High-Intensity Exercise as a Dishabituating Stimulus Restores Counterregulatory Responses in Recurrently Hypoglycemic Rodents

Impaired awareness of hypoglycemia (IAH), defined as “a diminished ability to perceive the onset of acute hypoglycemia,” affects 20–25% of all people with type 1 diabetes (T1D) and increases the risk of severe hypoglycemia sixfold (1). IAH develops primarily in response to repeated episodes of hypoglycemia (2). Individuals with or without T1D, who experience one or more episodes of hypoglycemia have impaired hormonal and symptomatic counterregulatory responses (CRRs) during a further episode of hypoglycemia, with the extent of suppression dependent on the depth, duration, and frequency of antecedent hypoglycemia (2). Conversely, the CRR can be restored if hypoglycemia is avoided (3). The mechanisms underpinning the development of IAH remain unknown but are likely a result of changes in specialized glucose-sensing regions of the body such as those found in distinct brain regions like the ventromedial hypothalamus (VMH) (4).

Habituation is a type of adaptive memory that occurs in many organisms in response to a repeated, often stressful stimulus (5). Habituation is defined as a “reduction of the psychological, behavioral or physiological response to a stimulus as a result of repeated or prolonged exposure” (5,6). Aspects of the physiological responses to repeated hypoglycemia, such as the progressive diminishment of CRR to hypoglycemia after repeated exposure, are consistent with the principal features of a habituated response (5–7), suggesting that this may provide an explanation for IAH. In this study, we directly address this hypothesis in a rodent model of recurrent hypoglycemia (RH) by demonstrating one of the defining features of a habituated process, namely that it is possible to rapidly restore the habituated response by dishabituation, the introduction of a novel strong stimulus (5,6).

Male Sprague-Dawley rats (250–300 g; Harlan UK) maintained on a 12-h day/night cycle and provided with food and drinking water ad libitum. Experimental procedures were approved by the University of Dundee Ethical Review Process and performed in accordance with U.K. Home Office regulations.

To test our hypothesis in rodents, high-intensity (HI) exercise was used as a dishabituating stimulus with RH as the habituated response. After 2 weeks of daily handling, rats underwent insulin-induced (0.75–1 units/kg i.p., Novorapid; NovoNordisk Ltd) RH (N = 24) or intraperitoneal volume-matched saline injections (Control; N = 16) three times weekly for 4 weeks (Fig. 1A and B). Animals were familiarized (5 m/min for 15 min daily) with the treadmill during the last 2 weeks, with exercise training in the morning and intraperitoneal injection in the late afternoon. Subsequently, vascular catheters were inserted under general anesthesia as previously described (8), and animals recovered over 5 days. On day 5 postsurgery, a further insulin or control injection was administered, and then on day 6, after feeding freely overnight, the animals were allocated to one of the following groups: 1) no exercise, 2) low-intensity (LI) exercise (15-min “walking” pace of 5 m/min with a 10% incline), or 3) HI exercise (5 min at 5 m/min followed by an incremental increase in speed at 2 m/min intervals to 5 min of “running” at 15 m/min). Where required, exercise was encouraged using a bottlebrush to the tail. Animals were then returned to their home cages and fed ad libitum. The following day, all animals underwent a hyperinsulinemic-hypoglycemic (2.8 mmol/L) clamp (8), with sampling for CRR hormones under euglycemic and hypoglycemic conditions (Fig. 1C). Animals were euthanized on completion of the study, and brains were removed for later analysis.

To examine for response recovery, an enhanced response to the novel stimulus and to gain insight into the potential mechanism, a further two groups of male Sprague-Dawley rats (n = 12–18 per group) were studied, and the impact of each exercise modality on counterregulatory hormones, cytokines, brain-derived neurotrophic factor (BDNF), and lactate was measured. The induction of hypoglycemia, familiarization with the treadmill, surgical procedures, and exercise were identical to the protocol described for experiment 1. However, immediately after each of the three exercise protocols blood samples were obtained from the carotid artery (N = 6–8 per group). The animals were then euthanized.

Hormones, cytokines, and metabolites were assessed as follows: glucose (Analox GM9 glucose analyzer; Analox Instruments Ltd.); insulin and glucagon (multiplex ELISA, Bio-Plex; Bio-Rad); epinephrine and corticosterone (ELISA; ALPCO and Demeditic Diagnostics, respectively); BDNF (DuoSet ELISA; R&D Systems); lactate levels (Sigma-Aldrich, Irvine, U.K.); and cytokines [V-PLEX Proinflammatory Panel 2 (rat) ELISA; Meso Scale Diagnostics].

Proteins extracted from each VMH sample were reduced, alkylated, and subjected to trypsin digestion. The tryptic peptides from each animal were labeled with iTRAQ (AB Sciex) labeling according to the manual of the manufacturer. All labeled samples were then pooled together. One-half of the pooled sample was subjected to high-pH off-line C18-based fractionation (eight fractions), and the other half was subjected to strong cation exchanger fractionation (five fractions). Liquid chromatography tandem mass spectrometry was carried out as previously described (9). iTRAQ quantification was carried out using Peaks software (Bioinformatics Solutions Inc.) with the International Protein Index rat database (2012–09–27). All samples were randomized, and the analyst was blinded to sample grouping during the processing.

All results are expressed as the mean ± SEM. Statistical analyses were performed using SPSS (version 21; SPSS). Data were analyzed by one-way ANOVA or repeated-measures ANOVA followed by post hoc testing (Bonferroni test) to localize significant effects. Statistical significance was set at P < 0.05.

CRRs to hypoglycemia after LI exercise or no exercise in both control and RH animals did not differ significantly (Supplementary Data) so were combined into the LI exercise grouping. Plasma glucose profiles were matched during hypoglycemia in all groups (Fig. 2A) (F = 0.4, df [2,39], P = NS). In contrast, mean glucose infusion rate (GIR) (F = 9.2, df [2,39], P < 0.05), epinephrine level (F = 34.5, df [2,39], P < 0.05), and glucagon level (F = 10.4, df [2,39], P < 0.05) differed significantly between groups. Post hoc analysis revealed that CRRs were significantly suppressed in RH and LI exercise rats compared with Control and LI exercise rats during hypoglycemia (Fig. 2B–E). In contrast, the mean (SEM) values for GIR (2.9 [0.6] vs. 4.4 [1.0] mg ⋅ kg⁻¹ ⋅ min⁻¹, Control and LI vs. RH and HI exercise rats; P = NS), epinephrine (7.9 [0.4] vs. 6.8 [0.6] ng ⋅ mL⁻¹, respectively; P = NS), and glucagon (245 [18] vs. 219 [31] ng ⋅ L⁻¹, respectively; P = NS) CRRs were not different between Control and LI exercise and RH and HI exercise groups (Fig. 2B–E). Corticosterone responses to the hypoglycemic challenge were not affected by any intervention. These findings indicate that a single episode of HI exercise can restore the CRR in rats exposed to RH over 4 weeks.

HI exercise resulted in significantly higher levels of epinephrine, glucagon, and lactate in both the Control and HI exercise and RH and HI exercise groups compared with their respective LI exercise studies (Figs. 3A–C). Moreover, consistent with response recovery after habituation, the mean (SEM) epinephrine (2.2 [0.3] vs. 1.4 [0.1] ng ⋅ mL⁻¹; RH and HI exercise vs. Control and HI exercise, P < 0.05), glucagon (46 [2] vs. 38 [2] pg ⋅ mL⁻¹; P < 0.05), and lactate (88 [3] vs. 77 [2] ng ⋅ μL⁻¹, P < 0.05] responses to HI exercise were all significantly greater after RH (Fig. 3A–C). Interestingly, the profile of cytokine release differed with the expected increase in cytokine release with HI exercise in Control groups (Fig. 3E–I) but had a more varied response after RH when the release of broadly proinflammatory cytokines (interleukin [IL]-6, IL-1β, and tumor necrosis factor-α) after HI exercise was significantly suppressed compared with Control and HI exercise, whereas the release of anti-inflammatory cytokines (IL-4 and IL-13) was generally increased; the latter increase also was evident after low-stress LI exercise (Figs. 3H and I).

Proteomic analysis of VMH samples from RH and LI exercise versus RH and HI exercise rats (Fig. 4A and B) revealed that DNAJB1 (or Hsp40) and Bassoon (Bsn) were most significantly affected by HI exercise (Fig. 3B), which was confirmed by Western blot analysis (Fig. 4C). Bsn is involved in glutamatergic signaling, as is BDNF (10), and BDNF levels in humans are significantly elevated in response to exercise (11). To determine whether BDNF may be involved in dishabituation, plasma BDNF level was measured and found to be augmented by HI exercise in RH animals (Fig. 3D). In addition, VMH expression of BDNF, TrkB, and pCREB were all significantly increased after RH and HI exercise (Supplementary Data).

This study demonstrates in an animal model of IAH that diminished CRR can be rapidly restored by a single bout of HI exercise. This finding supports the hypothesis that diminished CRR after RH results from habituation. Habituation is a form of adaptive learning where there is a decrease or cessation of responses to a stimulus after repeated presentations. Habituation is usually considered in the context of innate behaviors and a reduced response to repeated external or internal stimuli (5,6). A classic model of habituation is the gill-withdrawal reflex in Aplysia (12). When the gill-withdrawal reflex is repeatedly evoked by a tactile stimulus to the siphon, the amplitude of the response shows a marked decrement (habituation). The habituated response can then restored by presenting a strong tactile stimulus to another part of the animal (dishabituation) (12). Habituation also occurs after repeated physiological stress such as endotoxin exposure or in models of shock (13). In the context of hypoglycemia, low glucose level can be considered the internal stimulus, which leads to a reflex CRR. RH then leads via habituation to a progressive diminution of this response. This concept of hypoglycemia as an internal sensory cue stimulating a counterregulatory reflex (motor) response is consistent with our current understanding of the neural circuitry of brain glucose sensing (14).

In the current study, HI exercise was chosen as the dishabituating stimulus, but other novel stressors may prove equally effective. For instance, CRRs to hypoglycemia and cold exposure have been shown to influence each other (15). The present findings differ from previous studies where antecedent exercise produced further suppression of CRR to hypoglycemia (e.g., [16]). However, these used continuous aerobic exercise providing a moderate stimulus. Stimulus generalization (or cross-tolerance) between different moderate stressors is a feature of habituation (5,6), distinct from dishabituation, which requires a novel strong stimulus. Interestingly, prior RH did not suppress the CRR to exercise, as reported by others (17), but did lead to a change in the inflammatory response to both exercise modalities. The differing intensity of exercises used in the current study likely explains the contrasting findings. This also suggests that the impact of exercise on the CRR to hypoglycemia may vary depending on the type and duration of exercise performed.

To gain some understanding of the mechanisms by which HI exercise acutely restored the CRR to hypoglycemia, an exploratory proteomic analysis of VMH tissue was performed. A number of candidate proteins emerged such as Glycogenin 1 and adenylate cyclase 10 that play roles in glycogen turnover, which may be relevant to glucose sensing (18), whereas dipeptidyl peptidase 10 affects K_v4 channel gating (19). DNAJB1 encodes for Hsp40, which helps to protect proteins during cellular stress (20). Bsn, which was most affected by HI exercise, is a presynaptic cytomatrix protein that contributes to synaptic plasticity and learning (10). Bsn colocalizes with BDNF in glutamatergic presynapses, and increased BDNF release helps to maintain levels of neural excitation (10). Loss of VMH glutamatergic Steroidogenic Factor 1 (SF1) activity suppresses the CRR to hypoglycemia (21), while specific deletion of BDNF in VMH SF1 neurons also results in a reduced CRR to hypoglycemia (22). VMH SF1 neurons also appear to contribute to the metabolic benefits of endurance exercise (23). Taken together with the findings in the current study, it is possible to speculate that dishabituation may act via BDNF-mediated restoration of glutamatergic activity (Fig. 3D). These findings are, however, observational and need to be directly tested in further studies. In addition, we cannot exclude the possibility that dishabituation occurs in other brain regions not studied.

Overall, the current study supports the hypothesis that suppressed CRR after RH in T1D is a result of habituation. Consistent with this hypothesis, we have demonstrated that the introduction of a dishabituating stimulus, HI exercise, acutely restores the CRR to subsequent hypoglycemia. If confirmed in humans, these findings may lead to an improved understanding of IAH in T1D and the development of novel treatment strategies designed to restore hypoglycemia awareness.

Acknowledgments. The authors thank the staff of the Biomarker and Drug Analysis Core Facility, University of Dundee, who analyzed all of the blood and hypothalamic samples, for their support.

Funding. This work was supported by an award to R.J.M. from JDRF (1-2011-574).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. A.D.M., J.R.G., and J.T.-J.H. performed the experiments and wrote the manuscript. M.L.J.A. wrote the manuscript. R.J.M. designed the studies and wrote the manuscript. R.J.M. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the 75th Scientific Sessions of the American Diabetes Association, Boston, MA, 5–9 June 2015.