Neural Regulation of Blood Glucose in Acute Stress: A Report on Research Supported by Pathway to Stop Diabetes

Abstract

There is significant evidence that acute stress, a challenge to an organism’s homeostasis, has dramatic effects on metabolic control. Acute stress impairs blood glucose control in people with both type 1 and type 2 diabetes. In addition, growing evidence suggests that metabolic responses to stress in people without diabetes may be a crucial determinant of health. Acute dysregulation of blood glucose in the hospital setting, including both hyper- and hypoglycemia, predicts short- and long-term morbidity and mortality in patients with critical illnesses. Animal studies indicate that exposure to physiological and psychological stressors activates a highly conserved network of neural circuits that ultimately coordinate the functions of multiple organs to increase blood glucose. In this article, we provide an overview of the neural populations and circuits that increase blood glucose in response to acute stress, including our research funded by the American Diabetes Association Pathway to Stop Diabetes program, highlighting the impacts on clinical outcomes and opportunities for the development of therapies for diabetes. This article is part of a series of perspectives that report on research funded by the American Diabetes Association Pathway to Stop Diabetes program.

Article Highlights

• Internal and external stressors rapidly increase blood glucose, a highly conserved metabolic response.

• Multiple stress-modulated neural populations in the brain stem, hypothalamus, and forebrain contribute to regulation of the hypothalamo-pituitary-adrenal axis and sympathetic nervous system to elicit hyperglycemia.

• Exaggerated or diminished glucose responses to acute stress are associated with increased risk of type 2 diabetes and poor health outcomes.

• A greater understanding of the neural circuitry contributing to stress hyperglycemia and how these circuits are disrupted has the potential to provide new approaches to improve glycemic control.

Introduction

Stress—a disruption or anticipated disruption in homeostasis—leads to highly conserved behavioral and metabolic responses (1) including a rapid increase in blood glucose. The effects of acute stress on metabolic regulation are commonly seen in clinical practice. Acute stress disrupts glucose control in individuals with type 1 and type 2 diabetes (2,3). In addition, stress hyperglycemia is frequent in hospitalized patients, including in individuals without preceding diabetes. While this adaptive metabolic response is thought to provide short-term benefits, exaggerated or diminished glucose responses to stress are associated with increased morbidity and mortality as well as predictive of increased risk of later type 2 diabetes. Despite the clinical importance of stress hyperglycemia, the neural mechanisms linking stress and hyperglycemia are still being elucidated. However, a significant body of work shows that brain-body cross talk is crucial for stress hyperglycemia. Stress-activated brain circuits control hormonal and neural signals that in turn regulate peripheral organs including the liver and pancreas to modulate blood glucose in response to stress (4). Here, we provide an overview of the central circuits contributing to stress hyperglycemia in the brain stem, hypothalamus, and forebrain and highlight the gaps in our understanding of these circuits and how we may begin to address them to eventually improve clinical outcomes.

Many forms of stress disrupt glucose homeostasis, from psychological stressors such as aggression or predator exposure to physiological stressors including marked hypoglycemia and infection. External and internal sensory information is transmitted to the central nervous system (CNS). These sensory signals, in turn, initiate multiple behavioral and metabolic adaptations (5). Stress triggers beneficial behaviors (e.g., escape or aggression) and inhibits conflicting actions (e.g., feeding) along with increasing metabolic demand. In parallel, acute stressor exposure results in rapid metabolic adjustments to meet these increased metabolic demands. As the physiological or psychological stress subsides, the behavioral and metabolic responses to acute stress decline, ultimately restoring homeostatic balance.

Stress-activated circuits are highly conserved across species (5), and published evidence suggests that a network of neural circuits with overlapping functions contributes to dynamic adaptation with an acute stressor (also termed allostasis). Stressors activate CNS circuits that drive rapid changes in the autonomic nervous system (ANS), comprised of sympathetic and parasympathetic pathways, and initiate slower, more sustained endocrine responses, such as modulation of the hypothalamo-pituitary-adrenal (HPA) axis to increase corticosteroids (6). Stressor exposure increases sympathetic nervous system (SNS) activity within seconds resulting in adrenal activation and increased circulating epinephrine as well as organ-specific alterations in norepinephrine. However, SNS activity is usually not sustained as parasympathetic activation increases to restore balance between the sympathetic and parasympathetic arms of the ANS. Stressors also result in activation of the HPA axis and a slower but more sustained increase in plasma glucocorticoids. Concomitantly, stress induces changes in circulating cytokines (7). Together, autonomic, hormonal, and cytokine adaptations blunt insulin release, increase hepatic glucose production, and reduce insulin-dependent glucose uptake, leading to a rapid increase in blood glucose and impaired glucose tolerance (8). The nature, severity, and duration (9) of the stressor as well as previous experiences (10) modulate ANS and HPA activation that contributes to behavioral, physiological, and metabolic responses to stress.

Unfortunately, the adaptations to a stressor can be disrupted in many ways (11): repeated stimulation from multiple stressors, a failure to adapt to repeated exposure to the same stressor, an exaggerated or prolonged response to stress exposure, or inadequate activation. These disruptions can have far-reaching effects on health. The associations between recurrent and chronic stress exposure and metabolic diseases such as diabetes are well recognized (12). However, exaggerated or blunted metabolic responses to acute stress are also associated with significant detrimental effects on health (13). Stress-associated hyperglycemia is linked with increased morbidity and mortality in hospitalized patients. In intensive care patients without diabetes, increased hyperglycemia was significantly associated with increasing length of hospitalization and mortality, particularly in those with neurological disease and head trauma (14). More recently, the stress hyperglycemia ratio (SHR) has been used to describe the effects of stress on glycemic control in hospitalized individuals (15). SHR, the ratio of the admission fasting blood glucose to HbA_1c, provides an index of stress hyperglycemia in relation to preadmission glucose control. Both stress-induced hyperglycemia and SHR have been associated with increased complications, short-term and longer-term mortality, and increased risk of later diabetes in multiple conditions: acute myocardial infarction (16), cerebrovascular accident (17), sepsis (18), and coronavirus disease 2019 infection (19).

The substantial clinical evidence outlined above highlights the urgent need for a better understanding of the neural circuits underlying metabolic adaptations to acute stress and the physiological roles of hyperglycemia with stress. While there are extensive preclinical studies examining the contributions of stress-activated neural circuits to stress-induced behavioral adaptations, fewer studies have been conducted to examine the roles of neural circuits and populations in metabolic responses to acute stress. The goals of this article are threefold: to briefly summarize our understanding of the major neural circuits in the brain stem, hypothalamus, and limbic system that may regulate metabolic responses to acute stress; to discuss the possible physiological roles of stress hyperglycemia; and, finally, to highlight the gaps in our knowledge that offer opportunities for future studies to understand whether and how abnormal stress hyperglycemia contributes to metabolic disease so that novel therapies targeted stress-induced metabolic abnormalities can be developed.

Neural Circuits Regulating Stress Hyperglycemia

Brain Stem Circuits

Brain stem circuits play significant roles in metabolic responses to acute stress. The brain stem is the site of multiple sensory inputs that can convey indications of stress. For example, brain stem regions receive interoceptive signals indicating changes in blood volume that may indicate hypovolemic stress, signals about nutrient status that may communicate metabolic stress such as hypoglycemia, and vagal and spinal sensory signals conveying visceral inflammation and pain. Brain stem regions also receive spinal sensory signals relaying somatic pain in response to external stressors. In parallel, descending circuits from extrahypothalamic and hypothalamic regions convey stress-related sensory information, such as visual cues and odors. These ascending and descending sensory inputs interact and modulate brain stem regions that, in turn, regulate the activity of descending sympathetic and parasympathetic circuits, particularly in the rostral ventrolateral medulla (RVLM) and dorsal motor nucleus of the vagus (DMV), respectively. In addition to descending control of peripheral organs via autonomic innervation, stress-responsive brain stem regions, such as the nucleus of the solitary tract (NTS) and locus coeruleus (LC), have ascending projections to multiple hypothalamic and limbic centers to modulate their activity (20,21). Although multiple stressors increase Fos in brain stem regions indicating neural activation, in existing studies investigators have been unable to identify whether different modalities of stressors activate overlapping or distinct neural populations within these areas and other stress-responsive brain regions. As a result, distinct stressors may regulate specific downstream circuits and organs to regulate metabolic control.

There is considerable evidence to support a role for RVLM neurons in metabolic regulation with stress. RVLM neurons are activated by hypotension (22), marked glucopenia (23), and restraint (24) (to a lesser extent), as well as by conditioned fear (25), a psychological stress where an animal learns to associate a particular environment with a stressful experience. Activating RVLM neurons using bicuculline (23) in the absence of hypoglycemia increased blood glucose. These findings suggest that stress-activated RVLM neurons likely contribute to stress-induced increases in blood glucose. Modulation of RVLM neurons in cats (26) using CO₂-bubbled saline to stimulate baroreceptors also significantly increased blood glucose. More recently, in male and female mice (27), chemogenetic activation of RVLM catecholaminergic neurons increased blood glucose, while ablation of ventrolateral medulla (VLM) catecholaminergic neurons blunted the glucose response to both psychological and physical stressors (28). While the majority of catecholaminergic RVLM neurons project to the spinal cord, there are also ascending circuits from the RVLM to the periaqueductal gray and hypothalamic regions, particularly the lateral hypothalamus (LH) (29). These findings suggest that RVLM catecholaminergic signaling may modulate the activity of other neural populations that contribute to autonomic regulation.

Brain stem regulation of parasympathetic activity is primarily via the DMV. Increased parasympathetic activity stimulates pancreatic insulin release to lower blood glucose. Therefore, stressors may inhibit DMV neurons to reduce insulin release and raise blood glucose or increased DMV activity with stress may act to restore blood glucose. Several studies have highlighted the roles of stress-activated circuits to regulate DMV neuron activity. Acute stress activates serotonergic neurons in the dorsal raphe (DR) that, in turn, increase activity in cholinergic DMV neurons. In these studies, investigators demonstrated effects of cholinergic DMV neurons on gastrointestinal function but did not examine blood glucose (30). Catecholaminergic neurons also influence parasympathetic regulation of metabolism. In rats (31), ablating catecholaminergic neural inputs into the DMV (which are primarily from the A5 region in the rubrospinal tract) using 6-hydroxydopamine increased plasma insulin, an effect that was lost with vagotomy. These findings suggest that stress-activated catecholamine neurons normally act to inhibit DMV neurons, reducing insulin to increase blood glucose. In addition to descending vagal circuits, neurons in the DMV also project to adjacent regions, the NTS and area postrema, which may allow interaction between afferent and efferent vagal pathways.

Parasympathetic regulation of metabolism in the DMV is also modulated by inputs from the NTS. NTS neurons receive and integrate inputs from many visceral organs via vagal sensory circuits. Published studies demonstrate that NTS GABAergic neurons are activated by acute stressors in both males and females (32) including visceral malaise after lithium chloride injection and psychological stress such as restraint or conditioned fear. Recent studies showed that chemogenetic activation of GABAergic (33) NTS neurons increased blood glucose in unstressed mice. Chemogenetic inhibition of GABAergic NTS neurons reduced blood glucose in a streptozotocin(STZ)-administered hyperglycemia mouse model, effects that were mediated in part by blunted inhibition of vagal efferent circuits in the DMV. These studies support a role for NTS GABAergic neurons in physiological glucose regulation and possibly in stress-induced increases in blood glucose, but further studies are needed to determine whether NTS GABAergic neurons are required for the full hyperglycemic response to stress. NTS neurons also project directly to the RVLM, suggesting that NTS neural populations may regulate sympathetic activity. These neural populations are likely to be activated by internal stressors; e.g., hypoxic stress activates RVLM-projecting (34) NTS neurons. The neurochemical identity and roles of RVLM-projecting neurons in the NTS are not completely understood. However, Preproglucagon (Ppg)-expressing NTS neurons, producing glucagon-like peptide 1 (GLP-1), have been reported to project to the RVLM (35). Central administration of the GLP-1 agonist, exendin-4, increased Fos expression in the RVLM (36) and altered heart rate. In keeping with this, GLP-1 receptor agonists (GLP-1 RA) increased daily heart rates in clinical studies (37) and have been reported to increase skeletal muscle sympathetic nerve activity (38). These findings suggest that NTS GLP-1 neurons may modulate RVLM activity to regulate SNS tone, but whether NTS-RVLM GLP-1 circuits contribute to stress hyperglycemia and the effects of GLP-1 RA on stress hyperglycemia in clinical settings have not fully been explored. However, the overall beneficial effects of GLP-1 RA on glycemic control may mask NTS-RVLM circuit-specific effects on blood glucose in response to stress.

In parallel to regulating downstream sympathetic and parasympathetic efferent projections, brain stem regions also provide ascending circuits to regulate metabolism via actions on the HPA axis. Both noradrenergic and GLP-1–expressing NTS neurons modulate stress-induced activation of corticotrophin-releasing hormone (CRH) neurons (39). Noradrenergic NTS neurons increase CRH neural activity via α1 receptors, but with greater noradrenergic NTS activity, CRH neurons are inhibited. Ablation of noradrenergic NTS neurons blunted the HPA response to interoceptive signals, suggesting that the major role of this circuit is to enhance CRH release. Similarly, loss of GLP-1 inputs from the NTS blunted the HPA responses to both psychological and physiological stressors, suggesting that this circuit also increases HPA activation. In addition, there is evidence that catecholaminergic neurons in the VLM are needed for the full HPA response to hypoglycemic stress and to IL-1 administration (40).

Hypothalamic Circuits

Several hypothalamic nuclei are activated by psychological and physiological stressors including the LH and paraventricular (PVH), dorsomedial (DMH), ventromedial (VMH), and posterior (PH) hypothalamus. PVH neurons are robustly activated by stress in male and female rats (41) and mice (42). The PVH plays a key role in regulating both SNS and HPA responses to stress. The PVH is highly heterogeneous with multiple neural populations defined by the expression of distinct subsets of neuropeptides and neurochemicals. Among these populations, expression of corticotrophin-releasing hormone (CRH) and neuropeptide Y receptor 1 (NPYR1) best predicted neural activation in stressful fear retrieval and CRH neuron activity increases with numerous stressors (43). Activation of PVH CRH neurons releases CRH from the median eminence into the pituitary portal circulation, stimulating the secretion of adrenocorticotrophin hormone (ACTH) to drive corticosterone release from the adrenal cortex. As expected, chemogenetic activation of CRH neurons in the absence of stress increased blood glucose. However, stress-activated PVH CRH neurons play roles beyond HPA axis regulation. Recent studies described a role for PVH CRH neurons in rapid activation of sympathetic innervation to the liver to increase blood glucose, independent of the HPA axis or plasma epinephrine (44), via projections to the VMH. CRH neurons also contribute to effects beyond metabolism, including modulating anxiety-related behavior. PVH neurons expressing arginine vasopressin (AVP) are also strongly implicated in regulation of the HPA axis. AVP neurons are activated in response to stress and AVP potentiates the effects of CRH on ACTH release. In addition, chemogenetic activation of PVH^AVP neurons increases plasma glucose and glucagon (45). PVH neurons expressing oxytocin (PVH^OXT) have also been identified as contributing to glucose regulation. Chemogenetic stimulation of PVH^OXT neurons suppressed insulin release to increase blood glucose (46). However, the effects of silencing OXT and AVP neural populations in the PVH on stress hyperglycemia, to determine whether they are required for the full glycemic response to stress, are not known.

In addition to regulating HPA responses, PVH neurons project to brain stem and spinal cord regions (47) that regulate both sympathetic and parasympathetic activity. These including significant projections from the PVH to the brain stem NTS, DMV and RVLM regions as well as projections to the intermediolateral nucleus of the spinal cord. These preautonomic PVH populations are primarily in the posterior PVH. Studies in the cardiovascular system suggest that PVH projections to the RVLM increase sympathetic activity, while projections to the NTS inhibit sympathetic and parasympathetic activity (48). It is not clear in these studies if the effects of sympathetic activation are directly at the level of the heart, via increased adrenal epinephrine release, or both. Interestingly, PVH neurons receive significant inhibitory GABAergic inputs (49), particularly from regions immediately adjacent to the PVH. This peri-PVH region receives inputs from cortical and limbic brain regions, potentially allowing top-down regulation of HPA and autonomic activity.

Additional stress-activated hypothalamic regions have also been implicated in metabolic regulation in stress. Neural populations, particularly orexin-expressing neurons, in the LH show increased expression of the neural activity marker cFos in response to stress (50) in male and female (51) rodents. LH neurons project to numerous downstream sites, including sites regulating sympathetic and parasympathetic activity in the brain stem such as the NTS and DR, and spinal cord (52). In addition, there are projections from the LH to the PVH. Pharmacological studies demonstrated that upregulated cholinergic signaling in the LH increased plasma epinephrine and blood glucose, suggesting sympathetic activation (53). More recent neuromodulation studies suggest that LH orexin circuits contribute to glucose regulation. For example, optogenetic stimulation of LH orexin neurons impaired glucose tolerance and insulin sensitivity (54). However, the contributions of these neurons to stress hyperglycemia remain unclear.

The DMH is also activated in response to stress, particularly emotional (55) stressors, though whether there are sex differences in these responses is not known. DMH neural populations likely regulate both HPA and autonomic stress responses via efferent projections (56). DMH neurons projecting to the PVH are both activating (glutamatergic) and inhibitory (GABAergic). In keeping with this, modulation of DMH populations has been reported to both activate and inhibit the HPA axis. Chemical stimulation of DMH neurons in rats increased ACTH (57). However, bilateral DMH lesioning also increased ACTH release in response to emotional, but not immune, stressors (58). DMH circuits may also be involved in mediating the negative-feedback effects of corticosterone on the HPA axis, as microinjections of corticosterone into the DMH blunted the stress-induced corticosterone response. In addition, recent studies suggest that DMH neural populations expressing GLP-1 receptors may act to alter autonomic activity and lower blood glucose. Targeted activation of both GLP-1R-expressing neurons in the DMH and of GLP-1R projections from the DMH to the parasympathetic DMV lowered blood glucose via increased insulin release in mice (59). Whether DMH GLP-1R-expressing neurons are activated by stress and the time course of any changes remain to be investigated.

Neurons in the VMH are also responsive to some, but not all, stressors. In male rats and mice (60), predator stress and restraint combined with water immersion (61) increased Fos expression (62) in the dorsomedial VMH but restraint (63) or hypoxia alone did not. VMH neurons are important regulators of autonomic function. Anterograde tracing (64) from dorsomedial VMH neurons identified projections to CNS regions crucial for HPA and autonomic regulation including the PVH, BNST, PH, RVLM, and NTS. Early VMH lesioning studies demonstrated a transient enhancement of the baseline HPA axis but normal corticosterone responses to stress, suggesting that VMH neurons are not required for the full stress-induced HPA response (65). Neuromodulatory studies have begun to dissect the roles of distinct VMH neurons (66) subpopulations. For example, activation of SF1-, glucokinase- and CCK receptor B (CCKRB)-expressing VMH neurons is sufficient to increase blood glucose, while inhibition of VMH SF1 but not CCKRB neurons blunted the corticosterone response to insulin-induced hypoglycemia (66), suggesting that SF1 neurons modulate HPA axis activity.

Published studies also suggest a role for the PH in stress-related metabolic adaptations. In male rats, PH neurons are activated by stress and project to both the PVH and the raphe pallidus, a brain stem region implicated in regulation of autonomic outflow (67). Moreover, silencing PH neurons with muscimol blunted the HPA response to acute stress, while activation of PH neurons by GABA antagonists enhances ACTH and corticosterone release with restraint (67). These findings suggest that the PH contributes to regulation of the HPA axis and, possibly, ANS in response to acute stress, but again the effects on stress hyperglycemia were not examined.

Forebrain Limbic Circuits

Stressors result in robust activation of limbic structures including amygdala nuclei, bed nucleus of the stria terminalis (BNST), hippocampus, habenula complex, and prefrontal cortex (PFC). These regions, in turn, modulate the activation of the autonomic and HPA responses to stress. Interestingly, while there were no differences in the restraint stress responses of amygdala, BNST, and PFC between male and female mice, stress-induced Fos responses in the hippocampus and medial habenula were greater in female than in male mice (42).

The amygdala, a cluster of multiple nuclei in the temporal lobe that are central to emotional processing, memory, and autonomic function, is modulated by stress. Amygdala nuclei integrate behavioral, endocrine, and autonomic responses. The central nucleus of the amygdala (CeA) is activated by stressors. CeA neurons project to multiple brain stem areas (LC, periaqueductal gray), limbic regions including the BNST, as well as sparse projections to the periventricular hypothalamus (68). In rats, electrolytic lesions of CeA neurons blunted corticosterone and catecholaminergic responses to foot shock (69). Interestingly, CeA lesions did not blunt the tachycardic response to stress but delayed the recovery of heart rate to baseline after stress. These results suggest that circuits involving the CeA may modulate parasympathetic activity. In keeping with these findings are those of recent studies that GABAergic CeA neurons project to the cholinergic neurons in the DMV (70) providing an anatomic circuit for possible parasympathetic inhibition. CRH receptors are expressed in CeA neurons, and microinjection of CRH into the CeA significantly blunts the glucose response to both social stress and isolation stress in rats (71). In keeping with the relatively sparse connections between the CeA and hypothalamus, studies suggest that CeA activity does not alter the HPA axis. Microinjection of muscimol, a GABA agonist, into the CeA of rats with and without chronic pain did not alter either ACTH or corticosterone (72). In contrast, optogenetic activation of CRH-expressing CeA neurons projecting to the LC increases LC activity and anxiety-like behavior seen in acute stress (73). Together, these findings suggest that the main effects of CeA neurons are via modulation of the ANS rather than the HPA axis, but further work is needed to link their role in ANS modulation to stress hyperglycemia.

Restraint stress and other stressors also robustly increase cFos expression in the basolateral amygdala (BLA) and basomedial amygdala (BMA) (74). BLA neurons project to the LH (75), while BMA neurons project to several hypothalamic nuclei including DMH and LH (76). However, the functional effects of BLA and BMA circuits on metabolic regulation in stress are conflicting. Injection of the GABA agonist, muscimol, into the BLA has been reported to increase (77) the HPA response to acute stress or to have no effect (72) on the HPA axis, while CRF injection into the BLA enhanced corticosterone responses to restraint. Similarly, the effects of BLA and BMA manipulation on autonomic function are also inconsistent. BMA inhibition has been reported to increase heart rate (78), while findings of recent studies suggest that both BLA inhibition with muscimol and activation with bicuculline reduced heart rate. Therefore, although BMA and BLA neurons are stress responsive, it remains unclear whether these regions contribute to stress-mediated changes in ANS and HPA function and the downstream effects on blood glucose.

Our own research, funded by the American Diabetes Association Pathway to Stop Diabetes program, highlights the contribution of the medial amygdala (MeA) to metabolic regulation with acute stress. Activity in MeA neurons, particularly those projecting to the VMH, increased with a broad range of physical and psychological stressors in both male and female mice (79). Using chemogenetic and optogenetic tools, we demonstrated that increasing activity in MeA neurons and in MeA-VMH circuits in unstressed mice significantly increased blood glucose and impaired glucose tolerance, while inhibiting MeA-VMH circuit activity in stressed mice blunted their stress hyperglycemia. The effects of MeA-VMH circuit activity were mediated by enhanced sympathetic drive to the liver that increased hepatic glucose production, independent of the HPA axis and plasma epinephrine. Repeated stress blunted activity in this circuit, and the combination of reduced circuit activity with high-fat diet resulted in hyperglycemia and glucose intolerance. While these studies were performed in mice, our spatial transcriptomic findings identified several genes enriched in the MeA-VMH circuit that are linked to increased risk of type 2 diabetes in humans. These insights suggest that disrupted activity in this circuit may contribute to the increased risk of metabolic disease with recurrent or chronic stress.

The hippocampus, a brain region vital for memory, learning, and mood, also contributes to stress responses. A number of studies suggest that hippocampal inputs modulate both HPA and autonomic function in stress. Acute stress increases activity in the ventral hippocampus (5) and is exaggerated in mice previously exposed to chronic stress (80). The ventral hippocampus and subiculum express glucocorticoid receptors, and lesions in these regions or their output circuits modulate the HPA (81) axis and blunt the ability of exogenous corticosteroids to suppress HPA activity. Recent studies show that optogenetic activation of hippocampal circuits to the BNST blunted the corticosterone response to restraint stress (82).

Hippocampal circuits also regulate autonomic outflow. Retroviral tracing studies demonstrate connections from the hippocampus via the medial PFC (mPFC) to brain stem regions controlling autonomic activity (83). Electrical stimulation of the hippocampus is sufficient to lower heart rate and respiratory rate (84) suggesting a role for these circuits in the regulation of autonomic tone. Interestingly, hippocampal output is greatest during sharp wave ripples, and these are strongly correlated with decreases in interstitial glucose (85). Further, optogenetic stimulation of dorsal hippocampi in rats to mimic sharp wave ripples was also associated with a decrease in interstitial glucose.

The HPA and autonomic responses to acute stress are also modulated by cortical inputs, particularly the mPFC. The PFC is crucial for information processing allowing adaptation to one’s environment and emotional control. Both infralimbic (IL) and prelimbic (PL) PFC connect to multiple brain regions regulating HPA axis and autonomic function. These include limbic areas such as the BLA and BNST (86), hypothalamic regions such as the LH and DMH, and brain stem regions including the NTS (87) and DR (88). Recent studies identified a PACAPergic circuit from the mPFC to PVH that enhances stress activation of CRH neurons (89). However, published evidence suggests that the IL and PL PFC may play opposite roles in regulating stress responses. Injection of the GABA antagonist, bicuculline, to disinhibit the IL PFC increased sympathetic nerve activity, heart rate, and core body temperature (90). Further, local ablation or synaptic blockade of the IL PFC modulates parasympathetic control of cardiovascular function (91). However, other studies suggest that inhibition of IL mPFC has no effect on baseline or stress-induced cardiovascular responses (92). Published data also suggest that IL PFC enhances the HPA axis responses to stress, as IL PFC lesioning attenuated PVH Fos responses to stress and blunted the ACTH response to restraint (93). In contrast, the PL PFC appears to blunt stress-induced HPA axis activation. Lesions of the PL PFC enhance stress-induced Fos expression in the PVH and ACTH and corticosterone secretion (93). In addition, PL PFC circuits have been implicated in limiting the cardiovascular responses to psychological stressors. Further, PL PFC circuits are reported to suppress activity in DR neurons (88), which are known to be targeted by VMH circuits to regulate blood glucose (94). Together these data suggest that the PFC circuits contribute to regulation of both HPA and autonomic responses to stress.

Information from hippocampal, PFC, and amygdalar brain regions is relayed by other structures to modulate autonomic and hormonal stress responses, particularly the BNST. The BNST is comprised of multiple subnuclei and neural populations that have been implicated in regulating metabolism, reproductive function, social behavior, stress, and fear (95). BNST neurons are primarily GABAergic and provide inhibitory inputs to downstream regions such as the PVH, as well as amygdala and NTS (96). BNST subnuclei have different effects on autonomic and HPA responses to stress. Anterior BNST lesions reduced PVH expression of CRH, suggesting that this area enhances stress responses (97). In contrast, posterior BNST lesions increased CRH expression, in keeping with inhibition of the HPA axis (97). BNST regions also modulate autonomic function. BNST neurons project to brain stem regions modulating autonomic function as well as to the PVH (98). Activation of CRH receptor–expressing neurons in the medial BNST was reported to activate both sympathetic and parasympathetic cardiovascular responses (99). In addition, pharmacological blockade of BNST neural activity increased tachycardic responses to stress, suggesting that BNST circuits act to inhibit stress-induced autonomic cardiovascular responses. Recent studies demonstrate that optogenetic activation of BNST neurons rapidly increased blood glucose, partly via a GABAergic circuit to the arcuate nucleus and sympathetic activation (100). The habenula complex, a node involved in regulating emotional behavior that is divided into medial and lateral nuclei, is also activated by acute stress (101). Recent studies demonstrated that chemogenetic activation of a medial habenula–to–interpeduncular nucleus circuit increased blood glucose via a polysynaptic sympathetic circuit to the pancreas (102). These findings suggest that stress-activated habenula circuits may also contribute to metabolic responses to acute stress. Regarding neural circuits regulating stress hyperglycemia see Fig. 1.

Figure 1.

A: Schema of stress-responsive brain regions in the hypothalamus, forebrain, and brain stem. B: Circuits regulating stress activation of the HPA axis. The hypothalamic paraventricular nucleus (PVH) is activated via excitatory inputs (black lines) from multiple hypothalamic, forebrain, and brain stem nuclei to increase HPA activation with acute stress. Inhibitory inputs (red lines) from forebrain and hypothalamic nuclei restrain HPA axis activity. C: Circuits regulating stress modulation of the ANS. Stress-induced changes in autonomic activity originate in brain stem regions, including the VLM and LC for sympathetic activity and DMV for parasympathetic activity. These regions receive both excitatory and inhibitory inputs from nearby brain stem areas as well as descending inputs from forebrain and hypothalamic nuclei to regulate autonomic activity in response to acute stress. Figure was created with BioRender (biorender.com). vHIP, ventral hippocampus.

Conclusions and Perspectives

The adaptations to stressors rely on a complex network of circuits that integrate external and internal sensory signals to orchestrate a multiorgan response regulating behavior and metabolism. As outlined above, we are beginning to piece together the circuits that contribute to stress-induced autonomic responses and HPA activation. However, many questions remain.

Most of the studies summarized above link specific brain regions to stress-induced circuit activity. However, detailed monitoring of neural or circuit activity with concurrent measurement of blood glucose is rare in published studies but would allow a more detailed understanding of the precise temporal relationship among stress presentation, neural activity, and changes in circulating glucose. As several stress-activated neural populations also respond directly and indirectly to physiological changes in blood glucose, it is important to identify the time course of stress-induced alterations in neural activity in relation to changes in blood glucose for determination of whether neural activity is driving glucose alterations or is a response to changes in blood glucose. For example, neural activity secondary to inflammation and should precede the hyperglycemic response. Approaches such as in vivo calcium imaging with fiber photometry or miniscopes, together with continuous glucose monitoring, begin to provide the detailed temporal information needed. However, one difficulty is that extreme blood glucose levels, particularly hypoglycemia, can be viewed as a stressor. Therefore, comprehensive studies with examination of neural activity in candidate populations not only in response to extreme hypo- or hyperglycemia but also with a range of stress modalities and over physiological glucose concentrations would be ideal.

Most of the stress-responsive brain regions described above are linked to modulation of the HPA axis or ANS. However, further work is needed to strengthen the case that these circuits contribute to stress hyperglycemia. For most of the stress-activated circuits outlined, studies to demonstrate the necessity of these circuits for stress hyperglycemia are required. For example, work to examine the effects of activating or silencing circuit activity on blood glucose responses, with and without stress, may indicate the relative contribution of each circuit to stress hyperglycemia. Such studies may reveal whether therapeutic approaches to blunt circuit activity could reduce blood glucose with stress.

As detailed above, a number of stress-responsive circuits regulate autonomic activity but a detailed understanding of the peripheral organs and signaling pathways mediating these effects on blood glucose in stress is largely missing. In particular, SNS activity is regulated in an organ-specific manner (103). Therefore, specific circuits may modulate SNS activity to one or more distinct organs, such as adrenals (104) to increase plasma epinephrine secretion, or via local norepinephrine release in other organs such as the pancreas (59), adipose tissue (105), or liver (79). Recent studies have begun to dissociate local SNS activation in organs from increased plasma epinephrine (44,79). However, for most stress-responsive circuits these detailed analyses are not yet available. This information is clinically important, as peripheral neural circuits and pathways may be more accessible for pharmacological therapies than are CNS populations.

Although current studies have identified key brain regions contributing to stress hyperglycemia, most are lacking in detailed information about the specific neural subpopulations recruited by stress, their downstream circuits, and the precise mechanisms by which blood glucose is regulated. This is particularly the case for extrahypothalamic circuits. Detailed circuit analyses and identification of genetic markers for stress-activated neural populations will allow us to determine the contributions of defined populations and their circuits to stress-induced glucose responses. The application of single-cell/nucleus RNA sequencing and tools that combine circuit mapping with spatial gene expression (106) offers an opportunity to identify the cell populations contributing to CNS regulation of both the HPA axis and autonomic function. This information, in turn, will allow precise functional circuit mapping allowing investigators to dissect the contributions of limbic, hypothalamic, and brain stem circuits to metabolic responses to stress. A crucial aspect of these studies will be examination of both behavior and metabolism to identify circuits and molecular targets that regulate metabolic function without adverse effects on affective or physical behavior. This is critical for translational studies to repurpose or develop therapies that can target the glucose response to stress without adverse effects on feeding and behavior.

What is the physiological role of stress hyperglycemia? There is evidence that hyperglycemia with critical illness is accompanied by alterations in glucose transporters, particularly upregulation of GLUT1 mRNA and protein in the CNS (107) and macrophages (108) and reduced insulin-dependent translocation of GLUT4 in muscle (109). More recent work suggests that acute hyperglycemia without increased insulin enhances skeletal and cardiac muscle blood volume (110). Together these findings indicate that stress hyperglycemia would likely preserve glucose supply to critical neural and immune tissues and may enhance cardiac blood flow. Clinical and basic studies suggest that the glucose response to stress is proportional to the intensity of the stressor. If so, then stress hyperglycemia may be a proxy for severity of illness, which is, in turn, associated with increased morbidity and mortality. However, there is also evidence that an insufficient glucose response to stress in hospitalized patients is associated with poor outcomes, suggesting that stress hyperglycemia plays a physiological role (111). In keeping with these findings, the Normoglycemia in Intensive Care Evaluation–Survival Using Glucose Algorithm Regulation (NICE-SUGAR) trial demonstrated increased mortality among surgical intensive care unit patients treated with intense glucose control (81–109 mg/dL) in comparison with those treated with conventional glucose control (<180 mg/dL) (112). These findings suggest that both an impaired and an exaggerated hyperglycemic response to stress have detrimental effects on morbidity and survival. In addition to its effects on short-term prognosis, stress hyperglycemia in hospitalized patients is associated with significant risk of later diabetes (113). These findings imply that defects in stress-activated circuits that result in impaired or exaggerated stress hyperglycemia may also contribute to homeostatic glucose regulation and the later risk of type 2 diabetes.

A better understanding of the underlying neural circuits, peripheral mechanisms, and physiological roles of stress hyperglycemia may have important clinical consequences. Understanding of the precise roles of stress-induced increases in blood glucose in preserving neural, immune, and other critical functions may guide appropriate glucose management in acute stress. Given the increased risk of later diabetes for people with stress hyperglycemia, further studies to dissect the components and functions of stress-activated limbic, hypothalamic, and brain stem circuits that contribute to metabolic regulation could identify new opportunities for therapies to prevent diabetes and improve glucose control. These advances will rely on continued advances in tools to dissect neural circuits regulating metabolism as well as multidisciplinary studies with integration of neuroscience, metabolism, and clinical expertise.

Funding Statement

This work was also supported by a previous American Diabetes Association Pathway to Stop Diabetes grant (1-17-ACE-31) and, in part, by grants from the National Institutes of Health (R01DK124461) and U.S. Department of Defense (HT9425-23-1-0244).