Autonomic-Immune-Vascular Interaction: An Emerging Concept for Neurogenic Hypertension

INTRODUCTION

Hypertension is the single most important risk factor for cardiovascular disease and thus remains a global public health challenge. Significant progress has been made during the last several decades in the treatment of hypertension through the use of inhibitors of the renin-angiotensin system (RAS), with either angiotensin converting enzyme inhibitors (ACEIs) or angiotensin receptor type 1 (AT1R) blockers (ARB), diuretics, α-adrenoreceptor antagonists, and calcium channel blockers. Despite these advances, it has been extremely difficult to manage hypertension in ~40% of hypertensive patients¹. A majority of these unresponsive patients exhibit increased sympathetic drive and display neurogenic components^2,3,4. The role of elevated sympathetic nervous system in drug-resistant hypertension is corroborated by the recent successes in its treatment by renal sympathetic denervation in humans^5,6,7. These observations have led to the realization that elucidation and understanding of cellular, molecular, and physiological mechanisms involved in the development and establishment of neurogenic hypertension is pivotal in advancing new therapeutic strategies which would be effective in a majority of patients. The objective of this article is to summarize recent advances in this field, underscoring the importance of coordinated neural-peripheral signaling in neural control of blood pressure (BP), contribution of inflammatory molecules, and to introduce novel and emerging concepts in the pathogenesis of neurogenic hypertension. It is advised to consult other reviewers for details on fundamental studies related to neural pathways, renin-angiotensin system (RAS), and involvement of other hormonal systems in neurogenic hypertension^{8,9,10,11,12,13,14}.

Neurogenic hypertension

An alteration in neural cardiovascular (vagal/sympathetic) control mechanism is the principal characteristic of neurogenic hypertension. Studies demonstrating a significant enhancement of the vasomotor and cardiac sympathetic drive, as well as a reduction in the parasympathetic drive in borderline hypertensive patients, support this premise^15,16,17. In addition, hypertensive patients show a greater plasma norepinephrine (NE) level, and greater NE spillover in both the peripheral circulation and blood draining from the brain as well as raised sympathetic postganglionic activity targeting the skeletal muscle vascular bed^18,19,20,21. Animal models corroborate the human data. Recent evidence shows that even the monogenic forms of hypertension, such as the apparent mineralocorticoid excess, which greatly depends on salt absorptive mechanisms of the kidney, have also been attributed to an elevated sympathetic drive indicating a very close association between renal dysfunction and sympathetic overdrive^22,23. Equally, the Goldblatt model partially depends on signaling within the rostral ventrolateral medulla (RVLM) for the maintenance of hypertension²⁴.

Evidence from the spontaneously hypertensive rat (SHR), a well-established animal model for human hypertension, supports the idea that sympathetic overdrive precedes hypertension²⁵. Therefore, molecular and neuronal changes at the level of the brainstem and hypothalamus may contribute to neurogenic hypertension. Our studies have demonstrated that elevated RAS in the brainstem and hypothalamus of juvenile SHR affects the regulation of arterial pressure and baroreceptor reflex gain in the adult SHR^26,27,28,29. Moreover, enhanced central respiratory-sympathetic coupling in rat models, of which are already present at a pre-hypertensive age, contributes to raise vasomotor tone^25,30,31,32. Thus, sympathetic activation appears to not only initiate hypertension, but also maintains it. Whether elevated sympathetic activity is causal to the initiation of hypertension remains equivocal, but it certainly occurs early in the condition.

Inflammation, the blood brain barrier and hypertension

Overwhelming evidence, both clinical and experimental, demonstrates that hypertension is associated with circulating levels of inflammation markers. In particular, there is strong support of the concept that the vascular inflammatory process plays an important role in the pathophysiology of hypertension (Figure 1). Cross-sectional and prospective studies have demonstrated that circulating levels of inflammatory molecules such as c-reactive protein (CRP), tumor necrosis factor alpha (TNF-α), interleukin-6 (IL-6), monocyte chemoattractant protein-1 (MCP-1), and adhesion molecules such as P-selectin and intercellular adhesion molecule-1 (ICAM-1), are increased in patients with primary hypertension^{33,34,35,36,37}. Women's Health Study indicated that CRP levels predicted the development of hypertension, an association that was independent of BP. Finally, inhibitors of the RAS, which are effective anti-hypertensive in some patients, are shown to decrease CRP, IL-6, TNF- α levels, and inflammatory status^{33,34,35,36,37}. Animal studies have supported these clinical association studies and have been instrumental in furthering our understanding of the involvement of inflammatory mechanisms in vascular dysfunction and the hypertensive condition. For example, total leukocyte counts in young, mature, and old SHR were 50-100% above the controls with the number of monocytes, activated monocytes, and the lymphocyte counts significantly elevated³⁸. Leukocyte-endothelial interactions were also reported to be altered in the SHR³⁸. Moreover, the elevated BP, above normal leukocyte counts, and elevated number of activated neutrophils in SHR, were all suppressed after adrenalectomy, indicating a relationship between altered leukocyte adhesiveness and adrenal corticosteroids in SHR³⁸. Blood transfusions from SHR to Wistar-Kyoto (WKY) rats increased BP due to the more sticky nature of the SHR leukocytes³⁹. The ‘stickier’ nature of leukocytes has been attributed to the elevated glucocorticoid levels in the SHR, which induces leukocyte pseudopod projections and results in slower cell passage through capillaries and elevated hemodynamic resistance in the SHR³⁹. Furthermore, elevated Ang II promotes leukocyte-endothelial interaction, thereby contributing to vascular inflammation⁴⁰. T cells play an important role in Ang II infusion-, DOCA-salt-induced hypertension, and associated vascular and renal dysfunctions^41,42,43. T cell modulating agent, mycophenolate mofetil, prevents hypertension in animal models⁴³. The SHR exhibits increased levels of activated monocytes whose levels are significantly decreased by reducing BP³⁸. Vascular inflammation in this rat model of hypertension was increased along with increases in expression of inflammatory cytokines interleukin-1β (IL-1β), IL-6, and TNF-α. AT1 receptor blocker, Candesartan, produced anti-hypertensive effects and decreases in inflammatory cytokines^34,44. Similarly, an association between inflammation and hypertension has been demonstrated in aldosterone models. For example, circulating levels of inflammatory markers are increased by aldosterone infusion in humans and patients with primary hyperaldosteronism, which is associated with vascular dysfunction, cardiac fibrosis, and increased risk of myocardial infarction and stroke³³. Chronic treatment of hypertensive rats with mineralocorticord receptor antagonist, which attenuates BP, decreases circulating levels of certain inflammatory cytokines, and renal and cardiac expression of inflammatory mediators such as IL-1, IL-6 and nuclear factor kappa beta (NFκβ)³³.

Figure 1. Proposed hypothesis of a dysfunctional autonomic-immune-vascular mechanism in neurogenic hypertension.

A coordinated peripheral-brain communication is critical in the development and establishment of neurogenic hypertension. AT1 receptors on SFO/OVLT and other CVOs are activated by increased plasma Ang II and an integrated signal is transmitted to the PVN. This results in the activation of microglia and increased generation of ROS and cytokines and decrease in NO. This leads to increased neuronal activity by regulation of ion channel function directly or indirectly. Ang II generated within the PVN can also activate neuronal AT1 receptors to compliment changes in neuronal activity. An imbalance in sympathetic/parasympathetic activities can directly influence bone marrow activity resulting in increase in pro-inflammatory cells (PICs, increase inflammation) and decrease in EPCs (decrease in vascular repair). All these events would lead to high BP and associated CV pathophysiology. These neural-vascular signals are further perpetuated by AT1 receptor-mediated changes in cerebral blood vessels, interrupting the integrity of BBB and allowing inflammatory cells to enter the brain parenchyma, contributing to microglial activation and inflammation in the PVN. A similar AT1-dependent mechanism may be present in other cardioregulatory brain regions such as the NTS and the RVLM.

The association between inflammation and inflammatory cytokines and chemokines and hypertension is evident from the above discussion. However, the concept of the involvement of the brain in inflammatory mechanisms is still evolving. For example, inflammatory cytokines such as TNF-α and IL-1β induce cyclooxygenase-2 (COX-2) activity in perivascular macrophages of the blood brain barrier (BBB) and generate prostaglandin E2, which enters the brain and stimulates paraventricular nucleus (PVN) neurons regulating adrenocorticotropic hormone (ACTH) release and sympathetic drive^45,46. Additionally, expression of junctional adhesion molecule-1 (JAM-1) is significantly increased in the nucleus of the solitary tract (NTS) of SHR relative to WKY rats⁴⁷. This increased expression of JAM-1 induces both hypertension (in normotensive rats) and leukocytes accumulation in the NTS microvasculature and may lead to enhancement of cell transmigration across the tight junctions, activation of platelet aggregation, and increased production of cytokines⁴⁷, thereby contributing to the inflammatory status of the NTS which may then directly or indirectly regulate NTS neuronal activity and baroreceptor functions. Saavedra and collaborators have also demonstrated that inflammation in the SHR is not restricted to the peripheral vessels but it is also extended to the cerebral vasculature⁴⁴. Treatment of SHR with angiotensin receptor blockers with access to the brain, reversed cerebrovascular inflammation by decreasing macrophage infiltration, normalization of eNOS/iNOS ratio, reversal of heat-shock protein up regulation, and reversal of increased TNF-α, IL-1β, and ICAM⁴⁴. These observations beg the question: are inflammatory conditions in the cardiovascular control regions of the brain a result of cytokines produced in the circulation and exerting their influence in the cerebral vasculature to initiate a cascade of events leading to increased neuronal activity and sympathetic tone? Alternatively, are cytokines produced within the brain to regulate neuronal activity? There is no conclusive answer to support or refute either contention at the present time. However, multiple studies support the latter view. They include the following: (i) increased expression of inflammatory cytokines has been demonstrated in cardiovascular relevant brain regions of various animal models of hypertension^48,49,50,51 and in the heart failure model^45,52; (ii) ICV infusion of inflammatory cytokines such as IL-1α activates the sympathetic nervous system and increases BP^53,54; (iii) overexpression of IL-10 in the CNS following ICV administration decreases TNF-α, IL-1α, PGE2, and COX-2 in the PVN and ameliorates sympathoexcitation⁵⁵; (iv) AAV-mediated neuronal expression of IL-10 in the PVN attenuates chronic Ang II-induced hypertension⁵¹; (v) IL-6 in the NTS depresses the baroreceptor reflex gain⁵⁶; (vi) Migratory inhibitory factor (MIF), an anti-inflammatory cytokine, has been shown to negatively regulate chronotropic actions of angiotensin II in the neurons⁵⁷. Thus, decreased MIF expression in the PVN in response to angiotensin II may be associated with the development of high BP in the SHR⁵⁰; and (vii) numerous studies have demonstrated that inflammatory cytokines have profound effects on neuronal activity directly. For example, TNF-α and IL-1α increases neuronal firing frequency by binding to their respective receptors and the activation of NFkβ⁵⁸. Consistent with this, is the observation that ICV infusion of a specific NFkβ inhibitor (pyrrolidine dithiocarbamate) significantly attenuates Ang II-induced hypertension and inflammatory cytokines in the PVN⁴⁹. Cytokines can modulate neuronal activity via production of reactive oxygen species (ROS) and activation of iNOS. Resulting nitric oxide (NO) diffuses to adjacent tissues to modulate neuronal activity by influencing various ion channels^59,60. (viii) Receptors for cytokines are found in several sites in the brain on multiple cell types including neurons, microglia and astrocytes⁶¹. There is the possibility that a combination of both mechanisms (i.e. contribution of peripheral cytokines via cerebral vasculature and generation of intrinsic cytokines within the brain) may contribute to the neuroinflammatory process in neurogenic hypertension. It appears that the inflammatory process in the cardiovascular regulatory regions of the brain is linked with modulation of the autonomic nervous system and the increased BP. However, we acknowledge that not all inflammation in the brain results in hypertension and propose unique inflammatory profiles relate to distinct disease states⁴⁸. Finally, animal and human studies indicate hypoperfusion of the hypertensive brain can occur due to increased cerebral vascular resistance^62,63. The relationship between hypoperfusion and inflammation within brain microvasculature is undetermined but one can trigger the other thereby presenting a possible mechanism.

Brain microglia, cytokines, and neurogenic hypertension

Microglia, the resident macrophages of the brain, have attracted significant attention in recent years as a result of their role in various neuropathological conditions. They are activated rapidly in response to brain injury and produce a variety of inflammatory mediators including cytokines. In ischemic stroke, for example, microglia exerts neurotoxic functions through production of ROS and cytokines. This damages the BBB integrity leading to infiltration of inflammatory cells into the brain^64,65. Activated microglial cells are present in the CNS from patients with chronic neurodegenerative diseases including Alzheimer's disease, Parkinson's disease, and amyotrophic lateral sclerosis^66,67. Their importance in these disease processes has been studied extensively.

The concept that microglial cells in the cardiovascular relevant brain regions may participate in hypertension-linked pathophysiology is derived, in part, from our studies which show that chronic Ang II-induced hypertension is associated with increases in activated microglia in the PVN⁵¹. This activation is associated with increases in the levels of PVN inflammatory cytokines. Similarly, we found an increase in activated microglia in the PVN of the SHR (Figure 2). Further evidence in support of this concept is provided by our experiments with minocycline, an antibiotic which can cross BBB and inhibit microglial activation⁵¹. ICV infusion of minocycline attenuates Ang II-induced high BP, decreases PVN cytokines, and attenuates cardiac hypertrophy⁵¹. Multiple examples in the literature indirectly support the role of brain microglia. Saavedra et al have demonstrated that acute brain inflammatory response to bacterial endotoxin lipopolysaccharide results in the activation of microglia, not only in the cardiovascular-relevant brain regions (PVN, SFO) but also in prefrontal cortex, hippocampus, and amygdala^68,69. This is associated with increases in brain expression of inflammatory cytokines and their receptors, adhesion molecules, and iNOS^68,69. Systemic administration of centrally acting AT1 receptor blockers decreases these responses. Carnevale et al studied interactions among hypertension, inflammation, and β-amyloid (Aβ) deposition in a hypertensive mouse with transverse aortic coarctation⁷⁰. They demonstrated that hypertension, per se, triggers microglia activation and neuroinflammation before the Aβ deposition takes place, suggesting that microglial activation precedes and may trigger neurodegenerative processes in Alzheimer's disease. Finally, Lanz et al provided evidence that Ang II sustains brain inflammation in the experimental autoimmune encephalomyelitis (EAE) mouse model of multiple sclerosis via the transforming growth factor β (TGF-β)⁷¹. Their studies indicate that neurons and microglial cells express AT1 receptors which are primarily responsible for Ang II-induced increases in TGF-β production. Collectively, these observations provide persuasive arguments in support of brain for transmitting inflammatory signals that spread across BBB into cardiovascular control regions of the brain.

Figure 2. Activated microglial cells in the PVN of WKY rat and SHR.

Immunostaining with OX-42 antibody was used to identify activated microglia as described elsewhere⁵¹. Quantitation based on morphological analysis revealed a 75% increase in activated microglial cells in the PVN of SHR vs. WKY rat.

Visceral control of inflammation and putative alterations in neurogenic hypertension

The brain is devoid of the “classical immuno-surveillance system” but is in constant vigilance of the inflammatory status of the body by a reflex afferent feedback system. In recent years, the concept of “vagal immunoflex” has been introduced. This neuronal circuit involves activation of the vagal nerve afferents sensitive to systemic inflammation and reflex release of acetylcholine (Ach) in organs such as the liver, spleen, bone marrow, and heart to dampen the tissue inflammation by reducing pro-inflammatory cytokines. Presence of specific Ach receptors (nAchRα 7) on inflammatory cells support this anti-inflammatory role^72,73,74. In addition, vagal afferents can directly sense cytokine levels in the blood via specific cytokine receptors such as IL-1R present on the glomus cells adjacent to vagus nerve afferent endings within paraganglia located within the gastrointestinal tract, spleen, bone marrow, liver, and heart^{75,76,77,78,79}. Thus, vagal immunoflex system presumably transmits signals to the NTS, the central site of vagal termination, before onward relay to dorsal vagal motoneurons.

It is concluded from the above discussion that the vagus nerve is vital in informing the brain of the peripheral immune status and controlling the activity of the immune cells. However, is this effective in hypertension? In hypertension, cardiac vagal tone and the cardiac vagal baroreflex gain are both depressed^80,81. Similarly, we propose that the anti-inflammatory function of the vagus is also suppressed in conditions of hypertension. The possibility that the raised levels of sympathetic activity (vasomotor nerves) associated with hypertension is also pro-inflammatory, is a hypothesis at this point.

Dysfunctional ANS, endothelial repair, and neurogenic hypertension

Could dysfunctional cardiovascular autonomic activity in neurogenic hypertension suppress endothelial repair? It is well-accepted that endothelial dysfunction is an early event in hypertension-induced vascular pathophysiology. The healthy endothelial cells lining the blood vessels have both mechanical and functional roles, producing NO to promote vasodilation and reducing oxidative stress and inflammation. Since mature endothelial cells have a limited regenerative capacity, endothelial progenitor cells (EPCs) contribute to the repair and maintenance of endothelial damage to maintain normal vascular homeostasis in healthy individuals. In hypertension, however, dysfunctional endothelial cells produce ROS and other inflammatory molecules, which result in vasoconstriction, platelet activation, inflammation, and fibrosis^82,83,84,85. In addition, EPC numbers, migratory ability and functions are all impaired in hypertension⁸⁶. A combination of these processes leads to accelerated vascular dysfunction and hypertension-associated pathophysiology; therefore, (i) are EPCs from hypertensives dysfunctional, and, (ii) is there a neural control mechanism for EPC release and function that are altered in neurogenic hypertension?

1. EPC numbers and function have been inversely correlated in patients with cardiovascular disease (CVD), obesity, diabetes, chronic kidney disease, and other immune diseases^87,88,89,90. Similarly, decreases in numbers as well as their functions have been demonstrated in human hypertensives and animal models of hypertension^91,83,84,86. Dysfunctional EPCs are associated with increases in EPCs ROS, NADPH oxidase, and decreases in NO production. This view is supported by evidence that antihypertensive drugs such as ACE inhibitors and ARBs increase EPCs numbers and improve their function^92,93.

2. Release and function of EPCs from bone marrow appears to be neurally-regulated. Bone marrow is densely innervated by the sympathetic nervous system which stimulates the release of bone marrow-derived stem cells into the circulation^{94,95,96,97,98}. Similarly, cholinergic fibers have been found in the rat femur bone marrow around hematopoietic cells, suggesting parasympathetic regulation bone marrow-derived cells as well⁹⁸. Finally, retrograde tracing studies with rabies virus have clearly indicated a connection between the bone marrow and many cardiovascular control regions of the brain⁹⁹.

Considered together, these data suggest that dysfunctional sympathetic/parasympathetic regulation of bone marrow could lead to impaired EPCs levels and their repair capacity in hypertension. An altered sympathetic (and/or parasympathetic) drive to the bone marrow may contribute to EPC dysfunction in hypertension. These proposals are supported by the following evidence: (i) Decrease in sympathetic activity to the bone marrow in diabetes is associated with decreases in circulating EPCs which are also dysfunctional^97,88, and (ii) while it is generally considered that sympathetic activity is increased in hypertension, separate end organs can be differentially regulated by the sympathetic nervous system, and certain vascular beds reportedly receive decreased sympathetic activity in hypertension^100,101. It is tempting to suggest that a decreased sympathetic drive to the bone marrow could contribute to its dysfunction. Additionally, a decreased parasympathetic (vagal) regulation of the bone marrow may instead contribute in its dysfunction (Figure 1).

Unifying hypothesis

Synthesis of all available evidence has led us to propose the following hypothesis as a mechanism for neurogenic hypertension (Figure 1). Increase in peripheral Ang II leads to a cascade of events: AT1 receptors in the neurons of CVOs are activated resulting in increases in ROS^102,13,103. This leads to the activation of microglia in the PVN (Figure 2) which increases production of inflammatory cytokines, ROS, and other inflammatory modulators⁵¹. These modulators would stimulate PVN neuronal activity which, via brainstem (NTS and RVLM), would be reflected in modulation of ANS leading to increases in blood pressure. An imbalance in sympathetic/parasympathetic activities also regulates bone marrow activity by increasing inflammatory cells and decreasing EPCs and their functions. Decreases in EPC numbers and their dysfunctions accelerate vascular dysfunction and damage. Increases in inflammatory cells would result in increased inflammatory mediators which would exaggerate vascular dysfunction leading to the development of pathophysiology of cardiovascular-relevant tissue characteristics of hypertension, as well as issues of hypoperfusion⁶³.

The vascular arm of the neurovascular communication involves activation of AT1 receptors on the cerebral vessels which have free access to increased levels of peripheral Ang II. Activation of these receptors stimulates cytokines, JAM-1, ROS, and adhesion molecules. This interrupts the integrity of the BBB^{104,105,106,107} and enhances adherence of inflammatory cells causing leukocytes and cytokines to enter the brain parenchyma, contributing to microglial activation and inflammatory status of the PVN. Similar process could occur in other cardioregulatory brain regions.

Future directions

Although some evidence, as presented in this review, exists in overall support of the above hypothesis, many important issues remain to be proven. Further evidence would be needed in order to provide conclusive validation for our hypothesis.

1. In our hypothesis we have proposed a central role for the PVN. This is primarily based on our data that shows activated microglia predominantly in this region (51, Figure 2). However, the simultaneous involvement of a similar mechanism in other brain regions (i.e. NTS, RVLM, etc.) cannot be ruled out and needs to be investigated.

2. The hypothesis proposes that increases in activated microglia in the brain are both resident and derived from the peripheral inflammatory cells. There is no evidence to support this in hypertension. However, recent evidence from Chen et al¹⁰⁸ has shown that defective microglia cause pathological grooming in Hoxb8 mutant mice. Bone marrow transplant from wild type mice increases brain microglia and cures Hoxb8 mice of pathological grooming, indicating that activated microglia in this animal model result from immune cells in the peripheral circulation. Additionally, Longo et al¹⁰⁹ has demonstrated that a significant number of activated microglia observed in the hippocampus of status epileptus mice have bone marrow origin. Similar transplantation studies are needed to determine the origin of activated microglia in the brain in hypertension.

3. Are microglia targets of Ang II actions? It is without dispute that neurons in the cardiovascular regulatory brain regions such as the PVN express AT1 receptors and Ang II interaction with these receptors profoundly affects neuronal activity by stimulating a cascade of signaling pathways^110,111. It remains to be proven, however, whether AT1 receptors are present on microglia and if so, how they contribute to increased neuronal activity in neurogenic hypertension. Recent studies have shown a co-localization of AT1 receptor with Iba-1 positive cells, a marker for microglia in the brain of mouse model of multiple sclerosis⁷¹. In addition, TGF-β is co-localized with AT1 receptors in microglia. Furthermore, Lanz TV et al⁷¹ showed that EAE-induced TGF-β expression is blocked by Candesartan, an AT1 receptor blocker concluding that microglia contains functional AT1 receptors. Evidence for this conclusion is provided by in vitro studies which demonstrate that microglial cells in culture increase TGF-β expression in response to Ang II⁷¹. How these receptors regulate generation of inflammatory cytokines, ROS, and other mediators to influence neuronal activity, remains to be investigated.

4. We have hypothesized that a functional balance between the inflammatory cells and EPCs from the bone marrow is important in the maintenance of normal cardiovascular physiology. Its imbalance leads to vascular pathophysiology associated with hypertension. This concept needs to be further investigated.

Perspectives

Despite recent advances, it has been extremely difficult to successfully manage and control hypertension in a significant number of hypertensive patients. A majority of these unresponsive patients exhibit increased sympathetic drive and display neurogenic hypertension. Lack of underlying mechanisms involving neural-peripheral communication associated with increased sympathetic activity has contributed to this. We have proposed a unifying hypothesis of a dysfunctional neural-immune-vascular communication pathway that may be responsible for neurogenic hypertension. Evidence to support/refute this hypothesis will be critical in the elucidation of the underlying impairment, and exploring the mechanism for the advancement of neurogenic hypertension therapeutics.