INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

GLUCOCORTICOID RELEASE IN response to stress has been well characterized in a variety of species (e.g., 44). The most common glucocorticoid in birds is corticosterone (Cort; Ref. 19), although the exact function of Cort in the avian stress response remains unclear.

One associated action of Cort in birds is the mobilization of energy substrates, especially during stressful situations. Glucocorticoids classically have been shown to inhibit glucose uptake into tissues of the gut and periphery (29). The majority of glucose mobilization resulting from glucocorticoid stimulation takes place in hepatocytes (24). Adipose tissue is also a site of glucocorticoid-promoted lipolysis, producing an increase in plasma free fatty acids (16, 25, 37). In this way, glucocorticoids appear to release energy stores to help an animal cope with a stressful stimulus, recover from a period of acute stress, or prepare for an upcoming stressor.

Cort has also been proposed as a major antagonistic hormone to the actions of insulin, the primary hormone responsible for glucose uptake throughout the body (14). Glucocorticoids and insulin have been widely shown to work in opposition on energy balance in traditional mammalian models. In rats, for example, glucocorticoids administered centrally stimulate hyperphagia, whereas insulin inhibits feeding (40). Perhaps more important, however, is evidence that glucocorticoids antagonize the effects of insulin at tissues where insulin performs its major storing actions (hepatocytes, adipocytes, and muscle tissue) by exerting opposite, gluconeogenic effects to increase plasma glucose levels (14, 20, 25, 40). Still, little is known about whether this insulin/Cort antagonism on plasma glucose exists in birds.

There are indications that lipids may be a more significant source of energy than glucose in birds, especially during migration and fasting (1, 3, 21, 32, 42). In white-crowned sparrows, feeding decreases when plasma lipids are elevated, but feeding is insensitive to changes in plasma glucose levels (5). Because fat yields twice as much energy and water per gram than does carbohydrate or protein (15, 21), this preference for lipids as the primary energy substrate is far more economical for an organism adapted for flight.

Glucocorticoids released during stress mobilize lipids from adipose tissue, which supports gluconeogenesis. Stored triglycerides are broken down into nonesterified fatty acids (NEFAs), which then can be processed by the liver and other tissues for synthesis of ATP (29). In this way, an increase in plasma NEFAs produces lower plasma triglyceride levels. Stress decreases plasma triglycerides in rats (18, 39), and exogenous ACTH administered to the domestic fowl elevates Cort levels and leads to increases in plasma NEFAs (17). In Japanese quail, both ACTH and dexamethasone (a synthetic glucocorticoid) administration caused an increase in plasma NEFAs (6). Thus the evidence for stress-induced hyperlipidemia is compelling, although lipid mobilization in response to stress remains relatively unstudied in passerines.

We recently showed that stress elicits both a robust Cort response and a marked hyperglycemic response in a common passerine, the European starling (Sturnus vulgaris; Refs. 33, 35). These two studies also demonstrated distinct daily (circadiel) and seasonal (circannual) rhythms in both plasma Cort and glucose levels. These results led us to focus on the following three questions: 1) What is the result of the antagonism between insulin and Cort on stress-induced hyperglycemia in starlings? 2) Do plasma triglycerides respond to stress and in what way does the insulin/Cort antagonism affect stress-induced hyperlipidemia? 3) Do these responses vary on either a daily basis or in response to changes in photoperiod? We continue to use starlings so results can be compared with our earlier studies on Cort and glucose.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Birds. Wild European starlings were captured using mist nets during the winter in Eastern Massachusetts. Juveniles were not captured for use in this study, but some birds may have been hatch-year or second-year birds (31), giving a mixture of young and older birds. Birds were housed communally for several months in large indoor flight aviaries on an 11:13-h light-dark cycle, and 10 birds were transferred to individual cages for 2 wk before injections were initiated (n = 10). All rooms were climate controlled and maintained at 25°C for the duration of the study. Light cycles were adjusted throughout the experiment to induce seasonal changes in the birds. Birds were initially maintained on short days (11:13-h light-dark cycle), and once all injections had been completed, birds were shifted to long days (19:5-h light-dark cycle) as described previously (33). Birds were allowed 2 wk to acclimate to the new photoperiod before injections were resumed. Food and water were provided ad libitum. Captive starlings undergo a predictable daily variation in body weight (33). Because body weight is correlated with food intake (27), we can assume that the basal nutritional state of each bird is similar when samples are taken at the same time of the day over multiple days. All experiments were performed according to American Association for Accreditation of Laboratory Animal Care guidelines and approved by the Institutional Animal Care and Use Committee at Tufts University.

Stress/sampling protocols. Initial blood samples were taken within 3 min of entering the room. These were considered basal samples and pooled for statistical purposes as Cort does not significantly respond to stress until after ~3 min (45) and stress-induced hyperglycemia is not evident until 15 to 20 min following acute stress (7, 11, 12, 33, 43). Control birds were administered a saline injection immediately following blood sampling. Basal samples are reported as 0 min; however, injections were administered following basal sampling. Because this is a within-subjects study, and because each bird was bled at exactly the same time of day for each trial, there should be little change in the nutritional state of each bird before each experiment. Consequently, basal samples were only taken for the saline-injected controls.

Injections. Within 5 min of entering the experimental chamber, birds received subcutaneous injections in the dorsal neck region of either insulin (human recombinant expressed in Escherichia coli; Sigma Chemical) or Cort (dissolved in peanut oil; Sigma Chemical) or a combination of the two hormones. Two doses of insulin were used (1.0 and 4.0 IU/kg), as well as one dose of Cort (200 µg), and 0.85% saline served as the vehicle control. These combined to give six treatments: saline; 1.0 IU insulin; 4.0 IU insulin; saline + 200 µg Cort; 1.0 IU insulin + 200 µg Cort; 4.0 IU insulin + 200 µg Cort. Each bird was given all six treatments at two times a day, with the individual treatment order randomized. Body weight of the birds in this study ranged from 70 to 85 g.

Justification of doses. The normal range of plasma glucose in birds is ~10-20 mmol/l (22, 23), so the target glucose levels we sought for the highest dose of injected insulin ranged from 6 to 12 mmol/l. Additionally, we sought a dose of insulin that would show biological activity for 2.5 h or more, because postmeal insulin (endogenous) takes ~3 h to store ingested carbohydrates in birds and mammals (29, 42). One study of insulin injections in a passerine, the white-crowned sparrow, demonstrated that low doses of human recombinant insulin (0.5 and 1.0 IU/kg) became ineffective at changing food intake by 30 min postinjection, whereas higher doses (2.0 and 4.0 IU/kg) remained biologically active up to 120 min following injection (4). We determined an optimal dosage range and time course for insulin injection into starlings during a pilot study (Fig. 1A). Two doses of insulin, 1.0 and 4.0 IU/kg, injected subcutaneously were found sufficient to show a dose response of plasma glucose levels to insulin administration over the course of 2.5 h.

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Plasma glucose (A) and corticosterone (Cort; B) responses to stress under influence of graded doses of exogenous insulin (Ins.) or Cort, respectively, in birds held on short days at 1100.

Sample processing and assays. Hematocrit tubes were sealed with clay at one end and centrifuged at ~400 g for 5 min. Plasma was removed and frozen until analyzed. Cort was extracted from plasma with dichloromethane and analyzed using a radioimmunoassay as described previously (47). Glucose levels were determined using a hexokinase reagent (Sigma Chemical) combined with spectrophotometry as described previously (33). Plasma triglyceride levels were measured using a lipoprotein lipase/ESPA assay combined with colorimetric spectrophotometry (GPO Trinder, Sigma Chemical). Interassay variability for Cort, glucose, and triglyceride assays was 9.3, 8.5, and 5.9%, respectively. Intra-assay variability for Cort was 6.8%

Statistics. Data were analyzed using repeated-measures ANOVA for the effects of injection, time of day, and photoperiod on basal and stress-induced (40, 70, and 150 min) levels of Cort, glucose, and triglycerides. P values were adjusted to control for performing multiple ANOVAs (38). In addition, the effect of stress on Cort release was analyzed by comparing basal and 40-min plasma Cort levels using paired t-tests.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Basic response to stress. Basal levels of Cort and glucose showed distinct circadiel variation for birds held on short days (Fig. 2A). Cort levels were higher at 2300 than 1100 [F(1,18) = 9.14, P < 0.01], but, in contrast, basal glucose concentrations were elevated at 1100 [F(1,16) = 9.07, P < 0.01]. Triglycerides, on the other hand, showed no circadiel variation. Plasma Cort levels significantly increased after saline injection between 0 and 40 min at 1100 (t = 4.42, df = 9, P < 0.002) and at 2300 (t = 3.83, df = 9, P < 0.005). At 1100, Cort levels remained at stress-induced concentrations for the duration of sampling, whereas at 2300, Cort began to approach basal concentrations by the end of the sampling period. Although plasma glucose did not change after saline injection at 1100, there was a significant hyperglycemic stress response at 2300 through 40 min, after which glucose levels return to baseline [F(3,12) = 4.98, P < 0.02]. Saline injection also produced a decrease in plasma triglycerides at both 1100 [F(3,15) = 37.92, P < 0.0001] and 2300 [F(3,15) = 14.25, P < 0.0001].

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Glucose, Cort, and triglyceride responses to stress at 1100 or 2300 in birds held on short (A) or long days (B) and Cort response to restraint in free-living starlings during short days (C). Points represent means ± SE for n = 10 for captives and n = 7 for free-living starlings. Note differences in scale for Cort graphs.

Responses to insulin. Although there were indications that Cort concentrations tended to be increased at 70 and 150 min following the high-dose insulin injection (4.0 IU), this response was not statistically significant (Fig. 3) [although the increase during long days was nearly significant, with F(1,15) = 5.00, P > 0.05 after adjusting for multiple comparisons]. In contrast, injections of insulin reduced plasma glucose levels (Fig. 4). There was an overall dose-dependent treatment effect as insulin suppressed glucose concentrations at 1100 on both the short-day [F(2,27) = 39.16, P < 0.0001; Fig. 4A] and long-day [F(2,23) = 9.15, P < 0.002; Fig. 4C] photoperiods. Insulin also prevented the stress-induced hyperglycemia at 2300 on both short [F(2,27) = 20.21, P < 0.0001; Fig. 4B] and long days [F(2,22) = 10.32, P < 0.001; Fig. 4D]. Plasma triglyceride concentrations, however, did not change in response to insulin regardless of photoperiod, time of day, or dose (data not shown).

View larger version (29K): [in this window] [in a new window]   Fig. 3.   Cort responses to stress following insulin injection in birds held on short days at 1100 (A), short days at 2300 (B), long days at 1100 (C), or long days at 2300 (D). Points represent means ± SE for n = 10. Saline-injected controls are repeated from Fig. 2.

View larger version (30K): [in this window] [in a new window]   Fig. 4.   Glucose responses to stress following insulin injection in birds held on short days at 1100 (A), short days at 2300 (B), long days at 1100 (C), or long days at 2300 (D). Points represent means ± SE for n = 10. Saline-injected controls are repeated from Fig. 2.

Responses to Cort. Injections of Cort produced supraphysiological levels of Cort (Fig. 5). In birds held on short days (Fig. 5A), Cort was at maximum 40 min after the injection at both 1100 [F(3,81) = 39.03, P < 0.0001] and 2300 [F(3,66) = 59.12, P < 0.0001]. Similarly, in birds held on long days (Fig. 5B), Cort peaked at 40 min at both 1100 [F(3,42) = 50.54, P < 0.0001] and 2300 [F(3,40) = 22.60, P < 0.0001] in response to Cort injection. Exogenous Cort did not alter the glucose response at either time of day or during either photoperiod (Fig. 6) nor did it alter plasma triglyceride concentrations at 1100 in either season (Fig. 7). However, at 2300 on both photoperiods, triglyceride levels appeared to recover after 150 min, and Cort administration seemed to counteract this recovery (Fig. 7). Unfortunately, there was insufficient statistical power to conclusively draw this conclusion [for short day, F(1,16) = 1.66, P = 0.22 for Cort effect; F(3,54) = 0.77, P = 0.51 for the interaction of Cort and sampling time; for long day, F(1,16) = 0.013, P = 0.91 for Cort effect; F(3,50) = 2.206, P = 0.1 for the interaction of Cort and sampling time].

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Cort responses to stress following injection of saline (unmarked) or 200 µg Cort (+Cort) at 1100 or 2300 in birds held on short (A) or long days (B). Points represent means ± SE for n = 10. Saline-injected controls are repeated from Fig. 2.

View larger version (26K): [in this window] [in a new window]   Fig. 6.   Glucose responses to stress following injection of saline (unmarked) or 200 µg Cort (+Cort) in birds held on short days at 1100 (A), short days at 2300 (B), long days at 1100 (C), or long days at 2300 (D). Points represent means ± SE for n = 10. Saline-injected controls are repeated from Fig. 2.

View larger version (28K): [in this window] [in a new window]   Fig. 7.   Triglyceride responses to stress following saline or saline + Cort injection in birds held on short days at 1100 (A), short days at 2300 (B), long days at 1100 (C), or long days at 2300 (D). Points represent means ± SE for n = 10. Saline-injected controls are repeated from Fig. 2.

Interaction between insulin and Cort. Insulin did not alter Cort concentrations in response to exogenous Cort injections, and insulin and Cort did not interact to alter the triglyceride response to stress (data not shown). Exogenous Cort, however, had a limited effect on glucose concentrations by hastening plasma glucose recovery from insulin-induced hypoglycemia (Fig. 8). This effect was only apparent during the day and did not occur at night during either photoperiod (Fig. 8, A and B). By 150 min at 1100, glucose concentrations in insulin + Cort birds were higher than concentrations in birds that received insulin alone and in birds maintained on short [F(3,32) = 5.54, P < 0.005] and long days [F(3,27) = 8.30, P < 0.0005].

View larger version (15K): [in this window] [in a new window]   Fig. 8.   Change in glucose recovery from insulin injection (1.0 IU/kg) in birds receiving insulin alone (Ins.) or insulin + Cort (Ins. + Cort) on short (A) or long days (B). Points represent means ± SE of difference between glucose levels at 70 vs. 150 min for n = 10. Some data are repeated from earlier figures.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cort. Acute handling and restraint stress caused plasma Cort levels to increase in both photoperiods and at both times of the day. This verifies findings from a previous study in starlings (35) and adds to the growing body of comparative data on stress in wild and captive birds (e.g., 36, 46). This study has shown that stress-induced Cort levels begin to attenuate at 70 min and approach basal (nonstressed) levels by 150 min following initiation of an acute stressful stimulus. To our knowledge, this is the first measure of Cort responses to stress for this length of time in a wild or captive bird. These findings suggest a window of importance for Cort release in response to stress in birds, because after ~2.5 h following a stressor, circulating Cort begins to approach basal levels. Thus even though the birds were repeatedly exposed to the stress of handling and restraint throughout the duration of sampling, it is likely that negative feedback on the hypothalamic-pituitary-adrenal axis (e.g., 13) has reduced the Cort stress response dramatically by 2.5 h after the initial stressor.

Glucose. Plasma glucose levels changed in response to handling and restraint stress only during scotophase sampling under both photoperiods. This result supported earlier findings in the same species (33) where stress elevated glucose only at night (not during the day) in birds held on short days, long days, and during a prebasic molt. This earlier study also showed a circadiel rhythm in plasma glucose, with levels highest in mid-day and lowest at night. The present study supported these results, as basal glucose levels were elevated at 1100 over 2300 in both photoperiods. One reason proposed for the lack of a mid-day hyperglycemic response to stress is the variability associated with food intake through the active period (33). Because the birds are feeding throughout the photophase, perhaps it is difficult to turn on a strong hyperglycemic response while plasma glucose levels are elevated at mid-day. The release of insulin to clear postmeal surges in glucose levels may prevent a bird from mobilizing a hyperglycemic stress response until after the storing activity of insulin has subsided. However, we chose to perform these studies on fed, rather than fasted, birds to more closely mimic natural conditions.

Triglycerides. Plasma triglyceride levels decreased in response to stress in all groups, independent of photoperiod, time of day, or injection. This result was expected in light of previous studies in mammals (18, 39). Whether this is an effect of glucocorticoids or another system responsive to stress (e.g., catecholamine release) remains to be tested. Glucocorticoids are known to inhibit synthesis of triglycerides from NEFAs (2), but epinephrine has also been shown to increase triglyceride lipase activity (29).

Perspectives