UVBR exposure may shape disease dynamics within populations by reducing the capacity of hosts to tolerate otherwise manageable microbial interactions. Conversely, UVBR has been shown to impair some resistance pathways, increasing pathogen/allergen tolerance, termed 'photo-tolerance' (Spellman et al., 1984;de Gruijl, 2008). While pathogen tolerance may be an effective strategy to minimize the negative effects of a pathogen on an individual, tolerant hosts can also serve as highly infectious vectors of disease (Medzhitov et al., 2012). This has potentially significant implications for the spread of pathogens within and between populations. Several amphibian species, including the American bullfrog (Lithobates catesbeianus) and the clicking froglet (Crinnia signifera), have been identified as highly virulent vectors of Bd capable of harbouring and spreading the fungus to more susceptible species in their environments (Schloegel et al., 2012;Brannelly et al., 2018). Further work is required to determine if UVBR exposure regimes can influence pathogen tolerance in amphibians and if this may contribute to inter and intraspecific differences in susceptibility to pathogens like Bd.
Global UVBR levels have increased by between 2 and 6% since the middle of the last century (Lemus-Deschamps and Makin, 2012) and are expected to remain significantly elevated for much of this century (Ball et al., 2018;Montzka et al., 2018). Elevated UVBR levels correlate broadly with patterns of disease-related amphibian biodiversity loss, particularly at high altitude (Kiesecker et al., 2001;Middleton et al., 2001), yet potential causal links between UVBR exposure and amphibian disease remain relatively unexplored. and likely affects amphibians in the same way. However, further work is needed to establish the mechanistic link between UVBR, immune function and disease susceptibility in amphibians. By taking a physiological approach to understanding the mechanisms through which the environment shapes amphibian immune function, we can better predict how environmental change is likely to influence disease dynamics with and between species and populations. Physiological studies are uniquely positioned in this regard, because environmental effects on physiology can be experimentally tested. Therefore, unlike correlative studies, empirical physiological studies can provide the compelling evidence needed to establish direct cause-and-effect (Carey, 2005;Cooke et al., 2013). The diversity of ways in which the environment can shape population and species persistence via impacts on individual physiology, highlights the importance of taking a holistic and integrative approach to address ecoimmunological problems in a rapidly changing world. We have identified a number of direct and indirect pathways through which elevated UVBR could shape individual immune function and influence disease processes in amphibians, which we hope may guide future research in this field. Currently, our understanding of how UVBR and disease influence amphibians is restricted to a relatively small number of 'model' species. This represents a critical knowledge gap which may bias our responses to population declines in under-represented species. There remains an urgent need for fundamental data on how the environment shapes amphibian immune responses across orders, species, populations and life stages, and how this influences their capacity to resist pathogen infections.