The quantification of reproductive hormones in the hair of captive adult brown bears and their application as indicators of sex and reproductive state

Abstract

Recognizing the potential value of steroid hormone measurements to augment non-invasive genetic sampling, we developed procedures based on enzyme-linked immunoassays to quantify reproductive steroid hormone concentrations in brown bear (Ursus arctos) hair. Then, using 94 hair samples collected from eight captive adult bears over a 2-year period, we evaluated (i) associations between hair concentrations of testosterone, progesterone, estradiol and cortisol; (ii) the effect of collecting by shaving vs. plucking; and (iii) the utility of reproductive hormone profiles to differentiate sex and reproductive state. Sample requirements (125 mg of guard hair) to assay all hormones exceeded amounts typically obtained by non-invasive sampling. Thus, broad application of this approach will require modification of non-invasive techniques to collect larger samples, use of mixed (guard and undercoat) hair samples and/or application of more sensitive laboratory procedures. Concentrations of hormones were highly correlated suggesting their sequestration in hair reflects underlying physiological processes. Marked changes in hair hormone levels during the quiescent phase of the hair cycle, coupled with the finding that progesterone concentrations, and their association with testosterone levels, differed markedly between plucked and shaved hair samples, suggests steroids sequestered in hair were likely derived from various sources, including skin. Changes in hair hormone concentrations over time, and in conjunction with key reproductive events, were similar to what has been reported concerning hormonal changes in the blood serum of brown bears. Thus, potential for the measurement of hair reproductive hormone levels to augment non-invasive genetic sampling appears compelling. Nonetheless, we are conducting additional validation studies on hair collected from free-ranging bears, representative of all sex, age and reproductive classes, to fully evaluate the utility of this approach for brown bear conservation and research.

Keywords: Brown bear; Ursus arctos; enzyme-linked immunoassay; hair cortisol; hair reproductive hormone profile; non-invasive genetic sampling.

Figure 1:

Sequence of procedures followed in the statistical analysis of reproductive hormone concentrations measured in 94 hair samples collected from April 2013 to April 2015 from eight captive adult brown bears housed at the WSU Bear Research, Education and Conservation Centre. Interaction terms are not provided in the candidate models shown in the figure, but were included in the analyses and are discussed in the text. Abbreviations used are linear mixed models (LMM), generalized linear mixed models (GLMM), Akaike information criterion (AIC), and difference in sample-size–adjusted Akaike information criterion (AIC_C) values between top model and another candidate model (∆AIC_C).

Figure 2:

Reproductive hormone concentrations measured in 94 hair samples collected from eight adult captive brown bears between April 11, 2013 (Day 21) and April 3, 2015 (Day 13). Twenty-eight samples were collected from two adult females (11.3–12.3 years) that bred in May 2014 (41–71 days since March 21st) and gave birth to four cubs in early to mid-January 2015 (285–300 days). Forty-one samples were collected from four adult females (8.3–10.8 years) that were administered megestrol acetate (orally at 40–160 mg/day) as a method of birth control from April 20 to June 19 (30–90 days) in each year. Twenty-five samples were collected from two adult males (11.8–13.1 years) that bred in May 2014. Sixty-four hair samples were collected using electric clippers to sever the hair shaft at the skin surface, thus removing the follicle. At 30 sampling times, a sample was collected by plucking hair (follicle intact) with a hemostatic clamp from the skin. At 20 of these sampling times, hair was collected as paired samples by both shaving and plucking from the skin at adjacent body locations. Physiological phases correspond with: (i) pre-breeding—Days 1–35; (ii) breeding—Days 36–91; (iii) post-breeding—Days 92–212; and (iv) hibernation—Days 213–365.

Figure 3:

The predicted mean testosterone concentration in the hair of captive adult brown bears in relation to the standardized hair cortisol concentration, and by sex and reproductive class. The means and 95% confidence intervals were estimated by resampling of data from 94 records using model T1 presented in Table S2a. The analysis was constrained to bears that were sampled during the hibernation phase on January 4th. Standardized continuous variables in model T1 were set at mean values as follows: estradiol = 0, progesterone = 0 and ordinal day = 1.306.

Figure 4:

The predicted mean testosterone concentration in the hair of captive adult brown bears in relation to the standardized hair estradiol concentration, and by sex and reproductive class. The means and 95% confidence intervals were estimated by resampling of data from 94 records using model T1 presented in Table S2a. The analysis was constrained to bears that were sampled during the hibernation phase on January 4th. Standardized continuous variables in model T1 were set at mean values as follows: cortisol = 0, progesterone = 0 and ordinal day = 1.306.

Figure 5:

The predicted mean testosterone concentration in the hair of captive adult brown bears in relation to the standardized ordinal day of sampling, and by sex and reproductive class. The means and 95% confidence intervals were estimated by resampling of data from 94 records using model T1 presented in Table S2a. Physiological phases correspond with standardized ordinal day ranges as follows: pre-breeding (Pr: −1.660 to −1.424), breeding (B: −1.414 to −0.824), post-breeding (Po: −0.813 to 0.474) and hibernation (H: 0.484–1.610). Standardized continuous variables in model T1 were set at mean values as follows: cortisol = 0, estradiol = 0 and progesterone = 0.

Figure 6:

The predicted mean progesterone concentration in the hair of captive adult brown bears in relation to the standardized hair cortisol concentration, and by sex and reproductive class. The means and 95% confidence intervals were estimated by resampling of data from 94 records using model P1 presented in Table S3a. The analysis was constrained to bears that were sampled by shaving hair (no follicles) during the hibernation phase on January 4th. Standardized continuous variables in model P1 were set at mean values as follows: testosterone = 0, estradiol = 0 and ordinal day = 1.306.

Figure 7:

The predicted mean progesterone concentration in the hair of captive adult brown bears in relation to the standardized ordinal day of sampling, and by sex and reproductive class. The means and 95% confidence intervals were estimated by resampling of data from 94 records using model P1 presented in Table S3a. The analysis was constrained to bears that were sampled by shaving hair (no follicles). Physiological phases correspond with standardized ordinal day ranges as follows: pre-breeding (Pr: −1.660 to −1.424), breeding (B: −1.414 to −0.824), post-breeding (Po: −0.813 to 0.474) and hibernation (H: 0.484–1.610). Standardized continuous variables in model T1 were set at mean values as follows: cortisol = 0, testosterone = 0 and estradiol = 0.

Figure 8:

The predicted mean progesterone concentration in the hair of captive adult brown bears in relation to the standardized hair (a) cortisol and (b) testosterone concentrations, by method of hair collection. The means and 95% confidence intervals were estimated by resampling of data from 94 records using model P1 presented in Table S3a. With shaved samples, the assay was used to determine the progesterone concentration for guard hair shafts only. With plucked samples, progesterone concentrations reflect guard hairs with intact follicles. The analysis was constrained to breeding female bears that were sampled during the hibernation phase on January 4th. Standardized continuous variables in model P1 were set at mean values as follows: testosterone = 0 for panel (a), cortisol = 0 for panel (b), estradiol = 0 for both panels and ordinal day = 1.306 for both panels.

Figure 9:

The predicted mean estradiol concentration in the hair of captive adult brown bears in relation to the standardized hair testosterone concentration, and by method of hair collection. The means and 95% confidence intervals were estimated by resampling of data from 94 records using model E1 presented in Table S4a. With shaved samples, the assay was used to determine the testosterone concentration for guard hair shafts only. With plucked samples, testosterone concentrations reflect guard hairs with intact follicles. Standardized continuous variables in model E1 were set at mean values as follows: cortisol = 0 and progesterone = 0.

Figure 10:

The predicted mean estradiol concentration in the hair of captive adult brown bears in relation to the standardized hair progesterone concentration, and by method of hair collection. The means and 95% confidence intervals were estimated by resampling of data from 94 records using model E1 presented in Table S4a. With shaved samples, the assay was used to determine the progesterone concentration for guard hair shafts only. With plucked samples, progesterone concentrations reflect guard hairs with intact follicles. Standardized continuous variables in model E1 were set at mean values as follows: cortisol = 0 and testosterone = 0.

Figure 11:

Reproductive hormone concentrations measured in plucked hair samples (includes follicles) collected from two female (bF) and two male (bM) captive brown bears that bred successfully in May 2014. The females each gave birth to two cubs in early to mid-January 2015. The physiological phases are hibernation (H), pre-breeding (Pr), breeding (B) and post-breeding (Po).

Figure 12:

Reproductive hormone concentrations measured in shaved hair samples (does not include follicles) collected from six captive brown bears from April 2014 to April 2015. Two female (bF) and two male bears (bM) that bred successfully in May 2014, and the females each gave birth to two cubs in early to mid-January 2015. Two non-breeding females (nbF) were administered megestrol acetate as a method of birth control from mid-April to mid-June. The physiological phases are hibernation (H), pre-breeding (Pr), breeding (B) and post-breeding (Po).