«Item type Thesis or dissertation Authors Davis, Nicolas Citation Davis, N., Schaffner, C. M., & Smith, T. E. (2005). Evidence that zoo visitors ...»
A stock urine pool was made up for the purpose of providing quality control samples to run on each plate. The quality controls were used to calculate the intraand inter-assay CV (see below). Two different pools were made up altogether, one for the assay validation and visitor effect study (Pool A, see Table 2.5 and Chapter 3 respectively) and one for the social impact study and case study (Pool B, see Table
2.4 and Chapters 4 and 5).
Pool A was made up by taking 100 µl from each of 35 different samples taken from each of the group members that took part in the immunological validation and visitor study (see Table 2.5). An individual contributed either five or ten samples to the pool depending on their contribution to the overall study. Overall, pool A was made up of 10 samples from the adult male and 25 from the adult females.
Pool B was made up by taking 100 µl from each of 88 different samples taken from each of the group members that were present during the social impact study. Fourteen samples were selected from each adult that remained in the group throughout the study period, six samples from an adult that was only present during part of the study and six from each of the two sub adults. In summary, the pool was made up from 62 samples from the adult females, 14 from the adult male and 12 from the sub adult males.
Table 2.5 Information on the pools used in the cortisol EIA for the different studies.
Pool Number of Time period Number and sex Experiment samples of animals
To compute inter- and intra-assay variation, a high quality control of 1:128 dilution and a low quality control of 1:1024 dilution for the appropriate Pools A and B was used for each plate. These dilutions were chosen since they represented approximately 33% and 66% binding of the AB. These dilutions were made up when required and stored in the freezer at -20°C.
18.104.22.168 Immunological validation method An enzyme-immunoassay is based on an immunological reaction in which the analyte of interest binds to a specific antibody. This binding process can be affected by variables capable of causing imprecision and inaccuracy, such as interfering substances that could cross-react with the samples. Immunological validation is necessary and followed Diamandis and Christopoulos (1996) suggestions for appropriate immunoassay protocols demonstrating accuracy, specificity, precision and sensitivity.
Assay accuracy was determined by adding a low (1:2048), a medium (1:1024) and a high concentration (1:32) of Pool A to six serial diluted commercial standard samples [n = 6, 500, 250, 125, 62.5, 31.25, 15.6 pg/50 ul)] to ascertain sample recovery. To determine assay specificity, two displacement curves of halving dilutions of a urine pool A (see Table 2.5), ranging from 1:32 to 1:4086, were compared with a displacement curve of cortisol standard preparation to determine parallelism. The F-statistic was used to compare the slopes of the linear regressions of the displacement curves.
Precision was ascertained for each urine Pool separately, by monitoring interand intra assay coefficient of variations by using two quality control samples [a high quality control at around 65% binding (1:1024), and a low quality control at around 37% binding (1:128)] on all plates assayed for each part of the study. The intraassay variation was calculated by averaging reported CV for high and low quality controls on each plate. The inter-assay CV was determined by dividing the overall standard deviation of the quality control CVs by the overall mean multiplied by 100 to get a percentage. Target values for intra- and inter assay CVs are 5 and 10% respectively, although values up to 10 and 20% respectively are acceptable (Diamandus & Christopoulos, 1996). Sensitivity was calculated as a mean from all assays of the lowest concentration of cortisol on the 90% binding point of the standard curve.
22.214.171.124 Immunological validation results For accuracy the recovery of the commercial standard preparations added to the low concentration pool was 123.6% (r = 1.0; Y = 1.26X + 0.99; p 0.0001), the medium concentration pool was 104.8% (r = 1.0; Y = 1.08X - 3.91; p 0.0001) and the high concentration pool was 97.8% (r = 0.993; Y = 0.97X - 4.51; p 0.0001).
For specificity two separated serial dilutions of the urine pool A gave displacement curves that were parallel to the serial dilutions of the commercially prepared cortisol standards [F (2, 16) = 0.53, NS; F (2, 20) = 2.94, NS]. Results are calculated as the ratio of the antibody-analyte complexes of standards or samples that have bound (B) versus that bound at zero concentration or maximum binding (Bo) and is expressed as % B/Bo (Figure 2.9).
For precision the intra-assay variation for the validation and visitor effect study (Chapter 3, Pool A) was 5.67% (n = 8) and 4.40% (n = 8) for the high and low quality control pools respectively. For the social impact study (Chapter 5, Pool B) the assays were carried out over three batches with new standards and controls used each time. For batch one intra- assay variation was 6.03% (n = 33) and 6.16% (n = 33), for batch two 8.96% (n = 39) and 5.93% (n = 39) and for batch three 7.09% (n = 45) and 3.24% (n = 45) for the high and low quality control pools respectively. Finally, for the new male case study (Chapter 6, Pool B) intra-assay variation was 5.59% (n = 5) and 3.83% (n = 5) for the high and low quality control pools respectively.
Figure 2.9 The binding ratio (%B/Bo) of two serial dilutions of the spider monkey urine Pool A and the cortisol standards to demonstrate parallelism.
The inter-assay variation for the validation and visitor effect study (Chapter 3, Pool A) was 2.03% (n = 8) and 11.51% (n = 8) for the high and low quality control pools respectively. For the social impact study (Chapter 5, Pool B) the assays were carried out over three batches with new standards and controls used each time. For batch one the inter-assay CVs were 15.67% (n = 33) for the high and 19.81% (n =
33) for the low quality control pools, for batch two 19.58% (n = 39) for the high and 18.21% (n = 39) for the low, and for batch three 17.7% (n = 45) for the high and 20.9% (n = 45) for the low quality control pools. Finally, for the new male case study (Chapter 6, Pool B), the inter-assay CV was 15.5% (n = 10) and 15.6% (n = 10) for the high and low quality control pools respectively. The assay sensitivity was 3.95 pg.
126.96.36.199 Biological validation method Five adult spider monkeys from the Chester Zoo group contributed a total of 53 urine samples to the circadian assessment. Samples were collected opportunistically from 08:00 hrs to 18:00 hrs for three days over a seven-day period from the five adult subjects (Table 2.6). The diurnal period was divided into five time blocks of two hours and the samples were collapsed for each subject across the three collection days. The mean concentration of urinary cortisol was calculated for each animal in each time slot and analysed using a one-factor repeated measure of analysis of variance. This test was followed by a linear trend analysis to determine if there was a significant decreasing pattern over time of day (Keppel, 1993).
Table 2.6 Number of samples used for each time block for the circadian rhythm validation.
06:00 – 07:59 12 1.7 08:00 – 09:59 13 1.9 10:00 – 11:59 11 1.6 12:00 – 13:59 12 1.7 14:00 – 15:59 12 1.7 188.8.131.52 Biological validation results A circadian variation in urinary cortisol excretion was demonstrated when samples collected over three days and across an 8-h period were analysed (Figure 2.10). The assay was thus effective in detecting diurnal variation in cortisol levels [F (4, 16) = 4.59, p 0.001] confirming that cortisol excreted in the urine, as measured in my assay accurately reflects levels of cortisol circulating in the plasma. Trend analyses revealed a significant decreasing linear trend across the five time periods [F (1, 4) = 9.75, p 0.035].
184.108.40.206 Creatinine assay The hormone concentration for each sample was corrected for creatinine concentration using a modified Jaffe end-point assay (Burtis & Ashwood, 2001). To identify the most appropriate dilution an initial assay was run using a series of urine 3.5
Figure 2.10 Mean ± SEM levels of urinary cortisol across the time of day.
dilutions from a high dilution of 1:2 to a very low dilution of 1:2048. Based on the latter assays, the working dilution for creatinine assays was run at a dilution of 1 part urine in 200 parts distilled water.
The working dilution was made in two steps involving a 1:64 dilution followed by a 1:3.125 dilution. The urine samples stored at -20°C were defrosted at room temperature prior to assaying and thoroughly mixed for around 10 seconds using a vortex mixer. To create the 1:64 dilution, 630 μl of distilled water was pippetted into individually labelled 1.5ml eppendorf tubes, added to 10μl of the sample and mixed thoroughly. For the second step 235 μl of the 1:64 dilution was mixed with a further 500 μl of distilled water in another labelled 1.5ml eppendorf tube to form the working dilution. For each plate a maximum 28 samples could be run in duplicate (See Appendix B for template).
The inter plate reliability was assessed by running on each plate a high quality control (HQC) of 1:128 dilution and a low quality control (LQC) of 1:1024 dilution of the appropriate pools. Again the dilutions were measured out in two steps, with a one to one dilution of the 1:64 urine dilution with distilled water for the HQC and a one to sixteen dilution of the 1:64 urine dilution with distilled water for the LQC.
Creatinine assays were carried out using 96 well non bonding microtiter plates (Maxisorp, NUNC™). Standards (n = 4, 6, 3, 1.5, 0.75 µg / 200 µl, Sigma) and samples (at 1:200) were diluted in dH20 and 200 µl pipetted into the appropriate wells in duplicate. Control wells contained 200 µl of dH20 as an indicator of nonspecific binding. For each plate a mixture of 5ml of NaOH (0.75M) and 5 ml of picric acid was required, with 100 µl of the mixture added using a multi pipette to all wells except the control wells. The plate was then placed on the shaker for 1-2 minutes then read using the software Revelation version 4.22 on a microplate reader (Dynatech MR700), when the optical density at 490 nm of the top standard measured around 1.7. All urinary cortisol concentrations were expressed relative to creatinine (ug urinary cortisol / mg Cr).
2.3.3 Summary An enzyme-immunoassay was successfully developed and validated to quantify urinary cortisol in spider monkeys. Immunological validation of the assay was achieved by showing specificity, accuracy, precision and sensitivity. The biological validation was confirmed with the detection of a typical diurnal pattern of cortisol excretion in the urine, which is evident in the plasma of primates (Coe & Levine, 1995; Czekala, Lance, & Sutherland-Smith, 1994; T. E. Smith & French, 1997a).
By developing a biologically valid assay to quantify cortisol I have added to the growing number of research studies that use physiological indices as a tool for measuring potential stressors and biological events (Boinski, et al., 1999; Crockett, et al., 2000; Dettling, et al., 2002; Whitten, et al., 1998; Ziegler, et al., 1995). Recently, studies using faecal steroid assays have assessed the relationship between puberty and dispersal in wild female muriqui monkeys (Bracyteles arachnoids) (Strier & Ziegler, 2000) and ovarian cycles in Geoffroy’s spider monkeys (A. geoffroyi) (Campbell, et al., 2001). My findings contribute to these advances in the study of steroid hormones in Ateline primates. Collectively, this research is relevant for the captive breeding and management of New World monkeys as it provides a mechanism to gain valuable information regarding general welfare and reproductive competence, as well as encouraging researchers to explore more refined questions, such as the impact of the zoo environment on physiology.
To conclude, I have validated an enzyme-immunoassay to quantify levels of urinary cortisol in spider monkeys. The biochemical technique can be applied to assess the relationship between various stressors and a physiological index in a primate species.
3.1 Introduction 3.1.1 Factors effecting welfare of zoo-housed animals Animals in captivity are exposed to a variety of potentially harmful stressors.
These include environmental sources of stress related to housing conditions such as artificial lighting, aversive sounds, odours, substrate and extremes of temperatures, as well as confinement-specific stressors such as reduced retreat space, forced proximity to humans, restricted movement, abnormal social groupings, reduced feeding opportunities and other restrictions on opportunities for natural behaviour (Morgan & Tromborg, 2007). Animals in captivity also have a reduced amount of control over their environment and an increased amount of predictability (Carlstead, 1996). It is this lack of control and variations in predictability that are potentially the greatest stressors for animals in captivity (Bassett & Buchanan-Smith, 2007;
Sambrook & Buchanan-Smith, 1997; Wiepkema & Koolhaas, 1993).
A zoo environment has been identified as being unique compared to other captive environments and is characterised by the combination of three specific factors (Hosey, 2005). Firstly, the physical available space for animals in zoological parks is much smaller than they would normally range over in the wild. The impact of this restricted space on the welfare of the animals is however complex. While sufficient quantity and quality of space must be provided to enable appropriate species-specific behaviours the provision of resources makes comparisons with the wild difficult. The second is that most aspects of a zoo animal’s life are managed to some degree by humans. Their accommodation, feeding, group composition, health and reproduction are all to a greater degree out of their control. Finally the constant presence of a large number of unfamiliar humans is also unique to a zoo environment. The impact of visitors on zoo animals has been investigated with a review of studies showing how various visitor characteristics can have an effect (Davey, 2007).