Somatostatin Receptor Antagonism Reverses Glucagon Counterregulatory Failure in Recurrently Hypoglycemic Male Rats

Abstract

Recent antecedent hypoglycemia is a known source of defective glucose counter-regulation in diabetes; the mechanisms perpetuating the cycle of progressive α-cell failure and recurrent hypoglycemia remain unknown. Somatostatin has been shown to suppress the glucagon response to acute hypoglycemia in rodent models of type 1 diabetes. We hypothesized that somatostatin receptor 2 antagonism (SSTR2a) would restore glucagon counterregulation and delay the onset of insulin-induced hypoglycemia in recurrently hypoglycemic, nondiabetic male rats. Healthy, male, Sprague–Dawley rats (n = 39) received bolus injections of insulin (10 U/kg, 8 U/kg, 5 U/kg) on 3 consecutive days to induce hypoglycemia. On day 4, animals were then treated with SSTR2a (10 mg/kg; n = 17) or vehicle (n = 12) 1 hour prior to the induction of hypoglycemia using insulin (5 U/kg). Plasma glucagon level during hypoglycemia was ~30% lower on day 3 (150 ± 75 pg/mL; P < .01), and 68% lower on day 4 in the vehicle group (70 ± 52 pg/mL; P < .001) compared with day 1 (219 ± 99 pg/mL). On day 4, SSTR2a prolonged euglycemia by 25 ± 5 minutes (P < .05) and restored the plasma glucagon response to hypoglycemia. Hepatic glycogen content of SSTR2a-treated rats was 35% lower than vehicle controls after hypoglycemia induction on day 4 (vehicle: 20 ± 7.0 vs SSTR2a: 13 ± 4.4 µmol/g; P < .01). SSTR2a treatment reverses the cumulative glucagon deficit resulting from 3 days of antecedent hypoglycemia in healthy rats. This reversal is associated with decreased hepatic glycogen content and delayed time to hypoglycemic onset. We conclude that recurrent hypoglycemia produces glucagon counterregulatory deficiency in healthy male rats, which can be improved by SSTR2a.

Recurrent hypoglycemia is a pervasive clinical complication of insulin therapy in type 1 diabetes (T1D), owing to imperfect (nonphysiological) insulin replacement in a setting of defective glucose counterregulation (1). The α-cell, which secretes the chief counter-regulatory hormone, glucagon, becomes unresponsive to hyperinsulinemic hypoglycemia with progressive β-cell failure and endogenous insulin deficiency (1). Within about 5 years of T1D diagnosis, the counterregulatory burden has shifted to sympathoadrenal and other autonomic mechanisms (1). However, these pathways are easily overwhelmed by intensive insulin therapy, commonly leading to recurrent hypoglycemia and the clinical syndrome of hypoglycemia-associated autonomic failure (HAAF) (1). In short, recent antecedent hypoglycemia lowers glycemic thresholds for sympathoadrenal activation to subsequent hypoglycemia, resulting in diminished counterregulatory and neuroglycopenic symptom responses (1).

It is important to note that the literature models HAAF as the mechanism perpetuating recurrent hypoglycemia (1). This model assumes a state in which the α-cell is already completely unresponsive to hypoglycemia; however, at least a partial glucagon response is typically present in cases of newly diagnosed T1D (≤5 years) and in late stage type 2 diabetes even before HAAF develops (2, 3). In these and healthy animal models, it is well established that glucagon counterregulation is attenuated by acute, insulin-induced hypoglycemia (4, 5) and diminishes further with repeat hypoglycemia exposure (6, 7). Unlike epinephrine, glucagon secretion in diabetic rats is not corrected by short-term (3-4 weeks) hypoglycemia avoidance (8, 9). Moreover, the glucagon response to hypoglycemia is preserved where neural inputs have been severed in vivo (cervical spinal cord transection, partial or complete pharmacological adrenergic blockade, pancreas transplant) (10–12) and in vitro (perfused rodent pancreas and perifused rodent and human islets) (13, 14). Conversely, a sustained rise in intra-islet insulin levels during hypoglycemia induced with the insulin secretagogue, tolbutamide, silences the glucagon response to hypoglycemia despite intact autonomic signaling (15). Collectively, these data suggest the involvement of non-neural mechanisms, potentially operating in parallel to HAAF, that underpin defective glucagon counterregulation to recurrent hypoglycemia.

Somatostatin-14 is an endocrine hormone secreted by δ-cells of the endocrine pancreas (16). As the gatekeeper of the pancreatic islets, somatostatin exerts tonic inhibitory control over neighboring α- and β-cells by activating somatostatin receptor type 2 (SSTR2) and types 3/5 (SSTR3/5), localized to α- and β-cells, respectively (17). Reported elevations of pancreatic and plasma somatostatin in diabetic humans (18), dogs (19), and rodents (20–22) have long implicated δ-cell dysfunction in the failed α-cell response to insulin-induced hypoglycemia in T1D (16). More recently, Rorsman’s group demonstrated that a therapeutic dose of insulin inhibited hypoglycemic glucagon secretion by an indirect, paracrine mechanism mediated by somatostatin—a feature that was reversible with somatostatin type 2 receptor antagonism (SSTR2a) (23).

Pharmacological SSTR2a recently advanced to phase 1 clinical trials as a therapeutic tool for hypoglycemia prevention in T1D (24). In preclinical trials, SSTR2a was shown to restore glucagon counterregulation in response to acute (25, 26) and recurrent (27) episodes of clamped hypoglycemia in rodent models of streptozotocin and biobreeding T1D, and in healthy, human pancreas slices (16).

In this study, we used a nondiabetic rat model of hypoglycemia-induced counterregulatory failure to determine whether somatostatin signaling mediates the suppressive effect of recent, antecedent hypoglycemia on subsequent glucagon counterregulation. To this end, we evaluated the restorative potential of pharmacological SSTR2 antagonism on hormonal (glucagon and C-peptide) and glycemic responses to hypoglycemia after 3 prior episodes in nondiabetic rodents. The absence of pre-existing α-cell dysfunction in our model enabled us to selectively target and reverse the glucagon secretory defect secondary to recurrent hyperinsulinemic hypoglycemia per se. We hypothesized that SSTR2 blockade would rescue the glucagon secretory response and its catabolic effects on hepatic glycogen stores to resist subsequent hypoglycemia in recurrently hypoglycemic, nondiabetic, male Sprague–Dawley rats.

Materials and Methods

This study was conducted in accordance with the recommendations of the Canadian Council for Animal Care guidelines and has been approved by the York University Animal Care Committee (Protocol # 2017-7).

Rodent Treatment and Experimental Design

Forty-five healthy, nondiabetic, male Sprague–Dawley rats (Charles River Laboratories, ~250 g body mass, postweaned, age ~9-10 weeks) were habituated in the York University Vivaria for 1 week before handling. Rats were housed in a light controlled (12-hour light/dark cycle) room with a humidity of 50% to 60% and temperature of 22°C to 23°C. Rats had ad libitum access to standard rodent chow (Purina Labdiet 5012, St. Louis, MO, USA) and water. Following the 7-day habituation period, body weight, blood glucose, and food intake were monitored daily. Figure 1A provides an outline of the study design, with an itemized breakdown of animal subgroupings, animal exclusions, and analytes measured for each condition.

Figure 1.

Study flow diagram and experimental protocols. (A) Flow diagram. Shaded boxes denote terminal conditions. (B) Hypoglycemia conditioning protocol (days 1-3). Thirty-nine healthy rats received a daily bolus injection of Humulin-R insulin to induce similar hypoglycemia on 3 consecutive days using oral D-glucose as necessary. Blood glucose was measured before insulin administration (t = 0 minutes) and every 10 minutes thereafter for the duration of the protocol. A venous blood sample was collected at t = 0 minutes and again during hypoglycemia (ie, first blood glucose measurement ≤3.5 mmol/L) for subsequent glucagon and insulin analysis. (C) Subsequent hypoglycemia (day 4) with or without SSTR2a treatment. Hypoglycemia-conditioned rats were administered SSTR2a (PRL-2903, 10 mg/kg; n = 17; IP) or vehicle (n = 12) 1 hour (t = –60 minutes) before a bolus injection of insulin (Humulin-R insulin, 5 U/kg, IP) at t = 0 minutes. A terminal blood sample was collected at blood glucose ≤3.5 mmol/L and animals were anesthetized for tissue and portal vein blood extraction prior to euthanizing.

Baseline Controls (Day 1)

Following habituation, a subset of rats (n = 6) was randomly selected for baseline analysis and excluded from all hypoglycemia protocols. A blood sample was collected via saphenous vein bleed for baseline (basal/fed) measurement of plasma glucagon and C-peptide concentrations. These rats (and all others in this study at their respective endpoints) were then anaesthetized using isoflurane gas for portal blood and tissue collection (see below).

Hypoglycemia Conditioning (Days 1-3)

Hypoglycemia has been clinically defined by the American Diabetes Association Workgroup on Hypoglycemia as a whole blood glucose concentration of ≤3.9 mmol/L (28). In this study, we undertook a 3-day hypoglycemia conditioning protocol to induce defective glucagon counterregulation in healthy rats. The experimental procedure for hypoglycemic conditioning is outlined in Fig. 1B. Thirty-nine rats underwent recurrent insulin-induced hypoglycemia, comprising 1, 120-minute hypoglycemic event per day, on 3 consecutive days (food was removed ~60 minutes prior to each event). A stepwise reduction in the insulin dose was used to induce a similar level of hypoglycemia on each day (day 1: 10 U/kg, day 2: 8 U/kg, day 3: 5 U/kg, Humulin-R, Lilly, Canada). This insulin dose reduction over days 1 to 3 was designed in anticipation of diminished counterregulatory function over the 3-day conditioning period (29). Blood glucose was measured in duplicate via tail prick using a hand-held glucometer (ΑTRAK, Abbott) before insulin administration (t = 0 minutes) and every 10 minutes thereafter until each animal had remained between 1.7 and 2.5 mmol/L for 120 consecutive minutes. This is an established protocol for reproducing overt and progressive glucagon failure in healthy rodents (30–32). Animals that failed to reach the glycemic target by 60 minutes of their initial insulin injection received a second dose of insulin. If blood glucose levels dropped below range at any time, 0.1-1.0 mL of 35% D-glucose was administered, as necessary, to restore target hypoglycemia. If physically capable, animals drank the dextrose solution from a syringe; otherwise, it was delivered via oral gavage. Any animal that showed signs of severe hypoglycemia (ie, seizures and/or loss of consciousness) was immediately withdrawn from the study and excluded from all data analyses (see Fig. 1A for 1 such exclusion).

Saphenous vein blood samples were collected for subsequent glucagon and C-peptide analysis prior to insulin injection (t = 0 minutes) and at the first blood glucose measurement ≤3.5 mmol/L (in duplicate samples). This glycemic target was chosen to ensure that the glucagon response was captured beyond its glycemic activation threshold of ~3.6 to 3.9 mmol/L (33), especially since this threshold can shift to lower blood glucose concentrations with recurrent hypoglycemia (34).

Hypoglycemia-conditioned Controls

Following 3 days of recurrent hypoglycemia, 7 rats were randomly selected for terminal basal analysis ~24 hours after their third hypoglycemic event (ie, on the morning of day 4). These rats were anaesthetized for tissue and portal blood collection in the euglycemic state for comparison with baseline controls naive to hypoglycemia. One animal suffered a seizure after hypoglycemia conditioning on day 3 and was euthanized without advancing to day 4 (Fig. 1A). Their data were excluded from all analyses.

Subsequent Hypoglycemia (Day 4)

The experimental protocol for day 4 is provided in Fig. 1C. To assess the effects of recurrent hypoglycemia on glucagon counterregulation, with and without SSTR2a, a subset of rats (n = 31) was subjected to the same insulin induction protocol as day 3, but without any oral glucose provided. More rats were allocated to the drug-treated group (n = 17) than the vehicle group (n = 14), based on our previous findings that the glucagon response to SSTR2a is more variable than to vehicle in both healthy and diabetic rodents (25). Two rats in the vehicle group developed baseline hypoglycemia prior to insulin bolus on day 4 and were excluded from all data analyses (Fig. 1A; see below for possible limitation of αTOS as drug vehicle).

Duplicate blood glucose measurements were obtained via tail prick using a hand-held glucometer at t = –60 minutes (baseline), t = –30 minutes, t = 0 minutes (before insulin administration), and every 10 minutes thereafter until a glycemic endpoint of ≤3.5 mmol/L. At this time, a terminal blood sample was collected via saphenous venipuncture for subsequent hormone analysis and rats were anaesthetized for portal blood and tissue extraction.

Experiments were conducted in batches and technicians were blinded to the treatment condition.

SSTR2a and Vehicle

The SSTR2a used (PRL-2903) has a half-maximal inhibitory concentration of 2.5 nmol/L and binds to SSTR2 with a K_i of 26 nmol/L (35). This peptide is selective for SSTR2 over SSTR3 and SSTR5 by 10- and 20-fold, respectively, and has negligible binding affinity for SSTR1 and SSTR4 (35). Rats received an intraperitoneal (IP) injection of SSTR2a (10 mg/kg PRL-2903, IP; formulated by CPC Scientific, Sunnyvale, CA, USA; supplied by CDRD, BC, Canada) with vehicle (0.5% α-tocopheryl succinate [αTOS] in pH 7 phosphate-buffered saline, IP) or vehicle alone (controls) 1-hour (t = –60 minutes) prior to a single intraperitoneal injection of 5 U/kg Humulin-R insulin. The vitamin E prodrug, αTOS, was chosen over acetic acid, the more commonly used vehicle for PRL-2903 delivery (25–27) based on observations that αTOS offers (1) improved drug solubility; (2) less pain on administration, which could elicit a confounding stress response; and (3) increased stability and predictability of drug absorption. Unlike its active tocopherol (vitamin E), αTOS does not have a redox potential and does not act as an antioxidant (36).

Plasma Analysis

Blood samples from saphenous vein bleed (or portal vein at endpoint on day 4) were collected in potassium-EDTA coated, microvette capillary tubes (Sarstedt, Cat # 16.444.100, Canada) and centrifuged at 12 000 rpm for 5 minutes. Plasma was removed and stored in polyethylene tubes at –80°C for subsequent quantification of glucagon (Mercodia Cat# 10-1271-01, RRID:AB_2737304, http://antibodyregistry.org/AB_2737304) and C-peptide (Crystal Chem Cat# 90055, RRID:AB_2893130) levels using ELISA.AB_2737304), and C-peptide (Crystal Chem Cat# 90055, RRID:AB_2893130, http://antibodyregistry.org/AB_2893130) levels using enzyme-linked immunosorbent assay.

Tissue Analysis

Skeletal muscle (tibialis anterior and extensor digitorum longus) and liver glycogen content were quantified using a method modified from Carr and Neff (37). Briefly, frozen tissue was digested in 0.5 mL of 1 M KOH at 70°C for 1 hour, and the pH of the digest was titrated to 4.8 prior to the addition of acetate buffer and 0.5 mg/mL amyloglucosidase. Glycogen was subsequently hydrolyzed at 40°C for 2 hours. Glucose was detected enzymatically, and the absorbance was read at 340 nm in a spectrophotometer (Ultraspec 2100 pro; Biochrom Ltd, Cambridge, UK).

Statistical Analysis

All data are expressed as means ± SD. Statistical tests were conducted against a significance criterion of P < .05 using Prism software (GraphPad Software, San Diego, CA). Because treatment was administered on day 4 only, we elected to use a 1-way, independent samples analysis of variance (ANOVA) (factor: study group; levels: day 1, day 3, vehicle, SSTR2a) rather than a 2-way, mixed-effects ANOVA (factors: day × treatment) to analyze the following outcomes: basal or hypoglycemic plasma hormone levels, time to hypoglycemia onset, and tissue glycogen content. Multiple comparisons were performed using Tukey’s HSD post hoc test. Day 4 glycemic responses were analyzed using a 2-way mixed-design ANOVA (factors: treatment × time), followed by a Sidak post hoc test. A 2-tailed, unpaired t-test was used to compare portal vein hormone levels between day 4 treatment groups. Survival curve analysis was performed using a log-rank (Mantel–Cox) test.

Results

Body Weight and Blood Glucose

Body weight increased steadily over the experimental period, from 321 ± 67 g at baseline on day 1 to 334 ± 66 g on day 4 (P < .001) but did not vary significantly between day 4 treatment groups (vehicle: 354 ± 59 g; SSTR2a: 340 ± 64 g). Blood glucose concentrations from the hypoglycemia conditioning phase of the protocol (days 1-3; n = 36) are presented in Fig. 2A, while Fig. 2B shows the glycemic responses to subsequent insulin challenge (day 4) with or without SSTR2a. Basal blood glucose levels at t = –60 minutes were not significantly different across experimental days or conditions (Fig. 2A and 2B). Glucose levels during hypoglycemia conditioning on days 1 to 3 were similar, with all animals remaining in target range for 120 minutes. As described in the methods, plasma and tissue samples were collected at the first blood glucose reading ≤3.5 mmol/L. As a result, hypoglycemic outcome variables corresponded to a mean blood glucose concentration of 3.2 ± 0.4 mmol/L on day 1, 3.1 ± 0.4 mmol/L on day 3, 3.0 ± 0.4 mmol/L in vehicle (day 4), and 3.1 ± 0.4 mmol/L in SSTR2a (day 4) (not significantly different).

Figure 2.

Blood glucose responses to hypoglycemic challenge on hypoglycemia conditioning days (A). On day 4 (B), hypoglycemia-conditioned rats were given SSTR2a (10 mg/kg; n = 17) or vehicle (n = 12) at t = –60 minutes, followed by a 5 U/kg IP bolus injection of insulin 1 hour later (t = 0 minutes, indicated by vertical dotted line). Terminal blood glucose values from each animal have been extended out to the x-axis limit. Horizontal dotted line indicates the threshold for clinical hypoglycemia (3.9 mmol/L). Treatment by time interaction: P < .001; P = .06 at t = 10, 20 minutes. (C) Time from insulin administration to hypoglycemia onset on days 1-3 (hypoglycemia conditioning) and day 4 (subsequent hypoglycemia) with or without SSTR2a pretreatment. *P < .05 vehicle vs SSTR2a. (D) Survival curve comparing the percentage of animals in each group that remained euglycemic (>3.9 mmol/L) postinsulin exposure on day 4. *P < .05. All data are mean + SD. Data in the tables below (A) and (B) indicate the number of animals remaining at each major interval time point.

In Fig. 2B, each animal’s terminal blood glucose concentration as measured on day 4 (at ≤3.5 mmol/L) was carried forward to all subsequent timepoints following their removal from the challenge. This data extrapolation procedure was done to illustrate glycemic trends over time, despite diminishing sample sizes, which would otherwise misrepresent group means. There was a significant treatment by time effect on glycemic responses to subsequent insulin challenge (day 4; P < .001) plotted in Fig. 2B, with elevations in the SSTR2a group at 10 and 20 minutes after insulin administration (P < .05 and P = .06, respectively).

Time to Hypoglycemia Onset

Time from insulin administration to the onset of clinical hypoglycemia (blood glucose ≤3.9 mmol/L) did not differ significantly during the hypoglycemic conditioning days, averaging 50 ± 39 minutes (days 1-3 combined) (Fig. 2C). On day 4, the drug-treated group reached hypoglycemia by an average of 58 ± 40 minutes compared with 33 ± 34 minutes in the vehicle group (P < .05). A comparison of survival curves in Fig. 2D revealed a significant overall reduction in the proportion of hypoglycemic animals in the treatment vs vehicle group (P < .05). Three of 12 rats (25%) in the vehicle group were euglycemic at 30 minutes compared with 9 of 17 rats (53%) in the treatment group.

Glucagon

Mean plasma glucagon concentrations on experimental days 1, 3, and 4 are shown in Fig. 3A. Basal glucagon levels did not change significantly throughout the experimental period. The stimulated glucagon response to blood glucose ≤3.5 mmol/L diminished by 32% across 3 days of conditioning (day 1: 219 ± 99 pg/mL vs day 3: 150 ± 75 pg/mL; P < .01) and by an additional 53% on day 4 in the vehicle group (day 4 vehicle: 70 ± 52 pg/mL; P < .001 vs day 1; P < .05 vs day 3). Plasma glucagon was ~2.4-fold higher in SSTR2a-treated animals (171 ± 122 pg/mL) than vehicle controls (P < .05) and similar to levels observed on days 1 and 3. Portal vein glucagon levels did not vary significantly between groups on day 4 (Fig. 3B).

Figure 3.

Basal and hypoglycemic plasma hormone levels. Baseline (squares) and endpoint (circles) plasma concentrations of (A) systemic glucagon; (B) portal vein glucagon (day 4 only); (C) systemic C-peptide; (D) portal vein C-peptide (day 4 only); and (E) the ratio of portal glucagon to portal C-peptide. Note: hypoglycemic measurements were obtained at the first blood glucose reading ≤3.5 mmol/L. *P < .05, **P < .01, ***P < .001 vs vehicle; ††P < .01 vs day 3; ‡‡‡P < .001 vs day 1 and day 3. All data are means + SD.

C-peptide

Basal concentrations of plasma C-peptide were comparable across all 4 days (Fig. 3C). Hypoglycemic levels (measured at blood glucose ≤3.5 mmol/L) of C-peptide were also comparable between days 1 and 3 but increased significantly on day 4 (vehicle: 0.59 ± 0.23 nmol/L; SSTR2a: 0.36 ± 0.20 nmol/L) relative to day 1 (0.15 ± 0.09 nmol/L; P < .001 vs vehicle or SSTR2a) and day 3 (0.14 ± 0.11 nmol/L; P < .001 vs vehicle or SSTR2a) (Fig. 3B). This marked elevation in the vehicle group was partially, yet significantly, offset by SSTR2a pretreatment, and was mirrored by a comparable, 1.6-fold reduction in portal vein C-peptide levels of SSTR2a-treated animals (vehicle: 0.32 ± 0.11 nmol/L vs SSTR2a: 0.19 ± 0.13 nmol/L; P < .01) (Fig. 3D).

Portal Vein Glucagon-to-Insulin Ratio

The calculated ratio of glucagon to C-peptide in the portal vein during hypoglycemia on day 4 was ~2.6-fold higher in the SSTR2a (2014 ± 1698) vs vehicle group (776 ± 406; P < .05) (Fig. 3E).

Hepatic and Skeletal Muscle Glycogen Content

The hepatic glycogen content of euglycemic, baseline controls was not significantly different after 3 days of recurrent hypoglycemia (ie, measured 24 hours after the third hypoglycemic exposure) (Table 1). On day 4, hepatic glycogen stores were significantly lower in hypoglycemic animals than in their euglycemic counterparts (ie, hypoglycemia-conditioned controls: P < .05 vs vehicle; P < .001 vs SSTR2a). Notably, glycogen levels were reduced by 35% in the SSTR2a vs vehicle group after hypoglycemia (P < .01). No difference was detected in the glycogen content of the tibialis anterior or extensor digitorum longus muscle across days or treatment conditions (Table 1).

Table 1.

Tissue glycogen content of control animals in the basal (euglycemic) state and of hypoglycemic animals at target blood glucose ≤3.5 mmol/L

	Day 1	Day 4
	Baseline controls	Hypoglycemia-conditioned controls	Subsequent hypoglycemia
			Vehicle	SSTR2a
Hepatic glycogen (µmol/g)	27 ± 0.4	26 ± 1.9	20 ± 7.0*	13 ± 4.4***††
TA glycogen (µmol/g)	12 ± 1.9	12 ± 2.5	13 ± 3.2	11 ± 1.9
EDL glycogen (µmol/g)	11 ± 1.5	11 ± 2.0	11 ± 3.4	11 ± 1.7

Data are means ± SD.

Abbreviations: TA, tibialis anterior; EDL, extensor digitorum longus.

*P < .05, ***P < .001 vs baseline controls or hypoglycemia-conditioned controls; ††P < 0.01 vs vehicle.

Discussion

The goal of this study was to examine whether somatostatin signaling plays a role in the pathophysiology of hypoglycemia-induced counterregulatory failure. As hypothesized, SSTR2a restored glucagon secretion to subsequent hypoglycemia, which in turn, increased hepatic glycogen utilization and delayed time to hypoglycemia onset.

Antecedent hypoglycemia is known to blunt the glucagon response to hypoglycemia in healthy (4-6, 38) and diabetic subjects (39). Here, our nondiabetic rat model showed a 32% attenuation in counterregulatory glucagon secretion after 3 days of hypoglycemia conditioning. This response echoed the ~40% to 60% impairment reported by Chan et al. across a similar 3-day protocol (40). By reducing the insulin dose on each successive day of conditioning, we achieved a consistent rate of glycemic decline despite failing counterregulation. Without this compensatory dosing measure, vehicle-treated rats showed a further 53% impairment to subsequent hypoglycemia on day 4.

We demonstrated that delivery of the SSTR2a (10 mg/kg PRL-2903) 1 hour before hypoglycemia induction normalized plasma glucagon levels and delayed the onset of clinical hypoglycemia (blood glucose ≤3.9 mmol/L) by 25 minutes. The absence of pre-existing α-cell dysfunction in our nondiabetic rodent model allowed us to selectively target and reverse glucagon counterregulatory failure induced by recurrent hyperinsulinemic hypoglycemia per se. The lack of an observable drug effect in acutely hypoglycemic healthy rodents further supports the conclusion that somatostatin signaling in the pancreatic islets may increase with repeated hypoglycemia exposure (26). Abnormal elevations in plasma and/or pancreatic somatostatin have traditionally been considered a diabetes-specific phenomenon (16, 26). The rescue of glucagon counterregulation by SSTR2a in various diabetic rodent models during acute (25, 26) and recurrent (27) hypoglycemia suggests that somatostatin signaling may attenuate glucagon counterregulation (16, 27). We propose that in addition to diabetes, antecedent hypoglycemia per se may elevate somatostatin signaling to further suppress glucagon counterregulation.

Proinsulin connecting peptide (C-peptide) is a byproduct of insulin biosynthesis, secreted with insulin in an equimolar ratio. It therefore serves as a reliable metric for measuring β-cell secretory activity in the presence of exogenous insulin. We unexpectedly observed significantly higher levels of systemic C-peptide during hypoglycemia on day 4 than on days 1 and 3, which was more pronounced in the vehicle-only group. This suggests that the vehicle, αTOS, may have acted as an insulin secretagogue prior to and/or during hypoglycemia. However, because the vehicle was given to both groups, any potential paracrine or glycemic interference would not modify relative treatment outcomes. Insulin secretagogues are known to suppress hypoglycemic glucagon secretion under healthy conditions (15), and we propose that the SSTR2a used here may have inadvertently opposed this insult by reducing pharmacologically stimulated insulin secretion. Any nonspecific binding of the antagonist to SSTR3 and/or SSTR5 on the β-cell would be expected to raise C-peptide levels by blocking somatostatin’s inhibitory input (41). We instead propose that by occupying somatostatin binding sites on the α-cell, the antagonist may have promoted the somatostatin-mediated activation of its receptors on the β-cell, thereby inhibiting (and partially normalizing) hypoglycemic insulin release. Confirmed elevations in pancreatic and plasma somatostatin during PRL-2903 exposure, based on in vitro studies of the perfused isolated rat pancreas (41) and hypoglycemia clamp experiments in healthy rats (25), may lend further support to this theory.

Glucagon-stimulated glycogenolysis and gluconeogenesis by the liver are the predominant counterregulatory pathways that defend against hypoglycemic stressors (42). T1D reduces liver glycogen stores (43), even in those under good glycemic control (44). The reduction in liver glycogen is associated with impaired rates of glycogenolysis during hypoglycemia in humans with T1D (44) and in dog models of altered hepatic glycogen content (45). Interestingly, we did not observe any change in basal hepatic glycogen stores following 3 days of hypoglycemic conditioning in these nondiabetic rodents. Diminished glycogenolytic capacity in recurrently hypoglycemic rats is supported by Shum et al., who reported impairments in both absolute and incremental glucose production during the first 45 minutes of insulin-induced hypoglycemia, despite an intact neuroendocrine response (46). Previous studies proposed that SSTR2a promotes homeostatic recovery from hypoglycemia by augmenting the rate of hepatic glycogenolysis (27), though none had directly tested this hypothesis. We observed a 45% depletion of hepatic glycogen in rats treated with SSTR2a, indicating its ability to protect against hypoglycemia following a period of untreated, recurrent hypoglycemia. While hepatocytes have been shown to express SSTR1 and SSTR3 (47), the SSTR2 antagonist used in this study (PRL-2903) is incompatible with the former and binds to the latter with 10-fold lower affinity than SSTR2 (35). This supports the conclusion that somatostatin modulates hepatic glycogen stores by an indirect mechanism—likely via suppression of glucagon counterregulation (27).

Insulin antagonizes glucagon’s catabolic effects on liver glycogen, so the ratio of glucagon to insulin in the portal vein (which serves as a direct conduit from the pancreas to the liver) is considered the primary mediator of hepatic glycogen breakdown during mild hypoglycemia and exercise (48). Accordingly, the elevated portal vein ratio observed in SSTR2a-treated vs vehicle control animals may account for the significant reduction in hepatic glycogen observed in this group. However, to conclude whether glycogenolytic capacity was truly attenuated by antecedent hypoglycemia in this model, future studies must establish the glycogen response to acute hypoglycemia for comparison against each day of subsequent exposure.

We found that the circulating, but not portal concentration of plasma glucagon was significantly higher in drug- vs vehicle-treated animals at terminal hypoglycemia on day 4. This finding may invite speculation of a SSTR2a-mediated reduction in hepatic glucagon clearance; however, no such effects have been described. Instead, studies in isolated rat islets and pancreas slices have consistently demonstrated the potent and direct α-cell glucagonotropic effects of selective SSTR2 antagonists (25, 49). We suspect that our sampling timepoint, which took place 60 minutes after insulin administration in the SSTR2a group (vs 30 minutes in the vehicle group), may have captured a time-dependent decline in glucagon secretion, shown to occur between ~30 and 40 minutes of hypoglycemic induction with or without SSTR2a (26, 27, 50).

This study has several limitations to acknowledge. First, we elected to use αTOS as the vehicle for SSTR2a delivery. Long-term supplementation of vitamin E in its active form, α-tocopherol, has been shown to compensate for insulin resistance by increasing glucose-stimulated insulin release and peripheral insulin sensitivity in hyperglycemic, type 2 diabetic rats (51, 52). Evidence of a single-dose effect, especially in healthy subjects, is less convincing (51, 53, 54). It is also worth clarifying that αTOS is a prodrug, which shows slow and continuous hydrolysis (activation) by a tissue esterase in pharmacokinetic studies of intravenously infused rats (55). A peak in serum tocopherol levels 7 to 8 hours after delivery and minimal nonhepatic tissue exposure overall (55) further reduce the likelihood of vehicle interference. We cannot discount the possibility that αTOS promoted insulin secretion and/or peripheral insulin sensitivity in both treatment groups, owing to the premature glycemic decline following SSTR2a/vehicle administration (Fig. 2B). The use of α-tocopheryl succinate should thus be avoided in future antihypoglycemic drug formulations. As a second limitation, plasma glucagon and C-peptide levels were measured at a single hypoglycemic timepoint, which did not allow us to profile these responses over time. Finally, the majority of glucagon’s counterregulatory response lies beyond the glycemic threshold for clinical hypoglycemia onset (3.9 mmol/L) (33), reflected in our glycemic sampling target of 3.5 mmol/L, which questions a role for glucagon in resisting hypoglycemia onset (vs its progression) under normal conditions. Thus, it remains to be determined how the elevation in plasma glucagon levels as measured at 3.5 mmol/L relates to a delay in hypoglycemia onset at 3.9 mmol/L. Notably, SSTR2a infusion prevents glucose-mediated glucagon suppression between 3.5 and 5 mmol/L in the isolated rat pancreas, suggesting that SSTR2a may promote earlier glucagon recruitment and glycemic recovery (41). Future studies are required to elucidate how SSTR2a may modify the glycemic threshold for glucagon counterregulation in this and other models of health and disease.

In summary, we demonstrated that the SSTR2a, PRL-2903, restored glucagon counterregulation and significantly delayed the onset of subsequent hypoglycemia in recurrently hypoglycemic, nondiabetic rats. Hypoglycemia resistance with SSTR2a treatment is associated with accelerated hepatic glycogenolysis. Our findings suggest a role for somatostatin in the glucagon counterregulatory defect that develops in parallel with HAAF. Further investigation into these findings stands to advance our understanding of intra-islet paracrine dynamics in health and disease, as well as the therapeutic applications of SSTR2a in diabetic patients with advanced counterregulatory failure.

Glossary

Abbreviations

αTOS

α-tocopheryl succinate

ANOVA

analysis of variance

HAAF

hypoglycemia-associated autonomic failure

IP

intraperitoneally

SSTR2a

somatostatin receptor 2 antagonism

T1D

type 1 diabetes

Additional Information

Disclosure Summary: Richard Liggins is an employee of Zucara Therapeutics who provided the SSTR2a. All other authors declare that there are no relationships or activities that might bias, or be perceived to bias, their work.

Data Availability

Some or all datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.