Brain Transforming Growth Factor-β Resists Hypertension Via Regulating Microglial Activation
Background and Purpose—Hypertension is the major risk factor for stroke. Recent work unveiled that hypertension is associated with chronic neuroinflammation; microglia are the major players in neuroinflammation, and the activated microglia elevate sympathetic nerve activity and blood pressure. This study is to understand how brain homeostasis is kept from hypertensive disturbance and microglial activation at the onset of hypertension.
Methods—Hypertension was induced by subcutaneous delivery of angiotensin II, and blood pressure was monitored in conscious animals. Microglial activity was analyzed by flow cytometry and immunohistochemistry. Antibody, pharmacological chemical, and recombinant cytokine were administered to the brain through intracerebroventricular infusion. Microglial depletion was performed by intracerebroventricular delivering diphtheria toxin to CD11b-diphtheria toxin receptor mice. Gene expression profile in sympathetic controlling nucleus was analyzed by customized qRT-PCR array.
Results—Transforming growth factor-β (TGF-β) is constitutively expressed in the brains of normotensive mice. Removal of TGF-β or blocking its signaling before hypertension induction accelerated hypertension progression, whereas supplementation of TGF-β1 substantially suppressed neuroinflammation, kidney norepinephrine level, and blood pressure. By means of microglial depletion and adoptive transfer, we showed that the effects of TGF-β on hypertension are mediated through microglia. In contrast to the activated microglia in established hypertension, the resting microglia are immunosuppressive and important in maintaining neural homeostasis at the onset of hypertension. Further, we profiled the signature molecules of neuroinflammation and neuroplasticity associated with hypertension and TGF-β by qRT-PCR array.
Conclusions—Our results identify that TGF-β–modulated microglia are critical to keeping brain homeostasis responding to hypertensive disturbance.
Substantial evidence accumulated in the past decade indicates that hypertension is associated with both systemic and neural inflammation.1–3 Consistent with this notion, immunosuppression reduces blood pressure (BP) in various hypertensive models.4,5 Another feature of hypertension is that it is always accompanied with increased sympathetic nerve activity, especially elevated renal sympathetic nerve activity.3 Therefore, clinical trials of renal sympathetic nerves denervation via surgical sympathectomy have been performed for the treatment of resistant hypertension.6 Inhibition of central immune system can effectively attenuate BP and sympathetic outflow in hypertensive animal models.7–9 These results suggest that central immune system is involved in the regulation of sympathetic outflow and BP. Our recent studies further show that in hypertension, microglia, the principal immune cells of the central nervous system (CNS), turn proinflammatory and participate in neuroinflammation and sympathetic nerve overactivation.8,10 Microglial activation in established hypertension is detrimental to neural homeostasis and exacerbates the disease, whereas depletion of activated microglia alleviates hypertension.11
Microglia are a distinct cell population in the brain parenchyma, derived from primitive myeloid progenitors on embryonic day 8.5.12 Adult microglia are self-maintained with the dependence of CSF1R (macrophage colony–stimulating factor 1 receptor). Their phenotype is determined by transcription factors like PU.1, IRF8, and Sall1.13,14 Recent studies unveiled that microglia participate in shaping neuronal behavior, such as modulating spine formation and synaptic transmission under both physiological and diseases conditions.15,16 These suggest that microglia are proactively involved in shaping neural function. A unique feature of microglia comparing to other brain cells is that their activity can be primed and altered according to the surrounding environment. In normal conditions, microglia function as surveilling cells to detect pathogens and maintain CNS homeostasis.17,18 With the progression of aging and neurodegeneration, microglia are induced to an active mode and result in morphological, gene profile, and polarity changes.19 However, the mechanism of microglial activation in the progress of hypertension especially at the onset stage is currently unknown. A hint from recent study shows that the inhibition of transforming growth factor-β (TGF-β) signaling promotes microglial activation.13
TGF-β family in mammals includes 3 highly homologous members, TGF-β1, TGF-β2, and TGF-β3, among which TGF-β1 is the most abundant isoform and is ubiquitously expressed in various cell types and tissues.20 The actions of TGF-β family members are mainly mediated via binding to the transmembrane serine/threonine kinase receptors I and II, followed by phosphorylation of the transcription factor SMAD protein subunits 2 and 3.20 In the CNS, TGF-β is produced by most cell types, including neurons, astrocytes, and microglia, and plays a pivotal role in brain development and keeping its homeostasis.18 Moreover, TGF-β is a major determinant of adult neurogenesis.21 Loss of TGF-β results in increased neuronal death and microgliosis in acute or chronic brain injuries.22 Intriguingly, recent studies reported that TGF-β is required for microglial survival and proliferation, and TGF-β receptor I is exclusively expressed on microglia under normal conditions.18 Specific deletion of the TGF-β receptor (TGF-βR) resulted in rapid conversion of microglia toward an inflammatory phenotype,13 which implies that central TGF-β is a determinant in controlling microglial phenotype fate. Consistent with this notion, intracranial administration of TGF-β modulates microglial phenotype and promotes recovery from hemorrhagic stroke.23 Collecting above findings that neuroinflammation is highly associated with hypertension,1,24 microglia are the cellular mediator of neuroinflammation and hypertension,11 and central TGF-β plays a vital role in keeping microglial homeostasis,13 we hypothesize that central TGF-β signaling contributes to central BP regulation.
In this work, we investigated the role of central TGF-β in BP regulation and found that brain TGF-β has a suppression effect on angiotensin II (Ang II)–induced hypertension. Then we verified the benefit of TGF-β1 in reducing neural overreaction and microglial inflammatory cytokine expression during hypertension both in vitro and in vivo. Using microglial depletion and adoptive transfer strategies, we demonstrated that the effect of TGF-β on BP modulation is via microglia. Finally, we identified a molecular signature profile associated with hypertension and TGF-β.
All animal procedures were approved by Cedars-Sinai Medical Center and Zhejiang University IACUC in accordance with the NIH Guide for Care and Use of Laboratory Animals. Male adult (8–10 weeks old) C57BL/6J (WT), Tg(ITGAm-diphtheria toxin receptor/EGFP)34Lan (CD11b-diphtheria toxin receptor), and B6.129P-Cx3cr1tm1Litt/J (CX3CR1-green fluorescent protein) mice were purchased from Jackson Laboratories and bred in Cedars-Sinai Medical Center and Zhejiang University.
Osmotic Pump Implantation
Mice were anesthetized by inhalation of 2% to 3% isoflurane/O2 mixture. For intracerebroventricular cannulation, the mouse was positioned in a stereotaxic apparatus. An infusion cannula was implanted into the left cerebroventricular. A second 2-week osmotic minipump infusing 500 ng kg−1 min−1 Ang II was implanted in the same animal subcutaneously 3 or 7 days later.
The BP data of the experiments of TGF-β intracerebroventricular infusion and microglial depletion were collected by telemetry transducer; the BP data of TGF-β neutralizing and signaling blocking were collected by Tailcuff system; and the BP data of microglial transfer were collected with acute BP recording.
Adoptive Transferring of Primed Cells
After 24-hour incubation with saline or TGF-β (4 ng/mL), mouse C8-B4 microglial cells or astrocytes were harvested and adoptively transferred into the recipient WT mouse brain via intracerebroventricular infusion. Each mouse received 5×105 cells via a pump-driven Hamilton microsyringe accordingly to the following coordinates: 0.5 mm posterior to Bregma; 1 mm lateral to midline; and 2 mm below the dura.
The brain tissues from hypothalamus were dissected for RNA purification. The expression profiles of targeting genes were analyzed by customized RT2 Profiler Array System (Qiagen, CAPM13540) following the supplier’s instructions. All the assays were performed on a StepOnePlus Real-Time PCR system (Applied Biosystems).
Brain cells were isolated from the mouse brains by mechanical and enzymatic dissociation; the microglia were purified and enriched by Percoll gradient centrifugation (37% and 70%). For the cellular staining, the following antibodies were used: MHC II-APC, TNFα-PerCP-Cy5.5, CD11b-APC-Cy7, and CD45-FITC. All the antibodies were purchased from either eBioscience, BioLegend, or Pharmingen. Afterward, the samples were analyzed on a Beckman Coulter CyAn ADP, and data were analyzed by FlowJo software.
Data are summarized as mean±SEM. Statistical analysis was performed with GraphPad Prism version 6.
Details of all experimental protocols are presented in the Methods in the online-only Data Supplement.
TGF-β Regulates the Development of Hypertension
Previous studies reported that there is increased TGF-β1 in both plasma and monocytic cells of hypertensive patients compared with normal controls.25,26 To investigate the role of intracranial TGF-β in hypertension, we first measured the level of TGF-β1 in the hypothalamic paraventricular nucleus (PVN), which is central in sympathetic nerve activity regulation. We observed that TGF-β1 is constitutively expressed in normotensive mice and slightly but significantly increased in Ang II–induced hypertensive mice (Figure I in the online-only Data Supplement). To understand whether TGF-β is involved in BP regulation in the CNS, we first adopted a depletion strategy by intracerebroventricular infusion of an anti-(pan)TGF-β neutralizing antibody. Mice were treated with TGF neutralizing antibody via 2-week intracerebroventricular minipumps to deplete brain TGF-β. Preliminary results showed that 50 µg/d by intracerebroventricular minipump produced a maximal effect and could decrease brain TGF-β by 27% (Figure II in the online-only Data Supplement), and this dose did not change kidney TGF-β level or baseline BP. Three days later, mice were induced hypertension by subcutaneous infusion of Ang II (500 ng kg−1 min−1) via minipumps; BP was monitored for the following 9 days (Figure 1A). The result showed that the mice receiving TGF-β–neutralizing antibody had significantly elevated BP on the first 4 days after Ang II treatment compared with the mice receiving an isotype antibody (Figure 1A). Afterward, there was no difference between these 2 groups, which was probably because of an incomplete depletion of TGF-β. To further investigate the role of central TGF-β in BP regulation, we intracerebroventricularly delivered a selective TGF-β receptor inhibitor kinase inhibitor SB52533427 before hypertension induction. The specificity of SB525334 in abrogating the effect of TGF-β was validated in C8-B4 microglial cell line (Figure III in the online-only Data Supplement). Blocking TGF-β signaling caused a continuously greater BP response to Ang II than that of the intracerebroventricular vehicle-treated group (Figure 1B).
Above findings suggest that brain TGF-β could be an important regulator of BP. We next examined whether supplementing TGF-β could suppress BP increase at the onset of hypertension. Mice received intracerebroventricular infusion of recombinant TGF-β1 (50 ng/d) or saline 3 days before Ang II treatment (Figure 1C). Intracranial treatment with TGF-β1 significantly retarded the BP response to Ang II infusion, evidenced by 20 mm Hg lower than that of the mice receiving saline (Figure 1C).
At the end of the TGF-β1 supplementation experiment, we examined the expression of N-methyl-d-aspartate (NMDA) receptor and glutamate decarboxylase Gad65, both of which are important for neurogenic hypertension.28,29 Moreover, we examined the renal norepinephrine level, which is an indicator of sympathetic nerve activity.30 Ang II–induced hypertension caused a significant upregulation of the NMDA receptor subunit NMDAR2A, a downregulation of Gad65 in the PVN and brain stem, and a significant increase in renal norepinephrine, the latter suggesting neuronal overexcitation (Figure 2A and 2B). Consistent with the BP measurements, intracerebroventricular TGF-β1 treatment maintained NMDAR2A, Gad65 (Figure 2A), and norepinephrine (Figure 2B) at normal levels despite Ang II treatment. Taken together, these data suggest that TGF-β in the CNS can suppress central overactivity in hypertension.
TGF-β Regulates Microglial Activation in Hypertension
Our previous work showed that activated microglia play a key role in neuroinflammation, which aggravates hypertension.11 Considering that TGF-β is critical in determining microglial phenotype,13 and TGF-β receptor inhibitor is exclusively expressed in microglia in the CNS,18 we postulated that TGF-β exerts direct effects on microglia in hypertension. Thus, at the end of the experiment described in Figure 1C, the brain tissues of each group were collected and performed immunohistochemistry staining. Microglia were detected by marker Iba-1. In contrast to the ramified appearance of naive microglia, microglia from hypertensive mice displayed soma enlargement and process retraction (Figure 3A). Moreover, there is a greater density of microglia in the PVN of the hypertensive mice, as quantified by fractional area analysis.8 These changes indicate microgliosis, a characteristic of microglial activation.8,11 In contrast, we observed that TGF-β1 pretreatment prevented microgliosis in the PVN after hypertension induction (Figure 3A).
Upregulation of MHC class II and TNF-α is a hallmark of activation of microglia.11 We next examined both markers on microglia dissociated from naive, hypertensive and hypertensive mice administered with TGF-β1 via the intracerebroventricular route. Flow cytometry analysis revealed that MHC class II expression and TNF-α production were significantly increased in microglia of mice infused with Ang II. However, in the TGF-β1 pretreatment group, the expression of MHC class II and TNF-α is maintained at normal level (Figure 3B and 3C), which indicated that TGF-β suppresses the upregulation of MHC class II and TNF-α induced by hypertension. Thus, we conclude that TGF-β negatively regulates activation of microglia during hypertension.
TGF-β regulates gene expression via the phosphorylation of SMAD2 and SMAD3.31 We thus examined SMAD2/3 phosphorylation (pSMAD2/3) in C8-B4 microglial cells and found that pSMAD2/3 was significantly elevated after TGF-β1 (20 ng/mL for 30 minutes) priming (Figure 4A). In support of this notion, there were 25% of microglia expressing pSMAD2 30 minutes after intracerebroventricular injection of TGF-β1 (20 ng in 1 μL) into CX3CR1-green fluorescent protein mice, examined by immunohistochemistry (Figure 4B). Of note, other nonmicroglia (green fluorescent protein) cells barely display phosphorylated SMAD2, confirming that microglia are the major cell type mediating the biological activity of TGF-β in the CNS.
Recent work revealed that TGF-β1 is essential in keeping microglia in a quiescent state; disruption of TGF-β signaling promotes microglia toward proinflammatory activation.13 To investigate the suppressive effects of TGF-β, murine naive C8-B4 microglia were pretreated with or without TGF-β1 (4 ng/mL) for 18 hours followed by challenge with either prorenin (50 nmol/L), a hypertensive factor, or lipopolysaccharide (0.1 ng/mL) for another 6 hours. Consistently, pretreatment with TGF-β1 significantly attenuated prorenin- or lipopolysaccharide-elicited TNF-α expression in C8-B4 cells (Figure IV in the online-only Data Supplement).
Central TGF-β Modulates BP Through Microglia
Previously, we successfully devised a microglial depletion strategy by intracerebroventricular infusion of diphtheria toxin (DT) to CD11b-diphtheria toxin receptor mice.11 We thus investigated whether microglial depletion would remove the suppressive effect of TGF-β on BP in hypertension. CD11b-diphtheria toxin receptor mice were delivered via the intracerebroventricular route with TGF-β1 (50 ng/d) mixed with or without DT (800 pg/g BW/d). Six days later, mice were induced hypertension by Ang II infusion. Compared with the saline control group, intracerebroventricular TGF-β1 treatment retarded the BP increase in experimental hypertension. However, in TGF-β1+DT group, microglial depletion not only completely abrogated the suppressive effect of TGF-β1 on the BP increase, but further significantly elevated the BP compared with the saline controls (Figure 5). Of note, intracerebroventricular infusion of TGF-β1 with or without DT did not change the baseline BP (Figure 5). These data are intriguing because they not only indicate that microglia mediate the effect of TGF-β in hypertension but also imply that microglia themselves are important in maintaining brain homeostasis on the initiation stage of hypertension. To further understand the role of microglia in hypertension, we monitored BP in a fourth group, in which microglia were depleted by DT infusion before hypertension induction. Predepletion of microglia resulted in a more acute BP response than that of the intracerebroventricular saline control group; the BP curve was coincident with that of the DT+TGF-β1 cotreated group (Figure 5). Taken together, our study indicated that microglia under different status show distinct even opposite function in hypertension; TGF-β drives microglia toward an immunosuppressive phenotype, suppressing BP increase.
To further confirm that microglia could mediate the effects of TGF-β on BP regulation, we adoptively transferred C8-B4 microglia with or without in vitro TGF-β1 priming into the cerebroventricular of naive mice. Twenty-four hours after their intracerebroventricular transfer, C8-B4 cells tagged with CFSE were distributed in the recipient brains, mostly localized around the third ventricles where PVN is in vicinity (Figure V in the online-only Data Supplement). There was no difference in baseline BP and heart rate between the groups (Table in the online-only Data Supplement). A single intracerebroventricular injection of Ang II (50 ng) elicited an acute pressor response in both groups (Figure VI in the online-only Data Supplement). However, the magnitude and area under the response curve elicited by Ang II were significantly attenuated in mice transferred with TGF-β1–primed microglia compared with the ones received naive microglia. This difference was not observed in mice receiving medium or TGF-β1–primed astrocytes.
Identification of Neuroplasticity and Neuroinflammation Signatures Associated With Hypertension and TGF-β
To identify the neural molecular signatures associated with hypertension and TGF-β, we performed qRT-PCR arrays (Qiagen, CAPM13540) of PVN and brain stem lysates isolated from naive mice and mice made hypertensive by Ang II for 10 days. Moreover, we also examined the samples from mice pretreated via the intracerebroventricular route with TGF-β1 followed by 11-day Ang II treatment in the presence (TGF-β1+Ang II) or absence of microglia (TGF-β1+DT+Ang II). These gene targets were selected based on their widely reported roles in synaptic alteration, pain, or neural injury. We compared the identified molecular signature of genes whose expression changed by >2-fold across all 4 groups and pooled the genes into 3 categories: inflammatory factors (Figure 6A), signaling regulators (Figure 6B), and ion channels (Figure 6C).
Previous studies revealed that Egr1, Egr3, JunB, CD34, and Cox-2 are downstream genes induced by TGF-β1.32–36 Consistently, in our qRT-PCR array analysis, these genes were all upregulated in the TGF-β1–treated brains (Figure 6A and 6B), which verifies the robustness of this assay. This is consistent with reports that TGF-β is required for microglial proliferation because the stem cell antigen CD34 is expressed in proliferating resident microglia,18 and our assay shows that upregulation of the CD34 by TGF-β1 is intrinsic to microglia because microglial depletion completely abrogated the increase (Figure 6A). Moreover, our results confirm that microglia are the major source of the neurohumoral BDNF (Figure 6B), consistent with the finding that microglia-derived BDNF promotes neuron synapse formation.15
Ion channel genes are mainly expressed by neurons in the CNS. The assay reveals 3 distinctive patterns of ion channel expression in the brains from mice made hypertensive by Ang II when compared with the normotensive brains (Figure 6C). Hypertension is associated with downregulation of the nonselective cation channels (Trpv1, Trpv4, and Trpa1), upregulation of the voltage-gated sodium channel Scn10a, and no changes in the glutamate receptor (NMDA receptor subunit Grin2b and AMPA receptor subunits Gria1 and Gria2). However, intracerebroventricular TGF-β1 treatment not only substantially offset the effects of Ang II on Trpv1, Trpv4, Trpa1, and Scn10a but also upregulated glutamate receptor Grin2b and Gria2. Depletion of microglia completely abolished these effects of TGF-β1, again indicating that microglia mediate the effects of TGF-β in the brain.
Other changes in genes associated with neuroplasticity include the neurotrophic receptor Ntrk1, the signaling of which promotes sympathetic neuron activity37 (Figure 6B). Ntrk1 was upregulated by Ang II treatment and by depletion of resting microglia; in the presence of microglia, TGF-β1 suppressed its increase induced by Ang II. Among the transcription factors, Egr1–4, JunB, and Nfatc4 were considerably upregulated by TGF-β1, and our data indicate that these processes occurred in microglia because deprivation of these cells normalized or even decreased expression of Egr1–4, JunB, and Nfatc4 (Figure 6B).
It was recently reported that macrophage-derived Cox-2, in the periphery, is important for antagonizing salt-sensitive hypertension.38 Our array data show that TGF-β1 treatment before Ang II raised Cox-2 expression by 5-fold, and microglia depletion completely abrogated the effects of TGF-β1 on Cox-2 (Figure 6A). To confirm these results, we examined Cox-2 expression in the PVN and brain stem of naive, Ang II–hypertensive and Ang II+TGF-β1–treated mice. RT-PCR confirmed that TGF-β1 treatment significantly increased Cox-2 mRNA levels (Figure VIIA in the online-only Data Supplement); further, we demonstrated that TGF-β1 can dramatically elevate Cox-2 expression in primary microglia (Figure VIIB in the online-only Data Supplement), suggesting that in the CNS, the microglia-derived Cox-2 may also play a critical role in antagonizing the BP-induced changes and that TGF-β1 induces its expression.
Mice lacking TGF-β die of multiorgan inflammation early in life, highlighting a crucial role for TGF-β in dampening self-harmful inflammatory responses and maintaining tissue homeostasis.39 Hypertension is associated with neuroinflammation.1,2,24 This study, as a continuum of our previous studies, sought to understand how central physiological regulation and neural immune response are balanced in the setting of hypertension.
Using multifaceted approaches, we demonstrated that TGF-β is a negative regulator of neuroinflammation and hypertension. These data strongly support the hypothesis that resting microglia are neuroprotective from hypertensive disturbance at the onset of hypertension and maintain normal BP. Depletion of homeostatic microglia results in CNS more susceptible to hypertensive stimulation and BP further elevated, which is different from our previous study showing that the microglia were depleted after hypertension was established. Therefore, removal of activated microglia alleviated neuroinflammation and BP. Moreover, we found that in the resting state, TGF-β is constitutively expressed in the CNS, and consistent with others reports, we show that microglia are the major responsive cells to TGF-β in the CNS. Intriguingly, a recent study revealed that TGF-β is crucial for microglial survival and proliferation,18 which indicates that similar to macrophages in the periphery, immunosuppressive microglia are more populous than their proinflammatory counterparts in the CNS in normal basal condition. Thus, in contrast to the promoting role of microglia in established hypertension, resting microglia exhibit a downregulated immune phenotype adapted to the immune-suppressive environment of the CNS. Consistent with our findings, very recent work by Taylor et al23 showed that intracerebroventricular TGF-β treatment induces microglial toward an anti-inflammatory phenotype, and TGF-β plasma levels positively correlate patient recovery after hemorrhagic stroke.
Our study suggests that TGF-β and its signaling could be potential targets in managing neurogenic hypertension. Thus, understanding the mechanisms of how TGF-β and resting microglia regulate neuronal activity seems crucial. For this purpose, we constructed an RT-PCR array to measure genes associated with neuroplasticity and neuroinflammation at the initiation stage of hypertension. Not surprisingly, hypertension is accompanied by upregulation of an array of proinflammatory cytokines, for example, Ccl2, TNF-α, IL-6, IL-1α, and IL-1β (Figure 6A). A striking finding of this assay is that resting microglia play a critical role in containing neuroinflammation because predepletion of microglia exacerbated inflammation, suggesting that it is not microglia but other (unknown) brain cells that initiate neuroinflammation in hypertension. It is conceivable that in such a proinflammatory environment, microglia are educated to become fully activated later on. Thus, identifying the inflammation initiator may provide pivotal insights into the development of neurogenic hypertension. In addition, the array results reveal for the first time 3 distinctive patterns of ion channel expression changes in the Ang II–induced hypertensive brain. Most importantly, intracerebroventricular TGF-β1 treatment substantially offset the effects of Ang II on the expression of these ion channels, and depletion of microglia completely abolished these effects of TGF-β1. The upregulation or downregulation of these ion channels correlates with the BP data and are bone fide neuron activation signatures associated with hypertension. Interestingly, our screen suggests that in the CNS, the microglia-derived Cox-2 may relay the effects of TGF-β in antagonizing BP changes in hypertension.
Taken together, this study unveils a novel immune regulatory pathway in the development of neurogenic hypertension.
In this study, we identified that TGF-β tunes microglia toward an immune-suppressor phenotype under resting conditions, which resist the BP increase during the initiation stage of hypertension. By comparing gene profiles related to neuroinflammation and neural plasticity in mouse brain tissue, we identified a plethora of molecules associated with hypertension, TGF-β, and microglia. These findings suggest that surveillant microglia are tightly regulated by TGF-β, which are critical in maintaining the homeostasis of the CNS and BP.
Sources of Funding
This work was supported by Chinese Fundamental Research Fund for the Central Universities (2016XZZ002-03 to Dr Shi), American Heart Association grant 11SDG6770006 (to Dr Shi), National Natural Science Foundation of China 31270950 and 81670378 (to Dr Shen), NIH grants P01 HL129941 (to Dr Bernstein), HL-14388 (to Dr Johnson) and R01 NS075930 (to Dr Lyden), and China State Scholarship Fund 201306260029 (to Y. Li).
Guest Editor for this article was Miguel Perez Pinzon, PhD.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.117.017370/-/DC1.
- Received March 22, 2017.
- Revision received June 8, 2017.
- Accepted June 22, 2017.
- © 2017 American Heart Association, Inc.
- DiBona GF,
- Kopp UC
- Guzik TJ,
- Hoch NE,
- Brown KA,
- McCann LA,
- Rahman A,
- Dikalov S,
- et al
- Masson GS,
- Nair AR,
- Silva Soares PP,
- Michelini LC,
- Francis J
- Taylor RA,
- Chang CF,
- Goods BA,
- Hammond MD,
- Mac Grory B,
- Ai Y,
- et al
- Porreca E,
- Di Febbo C,
- Mincione G,
- Reale M,
- Baccante G,
- Guglielmi MD,
- et al
- Zhu S,
- Liu Y,
- Wang L,
- Meng QH
- Walshe TE,
- dela Paz NG,
- D’Amore PA
- Glass MJ,
- Wang G,
- Coleman CG,
- Chan J,
- Ogorodnik E,
- Van Kempen TA,
- et al
- Chen SJ,
- Ning H,
- Ishida W,
- Sodin-Semrl S,
- Takagawa S,
- Mori Y,
- et al
- Gervasi M,
- Bianchi-Smiraglia A,
- Cummings M,
- Zheng Q,
- Wang D,
- Liu S,
- et al
- Marone M,
- Scambia G,
- Bonanno G,
- Rutella S,
- de Ritis D,
- Guidi F,
- et al