Non-invasive monitoring of glucocorticoid metabolite concentrations in urine and faeces of the Sungazer (Smaug giganteus)

Study site and animals

The study was conducted at the SANBI National Zoological Garden (NZG), Pretoria, South Africa (25.73913°N, 28.18918°E) from the 24th of November 2017 to the 5th of December 2017. The study animals, consisting of one male (M1:291 g) and two females (F1: 295 g) and F2: 344 g), were housed in individual enclosures within the Reptile and Amphibian Section of the NZG. Individuals were separated by a 1.5 m high wall, which resulted in study animals not being in visual contact with one another. Each enclosure (2 m × 1.5 m) was covered in coarse river sand and included an artificial burrow constructed from fiberglass, UV-light and a water bowl with water available ad libitum. The light regime (13 L: 11 D) and humidity (range: 44–50%) were kept constant throughout the study period. A combination of meal worms and fresh, green vegetables were provided daily to all individuals. Prior to the start of the study, all individuals were given a two-week acclimatization period to the new enclosure and presence of researchers. The limited number of individuals used during the study reflects the availability of study animals in a suitable setting, as well as the difficulty in receiving approval to conduct research on vulnerable and endangered species.

Sample collection and ACTH challenge

In reptiles, urine and faeces can be excreted in unison, though not mixed (Fig. S2; Singer, 2003); urine is a white, solid substance, compared to the dark, solid faecal component, which allows for the separation of the two matrices with limited levels of cross-contamination (Kummrow et al., 2011). During the entire monitoring period, collected urine and faeces were separated during collection, and the two parts placed into separate 1.5 ml microcentrifuge tubes, sealed, and immediately stored at −20 °C until further processing. Following a two-week acclimatisation period, enclosures were checked for urine and faecal samples during the active period of S. giganteus (6 am–6 pm), for seven days. Cages were checked hourly to limit the effect of bacterial and environmental degradation of urine and faecal samples. In the morning hours of the eighth day all three individuals were injected intramuscularly with 0.45 µg synthetic ACTH g⁻¹ bodyweight (SynACTH^®, Novartis, South Africa Pty Ltd) in a 100 µl saline transport. This ACTH dose was chosen as it has been used successfully by a number of studies conducted on amphibian species such as the Fijian ground frog (Platymantis vitiana, Narayan et al., 2010), tree frog (Hypsiboas faber, Barsotti et al., 2017) and the American bullfrog (Rana catesbeiana, Hammond et al., 2018) to evoke a stress response. Subsequently, the individuals were released back into their individual enclosures, with faecal and urine collection continuing until day 15 of the study. The study was performed with the approval of the National Zoological Garden’s Animal Use and Care Committee (Reference: P16/22).

Steroid extraction in urine and faecal samples

Urine and faecal samples were lyophilized, pulverized and sifted through a mesh strainer to remove any undigested material, resulting in a fine faecal and urine powder (Heistermann, Tari & Hodges, 1993). Subsequently, 0.050–0.055 g of the respective urine and faecal powder was extracted with 1.5 ml 80% ethanol in water. After vortexing for 15 min, the suspensions were centrifuged for 10 min at 1,600 g and the resulting supernatants transferred into new microcentrifuge tubes and stored at −20 °C until analysis.

Enzyme immunoassay analyses

Depending on the original matrix, steroid extracts were measured for immunoreactive faecal glucocorticoid metabolite (fGCM) or urinary glucocorticoid metabolite (uGCM) concentrations using four different EIAs: (i) an 11-oxoaetiocholanalone (detecting fGCMs with a 5β-3α-ol-11-one structure), (ii) a 5α-pregnane-3β-11β-21-triol-20-one (measuring 3β-11β–diol-CM), (iii) a tetrahydrocorticosterone, and (iv) corticosterone EIA. Details about assay characteristics, including full descriptions of the assay components including cross-reactivities, can be found in Möstl et al. (2002) for the 11-oxoaetiocholanalone EIA, Touma et al. (2003) for the 5α-pregnane-3β-11β-21-triol-20-one, Palme & Möstl (1997) for the corticosterone EIA and in Quillfeldt & Möstl (2003) for the tetrahydrocorticosterone EIA. Assay sensitivities, which indicates the minimum amount of respective immunoreactive hormone that can be detected at 90% binding, as well as the intra- and inter-assay coefficients of variation of high and low quality controls for each EIA is shown in Table 1. Serial dilutions of extracted faecal and urine samples gave displacement curves that were parallel to the respective standard curves in the two assays of choice (5α-pregnane-3β-11β-21-triol-20-one and tetrahydrocorticosterone EIAs), with a relative variation in slope of <4%. All EIAs were performed at the Endocrine Research Laboratory, University of Pretoria, South Africa, as described previously (Ganswindt et al., 2002).

Data analysis

A total of six faecal and urine samples were analyzed for each individual. Individual median fGCM and uGCM concentrations from pre-injection samples were calculated, reflecting individual baseline concentrations. To determine the effect of the ACTH injection on the HPA axis, the fGCM and uGCM concentrations from post-injection samples were converted to percentage response, by calculating the quotient of individual baseline and related fGCM/uGCM samples. In this regard, a 100% (1-fold) response represents the baseline value.

Furthermore, the mean absolute deviation (MAD) was calculated for the baseline sample set (pre-injection). The MAD of the particular dataset shows the average distance between each baseline period data point and the calculated mean thereof, which represents the variability of the baseline samples collected. Thus, the lower a MAD value is for a specific EIA, the more stable the assay. Here, the individual baseline uGCM/fGCM concentration was subtracted from all pre-injection fGCM/uGCM values for each EIA-specific data set. The differences were noted as absolute values and the mean of the absolute values calculated, representing the MAD value for each EIA. The MAD values were converted to a percentage deviation value (MAD/Baseline Value*100) to allow for the comparison between EIAs. To determine the effect of the ACTH injection, the absolute change in fGCM and uGCM concentration was determined by calculating the quotient of baseline and post-injection peak fGCM and uGCM samples. MAD values below 15% were regarded as preferable.

The most appropriate EIA for measuring fGCM and uGCM concentrations in the species was chosen by comparing (1) the highest post-injection signal and (2) lowest MAD values observed. Values are given as mean ± standard deviation (SD) where applicable. Analytical statistics and graphical designs were performed using R software (R 3.2.1; R Development Core Team, 2013).

Defecation rate and MAD results

The average defecation rate (time between defecation events) showed considerable variation between individuals and matrices (Table 2).

The percentage MAD values were considerably lower in all EIAs when analyzing faecal (range: 3.17–15.67%) compared to urine (range: 13.31–56.52%) samples. For faeces, although the corticosterone EIA showed the lowest average percentage MAD value (mean ± SD = 8.37 ± 5.54%), the remaining three EIAs all showed comparable low average MAD levels (range: 9.95–11.02%). In contrast, all four EIAs showed high average percentage MAD levels in urine, with the 11-oxoaetiocholanalone EIA having the lowest average MAD value (mean ± SD = 17.17 ± 5.77%).

Faecal glucocorticoid metabolite analyses

All four EIAs showed a considerable response (178.46–744.84%) in fGCM concentration following the ACTH injection in all three study animals (Table 3). For F1, the 5α-pregnane-3β-11β-21-triol-20-one and tetrahydrocorticosterone EIAs showed the highest response, exceeding 450%, at 24 h post ACTH administration (Tables 2 and 3, Fig. 1A). Following this, fGCM concentrations did not return to baseline levels during the monitoring period. For F2, the tetrahydrocorticosterone and corticosterone EIA performed best, with responses exceeding 300%, at 24 h post ACTH injection (Tables 2 and 3, Fig. 1B). The fGCM concentrations returned to baseline levels 80 h post ACTH administration for F2. For M1, the 5α-pregnane-3β-11β-21-triol-20-one EIA showed the highest response, exceeding 700%, at 105 h post ACTH administration (Tables 2 and 3, Fig. 1C). Similar to F2, the respective fGCM concentrations returned to baseline levels 120 h post ACTH administration (Fig. 1C).

Urinary glucocorticoid metabolites analysis

Similar to the fGCM findings, all four EIAs showed a considerable response in uGCM concentrations (149.53%–651.82%) following the ACTH injection (Table 3). For the two females, the 5α-pregnane-3β–11β–21-triol-20-one and tetrahydrocorticosterone EIA showed the highest response, exceeding 340%, in the first collected faecal sample 27 h post ACTH administration (Tables 2 and 3, Fig. 2A & Fig. 2B). Respective uGCM concentrations returned to baseline levels at 51 h and 80 h post ACTH administration for F1 and F2, respectively. In contrast to the female profiles, the highest response to the ACTH administration in the male was found using the 11-oxoaetiocholanalone EIA (391.37% %), with both the 5α-pregnane-3β-11β-21-triol-20-one and tetrahydrocorticosterone EIAs also showing suitable responses (347.36%–347.87%; Fig. 2C). The peak response occurred 97 h post ACTH injection (Table 2), and thus male uGCM concentrations did not return to baseline levels during the monitoring period.

Conclusion

The ability to monitor physiological stress patterns in endangered reptile species, through non-invasive hormone monitoring techniques, offers conservationists an ideal tool which can be implemented within both free-ranging and captive setups with limited effort. With the increase in human-driven factors leading to substantial decreases in reptile populations, the need for such techniques are becoming more important. This study has successfully validated such a technique for monitoring the stress response in S. giganteus in both urine and faeces by using the 5α-pregnane-3β-11β-21-triol-20-one or tetrahydrocorticosterone EIA. Both assays showed low MAD values as well as a considerable response in fGCM and uGCM concentrations following ACTH injection. As such, both sample matrices can be used to monitor physiological stress in S. giganteus. Despite the results of this study, a number of uncertainties need to be addressed by researcher conducting further studies on the topic. Of greatest concern is the observed gut passage time and time to peak fGCM and uGCM concentrations between individuals. Although the time to peak fGCM (24 h) and uGCM (27 h) responses were similar in both females, the monitored male showed a prolonged gut passage time with peak fGCM and uGCM concentrations 81 h and 70 h later, respectively. However, if in fact differences in gut passage time or GC metabolite patterns between individuals or sexes of the species exist is yet to be determined by examining larger study populations. Currently, we recommend collecting only the faecal or urine component for GC metabolite monitoring in S. giganteus. Despite the limitations of this study the findings increased our understanding of stress hormone production, metabolism and excretion pattern in the species. We hope this will encourage and stimulate future research not only on this species, but reptiles in general, especially concerning the non-invasively examining the physiological stress response linked to a host of anthropogenic and natural factors.