The kallikrein-kinin system (KKS) has multiple functions in the maintenance of cardiovascular system, including vasodilation of peripheral and coronary arteries (Madeddu et al., 1997; Heitsch, 2003), inhibition of endothelin release (Momose et al., 1993), natriuresis and diuresis (Lortie et al., 1992), inhibition of superoxide production (Madeddu et al., 2007), release of nitric oxide (NO·), and synthesis of prostaglandins and endothelium-derived hyperpolarizing factor by endothelial cells (Vanhoutte, 2004; Alhenc-Gelas et al., 2011). Altogether, these studies support the notion that KKS acts as a unique modulator of the cardiovascular system under physiological and pathophysiological conditions (Regoli et al., 2012).

Kallikrein mRNA and protein expression have been identified in blood vessels, which indicate the presence of KKS components in vascular tissue (Saed et al., 1990). KKS biologically active peptides act through two different types of G protein-coupled receptors, B₁- and B₂-kinin receptors (B₁R and B₂R), that are expressed at different molecular levels. In the vasculature, both receptors have been implicated in the activation of a wide spectrum of vasoactive pathways (Tsutsui et al., 2000; Felipe et al., 2007; Regoli et al., 2012). Indeed, B₂R is ubiquitously expressed whereas B₁R is virtually absent in healthy tissues, but its expression is rapidly increased under certain pathological conditions (McLean et al., 2000). However, in spite of the degree of expression, there are compelling evidence supporting that even barely expressed, B₁R plays a physiological role and its activation induces vasoconstriction (Felipe et al., 2007) or vasorelaxation (Tsutsui et al., 2000) under unstressed conditions.

It is well-known that the endothelium-dependent vasodilatation in mouse aorta can be primarily attributed to NO· release as the result of endothelial and neuronal nitric oxide synthases (eNOS and nNOS) activity (Pohl et al., 2000; Capettini et al., 2010), which are highly regulated by changes in intracellular Ca²⁺ concentration, phosphorylation levels and, less known, by protein-protein interactions (Kone et al., 2003). Although previous studies have demonstrated molecular interactions between B₂R/eNOS and B₂R/nNOS (Ju et al., 1998; Golser et al., 2000), the functionality of such interactions in vascular tissue have not yet been described. Moreover, it has been shown that mice lacking B₁- or B₂-kinin receptors (B₁R^−/− or B₂R^−/−) exhibit impairments in acetylcholine (ACh)-induced vasodilation (Loiola et al., 2011). Therefore, based on the above considerations, the present study was designed to evaluate the participation of kinin receptors in ACh-induced vascular relaxation, focusing on protein-protein interactions involving kinin receptors with eNOS and nNOS.

The main findings of this study are as follows: (1) both B₁- and B₂-kinin receptors form heteromers with nNOS and eNOS in thoracic aorta; (2) B₁- and B₂-kinin receptors equally play an important role in the regulation of endothelium-dependent vasodilation of ACh through nNOS activity; (3) the oxidative stress is closely related with the onset of endothelial dysfunction in aorta of B₁R^−/− and B₂R^−/− mice.

Kinin receptors have been reported to play a pivotal role in numerous vascular responses (Berthiaume et al., 1997; Felipe et al., 2007). Although it remains controversial whether B₁R^−/− and B₂R^−/− mice are normotensive (Milia et al., 2001; Trabold et al., 2002; Abadir et al., 2003) or hypertensive (Madeddu et al., 1997, 1998; Emanueli et al., 1998, 1999), these mice exhibit increased susceptibility to develop hypertension in response to different stressor stimuli (Alfie et al., 1997; Madeddu et al., 1997; Duka et al., 2001). Furthermore, it was previously demonstrated that mice lacking B₁- or B₂-kinin receptors display a decreased NO· bioavailability, despite the increased enzymatic activity of the constitutive NOS isoform (Loiola et al., 2011). Thus, these findings indicate that the vascular NOS activity might be uncoupled in B₁R^−/− and B₂R^−/− mice. Moreover, whether ROS generation is involved, or not, in this vascular dysfunction remains unknown. Besides of these important issues, identification of the uncoupled NOS isoform represents another relevant question that could explain the underlying mechanisms and molecular basis of NOS uncoupling in B₁R^−/− and B₂R^−/− mice.

Ca²⁺-dependent NOS, which include eNOS and nNOS, are tightly regulated by several regulatory sites. Furthermore, new molecular regulators have been identified affecting their activities and, consequently, the amount of critical end-products for vascular function. Notably, protein-protein interactions have been proposed as an important mechanism that regulates the activity of NOS (Kone et al., 2003). Indeed, several studies have already described the functional relevance of physical interactions between NOS with a wide variety of structural and regulatory proteins such as, calmodulin (Rizzo et al., 1998), caveolin-1 (Ghosh et al., 1998), heat shock protein 90 (Russell et al., 2000), α1A adrenergic receptors (Pupo and Minneman, 2002), and angiotensin II receptor type 1 (AbdAlla et al., 2000). Although the heteromerization of B₂-kinin receptor with nNOS and eNOS was previously reported (Ju et al., 1998; Golser et al., 2000), this study is the first to demonstrate B₁R–eNOS/nNOS heteromerization. Moreover, although we are the first group to actually demonstrate such macromolecular protein complexes in native vascular tissue, if these interactions are functionally coupled to each other and, more importantly, whether this affects vascular function, it has remained unknown until now.

Here, we showed that eNOS and nNOS inhibition strongly decreases the ACh-induced vasorelaxation, confirming the large contribution of both constitutive NOS as the main source of NO· production. Intriguingly, pharmacologic antagonism or genetic ablation of B₁- or B₂-kinin receptors decreased the endothelium-dependent vasorelaxation response to ACh. Therefore, our results show that kinin receptors act as regulators of NOS activity in endothelial cells. To identify whether kinin receptors differently modulate eNOS and nNOS activities, the vasorelaxation response to ACh was assessed combining kinin receptors antagonists and NOS inhibitors. Our results demonstrated that B₁- and B₂-kinin receptors blockage fully abolished the relaxation upon non-selective NOS inhibition whereas nNOS inhibition did not potentiate the reduction of vasorelaxation induced by kinin receptors antagonism. In addition, we found similar pattern of vascular response in aorta from B₁R^−/− and B₂R^−/− mice. It is worth noting that non-selective NOS inhibition and B₁- or B₂-kinin receptors antagonists led to vascular contraction. These vascular responses further support that inhibition of the main vasodilation pathway favors contractile mechanisms, in which the oxidative stress is well-known to play a dysfunctional role in several vascular disorders (Johansson et al., 2005; Loria et al., 2014). Moreover, a reasonable explanation for selecting ACh as the agonist to trigger the vasorelaxation in the present study is the well-known NOS activation by muscarinic pathway whereas, the vascular response of agonists of kinin receptors is still controversial in aorta of different species, where it produces vasoconstriction (Levesque et al., 1993; Felipe et al., 2007) or vasodilation (Tsutsui et al., 2000). Furthermore, previous evidence has already demonstrated the absence of protein-protein interactions between kinin and muscarinic receptors (Willars et al., 1999). Importantly, we also validated the vascular dysfunction in B₁R^−/− and B₂R^−/− mice assessing the vasodilation induced by another well-known NOS activator, substance P. Taken together, our results show that B₁- and B₂-kinin receptors represent a functional regulatory domain of NOS activity. Moreover, our findings indicate that eNOS activity remained unaffected whereas the loss of nNOS function may represent a potential uncoupled activity.

Taking into account that B₁R and B₂R receptors are expressed in endothelial and smooth muscle layers (Leeb-Lundberg et al., 2005), kinin receptors may also affect the vascular relaxation by changing the responsiveness to NO· or, by acting on other targets resident in the vascular smooth muscle. To test this hypothesis, the vasodilator response to an exogenous NO· donor was used. Our results showed that kinin receptors did not affect the vasodilator response to NO· donor, which indicate a preserved smooth muscle function. Therefore, our findings strongly indicate that the endothelial dysfunction is the main cause of impaired vascular relaxation in aorta of B₁R^−/− and B₂R^−/− mice. Importantly, a cross-talk between B₁R and B₂R has been reported (Barki-Harrington et al., 2003), indicating an interdependence with each other for the proliferative response. Moreover, it has been also shown that B₁B₂R^−/− mice display protection against endotoxin-induced hypotension (Cayla et al., 2007) and obesity induced by high fat diet (Morais et al., 2015), while there are evidence showing amplified renal injury after ischemia/reperfusion (Kakoki et al., 2007) and increased nephropathy, neuropathy and bone mineral loss in Akita diabetic mice (Kakoki et al., 2010). Thus, in the present study we show that B₁B₂R^−/− mice present similar level of vascular dysfunction as found in mice lacking B₁R or B₂R, and therefore ruling out a possible synergistic response between these receptors in this mechanism. Although still under debate whether endothelial dysfunction is a primary cause or consequence of vascular diseases, such vascular dysfunction is generally associated with decreased NO· bioavailability and increased superoxide generation which ultimately lead to an impaired endothelium-dependent vasorelaxation (Vanhoutte et al., 2017).

Under physiological conditions, the activity of eNOS is able to produce NO·, whereas nNOS activity produces NO· and hydrogen peroxide (H₂O₂; Rosen et al., 2002; Gao et al., 2007). However, during certain pathological stimuli, the activity of eNOS and nNOS is drastically modified and becomes primarily uncoupled, producing a large amount of ROS, mainly, superoxide (Vasquez-Vivar et al., 1998; Weaver et al., 2005). Consistent with impaired vasodilation, acute blockage of kinin receptors markedly potentiate ACh-induced superoxide generation in aorta of WT mice. Additionally, we also demonstrated that B₁R^−/− and B₂R^−/− mice show a reduced NO· bioavailability associated with marked increase in intracellular levels of superoxide. Therefore, although the enhanced constitutive NOS activity in vascular tissue of B₁R^−/− and B₂R^−/− mice, our findings indicate that the uncoupled activity of nNOS represents a remarkable source of superoxide and, may be the major cause of the endothelial dysfunction in B₁R^−/− and B₂R^−/− mice.

Indeed, the oxidative stress is the major cause of endothelial dysfunction in several forms of vascular diseases, mainly via disruption of NO· signaling pathway (Drummond et al., 2011; Silva et al., 2016). Therefore, enhanced superoxide anion dismutation through the SOD mimetic can prevents the inactivation of NO· and formation of the highly reactive intermediate peroxynitrite (Weaver et al., 2005), and thus yielding augmented H₂O₂ concentration, an endothelium-dependent relaxing factor in aorta (Capettini et al., 2008, 2010). Accordingly, the left-shifted ACh-induced vasodilation in WT may be explained by the above factors. Moreover, the stronger enhancement of vascular relaxation induced by the SOD mimetic in B₁R^−/− and B₂R^−/− mice than in WT mice further support our hypothesis of oxidative stress involvement in the endothelial dysfunction. These findings are consistent with previous studies, in which the exogenous SOD mimetic restores endothelium-dependent vasodilation due increased NO· bioavailability (Fontana et al., 1999; Braga et al., 2015).

nNOS uncoupling has been implicated in several endothelium-dependent vascular disorders such as, in mesenteric artery of DOCA-salt hypertensive mice (Silva et al., 2016), thoracic aorta subjected to experimental atherosclerosis model (Capettini et al., 2011) and penile arteries of obese Zucker rat (Sánchez et al., 2012). Accordingly, deficient mice of B₁-kinin receptor has been described to aggravate atherosclerosis and the development of abdominal aorta aneurysms in ApoE^−/− mice under cholesterol rich-diet (Merino et al., 2009). B₂-kinin receptor deficient mice also demonstrate higher blood pressure in response to chronic excess of angiotensin II or dietary salt (Madeddu et al., 1997) and chronic mineralocorticoid excess (Emanueli et al., 1998). Moreover, pharmacologic blockade or genetic disruption of the B₂-kinin receptor accelerates the development of renovascular hypertension induced by clipping of the left renal artery (Madeddu et al., 1998). On the other hand, a recently study demonstrated that B₂-kinin receptor antagonist inhibited AngII-induced neutrophil activation and inflammatory phenotype in ApoE^−/− mice, thus suggesting B₂-kinin receptor antagonism as potential therapy for abdominal aortic aneurysm (Moran et al., 2016). Therefore, further studies are necessary in disease models to test the hypothesis whether the conformational rearrangement of protein-protein interactions, involving kinin receptors and NOS isoforms, is relevant to vascular dysfunction.

In summary, we provide data supporting a novel implication of kinin receptors in the endothelial dysfunction. We report here that B₁- and B₂-kinin receptors regulate the endothelium-dependent vasodilation of ACh through nNOS activity and indicate that molecular disturbance of short-range interactions between B₁- and B₂-kinin receptors with nNOS is involved in the oxidative pathogenesis of endothelial dysfunction.

TM participated in all steps of this study. GC and IJ performed experiments. SL, LC, VL, JC, EC, JLP, and JBP contributed to the experimental design, data analyses, data interpretation, and the preparation and revision of the manuscript.

This study was supported by Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, Brazil), PROMOB–Coordenação de Aperfeiçoamento de Pessoal de Nível Superior e Fundação de Apoio à Equipe e à Inovação Tecnológica do Estado de Sergipe (CAPES/FAPITEC/SE, Brazil), Fundação de Amparo à Pesquisa do estado de Minas Gerais (FAPEMIG/MG, Brazil). JC and EC are CNPq research fellow.