Effects of parasitic infection and reproduction on corticosterone plasma levels in Galápagos land iguanas, Conolophus marthae and C. subcristatus

Abstract In vertebrates, one main feature of stress response is the release of glucocorticoids (corticosterone in reptiles), steroid hormones whose synthesis is regulated by the hypothalamic–pituitary–adrenal axis (HPA). In the Galápagos Islands, populations of land iguanas are differentially impacted by a tick‐transmitted apicomplexan hemoparasite of genus Hepatozoon, which could cause diseases and ultimately reduce fitness. Using competitive enzyme‐linked immunosorbent assays (ELISA), we examined baseline plasma corticosterone levels of two syntopic and highly parasitized populations of the land iguana species Conolophus marthae and C. subcristatus in Wolf volcano (Isabela Island). We also used a poorly parasitized population of C. subcristatus from the same island (Bahia Urbina) as a reference. To better interpret the observed glucocorticoids patterns, we simultaneously performed the count of white blood cells (WBCs) in all individuals and investigated the reproductive status of females. We did not find evidence in support of either a positive or negative relationship between the tick load, hemoparasite infection, and glucocorticoid plasma concentration in C. marthae and C. subcristatus at Wolf volcano. The comparison between parasitized and non‐parasitized sites (V. Wolf and Bahia Urbina) would instead suggest an inverse relationship between corticosterone and parasites. Our findings support association between corticosterone plasma levels and reproduction.

When an internal and/or external environmental change occurs, the hypothalamic-pituitary-adrenal (HPA) axis stimulates the secretion of GCs by the adrenal glands to help organism in responding to stressful conditions (McEwen & Wingfield, 2003;Wingfield, 2013;Wingfield & Ramenofsky, 1999;Wingfield & Romero, 2001;Wingfield & Sapolsky, 2003;Wingfield et al., 1997). Glucocorticoids are the final product of the HPA axis and participate in the control of homeostasis, activating immediate life-saving processes (Romero, Dickens, & Cyr, 2009).
Normally, short-term glucocorticoid releases are helpful for individual survival because they stimulate both physiologic and behavioral emergency mechanisms, exclusively oriented to overcome the perturbation (Wingfield & Romero, 2001;Wingfield & Sapolsky, 2003). However, long-term activation of the stress response with chronically elevated GCs concentrations could be prejudicial. Prolonged elevated concentrations could expose the individual to a long-term overstimulation of survive mechanisms with consecutive inhibition of many fundamental functions including reproduction, growth, and immunocompetence (Dallman & Bhatnagar, 2001;Dhabhar, 2000;Sapolsky, 1987;Sapolsky et al., 2000;Wingfield et al., 1997). Therefore, persistent high levels are usually detrimental to health, as they may augment the stress-related disease and pathology ).
The effects of parasites are energetically demanding (e.g., reduction of growth and long-term survival; Madsen, Ujvari, & Olsson, 2005), which is why the increase in CORT concentrations is necessary to better cope with the challenges of parasitism (Raouf, Smith, Brown, Wingfield, & Brown, 2006). Although hormonal alterations accompany parasitic infection, the interpretation of stress response using CORT requires always a careful interpretation. In fact, CORT plasma levels vary not only in response to stressor-dependent factors (duration and intensity), but they are also susceptible to individual-dependent factors (e.g., reproductive status) which should be considered when GCs are used as physiologic indices of condition in wild populations (Breuner, Wingfield, & Romero, 1999;Moore & Jessop, 2003;Romero, 2002).
The role of CORT during reproduction is very complex. If, on one hand, CORT elevations due to a stress condition shift the individuals into an "emergency life-history stage" in which they reduce or even abandon their parental activities as reproduction (Greenberg & Wingfield, 1987;Wingfield et al., 1998); on the other hand, an increasing number of studies report positive associations between reproduction and CORT as a result of the involvement of such hormone in the energetically demanding breeding/nesting activities (Rubenstein & Wikelski, 2005;Tyrrell & Cree, 1998), or specifically in oogenic processes (Wilson & Wingfield, 1992). Although a cause-effect link exists between corticosterone and reproduction, the interpretation of such relationship may be difficult, as different factors influence glucocorticoids production.
In this study, we related plasma CORT concentration to levels of tick load and hemoparasitic infections in two populations of the two Galápagos land iguana Conolophus marthae (CM) and C. subcristatus  (Al-Ghamdi et al., 2011;Cook, Smit, & Davies, 2009;Roca & Galdón, 2010;Sloboda, Kamler, Bulantová, Votýpka, & Modrý, 2007). Generally, they are transmitted by ticks of the genus Amblyomma (Baneth et al., 2003;Vilcins, Ujvari, Old, & Deane, 2009), which occur at high density in Wolf volcano (Schatz, 1991). However, Hepatozoon was also PCR-detected in marine iguana blood extracted from the mosquito Aedes taeniorhynchus (Bataille et al., 2012), suggesting that transmission (the dynamics of which are still unclear) may occur via more than one vector. Since 2005 (when we started investigating CM), tick abundance at Wolf volcano has been steadily high. A comparably high density of ticks affecting iguanas is not found anywhere else in Galápagos and high tick density is associated with high prevalence of Hepatozoon at Wolf volcano (Fulvo, 2010), suggesting that, in Galápagos, ticks play a major role in the transmission of Hepatozoon to iguanas.
Previous studies of the relationship between ecto/hemoparasites and CORT in reptiles returned contradictory results because of the many causal relationships linking glucocorticoid levels to different independent factors. For this reason, experimental studies aimed at discriminating cause and effect have been increasingly performed. Given the conservation concern of the focal species, we were not allowed to use practices such as artificial infection or hormonal manipulations, considered invasive by the Galápagos National Park, the governmental authority that administrates biodiversity in Galápagos. Nevertheless, we used a comparative approach, although correlative, to study the role of parasites on endocrine activity, accounting for a factor related to CORT: reproductive state. In fact, according to the energy mobilization hypothesis that describes glucocorticoids increasing during period energetically demanding such as the gravidity (Romero, 2002), we hypothesized an increase in CORT levels in reproductive animals.
Thus, under the classical hypothesis that high parasite load is often associated with elevated glucocorticoid levels (Maier & Watkins, 1999), we tested whether CORT plasma levels increase with the intensity of infection by ectoparasites and/or Hepatozoon using two approaches. We first investigated CORT and parasite load in the two species at Wolf volcano (W), with females at different reproductive state. Although a replication of a site analogous to Wolf volcano would be recommendable, such a condition cannot be met in Galápagos. Thus, in a second approach, we attempted at a cross-site comparison by comparing two populations of CS exhibiting different levels of parasite infection (highly parasitized versus non-parasitized). To do so, we selected a non-parasitized, and at a non-reproductive-state population of CS (Bahia Urbina, BU). Of course, by selecting a different site, we might introduce local confounding factors that may affect CORT levels. Aware of this possible bias, we considered BU the best reference as it is the CS population geographically closest to Wolf volcano, in Isabela Island. As additional peripheral symptoms that characterize infection and correlate with it, we also investigated the heterophils/lymphocytes (H/L) ratio, commonly used as hematologic marker of stress in several reptile species (Duggan, 1981;Moberg, 1985;Xuereb, Row, Brooks, MacKinnon, & Lougheed, 2012).

| Ethic statement
Animal manipulation and blood sampling were performed according to a protocol that minimized animal stress, in accordance with the European Community guidelines and with the approval of the

| Field sites and sampling
The study was conducted in two different areas of Isabela  During all field sessions, 2 ml of blood was collected from the caudal vein of each individual using a 5-ml heparinized syringe. Blood samples were collected within 3-5 min from capture, under the assumption that this represent a sufficiently short time for CORT levels to represent baseline concentrations (Romero & Romero, 2002). Previous studies have shown that this time interval is sufficiently short to prevent that plasma levels of CORT be biased by capture stress (Cash, Holberton, & Knight, 1997;Romero, 2004;Romero & Wikelski, 2001;Sapolsky et al., 2000;Tyrrell & Cree, 1998;Wingfield et al., 1997).
We placed approximately 10 microliters of blood on the top of a slide and created a smear. Blood smears were air-dried. Blood samples were placed on ice immediately after collection and later centrifuged for 2 min at 135 g to separate plasma. Plasma was kept at −10°C while in the field and then stored at −80°C until it was processed. Each iguana was weighed and snout-vent length (SVL) was measured. The body condition index (BCI) was then estimated as the ratio of body mass/snout-vent length (SVL) 3 × 10 6 (the ratio was multiplied by 10 6 to reduce the number of decimals). This index has been already used for iguana species (Costantini et al., 2009;Laurie, 1989;Romero & Wikelski, 2001;Wikelski & Trillmich, 1997
We determined the parasitemia recording the number of erythrocytes infected by Hepatozoon (so far the only known hemoparasite infesting Conolophus spp., Fulvo, 2010) observed in 20 min, the time required to totally analyze approximately 10,000 erythrocytes and considered sufficient to obtain information about the intensity of infection (Valkiūnas, Iezhova, Križanauskienė, Palinauskas, & Bensch, 2008). Blood smears were scanned by the same investigator, at the same pace. If no hemoparasites were observed after this time, the individual was classified as uninfected. Trial sessions were blindly conducted prior to data collection to ensure consistency and repeatability.

| Hormonal analysis
We determined plasma levels of CORT by competitive enzymelinked immunosorbent assays (ELISA). All ELISA immunoassays were performed at the Laboratory of Clinical Biochemistry (Tor Vergata University Hospital). On the whole, for analyzing CORT of all blood samples, we used kits ELISA (KA0468) pre-coated with a polyclonal antibody. We used 10 μl of plasma diluted with 90 μl of assay buffer.
The detection limit was established to be 0.28 ng/ml. The intra-assay variation was 4.1% and the inter-assay variation 10.1%.
All samples were assayed in duplicate and randomly distributed between plates. All assays were performed according to the instructions of the kit manufacturers. Generalized linear models (GLZs) with an identity-link function were performed to evaluate which factors better explained the variation of CORT plasma levels. Females and males were analyzed separately in GLZ models, as in vertebrates sex differences in adrenocortical activity have been described (Kirschbaum, Wüst, & Hellhammer, 1992;Kudielka & Kirschbaum, 2005). For CORT, to ensure proper evaluation of interactive effects, for both sexes we considered two different models. Model 1 included only CM and CS from Wolf volcano, whereas Model 2 did not include CM and included both populations of CS (W+BU).

| Statistical analysis
For males and females, all models included species (or site for the model including only CS populations) as categorical factor and body condition index, parasitemia, ticks number, and H/L ratio as covariates.
For females, reproductive state was also included as categorical factor (yes or no). We tested also for the interaction between species and reproductive state.

| Corticosterone plasma levels in females
The reproductive status of females is reported in Table 3.
In females of CM, CORT plasma levels ranged from 0.16 to 74.11 ng/ml, whereas in CS it ranged from 0.22 to 158 ng/ml (mean and medians are reported in Table 4).

| Corticosterone plasma levels in males
In males CM, CORT plasma levels ranged from 0.21 to 13.9 ng/ml, whereas in CS it ranged from 0.19 to 153 ng/ml (mean and medians are reported in Table 4). In both species, no differences among

| DISCUSSION
In this study, we used CORT plasma levels to investigate possible impacts of ticks and Hepatozoon on the stress physiology of Galápagos land iguanas. Despite the fact that an inverse relationship between the level of hemoparasite/ectoparasite infection and glucocorticoid plasma concentration was previously found in iguanas (Hanley & Stamps, 2002), we did not find evidence in support of a such a relationship in the populations of C. marthae and C. subcristatus from Wolf volcano. In fact, in both species from the site, we did not observe significant differences in baseline CORT levels or  (Nardini, Leopardi, & Bielli, 2013).
However, it is also possible that the hemo-and non-hemo-parasitized individuals show similar low levels of CORT because they are both highly impacted by the chronically high tick density that affects both hemo-and non-hemo-parasitized individuals. Wolf volcano is the site where ticks occur at the highest density in Galápagos islands, with the minimum tick load per individual being in the order of tens. This load could be high enough to mask a correlation between CORT levels and number of ticks. Possibly, these results can be considered in the light of the immune redistribution hypothesis (Braude, Tang-Martinez, & Taylor, 1999;Maier & Watkins, 1999), which would allow to explain low levels of CORT in highly tick-infected iguanas (Wolf volcano) as a way for the immune system to focus on internal parasites such as hemogregarines (Hanley & Stamps, 2002 Although CORT did not correlate with parasites or body condition index, it positively correlated with reproductive condition in females. CORT levels appeared elevated in females carrying eggs (small, not fully formed or fully formed) in both land iguana species. Also in marine iguanas, CORT was mostly elevated during the gestation and nesting period before eggs were laid and it declined significantly immediately after egg-laying (Rubenstein & Wikelski, 2005). In our study, F I G U R E 3 Corticosterone and reproductive status. Vertical bars denote 95% confidence intervals F I G U R E 4 Corticosterone variation in relation to the interactive effect between species and reproductive status. Vertical bars denote 95% confidence intervals F I G U R E 5 Corticosterone variation in Conolophus subcristatus in relation to site. Vertical bars denote 95% confidence intervals the observed increase in CORT in reproductive females could reflect the energetic demands of reproduction (Wingfield, 1988 Although all non-reproductive females showed similar CORT baseline levels without a difference between species, in reproductive females we observed a difference in favor of CS, in which most individuals (90%) showed mature eggs. Thus, in case no biologic difference in the physiological process of CORT production between the two congeneric species exists, the CORT increase in reproductive CS females with fully formed eggs could also be associated with the metabolic change specifically required for egg development, as already described in other reptiles (Wilson & Wingfield, 1992). Indeed, generally a positive association between reproductive state and glucocorticoids level has been observed for many egg-laying vertebrates (Silverin & Wingfield, 1982;Wack, Fox, Hellgren, & Lovern, 2008;Wilson & Wingfield, 1992). Many studies (Grassman & Crews, 1989;Moore & Jessop, 2003;Taylor, DeNardo, & Jennings, 2004) demonstrated that elevated CORT plasma levels facilitate reproduction by mobilizing energy stores for egg production processes such as vitellogenesis, oocyte maturation, and ovulation. This hypothesis awaits further investigation.

ACKNOWLEDGMENTS
Data presented here result from a long-term program in collaboration between the Galápagos National Park and the University of Rome "Tor Vergata," aimed at the conservation of Galápagos land iguanas.
We thank the Parque Nacional Galápagos for the support that given in this study. We gratefully thank Sonosite Inc. for assisting us with portable ultrasound devices.