“Intersection of Metabolic and Blood Pressure Regulatory Circuits in the CNS: Recent Advances in Hypertension”

Abstract

Obesity represents the single greatest ongoing roadblock to improving cardiovascular health. Prolonged obesity is associated with fundamental changes in the integrative control of energy balance, including the development of selective leptin resistance, which is thought to contribute to obesity-associated hypertension, and adaptation of resting metabolic rate (RMR) when excess weight is reduced. Leptin and the melanocortin system within the hypothalamus contribute to the control of both energy balance and blood pressure. While the development of drugs to stimulate RMR and thereby reverse obesity through activation of the melanocortin system has been pursued, most of the resulting compounds simultaneously cause hypertension. Evidence supports the concept that although feeding behaviors, RMR, and blood pressure are controlled through mechanisms that utilize similar molecular mediators, these mechanisms exist in anatomically dissociable networks. New evidence supports a major change in molecular signaling within Agouti-related peptide (AgRP) neurons of the arcuate nucleus of the hypothalamus during prolonged obesity, and the existence of multiple distinct subtypes of AgRP neurons that individually contribute to control of feeding, RMR, or BP. Finally, ongoing work by our lab and others support a unique role for angiotensin II signaling through its type 1 receptor (AT₁) within one specific subtype of AgRP neuron for the control of RMR. We propose that understanding the unique biology of the AT₁-expressing, RMR-controlling subtype of AgRP neurons will help to resolve the selective dysfunctions in RMR control that develop during prolonged obesity, and potentially point toward novel druggable anti-obesity targets that will not simultaneously cause hypertension.

Graphical Abstract

Introduction

Obesity is caused by a complex interaction of genes, behavior, and environmental factors. Increased body weight is associated with increased blood pressure (BP), while weight loss is associated with reduced BP.^1–3 Obesity is an independent risk factor for the development of hypertension, and the Framingham Heart Study suggests that roughly 70% of primary hypertension is linked to obesity.²

Although the American Heart Association (AHA) made great strides toward its ambitious 2020 Impact Goal “to improve the cardiovascular health of all Americans by 20% while reducing deaths from CVDs and stroke by 20% by 2020”, we failed to achieve this goal primarily due to metabolic diseases, including obesity. In the AHA 2030 Impact Goal, Angell et al.⁴ state:

“Over the past decade (NHANES 2007–2016), there has been significant progress in the United States toward the 2020 Impact Goal, with less smoking in adults, increased physical activity in adults, improved dietary habits in both adults and youth, lower total cholesterol in adults and youth, and lower blood glucose in adults. However, these areas of progress have been offset by adverse forces: increases in high blood pressure and body mass index among adults and youth, higher blood glucose in youth, and lower physical activity in youth. The continued worsening of overweight and obesity rates among adults and youth throughout the decade is especially concerning and largely accounts for worsening blood pressure in adults and youth and fasting blood glucose among youth.”

Thus, it is increasingly clear that cardiovascular disease and metabolic disease are inextricably intertwined, and that future progress in hypertension and cardiovascular health is dependent upon improvements in our understanding and treatment of metabolic disease, especially obesity.

Obesity: Not ‘just’ overeating?

In 1985, the NIH Consensus Development Conference recognized obesity as a disease with high morbidity and mortality.⁵ Unfortunately, it took almost 30 years for obesity to be formally appreciated by major medical associations as a ‘disease’ and a global public health burden rather than ‘simply’ a consequence of overeating.⁶ It is well-recognized that obesogenic environmental factors abound in Western societies,⁷ yet obesity is the result of long-term caloric imbalance which may be caused by overfeeding, reduced energy expenditure, or a combination of both.

Energy balance is achieved through a complex interplay among systems controlling input versus output. Ingestive behaviors and the efficiency of the digestive tract to extract calories from consumed food both modulate caloric input into the system. In contrast, resting metabolic rate (RMR) and activity-based energy expenditure both contribute to caloric output. Dysfunctions in any of these systems can mediate the development and/or maintenance of obesity and targeting any of these systems may be utilized to promote weight loss, at least in the short-term.

One of the critical challenges to correcting obesity is that, after obesity is established, the body appears to defend this increased mass against targeted efforts to lose weight. Indeed, individuals enrolled in structured behavioral interventions for weight loss often regain roughly 80% of lost weight within 5 years.⁸ Similarly, although reductions in digestive efficiency and subsequent weight loss can be achieved with orlistat, weight relapse is common within a few years.⁹

Increasing evidence supports the concept that weight loss triggers an adaptative but dysfunctional RMR response, such that RMR is suppressed relative to body mass and composition (TABLE 1), which hinders long-term reduced weight maintenance.^10–12 Scattered evidence also supports the concept that this adaptation of RMR may predict subsequent weight regain.^{12, 13} Stimulating RMR via mitochondrial ionophores works to combat obesity in humans and animal models, however, because such compounds are highly toxic their use in humans is strongly prohibited.^14–19 Such compounds, however, provide critical proof-of-concept data regarding the potential utility of targeting RMR to treat obesity, while highlighting the ongoing need for a better understanding of the biology of RMR control (during both health and disease) to support the development of safe drugs to stimulate RMR.

Table 1:

Clinical studies reporting on metabolic adaptation after weight loss

Author	Metabolic adaptation, kcal/day (mean±SD)	n	Baseline BMI, kg/m2 (mean ± SD or range)	Sex	Max weight loss	Follow-up (months)	Daily calorie intake or calorie restriction	Duration of dietary intervention	Procedure
LIFESTYLE CHANGES
Short term studies (≤ 6 months)
Leibel88	-137±305	9	n/r	n/r	10%	2.5?	800 kcal/day	until weight goal achieved	-
Leibel88	-79±294	10	n/r	n/r	20%	n/r	800 kcal/day	until weight goal achieved	-
Doucet89	M ~ -112 F ~ -143	35	34.2±0.6 36.8±0.9	15 M 20 F	10% 9%	4.5	-700 kcal/day	15 weeks	-
Knuth90	-419±13	35	46.7±9.5	7M 9F	35%	7	>70% caloric needs	28 weeks	-
Rosenbaum91	NS	10	35.3±3.8	5 M 5 F	10%	3.75	800 kcal/day	until weight goal achieved	-
Bosy-Westphal92	-55±165	45	35.9±4.2	F	9%	3	~900 kcal/day	13 weeks	-
Müller93	-80±97	32	20.7–29.3	M	7.5%	0.75	~1400 kcal/day#	3 weeks	-
Nymo10	-111±165	31	36.7±4.5	18 M 13 F	10%	2	~600 kcal/day	8 weeks	-
Martins94	−49±128	71	34.6±3.4	33 M 38 F	13%	3.25	1000 kcal/day	8 weeks	-
Long term studies (> 6 months)
Rosenbaum95	-150±50	7	n/a	2 M 5 F	≥10%	>12 (R)	n/r	n/r	-
Fothergill12	-500±150	14	34.9±10.3	M, F	39%*	72	>70% caloric needs	28 weeks	-
Ostendorf96	NS	34	33.0±4.6	8 M 26 F	29%	72 (R)	n/r	n/r	-
Martins94	NS	45	34.6±3.4		14%**	12	n/r	n/r	-
Thom97	-150±162	15	39.4±4.3	F	13.5%	6	~850 kcal/day	12–20 weeks	-
BARIATRIC SURGERY
Short term studies (≤ 6 months)
Carrasco98	-85±124	38	44.4±4.8	4 M 34 F	29%	6	n/r	n/r	RYGB
Browning99	NS	8	47.1±6.8	2 M 6 F	28%	6	n/r	n/r	RYGB
Knuth90	-201±182	13	47±7.6	4 M 9 F	25%	6	~1000 kcal/day	24 weeks	RYGB
Golzarand100, 101	-230±234 -308±210	22 21	42.6±5.4	1 M, 21 F 21 F	23%	6	n/r	n/r	RYGB SG
Long term studies (> 6 months)
Das102	NS	30	50.1±9.3	4 M 26 F	38%	14	n/r	n/r	RYBG
Knuth90	NS	13	47±7.6	4 M 9 F	30%	12	~1000 kcal/day	52 weeks	RYGB
Tam103	-134±28 -255±122	5 9	45.1±2.5 49.9±3.4	5 F 2 M, 7 F	35%	12	800	8 weeks	RYGB SG

Metabolic adaptation calculated as measured RMR – predicted RMR; RYGB: Roux en Y gastric bypass; SG: sleeve gastrectomy; (R): retrospective study; n/r: not reported; NS: not significant;

^#50% energy requirements;

^*regained 70% of weight lost by end of follow-up (final weight loss 12%);

^**regained 30% of weight lost by end of follow-up (final weight loss 10%)

Leptin & Melanocortin Systems

Leptin serves as an afferent signal to communicate fat stores to the brain, thereby influencing the integrative control of energy stores.^20–22 Leptin modulates the activity of several types of cells, including Proopiomelanocortin (POMC) and Agouti-related peptide (AgRP) neurons, to influence feeding behavior, RMR, and BP.^23–27 POMC neurons stimulate the melanocortin receptor 4 receptor (MC₄R) via α-melanocyte stimulating hormone (αMSH), and also release γ-aminobutyric acid (GABA), adrenocorticotropic hormone, β-endorphin, and the cocaine and amphetamine regulated transcript. AgRP neurons utilize a combination of AgRP, Neuropeptide Y (NPY) and GABA to influence postsynaptic neurons. Importantly, AgRP acts as an inverse agonist at MC₄R.^{28, 29} Furthermore, MC₄R exhibits an unusually strong constitutive intrinsic (ie, ligand-independent) activity, highlighting the complex biochemistry of this signaling pathway and the importance of changes in AgRP neurotransmission even in the absence of αMSH signaling.³⁰ The MC₄R is expressed in multiple hypothalamic and extra-hypothalamic regions that are variably implicated in cardiovascular and metabolic control.^{30, 31} Activation of MC₄R by αMSH increases cardiovascular sympathetic activity, BP and heart rate in mice, whereas pharmacological inhibition of MC₄R or Mc4r deletion cause increased body weight, without affecting BP.³²

Because MC₄R is implicated in energy balance and BP control, it represents an obvious target for manipulating cardiometabolic functions.^{33, 34} Unfortunately, because MC₄R is involved in multiple pre-autonomic networks that contribute to cardiovascular versus metabolic control, such stimulation of sympathetic drive is typically non-selective, leading to simultaneous stimulation of both RMR and BP. For example, first-generation synthetic MC₄R agonists were abandoned as potential anti-obesity treatments because although they caused weight loss in rodents, monkeys, and humans, they also caused significant increases in BP.³⁵ Next-generation MC₄R agonists, represented by Setmelanotide,³⁶ reduce food intake and body weight in several species (rodents, dogs, and non-human primates), without affecting BP.^{37, 38} Setmelanotide increases resting energy expenditure in C57BL/6 mice supplied a high fat diet (HFD),³⁸ and decreases food intake and increases energy expenditure without causing hypertension in Rhesus monkeys fed obesogenic diets.³⁷ In a phase 1 human clinical trial (NCT01867437), it increased RMR (~111 kcal/d), with no effect on BP or heart rate.³⁹ Similarly, in subjects carrying MC4R mutations, Setmelanotide induced weight loss (~0.6 kg/week) without any effects on cardiovascular parameters (NCT02431442).³⁶ It also reduced hyperphagia and body weight in obese individuals with POMC and/or leptin receptor (LEPR) mutations without hemodynamic effects (NCT02507492).^{40, 41} As such, Setmelanotide received FDA approval for chronic weight management in individuals with selected mutations in the POMC, PCSK1, or LEPR genes on November 27, 2020. Future work to explore the utility of this compound for treating obesity that results from other mutations and causes (ie, beyond dysfunctions primarily attributed to the POMC neuron) is clearly warranted.

The mechanisms by which Setmelanotide induces weight loss without influencing BP are unclear. It may act as a biased ligand for MC₄R, differentially stimulating Gαs/cAMP versus Gαq second messengers.^{36, 41} Additionally, it may exhibit selective distribution/compartmentalization, bringing the compound into contact with cardiovascular versus metabolic control regions at different effective concentrations. Based on in silico docking and crystallization studies, it may alter binding kinetics of endogenous ligands.⁴² Regardless of mechanism, this differential action provides exciting proof-of-concept support for the idea that metabolic and cardiovascular control networks are dissociable, even if they utilize the same molecular mediators.

Selective Leptin Resistance

Although obesity is associated with increased adiposity and circulating leptin, metabolic responses to leptin are attenuated while cardiovascular responses to leptin are maintained. This phenomenon, termed Selective Leptin Resistance (SLR), is proposed to contribute to obesity-associated hypertension.⁴³ Studies of C57BL/6J mice fed HFD for 10 weeks to cause diet-induced obesity (DIO) demonstrate that SLR involves changes in hypothalamic leptin signaling. Indeed, cardiovascular (ie, renal) sympathetic responses to acute ICV leptin injection remain intact in DIO mice, while metabolic (ie, brown adipose) sympathetic responses are blunted.⁴⁴

To explore changes in hypothalamic leptin signaling during obesity, we performed single nucleus RNA-sequencing (snRNAseq) on arcuate nucleus of the hypothalamus (ARC) from mice after 10 weeks of HFD feeding.⁴⁵ Mice displayed expected body and fat mass gains, increased plasma leptin concentrations, and changes in food intake and energy expenditure metrics. Interestingly, the gene expression signature for leptin signaling was only significantly altered in 5 of the 23 unique clusters of cell types, including astrocytes, oligodendrocytes, and AgRP neurons – but notably and iconoclastically, not POMC neurons. In addition, the canonical signature for leptin signaling in AgRP neurons was awarded a negative Z-score, indicating that the canonical leptin-sensitive genes responded opposite to leptin’s typical effects. Thus, molecular evidence of SLR is observable within the ARC, and SLR is most obviously occurring within AgRP neurons.

Multiple groups have demonstrated in various rodent models that, while short-term HFD feeding suppresses expression of Agrp within the ARC, ^46–54 prolonged HFD leads to attenuation of this effect or even stimulation of Agrp expression.^{48, 55–57} Similarly, in our study, Agrp expression was significantly increased within AgRP neurons after DIO despite large increases in circulating leptin.⁴⁵ Further, within AgRP neurons we identified derangements of multiple signaling pathways implicated in transcriptional control of Agrp, including CREB and MAPK/ERK cascades.

Although it is classically considered a ‘hunger’ hormone, evidence suggests that AgRP is also important for the control of energy expenditure. Egan et al. previously demonstrated that genetic deletion of the leptin receptor from cells expressing Agrp resulted in increased body mass without effects on feeding behaviors, implying a suppression of energy expenditure.⁵⁸ Similarly, Makimura et al. demonstrated that RNAi-mediated knockdown of Agrp in the mouse hypothalamus increased heat production,⁵⁹ Small et al. reported that ICV injection of recombinant AgRP protein suppressed oxygen consumption in Wistar rats,⁶⁰ and Krashes et al. showed reduced oxygen consumption in mice after DREADD-mediated stimulation of AgRP neurons.⁶¹

Collectively, these findings support the concepts that SLR involves fundamental changes in the behavior of AgRP neurons, and that AgRP neurons contribute to the integrative control of energy homeostasis. Thus, understanding the unique biology of these cells in both health and disease may help to explain the pathogenesis of SLR and weight loss-associated adaptation of RMR.

The ARC renin-angiotensin system

Local paracrine versions of the RAS have been identified in the brain and other tissues, which operate independently of the circulating RAS .^62–65 We and others have implicated the brain RAS in the control of energy homeostasis and ingestive behaviors, including roles of the RAS within the ARC, the paraventricular nucleus, the nucleus tractus solitarius, the lateral parabrachial nucleus, and other regions.^66–70 Pharmacological activation of the brain RAS by angiotensin II (ANG) infusion^{71, 72} or deoxycorticosterone acetate (DOCA)-salt treatment,⁷³ or transgenic activation of the RAS within the brain⁷⁴ all stimulate RMR with mechanisms dependent upon angiotensin II type 1 receptors (AT₁). Pharmacological blockade of the brain RAS or genetic disruption of AT_1A (Agtr1a) receptors attenuate thermogenic brown adipose sympathetic nerve activity (BAT SNA) responses to leptin.⁷⁵ We demonstrated that leptin-mediated activation of BAT SNA and heat production responses to an array of stimuli are all dependent upon AT_1A specifically on AgRP neurons.^{31, 76, 77} Interestingly, however, disruption of the Agtr1a gene in AgRP neurons had no major effect on feeding behaviors or BP control,⁷⁶ whereas others have demonstrated that deletion of Agtr1a in other brain regions does modulate feeding behavior.⁷⁸ This observation suggests the existence of multiple distinct subtypes of AgRP neurons within the ARC; some of which contribute to feeding, some to BP control, and others to RMR control. Indeed, in silico reanalysis of single-cell RNA sequencing datasets and RNAscope-based in situ hybridization experiments support the concept that Agtr1a is only expressed in a specific subset of AgRP neurons within the ARC.^{31, 76, 77} These findings lead us to hypothesize the distinct involvement of Agtr1a-expressing AgRP neurons in the control of RMR, but not feeding or BP (FIGURE 1).

Figure 1. Working model: labelled-line encoding of feeding, BP and RMR control in the ARC.

Evidence supports the working hypothesis that a specific subset of AgRP neurons express the AT_1A receptor (Agtr1a), and that these receptors simultaneously (i) identify the subset of AgRP neurons involved in RMR control, and (ii) serve a critical molecular role in the integrative control of RMR. We hypothesize that molecular dysfunctions specifically within the Agtr1a-expressing subset of AgRP neurons contributes to the development of SLR and weight loss-associated adaptations in RMR control. Pomc, proopiomelanocortin; αMSH, α-melanocyte stimulating hormone; Mc4r, melanocortin type 4 receptor; Agrp, Agouti-related peptide; ANG, angiotensin II; BAT SNA, brown adipose tissue sympathetic nerve activity; RMR, resting metabolic rate.

Ongoing questions

Together, these findings prompt many ongoing questions:

1. How do changes in the local ARC RAS contribute to the pathogenesis of SLR and obesity-associated adaptations in RMR control?

2. How do anatomically distinct neural pathways projecting from the ARC differentially contribute to feeding, RMR, and BP control, and how do these pathways functionally or structurally change with prolonged obesity? How can these pathways be targeted individually for therapeutic applications?

3. What unique molecular pathways exist and differentially contribute to cellular functions in each of the dissociable neural pathways controlling feeding, RMR, and BP? What druggable targets can be identified to selectively stimulate each individual pathway?

4. It is known that there are sex differences in the biology of the ARC and AgRP function.^{58, 79–84} How do sex differences in AgRP function contribute to sex differences in RMR and BP control?

5. Ghrelin, glucagon-like peptide-1, cholecystokinin, and other circulating peptides also act at the ARC in parallel or in opposition to leptin to control energy balance, and are also implicated in BP control.^{85, 86} How do alterations in ARC ANG signaling impact these other systems?

6. Do these results translate to human physiology? Circulating AgRP levels appear to correlate with hypothalamic Agrp expression, which may identify AgRP as a useful biomarker for cardiometabolic dysfunctions.⁸⁷ Is circulating AgRP a valid marker of hypothalamic melanocortin activity in humans and does it correlate with RMR control? Can it be used to identify subjects at risk of metabolic adaptation?

Conclusions

Obesity represents the primary hurdle for continued improvement in cardiovascular health. The local melanocortin and renin-angiotensin systems within the hypothalamus contribute to both cardiovascular and metabolic control, and ongoing work to dissect the anatomical and molecular mechanisms by which these systems contribute to normal and pathological control of energy homeostasis promises to identify novel therapeutic targets for both obesity and cardiovascular disease. These findings also highlight the need for greater integration of cardiovascular and metabolic endpoints in preclinical research, as the development of next-generation therapeutics for obesity versus hypertension in isolation is often derailed by failure to appreciate the simultaneous ‘off target’ consequences of manipulations of one system upon the other.