Role of Corticotrophin-Releasing Hormone in the Impairment of Counterregulatory Responses to Hypoglycemia

Male Sprague-Dawley rats that weighed 250–300 g (Charles River Breeding Laboratories, Wilmington, MA) were studied. Animals were housed in an environmentally controlled room with a 12-h light/dark cycle and were maintained on standard ad libitum rat diet (Prolab 3000; AGWAY, Waverley, NY). Animals were given 10 days to acclimate to the animal facility before undergoing surgery. Animals weighed 334 ± 4 g before surgery and 340 ± 4 g on the first day of the treatment protocol. There was no weight difference between groups at either time. All procedures were reviewed and approved by the Yale Animal Care and Use Committee.

All animals underwent an aseptic surgical procedure for the placement of internal jugular vein and carotid artery catheters under intraperitoneal pentobarbital anesthesia (Nembutal 35 mg/kg body wt; Abbott Laboratories, Chicago, IL). The polyethylene carotid artery catheter was extended to the level of the aortic arch, and the silicone internal jugular vein catheter was extended to the level of the right atrium. At the end of the procedure, both catheters were flushed and filled with heparin (42 units/ml) and polyvinylpyrrolidone (1.7 g/ml) solution (Sigma Chemical, St. Louis, MO). Catheters were tunneled subcutaneously and externalized behind the head. Catheters then remained sealed until the study day. Only animals that had recovered completely from surgery, were behaving normally, and were gaining weight with no evidence of infection were studied.

Seven days after surgery, animals were divided into five groups. The protocol consisted of three mornings of antecedent treatment followed by a hyperinsulinemic-hypoglycemic clamp on day 4. Animals were allocated to a treatment group in random order over the course of the study. All groups of animals continued to gain weight during the 3 days of antecedent treatment with no difference between groups.

Animals in the antecedent hypoglycemia group (Insulin n = 12) received 10 units/kg short-acting insulin (Humulin R; Eli Lilly, Indianapolis, IN) as an intraperitoneal injection at 9:30 each day. Animals were then fasted for 3 h. This produced an ∼2- to 2.5-h period of hypoglycemia on the 3 days before a hyperinsulinemic-hypoglycemic clamp. Animals were given access to food at 150 min. If animals were unable to eat at that time, then they were given an intraperitoneal injection of glucose (400 mg). The intraperitoneal route was selected because we observed, in preliminary experiments, that subcutaneous insulin produces a period of hypoglycemia of variable length and intensity. Animals were handled on each day to acclimate to the laboratory surroundings, but the vascular catheters were not opened. A separate group of animals were used to characterize the glucose and corticosterone response to insulin injections on each of the antecedent days. Results for corticosterone are shown for day 1 of hypoglycemia (n = 7); days 2 and 3 showed similar profiles (data not shown).

Animals in the control group (Saline n = 14) were handled in the same way as the Insulin group, but the intraperitoneal injections were replaced with 0.5 ml of normal saline. Again, a separate group of animals (n = 7) were used to quantify the glucose and corticosterone response to each day of this treatment.

Having characterized the corticosterone response to insulin-induced hypoglycemia in the insulin hypoglycemia group described above, a dose of subcutaneous corticosterone (Cort n = 9) was empirically calculated to produce a similar elevation in plasma corticosterone concentrations. Solutions were prepared by dissolving 20 mg of corticosterone (Sigma Chemical) in 2 ml of absolute ethanol and then diluted to 10 ml with normal saline. Compared with the insulin group, subcutaneous corticosterone injection resulted in a faster rise in plasma corticosterone with a rapid fall. To compensate for the latter, the final regimen chosen involved two injections of 500 μg of corticosterone at time 0 and again at 60 min. Separate groups of animals were again used to characterize the glucose and corticosterone responses to these injections (n = 7).

A dose of subcutaneous ACTH (ACTH n = 8) was empirically calculated to produce a similar elevation of plasma corticosterone concentration as seen during insulin induced hypoglycemia. A dose of 2 units of synthetic ACTH (Organon, West Orange, NJ) was administered on each of the three antecedent mornings. Animals were handled and fasted as previously described. A separate group of animals (n = 12) were used to characterize the glucose and corticosterone response to the injections.

A dose of 10 μg of rat CRH (CRH n = 10; Sigma Chemical) was administered subcutaneously on each morning to produce a peak plasma corticosterone similar to that of antecedent insulin-induced hypoglycemia. Animals were handled and fasted as previously described, but vascular catheters were not opened until the day of the clamp. A separate group of animals was again used to characterize the glucose and corticosterone response to CRH (n = 9).

For further elucidating the actions of antecedent CRH on the sympathoadrenal axis, two additional groups of animals were studied. Animals underwent surgery as described above and if healthy at 1 week were randomly assigned to receive 3 days of antecedent CRH (as described previously) in combination with the CRH receptor 1 (CRHr1) antagonist antalarmin (n = 7) or a control group of CRH in combination with the diluent vehicle for antalarmin (n = 9). Antalarmin is a lipophilic CRHr1-specific antagonist (supplied by the laboratory of Dr. G. Chrousos) and is dissolved in an ethanol/emulphor solution (Sigma Chemical), giving a concentration of antalarmin of 100 mg/ml. Immediately before injection, the solution is diluted with water to a concentration of 20 mg/ml and then injected at a dose of 2 mg/100 g for both groups. Animals then underwent the hyperinsulinemic-hypoglycemic clamp.

The hyperinsulinemic glucose clamp technique as adapted for the rat (26) was used to provide a standardized hypoglycemic stimulus. On day 4 of the study, all groups of animals underwent the same experimental protocol. Rats were fasted overnight before the study. On the morning of the study, the catheters were flushed with saline/heparin (1–2 units/ml heparin). Animals were fully awake and freely moving about the cage untethered. Animals were allowed to rest for 60 min after the catheters were opened. Baseline blood samples were then withdrawn for measurement of glucose, insulin, epinephrine, and norepinephrine concentration. A continuous infusion of human insulin at a rate of 20 mU · kg⁻¹ · min⁻¹ (Humulin R; Eli Lilly) was then started and maintained for 90 min. A variable infusion of exogenous glucose (dextrose 10%) was then adjusted based on 5-min plasma glucose measurements to achieve a mean plasma glucose concentration of 50 mg/dl for the duration of the study. Studies in which plasma glucose fell below 40 mg/dl were excluded from analysis. Blood samples for measurement of insulin, epinephrine, norepinephrine, corticosterone, and glucagon were taken at t = 60 and 90 min.

Plasma glucose was measured in duplicate using the glucose oxidase method (Beckman Instruments, Fullerton, CA). Plasma insulin (Binax, South Portland, ME) and glucagon and corticosterone (ICN Biomedicals, Carson, CA) were determined by double-antibody radioimmunoassays. Catecholamine analysis was performed by high-performance liquid chromatography using electrochemical detection (ESA, Acton, MA).

All data are expressed as means ± SE. Comparison between the study groups was made using ANOVA, followed by a Student’s t test to localize effects if differences achieved significance. All analyses were performed using SPSS for Windows version 10.0.0.

Figure 1 shows the plasma glucose response to insulin injection over the 3 antecedent days. Plasma glucose fell further on each of the antecedent days. Glucose levels were significantly lower on day 3 than on day 1 at the 30-, 60-, and 90-min time points (for t = 90 min, glucose values were 58 vs. 44 mg/dl; P = 0.005).

Table 1 compares the corticosterone response to insulin-induced hypoglycemia with the corticosterone responses to injection of corticosterone, ACTH, and CRH. The aim was to produce a corticosterone profile similar to that induced by hypoglycemia but with the minimum of animal handling. Basal corticosterone did not differ significantly between the groups before injection. The magnitude of the peak rise in corticosterone achieved by antecedent treatment did not differ significantly between groups (P = 0.051); the corticosterone- and CRH-treated groups, however, showed a trend toward higher values. The area under the corticosterone curve was virtually identical in the insulin hypoglycemia, corticosterone, and CRH groups, whereas the ACTH-treated group showed a significantly reduced area under the curve as compared with the other groups (P < 0.05). Apart from the insulin group, none of the animal groups showed a significant change in plasma glucose in the 180 min after any of the 3 days of antecedent treatment. There were no significant differences in baseline glucose measurements on any of the antecedent days.

Table 2 compares fasting plasma glucose and insulin concentrations as well as glucose and insulin levels achieved during the clamp. There were no differences among the five groups in fasting plasma glucose or insulin. In addition, basal concentrations of corticosterone (mean 236 ± 20 ng/ml), epinephrine (mean 142 ± 16 pg/ml), or norepinephrine (mean 498 ± 75 pg/ml) were not significantly different among the groups before the hypoglycemic clamp study. Insulin concentrations achieved during the clamp were also similar among groups. Figure 2 displays mean glucose levels achieved between 0 and 90 min for each group. The hypoglycemic plateau was virtually identical in each group, except for corticosterone-treated animals, which showed significantly higher mean glucose levels during the clamp than controls (P = 0.02). Although the glucose infusion rates (Table 2) at the end of the clamp period tended to be higher in the insulin and CRH groups and lower in the corticosterone and ACTH groups as compared with the saline control group, the differences were not statistically significant.

In response to the hypoglycemic stimulus, plasma concentrations of all counterregulatory hormones measured increased in all animal groups. Figure 3A shows the peak epinephrine response achieved by each group during the clamp. As expected (7), the insulin-treated group (recurrent antecedent hypoglycemia) showed a reduced epinephrine response to subsequent hypoglycemia (insulin 1,673 ± 241 vs. control 3,559 ± 563 pg/ml; P = 0.007). In contrast, despite the reduced hypoglycemic stimulus, the increment in epinephrine in the corticosterone group tended to be higher as compared with controls (5,704 ± 912 pg/ml; P = 0.063). Moreover, ACTH animals showed a nearly twofold higher peak epinephrine response than controls (6,681 ± 814 pg/ml; P = 0.007 versus controls). The epinephrine response in the CRH-treated group, however, was markedly reduced and similar to that of the insulin-treated group (1,276 ± 404 pg/ml; P = 0.034 versus controls). Figure 3B shows the peak norepinephrine response achieved by each group during the clamp. Again, the insulin-treated group showed a reduced response as compared with controls (insulin 567 ± 69 vs. control 821 ± 80 pg/ml; P = 0.024). Similarly, norepinephrine responses were suppressed in the CRH animals (580 ± 43 pg/ml; P = 0.016), whereas the corticosterone and ACTH groups showed no significant differences from the controls. Figure 4 shows the peak glucagon response to clamped hypoglycemia (glucagon data were not available for the ACTH group). Although the glucagon response tended to be higher in the corticosterone group and lower in the antecedent insulin hypoglycemic groups, these differences did not reach statistical significance. However, glucagon release in the CRH-treated rats was again reduced (CRH 68 ± 10 vs. control 157 ± 21 pg/ml; P = 0.004). The corticosterone response for each group during the hypoglycemic clamp is shown in Fig. 5. Neither the corticosterone nor the ACTH group showed a difference from controls, whereas insulin and CRH treatment enhanced the response (control 486 ± 19 vs. insulin 554 ± 22 [P = 0.03] and CRH 613 ± 24 ng/ml [P = 0.004]).

Table 3 compares the two additional groups of animals treated with either a combination of CRH and the CRHr1 antagonist antalarmin or a control group of CRH with the diluent vehicle for antalarmin. There was no difference in fasting glucose, fasting insulin, or glucose and insulin levels achieved during the clamp. Animals in the CRH/vehicle group showed suppression of peak epinephrine response to hypoglycemia. The degree of suppression was similar to that in the group of animals treated with antecedent CRH as described earlier. In animals treated with CRH in combination with the CRHr1 antagonist, this suppression in epinephrine response was not seen. Epinephrine responses in this group were similar to those of the previous saline control group. In contrast, the addition of the CRHr1 antagonist had no significant effect on the norepinephrine response to hypoglycemia.

We have shown, using a chronically catheterized rodent model, that 3 days of insulin-induced hypoglycemia can produce downregulation of the sympathoadrenal response to subsequent hypoglycemia. This deficit is similar to that seen in patients who experience recurrent hypoglycemia and hypoglycemia unawareness. Remarkably, the administration of CRH but not ACTH or corticosterone for 3 days, in doses that raised glucocorticoid levels to the same extent as was seen with insulin-induced hypoglycemia, produced similar sympathoadrenal downregulation.

Deficient secretion of glucagon and catecholamines plays a major role in the morbidity associated with iatrogenic hypoglycemia in patients with type 1 diabetes (7,27,28). The glucagon response to hypoglycemia is commonly diminished after some years of treatment of type 1 diabetes (29,30). Consequently, the adrenergic response becomes the critical defense mechanism against iatrogenic insulin-induced hypoglycemia. It is now well-established the antecedent hypoglycemia per se is a key contributory factor to the diminished adrenergic response to hypoglycemia seen in patients with diabetes, particularly those who receive intensive insulin treatment (17). It is not known how insulin-induced hypoglycemia produces this syndrome, although glucocorticoid excess has been implicated (23,24).

We chose an awake rat model to perform this study because of the morbidity associated with recurrent hypoglycemia and because of the greater ability to manipulate different treatment regimens over time. The use of an animal model, however, does have disadvantages and imposes limitations on the study design. Handling of the animals is in itself a stressful stimulus; therefore, all animal groups were handled in a similar way before the clamp study to minimize stress. It is noteworthy in this regard that all five animal groups showed continued weight gain during the 3 days of antecedent treatment. We attempted to minimize the impact of the surgery by studying only those animals that had fully regained the weight they had lost as a result of the procedure. The observation that antecedent iatrogenic hypoglycemia decreases epinephrine and norepinephrine responses to acute hypoglycemia agrees with a number of previous human and animal studies (6–12). It is interesting that antecedent corticosterone treatment tended to have the opposite effect, a slightly greater epinephrine response in the face of a reduced hypoglycemic stimulus. These findings are in agreement with previous reports in rodents (31,32) but contrast with clinical studies in which antecedent cortisol or ACTH infusion suppressed epinephrine responses to hypoglycemia (23,33). This may reflect differences in the protocols used. The clinical studies (23,33) used either 1 day of previous glucocorticoid exposure or 1 day of ACTH infusion, whereas the current study and the study by Shum et al. (31) in rodents employed 3–4 days of exposure to glucocorticoid before instituting hypoglycemia. It has been suggested that the interactions between the HPA and sympathoadrenal axes vary with time. Cunningham et al. (34) showed that central administration of CRH initially stimulated HPA and sympathoadrenal activity but after several days attenuation of CRH responsiveness occurred. Recent work by Evans et al. (35) supports the view that the paraventricular nucleus (the major site of CRH expression within the brain) plays a key role in the counterregulatory impairment associated with recurrent hypoglycemia but that antecedent glucocorticoid exposure has little effect on paraventricular nucleus (PVN) activation (32).

An alternative explanation is that the responses differ between species, although previous studies have shown the awake rat to be an excellent model of glucose counterregulation in humans, including the influence of antecedent hypoglycemia on sympathoadrenal responses (7). The animal and human data are not necessarily contradictory. Both the site and the timing of the many HPA/sympathoadrenal interactions seem to be important (32,34,35). For example, the divergence between the human and rodent data might represent differences in the sensitivity and timing of the inhibitory effects of the species-specific glucocorticoids on CRH production. Thus, the final result of both glucocorticoid infusion in humans and repeated CRH injection in the rat could be similar—that is, lower exposure of the sympathoadrenal axis to the actions of CRH in response to hypoglycemia with consequent impairment of the counterregulatory response.

Antecedent ACTH treatment tended to increase further the epinephrine responses (versus corticosterone), but this difference was not significant. The antecedent ACTH group did, however, show a significant twofold greater epinephrine response as compared with control animals. This difference was not seen with the norepinephrine, perhaps because of differential effects of glucocorticoids and ACTH on the adrenal medulla (35). Secretion of epinephrine from the adrenal medulla may be suppressed by corticosterone infusion (36,37), whereas ACTH has the opposite effect (37,38). This may be explained by the differential effects of ACTH (stimulation) and corticosterone (inhibition) on local glucocorticoid levels within the portal circulation that are bathing the adrenal medulla.

The finding that recurrent exposure to CRH unlike ACTH and corticosterone results in decreased epinephrine, norepinephrine, and glucagon responses to hypoglycemia has not previously been reported. It is clear from our studies that the mechanism cannot simply be a direct effect on the production of ACTH, and in turn adrenal glucocorticoids as this would, on the basis of our data, be expected to increase rather than decrease the epinephrine response. It is noteworthy that CRH, in addition to its role within the HPA axis, plays an important part in the sympathoadrenal response to stressful stimuli. CRH, if given as an intracerebroventricular (ICV) injection, results in an elevation in plasma catecholamines, predominantly norepinephrine (39,40). ICV CRH exerts differential effects on noradrenergic activity in a variety of tissues that can be blocked with the ganglion blocker chlorisondamine (39). Acute ICV CRH administration has also been shown to increase glucagon production via an autonomic pathway (39). It has been shown that CRH acts at least in part by a direct effect on the sympathetic nervous system within the brainstem (41). A number of studies have explored the effect of local administration of CRH or CRHr blockers on specific brain nuclei (41,42). As a result, CRH has been shown to play a key role in the coordination of behavioral, endocrine, and neurological responses to stress (22,43). CRH is therefore a very strong candidate as a modulator of the stress response to hypoglycemia. In keeping with this view, it is known that the levels of CRH in the systemic circulation rise significantly in response to insulin-induced hypoglycemia (44). Moreover, whereas acute central administration of CRH activates the HPA and sympathoadrenal responses, long-term administration of CRH has been reported to reduce basal HPA activity (34) at the level of the hypothalamus. Of particular interest, both epinephrine and norepinephrine responses to CRH also have been shown to diminish with time. It is intriguing to speculate that the downregulation of the acute effect of CRH on the sympathoadrenal system might contribute to the impaired catecholamine response caused by recurrent antecedent hypoglycemia.

The site of action of the administered CRH is uncertain. Previous studies have suggested a unidirectional transport of CRH across the blood-brain barrier from the brain to the periphery, suggesting that the peripheral CRH given in this study may be acting peripherally (45). In contradiction to this view, other studies have shown that peripheral CRH administration stimulates fos activation in a variety of nuclei within the brain (46). Although the blood-brain barrier may prevent the transport of CRH into most of the brain, there are a number of areas where the blood-brain barrier is deficient, such as the median eminence and area postrema to which CRH may gain access. CRH receptors have also been shown to have a wide distribution in areas outside the brain, including the adrenal medulla and sympathetic ganglia (22,47). It is noteworthy in this regard that recent studies from our laboratory examining the influence of antecedent ICV CRH on subsequent hypoglycemic counterregulation indicate a similar level of counterregulatory impairment to that seen with peripherally administered CRH (unpublished data).

CRH acts via two receptors. The individual roles and distributions of the two receptors have not been fully elucidated, but there is some evidence that the CRHr2 receptor may be involved in downregulating the stress response. For example, CRHr2 knockout mice show enhanced anxiety behavior, an effect that was not due to changes in HPA axis activity (48). Our data would, however, support the view that the CRHr1 receptor is involved in the sympatho-adrenal downregulation that we have demonstrated. The use of a specific CRHr1 antagonist (49) in combination with antecedent CRH effectively abolishes the downregulation of the epinephrine response seen with subsequent hypoglycemia. Because the CRHr1 antagonist antalarmin readily crosses the blood-brain barrier, the site of its action is uncertain. The failure of antalarmin to increase norepinephrine levels is perhaps suggestive of an effect at the level of the adrenal medulla. CRH is known to stimulate catecholamine secretion via specific receptors in the adrenal medulla (50), although the type of receptor has not been characterized. Regardless of the site of effect, recurrent exposure to CRH seems to be downregulating the CRHr1 and antalarmin seems to be blocking this effect. Equally as there is central neural control of adrenal medullary activity, these results do not necessarily suggest that the effects that we are seeing are entirely peripheral (39).

In summary, our data implicate CRH as the mediator of the sympathoadrenal downregulation seen after recurrent antecedent hypoglycemia in a rodent model of hypoglycemia-associated autonomic failure and provide evidence that this effect of CRH is not simply a consequence of activation of the HPA axis.

This work was supported by grants from the National Institutes of Health (DK 20495 and DK45735) and by the Juvenile Diabetes Research Foundation Center for the study of Hypoglycemia. D.E.F. was supported by a mentor-based fellowship from the American Diabetes Association.

We are grateful for the assistance of Aida Groszman and Andrea Belous for hormone analysis and of Jianying Dong and Hiahong Zong for performing the surgical procedures.