Introduction

See Editorial Commentary, pp 265–266

Hypertension, particularly resistant hypertension, is associated with enhanced sympathetic tone and can be substantially managed by renal nerve ablation¹ and baroreceptor nerve stimulation.² The sympathetic outflow is controlled by several important nuclei and their circuits in the central nervous system (CNS), especially the hypothalamic paraventricular nucleus (PVN) and the rostroventrolateral medulla and the nucleus tractus solitaries in hindbrain.^3,4 Perturbations of these nuclei have been implicated in hypertension. Although neurons in those regions are the major cells in modulating sympathetic outflow, what factors mediate the elevation of neuronal activity in hypertension stay elusive. Emerging studies indicate that hypertension is accompanied with extensive neuroinflammation and that central anti-inflammatory treatment significantly alleviated hypertension.^5–7 Thus, determining the cellular mechanism of neuroinflammation and neuronal modulation in hypertension is critical to fully understand central regulation of blood pressure.

The CNS has long been considered as immune privileged because of the blood–brain barrier, the brain’s lack of lymphatic drainage to lymph nodes, and suboptimal capacity to present antigen.⁸ Microglia are the primary immune cells in the CNS. They are derived from primitive macrophages emanating from the embryonic yolk sac.⁹ In the CNS, microglia proliferate and maintain homeostasis with limited contribution from peripheral blood-borne cells.¹⁰ Recent transcriptome analyses revealed that microglia have a distinct phenotype from macrophages in other tissues, suggesting unique characteristics of microglia.^11,12 As surveillance cells, microglia are highly sensitive to pathological disturbance in the brain and play major roles in the progressive pathology of neurodegenerative diseases, such as Alzheimer’s disease.¹³ On stimulation, microglia promptly undergo a series of morphological and phenotypic changes, eventually releasing mediators that can directly modulate neuronal activities.^13,14 Moreover, many lines of evidence indicate that microglia are highly involved in shaping neuronal behavior via sculpting dendritic spine formation and modulating neurotransmitter receptor presentation on the synaptic terminals in physiological conditions.^15,16

In the present study, we examined microglia in hypertension and found that microglia were activated in a different pattern from peripheral monocytes. When microglia were depleted by intracerebroventricular (ICV) administration of diphtheria toxin (DT) into the transgenic CD11b-diphtheria toxin receptor (DTR) mice, the neuroinflammation and blood pressure increase induced by either angiotensin (Ang) II or l-N^G-nitro-l-arginine methyl ester (l-NAME) were significantly attenuated. In contrast, adoptive transfer of activated microglia prolonged pressor responses to central application of Ang II. Taken together, our findings indicate that microglia are the key players in the neurogenic regulation of hypertension.

Methods

All surgical and experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Cedars Sinai Medical Center. A detailed Methods section is available in the online-only Data Supplement.

Results

Microglial Activation Pattern in Hypertension

To characterize activation states of microglia in established hypertension, C57BL/6 mice were treated with subcutaneous infusion of Ang II or by feeding l-NAME in drinking water for 4 weeks. Systolic blood pressure reached 130 mm Hg in the 1st week and sustained in the following 3 weeks (Figure S1A in the online-only Data Supplement). Four weeks after the induction of hypertension, mice were euthanized and microglia were analyzed. In both models, there was a significant increase of microglia in the PVN and motor cortex of hypertensive brains compared with the normotensive brains, as manifested by an increased area of Iba1 staining (Figure S1C–S1D). In contrast to the ramified appearance of naïve microglia, hypertensive microglia showed soma enlargement and process retraction. Thus, hypertension is associated with microgliosis, a characteristic of microglial activation.¹⁷

To define the characteristics of microglia in hypertension, we dissociated microglia from the brains of normotensive mice or mice made hypertensive with Ang II or l-NAME. Because we previously found that there were increases of tumor necrosis factor α (TNFα) and interleukin (IL)-1β expression in the brain of hypertensive rats,^5,16 we first evaluated microglial expression of these proinflammatory cytokines using intracellular staining and flow cytometry analysis. After dissociated from the mouse brains, microglia were cultured in vitro for 6 hours in the presence of bredfeldin A, which blocks the secretion of protein from cells.¹⁸ There was elevation of TNFα- and IL-1β-expressing, as well as mild but significantly increased IL-6-expressing, microglia in l-NAME-treated mice compared with those in normotensive animals (Figure 1A). Although the numbers of TNFα-, IL-1β-, or IL-6-expressing microglia from Ang II–treated mice were not conspicuously altered in the resting state (data not shown), there were remarkably more cells expressing these cytokines after lipopolysaccharide (LPS) treatment in Ang II hypertensive microglia than normotensive microglia, which indicates their preactivation (Figure 1B).

• Download figure
• Download PowerPoint

Current concepts of microglial activation arise, in part, from research into macrophage biology. In macrophages, M1 (proinflammatory classical activation) and M2 (alternative activation) represent extremes in a continuum of activation states.¹⁹ We thus investigated the M1-associated markers (MHC class II, CCR7, IFNγR, and iNOS) and the activation markers representing M2 state (CD36, mannose receptor, Tie2, CCR2, and IL-4Rα) in hypertension-associated microglia. After 4 weeks of Ang II or l-NAME treatment, microglia were dissociated and enriched by Percoll gradient centrifugation followed by flow cytometry analysis. Intriguingly, all these molecules except iNOS were upregulated in the hypertensive microglia (Figure 2A and 2B). This activation profile is specific to microglia because the monocytes from Ang II–induced hypertensive mice had only increased IL-4Rα and decreased MHC class II expression compared with their normotensive counterparts (Figure 2B).

• Download figure
• Download PowerPoint

Loss of Microglia-Alleviated Blood Pressure and Neuroendocrinological Factors Associated With Hypertension

To investigate the role of microglia in hypertension development, we first used a microglial depletion strategy. Transgenic CD11b-DTR mice express the DTR under the control of the endogenous CD11b promoter.²⁰ In the CNS, only microglia but not neuronal or other glial cells express DTR in these mice. A single ICV injection of DT resulted in a dose-dependent reduction of microglia in both the PVN and motor cortex (Figure 3A). At a dose of 1000 pg/g, DT caused an over 90% loss of microglia, which was confirmed by both immunohistochemistry (Figure 3A) and flow cytometry analysis (the CD11b⁺CD45^low population in Figure 3B). To be noted, ICV DT injection (1 000 pg/g/d) did not change total blood monocytes (CD11b⁺Ly6G⁻F4/80^low) or inflammatory monocytes (CD11b⁺Ly6G⁻Ly6C^high; Figure 3B). At the dose we used, DT was not toxic to neurons or astrocytes in the CD11b-DTR mice because NeuN or GFAP staining of DT-treated PVN showed no change in the density or distribution of neurons or astrocytes, respectively, in comparison to the untreated CD11b-DTR mice (Figure 3C). To be noted, F4/80 and CD31 staining showed no difference in the density of perivascular macrophages between DT-treated and saline-treated CD11b-DTR mice (Figure 3C). Further, DT injection did not cause any change in body weight and general behavior of mice. Our results showed that ICV infusion of DT to CD11b-DTR mice is a valid model to investigate physiological changes in microglial loss without causing change in the brain or circulating monocytes. Because of the efficiency of microglial depletion with the dose of 1000 pg/g/d, we used this dose in all following studies.

• Download figure
• Download PowerPoint

CD11b-DTR mice were treated with Ang II or l-NAME. After 2 to 3 weeks, when hypertension was established and neuroinflammation had developed, microglia were depleted by ICV infusion of DT. In both models, microglial depletion caused a gradual reduction in blood pressure and by 2 weeks, the mice had a 20 mm Hg lower blood pressure than the ICV saline-infused CD11b-DTR mice (Figure 4A). Notably, ICV DT did not alter resting blood pressure in naive CD11b-DTR or C57BL/6J mice, nor did it change the blood pressure responses when C57BL/6J mice were infused with Ang II (Figure S2). To examine the effects of microglial depletion on neuroinflammation induced by hypertension, brains were harvested and the PVN were dissected at the end of the protocol. The expression of TNFα and IL-1β were analyzed by ELISA. As shown in Figure 4B, hypertension resulted in significant increases of both cytokines. Remarkably, loss of microglia reduced these cytokines to normal levels. These data strongly support our hypothesis that microglia are central in neuroninflammation and blood pressure regulation.

• Download figure
• Download PowerPoint

To understand whether microglia affect neuronal plasticity in hypertension, we examined the expression of the N-methyl-d-aspartate (NMDA) receptor, an excitatory synaptic receptor reported to be critical for neurogenic hypertension.²¹ In the hypertensive PVN, the expression of NMDA subunit GluN2A was upregulated by 2-folds (Figure 4C). Importantly, microglial depletion abolished such an increase, suggesting that loss of microglia may prevent hypertension-associated neuronal excitation. Vasopressin and norepinephrine are 2 CNS-regulated hormones, which are associated with blood pressure increase.³ To better understand the downstream events of neuronal excitation, we examined the levels of plasma vasopressin and kidney norepinephrine in hypertensive mice. Consistent with NMDA receptor expression, both hormones were significantly increased in Ang II–treated hypertensive mice (Figure 4D). Microglial depletion suppressed vasopressin and norepinephrine levels in Ang II–treated mice to the normal levels.

Transfer of Activated Microglia Changed Blood Pressure Response

To further confirm the central role of microglia on blood pressure regulation, we adoptively transferred N9 cells, a murine microglial cell line, to the cerebroventricle of naïve C57BL/6J mice. N9 cells are homogeneousness and they lack the contamination of astrocytes, which is a concern when using cultured microglia from newborn brain. Some cells were primed in vitro with either Ang II (100 nmol/L for 12 h) or LPS (10 ng/mL for 6 h). Twenty-four hours after their ICV transfer, the recipient mice were anesthetized, and their basal blood pressure and heart rate were recorded (Outline in Figure 5A). There was no difference in baseline blood pressure and heart rate across the groups (Table S1). When we induced a transient blood pressure increase by a single ICV injection of Ang II (50 ng), there was an acute pressor response with a 5 to 10 mm Hg rise (Figure 5B) in all groups. Interestingly, there was a significantly prolonged pressor response in mice receiving either Ang II- or LPS-primed microglia compared with the mice receiving naïve microglia or saline. This was specific for activated microglia because transferring Ang II–primed astrocytes did not change the duration of the pressor response. Minocycline has been widely used as an anti-inflammatory reagent affecting microglia.²² We incorporated another 2 cohorts of mice which were transferred with microglia primed with Ang II or LPS in the presence of minocycline. The prolonged pressor responses observed with activated microglia were completely abolished when microglia were coincubated with minocycline before transferring (Figure 5C).

• Download figure
• Download PowerPoint

To confirm these blood pressure changes were associated with changes in neuronal profiling, PVN tissues were dissected from the recipients 24 hours after transfer. There was a 2-fold increase in GluN2A level in LPS-primed group compared with the group receiving naïve microglia (Figure 5D). Minocycline coincubation with LPS fully abrogated this increase. These experiments suggest that activated microglia may potentiate neuronal responses to hypertensive stimulant by enhancing the expression of NMDA subunit GluN2A in the PVN.

Discussion

Hypertension is associated with neuroinflammation; however, the cellular mechanism of neuroinflammation is unknown, and it is unclear whether neuroinflammation contributes to the progression of hypertension. Here, we studied murine Ang II and l-NAME models and found that rampart microglial activation is a hallmark of hypertension-associated neuroinflammation. Because Ang II and l-NAME induce hypertension through different mechanisms, we surmise that microglial activation is a characteristic of hypertension. To approve this hypothesis, future study of human samples will be critical. In both hypertensive models, microglial depletion significantly decreased blood pressure, neuroinflammation, and the levels of peripheral hypertensive hormones norepinephrine and vasopressin. In contrast, adoptive transfer of activated microglia predisposed recipients to hypertensive stimulant. In conclusion, this study shows that microglia are key modulators in the development of neurogenic hypertension.

The cause-and-effect relation between hypertension and neuroinflammation is under debate. In this present study, we focused on the role of microglia in established hypertension and in pressor responses to ICV Ang II. Our depletion data and adoptive transfer data clearly show that activated microglia contribute to blood pressure regulation. However, depletion of microglia did not correct blood pressure to normal level in established hypertension and transfer of microglia did not by itself gave rise to blood pressure increase. Therefore, microglia activation is secondary to hypertensive insults but not an initiator of hypertension in our models. Microglia activation and hypertension forms a feed-forward loop.

Our study unveils a unique activation pattern of microglia associated with hypertension. Both M1 (classical activation) and M2 (alternative activation) markers are upregulated in hypertensive microglia but not in hypertensive monocytes. M1 and M2 represent extremes of a continuum in a universe of macrophage activation states. The upregulation of both M1 and M2 markers have not been reported in any other tissue macrophages in any other models. Such a distinctive phenotype of microglia echoes recent transcriptome analyses, which revealed that microglia, although considered the macrophages in brain, are distinguishable from peripheral myeloid cells.^11,12 Whether this is caused by the unique brain environment or the hypertensive factors or both needs further investigation.

In this study, we devised a microglial depletion strategy by infusing DT ICV to CD11b-DTR mice. A previous study also investigated microglia biology through a DT depletion strategy. In that study, the authors subcutaneously injected DT to CD11b-DTR mice at the age of P3 when their blood–brain barrier was incomplete.²² However, peripheral monocytes and macrophages were also ablated by that approach. Our approach with ICV DT infusion efficiently depleted microglia but leaving other CNS cells and circulating monocytes intact, suggesting the exclusive depletion of microglia. Thus, our study may provide a unique model to investigate microglia biology to a variety of pathophysiological settings in the future.

Microglia are the major sources to produce inflammatory molecules and neurotrophic factors. Proinflammatory cytokines, such as IL-1β and TNFα, can directly or indirectly modulate neuronal activities.^23,24 One mechanism is that they can enhance neuronal excitation by increasing NMDA receptor expression in the postsynapes.^25,26A recent study shows that brain-derived neurotrophic factor produced by microglial are of importance in the regulation of neuronal activity by increasing the expression of NMDA receptors on the postsynaptic terminals.¹⁶ Given the essential role of NMDA receptors in neuronal excitation and plasticity,²¹ signals triggering NMDA receptor presentation are the key factors on neuronal activity. NMDA receptor is a tetrametric glutamate-gated ion channel, assembled by an essential subunit, NR1, and ≥1 components of a second family of subunits NR2A to D and NR3A/B.²⁷ Kinase-mediated phosphorylation directly changes the activity and density of NMDA receptor subunits on the membrane.²⁷ We found that the expression of the NMDA receptor in the PVN was significantly upregulated in the presence of microglia but was decreased when microglia were depleted, suggesting that activated microglia could exacerbate hypertension through augmenting neuronal excitation. Future studies of the molecules mediating microglia and neuronal activation in hypertension will provide potential targets for the management of essential hypertension.

Perspective

Our work provides direct evidence that hypertension elicits a unique activation pattern of microglia. Using depletion and adoptive transfer strategies, we evidenced that microglia are the major cellular modulators of neurogenic hypertension. Hypertension is generally accompanied with elevated sympathetic tone. To understand the central regulation of hypertension is critical to develop efficient treatment to essential hypertension. Our work unveils that microglia are central to neuroinflammation and neuronal regulation of hypertension, which provides a mechanistic insight to this disease. To pinpoint the molecules which mediates microglia–neuron interaction may provide potential targets for the management of essential hypertension.

Footnotes