Defects in Host Immune Function in Tree Frogs with Chronic Chytridiomycosis

The amphibian chytrid fungus Batrachochytrium dendrobatidis (Bd) has caused mass mortality leading to population declines and extinctions in many frog species worldwide. The lack of host resistance may be due to fungal immunosuppressive effects that have been observed when Bd is incubated with cultured lymphocytes, but whether in vivo host immunosuppression occurs is unknown. We used a broad range of hematologic and protein electrophoresis biomarkers, along with various functional tests, to assess immune competence in common green (Litoria caerulea) and white-lipped (L. infrafrenata) tree frogs experimentally infected with Bd. Compared with uninfected frogs, Bd infection in L. caerulea caused a reduction in immunoglobulin and splenic lymphocyte responses to antigenic stimulation with sheep red blood cells, along with decreased white blood cell and serum protein concentrations, indicating possible impaired immune response capability of Bd-infected frogs. This is the first in vivo study suggesting that infection with Bd causes multiple defects in systemic host immune function, and this may contribute to disease development in susceptible host species. Although L. infrafrenata failed to maintain Bd infection after exposure, white blood cell and serum globulin concentrations were lower in recovered frogs compared with unexposed frogs, but antigen-specific serum and splenic antibody, and splenic cellular, responses were similar in both recovered and unexposed frogs. This may indicate potential systemic costs associated with infection clearance and/or redirection of host resources towards more effective mechanisms to overcome infection. No clear mechanism for resistance was identified in L. infrafrenata, suggesting that localized and/or innate immune defense mechanisms may be important factors involved in disease resistance in this species.


Introduction
The recent global spread of the emerging infectious disease chytridiomycosis has caused declines and extinctions of many amphibian species [1][2][3]. The causative fungal skin pathogen, Batrachochytrium dendrobatidis (Bd), has had the most devastating impact in remote and protected mountainous regions, where abundant populations crashed within months of its arrival [1,2,4]. Environmental changes such as pollution and climate are generally not considered primary factors in its emergence -Bd can clearly cause high mortality rates in healthy, immune competent populations [3]. Combined with its ability to spread rapidly through host populations and persist even at low host densities, it has had an unprecedented effect on amphibian biodiversity [1,2,[5][6][7]. If naïve susceptible amphibian populations survive introduction of Bd, it becomes endemic with reduced mortality rates and sometimes partial recovery, suggesting selection for host resistance and/or waning pathogen virulence [8].
Wide variation in susceptibility to chytridiomycosis exists between species, populations and individuals. Within a species or population, local environmental conditions and specific behavioural characteristics can influence disease dynamics; Bd is susceptible to heat and desiccation, and frogs inhabiting unfavourable habitats have improved survival [8,10,14,15]. Although recent progress has been made in understanding aspects of resistance to Bd, the mechanisms of immunity appear complex and much remains unknown. The post-metamorphic amphibian immune system is fundamentally similar to that of mammals, demonstrating innate and adaptive responses including specific cell-mediated and antibody responses, and immunoglobulin isotype heterogeneity [16][17][18][19]. Innate host defense mechanisms, such as antimicrobial skin peptides and symbiotic bacteria, may influence susceptibility to Bd infection [20,21].
To date, little evidence of an effective localized or systemic adaptive immune response in Bd-infected Rana, Silurana or Litoria species has been found [9,[21][22][23][24]. Activation of innate and adaptive immunity has recently been suggested to be an important component of natural Bd resistance in Xenopus laevis [25]. Knowledge of amphibian immune responses to fungal pathogens is extremely limited and the contrasting findings of the few studies available highlight the need to broaden the taxonomic focus of future immunologic studies [23].
Furthermore, there is a critical knowledge gap about why Bdsusceptible amphibians fail to mount an effective immune response: is it due to pathogen immune evasion, host immunosuppression or a combination of the two? Negligible cellular inflammation occurs in the skin of infected frogs, suggesting sporangia may evade host immune recognition due to their intracellular location within the superficial epidermis [26]. Recent genetic, stress hormone and in vitro immune function studies indirectly suggest Bd may actively suppress the host immune response [27][28][29][30][31].
The overall aim of our study was to determine whether Bd infection suppresses systemic innate and adaptive host immune responses. We used diverse methods, previously established in mammals and birds, to study immune structure and function in Bd-infected and control frogs. Methodology involved measuring 1) mass and cellularity of immune organs, 2) total and differential peripheral white blood cell (WBC) counts, 3) serum protein fraction concentrations via gel electrophoresis, 4) in vivo phytohemagglutinin (PHA) skin response, and 5) in vivo antisheep red blood cell (SRBC) antibody response. Immunization with SRBC to evaluate humoral immunity via serum and splenic hemolytic antibody production has been previously reported in three anuran species: Rana pipiens, Bufo arenarum and X. laevis [32][33][34][35]. The T-cell mitogen PHA has been used to evaluate anuran splenocyte, thymocyte and lymphocyte proliferative responses in vitro [31,[36][37][38]. Only two reports describe the in vivo PHA skin response test to assess cell-mediated immunity in adult anurans [39,40], and apart from assessment of skin peptide profiles [41], there are no reports describing innate or adaptive immunity in Litoria species. Our results showed that all of the methods, with the exception of the PHA skin test, were reliable for assessing immune function in the species studied, and that chronic Bd infection in L. caerulea caused multiple systemic immune function defects. Litoria infrafrenata failed to maintain infection with Bd after experimental exposure, but recovered frogs had lower white blood cell and serum globulin responses compared with unexposed frogs, suggesting potential costs associated with infection clearance and/or redirection of host resources towards more effective mechanisms to combat infection.

Ethics Statement
This study was carried out in strict accordance with the recommendations in the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes of the National Health and Medical Research Council. The research protocols were approved by the James Cook University Animal Ethics Committee (A1085) and the Queensland Parks and Wildlife Service (Scientific Purposes Permit WISP03866106). All blood sampling and initiation of in vivo tests were performed under tricaine methanesulfonate general anaesthesia; frogs were euthanized at the end of the study by cardiac exsanguination following tricaine methanesulfonate general anaesthesia; all efforts were made to minimise suffering throughout the study. All field locations and activities, including collection of frogs from the wild, were approved by the Queensland Parks and Wildlife Service (Scientific Purposes Permit WISP03866106).

Animals
Free-ranging clinically healthy adult individuals of the common green tree frog (L. caerulea, n = 20) and the white-lipped tree frog (L. infrafrenata, n = 20) were collected from widespread residential and semi-rural areas in and around Cairns and Townsville in far northern Queensland, Australia. The two species were selected based on their large body size, relative ease of capture, wide distribution and stable conservation status. Each frog was placed, using a new powder-free nitrile glove, into an individual plastic holding container (706956150 mm 3 ) for transport. Frogs were housed in individual plastic containers (23062306350 mm 3 ) in temperature (20-22uC) and light (12L/12D) controlled quarantine facilities at James Cook University (Cairns). Aged tap water was changed daily and frogs were fed large domestic crickets (Acheta domestica) dusted with superfine calcium carbonate (Cattlekare, Dandenong, VIC) and multivitamin powder (Reptivite, Zoo Med Laboratories Inc., San Luis Obispo, CA), ad libitum each day.
Frogs from each species were randomly assigned equally between two experimental trials. Before the trials commenced, each frog was clinically examined by a veterinarian, weighed, and a swab sample was collected from the ventral skin surfaces for determination of Bd zoospore equivalents by real-time polymerase chain reaction (PCR) analysis (James Cook University, Townsville) [42]. All swab samples were analyzed in triplicate and compared with James Cook University zoospore standards. All frogs were negative for Bd prior to commencement of the experimental trials.

Experimental Design 1: Uninfected Tree Frogs
This was designed as a pilot study to validate functional immune tests in healthy frogs from the two species. At the start of Experiment 1 (day 0), L. caerulea (n = 10) and L. infrafrenata (n = 10) were anesthetized for sampling for general immunological and hematological biomarkers and for initiation of functional tests for immune competence. Anesthesia was induced by shallow immersion in 0.20% (L. infrafrenata) or 0.25% (L. caerulea) ethyl 3-aminobenzoate methanesulfonic acid solution (tricaine methanesulfonate, Sigma-Aldrich Inc., St Louis, MO) buffered with 10 mEql 21 sodium bicarbonate solution (8.4%, Pro Care Animal Health, Dandenong, VIC).
Blood samples (250-500 ml, ,1% body weight) were collected for hematologic, plasma biochemical and serum protein electrophoretic analysis from dorsally recumbent frogs via cardiocentesis with a 1 ml syringe and 25 g needle (Terumo Corporation, Binan, Laguna). The PHA skin and the SRBC antibody response tests were also initiated at this time.
On day 7, each frog was euthanized by cardiac exsanguination following induction of anesthesia as previously described. Blood samples were collected for hematologic, plasma biochemical and serum protein electrophoretic analysis and for SRBC antibody assay. Spleen, liver and kidneys were dissected, weighed and recorded as % body weight. The spleen was immediately processed for determination of total lymphocyte count, cell viability and rosette formation by antibody-producing cells. These measurements are described below.

Experimental Design 2: Bd-infected Tree Frogs
At the start of Experiment 2 (day 0), L. caerulea (n = 10) and L. infrafrenata (n = 10) were anesthetized and blood samples were collected for analysis as described in Experiment 1. Frogs were then exposed to Bd via shallow immersion in a 25 ml bath of dilute electrolyte solution (mmol l 21 : KH 2 PO 4 1, CaCl 2 .H 2 O 0.2, MgCl 2 0.1) inoculated with 250,000 zoospores for 24 h, after which they were returned to their holding containers with aged tap water. During the exposure period, frogs were held in small individual plastic containers (5061006150 mm 3 ) with a lid to ensure continuous contact of the ventral skin surfaces with the inoculum.
Frogs were weighed and swabs collected for PCR at 10, 20, 30, 40, 50, 60, 75 and 82 d post-exposure. On day 30 post-exposure, frogs were anesthetized and blood samples collected for hematologic and plasma biochemical analysis. On day 75 post-exposure (corresponding to day 0 in the uninfected frogs from Experiment 1), Bd-exposed frogs were again anesthetized for blood sampling and for initiation of the PHA skin and SRBC antibody response tests. One Bd-exposed L. infrafrenata failed to recover postanesthesia and was excluded from the trial. On day 82 postexposure (corresponding to day 7 in the uninfected frogs from Experiment 1), each Bd-exposed frog (L. caerulea n = 10, L. infrafrenata n = 9) was euthanized by cardiac exsanguination following anesthesia. Blood, spleen, liver and kidney samples were collected as per Experiment 1. This time point for testing was chosen to ensure that frogs infected early in the trial maintained infection, but had not yet developed severe clinical signs of disease which would confound our immune function assays.

Hematologic and Plasma Biochemical Analysis
Blood samples from each frog were processed according to standard amphibian procedures [43,44]. Fresh blood smears were air dried and fixed with 100% methanol; 200 ml was collected into a 0.6 ml Microtainer pediatric lithium heparin tube (Becton and Dickinson, Franklin Lakes, New Jersey); and 150-200 ml was collected into a plain 1.0 ml microcentrifuge tube (Eppendorf AG, Hamburg), centrifuged (5,5906g for 10 min) and the supernatant decanted and refrigerated at 4uC until submission for serum protein electrophoresis. Additional blood (500-1000 ml) from the final sample on day 7 (healthy uninfected frogs, Experiment 1) and day 82 (Bd-exposed frogs, Experiment 2) was collected into a plain 1.0 ml microcentrifuge tube, allowed to clot at room temperature for 1 h, then centrifuged (5,5906g for 10 min) and the supernatant decanted and frozen at 270uC for later SRBC antibody microtiter assay.

Serum Protein Electrophoresis
Serum samples (n = 97) were submitted to a commercial reference laboratory (Gribbles Veterinary Pathology, Clayton, VIC) for determination of total protein and protein fraction (albumin, total globulins and a-1, a-2, b and c globulin) concentrations. Electrophoresis was conducted according to the manufacturer's recommendations using the semi-automated agarose gel electrophoresis system (Hydrasys, Sebia Inc., Norcross, GA) and the split protein b1 and b2 gel reagent (Hydragel 30, Sebia Inc., Norcross, GA). The resultant gel was fixed, stained and scanned using the same equipment. Densitometer laser tracings were used to measure protein fraction percentages [49], and absolute values were determined on the basis of biuret total protein measurement. The albumin-globulin (A-G) ratio was calculated by dividing the albumin value by the sum of the globulin fraction values.

PHA Skin Response Test
The PHA skin response test for T-cell mediated immunity was initiated following standard avian and mammalian procedures [50][51][52] adapted for anurans. A 0.1 ml dose of 0.5% PHA-P (Sigma-Aldrich Inc., St Louis, MO) in phosphate-buffered saline (PBS) (pH 7.4, Sigma-Aldrich Inc., St Louis, MO) was injected intradermally in the interdigital webbing of the left hind foot between the second and third phalanges with a 1 ml syringe and 27 g needle. The same volume of PBS was injected intradermally as a control in the right hind foot interdigital webbing. The thickness of each injection site was measured to the nearest 0.02 mm using manual vernier callipers (Mitutoyo Corporation, Kanagawa) immediately before and then 6, 12, 24 and 48 h postinjection. The PHA stimulation response was calculated as the change in the thickness (mm) of the PHA-injected interdigital site minus the change in thickness of the control site.

Serum SRBC Antibody Assay
The SRBC antibody response test was initiated following standard avian and mammalian procedures [51,53] adapted for anurans. Each frog was injected intracelomically with 0.5 ml of a 10% suspension of SRBC (Sigma-Aldrich Inc., St Louis, MO) in PBS with a 1 ml syringe and 23 g needle.
Total (IgM and IgY) and 2-mercaptoethanol-resistant (IgY) SRBC antibody activities were measured 7 d post-immunization using a standard microtiter method [52,53]. Saline (PBS, 50 ml) was added to each well in 96-well round-bottomed microtiter plates (Eppendorf AG, Hamburg). Serum (50 ml) was added to the first well of each row, and serial two-fold dilutions were performed across rows. Fifty ml of 0.5% SRBC suspension in PBS was then added to each well and the plates incubated at 37uC for 3 h and then overnight at room temperature. Titers were recorded as LOG 10 of the reciprocal of the highest dilution showing agglutination. To measure IgY titers, serum samples were incubated for 60 min with 0.2 M 2-mercaptoethanol (Sigma-Aldrich Inc., St Louis, MO) before dilution. All serum samples were assayed in duplicate and the same batch of SRBC was used for all immunizations and assays. Table 1. Functional test results for immune competence in uninfected and Batrachochytrium dendrobatidis-exposed Litoria caerulea and L. infrafrenata following stimulation with intradermal phytohemagglutinin (PHA) and intracelomic sheep red blood cells.

Splenic Lymphocyte Count and Cell Viability Determination
Each spleen was divided equally by weight and processed following standard avian and mammalian procedures [51][52][53]. One half was fixed in 10% neutral buffered formalin, the other was homogenised into a single cell suspension in 1.0 ml Hanks balanced salt solution (HBSS) (Sigma-Aldrich Inc., St Louis, MO) with a scalpel blade. Total splenic lymphocytes in the cell suspension were counted manually in a modified Neubauer hemocytometer at 4006 magnification.
Spleen cell viability was determined by incubating 0.2 ml spleen cell suspension with 0.3 ml HBSS and 0.5 ml trypan blue solution (Sigma-Aldrich Inc., St Louis, MO) for 10 min [54,55]. Stained (non-viable) and unstained (viable) cells were counted manually in a modified Neubauer hemocytometer at 4006 magnification. Cell viability % was calculated by dividing the mean number of unstained cells by the mean total number of stained and unstained cells.

Splenic Rosette-forming Cell Assay
This method was adapted from a combination of techniques previously described for rosette-forming [56,57] and plaqueforming [58,59] cell assays. Rosette-forming IgM-producing splenic lymphocytes sensitized to SRBC were counted in two monolayer chambers following incubation of 0.04 ml spleen single cell suspension with 0.2 ml HBSS and 0.2 ml 10% SRBC suspension at 37uC for 1 h. Rosette-forming cells (RFC) were counted manually at 506 magnification and the total number of RFC calculated per spleen.

Statistical Analysis
Independent-samples t-tests were used to compare functional tests for immune competence in healthy frogs between the two species, between healthy and Bd-exposed frogs within each of the two species, and between infected L. caerulea with low Bd loads (,1,000 zoospores) and high Bd loads (.1,000 zoospores). Variables analyzed included skin PHA stimulation, serum IgY and combined serum IgM/IgY titers, total splenic cell count, splenic RFC count, splenic cell viability, ratios of kidney, liver and spleen to body weight, and various hematologic, plasma biochemical and protein electrophoretic parameters.
Paired-samples t-tests were used to compare various hematologic, plasma biochemical and protein electrophoretic variables pre-and post-immune stimulation within each experimental group (healthy and Bd-exposed) for each species.
The software package PASW Statistics (Version 18, 2009, SPSS Inc., Chicago, IL) was used for all analyses, and statistical significance was set at #0.050 in all cases.

Bd-infected Litoria caerulea
At day 75 post-exposure when immune function tests were initiated, 100% of exposed L. caerulea (10/10) tested positive for Bd on PCR. Zoospore counts per sample ranged from 24 to . 10,000; six frogs had low Bd loads (,1,000 zoospores) and four had high Bd loads (.1,000 zoospores). One frog with .10,000 zoospores showed mild clinical signs of disease including lethargy and cutaneous erythema; all other frogs were clinically normal.
Mean splenic total lymphocyte and RFC counts, splenic cell viability and liver-body weight ratio post-immune stimulation were lower in Bd-infected L. caerulea (n = 10) compared with the uninfected frogs (n = 10) ( Table 1). Mean responses to all of the functional tests for immune competence did not differ between infected frogs with low and high Bd loads (P.0.050 in all cases).
Following immune stimulation of Bd-infected L. caerulea, mean thrombocyte count decreased and A-G ratio increased; all other hematologic, plasma biochemical and serum protein electrophoretic parameters did not differ significantly pre-and poststimulation (Tables 2 and 3).  (Table 2).

Bd-infected Litoria infrafrenata
At day 75 post-exposure, 0% of the nine exposed L. infrafrenata tested positive for Bd on PCR. Five frogs had low Bd loads at either day 10 (n = 3) or 20 (n = 2) post-exposure, but all of these were negative from day 30 onwards post-exposure, and the other four frogs tested negative throughout the experiment.
Mean skin PHA response and kidney-body weight ratio were lower in exposed uninfected L. infrafrenata compared with healthy uninfected individuals post-immune stimulation; mean responses to the other immune function tests did not differ between the two groups (Table 1).
Following immune stimulation of Bd-exposed uninfected L. infrafrenata, mean Hb concentration and RBC and thrombocyte counts decreased, while glucose concentration increased; all other hematologic, plasma biochemical and serum protein electrophoretic parameters did not differ significantly (Tables 4 and 5).
Compared with Bd-exposed uninfected individuals, immune stimulation of healthy uninfected L. infrafrenata caused a greater  (Table 4).

Uninfected Tree Frogs
Functional immune competence test results for uninfected L. caerulea and L. infrafrenata are presented in Table 1. Following immune stimulation, mean total splenic RFC (P = 0.006) and liver to body weight ratio (P = 0.003) were greater in L. caerulea, while mean spleen to body weight ratio was greater in L. infrafrenata (P = 0.001). Mean responses to the other immune function tests did not differ between the two species.
Pre-and post-immune stimulation hematologic values for uninfected L. caerulea and L. infrafrenata are presented in Tables 2 and 4 respectively. Following immune stimulation of L. caerulea, PCV, Hb, MCH and MCHC decreased, while total WBC, lymphocyte and monocyte counts and polychromasia % increased. In L. infrafrenata, total RBC and thrombocyte counts decreased following immune stimulation, while total WBC, neutrophil and monocyte counts, neutrophil-lymphocyte ratio, and polychromasia % increased.
Pre-and post-immune stimulation plasma biochemical and serum protein electrophoresis values for uninfected L. caerulea and L. infrafrenata are presented in Tables 3 and 5 respectively. Five protein fractions were defined in all samples submitted for electrophoresis. Following immune stimulation of L. caerulea: total globulins along with a-1 and c globulin fraction concentrations increased, while the A-G ratio decreased; none of the plasma biochemical values changed significantly. In L. infrafrenata: plasma CK increased while glucose decreased post-immune stimulation; total protein, albumin, total globulins and a-1 globulin concentrations, and A-G ratio, all increased.

Discussion
Our results provide the first direct evidence suggesting that Bd infection causes multiple in vivo systemic immune function defects in a susceptible amphibian host, which likely enables disease development. Antigenic stimulation of L. caerulea chronically infected with Bd resulted in lower splenic lymphocyte, WBC, serum protein and immunoglobulin responses compared with uninfected frogs. Although L. infrafrenata failed to maintain infection with Bd after experimental exposure, recovered frogs also had reduced WBC and serum globulin concentrations compared with unexposed frogs. Our work expands on recent studies showing in vitro lymphocyte proliferation was impaired by Bd cells and supernatants [31].
We successfully used diverse methods, many of which have not previously been used outside of applied veterinary research, to  (Fig. 2A), a-1 globulin (Fig. 2B) and c globulin (Fig. 2C) concentrations (g l 21 ) pre-and post-immune stimulation in healthy uninfected (n = 10) and Batrachochytrium dendrobatidis-infected (n = 10) Litoria caerulea. Bars are mean 6 SEM. *P,0.050 within each group. doi:10.1371/journal.pone.0107284.g002 Table 4. Pre-and post-immune stimulation hematologic values for uninfected and Batrachochytrium dendrobatidis-exposed but uninfected Litoria infrafrenata.  Table 5. Pre-and post-immune stimulation plasma biochemical and serum protein electrophoretic values for uninfected and Batrachochytrium dendrobatidis-exposed but uninfected Litoria infrafrenata. assess general innate and adaptive host immune competence in two tree frog species with and without Bd infection. All of the methods, with the exception of the PHA skin test, appeared to be reliable, and we present comprehensive baseline immune function data for the two frog species. Following stimulation of uninfected L. caerulea, immune organ activation resulted in circulating WBC and protein responses, total (IgM and IgY) and IgY-specific serum antibody production, and splenic lymphocytic IgM-specific antibody production. These indicate active cellular and humoral responses consistent with those seen in other taxonomic groups [51,52]. The change in RBC indices (reduced PCV, Hb, MCH and MCHC, concurrent with increased polychromasia) were not unexpected due to repeated blood sample collection within a short period of time, and indicate an adequate bone marrow regenerative response to relative anemia [48]. This range of data, combined with additional splenic cell measurements, provides comprehensive baseline immune function data for this species.

Status
Compared with uninfected L. caerulea, chronic Bd infection in this species caused a significant reduction in total spleen cell concentration, splenic cell viability, splenic lymphocyte antibody production and liver-body weight ratio. The WBC, lymphocyte, monocyte and serum protein responses to immune stimulation were also reduced. All of these changes indicate an impaired ability of Bd-infected L. caerulea to respond adequately to antigenic stimulation. Inflammatory cells do not migrate to the sites of infection [26], hence the reductions in circulating white cell numbers in Bd-infected L. caerulea were most likely caused by defects in production and/or survival of these immune cells. The significance of the A-G ratio increase without a concurrent change in any of the five protein fractions remains unclear but may indicate a subtle serum protein response. Infected frogs had a lower Hb concentration decrease and polychromatophilic RBC response following immune stimulation. Although these two changes are likely to be a direct regenerative response to recent blood collection, the lower magnitude of these responses in the Bdinfected frogs may indicate reduced organ response to physiologic stimulation. The decreased mean thrombocyte count is most likely a response to recent blood collection and relative anemia, but may also represent depressed bone marrow function. Mean thrombocyte count also decreased in the uninfected frogs but this was not statistically significant.
Active cellular, humoral and splenic responses in unexposed L. infrafrenata were generally consistent with those in the uninfected Figure 3. White blood cell counts of Litoria infrafrenata. Total white blood cell (WBC) (Fig. 3A) and neutrophil (Fig. 3B) counts (x10 9 l 21 ) preand post-immune stimulation in healthy unexposed (n = 10) and Batrachochytrium dendrobatidis-exposed but uninfected (n = 9) Litoria infrafrenata. Bars are mean 6 SEM. *P,0.050 within each group. doi:10.1371/journal.pone.0107284.g003 Figure 4. Serum a-1 globulins of Litoria infrafrenata. Serum a-1 globulin concentration (g l 21 ) pre-and post-immune stimulation in healthy unexposed (n = 10) and Batrachochytrium dendrobatidis-exposed but uninfected (n = 9) Litoria infrafrenata. Bars are mean 6 SEM. *P,0.050 within each group. doi:10.1371/journal.pone.0107284.g004 L. caerulea, and in other taxonomic groups [52,53]. Notable interspecies differences in our results include the neutrophilic versus lymphocytic WBC response and the hyperglycemia and elevated CK post-immune stimulation in L. infrafrenata. Neutrophil reference values in L. infrafrenata vary with season, and are significantly higher compared with L. caerulea [44]. Furthermore, the two species differ in temperament, with L. caerulea generally calm and tolerant of handling, and L. infrafrenata often exhibiting clinical signs of stress associated with handling and confinement. This is likely to account for the neutrophilic WBC response and the increased glucose and CK values in L. infrafrenata. The reduced RBC count concurrent with increased polychromatophilic response again was not unexpected due to repeated blood sample collection, and indicate an adequate bone marrow regenerative response.
All Bd-exposed L. infrafrenata failed to maintain infection despite identical experimental conditions to those for L. caerulea, and the five of nine frogs that tested positive early in the trial selfcured. The sampling protocols were completed despite the loss of infection, and results showed immunologic effects on some systemic responses resulting from Bd exposure. Post-immune stimulation WBC, neutrophil, polychromatophilic RBC, thrombocyte and a-1 globulin concentrations, and skin PHA response, were reduced in the Bd-exposed L. infrafrenata compared with uninfected frogs. However, immune stimulation caused similar total and IgY-specific serum antibody, splenic lymphocytic IgMspecific antibody, and splenic cellular, responses in both uninfected and Bd-exposed L. infrafrenata. Variations in RBC indices and plasma glucose were not unexpected as previously discussed.
In previous studies using the model species Silurana tropicalis, Bd infection appeared to cause down-regulation of some immune genes including those associated with Toll-like receptors, complement pathways, and B-and T-lymphocytes [27,28]. Recently, soluble factors in Bd culture supernatant inhibited in vitro lymphocyte proliferation assays, but did not reduce macrophage activity [31]. Other studies found higher concentrations of urinary corticosterone metabolites and plasma corticosterone, in Bdinfected L. wilcoxii and L. caerulea respectively, indicating a physiological stress response [30,62].
Stress hormones are known to alter normal WBC distribution, and peripheral leukocyte profiles consistent with a classical mammalian stress-related response include a relative neutrophilia, lymphopenia and eosinopenia [63]. Peripheral neutrophilic and eosinopenic responses have previously been reported in Bdinfected larval anurans (Rana catesbeiana), although lymphocyte abundance did not change [64]. Conflicting findings have been reported in post-metamorphic anurans: juvenile L. chloris showed relative peripheral neutropenic, eosinopenic and basophilic responses to Bd infection [20], while L. caerulea showed a lymphopenic response [62,65]. The amphibian leukocyte response to stress varies according to species and many other intrinsic and extrinsic factors, including season and sex [44,66,67], making interpretation of leukocyte profiles alone difficult.
Previous laboratory and field studies show that L. caerulea is highly susceptible to Bd [11,12,68], but little data is available for L. infrafrenata. Our results indicate that L. infrafrenata is a naturally resistant host. Pre-exposure data collected from L. infrafrenata did not differ greatly from L. caerulea, hence our results do not suggest a clear mechanism for greater natural resistance in L. infrafrenata, which may be related to localized and/or innate immune defense mechanisms that we did not measure. The skin peptide profiles of L. infrafrenata and L. caerulea differ greatly [41] and this could partly explain interspecies differences in host resistance, although L. infrafrenata produces no known major antibiotic peptides compared with at least five that have been identified in L. caerulea [41].
The reduced WBC, neutrophil and globulin concentrations in exposed L. infrafrenata may indicate direct host immune system costs associated with infection clearance in frogs challenged postmetamorphosis and/or redirection of host resources away from systemic adaptive immune responses towards alternate immune mechanisms involved in combating infection. Larval common toads (Bufo bufo) experimentally exposed to low Bd doses usually died at or soon after metamorphosis without detectable infections, suggesting fitness costs attributable to exposure, control and clearance in the absence of extensive pathogen proliferation [69]. However, exposed L. infrafrenata were still able to produce serum antibody, and splenic cellular and antibody, responses of similar magnitude to uninfected frogs following immune stimulation. This suggests that adaptive immune responses may also play an important role in Bd-resistant host species.
Despite our findings that Bd causes multiple defects in systemic immune function in L. caerulea, an Australian survey found only 6.5% (13/199) of frogs with severe chytridiomycosis had concurrent acute secondary infectious diseases [68]. The low incidence of secondary infections may indicate that systemic immune suppression is not generalized and distinct components remain effective against opportunistic pathogen invasion, and/or that systemic immune response capability is not reduced below a critical point until late in the disease process. Total and IgYspecific serum antibody responses did not differ between uninfected and infected frogs of either species, suggesting that serum immunoglobulin response as a component of systemic adaptive immunity is not suppressed in chronic chytridiomycosis and that it may not be an important component of disease resistance. These findings further support the hypothesis that localized and/or innate immune defense mechanisms may be key factors.
Skin PHA stimulation was minimal and variable in contrast to other taxa and this test was not found to be a sensitive indicator of immune function in either of the two Litoria species in our study. There was a slightly greater response in L. infrafrenata compared with L. caerulea, but this was still minimal compared with other species and quite variable. The PHA skin response in adult R. pipiens was also found to be highly variable and less sensitive for detecting pesticide-related immune suppression compared with hemocyanin-specific antibody and whole blood chemiluminescence tests [39], although in the cane toad (Rhinella marina) the test was found to reliably quantify immune response in an assay validation study [40]. Splenic cell viability assessed by trypan blue exclusion resulted in lower mean viable cell counts in both of the Litoria species in our study compared with avian and mammalian species, but this test was still useful for comparing immune function between Bd-infected and uninfected frogs. Future adaptations of this method to improve viable cell harvest should include the use of an amphibian-specific isotonic diluent; mammalian isotonic solutions such as Hanks balanced salt solution are hypertonic to amphibian cells [70] and may cause cell damage. Although bisection of the spleen was necessary to perform multiple splenic tests, improved cell viability may also be achieved by processing the whole spleen for one single method.
A limitation associated with our experimental design is that the two experimental trials were not conducted simultaneously. The trial in uninfected frogs commenced as a pilot study to validate methodology prior to application in the trial with Bd-infected frogs, but insufficient specimens were available to include a negative control group in the infection trial. This limitation was largely compensated for by 1) standardizing timing of sample collection relative to initiation of immune tests, 2) ensuring experimental procedures were performed by the same person (SY) using identical analytic equipment, reagents and methodology, and 3) maintaining identical experimental laboratory and husbandry conditions. Additionally, a large concurrent experiment in our laboratory showed that immunity did not decrease in uninfected L. caerulea held in captivity and blood sampled three times over this period -with the exception of a mild but significant decrease in WBC and lymphocyte counts on day 30, no hematologic immune biomarkers varied significantly after 60 or 125 days [24]. Our baseline immune function data provide a valuable tool for progressing future immunologic and Bd pathogenicity studies in amphibians. General mechanisms of fungal immune suppression include inducing anti-inflammatory cytokines, decreasing pro-inflammatory cytokines and complement evasion [71]. Further work on understanding the mechanisms of Bd immune suppression is needed and may involve exposing frogs directly to identified pathogen-secreted factors and assessing immune function using the in vivo methods that we have shown here to be sensitive indicators.
Our results may explain the lack of an adaptive immune response to infection in some species [22,24,72], and suggest that if vaccine development or other immune modulation is attempted in future trials, it will be necessary to better understand the mechanisms of immune suppression in order to overcome it. While many of the immune tests we describe here will be valuable for use in future immunologic studies, the total and differential WBC and serum electrophoresis measurements, along with antigen-specific serum antibody assays, have particularly wide potential future applications as repeatable, ante-mortem methods for assessment of amphibian immune function.