Sex differences in the sympathoexcitatory response to insulin in obese rats: role of neuropeptide Y

Abstract

Key points

• Intracerebroventricular insulin increased sympathetic nerve activity (SNA) and baroreflex control of SNA and heart rate more dramatically in obese male rats; in obese females, the responses were abolished.

• In obese males, the enhanced lumbar SNA (LSNA) responses were associated with reduced tonic inhibition of LSNA by neuropeptide Y (NPY) in the PVN. However, PVN NPY injection decreased LSNA similarly in obesity prone/obesity resistant/control rats. Collectively, these results suggest that NPY inputs were decreased. In obese females, NPY inhibition in the PVN was maintained. Moreover, NPY neurons in the arcuate nucleus became resistant to the inhibitory effects of insulin.

• A high‐fat diet did not alter arcuate NPY neuronal InsR expression in males or females.

• Obesity‐induced ‘selective sensitization’ of the brain to the sympathoexcitatory effects of insulin and leptin may contribute to elevated basal SNA, and therefore hypertension development, in males with obesity.

• These data may explain in part why obesity increases SNA less in women compared to men.

Abstract

Obesity increases sympathetic nerve activity (SNA) in men but not women; however, the mechanisms are unknown. We investigated whether intracerebroventricular insulin infusion increases SNA more in obese male than female rats and if sex differences are mediated by changes in tonic inhibition of SNA by neuropeptide Y (NPY) in the paraventricular nucleus (PVN). When consuming a high‐fat diet, obesity prone (OP) rats accrued excess fat, whereas obesity resistant (OR) rats maintained adiposity as in rats eating a control (CON) diet. Insulin increased lumbar SNA (LSNA) similarly in CON/OR males and females under urethane anaesthesia. The LSNA response was magnified in OP males but abolished in OP females. In males, blockade of PVN NPY Y1 receptors with BIBO3304 increased LSNA in CON/OR rats but not OP rats. Yet, PVN nanoinjections of NPY decreased LSNA similarly between groups. Thus, tonic PVN NPY inhibition of LSNA may be lost in obese males as a result of a decrease in NPY inputs. By contrast, in females, PVN BIBO3304 increased LSNA similarly in OP, OR and CON rats. After insulin, PVN BIBO3304 failed to increase LSNA in CON/OR females but increased LSNA in OP females, suggesting that with obesity NPY neurons become resistant to the inhibitory effects of insulin. These sex differences were not associated with changes in arcuate NPY neuronal insulin receptor expression. Collectively, these data reveal a marked sex difference in the impact of obesity on the sympathoexcitatory actions of insulin and implicate sexually dimorphic changes in NPY inhibition of SNA in the PVN as one mechanism.

Key points

• Intracerebroventricular insulin increased sympathetic nerve activity (SNA) and baroreflex control of SNA and heart rate more dramatically in obese male rats; in obese females, the responses were abolished.

• In obese males, the enhanced lumbar SNA (LSNA) responses were associated with reduced tonic inhibition of LSNA by neuropeptide Y (NPY) in the PVN. However, PVN NPY injection decreased LSNA similarly in obesity prone/obesity resistant/control rats. Collectively, these results suggest that NPY inputs were decreased. In obese females, NPY inhibition in the PVN was maintained. Moreover, NPY neurons in the arcuate nucleus became resistant to the inhibitory effects of insulin.

• A high‐fat diet did not alter arcuate NPY neuronal InsR expression in males or females.

• Obesity‐induced ‘selective sensitization’ of the brain to the sympathoexcitatory effects of insulin and leptin may contribute to elevated basal SNA, and therefore hypertension development, in males with obesity.

• These data may explain in part why obesity increases SNA less in women compared to men.

Introduction

Obesity increases the risk for cardiovascular diseases, such as hypertension. Given the association between visceral adiposity and high blood pressure, mechanisms have been extensively investigated. The consensus is that one mechanism may be increased sympathetic nerve activity (SNA) (Hall et al. 2010; Grassi et al. 2015) partly as a result of the central actions of leptin and insulin (Lim et al. 2013). Not only does obesity increase the plasma levels of these metabolic peptides, but also leptin increases SNA to the same extent or more in obese animals (Prior et al. 2010; Mark, 2013). However, whether obesity similarly amplifies the sympathoexcitatory action of insulin is unknown.

The impact of obesity on SNA and arterial pressure (AP) differs between the sexes (Brooks et al. 2015; Lambert et al. 2015; Faulkner & Belin de Chantemele, 2018). In men, obesity increases muscle SNA (MSNA). In women, MSNA either does not relate (Lambert et al. 2007; Tank et al. 2008; Maqbool et al. 2010; Brooks et al. 2015; Straznicky et al. 2017) or relates weakly (Ribeiro et al. 2001) to indices of adiposity such as body mass index. One explanation for this sex difference may be that, in females, the sympathoexcitatory effect of leptin varies during the reproductive cycle (Shi & Brooks, 2015). Another potential mechanism is that obesity amplifies the SNA responses to insulin (and leptin) in males but not females. Therefore, one purpose of the present study was to investigate whether intracerebroventricular (i.c.v.) infusion of insulin increases lumbar SNA (LSNA) more in male compared to female rats made obese by consumption of a high‐fat diet (HFD). Insulin also enhances baroreceptor reflex control of SNA, by increasing baroreflex gain (the change in SNA elicited by a specific change in AP) and the maximum SNA reached at low AP levels (Pricher et al. 2008; Young et al. 2010; Cassaglia et al. 2011). Following ingestion of food, the action of insulin to enhance reflex SNA activation serves to maintain AP in response to the baroreceptor unloading as a result of mesenteric vascular dilatation (Aronow, 1995) and to insulin‐induced vasodilatation (Young et al. 2010). In addition, the increased SNA stimulates glucose uptake (Nonogaki, 2000). However, eating is a well‐established trigger for myocardial infarction as a result of SNA activation (Culic, 2007; Nawrot et al. 2011). Therefore, we also investigated whether obesity enhances insulin‐induced increases in baroreflex gain in male rats and whether this action is less in female rats.

We next considered mechanisms for potential sex differences in insulin responses. Accumulated information indicates that insulin increases SNA by binding to receptors in only one brain site, the arcuate nucleus (ArcN), via a neuropathway that includes the PVN: (i) ArcN blockade with muscimol (Cassaglia et al. 2011) or inhibition of ArcN insulin actions with an insulin affibody (Luckett et al. 2013) prevents or reverses the sympathoexcitatory effects of i.v. insulin; (ii) ArcN insulin dose dependently increases SNA (Cassaglia et al. 2011); (iii) PVN blockade with muscimol reverses the effects of i.v. (Ward et al. 2011) or ArcN (Cassaglia et al. 2011) insulin; and (iv) PVN insulin does not increase SNA (Cassaglia et al. 2011; Ward et al. 2011). The primary ArcN neurons that project to the PVN are excitatory pro‐opiomelanocortin (POMC) neurons, which release α‐melanocyte stimulating hormone (α‐MSH), and sympathoinhibitory neuropeptide Y (NPY) neurons. ArcN insulin increases SNA by both increasing excitatory α‐MSH inputs (Ward et al. 2011) and suppressing tonically inhibitory NPY inputs (Cassaglia et al. 2016) into the PVN.

Although no information is available in obese females, in obese males, POMC neurons (that release α‐MSH) have been implicated in the increases in SNA and AP (Hall et al. 2010). However, α‐MSH‐induced sympathoexcitation in the PVN requires simultaneous withdrawal of tonic NPY inhibition, mediated via Y1 receptors (NPY1R) (Shi et al. 2015). Therefore, we began our investigation of the mechanisms mediating sex differences in the impact of obesity on insulin‐induced sympathoexcitation by focusing on ArcN NPY neurons. We tested the hypothesis that insulin increases LSNA more in obese males, in part because of reduced tonic NPY inhibition of PVN presympathetic neurons, whereas NPY inhibition is maintained in obese females. This hypothesis was tested in part by determining whether the increase in SNA elicited by acute blockade of PVN NPY1R is reduced in obese males but normal in obese females.

Methods

Animals and ethical approval

Male (n = 139) and female (n = 136) Sprague–Dawley (SD) rats (Charles River Laboratories, Inc., Wilmington, MA, USA) were housed individually under a 12:12 h light/dark cycle in a temperature‐controlled (22 ± 2°C) room. Food and water were available ad libitum. All procedures were conducted in accordance with the National Institutes of Health Guide for the Health and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee of Oregon Health & Science University (reference no. IS00002352).

Diet‐induced obesity (DIO) rat model

Rats aged 8 weeks (males weighing ∼250 g; females weighing ∼200 g) were placed on a purified moderately high‐fat diet (HFD) [33% kcal as fat; LabDiet 571R (5001 with 10% lard) (Purina, Richmond, IN, USA) or D16030908 (Research Diets, New Brunswick, NJ, USA)] or a low‐fat control diet (CON) [13.5% kcal as fat; LabDiet 5001 (Purina) or D16030911 (Research Diets)]. HFD fed rats diverge into populations with high body weight (BW) gain (obesity prone; OP) or BW gain similar to those fed the low‐fat CON diet (obesity resistant; OR). Consistent with previous work (Dobrian et al. 2000; Boustany et al. 2005; Levin & Strack, 2008; McCully et al. 2012; Zhao et al. 2012), after ∼4 weeks on the diet, rats in the top tertile of BW gain were defined as OP and the bottom tertile were defined as OR.

After 4 weeks on the CON diet or HFD, body fat content was assessed in a subset of male and female rats (three sets each), using an Echo MRI 3‐in‐1 Animal Composition Analyzer (Echo Medical Systems, Houston, TX, USA) and related to the increment in BW. Two indices of insulin sensitivity were derived from fasting insulin and glucose levels in some rats 4–5 weeks after initiating the HFD or CON diet: the homeostasis model assessment of insulin resistance (HOMA‐IR) and the quantitative insulin sensitivity check index (QUICKI), both of which have been validated in SD rats against the ‘gold standard’ hyperinsulinaemic‐euglycaemic clamp (Cacho et al. 2008). Briefly, rats were fasted for 6–7 h. Then, the lateral saphenous vein was punctured with a 20 G needle, and blood was collected into heparinized microhaematocrit capillary tubes and centrifuged. The plasma was frozen (–20°C) until assayed for plasma insulin levels using an enzyme immunoabsorbent assay (#10‐1250‐01; Mercodia, Uppsala, Sweden) by the Endocrine Technologies Support Core at Oregon National Primate Research Center (Beaverton, OR, USA). Blood was also collected onto a OneTouch Ultra blood glucose strip and blood glucose level was determined using a One Touch Ultra2 glucometer (LifeScan Europe, Zug, Switzerland). The HOMA‐IR and QUICKI were calculated from these fasting glucose and insulin levels, as described previously (Cacho et al. 2008).

Experimental protocols

Data acquisition

Pulsatile and mean arterial pressure (MAP), heart rate (HR) and SNA were continuously recorded using an MP100 data acquisition and analysis system (BIOPAC, Diss, UK) sampling at 2000 Hz and a Grass tachograph amplifier (Grass Instruments, Quincy, MA, USA). SNA was bandpass filtered (100–3000 Hz), amplified (10,000×), rectified and integrated in 1 s bins. MAP and HR data were also grouped into 1 s bins from which mean values were obtained. In each rat, post‐mortem SNA was quantified at the end of the experiment, and this background level was subtracted from values of SNA recorded during the experiment. SNA was normalized to control SNA, which was the 30 s average just before the first baseline baroreflex curve was produced or before experimental infusions or injections were initiated (% of control).

i.c.v. insulin infusion

Surgery

Female rats were studied without establishing the oestrous cycle stage because responses to i.c.v. insulin do not vary throughout the cycle (Shi & Brooks, 2015). Male and female rats were anaesthetized with isoflurane, and femoral arterial and venous catheters, a lateral cerebroventricular cannula, and a lumbar nerve electrode were implanted as described previously (Cassaglia et al. 2011; Li et al. 2013; Shi & Brooks, 2015). Then, the rats were transitioned to urethane anaesthesia (CON and OR rats: 1.1 g kg^–1; OP rats: the same absolute amount of urethane as their OR counterparts). Importantly, a similar level of anaesthetic depth was confirmed in all groups by lack of a withdrawal or pressor response to toe pinch.

Experimental protocol

After at least 1 h of stabilization following urethane infusion, baseline measurements of MAP, HR, LSNA and baroreflex function were made as described previously (Cassaglia et al. 2011; Li et al. 2013; Shi & Brooks, 2015). Then, an i.c.v. infusion of insulin [100 μU min^–1 (Muntzel et al. 1994; Pricher et al. 2008)] or the artificial cerebrospinal fluid (aCSF) vehicle (0.6 μL min^–1) was initiated and continued for 2 h (Pricher et al. 2008). Basal levels of MAP, HR and LSNA and baroreflex control of LSNA and HR were re‐examined 1 and 2 h later.

Baroreflex curve generation

Complete sigmoidal baroreflex curves were generated by first quickly lowering MAP to ∼50 mmHg by i. v. infusion of nitroprusside (1 mg mL⁻¹; 20 μL min⁻¹), and then by steadily and smoothly raising MAP to ∼175 mmHg over 3–5 min by both withdrawing nitroprusside and infusing phenylephrine at increasing rates (1 mg mL⁻¹; 1–35 μL min⁻¹). Baroreflex curves relating SNA to MAP were constructed from data obtained during the MAP upswing from 50 to 175 mmHg. The sigmoidal baroreflex relationships were fitted and compared using the Boltzman equation: LSNA or HR = (P ₁ – P ₂)/[1 + expP₄(MAP – P ₃)]. P ₁ is the maximum LSNA or HR, P ₂ is the minimum LSNA or HR, P ₃ is the MAP associated with the LSNA/HR value midway between the maximal and minimal values (BP50; denotes position of the curve on the x‐axis) and P ₄ is the coefficient used to calculate maximum gain, – (P ₁ – P ₂)/(P₄ × 4), which is an index of the slope of the linear part of the sigmoidal baroreflex curve. Absolute values of gain, the maximum and minimum LSNA or HR, and the BP50 are shown in Figs 3, 4, 5.

Figure 3.

The effect of i.c.v. insulin infusion on baroreflex control of LSNA

A and C, representative experiments. B and D, the grouped data. In the representative experiments (A and C), black closed squares illustrate the control curve; open grey triangles, 1 h after i.c.v. insulin; and closed grey squares, 2 h after i.c.v. insulin. In males (A and B), i.c.v. insulin enhanced baroreflex control of LSNA (maximum and gain) in OR and CON rats, and this effect was markedly greater in OP rats. In females (C and D), insulin similarly enhanced baroreflex control of LSNA (maximum and gain) in OR and CON rats, although these effects were not observed in OP rats. Baseline curves, solid circles and line; 1 h i.c.v. insulin, open squares and solid line; 2 h i.c.v. insulin, open triangle and dashed line. Insets: absolute values of baroreflex gain. Filled bar, control values; open bar, 1 h i.c.v. insulin; hatched grey bar, 2 h i.c.v. insulin. Note different scale for OP males. For n, see Fig. 2. ^* P < 0.05 compared to baseline within group. †P < 0.05, OP different from OR and/or CON at the same time.

Figure 4.

The effect of i.c.v. insulin infusion on baroreflex control of HR

In males, i.c.v. insulin did not alter baroreflex control of HR in OR or CON rats but increased the HR baroreflex maximum and gain in OP rats. In females, insulin did enhance baroreflex control of HR (maximum and gain) in OR and CON rats, although these effects were abolished in OP rats. Baseline curves, solid circles and line; 1 h i.c.v. insulin, open squares and solid line; 2 h i.c.v. insulin, open triangle and dashed line. Insets: absolute values of baroreflex gain. Filled bar, control values; open bar, 1 h i.c.v. insulin; hatched grey bar, 2 h i.c.v. insulin. For n, see Fig. 2. ^* P < 0.05 compared to baseline within group. †P < 0.05, OP different from OR and/or CON at the same time.

Figure 5.

Infusion of aCSF i.c.v. did not alter baroreflex control of LSNA or HR in male or female OP, OR or CON rats

Baseline curves: filled black circles, solid black line; 1 h i.c.v. aCSF: open square, solid black line; 2 h i.c.v. aCSF: grey open triangle, dashed line. Insets: absolute values of baroreflex gain. Filled bar, control value; open bar, 1 h i.c.v. aCSF; hatched grey bar, 2 h i.c.v. aCSF. For n, see Fig. 2 D.

PVN nanoinjections

Surgery

After surgery under isoflurane anaesthesia to implant femoral catheters and a lumbar nerve electrode, the rats were transitioned to urethane anaesthesia as above.

Blockade of PVN NPY1R

After a 60 min equilibration period, the NPY1R antagonist, BIBO3304 (60 nL of 1 mm L^–1) (Cassaglia et al. 2014) or aCSF (60 nL) was nanoninjected bilaterally into the PVN of male and female CON, OR and OP rats, as described previously (Cassaglia et al. 2014) and measurements of MAP, HR and LSNA were continued for 10–15 min. In some rats that received PVN aCSF injections, another protocol was performed at least 30 min later (i.e. i.c.v. aCSF or PVN BIBO3304).

Responses to NPY

To determine whether obesity suppresses the responses to NPY in the PVN in males, another set of CON, OR and OP rats received bilateral PVN injections of NPY (60 nL of 0.1 mm L^–1; Tocris Bioscience, St Louis, MO, USA) as described previously (Cassaglia et al. 2014).

Blockade of PVN NPY1R after i.c.v. insulin

To investigate whether insulin suppresses tonic NPY sympathoinhibition in obese females, BIBO3304 was injected into the PVN in another group of female CON, OR and OP rats, 2 h after initiating an i.c.v. insulin infusion, as described above.

Quantification of insulin receptor (InsR) expression in NPY neurons

To begin to investigate the mechanisms by which obesity alters responses to insulin, InsR mRNA was quantified in NPY neurons throughout the ArcN using fluorescent in situ hybridization (RNAScope; ACDBio; San Francisco, CA, USA) in male and female CON, OR and OP rats. We did not identify the reproductive cycle stage of the females because the sympathoexcitatory response to insulin does not vary with the cycle (Shi & Brooks, 2015) and because a HFD inhibits cycling in SD rats (Z. Shi and V. L. Brooks, unpublished observations). Briefly, male and female CON, OR and OP rats (4–6 weeks on HFD) were deeply anaesthetized and perfused transcardially with physiological saline followed by ice cold 4% paraformaldehyde in 0.1 m sodium phosphate buffer. Brains were removed, post fixed overnight at 4°C, incubated in sucrose, snap‐frozen, sectioned (15–20 μm) and mounted onto SuperFrost Plus slides (Fisher Scientific, Waltham, MA, USA) and stored at –80°C. In situ hybridization was performed in accordance with the manufacturer's instructions using the RNAScope® Fluorescent Multiplex Kit (320850; ACDBio) and ACDBio probes: Rn‐Npy‐C2 (450971‐C2; Accession No. NM_012614.2; target region 8–498) and Rn‐Insr (406421; Accession No. NM_017071.2; target region 431–1287). Generally, brains from a set of CON, OR and OP male and/or female rats were processed in parallel. Sections were imaged on an AxioImager fluorescent Apotome 2 (Zeiss, Jena, Germany) with a 20 × 0.8 PlanApo objective; Rn‐Npy‐C2 probe was detected using settings for fluorophore Atto 647 nm and Rn‐Insr probe was detected using Atto 550 nm.

Images were analysed using Zen 2 (Blue Edition) software (Zeiss): the ArcN region of interest was outlined, the threshold for A647 signal was manually set to select only neurons with NPY mRNA, and then the threshold for A555 signal was set to count individual puncta representing InsR mRNA within NPY positive neurons. For each section, one side was analysed: the total area of NPY positive neurons was summed, and the number of InsR mRNA puncta within NPY positive neurons was counted and summed. NPY neuronal areas with 0–2 InsR puncta (background) were excluded in the InsR count. Then, for each animal, the values obtained from all sections within three regions of the ArcN (mm caudal from bregma): 2.3–3.0 (rostral); 3.0–3.4 (middle); and 3.4–4.1 (caudal) were averaged and compared between groups. In Fig. 10, to ensure visible NPY mRNA signal, the signal was optimized slightly in all images equally in Zen 2.

Figure 10.

ArcN NPY InsR expression is unaltered by a HFD in males or females

In both males (A) and females (C), NPY and InsR expression varied in parallel from rostral to caudal, reaching a peak in the middle region (3.0–3.4 mm caudal to bregma). As a result, the number of InsR puncta per total NPY area did not vary. B, detailed quantification throughout the ArcN in representative CON male and female rats. NPY and InsR expression and InsR/NPY were not different between the sexes or between CON, OR and OP rats. Representative images from the three ArcN regions are shown for CON males (D) and females (F). Images in (E) are expanded from the boxed areas in (D). n = 3 or 4 per group. ^* P < 0.05, middle level compared to rostral and caudal levels.

Drugs

The dose of (human) insulin (100 units mL^–1; Novo Nordisk, Bagsværd, Denmark; Eli Lilly and Company, Indianapolis, IN, USA) was selected based on previous dose–response studies (Muntzel et al. 1994; Pricher et al. 2008; Cassaglia et al. 2011). BIBO3304 (Tocris) is a highly selective NPY1R antagonist (Cassaglia et al. 2014). The dose of BIBO3304 was previously shown to block the sympathoinhibitory effects of exogenous NPY in PVN (Cassaglia et al. 2014). A dose of NPY (Tocris) previously shown to elicit rapid decreases in SNA, MAP and HR (Cassaglia et al. 2014) was used. These drugs were dissolved in aCSF containing (in mmol L^–1): 128 NaCl, 2.6 KCl, 1.3 CaCl₂, 0.9 MgCl₂, 20 NaHCO₃ and 1.3 Na₂HPO₄; pH was corrected to 7.4, and the aCSF was filtered before use.

Histological verification of PVN nanoinjections

Histological maps [adapted from (Paxinos & Watson, 2007)], illustrating PVN injection sites, and those that missed the PVN, are provided in Fig. 9.

Figure 9.

Histological maps

Maps illustrating PVN (and missed) injection sites in male and female rats.

Statistical analysis

Differences between groups in baseline and insulin sensitivity measurements were determined using two‐way ANOVA and the post hoc Newman–Keuls. For experiments involving i.c.v. infusions, between group differences within sex were first tested with a three‐way ANOVA [group (OP, OR, CON), treatment (i.c.v. insulin or aCSF) and time] for repeated measures (RM). If this analysis revealed a significant interaction, further within treatment two‐way RM ANOVA and the post hoc Newman–Keuls test was used to identify specific within and between group differences. Two‐way RM ANOVA was also used to identify between group differences in responses to PVN nanoinjections. Finally, differences in NPY, InsR and InsR/NPY expression were determined using three‐way RM ANOVA (with factors sex, group and ArcN level). In some cases, the data were log‐transformed before analysis to normalize between group differences in variances. Data are expressed as the mean ± SEM. P < 0.05 was considered statistically significant.

Results

Characteristics of the DIO model in males and females

After 4 weeks on a HFD, male and female OP rats weighed more than their OR counterparts, which had a weight similar to CON rats (Fig. 1 A). Nevertheless, the HFD elicited a range of increases in BW, and fat accrual was well correlated with the increment in BW, in both males and females (Fig. 1 B). Interestingly, as reported previously (Levin et al. 1997), some male and female rats on the CON diet also added excess fat, although the consumption of a HFD clearly magnified the range in BW produced (Fig. 1 B).

Figure 1.

Baseline metabolic measurements

A, in both males (n = 14–17) and females (n = 15–18), OP rats gained more BW than OR or CON rats. ^* P < 0.05, OP compared to OR and CON. B, accrued body fat relates to the increment in BW in both male and female rats on a CON diet or HFD. C, male OP rats are more insulin resistant than OP female rats. HOMA‐IR and QUICKI were calculated as described previously Cacho et al. (2008). Two‐way ANOVA revealed significant sex differences for QUICKI, HOMA‐IR and plasma insulin levels (†P < 0.05 between sexes) and a significant effect (P < 0.002) of diet for all variables. ^* P < 0.05 compared to CON within group. ^** P < 0.05 compared to OR within group.

As expected (Zhao et al. 2012), male OP rats were insulin resistant, as indicated by elevated fasting plasma insulin levels, fasting plasma glucose levels that increased with increments in BW, decreases in QUICKI, and increases in HOMA‐IR (Fig. 1 C). By contrast, in females, plasma glucose levels did not correlate with the increases in BW, and plasma insulin levels were not increased in OP rats (although insulin did relate weakly to increases in BW; r ² = 0.15) (Fig. 1 C). In females, only the QUICKI suggested a degree of insulin resistance in OP animals. Most importantly, the indices of insulin sensitivity and plasma insulin levels were lower in OP female compared to OP male rats (Fig. 1 C).

Baseline MAP and HR values were not different between OP and OR/CON rats, in either males or females (Table 1). Baroreflex control of LSNA or HR (Figs 3 and 4) was also unaltered by obesity in either sex in the anaesthetized state, except that the baroreflex LSNA maximum was elevated in OP males compared to OR and CON males (Fig. 3 B) (P < 0.05).

Table 1.

Baseline levels of MAP and HR in anaesthetized male and female OP, OR and CON rats

	Males			Females
	OP (n = 27)	OR (n = 26)	CON (n = 24)	OP (n = 25)	OR (n = 25)	CON (n = 26)
MAP (mmHg)	111 ± 3	112 ± 2	112 ± 3	110 ± 2	105 ± 4	112 ± 3
HR (beats min–1)	374 ± 6	379 ± 9	354 ± 7	376 ± 8	367 ± 7	379 ± 8

The action of i.c.v. insulin to increase LSNA and baroreflex function is enhanced in obese male rats but is abolished in obese female rats

In lean male rats (OR and CON), i.c.v. insulin similarly increased LSNA (Fig. 2 B and C) and the gain and maximum of baroreflex control of LSNA (Fig. 3 B). However, as reported previously (Pricher et al. 2008), i.c.v. insulin had little effect on baroreflex control of HR (Fig. 4). In OP males, the insulin‐induced rise in LSNA (Fig. 2 A, left) and enhancement of baroreflex control of LSNA (Fig. 3 A and B) were magnified greatly. Moreover, in obese males, i.c.v. insulin now increased the gain and maximum of baroreflex control of HR (Fig. 4). The i.c.v. insulin infusions had no effects on baseline MAP or HR in any group of males (Table 2).

Figure 2.

The sympathoexcitatory response

The sympathoexcitatory response to i.c.v. insulin is amplified in obese males (A, left) but abolished in obese females (A, right) compared to male and female OR and CON rats (B and C). Note different scales in OP vs. OR/CON rats. Infusion of aCSF into the lateral ventricle had no effects in male or female CON, OR and OP rats (D). ^* P < 0.05 compared to baseline (time zero) within group. †P < 0.05, OP different from OR and/or CON at the same time.

Table 2.

Effect of i.c.v. insulin or aCSF infusions on MAP and HR in male and female OP, OR and CON rats

Males	OP			OR			CON
Insulin	Base	1 h	2 h	Base	1 h	2 h	Base	1 h	2 h
MAP (mmHg)	119 ± 8	108 ± 7	109 ± 6	107 ± 5	99 ± 7	96 ± 5	98 ± 4	84 ± 9	91 ± 9
HR (beats min–1)	357 ± 11	372 ± 24	357 ± 23	350 ± 12	353 ± 19	364 ± 17	330 ± 15	329 ± 12	343 ± 15

aCSF	Base	1 h	2 h	Base	1 h	2 h	Base	1 h	2 h
MAP (mmHg)	122 ± 3	126 ± 4	124 ± 2	122 ± 1	115 ± 3	109 ± 4	124 ± 6	122 ± 7	124 ± 6
HR (beats min–1)	350 ± 17	363 ± 17	368 ± 14	406 ± 20	402 ± 17	372 ± 11	381 ± 20	380 ± 15	370 ± 17

Females
Insulin	Base	1 h	2 h	Base	1 h	2 h	Base	1 h	2 h
MAP (mmHg)	108 ± 9	113 ± 7	111 ± 6	111 ± 9	113 ± 8	109 ± 11	123 ± 3	124 ± 2	135 ± 4
HR (beats min–1)	403 ± 15	393 ± 21	385 ± 27	357 ± 8	414 ± 29*	438 ± 27*	370 ± 15	387 ± 10	433 ± 17a

aCSF	Base	1 h	2 h	Base	1 h	2 h	Base	1 h	2 h
MAP (mmHg)	106 ± 6	106 ± 6	105 ± 7	107 ± 7	108 ± 3	105 ± 4	114 ± 8	114 ± 5	115 ± 7
HR (beats min–1)	371 ± 10	374 ± 7	370 ± 15	382 ± 23	374 ± 28	383 ± 24	374 ± 17	385 ± 21	394 ± 16

^a P < 0.05, compared to baseline within a group. For n, see Fig. 2.

In OR and CON females, i.c.v. insulin increased basal LSNA (Fig. 2 B and C) and the gain and maximum of baroreflex control of LSNA (Fig. 3 D) similarly to lean males. We also confirmed (Pricher et al. 2008) that, unlike males, i.c.v. insulin infusion increases the gain and maximum of baroreflex control of HR in lean female (OR and CON) rats (Fig. 4). The responses of OP females to i.c.v. insulin were dramatically distinct from both lean females and OP males: (i) i.c.v. insulin failed to increase LSNA (Fig. 2 A, right) and its baroreflex regulation (Fig. 3 C and D); and (ii) the effect of insulin to enhance baroreflex control of HR was similarly abolished (Fig. 4). In females, as in males, i.c.v. insulin did not alter MAP; however, insulin increased HR in lean rats but not OP rats (Table 2).

Infusion of aCSF i.c.v. had no effects on baseline variables (Fig. 2 D and Table 2) or baroreflex control of LSNA or HR (Fig. 5) in any group of males or females. Therefore, we conclude that obesity differentially alters the sympathoexcitatory responses to insulin in males vs. females: insulin responses were markedly enhanced in obese males but eliminated in obese females.

Obesity reduces tonic PVN NPY sympathoinhibition in obese males but not obese females

We next investigated whether the dramatic sex differences in the responses of obese rats to insulin were mediated in part by differential changes in tonic inhibition of PVN pre‐sympathetic neurons by NPY. As expected (Cassaglia et al. 2014; Shi et al. 2015), bilateral PVN nanoinjections of the NPY1R antagonist BIBO3304 increased LSNA in CON and OR rats (Fig. 6, left); injections outside the PVN were ineffective (to 95 ± 3% control; n = 3 CON males and 3 CON females) (Fig. 9). However, in male OP rats, PVN BIBO3304 did not alter LSNA (Fig. 6, left). By contrast, in CON, OR and OP females, PVN BIBO3304 triggered similar increases in LSNA (Fig. 6, right). PVN BIBO3304 did not significantly alter MAP or HR in any group (data not shown). PVN nanoinjections of aCSF also had no effects in any group (Table 3). Therefore, we conclude that, in males, obesity suppresses tonic inhibition of LSNA by NPY in the PVN, although this inhibition is maintained in obese females.

Figure 6.

Effects of PVN nanoinjections of BIBO3304 on LSNA in male and female OP, OR and CON rats

Top: representative experiments. Bottom: grouped data. In males (left), PVN BIBO3304 increased LSNA in OR and CON but not OP rats. In females (right), PVN BIBO3304 increased LSNA in all groups. solid circles and line, CON; open circles and solid line, OR; solid grey circle and dashed line, OP. ^* P < 0.05 compared to baseline. †P < 0.05 compared to OR and CON under the same conditions.

Table 3.

PVN nanoinjections of aCSF have no effects on LSNA, MAP or HR in any group of male or female rats

	CON (n = 6)		OR (n = 5)		OP (n = 6)
Males	Baseline	PVN aCSF	Baseline	PVN aCSF	Baseline	PVN aCSF
LSNA (% control)	100 ± 0	104 ± 2	100 ± 0	101 ± 2	100 ± 0	100 ± 3
HR (beats min–1)	358 ± 17	357 ± 18	416 ± 15	409 ± 14	373 ± 14	375 ± 15
MAP (mmHg)	116 ± 6	114 ± 6	116 ± 9	114 ± 14	113 ± 7	112 ± 6

	CON (n = 5)		OR (n = 5)		OP (n = 5)
Females	Baseline	PVN aCSF	Baseline	PVN aCSF	Baseline	PVN aCSF
LSNA (% control)	100 ± 0	102 ± 4	100 ± 0	99 ± 1	100 ± 0	104 ± 3
HR (beats min–1)	392 ± 32	393 ± 33	370 ± 16	372 ± 18	368 ± 24	366 ± 21
MAP (mmHg)	112 ± 8	114 ± 7	95 ± 11	96 ± 11	117 ± 2	114 ± 2

In males, the reduced tonic PVN NPY sympathoinhibition could be due to decreased NPY release in PVN or decreased responsiveness to NPY. To test the latter hypothesis, we injected NPY into the PVN of CON, OR and OP male rats. As shown in Fig. 7, NPY decreased LSNA similarly in all three groups. Therefore, in obese male rats, the reduced tonic inhibition of SNA by NPY in the PVN is not a result of reduced responsiveness of PVN presympathetic neurons to NPY but, instead, is probably the result of decreased NPY release (or inhibition of NPY neurons in the ArcN).

Figure 7.

PVN NPY nanoinjections decrease LSNA similarly in OP, OR and CON male rats

Top: representative experiments. Bottom: grouped data. PVN NPY did not significantly alter MAP (in mmHg: CON, 119 ± 3 to 119 ± 2; OR, 111 ± 6 to 114 ± 5; OP, 108 ± 8 to 108 ± 7) or HR (in beats min^–1: CON, 372 ± 21 to 331 ± 16; OR, 337 ± 15 to 338 ± 16; OP, 360 ± 0 to 357 ± 7) in any group. ^* P < 0.05, time effect; no between group differences by ANOVA (P > 0.05).

In OP females, if basal tonic NPY inhibition is normal, there is the question of why insulin fails to increase LSNA. We tested the hypothesis that this failure is partly the result of an inability of insulin to inhibit NPY neurons. As shown above in Fig. 2, i.c.v. insulin infusion increased LSNA in female OR and CON rats but not OP rats (Fig. 8); HR also increased in lean rats (CON: by 25.0 ± 6.5 beats min^–1; OR: by 29.4 ± 4.8 beats min^–1; P < 0.05) but not OP rats (by 0.9 ± 9.9 beats min^–1). As expected (Cassaglia et al. 2016), subsequent PVN BIBO3304 nanoinjections did not increase LSNA in lean CON and OR rats (because insulin had inhibited NPY inputs). However, BIBO3304 increased LSNA in OP rats (Fig. 8). Thus, in obese females, ArcN NPY neurons become resistant to the inhibitory effects of insulin. Nevertheless, in female OP rats after insulin and BIBO3304, LSNA remained lower (P < 0.05) than in CON or OR rats after insulin and BIBO3304. Because the increase in LSNA evoked by insulin is mediated largely by POMC neurons and release of α‐MSH into the PVN (Cassaglia et al. 2016), these results also suggest that insulin does not activate ArcN POMC neurons normally in OP females.

Figure 8.

Infusion of insulin i.c.v. suppresses tonic inhibition of LSNA by PVN NPY in female CON and OR rats but not female OP rats

i.c.v. insulin infusion was begun at time zero (first arrow) and continued for the remainder of the experiment; BIBO3304 was nanoinjected bilaterally into the PVN 2 h after starting the insulin infusion (second arrow). Solid circles and line, CON; open circles and solid line, OR; solid grey circle and dashed line, OP. ^* P < 0.05 compared to baseline; †P < 0.05, compared to 120 min (2 h after starting i.c.v. insulin). OP was different from OR/CON at all times except baseline (time 0).

NPY InsR expression in male and female CON, OR and OP rats

Because insulin acts solely in the ArcN to increase SNA, collectively, these data suggest that the sexually dimorphic responses to insulin in obese females vs. obese males are a result of changes in the ArcN: in OP males, inhibitory NPY inputs are suppressed; in OP females, insulin fails to inhibit ArcN NPY inputs to the PVN. Therefore, we next investigated whether obesity altered ArcN InsR (or NPY) expression in either sex. NPY expression in CON male and female rats increased from the rostral to caudal levels of the ArcN, peaking in the mid‐ArcN, and then decreased in the most caudal sections (Fig. 10 B, D, E and F). InsR expression changed in parallel, such that the NPY InsR expression level remained constant throughout the nucleus (Fig. 10 A–C). This pattern of NPY, InsR, and InsR/NPY expression was not modified by a HFD in either male or female rats (Fig. 10 A and C).

Although NPY expression in the DMH is normally low and present almost exclusively in the compact division, orexigenic states like obesity and lactation can induce DMH NPY expression within and around the compact area (Guan et al. 1998; Li et al. 1998). Moreover, DMH NPY neurons project to the PVN (Lee et al. 2013). The DMH was present in the slices used to quantify ArcN NPY/InsR; therefore, we also examined DMH NPY expression in CON/OR/OP male and female rats. We confirm that, in CON male and female rats, NPY expression was low (compared to the ArcN) and found almost exclusively in the compact area, with only a scattering of cells outside (∼10–15 positive neurons for the entire DMH); however, a similar expression level was found in OR and OP animals (data not shown). Therefore, it appears that a decrease in DMH NPY expression in OP males does not explain the reduced tonic NPY inhibition of LSNA in the PVN.

Discussion

The present study tested whether the sympathoexcitatory response to insulin is exaggerated in obese males but not obese females, partly because of sexually dimorphic changes in NPY inhibition of SNA in the PVN. Our major new findings are that, in OP compared to OR/CON male rats: (i) i.c.v. insulin increased LSNA and its baroreflex regulation dramatically more; (ii) i.c.v. insulin enhanced baroreflex control of HR; and (iii) although PVN NPY‐induced sympathoinhibition was similar in all groups, blockade of PVN NPY1R failed to increase LSNA in OP males. By contrast, in female rats: (i) i.c.v. insulin increased basal and baroreflex control of LSNA in OR/CON, but not OP, rats; (ii) the sympathoexcitatory response to blockade of PVN NPY Y1R was similar in OP, OR and CON rats; and (iii) after i.c.v. insulin infusion, blockade of PVN NPY1R increased LSNA only in OP rats. In both males and females, a HFD failed to alter NPY expression or InsR expression in NPY neurons throughout the ArcN. Collectively, these data reveal marked sex differences in the impact of obesity on the control of SNA: obesity amplifies insulin‐induced increases in LSNA in males but abolishes this SNA response in females. We identify ArcN NPY neurons, via a pathway to the PVN, as a pivotal cellular mediator of this sexual dimorphism. However, the ArcN NPY cellular mechanism underlying this sex difference does not appear to involve differential changes in NPY content or InsR expression.

Obesity amplifies the sympathoexcitatory response to insulin in male rats

In males, leptin is widely considered a major contributor to obesity‐induced elevations in SNA and AP (Hall et al. 2010; Harlan & Rahmouni, 2013). Not only are plasma leptin levels elevated, but also leptin‐induced increases in SNA are preserved or enhanced, despite muted anorexic actions of leptin, so‐called ‘selective leptin resistance’ (Prior et al. 2010; Mark, 2013). In parallel, insulin‐induced anorexia is reduced (Clegg et al. 2005; Vogt & Bruning, 2013) and, as shown in the present study, the sympathoexcitatory effects of i.c.v. insulin are greatly magnified in obese male rats. In addition, the ability of insulin to increase the gain and maximum of baroreflex control of LSNA and HR was also exaggerated in male OP rats. These results suggest that SNA could reach dangerously high levels in obese males when insulin levels are also high, as with eating, and that the central actions of insulin may contribute to obesity‐induced hypertension.

These findings may appear to conflict with previous studies in obese humans demonstrating blunted sympathoexcitatory responses to increments in plasma insulin (Vollenweider et al. 1994; Straznicky et al. 2009). As discussed previously (Vollenweider et al. 1994; Limberg et al. 2014; Brooks et al. 2015), the rate of insulin transport from plasma into the brain may explain why increases in brain (aCSF) insulin greatly increase SNA, yet, systemically, the effects of insulin are reduced. More specifically, plasma insulin was elevated in these studies of obese humans (Vollenweider et al. 1994; Straznicky et al. 2009) and the transport of insulin across the blood–brain barrier becomes saturated at plasma levels close to normal euglycaemic levels, at least in mice (Banks et al. 1997a; Banks et al. 1997b). Moreover, obesity impairs transport of insulin across the blood–brain barrier (Kaiyala et al. 2000; Banks, 2004; Kern et al. 2006; Meijer et al. 2016), thus limiting access of blood insulin to its central site of action. Alternatively, if the results of the present study obtained in rats may be extrapolated to humans, the inclusion of both sexes (Straznicky et al. 2009) may have obscured detection of a significant sympathoexcitatory effect of i.v. insulin. Yet, Lim et al. (2013) showed that acute i.c.v. injection of an insulin antagonist decreases MAP and renal SNA in obese rabbits. In humans, noradrenaline spillover was directly related to the degree of insulin resistance, and diet‐induced weight loss decreased noradrenaline spillover only in insulin resistant subjects (Straznicky et al. 2005). Our demonstration that the central sympathoexcitatory effects of insulin are markedly amplified in obese males might resolve this paradox. If the brain were sensitized, then normal or even reduced brain insulin levels could contribute to the associated net increases in basal SNA in obese male subjects.

Obesity abolishes the sympathoexcitatory response to insulin in female rats

Several studies in humans have demonstrated that obesity increases MSNA in men but not women (Lambert et al. 2007; Tank et al. 2008; Maqbool et al. 2010; Brooks et al. 2015; Straznicky et al. 2017). We therefore aimed to test whether this sex difference can be explained in part by differences in the central sympathoexcitatory responses to insulin in obese rats. As expected (Levin & Strack, 2008), we found that, in females, similar to in males, a HFD engenders a range of BW, allowing selection of OP and OR cohorts. In addition, in both sexes, the increment in BW was highly correlated to the accrued fat mass. However, to our knowledge, sex differences in the impact of diet‐induced obesity on fasting insulin and glucose levels have not been documented in SD rats. We previously reported that male OP rats become insulin resistant, based on the hyperinsulinaemic–euglycaemic clamp approach (Zhao et al. 2012), for which changes in the insulin‐sensitivity indices HOMA‐IR and QUICKI are confirmatory in the present study. On the other hand, OP females, as in humans (Brooks et al. 2015; Straznicky et al. 2017), became far less insulin resistant. Fasting plasma insulin levels were also lower in obese females compared to males. Thus, this model replicates sexually dimorphic features of obesity in humans. Yet, despite near normal systemic insulin sensitivity, i.c.v. insulin failed to increase LSNA in OP female rats, in sharp contrast to lean OR and CON rats, in which responses to i.c.v. insulin were similar to males [Figs 1 and 2 and Shi & Brooks, (2015)]. In parallel, although i.c.v. insulin enhanced baroreflex control of LSNA and HR in lean females, this action was abolished in OP females. Therefore, we conclude that obesity may not increase SNA in females as much as in males, in part because insulin levels increase less, and because the central sympathoexcitatory response to insulin is muted in obese females but exaggerated in obese males.

Role of PVN NPY in the sex difference

Why does obesity amplify LSNA responses to i.c.v. insulin in males but abolish this response in females? The mechanism is probably multifactorial but, because a reduction in tonic NPY inhibition of SNA in the PVN is required before insulin‐induced sympathoexcitatory inputs such as α‐MSH can increase SNA (Shi et al. 2015), we first focused on ArcN NPY neurons.

Obese males

In males, obesity eliminated tonic PVN NPY inhibition of SNA. The consequence is that excitatory inputs into the PVN, such as by α‐MSH, would prevail. We next began to explore the mechanisms. Initially, as reviewed by Beck (2006), within the first few weeks after initiating a HFD in males, ArcN NPY levels remain relatively stable but then decrease weeks later as a HFD continues. In the present study, male rats were studied relatively early in obesity development (4–6 weeks of a HFD), yet, despite elevated insulin levels and excess fat mass (Fig. 1), ArcN and DMH NPY expression was unaltered. Second, previous reports showed that obesity or leptin treatment decreases PVN NPY1R expression in male but not female mice (Zammaretti et al. 2007; Mele et al. 2016). Therefore, we tested whether NPY inhibition of PVN presympathetic neurons is eliminated in OP male rats because of a decrease in response to NPY in the PVN. However, PVN NPY evoked similar decreases in LSNA in male CON, OR and OP rats. Collectively, these results suggest that a reduction in ArcN NPY expression or PVN NPY receptor number/signalling does not mediate the loss of tonic PVN NPY inhibition in males early in obesity development. Instead, obesity may act upstream, in the ArcN, to decrease basal release of NPY that tonically inhibits presympathetic neurons in the PVN.

We therefore next considered mechanisms that could decrease ArcN NPY neuronal activity. We tested the hypothesis that, in males, obesity increases NPY neuronal InsR expression; because insulin inhibits NPY neurons, increased InsR would suppress NPY neuronal activity and tonic NPY inhibition of PVN pre‐sympathetic neurons would decrease. However, it was previously shown in males that obesity failed to modify InsR expression in blocks of ArcN tissue (Clegg et al. 2005). Moreover, rather than increased insulin signalling (and NPY neuronal inhibition), decreased signalling in hypothalamic blocks has been observed (Carvalheira et al. 2003; De Souza et al. 2005). These results indicate that the ArcN subset of NPY neurons that influence SNA is altered differentially by a HFD. Therefore, in view of the diversity of the ArcN, to test this hypothesis, we performed a detailed quantification of NPY neuronal InsR expression throughout the ArcN in male and female CON, OR and OP rats. In agreement with previous stuides that used classical in situ hybridization (Marks et al. 1992; Urban et al. 1993; Baker & Herkenham, 1995), NPY expression increased from the rostral to caudal segments of the ArcN, before decreasing. InsR expression changed in parallel, such that the InsR/NPY ratio remained relatively constant throughout the ArcN. Of note, in males, InsR expression in NPY neurons was unchanged by a HFD. Thus, increased InsR expression in ArcN NPY neurons cannot explain why, with obesity in males, tonic NPY sympathoinhibition is lost.

If InsR number is unaltered, then chronically increased insulin levels (Fig. 1) or increases in insulin signalling in those ArcN NPY neurons that influence SNA, could suppress tonic NPY sympathoinhibition. Alternatively, obesity may inhibit NPY pre‐sympathetic neurons via another mechanism. One possible mechanism may involve leptin, which also inhibits the NPY neurons that tonically suppress SNA (Shi et al. 2015). However, as with insulin, leptin increased SNA more in obese rabbits, even though leptin‐induced c‐fos expression in the ArcN was reduced (Lim et al. 2013). These results support the concept that ArcN neurons controlling feeding respond weakly to further elevations in leptin, although the neurons that control SNA respond more. Thus, in obese males, higher leptin levels or actions could suppress the activity of ArcN NPY neurons that inhibit SNA. Alternatively, obesity appears to favour the hypertensive (AngII) over the antihypertensive [Ang‐(1–7)] arm of the renin–angiotensin system in males, with the opposite occurring in females (Gupte et al. 2012). ArcN AngII increases BP and SNA (Arakawa et al. 2011), whereas ArcN Ang‐(1–7) can decrease BP and SNA in females (Z. Shi and V. L. Brooks, unpublished observations). ArcN AngII type 1 receptors are expressed largely in NPY neurons (Claflin et al. 2017) and ArcN AngII decreases NPY inputs to the PVN (Shi & Brooks, 2018). Therefore, in males, high AngII could suppress the activity of ArcN NPY neurons that influence SNA, although this action may be counteracted by Ang‐(1–7) in females. Further work is required to test these hypotheses.

Obese females

By contrast to males, in OP females, tonic NPY inhibition of PVN presympathetic neurons was maintained. Moreover, this tonic inhibition persisted after 2 h of i.c.v. insulin infusion in female OP but not OR or CON rats. This result suggests that, in obese females, in addition to a failure of insulin to invoke enhanced sympathoexcitation like obese males, another mechanism renders ArcN NPY neurons unresponsive to insulin.

What is the mechanism that is responsible for this? In obese females, as in males, neither ArcN NPY, nor NPY cellular InsR levels changed significantly. Therefore, lower NPY InsR expression does not appear to explain the inability of insulin to reduce tonic NPY suppression of LSNA in females. Instead, downstream signalling pathways may be inhibited. A HFD in SD rats suppresses cycling and lowers plasma oestrogen levels (Balasubramanian et al. 2012) (Z. Shi and V. L. Brooks, unpublished observations). Loss of oestrogen uncouples the insulin receptor from downstream signalling in ArcN POMC neurons in obese mice (Qiu et al. 2018). Further work is required to determine whether a similar phenomenon occurs in NPY neurons.

Summary

The present study reveals a dramatic sex difference in central sympathoexcitatory responses to insulin in obese rats: the LSNA response was amplified in males but abolished in females. Moreover, we identify sex‐dependent changes in tonic NPY inhibition of SNA in the PVN of obese animals as a key contributor to this difference. Normally (Fig. 11, top), insulin increases SNA by acting in the ArcN to inhibit NPY neurons and excite POMC neurons, both of which project to the PVN. In obese males, basal PVN NPY inhibition was lost (Fig. 11, bottom left). Because the slowly developing rise in SNA in response to insulin is mediated largely by PVN α‐MSH (Cassaglia et al. 2016), we propose that, in obese males, insulin activates POMC neurons more, leading to a greatly magnified action of α‐MSH in PVN, unfettered by the inhibitory effects of NPY. In addition, downstream excitation in the RVLM may be enhanced. In OP females (Fig. 11, bottom right), tonic NPY suppression of SNA was maintained but was not inhibitable by insulin, as in lean animals. Thus, excitatory inputs to the PVN, such as α‐MSH, would be blocked. After i.c.v. insulin and PVN NPY1R blockade, SNA was less in OP females compared to OR and CON rats. Therefore, we also speculate that insulin activates POMC neurons less in OP females. Further studies are required to test whether POMC neuronal responsiveness to insulin is also altered in a sexually dimorphic manner by obesity.

Figure 11.

Proposed mechanisms by which obesity amplifies the sympathoexcitatory response to insulin in males but abolishes the response in females

Top: previous and present studies have shown that, in lean males and females, insulin increases SNA by binding to receptors in the ArcN (Cassaglia et al. 2011; Luckett et al. 2013) to suppress tonic sympathoinhibitory NPY inputs (Cassaglia et al. 2016) (Fig. 8) and activate excitatory α‐MSH and glutamatergic inputs (Ward et al. 2011; Stocker & Gordon, 2015) to the PVN. Then, activated PVN glutamatergic neurons project downstream to stimulate presympathetic neurons in the rostral ventrolateral medulla (RVLM) (Bardgett et al. 2010). Bottom: in obese males (left), tonic inhibition of PVN presympathetic neurons by NPY is lost (pink neuron). Although not yet demonstrated explicitly, indirect evidence (Cassaglia et al. 2016) suggests that, in obese males, insulin activation of ArcN POMC neurons may be amplified (thicker arrow) (da Silva et al. 2014; Rahmouni, 2014). In addition, the excitation of the RVLM by glutamate may be enhanced (Huber & Schreihofer, 2016). By contrast, in obese females (right), tonic NPY inhibition is maintained (Fig. 5) and ArcN NPY neurons become resistant (X) to the inhibitory effects of insulin (Fig. 8). Our data also indirectly suggest that obesity in females impairs responses of POMC neurons to insulin (X). These sexually dimorphic changes do not appear to involve alterations in the expression of the insulin receptor in ArcN NPY neurons (Fig. 10). [Color figure can be viewed at wileyonlinelibrary.com]

Thus, the stage is set to investigate the mechanisms that underlie this sexually dimorphic response in the ArcN at a more cellular/molecular level. Potential players include gonadal hormones, leptin, or differing responses or actions of the renin–angiotensin system. This information is required to best inform treatment of hypertension in obese men (and probably postmenopausal women) vs. premenopausal women.

Additional information

Competing interests

The authors declare that they have no competing interests.

Author contributions

This work was conducted in the Department of Physiology and Pharmacology, Oregon Health & Science University, Portland, OR 97239. VLB was responsible for the conception/design of the work. ZS, PAC, NEP were responsible for the experimental design. ZS, PAC, NEP were responsible for the acquisition of data. ZS, PAC, NEP and VLP were responsible for the analysis and interpretation of data. ZS, PAC, NEP were responsible for manuscript/figure development. VLB was responsible for the production of manuscript. All authors have approved the final version of the manuscript submitted for publication and agree to be accountable for its content.

Funding

This work was supported in part by funding from the NIH (HL088552 and HL128181) and the AHA (09GRNT2060630, 12GRNT11550018 and 15POST23040042). Use of the Advanced Light Microscopic Core was funded in part by P30 NS061800 (PI, S. Aicher).

Biography

Zhigang Shi obtained his MD and MS degrees from Shanxi Medical University, Taiyuan, China.  His postdoctoral work focused on the generation and cardiovascular impacts of obesity. As a first step towards understanding the mechanisms that contribute to obesity‐induced hypertension, he identified the neuronal pathways by which leptin and insulin increase sympathetic activity. More recently, he has begun to investigate the profound sex differences in the sympathoexcitatory actions of these metabolic peptides. This work has culminated in the current paper, which begins to unravel the mechanisms by which obesity increases sympathetic nerve activity in men but not premenopausal women.