Reduced cerebrovascular remodeling and functional impairment in spontaneously hypertensive rats following combined treatment with suboptimal doses of telmisartan and ramipril: is less really more?

Abstract

Systemic arterial hypertension, which is the major risk factor for stroke [1], is accompanied by structural remodeling of cerebral arteries that initially serves to increase vascular resistance, thereby protecting the brain microvasculature from elevated mean arterial pressure and reducing the immediate threat of hemorrhagic stroke [2]. However, with structural remodeling of the cerebral vasculature comes reduced vasodilator capacity [2]. This not only increases the likelihood of cerebral ischemia and thus the incidence of ischemic strokes in hypertensive individuals, but also compromises the ability of cerebral vessels to support reperfusion after stroke, thereby worsening outcomes [3]. In this issue of Journal of Hypertension, Dupuis et al.[4] demonstrate that treatment of spontaneously hypertensive rats (SHRs) with a combination of ‘suboptimal’ doses of telmisartan (an angiotensin AT1 receptor antagonist) and ramipril [an angiotensin-converting enzyme (ACE) inhibitor], normalized mean arterial and cerebral arteriolar blood pressures and completely prevented cerebrovascular remodeling such that wall thicknesses and internal diameters of cerebral arteries were comparable to those of normotensive Wistar–Kyoto (WKY) rats. Importantly, at these suboptimal doses, each drug on its own only partially lowered systemic blood pressure, had minimal effects on cerebral arteriolar pressure, and was totally ineffective at preventing cerebral artery remodeling. Furthermore, combination but not monotherapy also improved the cerebral artery dilator response following acute induction of hypotension by controlled removal of blood from the femoral artery of SHRs. It is important to emphasize that the suboptimal doses of telmisartan and ramipril used by Dupuis et al. had relatively minor effects on blood pressure in SHRs (i.e. ≤50% reduction). By contrast, combination therapy with these ‘suboptimal’ doses lowered blood pressure in SHRs back to the normotensive levels of WKY rats. Thus, one obvious question is whether the additional benefits of combined versus monotherapy with ramipril and telmisartan were merely related to their markedly different effects on blood pressure. To shed some light on this it is worth comparing the current findings of Dupuis et al. with those from their earlier study looking at the effects of monotherapy with doses of telmisartan or ramipril that were fully effective at reducing elevated blood pressure in SHRs. At those higher doses, telmisartan and ramipril alone each appeared to be as effective as the combination therapy with suboptimal doses at preventing vascular remodeling and the impairment in the vasodilator responses to acute systemic hypotension in SHRs [5]. Thus, notwithstanding the limitations of comparing findings from separate studies, at face value there would appear to be no particular benefit, in terms of prevention of cerebral artery remodeling and functional impairment in SHR, of the combination therapy with suboptimal doses of telmisartan and ramipril over monotherapy with higher doses of either drug. An additional question from the current study by Dupuis et al. is whether the beneficial effects of combination therapy with suboptimal drug doses on cerebral artery structure and function would actually translate to a reduced incidence or improved outcome in ischemic stroke. Thus, a pertinent follow-up study might involve examining the effects of the different dosing regimes in relevant animal models of stroke. The most widely accepted model of experimental stroke involves the insertion of a surgical filament through the internal carotid artery of rodents until the tip occludes the middle cerebral artery, resulting in cessation of flow [6,7]. The filament is typically left in place for 30–60 min and then removed to restore blood flow. The impact of the ischemia–reperfusion insult can then be assessed functionally by a series of neurological scoring tests, and pathologically by post mortem assessment of brain infarct and edema volume. Indeed, a previous study that employed this model showed that chronic treatment of normotensive mice with a nonhypotensive dose of the AT1 antagonist, candesartan, reduced infarct volume and prevented neurological deficit 24 h after temporary occlusion of the middle cerebral artery. By contrast, that same study reported that an ACE inhibitor, enalapril, had no protective effects whether used at either a hypotensive or nonhypotensive dose [8]. Thus a logical extension of the present study could involve investigating the effects of chronic treatment with telmisartan and/or ramipril on outcomes after middle cerebral artery occlusion in SHRs, with the hypothesis being that the improvements in cerebrovascular structural and vasodilator properties afforded by combination therapy with suboptimal doses [4] or by monotherapy with higher doses of each drug [5] will result in enhanced reperfusion, greater salvage of brain tissue, and reduced neurological deficit, compared to vehicle treatment, or treatment with suboptimal doses of each drug alone. However, whereas models such as that described above make assessment of drug interventions on outcome after experimental stroke feasible, such models are, of course, not appropriate for investigating the effects of these treatments on stroke incidence (i.e. since the models result in stroke 100% of the time!). In fact, as it is particularly challenging to study the incidence of ischemic stroke in experimental models, it would be difficult to assess whether the current findings of Dupuis et al. translate to reduced incidence of ischemic stroke [4]. Were one to assume that the current and previous findings of Dupuis et al. do actually translate to a reduction in stroke incidence then they would seemingly be at odds with several large-scale clinical trials including HOPE, ONTARGET and TRANSCEND [9–11]. These trials, which examined the effects of telmisartan (80 mg) and/or ramipril (10 mg) on cardiovascular events, including stroke, in high-risk patients, showed that when used alone each drug was equally protective. Surprisingly, ONTARGET found that, despite more pronounced blood pressure lowering (i.e. a 2.4/1.4 mm Hg greater reduction) combination therapy with telmisartan and ramipril did not lead to a further reduction in the number of events. Combination therapy with conventional doses of telmisartan and ramipril was, however, associated with significantly more side effects than treatment with either drug alone. Thus, in light of the current observation in SHRs that combination therapy with suboptimal doses was as effective at reducing blood pressure and preventing cerebrovascular remodeling and dysfunction as treatment with conventional doses of each drug alone, it would be interesting in future clinical trials to compare the adverse events profile of each treatment regime. In addition to the cerebrovascular protection afforded by their blood pressure-lowering effects, it is plausible that telmisartan and ramipril mediate some of their beneficial actions locally by modulation of the components of the renin–angiotensin system (RAS) that are intrinsic to cerebral vessels. In this regard, a particularly interesting observation in the present study was that whereas in untreated SHRs, and in animals treated with either telmisartan or ramipril alone, direct application of Ang II to cerebral vessels in vivo via a cranial window elicited a vasoconstrictor response, in rats treated with the combination of suboptimal doses of each drug, Ang II was a powerful and potent (i.e. near maximal response at 10−9 mol/l) vasodilator. There are several explanations for how chronic treatment with the combination of ramipril and telmisartan could have converted Ang II from a vasoconstrictor to a vasodilator in the cerebral vasculature of SHRs and these will be discussed below, in light of a brief update on the current state of knowledge surrounding the various components of the RAS. The RAS is responsible for the production of a series of angiotensin (Ang) peptide fragments which act in either an endocrine, paracrine or autocrine fashion by stimulating one of four receptor subtypes. The initial step in the pathway involves the release of the enzyme, renin, from the kidneys in response to stimuli such as reduced intrarenal blood pressure or decreased delivery of Na+ and Cl− to the macula densa [12]. Renin catalyzes the conversion of angiotensinogen, into the relatively inert decapeptide Ang I which may in turn be converted to Ang II by removal of two C-terminal residues by the carboxypeptidase, ACE. The effects of Ang II on the cardiovascular system are primarily mediated by stimulation of either AT1 or AT2 receptors [12]. AT1 receptors, which are highly expressed on vascular smooth muscle cells (VSMCs), are responsible for the ‘classical’ actions of Ang II on the vessel wall including vasoconstriction, VSMC proliferation, increased production of extracellular matrix proteins and activation of NADPH oxidase, all of which are generally considered to be ‘disease-promoting’ [13]. By contrast, the AT2 receptor, which has similar affinity for Ang II as the AT1 receptor, is expressed on endothelial cells in addition to VSMCs, and is linked to vasoprotective effects such as vasodilatation and suppression of VSMC proliferation and oxidative stress [14,15]. Stimulation of AT2 receptors in the vessel wall has also been shown to cause the release of bradykinin from kininogen, which acts in an autocrine/paracrine fashion on endothelial B2 receptors causing activation of endothelial nitric oxide synthase (eNOS) and release of the powerful vasodilator, anti-inflammatory and antithrombotic molecule, nitric oxide [12,16]. In cerebral arteries, bradykinin may also normally elicit dilatation via generation of cyclooxygenase-derived reactive oxygen species (ROS) [17]. In these ways, AT2 receptors may be considered as ‘physiological’ inhibitors of AT1 receptors. Furthermore, given the comparable affinities of each receptor subtype for Ang II, the ratio of AT1: AT2 receptor expression in the vascular wall is likely to be a major determinant of whether Ang II is damaging or protective. Several additional angiotensin peptide fragments have been discovered more recently and suggested to mediate at least some of the cardiovascular effects of the RAS. The most notable of these is Ang 1–7, which may be formed directly from Ang I by the actions of neutral endopeptidase, or from Ang II by the recently discovered homolog of ACE, ACE-2 [12,16]. Ang 1–7 has a higher affinity for AT2 receptors than AT1 receptors, although its affinity at the AT2 receptor is still substantially lower than that of Ang II. More recently it has been postulated that the MasR is the main target of Ang 1–7, although the significance of this receptor type in the context of cardiovascular disease is largely unknown. Ang II can also be metabolized by aminopeptidase A into the hexapeptide fragment Ang 2–8 (also known as Ang III), which is generally regarded as a less potent agonist than Ang II of both AT1 and AT2 receptors but which may be in fact be the major effector peptide of the RAS in the brain. Ang III may, in turn, be converted into an Ang 3–8 fragment (Ang IV) by aminopeptidase B [12]. Ang IV, which is a weak agonist at AT1, AT2 and MasR receptors, is thought to mediate its effect by stimulation of a fourth binding site. Although originally designated as the ‘AT4’ receptor, this site was recently shown to be identical to the enzyme insulin-responsive aminopeptidase (IRAP). Like the MasR receptor, little is known of the cardiovascular relevance of IRAP, although we have provided some recent evidence to suggest that activation of IRAP confers protection against the development of atherosclerosis in hypercholesterolemic apolipoprotein E-deficient mice [18]. The major components of the RAS are summarized in Fig. 1.Fig. 1Coming back to the issue of how Ang II application to a cranial window in SHRs treated with a combination of telmisartan and ramipril elicits vasodilatation, we agree with the authors that the ‘mechanism behind the… effect (likely) involves the AT2 receptor’ for reasons outlined in the next paragraph. However, it is important to note that there are several other, although less likely, possibilities and only further experimentation will definitively test which of these provide the actual explanation. Furthermore, Dupuis et al. do not offer an explanation as to why the vasodilator response to exogenous Ang II was only observed following combined treatment with telmisartan and ramipril and not after treatment with either drug alone. Hence the following section outlines what we perceive to be the logical next steps to clarify whether Ang II elicits cerebral vasorelaxation in SHRs treated with a combination of telmisartan and ramipril. We also provide a hypothesis to explain why this vasodilator effect is not observed in untreated SHRs, or in SHRs treated with ramipril or telmisartan alone, again with a view to stimulating further experimentation. The first issue that needs to be addressed in order to explain why Ang II is a potent dilator of cerebral vessels only in SHRs treated with the combination of suboptimal doses of telmisartan and ramipril is which angiotensin receptor subtype mediated the response. Given that chronic systemic treatment with telmisartan, even at a suboptimal dose, was likely to have resulted in some degree of antagonism of AT1 receptors in the cerebral vessels exposed in the cranial window preparation; the relative affinities of Ang II for the different receptor subtypes (i.e. AT1=AT2>>AT3>>>AT4 [14]); and evidence from previous studies that AT2 receptor stimulation elicits vasodilatation in various vascular beds via the release of bradykinin and nitric oxide leading to elevated cGMP levels in VSMCs [19], one would predict that the vasodilator response observed by Dupuis et al. is most likely to be AT2-dependent. Thus, the response should be inhibited – and potentially converted back into a contraction – by acute addition to the cranial window of either a selective antagonist of the AT2 receptor (e.g. PD123319 [20]), or possibly by an antagonist of the bradykinin B2 receptor (e.g. incatibant), or an eNOS inhibitor (e.g. L-NAME). Nonetheless, one cannot exclude completely the possibility that the amounts of Ang II added to the cranial window in the study by Dupuis et al. (i.e. 10−9 and 10−6 mol/l) were sufficient to overcome the chronic AT1 receptor antagonism afforded by systemic treatment with telmisartan and thus activate signaling events downstream of this receptor subtype. Hence there may be justification (particularly if an AT2 antagonist had no effect) to examine the effect of ‘topping up’ the level of AT1 antagonism achieved by chronic treatment with telmisartan, by direct application of a higher concentration of this drug to the cranial window preparation. Indeed, we have previously reported a mechanism whereby AT1 receptor stimulation could lead to vasodilatation of cerebral vessels. Specifically we showed that cerebral arteries in rats appear to be unique from systemic vessels in that they utilize NADPH oxidase-derived ROS as an endogenous vasodilator mechanism [21,22]. Although it is well established that Ang II is a powerful activator of NADPH oxidase and that this effect is mediated downstream of AT1 receptor stimulation, we demonstrated that inhibition of NADPH oxidase activity in basilar arteries isolated from normotensive Sprague–Dawley rats converted Ang II from having relatively little impact on vascular tone to it becoming a potent vasoconstrictor and concluded that activation of NADPH oxidase and subsequent release of NADPH oxidase-derived ROS is a novel mechanism of vasodilatation in cerebral vessels that may offset the ‘classical’ vasoconstrictor actions of Ang II in this vascular bed [21]. It is also possible that, following its application to the cranial window preparation in SHRs, Ang II was rapidly converted to a MasR and/or IRAP agonist by the actions of ACE2 and/or aminopeptidase B. Although there is no information surrounding the effects of MasR activation on cerebral artery tone (or indeed in vessels from other vascular beds), there is evidence that Ang IV increases cerebral blood flow after subarachnoid hemorrhage in rats, presumably by dilating cerebral vessels via stimulation of IRAP [23]. Thus, it would be interesting to examine whether an inhibitor of aminopeptidase B activity, such as bestatin, influences cerebral vasodilator responses to Ang II in SHRs treated with telmisartan and ramipril, and also whether the direct application of Ang IV to the cranial window of these animals evokes a vasodilator effect. Whichever of the above scenarios proves to be the actual mechanism underlying the vasodilator effect of Ang II reported by Dupuis et al., we are still left to ponder why Ang II was a vasodilator only in SHRs treated with the combination of suboptimal doses of telmisartan and ramipril, and not in untreated SHRs or those animals treated with either drug alone. As mentioned above, previous literature indicates that AT2 receptor stimulation is the most likely explanation for the vasodilator response. Thus, again for the purpose of forming a framework against which to design future studies Fig. 2 outlines a working hypothesis that we would like to put forward to explain the current findings of Dupuis et al.Fig. 2In conclusion, comparisons between the current [4] and previous findings [5] of Dupuis et al. in SHRs indicate that combination therapy with suboptimal doses of telmisartan and ramipril is at least equivalent to monotherapy with higher doses of each drug at reducing elevated blood pressure. However, this has not been established in humans and thus there may be a rationale to directly compare the antihypertensive effects of these different dosing regimes in future clinical trials. Furthermore, although combination therapy with suboptimal doses of telmisartan and ramipril was effective at preventing structural remodeling and functional impairment in the cerebral vasculature of SHRs, it remains unclear whether these effects would extend to a reduction in stroke risk, as has been established clinically for telmisartan and ramipril when used alone at conventional therapeutic doses. Thus, much work needs to be done before we can answer the question posed at the beginning of this discussion, that is ‘is less really more’ when it comes to protection against cerebrovascular remodeling and incidence of ischemic stroke with telmisartan and ramipril? Acknowledgements Financial support: We thank the National Health and Medical Research Council of Australia for financial support in the form of fellowships to K.B. (465150), C.G.S. (606472) and G.R.D. (465109).