Subfornical organ insulin receptors tonically modulate cardiovascular and metabolic function

Abstract

Insulin acts within the central nervous system through the insulin receptor to influence both metabolic and cardiovascular physiology. While a major focus has been placed on hypothalamic regions, participation of extrahypothalamic insulin receptors in cardiometabolic regulation remains largely unknown. We hypothesized that insulin receptors in the subfornical organ (SFO), a forebrain circumventricular region devoid of a blood-brain barrier, are involved in metabolic and cardiovascular regulation. Immunohistochemistry in mice revealed widespread insulin receptor-positive cells throughout the rostral to caudal extent of the SFO. SFO-targeted adenoviral delivery of Cre-recombinase in insulin receptor^lox/lox mice resulted in sufficient ablation of insulin receptors in the SFO. Interestingly, when mice were maintained on a normal chow diet, deletion of SFO insulin receptors resulted in greater weight gain and adiposity, relative to controls, independently of changes in food intake. In line with this, ablation of insulin receptors in the SFO was associated with marked hepatic steatosis and hypertriglyceridemia. Selective removal of SFO insulin receptors also resulted in a lower mean arterial blood pressure, which was primarily due to a reduction in diastolic blood pressure, whereas systolic blood pressure remained unchanged. Cre-mediated targeting of SFO insulin receptors did not influence heart rate. These data demonstrate multidirectional roles for insulin receptor signaling in the SFO, with ablation of SFO insulin receptors resulting in an overall deleterious metabolic state while at the same time maintaining blood pressure at low levels. These novel findings further suggest that alterations in insulin receptor signaling in the SFO could contribute to metabolic syndrome phenotypes.

INTRODUCTION

Impairments in insulin receptor-mediated signaling, characterized as insulin resistance, are directly associated with metabolic syndrome phenotypes including obesity, diabetes, hypertriglyceridemia, and hypertension (6, 7, 18, 25, 26, 31). In addition to peripheral insulin action, central nervous system (CNS) insulin-mediated signaling has emerged as critical in the control of cardiometabolic function. An accumulating body of literature indicates that insulin signaling within the CNS influences satiety, glucose homeostasis, and energy balance (4, 6, 14, 15, 26, 28). Findings from animal models and humans provide additional evidence that brain insulin modulates the cardiovascular system, in part, by influencing sympathetic neural outflow to cardiovascular organs (1, 2, 9, 10, 40, 43, 46, 59, 64). However, the understanding of the integrated neural networks involved in the cardiovascular and metabolic effects of insulin remain incomplete.

Mice with neuron-specific deletion of insulin receptors (NIRKO mice) demonstrate elevated adiposity, hypertriglyceridemia, and insulin resistance under normal chow conditions, indicating that brain insulin signaling is necessary for the tonic regulation of energy homeostasis (6). Given the well-accepted role for the hypothalamus in metabolic control, along with robust expression of insulin receptors in hypothalamic nuclei, numerous investigations have examined a role for hypothalamic insulin receptor-mediated control of energy homeostasis. Interestingly, selective removal of insulin receptors from brain Nkx2.1 or single-minded homolog-1-positive cells (11), steroidogenic factor 1-expressing neurons in the ventromedial hypothalamic nucleus (VMH) (26), melanin concentrating hormone-expressing neurons in the lateral hypothalamus (18), and proopiomelanocortin-producing or agouti-related peptide cells in the arcuate nucleus of the hypothalamus (29) does not fully recapitulate the phenotype observed with global CNS insulin receptor deletion. Collectively, these findings suggest that brain insulin receptor-mediated signaling outside of these neuronal populations is involved in whole body metabolic regulation, although the region in which this occurs remains unknown.

In a variety of animal species and humans, insulin administration has been reported to increase, decrease, or not change blood pressure (1–3, 21, 24, 37); findings that are likely variable due to contrasting peripheral (e.g., direct vascular) and CNS influences of insulin. When focusing on central insulin effects, systemic or intracerebroventricular (ICV) administration of insulin evokes large increases in regional sympathetic outflow, including to cardiovascular organs (7, 9, 10, 40, 43, 45, 59), indicating that CNS insulin signaling is likely involved in blood pressure regulation. However, investigations into the hemodynamic and sympathetic effects of insulin have relied primarily on acute and/or chronic insulin dosing. While beneficial, these approaches do not allow for evaluation of the tonic cardioregulatory role of CNS insulin signaling, which necessitates blockade or disruption of brain insulin receptor pathways (13). Similarly, although a hypothalamic network (arcuate nucleus to paraventricular nucleus) has been hypothesized to mediate the acute cardiovascular sympathetic actions of insulin (9, 10, 45, 59), the CNS nuclei through which insulin chronically modulates the cardiovascular system remain undefined.

The subfornical organ (SFO) is a sensory circumventricular organ located outside the blood-brain barrier. With dense multidirectional projections to cardiovascular and metabolic nuclei, including hypothalamic regions (16, 23, 38, 42), the SFO is intimately involved in cardiometabolic regulation through the sensing of circulating peripheral and cerebrospinal fluid factors (19, 67). Immunocytochemical and transcriptomic analyses indicate that insulin receptors are expressed within the SFO (19, 34, 39, 55–57, 60). Based on this, we reasoned that the SFO is involved in both the cardiovascular and metabolic actions of insulin. Using a combined genomic and integrative physiological approach, our findings collectively highlight a previously undetermined role for SFO insulin receptors in tonic cardiometabolic control.

METHODS

Animals.

Male C57BL/6J mice as well as male mice in which exon 4 of the insulin receptor is flanked by loxP sites (IR^lox/lox mice) were initially obtained from Jackson Laboratories (stock #006955) and maintained as an in-house colony. Mice were provided access to a normal chow diet (percent kcal: 5% fat, 20% crude protein, 4% crude fiber; catalog #0039980, PMI Nutrition International, Brentwood, MO) and water ad libitum under a 12 h light-dark cycle. All experimental procedures were approved by the Institutional Animal Care and Use Committees at the George Washington University and Cornell University following the standard guidelines of the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Adenoviral targeting of the SFO.

An adenoviral vector (serotype 5) encoding Cre-recombinase (AdCre) (or titer-matched control; AdLacZ) driven by the CMV promoter was obtained from the Iowa Gene Transfer Vector Core and used to remove insulin receptors in the SFO as we have previously described (50, 63, 67). With this approach, maximal transgene expression has been shown to occur within 1 wk postviral delivery (49, 50). In brief, 6 wk old IR^lox/lox mice were anesthetized ip with a mixture of ketamine (150 mg/kg) and xylazine (15 mg/kg) and provided ketoprofen (5 mg/kg sc) as an analgesic before surgical procedures. Mice were then placed in a stereotaxic apparatus, and a hole was drilled in the skull under a dissecting microscope. Adenoviral vectors (500 nl over 5 min) were targeted to the SFO via the lateral ventricle (0.3 mm rostral and 1.0 mm lateral to Bregma, 3.0 mm ventral from the dorsal surface of the skull) using pulled glass pipettes and a pressure injection system. Importantly, we and others (12, 50, 65, 67) have shown that this is a highly effective approach to target the SFO, while minimizing damage to the nucleus (see results for further description). In some mice, in the same surgical setting, a 26-gauge cannula (#8IC315GSPCXC; PlasticsOne, Roanoke, VA) was placed into the lateral ventricle (ICV) and fixed to the skull with dental cement, as previously described (63). A subsequent dose of ketoprofen was provided at 24 h postsurgery.

Measurement of body weight, food intake, and tissue sampling.

IR^lox/lox mice were individually housed after viral administration. Weekly body weights, as well as food intake, were measured over a 24 h period at the same time each day for 10 wk, and adipose and liver tissue were collected and weighed when mice were euthanized as we described previously (32, 67). Tissue weights were evaluated both as an absolute value and normalized to body weight.

Radiotelemetry.

In a subset of IR^lox/lox mice, radiotelemeters were implanted for conscious recordings of arterial blood pressure and heart rate, as previously reported (20, 63, 67). Briefly, in the same surgical session as adenoviral gene transfer, mice were anesthetized and provided analgesic, as described above. The radiotelemeter catheter was implanted in the thoracic aorta via the left common carotid artery, and the body of the radiotelemetry probe (TA11PA-C10; Data Sciences International, St. Paul, MN) was placed sc in the right flank and incisions closed and sutured. We housed mice individually after surgery and provided them with 1 wk of recovery before obtaining recordings. Continuous high frequency recordings (1,000 Hz) were performed over a 24 h period once a week throughout the experimental period. Data were analyzed over the 24 h period as well as during the light (06:00–18:00) and dark (18:00–06:00) cycles.

ICV insulin administration.

Following adenoviral delivery, lateral ventricle cannulation, and surgical recovery, mice were anesthetized with isoflurane, and human insulin (35) (0.4 µU/2 µl, #ND-1048; Eli Lilly, Indianapolis, IN) was slowly administered into the lateral ventricle through the ICV cannula with a Hamilton syringe. At 1 h following insulin administration mice were euthanized, and brains prepared for c-Fos immunohistochemistry as described below.

Immunohistochemistry.

Immunohistochemistry procedures were performed as we have previously described (22). Briefly, mice were anesthetized with ketamine/xylazine as above and perfused transcardially with 100 ml of phosphate-buffered saline (PBS, pH 7.4) followed by 100 ml of 3% paraformaldehyde in PBS. Brains were then removed, postfixed with 20% sucrose overnight at 4°C, and embedded in optimal cutting temperature (OCT) compound. A series of coronal SFO sections were obtained using a cryostat with a thickness of 10 μm and stored at –80°C until use. The sections were air-dried before immunohistochemical procedures and then incubated sequentially for 30 min in PBS and 1 h in blocking buffer (3% BSA, 0.3% Triton X-100 in PBS) at room temperature (RT). The sections were then incubated in a humidified chamber overnight in a cold room with primary antibodies diluted in blocking buffer (insulin receptor β: #bs-0681R-Biotin, 1:500, Bioss Antibodies, Woburn, MA; c-Fos: #226004, 1:1,000, Synaptic Systems, Göttingen, Germany; Cre-recombinase: #NB100-56133, 1:4,000, Novus Biologicals, Littleton, CO). Sections were then washed 3×10 min with slow agitation in PBS followed by 2 h incubation at RT with secondary antibodies (Goat anti-rabbit IgG with Alexa Fluor-488: #ab150077, 1:1000, Abcam; Goat anti-Guinea pig IgG with horseradish peroxidase for c-Fos: #ab97155, 1:1000, Abcam, Cambridge, MA). 3,3′-Diaminobenzidine (#E733; Amresco, Solon, OH) was applied to visualize the SFO c-Fos signal following 3×10 min washing in PBS. The sections were coverslipped with Vectashield mounting medium (#H1500; Vector Laboratories, Burlingame, CA) and subjected to microscopy. Images were obtained under a BX43F Olympus fluorescent microscope (Center Valley, PA) with a DP80 camera using the same exposure time across samples. The appropriate working concentration and incubation time for each antibody were determined in a series of preliminary experiments. In particular, for the antibody against the insulin receptor, different working concentrations (1:100–1:2,000) were tested on adjacent brain sections to validate the specificity of the antibody. Immunoreactive signals for the insulin receptor were found according to an antibody concentration-dependent density gradient (data not shown). After validating the optimal working concentration of the antibody, slides including the rostral to caudal extent of the hypothalamus, including known regions of high insulin receptor expression, were utilized to verify the signal specificity of insulin receptors (Fig. 1). Insulin receptors were found to be robustly present in multiple hypothalamic regions as previously reported (39), while no immunoreactive signals were visualized on control slides incubated with blocking buffer without primary antibody (Fig. 1), indicating that immunoreactive signals for the insulin receptor are neither tissue autofluorescence nor noise. For presentation purposes, images were linearly modified with Adobe Photoshop under the same set of adjustments for all images.

Fig. 1.

Verification of the insulin receptor antibody. Immunohistochemistry throughout hypothalamic regions known to express the insulin receptor revealed clear cytosolic and membrane expression of insulin receptors, while no immunoreactivity was detected from negative control slides incubated without primary antibody. Representative of n = 3. Scale bar = 100 µm. ARC, arcuate nucleus; OVLT, organum vasculosum lamina terminalis; SON, supraoptic nuclei; SFO, subfornical organ; 3V, third ventricle.

Hepatic and circulating triglyceride measurements.

Upon euthanization, livers were weighed and frozen immediately on dry ice. At the same time, trunk blood samples were collected via decapitation and stored on ice. Whole blood and livers were stored at –80°C until use. Extraction and quantification of hepatic and blood triglycerides were performed with a triglyceride assay kit (#K622; Biovision, Milpitas, CA) following the manufacturer’s protocol, as described (20). In brief, ~50 mg of liver tissue was homogenized in 5% Triton X-100 solution, incubated in a water bath at 90°C for 5 min, and microcentrifuged for 2 min after cooling down to RT. The supernatant was obtained, and a triglyceride probe, enzyme mix, and lipase were applied. For quantification of circulating triglycerides, the same approach was used on 50 μl of blood. Absorbance at 570 nm was measured with a microtiter plate reader (BioTek, Winooski, VT), and the optical density of each sample was normalized to blank and lipase controls. Liver triglyceride concentrations were interpolated from a linear regression of standard control samples and normalized to the liver weight used for the assay.

Oil Red O staining.

Standard Oil Red O staining was used to visualize lipid deposition in liver. Fresh frozen livers were dehydrated in 30% sucrose solution for 1 h, placed into tissue embedding molds (#18646A; Polysciences, Warrington, PA) with OCT compound, and frozen in a dry ice/isopropanol mixture. Cryosectioning was performed to obtain liver tissues with 30 μm thickness. The sections were equilibrated with 60% isopropyl alcohol for 5 min and then incubated in filtered Oil Red O solution (#A12989; Alfa Aesar, Haverhill, MA) for 7 min. The sections were placed under running tap water for 10 min, counterstained with hematoxylin, and coverslipped with aqua-mounting medium. Images were obtained under a light microscope (Olympus BX43).

Statistical analysis.

Comparisons between two groups over time were analyzed by two-way repeated-measures ANOVA with a Tukey’s multiple comparison post hoc test, and single comparisons were conducted with an unpaired two-tailed student’s t-test (GraphPad Prism 7, San Diego, CA). The significance level was set at P < 0.05. Data are presented as means ± SE.

RESULTS

Insulin receptors are expressed throughout the SFO.

Given limited in-depth investigations of SFO insulin receptor expression, we first confirmed the presence of the insulin receptor in this circumventricular region in C57BL/6J mice. As shown in Fig. 2, immunohistochemistry revealed broad and robust distribution of the insulin receptor throughout the rostral to caudal extent of the SFO. Importantly, the specificity of this signal (detailed in methods) was further confirmed through immunohistochemistry with exclusion of the primary antibody (Fig. 1), indicating that our findings of insulin receptor expression are not due to autofluorescence background that is generally seen in circumventricular sections.

Fig. 2.

Insulin receptors are robustly expressed throughout the SFO. Immunohistochemistry of insulin receptors in the rostral, medial, and caudal portions of the SFO. DAPI is presented as a nuclei marker. The arrows indicate clear insulin receptor-immunoreactivity on DAPI-positive cells. Representative of n = 7. Scale bar = 100 µm.

Cre-mediated deletion of SFO insulin receptors.

We subsequently utilized stereotaxic-targeted delivery of AdCre to remove the SFO insulin receptor in IR^lox/lox mice. In line with our previous reports (63, 67), immunohistochemistry for Cre-recombinase indicated robust Cre expression in the SFO of experimental (AdCre), but not control vector treated (AdLacZ) animals (Fig. 3, top). This Cre-recombinase expression was limited to the SFO as Cre-immunoreactivity was absent around SFO/lateral ventricular regions including the ventral hippocampal commissure and anterior part of the paraventricular thalamic nucleus, as well as in third ventricular regions including the paraventricular nucleus of the hypothalamus (Fig. 3, middle). These findings suggest that our approach for Cre-recombinase delivery was SFO specific. As expected, SFO-targeted delivery of Cre-recombinase resulted in a clear reduction in SFO insulin receptor expression (Fig. 3, bottom). To quantify the degree of insulin receptor knock-down in the SFO with AdCre, we measured the expression of c-Fos, an activity-dependent cellular marker, in the SFO after insulin stimulation. Acute ICV administration of insulin was associated with marked c-Fos expression in the SFO of control animals, while a blunted insulin-induced c-Fos expression was evident in the SFO following AdCre-mediated removal of the SFO insulin receptor. Quantitative analysis demonstrated that removal of SFO insulin receptors resulted in an approximate 45% lower insulin-mediated SFO c-Fos immunoreactivity when compared with control counterparts (c-Fos positive cells per section: 8.6 ± 0.6.4 versus 4.7 ± 1.0, AdLacZ versus AdCre; P < 0.05; n = 3/group, 3–6 sections/animal). Together, these data demonstrate the effectiveness of AdCre-mediated removal of the insulin receptor in the SFO.

Fig. 3.

Cre-recombinase mediated targeting of SFO insulin receptors. Immunohistochemistry of Cre-recombinase in the SFO (top) and brain regions around the lateral ventricle (LV) and third ventricle (3V regions, middle), as well as insulin receptor expression in the SFO (bottom) following SFO-targeted adenoviral delivery of Cre-recombinase (AdCre) to remove insulin receptors, or control vector (AdLacZ). Representative of n = 3/group. Scale bar = 100 µm. PVA, anterior part of paraventricular thalamic nucleus; PVN, paraventricular nucleus of the hypothalamus; vhc, ventral hippocampal commissure.

Targeted deletion of the SFO insulin receptor predisposes to body weight gain and adiposity independent of food intake.

To evaluate a functional role of the SFO insulin receptor, we monitored body weight and food intake following SFO-targeted delivery of AdCre, or control AdLacZ, in IR^lox/lox mice. Early following adenoviral transfer there were no differences in body weight or food intake between groups. However, when mice were maintained on a normal chow diet over the next 10 wk, removal of the SFO insulin receptor resulted in an enhanced body weight gain (Fig. 4A; weekly body weight gain: 0.5 ± 0.1 versus 1.7 ± 0.4 g, AdLacZ versus AdCre, P < 0.05). This did not appear to be due to differences in caloric intake as food intake was comparable between groups across the course of the study (Fig. 4B). Evaluation of regional adipose tissue depots upon euthanization indicated significant elevations in abdominal, gonadal, and subcutaneous inguinal white, as well as subscapular brown adipose tissue in AdCre relative to AdLacZ targeted animals. This was evident when evaluated as absolute tissue mass (not shown) or normalized to body weight (Fig. 4, C–F) to account for the differences in total body mass between groups. Together, these findings demonstrate that under normal chow conditions, the SFO insulin receptors are necessary to protect against excess weight gain and adiposity in a food intake-independent manner.

Fig. 4.

SFO insulin receptor ablation results in excess weight gain and adiposity. Body weight (A) and food intake (B) over 10 wk under normal chow conditions following selective removal of SFO insulin receptors (AdCre) or control (AdLacZ). Abdominal (C), gonadal (D), subcutaneous inguinal (E), and subscapular brown (F) adipose tissue mass 10 wk following SFO targeted AdCre of AdLacZ normalized to body weight. n = 7–11. *P < 0.05 versus AdLacZ.

SFO insulin receptor ablation results in hepatic steatosis and hypertriglyceridemia.

The hepatic manifestation of the metabolic syndrome is commonly characterized by the development of hepatic steatosis (17). At 10 wk following viral targeting of the SFO, liver weight was not different between groups (Fig. 5A). However, hepatic triglyceride levels were ~4-fold higher following removal of the SFO insulin receptor, relative to controls (Fig. 5B). This robust hepatic steatosis was further confirmed at the histological level, with Oil Red O staining revealing significant lipid droplet accumulation in AdCre-targeted animals, whereas minimal hepatic lipid was evident in AdLacZ controls (Fig. 5C). In parallel, circulating triglycerides were markedly elevated following removal of the SFO insulin receptor (Fig. 5D). Together these findings indicate that in addition to elevations in body weight and adiposity, removal of SFO insulin receptors results in hepatic steatosis and hypertriglyceridemia.

Fig. 5.

Removal of the SFO insulin receptors results in hepatic steatosis and hypertriglyceridemia. Liver mass (A), liver triglyceride levels (B), Oil Red O staining of hepatic neutral lipids (C), and blood triglyceride levels (D) 10 wk following removal of SFO insulin receptors (AdCre) relative to control animals (AdLacZ). Scale bar = 200 µm. n = 7–11. *P < 0.05 versus AdLacZ.

Removal of the SFO insulin receptor results in chronic lowering of arterial blood pressure.

Next, we evaluated the cardiovascular role of insulin receptor signaling in the SFO. Following viral targeting and surgical recovery, conscious 24 h radiotelemetry measurements were obtained in AdCre and AdLacZ cohorts for up to 10 wk under normal chow conditions. Heart rate was unchanged following ablation of the SFO insulin receptor (Fig. 6, A and C). However, removal of insulin receptors in the SFO resulted in a reduction in mean arterial pressure (Fig. 6B), which was evident as early as 1 wk postsurgery (104 ± 1 versus 99 ± 2 mmHg, AdLacZ versus AdCre, P < 0.05) and persisted throughout the experimental period. The reduction in 24 h mean arterial blood pressure (Fig. 6, B and D) following removal of SFO insulin receptors appeared to be derived from a reduction in blood pressure during both the dark and light phases (Fig. 6D), although the reduction in light phase blood pressure did not reach statistical significance (P = 0.08). In line with this, diastolic blood pressure was lower in AdCre targeted animals regardless of the light cycle phase (Fig. 6F). In contrast, systolic blood pressure was not different between the two groups (Fig. 6E). In addition, indirect estimates of locomotor activity obtained from the radiotelemetry recordings indicated that removal of SFO insulin receptors resulted in a slight, albeit significant, reduction in 24 h activity. This change in activity was most prominent during the dark phase, whereas light-phase activity was not altered (Fig. 6G). Collectively, these findings demonstrate that under normal chow conditions insulin receptors in the SFO are involved in the tonic regulation of arterial blood pressure and activity, but not heart rate.

Fig. 6.

SFO insulin receptor removal results in a lowering of arterial blood pressure. Twenty-four hour radiotelemetric measurements of heart rate (A) and mean arterial blood pressure (MAP; B) over 10 wk following selective removal of SFO insulin receptors (AdCre) or control (AdLacZ). Heart rate (C), MAP (D), systolic blood pressure (SBP; E), diastolic blood pressure (DBP; F), and activity (G) averaged over the 10 wk period in AdCre and AdLacZ SFO-targeted animals over 24 h, as well as during the light and dark cycles. n = 7–11. *P < 0.05 versus AdLacZ.

DISCUSSION

As a sensory circumventricular organ, the SFO is uniquely situated to sense and integrate circulating information in response to metabolic and cardiovascular states. The current findings further highlight this critical role of the SFO and indicate that: 1) insulin receptors are expressed throughout the SFO; 2) SFO insulin receptor ablation results in excess weight gain, adiposity, hepatic steatosis, and hypertriglyceridemia; and 3) removal of SFO insulin receptors results in a chronic lowering of arterial blood pressure. Collectively, these data highlight a multiple systems influence of SFO insulin receptors in cardiometabolic regulation.

Autoradiographic binding and histological studies initially pointed to the presence of SFO insulin receptors (39, 55, 57, 60) with more recent verification by transcriptomic analysis (19). Our data support and extend these previous findings by revealing that SFO insulin receptors are expressed throughout the rostral to caudal axis of this nucleus with immunoreactivity detected in both the peripheral (i.e., outer shell) and ventromedial core part of the SFO. These findings not only indicate robust insulin receptor density in the SFO but further suggest that SFO insulin receptor-expressing cells are situated to influence a broad array of downstream nuclei. For example, efferent fibers from the outer shell portion of the SFO have been shown to project to hypothalamic nuclei (e.g., paraventricular nucleus, lateral hypothalamus, supraoptic nucleus), whereas the ventromedial core sends efferent projections to the bed nucleus of the stria terminalis (23, 36, 52). While additional investigations are needed to determine the precise efferent targets of SFO insulin receptor-expressing cells, these histological findings point to a potential widespread physiological influence of SFO insulin action.

In addition to its well-recognized role in fluid balance and cardiovascular regulation, it is becoming readily apparent that the SFO participates in whole body metabolic homeostasis. For example, the SFO has been shown to influence feeding behavior (51), weight loss in response to the adipokine leptin, and brown adipose tissue thermogenesis (67). Furthermore, we recently demonstrated that SFO mechanisms contribute to obesity-related conditions, namely hepatic steatosis (20). In line with this growing body of literature, the current data indicate that SFO insulin receptors play a functional metabolic role. Removal of insulin receptors specific to the SFO under normal chow conditions resulted in a metabolic syndrome-like phenotype characterized by an elevated body weight, adiposity, hepatic steatosis, and hypertriglyceridemia, all similar metabolic characteristics to the aforementioned whole brain insulin receptor knockout (NIRKO) model (6). Interestingly, SFO insulin receptor ablation did not influence food intake under normal chow conditions; similar findings have been observed following knockout of insulin receptors from the ventromedial hypothalamus (26), lateral hypothalamus (18), and arcuate nucleus proopiomelanocortin-producing or agouti-related peptide neurons (29). In contrast, knockout of insulin receptors in arcuate nucleus neuropeptide Y (33) or ventrotegmental area dopaminergic neurons (28) results in hyperphagia. Taken together, these findings point to the SFO as part of a distributed network of insulin action in which numerous CNS regions likely work independently or in parallel to influence the diverse metabolic actions of insulin.

In addition to modulation of energy homeostasis, central insulin and insulin receptor signaling evoke changes in cardiovascular sympathetic outflow, indicative of a role in blood pressure regulation. However, discrepant findings exist. ICV administration of low doses of insulin (0.1 to 10 µU/min) in anesthetized rats failed to evoke changes in blood pressure despite insulin-induced elevations in lumbar sympathetic nerve activity (SNA) (40). Similarly, central administration of higher doses of insulin (100 µU/min or greater in anesthetized animals) has been shown to increase adrenal, renal, and lumbar SNA (43, 46, 59) and enhance baroreflex control of SNA and heart rate, with no accompanying change or a fall in blood pressure (43). Thus, while it is well accepted that insulin administration elicits elevations in sympathetic outflow (9, 10, 40, 43, 46, 59, 64), importantly, findings in animal models have been limited to anesthetized conditions that may mask changes in blood pressure. Moreover, the extent to which acute insulin administration reflects chronic CNS influences of circulating insulin is unclear. The current findings using selective Cre-mediated targeting reveals a tonic regulatory role for SFO insulin receptors in cardiovascular regulation and suggests that insulin receptor signaling in the SFO is necessary for the maintenance of arterial blood pressure. Moreover, a modest fall in blood pressure was noted within 1 wk following targeting of SFO insulin receptors (when maximal transgene expression occurs) (49, 50). This occurred at a time at which body weight was not yet altered, in line with an emerging body of evidence supporting dissociation between the central mechanisms involved in metabolic and cardiovascular regulation (20, 44, 66). In addition, arterial blood pressure remained low across the experimental period in Cre-targeted animals despite increases in adiposity, steatosis, and hypertriglyceridemia, all factors that would be expected to contribute to elevations in blood pressure.

Although not a direct focus of the current studies, our findings suggest that SFO insulin receptors may also participate in the regulation of activity. While measurements of locomotor activity via radiotelemetry are indirect estimates, a reduction in overall activity was found following SFO insulin receptor removal. Previous findings have also suggested an influence of CNS insulin receptors on activity and behavior, although the direction of influence (i.e., increase or decrease) appears to differ depending on the brain region that was investigated (8, 18, 28, 33). Given the strong influence of locomotor activity on cardiometabolic investigation, it is possible that the changes in “metabolic” parameters and arterial blood pressure in the current study could be explained, at least in part, by the changes in activity following SFO insulin receptor ablation. That is, a reduction in activity would be expected to increase body weight/adiposity, contribute to hepatic steatosis, and lower arterial blood pressure (particularly during the dark phase in mice). However, future in-depth investigations are needed to determine the activity-dependent and -independent influences of SFO insulin receptors on cardiometabolic regulation.

The SFO is highly connected to a number of endocrine and autonomic nuclei, notably within the hypothalamus, that influence the cardiovascular and metabolic systems (16, 23, 38). Our findings of a divergent influence of SFO insulin receptors on metabolic versus cardiovascular regulation suggests that these physiological influences may occur through different efferent networks. Alternatively, it is possible that the metabolic and cardiovascular actions of insulin in the SFO are mediated by discrete neuron subpopulations that project to the same and/or different downstream nuclei. In support of this, electrophysiological findings have revealed a clear delineation in the responsiveness of SFO neurons to insulin. Specifically, whole-cell current clamp experiments indicated that in response to exogenous insulin, one-third of tested SFO neurons were insensitive, one-third depolarized, and one-third hyperpolarized (30). Thus, these excitatory and inhibitory effects may relate to the divergent cardiometabolic outcomes of SFO insulin receptor signaling, although future experiments are warranted in this regard, as well as investigation of integrated downstream circuits.

Importantly, our findings are limited to male mice. This is critical given sex-specific influences on whole body insulin signaling and sensitivity (41, 47). In addition, brain insulin receptor levels fluctuate with the estrus cycle (61), suggesting that female hormones are closely associated with central insulin receptor expression. A number of diverse animal models that are used to investigate components of the metabolic syndrome also suggest that females may be, at least partially, protected from metabolic syndrome development (53, 54, 58, 62). Moreover, sexually dimorphic influences in insulin-induced increases in sympathetic outflow have been recently demonstrated (48). Thus, it is possible that different and/or similar findings would be present if the current studies were replicated in female mice; an area that requires critical attention.

Perspectives

A common consensus is that insulin resistance is central to the development and progression of obesity-related comorbidities (i.e., metabolic syndrome) (27). However, it is important to consider that in metabolic syndrome conditions, selective insulin resistance occurs in which certain insulin-mediated responses are blunted whereas others are maintained (5, 30). This divergent and selective process has been suggested to occur in the CNS; during obesity the cardiovascular actions of CNS insulin are maintained while the metabolic responses are lost (26, 31), resulting in a deleterious metabolic state in concert with cardiovascular disease, including hypertension. Our findings point to a dual physiological role of SFO insulin receptor signaling, suggesting that selective insulin resistance in the SFO could contribute to cardiometabolic disease development. While speculative, if “metabolic” SFO insulin receptor responses were to become insulin resistant, whereas “cardiovascular” insulin receptor signaling was preserved or enhanced, this could contribute to damaging metabolic conditions accompanied by elevations in blood pressure. Overall, these findings highlight a novel CNS site of insulin action in the SFO, which is uniquely situated to broadly impact cardiometabolic regulation.

GRANTS

Funded in part by National Heart, Lung, and Blood Institute Grants R00HL-116776 and R01HL-141393.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.K.J., J.A.H., H.S., K.A.B., S.D.B., and C.N.Y. conceived and designed research; J.K.J., J.A.H., H.S., K.A.B., S.D.B., and C.N.Y. performed experiments; J.K.J., J.A.H., H.S., K.A.B., S.D.B., and C.N.Y. analyzed data; J.K.J., J.A.H., H.S., K.A.B., S.D.B., and C.N.Y. interpreted results of experiments; J.K.J. and C.N.Y. prepared figures; J.K.J. and C.N.Y. drafted manuscript; J.K.J., J.A.H., H.S., K.A.B., S.D.B., and C.N.Y. edited and revised manuscript; J.K.J., J.A.H., H.S., K.A.B., S.D.B., and C.N.Y. approved final version of manuscript.