Osmotic minipumps containing vehicle (saline) or Ang II (600 ng/kg/min) were implanted subcutaneously in mice (n=5/group) under isoflurane anesthesia. Systolic blood pressure was monitored daily in awake mice using tail-cuff plethysmography, as previously described (Kazama et al., 2004; Capone et al., 2010; Capone et al., 2011). In some experiments the AVP V1a receptor antagonist SR49059 (4 mg/100 ml) (Thibonnier et al., 2002) was administered via the drinking water 7-14 days after pump implantation.

Mice underwent SFO-targeted stereotaxic microinjection of an adenovirus encoding CuZnSOD (AdCuZnSOD) or control LacZ (AdLacZ) as described previously (Zimmerman et al., 2004; Peterson et al., 2009). To avoid damage to the neocortex in which CBF was recorded, the micropipette for viral vector delivery was inserted in the hemisphere contralateral to the site of CBF measurement. With this method, viral transduction is largely restricted to the SFO and does not involve other brainstem cardiovascular nuclei (Peterson et al., 2009). In the present study, the efficiency and regional selectivity of the viral transduction of the SFO was verified post-mortem by SOD immunocytochemistry (fig. 1). One week after gene transfer mice were implanted with minipumps delivering slow-pressor doses of AngII or vehicle as described above, and mice were instrumented for CBF recording at day 14. Separate mice were instrumented for assessment of cerebrovascular reactivity before and 30 min after acute i.v. administration of pressor doses of AngII (7.5 ng/min/mouse; i.v.) as previously described (Kazama et al., 2003; Kazama et al., 2004).

Mice were anesthetized with isoflurane in a mixture of N₂ and O₂ (Induction: 5%; Maintenance: 2%). The trachea was intubated and mice were artificially ventilated with an oxygen-nitrogen mixture. The O₂ concentration in the mixture was adjusted to provide an arterial pO2 (paO2) of 120-130 mmHg (Capone et al., 2009; Capone et al., 2010; Capone et al., 2011). One of the femoral arteries was cannulated for recording mean AP (MAP) and collecting blood samples. Rectal temperature was maintained at 37° C using a thermostatically controlled rectal probe connected to a heating pad. End-tidal CO₂, monitored by a CO₂ analyzer (Capstar-100, CWE Inc.), was maintained at 2.6-2.7% to provide a pCO2 of 33-36 mmHg (Capone et al., 2009; Capone et al., 2010). After surgery, isoflurane was discontinued and anesthesia was maintained with urethane (750 mg/kg; i.p.) and chloralose (50 mg/kg; i.p.). Throughout the experiment the level of anesthesia was monitored by testing corneal reflexes and motor responses to tail pinch.

A small craniotomy (2×2 mm) was performed to expose the parietal cortex, the dura was removed, and the site was superfused with Ringer’s solution (37° C; pH: 7.3-7.4) (Composition in mM: NaCl: 137, KCl: 5, MgCl²: 1, Na²HPO³: 1.95, NaHCO³: 15, CaCl²: 2). CBF was continuously monitored at the site of superfusion with a laser-Doppler probe (Perimed) positioned stereotaxically on the cortical surface. The outputs of the flowmeter and blood pressure transducer were connected to a data acquisition system (PowerLab) and saved on a computer for off-line analysis. CBF values were expressed as % increases relative to the resting level. Zero values for CBF were obtained after the heart was stopped by an overdose of isoflurane at the end of the experiment. Although LDF is not quantitative, it monitors relative changes in CBF quite accurately (see ref. (Iadecola, 1997) for a review).

ROS production in the somatosensory cortex was assessed by hydroethidine (DHE) microfluorography as previously described (Girouard et al., 2006, 2007; Capone et al., 2010; Capone et al., 2011). DHE (2 μM; Molecular Probes) was superfused on the somatosensory cortex for a total of 60 min. The brain was removed, frozen, and coronal sections (thickness: 20 μm) were cut through cortex underlying the cranial window using a cryostat. Sections were analyzed using a Nikon Eclipse E800 fluorescence microscope equipped with a custom filter set for detection of DHE oxidation products (Girouard et al., 2006, 2007; Capone et al., 2010; Capone et al., 2011). Images were acquired with a digital camera (Coolsnap, Roper Scientific) and analyzed in a blinded manner using the IPLab software (Scanalytics), as described (Capone et al., 2011). Fluorescent intensities of all sections (20 per animal) were added, divided by the total number of pixels analyzed, and expressed as relative fluorescence units (RFU). Methods for assessing ROS production in the SFO using DHE have been published (Zimmerman et al., 2004). After 14 days of AngII infusion, brains were removed, frozen, and sectioned in a cryostat (thickness: 30 μm). Brain sections including the SFO were incubated for 5 minutes with DHE (1 μM). After washing with phosphate-buffered saline, DHE fluorescence was visualized by confocal microscopy (Zeiss LSM 510).

For light and confocal microscopy, mice (n=4/group) were perfused transcardially with phosphate buffered saline (PBS) followed by 4% paraformaldehyde in PBS. Brains were removed, stored overnight in the same fixative at 4°C, and then submerged in 30% sucrose solution for at least 2 days. Coronal brain sections (thickness: 20 μm) were cut through the somatosensory cortex and the PVN using a cryostat. After antigen retrieval with proteinase K, sections were permeabilized in PBS with 0.5% Triton X-100 (Sigma) and then blocked in PBS with 0.1% Triton X-100 (Sigma) and 5% of normal donkey or goat serum. Sections were incubated with primary antibodies (ET1: 1:200, rabbit polyclonal antibody, Peninsula Laboratories; AVP: 1:16,000, guinea pig, Peninsula Laboratories; Chemicon International; CD31: 1:200, rat monoclonal antibody, BD Biosciences) at 4°C overnight. The specificity of the immunolabel was tested previously (Tharaux et al., 1999) and was verified by omitting the primary antibody. After rinsing, sections were incubated with secondary antibodies. AVP immunolabeling was revealed by a biotylinated goat anti guinea pig IgG (1:400; Inc Star) whereas ET-1 and CD31 immunolabeling were revealed by donkey FITC and Cy5-conjugated antibodies (1:200; Jackson ImmunoResearch Laboratories), respectively. A Nikon light microscope or a Leica confocal microscope were used to visualize the signal associated with each antibody(Capone et al., 2010). For immunolectronmicroscopy mice were perfusion fixed with heparin-saline followed by 3.75% acrolein and 2% paraformaldehyde in PBS (Milner et al., 2010). Coronal sections (40 μm thick) were cut through the somatosensory cortex on a Vibratome and processed for ET1 immunoreactivity using the immunogold-silver method (Chan et al., 1990). Briefly, sections were incubated in anti-ET1 (1:1000; Peninsula) in 0.1% bovine serum albumin 0.1M tris saline (pH7.4) one day at room temperature. Sections then were rinsed, incubated 2 hrs in goat anti-rabbit IgG conjugated to 1-nm gold particles (AuroProbe One; Amersham), rinsed in PB-saline, post-fixed in glutaraldehyde and rinsed again. Gold particles were intensified by incubation for 6 minutes with silver solution (Amersham). Sections then were cut (70 nm thick) on a Leica ultratome, collected on copper grids, counter-stained and imaged on a Tecnai transmission electron microscope.

ET1 was measured in cerebrovascular preparations using commercially available ELISA kit (Enzo Life Sciences). Cerebral blood vessels isolated from the brain surface by stripping the pia under a surgical microscope. This preparation includes large, medium size and small cerebral blood vessels (Park et al., 2011). Vessel homogenates were prepared in 1M acetic acid containing 15 μM Pepstatin A (Calbiochem) and purified over C18 columns as suggested by the manufacturer. Plasma AVP was measured using a competitive EIA kit (Cayman Chemicals) after purification over a phenyl matrix (Bond elut-PH, Varian). qRT-PCR was used to measure mRNA for endothelin converting enzyme-1 (ECE1), a protease that cleaves the ET precursor (big ET) into ET. Primers were 5’-GCCTACCGGGCGTACCAGAAC-3’ and 5’-GGTGTGCGGACAGAGCACCAG-3’. RNA was prepared and qRT-PCR was carried out as previously described (Kunz et al., 2008). ECE1 mRNA levels were normalized to hypoxanthine phosphoribosyltransferase mRNA and relative expression levels were calculated as described (Kunz et al., 2008)

The observation that the cerebrovascular effects of AngII depend on ROS production in the SFO raises the possibility that neurohumoral mechanisms involving in the SFO play a role. The PVN is the target of a major efferent pathway from the SFO and AngII is well known to release AVP through activation of this pathway (Ferguson and Kasting, 1986; Anderson et al., 2001). Therefore, we examined whether slow-pressor AngII leads to AVP release, and, if so, whether AVP participates in the cerebrovascular dysfunction. AngII infusion increased AVP immunoreactivity in the PVN (fig. 4A, B) and elevated plasma levels of AVP at 14 days, but not at 3 or 7 days (fig. 4C). The increase in plasma AVP was abolished by SFO viral gene transfer of CuZnSOD (fig. 4C). The AVP V1a receptor antagonist SR49059, administered orally during day 7-14 of vehicle or AngII infusion, reduced baseline AP in mice receiving vehicle (saline), but did not block the elevation in AP induced by AngII (fig. 4D). However, SR49059, at a dose effective in completely blocking the effects of AVP on CBF (fig. 4H), prevented the cerebrovascular dysfunction induced by AngII only partially (fig. 4E, F). Next, we examined the possibility that plasma AVP alters cerebrovascular regulation acting directly on V1a receptors in cerebral blood vessels (Faraci et al., 1994; Fernandez et al., 2001). Application of SR49059 to the cranial window, at a concentration (10μM) effective in preventing the cerebrovascular effects of AVP (fig. 4H), did not attenuate resting CBF (table 1) or the neurovascular dysfunction induced by AngII (fig. 4G). Thus, V1a receptors are not the final mediator of the CBF dysfunction, but they are required for the cerebrovascular effects of slow-pressor AngII hypertension.

AVP upregulates the potent vasoactive peptide ET1 in systemic vessels (Li et al., 2003a). Therefore, we investigated whether AVP might contribute to the cerebrovascular effects of chronic AngII by increasing ET1 in cerebral blood vessels. AngII infusion induced a profound increase in ET in pial vessel preparations, assessed by ELISA (fig. 5A). The ET elevation was observed between 7 and 14 days of AngII infusion, and was suppressed by oral treatment with the V1a receptor antagonist SR49059 (fig. 5A). AngII infusion increased mRNA expression of endothelin converting enzyme-1 (ECE-1) in pial vessel preparations between 7 and 14 days (fig. 5B), suggesting local vascular conversion of the inactive precursor peptide (big ET) into ET. In support of this hypothesis, cofocal microscopy showed an abundance of ET1 immunoreactivity in pial arterioles of the somatosensory cortex of mice infused with AngII (fig. 5C). Similarly, electronmicroscopy demonstrated an increase in ET1 immunoreactivity in cerebral microvascular endothelial cells of AngII-infused mice (fig. 5D). To determine whether ET1 contributes to the cerebrovascular alterations induced by AngII the ET_A receptor antagonist BQ123 was superfused in the cranial window at a concentration (10μM) sufficient to block the vascular effects of ET1 (data not shown). BQ123 did not affect the AP elevation (Saline: 74±2; AngII: 91±5 mmHg; p<0.05) or resting CBF (Table 1), but it partially rescued the cerebrovascular dysfunction (fig. 6A, B). These observations, collectively, raise the possibility that AVP induces ET1 expression in cerebral blood vessels, which, in turn, mediates the alteration in cerebrovascular regulation. If so, the rescuing effects of BQ123 and SR49059 on the cerebrovascular dysfunction should not be additive. Consistent with this prediction, topical application of BQ123 in mice treated with SR49059 (oral administration, days 7-14 of AngII infusion) did not enhance the protection (fig 6A, B), indicating that V1a and ET_A receptors do not act via distinct mechanisms resulting in additive effects.

The findings presented above indicate that AVP and ET1 can account for only part of the alterations in neurovascular regulation induced by AngII. Therefore, we tested whether AngII mediates the remainder of the cerebrovascular dysfunction by activating AT1 receptors in cerebral blood vessels. To this end, we examined the effect of topical application of the AT1 receptor antagonist losartan in mice treated with slow-pressor AngII. Losartan partially reversed the cerebrovascular effects of AngII (fig. 6C,D), without altering the AP increase (Saline: 75±2; AngII: 92±6 mmHg; p<0.05) or resting CBF (Table 1). In contrast, losartan completely counteracted the cerebrovascular dysfunction induced by acute administration of pressor doses of AngII (fig. 6E,F), attesting to its ability to fully inhibit AT1 receptors. Next, we investigated whether the component of the cerebrovascular dysfunction mediated by AVP and ET1 is distinct from that mediated by AngII. If so, the rescue of the dysfunction afforded by losartan should be additive with that conferred by SR49059 or BQ123. As anticipated, topical application of losartan in mice treated orally with SR49059 completely rescued the dysfunction (fig. 4E, F). Similarly, co-application of BQ123 plus losartan rescued the dysfunction in full (fig. 6G, H). Thus, AVP and ET1 exert their effect through pathways distinct from those of AngII.

The SFO participates in the increase in AP induced by chronic administration of AngII, an effect mediated by ROS (Zimmerman et al., 2004). Circulating AngII does not penetrate the BBB, but the ventromedial core region of the SFO is devoid of BBB (McKinley et al., 2003). Therefore, circulating AngII gains access to the SFO and activates local neurons (Ferguson and Renaud, 1986; McKinley et al., 1992; McKinley et al., 2003). Evidence indicates that AngII in the SFO activates AT1 receptors triggering ROS production by the enzyme NADPH oxidase (Zimmerman et al., 2004; Peterson et al., 2006). The resulting alteration in neuronal calcium homeostasis (Zimmerman et al., 2005) is thought to lead to activation of efferent projections from the SFO, including the PVN. We have shown here that suppression of ROS in the SFO also prevents the alterations in functional hyperemia and endothelium-dependent responses induced by slow-pressor AngII. The attenuation of the dysfunction cannot be explained by the lack of AP increase in mice treated with AdCuZnSOD since we have previously shown that the cerebrovascular effects produced by slow-pressor AngII administration are independent of the associated elevations in AP. Accordingly, the cerebrovascular dysfunction is not observed in phenylephrine hypertension and is present even if doses of AngII that do not increase AP are infused for 2 weeks (Capone et al., 2011). Furthermore, in the present study we observed that the CBF dysfunction is dampened by V1a, ETA or AT1 receptor antagonists despite persistent elevations in AP. Therefore, the suppression of the cerebrovascular dysfunction by AdCuZnSOD gene transfer to the SFO cannot be attributed to the associated effects on AP.

How can the SFO lead to cerebrovascular dysfunction? The SFO is unlikely to mediate the cerebrovascular responses to AngII through direct neural projections to the somatosensory cortex because the SFO projects predominantly to hypothalamus and other autonomic brainstem nuclei (McKinley et al., 2003). Furthermore, we have previously demonstrated that the neurovascular dysfunction induced by slow-pressor AngII is unrelated to effects on cortical neural activity (Capone et al., 2011). Therefore, it is unlikely that the SFO exerts its effect on neurovascular regulation through direct neocortical projections altering neural activity in the somatosensory cortex. On the other hand, the SFO projects to the PVN (McKinley et al., 2003), and is required for the AVP release evoked by icv or systemic injection of AngII (Iovino and Steardo, 1984; Mangiapane et al., 1984). Accordingly, we found that Ang II increases circulating AVP and that inhibition of V1a AVP receptors did not affect resting CBF or baseline CBF responses, but it prevented the cerebrovascular dysfunction induced by AngII. These findings demonstrate for the first time an involvement of AVP in the cerebrovascular dysfunction induced by hypertension. AVP, a hormone with powerful cerebrovascular effects (Faraci et al., 1988; Faraci et al., 1994), is not directly responsible for the cerebrovascular dysfunction by acting on V1a receptors on cerebral blood vessels, since acute neocortical application a V1a receptor antagonist does not reverse the cerebrovascular dysfunction. Rather, AVP induces ET1 expression in cerebral blood vessels, which in turn, is responsible for the cerebrovascular alterations. This conclusion is supported by the observations that: (a) chronic systemic administration of the V1a antagonist blocks the ET1 upregulation in cerebral vessels and ameliorates the dysfunction, and (b) acute neocortical application of the ET_A antagonist BQ123 also ameliorates the dysfunction. Our data, taken together, provide evidence that slow-pressor AngII activates the PVN to release AVP, which, in turn, induces the expression of ET1 in resistance arterioles resulting in cerebrovascular dysfunction.

However, AVP and ET1 do not account for the totality of the cerebrovascular effects induced by slow-pressor AngII since another component of the dysfunction is related to activation of AT1 receptors by AngII. In our model of AngII hypertension cerebrovascular dysfunction does not occur immediately after the start of the AngII infusion, but it develops after 7 days (Capone et al., 2011). Thus, the AngII delivered by the osmotic minipumps is not sufficient to induce cerebrovascular dysfunction, at least initially. However, slow-pressor AngII infusion increases endogenous AngII synthesis by the kidney and increases plasma levels of AngII (Gonzalez-Villalobos et al., 2009). Therefore, it is conceivable that higher levels of circulating AngII are produced endogenously during AngII slow-pressor infusion, leading to the observed AT1-dependent cerebrovascular effects. Interestingly, our finding that CuZnSOD gene transfer blocks both the AVP/ET1 and AngII components of the dysfunction indicates that central pathways represented in the SFO may also be critical for the elevation in endogenous AngII. Another possibility is that the sensitivity of cerebral blood vessels to the effects of AngII increases progressively during the AngII infusion period, but there are no data in support of this hypothesis. Irrespective of the mechanisms of the involvement of AngII, central pathways involving ROS signaling in the SFO are upstream of both the AVP/ET1 and AngII components mediating the dysfunction.

ET1 and AngII mediate the cerebrovascular dysfunction by increasing ROS production and leading to vascular oxidative stress. We have previously demonstrated that the cerebrovascular dysfunction induced by slow-pressor AngII is mediated by ROS produced by the enzyme NADPH oxidase (Capone et al., 2011). It is therefore likely that AngII and ET1 activate NADPH oxidase resulting in increased vascular ROS production, a possibility supported by extensive in vivo and in vitro evidence linking these peptides to NADPH oxidase-dependent ROS production (Bedard and Krause, 2007; Faraci, 2011). The mechanisms by which vascular oxidative stress alters the regulation of the cerebral circulation remain to be defined. Our previous data indicate that the dysfunction induced by acute AngII hypertension is mediated by vascular nitrosative stress(Girouard et al., 2007), but it remains to be determined whether a similar mechanism is also involved in slow-pressor AngII hypertension.