Critical review of the phytohemagglutinin assay for assessing amphibian immunity Authors: Lauren Hawley, Kelly L Smalling, Scott Glaberman Published: 2023 ABSTRACT Infectious diseases are contributing to global amphibian declines. Scientists need better immune tests to identify vulnerable amphibian populations. This study reviews one such test in amphibians called the phytohemagglutinin (PHA) assay. The authors identify knowledge gaps and recommend standardized procedures to enhance the PHA assay’s effectiveness for conservation efforts. Introduction In an era of unprecedented biodiversity loss, the ongoing global decline of amphibians is among the most severe (Bishop et al., 2012). Forty percent of amphibian species are threatened and 70% are in decline—a downward trend that is expected to continue in the coming decades (Collins and Halliday, 2005; Hayes et al., 2010; Bolochio et al., 2020). Infectious disease has been repeatedly identified as a primary stressor (Vredenburg et al., 2010; Bienentreu and Lesbarrères, 2020). There are a variety of pathogens in amphibians, including bacteria such as Aeromonas hydrophila, viruses such as Ranavirus, fungi such as Batrachochytrium dendrobatidis (Bd), and macroparasites such as Ribeiroia spp., that have impacted amphibian populations (Mauel et al., 2002; Daszak et al., 2003). Certain pathogens, notably Bd and Ranavirus, have been directly implicated in the mortality, decline, extirpation, and even extinction of amphibian species (Fernández-Benéitez et al., 2008; Vredenburg et al., 2010; Miller et al., 2011; Scheele et al., 2019). Chytridiomycosis represents what some experts believe is the greatest disease-driven extinction event in vertebrate history (Skerratt et al., 2007; Vredenburg et al., 2010). Previous research indicates that amphibians vary in their ability to defend against pathogens (Knapp et al., 2016; Lips, 2016; Scheele et al., 2019). Genetics, body size and microbiota all play roles in disease susceptibility (Lips, 2016). At the same time, external factors, such as climate change (Rohr and Raffel 2010; Brunner et al., 2015), habitat fragmentation (Greer and Collins, 2008) and pollution (Forson and Storfer, 2006; Brown et al., 2021), impact the immune response to disease in complex ways (Hayes et al., 2010; North et al., 2015; Rollins-Smith, 2017; Ohmer et al., 2020). For instance, a study by Rohr and Raffel (2010) suggests that increased climatic variability can suppress amphibian immune systems as they ‘lag behind’ rapid changes in temperature, and may simultaneously increase pathogen abundance by inducing Bd spore release. Habitat fragmentation, meanwhile, may alter disease dynamics via changes in host density (Greer and Collins, 2008), limitations to host gene flow and stress caused by factors such as predator density (North et al., 2015). Contaminants may cause direct immune suppression via disruption of genetic and chemical processes and direct destruction of cell membranes (Hayes et al., 2010), or indirect immunosuppression by causing changes in behavior and development, which increase the probability of exposure to pathogens (Rohr et al., 2013) or reducing energy available for the immune response (Rohr et al., 2008; Langerveld et al., 2009). The relationship between environmental factors, immune system function and disease outcomes—whether an infection develops, the severity of infection and mortality rate—is vital to understanding and managing amphibian declines (Lips, 2016; Scheele et al., 2019). Therefore, the ability to assess immunocompetence is critical for determining which individuals or populations are most at risk from disease and other environmental stressors (Young et al., 2001; Collins, 2010; Blaustein et al., 2012). For the purposes of this paper, immunocompetence refers to the ability of an individual’s immune system to respond to and defend against pathogens or other foreign substances. Martin et al. (2006) discusses the different components of the immune system and the importance of considering the broader ecological context when studying immune function. They also suggest that measuring a set of immune characteristics can provide a more nuanced understanding of an individual’s immunocompetence, since no test can evaluate every aspect of a vertebrate immune system. While measuring amphibian immunity is essential for identifying at-risk populations, there are few immunological tools available for this purpose (Clulow et al., 2015). Most methods, including leukocyte counts, complement protein quantification tests, bacterial-killing assays and immunoglobulin levels, require collecting blood samples (Christin et al., 2003; Demas et al., 2011; Priyadarshani et al., 2015; Goetz et al., 2017; Rodriguez and Voyles, 2020), which is technically challenging and can be harmful to amphibians due to their small size and limited venipuncture sites (Heatley and Johnson, 2009). Cardiac puncture is often used to obtain adequate blood volume, despite substantial risk to the animals (Heatley and Johnson, 2009). Other obstacles to applying most immune testing tools to amphibians include the need for rapid transportation of samples to the lab for processing (e.g. macrophage phagocytic ability assays), expensive equipment (e.g. natural killer cell toxicity assays), species-specific reagents (e.g. cytokine profiles), sterile conditions (e.g. lymphocyte proliferation assays) and/or extensive training (e.g. anti-microbial peptide analysis) (Christin et al., 2003; Mangoni, 2006; Demas et al., 2011; Clulow et al., 2015; Goetz et al., 2017). In addition to a lack of appropriate immunological tools, there are also few well-studied amphibian species (Carey et al., 1999). Several species such as Xenopus laevis, X. tropicalis, Rana pipiens and Ambystoma mexicanum have been extensively used as immunological models, mainly because they are easy to breed and/or maintain in the laboratory (Womack et al., 2022). However, these model species only represent a small fraction of the 53 described amphibian families, and do not necessarily represent the full range of amphibian immune system structure or function (Rollins-Smith and Woodhams, 2011; Levy and Heald, 2016; Lips, 2016; Bolochio et al., 2020). The phytohemagglutinin (PHA) assay is a method for assessing the immune response that avoids the limitations of other immune assays, as it is affordable, technically straightforward and does not require a blood sample (Martin et al., 2006). PHA is an extract derived from red kidney beans (Phaseolus vulgaris), which stimulates the innate immune system through inflammation and cell-mediated adaptive immune response with the recruitment of T cells and other leukocytes (Martin et al., 2006; Salaberria et al., 2013). This process induces a swelling response that is measured and converted into an index of immunocompetence (Tella et al., 2008; Al-Khalaifah and Al-Nasser, 2020). While the PHA assay is primarily considered a test of adaptive immunity, cells associated with the innate immune system also play a vital role in swelling, particularly in the initial response, and antigen-presenting cells are not needed to stimulate the T-cell response (Martin et al., 2006; Vinkler et al., 2010). Thus, the PHA assay provides a complex demonstration of both innate and adaptive immune responses, making it a particularly valuable test for evaluating immunocompetence (Tella et al., 2008; Vinkler et al., 2010). The PHA assay is affordable, relatively non-invasive and has been used extensively over the last 5 decades (Martin et al., 2006; Tella et al., 2008). Studies in birds have associated PHA responses with a range of other measures, including contaminant levels, noise pollution and coloration (Grasman et al., 1996; Smits and Williams, 1999; Møller and Cassey, 2004; Møller and Saino, 2004; Moreno et al., 2005; Grasman, 2010; Sivaraman and Kumar, 2013; Lopez-Antia et al., 2015; Kerimov et al., 2018; Pusch and Navara, 2018; Arct et al., 2019; Minias et al., 2019; Obomsawin et al., 2021), and have assessed the impact of immunocompetence on health outcomes (Mougeot et al., 2004; Moreno et al., 2005; Sivaraman and Kumar, 2013; Martínez et al., 2018). While the PHA assay is still often used for avian studies (O’Connor et al., 2014; LaMonica et al., 2021), it has also grown increasingly popular among other taxa, including mammals, reptiles and fish (Finger et al., 2015; Hernández-Arciga et al., 2018; Merlo et al., 2018; Plasman et al., 2019; Nasri et al., 2020; LaMonica et al., 2021; Stienbarger et al., 2021; Miyazawa et al., 2022). The use of PHA assay in amphibians is relatively recent, with substantially fewer studies compared with birds. Published amphibian studies vary widely in terms of species, methodology (e.g. dosage), experimental design (e.g. controls) and results (e.g. swelling response), making it challenging to develop consistent best practices across a range of amphibian species. Moreover, most amphibian PHA studies have been conducted under controlled laboratory conditions, and no formal protocol has been developed for wild populations. To date, there has never been a critical review of PHA testing in amphibians to evaluate previous studies and highlight future needs. Our review fills this gap with three main objectives: (1) capture variation in both organismal and experimental aspects of published amphibian PHA studies; (2) identify uncertainties or gaps in the overall literature that limit the use of the PHA assay in amphibians; (3) propose future research directions to fill these gaps, with the goal of utilizing the PHA assay with amphibians under field conditions. It is our hope that this review will improve the use of the PHA assay for assessing the effects of disease and co-stressors on amphibian populations, which will benefit amphibian conservation. Literature Search & Data Collection All amphibian PHA studies were compiled through a systematic literature search using the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement (Moher et al., 2009; Page et al., 2021). PRISMA provides a set of criteria for carrying out and reporting the outcome of systematic reviews (Sarkis-Onofre et al., 2021). Full details of the literature search process and keywords using PRISMA are provided (Supplementary Fig. S1). After capturing all known amphibian PHA studies, we extracted the following types of data from each study for further review: author(s), publication year, study aim, species, source of specimens, sample size, inoculation, control design, injection site, PHA concentration and injection volume, swelling response, and experimental time points (Table 1). We focused on these data types because they allow comparison of methodologies across studies and species. Table 1Summary of amphibian PHA assay studiesStudySpeciesAverage adult weighta (g)Average adult SVLa (mm)Life stageInjection siteConc. (mg/ml)Injected volume (ml)PHA per injection (g)Times measuredInoculationPHA control typeSpecimen sourceSex tested Abu Bakar et al., 2016 L. aurea 5093.9Juvenile, adultSubcutaneous anterior lateral surface lower leg below knee10.040.0435 hNoBilateralCaptive bredF, M Bakewell et al., 2021 A. punctatus 6076AdultSubcutaneous anterior lateral surface lower leg below knee20.050.124 hNoBilateralField collected from a large areaF, M Brannelly et al., 2019 L. pipiens 39.2165JuvenileHind thigh250.020.512, 24, 48 hYesBilateralEgg masses from fieldF, M Brown and Shine, 2014 R. marina 106.25241AdultToe web between 2nd and 3rd digits in hind foot20.050.16, 12, 24, 48, 72 hNoBilateralField collected from 1 siteF, M Brown et al., 2011 R. marina 106.25241AdultToe web between 2nd and 3rd digits in hind foot2.670.050.136, 12, 24, 48, 72 hSubsetBilateralField collected from 1 siteF, M Brown et al., 2015 R. marina 106.25241AdultToe web between 2nd and 3rd digits in hind foot20.050.124, 48, 72 hNoBilateralField collected from 1 siteF, M Cassettari et al., 2022 Rhinella granulosa —90AdultSubcutaneous hind fleshy base of foot200.010.26, 12, 24 hNoBilateralField collected from 1 siteF, M Ceccato et al., 2016 L. peronii 12.1772.8Larvae, juvenileRight side of tail OR hind thigh1200.0524, 48 hNoNoneEgg masses from fieldF, M Clulow et al., 2015 L. aurea 5093.9Juvenile, adultSubcutaneous anterior lateral surface lower leg below knee10.040.046, 12, 24, 48 hNoBilateralVaried field collected and bred in captivityF, M L. revelata 0.9335No L. peronii 12.1772.8No L. tasmaniensis 3.3750.5No de Assis et al., 2015 R. icterica 120190AdultSubcutaneous hind fleshy base of foot200.010.212 hNoBilateralField collected from 1 siteM Desprat et al., 2015 H. arborea 9.550AdultRight hind leg250.020.518 hNoBilateralField collected from 1 siteM Desprat et al., 2017 H. arborea 9.550AdultLeg250.020.518 hNoBothField collected from 1 siteM Fites et al., 2014 X. laevis 130147AdultIntramuscular middle plantar side hind foot20.10.224, 48 hYesBilateralPurchased from vendorF Gervasi and Foufopoulos, 2008 Lithobates sylvatica 17.683JuvenileSubcutaneous ventral hind thigh10.050.0525, 48 hNoNoneEgg masses from fieldF, M Gilbertson et al., 2003 L. pipiens 39.2165AdultMiddle toe hind foot where webbing ends20.060.1324, 48, 72 hYesBilateralField collected from different sitesF, M Iglesias-Carrasco et al., 2018 P. perezi 31.886AdultHind footpad20.010.0224 hNoNoneField collected from 4 sitesM Josserand et al., 2015 H. arborea 9.550AdultHind leg muscle250.020.512, 14, 16, 18, 20, 22, 24 hSubsetMixedField collected from 1 siteM Madelaire et al., 2017 Rhinella jimi ——AdultHind fleshy base of foot200.010.212, 24 hNoBilateralField collected from 1 siteM R. granulosa —90No Pleurodema diplolister —20No Madelaire et al., 2019 R. jimi ——AdultHind fleshy base of foot200.010.212, 24 hNoBilateralField collected from 1 siteF, M R. granulosa —90No P. diplolister —20No Murillo-Rincón et al., 2017 Rana arvalis 10.8—LlarvaeBetween 4th and 5th muscular segment of tail10000.022048 hNoWhole specimenCaptive bred from field-collected frogsF, M Seiter, 2009 L. sylvatica 17.683LarvaeSubcutaneous base of tail20.020.0324, 48 hNoNoneEgg masses from fieldF, M Szuroczki et al., 2016 L. pipiens 39.2165LarvaeTail10.010.0124, 48 hNoWhole specimenPurchased from vendorF, M Szuroczki et al., 2019 L. sylvatica 17.683LarvaeTail10.010.0124 hNoWhole specimenEgg masses from fieldF, M Titon et al., 2016 Hypsiboas albopunctatus 7540AdultSubcutaneous hind fleshy base of foot200.010.212, 24 hNoBilateralField collected from 1 siteF, M Titon et al., 2019 R. ornata 12.184.29AdultHind fleshy base of foot200.010.212, 24 hNoBilateralField collected from 1 siteM Titon et al., 2021 R. icterica 120190AdultHind fleshy base of foot200.010.212 hNoBilateralField capturedM Troïanowski et al., 2017 H. arborea 9.550AdultHind leg500.02116 hNoBilateralField collected from 1 siteM Vasconcelos-Teixeira et al., 2022 Rhinella diptycha ——AdultHind fleshy base of foot200.010.212, 18 hNoWhole specimenField collected from 3 sitesF, M Venesky and Laskey, 2022 P. cinereus 65412AdultTail40.010.0424 hNoNoneField collected from 1 siteF, M Venesky et al., 2012 Lithobates sphenocephalus 65.7127Larvae2 mm from tip of tail, right side, at junction of tail and muscle10.020.0248 hNoBilateralEgg masses from fieldF, M Wang et al., 2019 X. laevis 130147adultmiddle plantar side of foot – random hind/fore20.10.224 hNoBilateralPurchased from vendorM Young et al., 2014 b Litoria caerulea 50100AdultToe web between 2nd and 3rd digits in hind foot50.10.56, 12, 24, 48 hNoNoneField collected from different sitesF, M Litoria infrafrenata 24.46110No Zamora-Camacho, 2019 E. calamita 62.7100AdultSubcutaneous forelimb sole pad1000.0111, 2, 3, 4, 5, 6 hNoBilateralField collected from 2 sitesF, M Zamora-Camacho et al., 2022 P. perezi 31.886AdultHind fleshy base of foot100.010.124 hNoNoneEgg masses from fieldF, M Zamora-Camacho and Comas, 2018 E. calamita 62.7100AdultSubcutaneous forelimb sole pad100.010.11, 2, 3, 4, 5, 6 hNoBilateralField collected from 1 larger areaF, M Zhang et al., 2017 P. nigromaculatus 56.490.8AdultToe web between 2nd and 3rd digits in hind foot3.30.050.176, 12, 24, 48, 72 hNoBilateralField collected from 1 larger areaF, M P. raddei ——No B. gargarizans 116.2134Noa aData are averages for adult of the species based on AmphBIO (Oliveira et al., 2017).bThis was the only study in our review that did not detect a positive response. Summary of amphibian PHA assay studies aData are averages for adult of the species based on AmphBIO (Oliveira et al., 2017). This was the only study in our review that did not detect a positive response. We found a total of 36 amphibian studies that have utilized the PHA assay. The first was published in 2003, and studies continued sporadically thereafter until 2014, at which point publications began to increase (Fig. 1). Timeline of published PHA studies in amphibians. Search was performed according to PRISMA (see Supplementary Fig S1 for full search details). Analysis of Existing Amphibian PHA Studies We divided our critical review of amphibian PHA studies into two sections: ‘Organismal Variables’, which focuses on multiple attributes of organisms used for testing (e.g. species, sex, life stage), and ‘Experimental Variables’, which focuses on aspects of experimental study design (e.g. PHA injection site and concentration). All information collected from amphibian PHA studies is summarized in Table 1. Organismal Variables Species Out of 36 PHA studies in amphibians, a total of 27 distinct species were examined representing seven different taxonomic families (Fig. 2; Table 1). All publications focused on frogs, except for a single study on salamanders, specifically Plethodon cinereus (Venesky and Laskey, 2022), and there have been no studies on caecilians. As expected, species-rich anuran families such as Ranidae (aquatic frogs), Hylidae (treefrogs and relatives), and Bufonidae (toads) are well represented. However, there are many large amphibian families or clades with substantial numbers of species that are not represented in PHA studies (Fig. 2), such as the Microhylidae, generally smaller, narrow-mouthed frogs, which contains >700 species. Amphibian phylogenetic tree by taxonomic family. Families in bold contain >100 species according to AmphibiaWeb (https://amphibiaweb.org). Families in italics contain >10 species listed as critically endangered by the IUCN Red List. Circles represent the number of species within a given family that have been tested in PHA studies (see Table 1). The phylogenetic tree was generated using all available amphibian taxa in TimeTree (http://www.timetree.org) Interspecific variation has been observed in swelling response (Clulow et al., 2015; Zhang et al., 2017) and lymphocyte proliferation (Wiley and Meisner, 1984). Therefore, analyzing a wide range of species is particularly important for the PHA assay. Variation in assay results, accounting for variation in sex and body size, has been found even among closely related species such as Limnodynastes peronii and Limnodynastes tasmaniensis (Clulow et al., 2015). Experts have cautioned against interspecific comparisons of PHA results (Martin et al., 2006), in part, due to varied investment in innate immunity among taxa (Woodhams et al., 2006). Martin. et al. suggests, based on much more robust avian literature, that these differences are due to a combination of the variety of immune cells involved in PHA oedema and variation in investment in production of different immune cells among species to support different environments, life histories and physiology (2006). The species tested in PHA studies are geographically unrepresentative as well. Species from Australia, China and North America are over-represented, while species from Africa and South America are substantially under-represented (Supplementary Fig. S2). This is problematic, not only because certain geographically limited taxa remain under-represented, but also because different regions have diverse ecological conditions that make it challenging to extrapolate results from species in other areas (Christie et al., 2020). Studying these species in a laboratory setting is a vital first step in investigating these differences and establishing standardized methods that are effective in diverse geographic areas. Both taxonomic and geographic biases are known problems for amphibian research (Christie et al., 2021). There is a clear imperative to test and/or extrapolate results to more species in underrepresented taxonomic groups and global regions. To further this goal, there are multiple species from understudied areas and taxa with known husbandry information (e.g. dendrobatid frogs) that could be used in future studies. Life stage and age Life stage, metamorphosis and age have complex effects on immune function and disease outcomes (Flajnik et al., 1987; Rollins-Smith, 1998; Haislip et al., 2011; Grogan et al., 2018). All PHA studies conducted so far have either tested a single life stage or placed alternative life stages into different experimental groups to avoid any confounding effects of age on results (Clulow et al., 2015). Most studies have been performed on adults (Table 1), although several studies have been performed on larvae (Venesky et al., 2012; Ceccato et al., 2016; Szuroczki et al., 2016, 2019; Murillo-Rincón et al., 2017; Venesky and Laskey, 2022) or juveniles (Clulow et al., 2015; Abu Bakar et al., 2016; Ceccato et al., 2016). Only two studies have directly compared age and swelling responses from PHA, while two others compared development time and PHA response (Abu Bakar et al., 2016; Zamora-Camacho and Comas, 2018). Abu Bakar et al. (2016) found that PHA-induced oedema in Litoria aurea exposed to Bd was higher in adults than in either subadults or juveniles. They also found that while Bd exposure reduced PHA-related swelling for all groups, the reduction was less pronounced in adults (Abu Bakar et al., 2016). The authors suggested that this response may illustrate maturation and recovery of the adaptive immune response after metamorphosis. In general, larval immune systems are less developed and robust than adult immune systems, and aspects of the immune system are suppressed during metamorphosis in anurans, although much less is known about urodeles (Rollins-Smith, 1998; Abu Bakar et al., 2016; Grogan et al., 2018). Another study on Epidalea calamita performed on adults exclusively found that swelling decreased with age, leading study authors to suggest that senescence may reduce immunocompetence after reaching adulthood (Zamora-Camacho and Comas, 2018). Two studies looking at the relationship between development time and immunocompetence found that increased development time in response to stress is associated with reduced PHA response (Gervasi and Foufopoulos, 2008; Seiter, 2009). Given the substantial remodeling of the amphibian immune system during metamorphosis, and the limited study across different age groups, further research in the laboratory is needed to elucidate the mechanism driving swelling-response variation between life stages (Flajnik et al., 1987; Grogan et al., 2018). Such study is an important precursor to field research since it would be impractical to limit field studies to similar-age individuals. Sex Variation in immunity among sexes is well known in vertebrates because of differences in physiology and life history traits (Nunn et al., 2009). In amphibians, males and females can differ substantially in size, behavior and hormone levels (Shine, 1979; Licht et al., 1983), all of which can impact immune function (French et al., 2007; Taneja, 2018; Vitousek et al., 2018). However, previous studies have found minimal differences in immunocompetence both among sexes and individuals of the same sex exhibiting variation in reproductive morphology and behavior during the mating season (Saad, 1992; Zapata et al., 1992; Cornuau et al., 2014; Thomas and Woodley, 2017). While some amphibian PHA studies looked at only males (de Assis et al., 2015; Desprat et al., 2015, 2017; Josserand et al., 2015; Madelaire et al., 2017; Troïanowski et al., 2017; Iglesias-Carrasco et al., 2018; Titon et al., 2019, 2021; Wang et al., 2019) or females (Fites et al., 2014) to avoid confounding sex-related effects, several other studies specifically investigated sex-related differences in PHA swelling responses (Bakewell et al., 2021; Zamora-Camacho et al., 2022). The results of amphibian studies investigating the influence of sex on PHA swelling have been inconsistent. A study of Anaxyrus punctatus reported greater swelling responses in females (Bakewell et al., 2021), while another study with E. calamita showed earlier swelling in females but statistically indistinguishable peak swelling among sexes (Zamora-Camacho, 2019). A recent study of Pelophylax perezi found that larval exposure to ammonium impacted immunocompetence in adult males, including PHA-induced oedema at 24 hours, but not in adult females (Zamora-Camacho et al., 2022). Several other studies with Rhinella marina, L. aurea, Litoria revelata, L. peronii, L. tasmaniensis, Pelophylax nigromaculatus, Pseudepidalea raddei and Bufo gargarizans showed no effect of sex on peak swelling (Brown and Shine, 2014; Brown et al., 2015; Clulow et al., 2015; Zhang et al., 2017). Each of these studies accounted for variation in individual body size in drawing conclusions about the effect of sex, although they did not measure or evaluate any additional measures of sex, sex hormones or reproductive behavior. Overall, there is limited support that sex and reproductive activity influence PHA swelling response, but further study is warranted given the conflicting results and lack of species examined. Given the current uncertainty regarding these effects, studies should at the very least record sex and integrate it into PHA swelling analysis. Body size The PHA assay, like many immunological tools, is simpler to perform on larger specimens, because swelling is easier to measure and less likely to require precise equipment (Szuroczki et al., 2016, 2019; Murillo-Rincón et al., 2017). As a result, PHA studies have been biased towards medium or large animals, with nearly 90% of species studied being medium-sized or larger in adulthood (see Table 1). Studying species across the full spectrum of body sizes may be particularly important for amphibians, as there is evidence that smaller sizes are associated with increased vulnerability to declines (Cardillo, 2021). Fortunately, progress has been made to develop methodologies and tools appropriate for smaller specimens, including altered injection sites and the use of microscopy. See Experimental Variables: PHA Injection for more information on relevant methodology, as well as Clulow et al. (2015), Murillo-Rincón et al. (2017) and Szuroczki et al. (2016). Body size variation within species may also affect PHA-related swelling, which experts have suggested may be due to either differing spleen size (Fites et al., 2014; Vasconcelos-Teixeira et al., 2022) or the higher cost of immune activation in smaller individuals (Venesky and Laskey, 2022). Six studies assessed the effect of conspecific body size on swelling response with mixed results: three studies reported no effect of body size on oedema (Brown and Shine, 2014; Brown et al., 2015; Clulow et al., 2015; Zamora-Camacho and Comas, 2018); two studies found that increased body size was associated with increased swelling (Brown et al., 2015; Titon et al., 2019); and one study found that decreased body size was associated with increased swelling (Bakewell et al., 2021). Generally, increased body size would be expected to be associated with an increased immune response (Fites et al., 2014; Vasconcelos-Teixeira et al., 2022; Venesky and Laskey, 2022). The results of Bakewell et al. (2021) are therefore unexpected. Authors of this study suggest it is possible that results may have been a result of inaccurate measurements. Therefore, further study using more precise measurements (see Experimental Variables: Swelling Response) would be beneficial (Bakewell et al., 2021). While many studies measure size using snout-vent length (SVL), measuring and reporting both SVL and weight of each individual tested would significantly improve future work by facilitating analysis of the impact of body size within and between studies. Experimental Variables PHA injection Injection procedures, including injection amount, site, and repetition, varied greatly among studies. Injection volumes of both PHA (diluted in saline solution) and PBS (control) varied from 4 μl (Ceccato et al., 2016) to 100 μl (Fites et al., 2014; Young et al., 2014; Wang et al., 2019). The most common injection volume of PHA or PBS solutions by far was 10 μl, which was used in 15 of 36 studies. The average injection volume across studies was approximately 30 μl. Given that injections of higher volumes can increase swelling regardless of content, maintaining an equivalent volume across all controls and treatments is important, as is minimizing injection volume when possible (Collins et al., 2020). The amount of PHA per injection ranged from 0.01 mg (Szuroczki et al., 2016, 2019) to 20 mg (Murillo-Rincón et al., 2017), with 0.2 mg being the most common (see Table 1). It should be noted that studies of large species such as R. marina and Rhinella icterica were successfully completed using a dose of only 0.1 mg, suggesting that higher doses may be unnecessary, and should be used only when lower doses are not successful (Brown et al., 2011, 2015; Brown and Shine, 2014). Only two studies to date have attempted to establish an ideal dose of PHA for the assay. Brown et al. (2011) tested three doses: 0.05, 0.1 and 0.25 mg on large R. marina; while the PHA amount was not related to peak swelling response, swelling response time was slower at the lowest dose, with no difference between the 0.1- and 0.25-mg doses (Brown et al., 2011). Josserand et al. (2015) found similar results comparing swelling responses of Hyla arborea injected with 0.25, 0.5 and 1 mg of PHA; while 0.25 mg was insufficient to generate a swelling response, there was no difference in peak swelling response between 0.5 and 1 mg. Based on these studies, it appears that once a minimum threshold is reached, increasing the PHA dose is unnecessary. Dosage also appears to vary substantially in avian studies (Smits et al., 2002); however, a general protocol has been developed for birds by Grasman at a dose of 0.1 ml of 1 mg/ml PHA solution (Grasman, 2010). A variety of injection sites were used across different PHA studies, species and life stages (Fig. 3). Injection sites should be tailored to the specimen type to generate an easily measurable swelling response (Clulow et al., 2015). However, many studies do not provide an adequate description of the injection site for proper duplication, using unspecific terms such as ‘tail’, ‘leg’ and ‘hind leg’ (Szuroczki et al., 2016; Desprat et al., 2017; Troïanowski et al., 2017). Still, some overall patterns can be observed. Subcutaneous injection of the toe web was often used for larger species but was not possible for smaller specimens or any specimens without toe webs (Clulow et al., 2015; Josserand et al., 2015). Methods used on smaller specimens included subcutaneous injections below the knee and intramuscular injections of the footpad and leg (Clulow et al., 2015; Josserand et al., 2015). The sole study that failed to generate a difference in swelling response between PHA and PBS controls used a subcutaneous toe web injection site (Young et al., 2014). On the other hand, Clulow et al. (2015) found that upper leg injections were not as successful with larger species, while toe web injections were not successful with smaller species and proposed that the lower leg and foot pad are the two most broadly applicable methods for PHA. Some research found greater success with intramuscular injection (Fites et al., 2014); however, there is histological evidence that PHA injections can cause muscle damage in some specimens, which contraindicates the use of intramuscular injection, unless necessary, until further research is conducted (Clulow et al., 2015). One study randomly injected hind or forelimbs, introducing possible problems comparing results within the study (Wang et al., 2019). Although avian studies were originally conducted using either the patagia or the foot as injection sites, there was no correlation found between swelling at these two sites (Kean and Lamont, 1994). Today, the patagia is used nearly universally (Martin et al., 2006; Hasselquist, 2007; Grasman, 2010). It is important to evaluate and address the lack of comparability among injection sites in amphibian studies, although it is unlikely that a single universal injection site can be established. This issue should be thoroughly examined and taken into consideration when designing and interpreting amphibian studies. Anatomical map of PHA injection sites used across amphibian studies. Colors represent types of injection. Numbers inside circles show the number of studies utilizing each injection location. This froglet image represents an amalgam of different amphibian species and life stages since injections were performed on adult and larval frogs and salamanders. For caudates and larval anurans, tail injections are easier because legs are either less prominent or absent altogether. For the smallest specimens, studies have adopted protocols with a high level of specificity of measurement to ensure comparability of swelling responses among individuals. Venesky et al. (2012), for instance, made injections and measurements 2 mm from the tip of the tail, while Murillo-Rincón et al. (2017) specified that injections and measurements took place between the fourth and fifth muscular segments of the tail. Such specificity is not found in publications of adult studies, although it would be beneficial. Studies suggest that while the first PHA injection is a broad measure of immune response that engages both the innate and adaptive immune system, a second PHA injection elicits a more substantial adaptive immune response, prompting some researchers to use inoculation to provide a more robust measure of adaptive immunity (Fites et al., 2014; Josserand et al., 2015). Five of the 36 studies inoculated specimens with PHA in the weeks leading up to the assay. These studies, which included Lithobates pipiens, R. marina, X. laevis and H. arborea, all used inoculation to elicit a stronger measure of adaptive immunity. The only exception was Brown et al.’s study of R. marina(2011), which examined whether inoculation increased swelling response as would be expected if the adaptive immune response was enhanced by multiple exposures (2011). Their results showed that the second injection of inoculated specimens did cause increased swelling, supporting this hypothesis. A study by Fites et al. (2014) found that, while swelling response to the first PHA injection was not affected by exposure to Bd, swelling response to the second exposure was significantly impacted, suggesting the possibility that Bd supernatants affect the adaptive arm of the immune system more than the innate, though it should be noted that the adaptive and innate immune systems are not truly separate; in fact, the innate immune system initiates and guides the adaptive immune response (Dempsey et al., 2003). Still, inoculation certainly warrants further study, as it could facilitate assessment of more specific functions of the amphibian immune system. However, it should be noted that this method is not ideal for field use, as it would necessitate a much longer holding time of animals, which may be impractical. Swelling response Interstudy variation in swelling response is difficult to assess, because swelling has been reported differently among studies. Many studies used the approach adopted by Brown et al. (2015) and Zhang et al. (2017), reporting a proportion as change in thickness over starting thickness (Brown et al., 2011, 2015; Brown and Shine, 2014; Clulow et al., 2015; de Assis et al., 2015; Abu Bakar et al., 2016; Szuroczki et al., 2016; Madelaire et al., 2017, 2019; Murillo-Rincón et al., 2017; Brannelly et al., 2019; Titon et al., 2019; Cassettari et al., 2022; Vasconcelos-Teixeira et al., 2022; Venesky and Laskey, 2022). Meanwhile, many other studies reported millimetres of swelling without providing a starting or ending thickness measure to allow comparison with percentage-based reporting (Gilbertson et al., 2003; Gervasi and Foufopoulos, 2008; Seiter, 2009; Venesky et al., 2012; Young et al., 2014; Desprat et al., 2015, 2017; Ceccato et al., 2016; Szuroczki et al., 2016, 2019; Titon et al., 2016; Troïanowski et al., 2017; Wang et al., 2019; Zamora-Camacho, 2019; Zamora-Camacho et al., 2022). Other methodologies have been used less often; for example, Zamora-Camacho and Comas (2018) calculated the ‘area under the curve’, by integrating swelling response measurements over multiple time points. Swelling responses in PHA studies are measured at set times after injection using a variety of measurement equipment including digital calipers (Fites et al., 2014; Wang et al., 2019; Venesky and Laskey, 2022) and pressure-sensitive or adjustable force micrometers (Clulow et al., 2015; Abu Bakar et al., 2016; Zhang et al., 2017; Zamora-Camacho and Comas, 2018) to measure relative width before and after swelling. While adjustable force micrometers are more accurate than digital calipers, they have also been tied to a ‘compression effect’, in which leg measurements decrease, or, in the treatment leg, increase less, due to compression of the tissue by the device. Compression effects should be accounted for via control measurements if these tools are used (Clulow et al., 2015) (see Experimental Variables: Experimental Controls). However, a study using digital calipers also found negative swelling responses in some individuals, suggesting digital calipers should also be evaluated for compression effect (Ceccato et al., 2016). Meanwhile, for smaller specimens, such as larvae, more specialized tools have been used. Murillo-Rincón et al. (2017) used photography, a reference grid and software to measure swelling, while Szuroczki et al. (2016, 2019) used a dissection microscope along with digital calipers. While the dissection microscope method requires anaesthesia (Szuroczki et al., 2016, 2019), the photography and reference grid method appears to avoid the compression effect and reduce handling time without the need for anaesthesia; however, authors indicated that some results had to be excluded due to larval positioning in the photographs (Murillo-Rincón et al., 2017). Therefore, while promising, a digital imaging protocol could be further developed in the future to ensure accuracy and feasibility among different species, life stages, body sizes and morphological features. While relatively expensive, magnetic resonance imaging (MRI) is still the best practice for digital imaging of soft tissues. However, advances in technology have allowed both analysis of 2D images, as in Murillo-Rincón et al. (2017), and more advanced photogrammetry, in which 2D images are used to generate a 3D model, though one less accurate than computed tomography (CT) scans (Ford et al., 2023). Ford et al. (2023) provides an excellent overview of current methods of anatomical study, including the latest advances in digital imaging. Measurement intervals, or time passed since injection when swelling measurements are taken, have differed among studies. Validation and optimization studies (i.e. studies designed to identify the timing of peak swelling) often involve taking many measurements at different time points, while immunology studies tend to choose one time point thought to coincide closely with peak swelling. Identifying the timing of peak swelling can be complex and varies considerably between species (Clulow et al., 2015; Madelaire et al., 2017, 2019; Zhang et al., 2017; Zamora-Camacho, 2019), sex, and population (Zamora-Camacho, 2019). For example, Josserand et al. (2015) found that swelling occurred in H. arborea ~14 hours after injection, while Clulow et al. (2015) found that peak swelling occurred at 24 hours for L. aurea, L. revelata and L. tasmaniensis, but occurred at 48 hours for L. peronii. Therefore, Clulow et al. (2015) cautioned that, before the time frame for a given study is set, species should be validated, and measurement protocols optimized according to a standard protocol. Twenty-four hours is the most common time point used to measure peak swelling in amphibian PHA studies (Table 1). Many studies have used 24 hours as a default for species without carrying out a validation step to empirically determine peak swelling time (Iglesias-Carrasco et al., 2018; Szuroczki et al., 2019; Wang et al., 2019; Bakewell et al., 2021; Zamora-Camacho et al., 2022). For avian studies, the standard protocol is to take measurements at ~24 hours (Grasman, 2010). A similarly consistent time point that is effective across taxa for amphibians would facilitate future study, though validation of such a time point across taxa beforehand is certainly advisable. Although studies aim to capture the approximate time of peak swelling, which is expected to occur and subside within 72 hours after injection, there is no consensus on the optimal frequency of measurements required to achieve this objective. A ‘24-hour regime’, in which triplicate measures are taken at 6, 12, 24, 48 and 72 hours (Clulow et al., 2015), was used in four methodology optimization studies (Brown et al., 2011; Brown and Shine, 2014; Young et al., 2014; Zhang et al., 2017), while an adapted ‘every 4-hour regime’, in which single measurements are taken every 4 hours for the first 72 hours of the study, was only used by Clulow et al. (2015). These two optimization approaches have benefits and drawbacks; while more frequent 4-hour measurements would more accurately determine peak swelling, taking single rather than triplicate measurements could reduce precision and accuracy of the measurements themselves. Conversely, an increased number of measurements could increase handling stress and researcher effort, which must be weighed against the benefits of more thorough validation. Nineteen studies, meanwhile, took measurements at two or fewer times, using species that had not been previously optimized with more granular measurements (Gervasi and Foufopoulos, 2008; Seiter, 2009; Fites et al., 2014; de Assis et al., 2015; Desprat et al., 2015; Szuroczki et al., 2016, 2019; Titon et al., 2016, 2019, 2021; Madelaire et al., 2017, 2019; Murillo-Rincón et al., 2017; Troïanowski et al., 2017; Iglesias-Carrasco et al., 2018; Wang et al., 2019; Bakewell et al., 2021; Vasconcelos-Teixeira et al., 2022; Zamora-Camacho et al., 2022). Studies by Zamora-Camacho (2019) and Zamora-Camacho and Comas (2018) are unique in that they tested ‘early swelling peak’, within just 6 hours of injection. This approach is noteworthy as it could help improve practicality in the field by reducing waiting time. While innate immune response swelling occurs rapidly, before the combined innate and adaptive-driven peak (Szuroczki et al., 2016), significant early swelling with injection of both PHA and PBS has been observed (Josserand et al., 2015). Given that both of Zamora-Camacho, 2019 studies were performed without a control, replication of these studies with a PBS control would help verify whether this approach is appropriate. The effects of PHA assay-induced swelling on amphibian mobility, health and stress are still uncertain. To date, no study has specifically sought to identify the post-assay effects of PHA itself on amphibians. One study by Iglesias-Carrasco et al. (2018) tested the effect of the PHA assay on hiding time and found no change from PHA treatment; however, this is an anti-predator behavior that relies relatively little on limb mobility. While a study of avian nestlings found that the PHA assay did not increase levels of stress proteins (Merino et al., 1999), no similar study has been conducted on amphibians. Meanwhile, some studies suggest potential impacts of the PHA assay on amphibians that warrant further research. Zhang et al. (2017) noted that the PHA assay caused a reduction in body mass across the course of the study. Murillo-Rincón et al. (2017) found that when performed on larvae, the PHA assay increased the larval period, which is potentially problematic given that development time has been a long-standing measure of fitness and can impact survival in ephemeral environments. Desprat et al. (2017) observed that the PHA assay caused a lightening of the vocal sac in starved males proportional to the PHA response, which the authors suggested may be an indicator to females that a male recently encountered a disease agent, reducing chances of breeding success. Finally, histological analysis by Clulow et al. (2015) found evidence of damage to the muscle fibers and fibrosis in some specimens injected with PHA. While none of these findings preclude the use of the PHA assay for field studies, further research is needed to ensure that it minimally impacts amphibian health, especially before the assay is performed on more vulnerable species. Experimental controls It is well established that appropriate controls are an essential part of scientific study. In the case of PHA studies, the approach to implementing controls has varied. Many studies evaluated PHA-associated swelling by injecting PHA into one limb, while injecting PBS into the other limb as a control (Table 1). It is possible that PHA injections can have a systemic effect, meaning that these bilateral controls are not ‘true’ controls. Other studies, primarily of anuran larvae and salamanders, made injections in the tail rather than legs. All but one of these studies (Ceccato et al., 2016) added a separate control group of animals that received PBS control injections instead of PHA, a practice referred to hereafter as ‘whole animal controls’. One larval study included a third control group that received no injection at all (Murillo-Rincón et al., 2017). Nearly 20% of studies did not use controls of any kind. These studies generally cited research by Smits et al. (1999), showing that eliminating controls reduced variation, sampling time, handling stress and measurement variability, and that controls only served to identify two outliers out of 608 specimens tested, leading the authors to suggest it would be best to eliminate controls in these studies. However, no such analysis has ever been performed on amphibians. Furthermore, controls are necessary to account for the compression effect, discussed above in Experimental Variables: Swelling Response (Clulow et al., 2015). While rarely acknowledged in the literature or accounted for by measurement technique or statistical analysis, the compression effect has been observed in studies in which negative swelling results were found. Gervasi and Foufopoulos (2008) found these negative differences negligible, and included them in their study; however, Ceccato et al. (2016) found negative swelling results in many measurements, and the choice by the authors to remove these anomalous results reduced sample size significantly. For all these reasons, steps are needed to either statistically control for compression or pursue non-compressive measurement tools (e.g. digital imaging). Until then, the use of experiment controls remains a best practice for amphibian PHA studies. Histology Histological analysis is essential to establish the mechanisms behind, and interpretation of, PHA-induced oedema (Zhang et al., 2017). However, only four amphibian PHA studies to date have published histological information (Brown et al., 2011, 2015; Clulow et al., 2015; Josserand et al., 2015). Josserand et al. (2015) verified results from H. arborea using a blood smear taken 24 hours after injection, stained using May Grünwald-Giemsa, and examined using light microscopy to identify white blood cells. Measurements included leukocyte ratios per 100 white blood cells, leukocytes per 10 000 red blood cells and the ratio of neutrophils to lymphocytes. Results showed that two injections of PHA induced higher leukocyte counts per 10 000 red blood cells, as well as higher lymphocyte ratios, than a single injection alone (Josserandet al., 2015). The study also found that white blood cell counts were correlated with the size of the swelling response. The authors suggested that their data support the hypothesis that initial swelling from PHA injection represents an innate response, while the second, which is heightened by inoculation, is more adaptive. Brown et al. (2011, 2015) performed histological analysis of injected tissue directly in R. marina. Biopsies of injected skin were taken post-injection for both PHA and PBS injection sites, fixed in phosphate-buffered formalin, embedded in paraffin, stained in haematoxylin and eosin and examined using light microscopy to identify white blood cells and generate a total cell count per set field of view. The first study (Brown et al., 2011) sought to use histological analysis to assess the mechanism behind PHA swelling at varying points in the process. This analysis found that the magnitude of the swelling response at PHA injection sites was positively correlated with the average number of lymphocytes per field of view, with eosinophils, neutrophils and macrophages peaking at 12 hours post-injection, and lymphocytes peaking at 24 hours post-injection, lending support to the effectiveness of the assay to assess both innate and adaptive immune system function, with optimal testing times depending on the response of interest (Brown et al., 2011). The authors noted, however, that R. marina did not show lymphocyte mitosis in response to PHA until 3 days after injection; therefore, the swelling response at 24 hours is largely representative of recruitment of pre-existing lymphocytes (Goldshein and Cohen, 1972; Brown et al., 2011). Meanwhile, PBS injection was associated with a brief increase in non-lymphocyte white blood cells, but not a significant increase in lymphocyte recruitment (Brown et al., 2011). The second study (Brown et al., 2015), which took histological samples only at 72 hours post-injection, found similar results, with white blood cell counts significantly higher at PHA injection sites. Finally, Clulow et al.’s (2015) study of L. aurea and L. peronii conducted histological analysis at 24 hours post-injection. Biopsies of injected skin were fixed in Bouin’s fixative, embedded in paraffin, marked with Blue Tissue Marking Dye, stained in haematoxylin and eosin and examined using light microscopy to assess inflammation and identify and count white blood cells (Clulow et al., 2015). Analysis showed significant vasodilation and infiltration of leukocytes in PHA injection sites (Clulow et al., 2015). Cell counts were significantly higher at PHA injection sites for lymphocytes, neutrophils and macrophages, though not for eosinophils, supporting engagement of both innate and adaptive immune responses (Clulow et al., 2015). Histological verification of the response to PHA provides important insights into the nature of the immune response, which is critical for understanding the interaction between amphibian hosts and their pathogens. However, given the need to sacrifice individuals to acquire this verification, it is essential to apply them only where they are most needed. Otherwise, analysis of blood samples may be preferable. The Role of the PHA Assay in Assessing Amphibian Immunity A primary objective of this review is to examine the existing body of work surrounding the PHA assay in amphibians. The assay is intended to evaluate the potential of an individual to mount a localized inflammatory reaction in response to an infection. In vertebrates, increased PHA swelling is generally regarded as an indicator of immune health (Seiter, 2009; Brown and Shine, 2014; Ceccato et al., 2016; Szuroczki et al., 2016, 2019). Studies have shown that the PHA test can be used to measure immunocompetence in birds (Martin et al., 2006; Tella et al., 2008; Salaberria et al., 2013). These tests indicate that the area of swelling is associated with the infiltration and counts of certain immune cells (Martin et al., 2006). Furthermore, these tests also demonstrate that the PHA assay produces a stronger response after inoculation, causing T-lymphocyte proliferation after a single dose, but even greater proliferation after a subsequent dose (Tella et al., 2008). There is also a positive correlation between PHA oedema and the phagocytic activity of heterophiles and monocytes (Salaberria et al., 2013). While research validating the PHA assay in amphibians is not as extensive, increased PHA oedema has been consistently associated with elevated blood leukocyte levels (Gervasi and Foufopoulos, 2008; Josserand et al., 2015; Szuroczki et al., 2019). Histological assessment in amphibians has found immune cell infiltration at the injection site (Brown et al., 2011; Josserand et al., 2015; Ceccato et al., 2016), and inoculation has been shown to induce larger swelling responses on subsequent injection (Brown et al., 2011). PHA assay results have also been associated with important measures of individual or population health status. PHA results in birds have been tied to body condition, survival rates, longevity, current infection, and recruitment (Navarro et al., 2003; Moreno et al., 2005; Sivaraman and Kumar, 2013; Bowers et al., 2014). In amphibians, higher levels of PHA swelling have been linked to protein- and anti-oxidant-rich diets during the larval stage, which authors attribute to the positive impact of these diets on immune health (Venesky et al., 2012; Szuroczki et al., 2016, 2019). Conversely, factors associated with reduced PHA swelling in amphibians include desiccation, density stress, captivity, and predator presence, attributes which the authors suggest would reduce immunocompetence (Gervasi and Foufopoulos, 2008; Seiter, 2009; Clulow et al., 2015; Titon et al., 2019). Despite strong evidence supporting PHA as a meaningful measure of immunity, the immune system is highly complex, and different immune measures that capture diverse aspects of the immune system can yield conflicting patterns (Kennedy and Nager, 2006; Martin et al., 2006). For instance, Brown and Shine (2014) found that toads that moved greater distances had diminished bacteria-killing ability (BKA) and phagocytic abilities, but increased PHA oedema. The authors suggest that this difference may be the result of a trade-off in which increased energy requirements lead to reduced investment in bacteria-killing and phagocytic abilities, while leukocyte proliferation and inflammation are increased to compensate. Similarly, in the avian PHA literature, Pigeon et al. (2013) found a significant negative correlation between BKA and PHA results in intensive farmland habitats. Other studies, meanwhile, found positive correlations between these same immune measures. A study of Rhinella ornata toads found that long-term captivity suppressed both PHA swelling and BKA (Titon et al., 2019), and a study of Passer domesticus found that PHA and BKA results were both positively correlated with the same major histocompatibility complex genes (Lukasch et al., 2017). Although these conflicting patterns are not surprising given the complexity of the immune system and the limitations of any given immune test (Kennedy and Nager, 2006; Martin et al., 2006), further research investigating the relationships between external factors and different aspects of the immune response is needed to elucidate the mechanisms driving variation. In addition, performing multiple immune assays, where possible, can help to identify how specific aspects of immunity are affected by internal and external factors (Martin et al., 2006). Adding to the complexity, different diseases have varying effects on immune measures, and pathogens themselves may be affected differently by specific aspects of immunity measured (Viney et al., 2005). For instance, certain pathogens or contaminants, like Bd, may suppress components of the host immune system (Fites et al., 2014). Studies on adult amphibians have shown reduced PHA swelling in response to infections caused by Bd and Echinostoma trivolvis (Fites et al., 2014; Szuroczki et al., 2016; Venesky and Laskey, 2022). However, some pathogens or contaminants can indirectly upregulate certain aspects of the immune system while suppressing others. Exposure to organochlorines in amphibians, for example, has resulted in increased PHA oedema, despite a suppressed antibody response to keyhole limpet haemocyanin linked to dinitrophenyl (KLH-DNP) (Gilbertson et al., 2003). The authors postulated that contaminants may stimulate certain facets of the immune system or trigger compensatory upregulation to counter reduced antibody production. This phenomenon has also been observed in birds. In a study on American kestrels, exposure to polychlorinated biphenyls (PCBs) increased PHA swelling and white blood cell counts but decreased plasma triiodothyronine levels, which have been associated with reduced antibody production and T-cell response (Smits et al., 2002). The researchers hypothesized that PCB exposure may directly or indirectly affect thyroid hormones, leading to a diminished antibody response and stimulated white blood cell proliferation. However, there is a lack of laboratory studies that directly link PHA results to specific disease outcomes, and such studies would greatly facilitate the interpretation of PHA studies (Owen et al., 2010). Furthermore, establishing expected swelling response in a given species would enable the detection of both unusually low and unusually high swelling responses, which could indicate immune dysregulation. Challenges and Recommendations There are now several dozen amphibian PHA studies in the literature that have yielded important insights into amphibian immunity. However, a lack of validation and optimization studies, substantial variation in methodology and reporting, and limited taxonomic and geographic diversity still limit its application. Here, we discuss areas where improvement is needed based on this review and make recommendations for addressing these areas with future research. Challenge 1:Challenge 1: Lack of validation and optimization reduces consistency and utility of resultsRecommendation: Validations should be performed, and protocols optimized for all unvalidated study species, methodologies, and control schemas, including more substantial histological study. Challenge 1: Lack of validation and optimization reduces consistency and utility of results Recommendation: Validations should be performed, and protocols optimized for all unvalidated study species, methodologies, and control schemas, including more substantial histological study. Validations are used to confirm the efficacy of the PHA assay for a given species or life stage. This typically involves optimizing technical variables such as measurement timing, injection concentration, and injection site, to effectively measure a swelling response to PHA. Therefore, the terms ‘validation’ and ‘optimization’ are often used interchangeably in the literature. Several validation studies have evaluated swelling measurement timing, which is essential for establishing a peak swelling response (Table 1). Since peak swelling varies from species to species, or even potentially among individuals, sexes, and life stages, this validation step increases confidence in results for a given species. Most validation studies have used the ‘24-hour regime’ using triplicate measurements, while one study used an ‘every 4-hour regime’ using single measurements. More frequent measurements are better for determining the timing of peak swelling, while recording multiple measurements increases confidence in results. However, increasing measurements also increases handling stress, risk of injury, and research effort. Ideally, the best approach is one that is more accurate and precise, and less invasive and potentially harmful. In this vein, researchers should consider working to develop digital imaging measurement techniques that require minimal handling stress and enable more accurate measurements without the risk of tissue compression. Until such a methodology is fully developed, researchers should perform optimizations on any newly tested species by measuring swelling every 4–6 hours for the first 12 hours after injection, and then at least every 12–24 hours thereafter, with preferably two or more replicates per time point. Since the ‘24-hour regime’ is the most used, and involves measurements in triplicate, it can be an effective way to keep future studies consistent. Ultimately, the establishment of a consistent measurement time, such as 24 hours used in the standard avian protocol, would be beneficial. However, such a protocol could do more harm than good if the measurement time fails to capture significant immunological data for some species, making validation of such a universal measurement time essential. Given that the majority of PHA studies in amphibians have used control limbs rather than whole animal controls, and some studies do not use controls at all, there is a need to delineate which types of controls are/are not needed under different scenarios. While studies in birds have suggested that PHA controls are not necessarily beneficial, the same conclusion has not been adequately drawn for amphibians. Demonstrating that controls are unnecessary for amphibians under certain conditions would lend increased flexibility to field studies. In addition, salamanders and larval amphibians typically require whole animal controls because their limbs are too small or absent to receive concurrent control injections, which greatly increases the sample size. Some have argued that controls are essential in amphibian PHA studies to account for the compression effect (Experimental Variables: Swelling Response). However, future improvements in digital imaging-based measurements of oedema could alleviate this concern. Until such time, true whole animal controls are still generally recommended, and at least using control limbs in post-metamorphic anurans remains best practice. In addition to general validation and optimization, more thorough histological investigation would improve our understanding of the specific lymphocytes accumulating in tissues surrounding the injection site. Although histological analyses have been performed in previous PHA studies in amphibians, there are still gaps to be addressed by future work. First, there has been no hour-to-hour analysis of cell infiltration into tissues. Second, B- and T-cell counts have not been differentiated in prior studies. Third, there has been no detailed histological comparison of immune responses between species. Finally, the influence of inoculation on histology has not been evaluated. All these comparisons would better indicate the nature and progression of the immune response to PHA and would help establish the extent to which swelling represents innate immunity, adaptive immunity or both at a given point in time. This would greatly improve the applicability of the PHA assay for predicting how a given population or species may respond to infectious diseases. The primary drawback of histological analysis is the need to sacrifice specimens; therefore, we recommend that such validations of swelling response be initially conducted at broader taxonomic levels , reserving histological analysis for detailed, species-by-species validation. Challenge 2: Inconsistent methodologies among studies reduces utility and may cause unnecessary harm Recommendation: Consistent, conservative methodologies are needed for measurement timing, injection site, and PHA dose. Certain methodological aspects of PHA studies in amphibians are highly inconsistent across studies. As discussed above, many studies have attempted to measure peak swelling, often at a ‘default’ time such as 24 hours, while others have measured swelling at earlier time points. Injection site has also varied substantially (Fig. 3). Toe and toe web injections have been found to be effective for larger adult frogs, but other body forms and sizes require alternative injection sites. PHA dose, however, is the area that would most benefit from greater consistency, varying dramatically from study to study, from 0.01 to 20 mg. While protocols exist for optimizing peak swelling timing for untested species, there is no protocol for determining PHA dose or injection site. Injection of 0.5 mg of PHA seems to be effective even for larger species, while 0.2 mg was most used, and 0.1 mg was successfully used in specimens as large as R. marina. Subcutaneous toe or toe web injections have been widely used in larger species, while smaller species tend to find greater success with injections further up the leg. Meanwhile, tail injections were used for salamanders and larval amphibians. Therefore, studies should consider establishing a swelling response with injections of 0.5 mg PHA either using the toe or toe web for large species, higher on the leg for smaller species, and on the tail for salamanders and larval amphibians. Only when a significant swelling response is not found using these methods do we recommend using higher concentrations of PHA. Although some research found greater success with intramuscular injections, given that there is histological evidence that PHA injections may cause muscle damage in some specimens, studies with smaller specimens should begin with subcutaneous injections on the leg, and only if unsuccessful turn to intramuscular injections. As in avian studies where the PHA protocols have been standardized, establishing a consistent dose and injection site for as many amphibian species as possible would facilitate comparisons among species and studies. Challenge 3: Lack of consistency in reporting impedes interpretation and replication Recommendation: Injection site, morphological features and swelling response reporting should be expanded and standardized across studies. Reporting of injection site, swelling response, and basic information about test animals (e.g. weight and SVL) has varied widely across studies, impeding study comparability and study design development. Many studies do not provide sufficient detail of injection sites to allow for duplication. Based on the avian literature, swelling responses are not likely comparable among swelling sites, further supporting the need for standardization. To address this problem, procedures like those found in Venesky et al. (2012) are recommended, in which a specific measurement from a consistent reference point (e.g. tip of the tail) was used. Ideally, a similar location would be generally used across studies within the same species and metamorphic stage, and, where possible, other species with similar morphological attributes. Swelling response reporting has also been inconsistent. To enable comparison between studies, a consistent reporting methodology, ideally coupled with a full dataset of starting and peak measurements for both PHA and PBS injection sites, should be reported. We recommend adopting the percentage methodology of Brown et al. (2015) and Zhang et al. (2017) (Equation1).(1)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} $$ \small\begin{align*} 100&\times \left(\frac{PHA\ thickness\ at\ x\ hour- PHA\ thickness\ at\ 0\ hour}{PHA\ thickness\ at\ 0\ hour}\ \right)\notag \\ &- \left(\frac{Saline\ thickness\ at\ x\ hour- Saline\ thickness\ at\ 0\ hour}{Saline\ thickness\ at\ 0\ hour}\right) \end{align*}\end{document} Some validation studies are performed as a pilot, but go unreported, making their findings less accessible for future study of the species (de Assis et al., 2015; Venesky and Laskey, 2022). Reporting these findings in supplemental information, as in Zhang et al. (2017), would facilitate future study and repeatability. Finally, studies should provide raw data in publications, when possible, in accordance with the FAIR (findable, accessible, interoperable and reusable) principles. It is common not only in PHA studies, but also many other biological disciplines, to present summary statistics for each experimental treatment group or replicate, but this makes it difficult to compare results among studies and improve the PHA assay through meta-analysis. More thorough reporting of basic information for each individual including size (both SVL and weight) and swelling measurements as supplemental materials would greatly enhance protocols for PHA testing in amphibians. We recommend that datasets be deposited in permanent data repositories upon publication. These datasets should be easy to find and freely accessible to increase their applicability in conservation biology. Challenge 4: Taxonomic and geographic biases reduce conservation applicability Recommendation: An effort should be made to apply the PHA assay to under-represented taxonomic groups and geographic areas. Given its affordability, ease of use and relatively minimal impact on specimens, the PHA assay is uniquely suited to testing a wide array of amphibian groups worldwide. Historically, however, the assay’s application has been limited, with many taxonomic groups and geographic areas underrepresented in the literature. More studies are needed on caudates, amphibians from South America and Africa, and data-deficient families such as Microhylidae, Ceratobatrachidae and Strabomantidae. To begin with, studying species that are easily obtainable or have known husbandry such as Leptopelis calcaratus (Africa) or Dendrobates pumilio (South America) could be a potential next step. Challenge 5: Further research needed to support field implementation and explore effects of PHA assay on amphibians Recommendation: Further laboratory study is needed to establish the impact of organismal and environmental variables on PHA assay results, and the effects of the PHA assay on animal welfare, both short- and long-term. Once these gaps have been addressed, field-specific protocols can be established, and field study undertaken. To expand the applicability of the PHA assay, a greater understanding of the impact of organismal and environmental variables on assay results must be established. Once these variables are sufficiently understood to interpret the results of the PHA assay in the field, field-specific protocols must be established and validated. In addition, the effects of the assay itself on specimen mobility, health and stress, especially during active swelling, should be studied to ensure it does not harm animals. The impact of organismal and environmental variables can be assessed via laboratory validation and targeted study. Once this gap has been addressed, field protocols may be established by (1) identifying and validating (via histology or similar immunological testing) the earliest possible swelling response to allow full testing within a single day (optimized in the lab), or alternatively, overnight housing in containers that generate a minimal stress response (optimized in the field), (2) creating consistent injection and swelling measurement protocols that can be replicated under field conditions and (3) identifying and, where possible, addressing environmental factors that could affect field results. In addition, effects on specimens, particularly in the first 72 hours, should be carefully evaluated. To date, no study has specifically sought to identify the effect of the PHA assay on mobility, health or stress. Future studies that directly measure mobility, stress hormones, food consumption and other measures of health and fitness, are essential to establish the field viability of this test and to assess the viability of its use in vulnerable species. Challenge 6: Further investigate relationships between the PHA assay and disease outcomes Recommendation: Further study should be undertaken to establish the relationships between PHA assay results and specific pathogens and contaminants across species. While a significant swelling response to PHA is generally associated with immunocompetence, and the lack thereof indicates immunosuppression, there are exceptions. The complexity of the innate immune response or disease conditions supports the need for further research on the relationships between disease outcomes and PHA results in a particular species. It is crucial to study both how various diseases and contaminants influence a species' PHA response, and what these PHA results can reveal about potential disease outcomes. To this end, we recommend that the following types of studies are conducted: 1) species-specific analysis to establish baseline PHA responses across a range of healthy individuals, 2) disease-focused investigations that evaluate the impact of diseases of interest on a species’ PHA response, and 3) contaminant studies that investigate the potential influence of contaminants and other environmental stressors on PHA responses. Where feasible, we also recommend that multiple measures of immunity be gathered to generate as complete a picture of immune health as possible. Conclusion In the past 2 decades, a foundation has been built for applying the PHA assay to amphibian immunology and conservation. Such tools are more important than ever given the rise of multiple emerging infectious diseases in amphibians and their many interactions with global environmental change. To date, there have been 36 PHA studies on 27 amphibian species, spanning a wide variety of research goals. In this review, we have outlined a set of improvements that would enable the PHA assay to be more broadly and consistently applied across amphibian taxa. Detailed protocols and optimization are needed for all test species, including the determination of swelling measurement times, injection sites, and PHA injection concentrations, confirmation of the swelling response via histological analysis and the use of appropriate experimental controls. The results of all PHA studies, including validation studies, should have consistent and detailed reporting, including enough information to allow replication or meta-analysis. Such information includes basic morphological characteristics such as size and weight, as well as swelling response for each individual tested, and clear communication of injection site. Additional effort should be made to conduct studies with species from underrepresented taxa and geographic areas. More work is needed to develop protocols that are amenable for field use, including confirming peak swelling timing for species and evaluating animal welfare post-test. Adaptation of the PHA assay to the field will help address ecological questions that may not be easily answered with model organisms in the lab. Finally, better communication among researchers utilizing the PHA assay in amphibians will help increase the consistency and utility of studies, which could be achieved, for example, through the development of society meeting sessions or workshops or the creation of common study protocols and templates. Supplementary Material