Cardiac and Vascular α1-Adrenoceptors in Congestive Heart Failure: A Systematic Review

As heart failure (HF) is a devastating health problem worldwide, a better understanding and the development of more effective therapeutic approaches are required. HF is characterized by sympathetic system activation which stimulates α- and β-adrenoceptors (ARs). The exposure of the cardiovascular system to the increased locally released and circulating levels of catecholamines leads to a well-described downregulation and desensitization of β-ARs. However, information on the role of α-AR is limited. We have performed a systematic literature review examining the role of both cardiac and vascular α1-ARs in HF using 5 databases for our search. All three α1-AR subtypes (α1A, α1B and α1D) are expressed in human and animal hearts and blood vessels in a tissue-dependent manner. We summarize the changes observed in HF regarding the density, signaling and responses of α1-ARs. Conflicting findings arise from different studies concerning the influence that HF has on α1-AR expression and function; in contrast to β-ARs there is no consistent evidence for down-regulation or desensitization of cardiac or vascular α1-ARs. Whether α1-ARs are a therapeutic target in HF remains a matter of debate.


Introduction
The sympathetic nervous system primarily regulates cardiac function via β-adrenoceptors (β-ARs) [1,2], whereas α 1 -adrenoceptors (α 1 -ARs) play an important role in the regulation of vascular tone [3]. Yet, α 1 -ARs are also expressed in various cell types of the heart and can contribute to the regulation of cardiac contraction [4,5], rhythm [6,7] and hypertrophy [8]. However, the role of cardiac α 1 -ARs appears to be species-dependent with a considerable contribution in rodents but a much smaller (if any) contribution in the healthy human heart [1]. On the other hand, cardiac expression of other subtypes such as β 3 -ARs also appears to be lower in the human heart than in that of other species, but such receptors become up-regulated in congestive heart failure (CHF) [9].
The sympatho-adrenal system is activated in acute and chronic heart failure (HF). Accordingly, plasma levels of noradrenaline and its co-transmitters are increased in CHF [10,11]. This appears to correlate with a higher risk of mortality [12]. In line with this, patients with reduced left ventricular function appear to be more prone to post-operative vasoplegia that additionally increases the mortality and morbidity risk [13,14]. The regulation of β-ARs in CHF has been reviewed extensively and appears distinct for the three β-AR subtypes [1]. There are three α1-AR subtypes: α1A-, α1B-and α1D-ARs which have been identified in both animal and human hearts at the mRNA level [1,39,40]. All studies agree that α1A-AR is the most abundant subtype in the heart but the presence of α1B-and α1D-AR mRNA were also confirmed [1].

mRNA Expression
There are three α 1 -AR subtypes: α 1A -, α 1B -and α 1D -ARs which have been identified in both animal and human hearts at the mRNA level [1,39,40]. All studies agree that α 1A -AR is the most abundant subtype in the heart but the presence of α 1B -and α 1D -AR mRNA were also confirmed [1]. Although animal hearts express more α 1 -AR in the heart, human and rodent hearts are reported to locate α 1 -AR subtypes similarly, as α 1A -and α 1B -subtypes are predominantly found in myocardium whereas the α 1D -subtype is expressed in coronary arteries [40,41]. Myagmar et al. claimed that α 1A -ARs are present in a myocyte subset while α 1B -ARs are in all cells [42]. Several studies reported changes in α 1 -AR subtype mRNA levels in CHF, but the results are heterogeneous (Table 1). Table 1. Summary of the changes in cardiac mRNA studies for total α 1 -adrenoceptors (AR) and their subtypes in chronic heart failure.
In human failing hearts, Fischer et al. demonstrated that mRNA of α 1A -and α 1B -ARs in the left ventricle was increased while no change in mRNA levels was observed in other cardiac regions [43]. Similarly, Jensen et al. reported that α 1A mRNA was increased in failing left ventricle LV and tended to increase in failing RV, but there was no change in α 1B and α 1D mRNA in any region [41]. On the other hand, Monto et al. did not confirm an altered expression of the three subtypes in human failing hearts; however, the relative abundance increased when considering the reduced β-AR expression [17]. They suggested that expression of α 1A -ARs in the LV correlates with the left ventricular ejection fraction (LVEF) suggesting that this subtype may contribute to maintain cardiac inotropy in the failing heart. Moreover, a study in a rat coronary ligation model of CHF reported α 1D -AR mRNA to be decreased; a numerical reduction of α 1A -and α 1B -AR mRNA did not reach statistical significance [2].

Protein Expression
Cardiac α 1 -ARs were first identified at the protein level in rats by Williams and Lefkowitz [44] and thereafter confirmed in many species. Their quantitative expression at the protein level ranges from a relatively high expression in rodents to a low expression in humans, where their role appears considerably lower than that of β-ARs [1,39,40]. Despite numerous studies, data on α 1 -AR binding properties in human HF have remained inconsistent ( Table 2). The density of α 1 -ARs was reported as increased [45][46][47][48], unaltered [41,45,[49][50][51][52], or decreased [43,53]. Table 2. Summary of cardiac α 1 -adrenoceptor (AR) protein expression in chronic heart failure.

Binding
Species Tissue α 1A -AR α 1B -AR α 1D -AR Total α 1 -AR Böhm et al. [49] human papillary n/a n/a n/a X Bristow et al. [50] human LV and RV n/a n/a n/a X Corr et al. [36] cat ventricle n/a n/a n/a ↑ Dixon et al. [54] rat LV n/a n/a n/a ↑ Erdmann and Böhm [51] human papillary n/a n/a n/a X Fischer et al. [43] human LV, RV, LA, RA ↓ ↓ ↓ n/a Gopalakrishnan et al. [32] rat LV and RV n/a n/a n/a X Hwang et al. [45] human explanted hearts n/a n/a n/a ↑ (in ICM) X (in DCM) Jensen et al. [41] human LV and RV X (↑ in RV) X undetected X Karliner et al. [33] GP whole heart n/a n/a n/a ↑ Limas et al. [52] human LV and RV n/a n/a n/a ↓ (in severe HF) X (in mild) Litwin et al. [25] rat LV n/a n/a n/a X Maisel et al. [34] GP LV n/a n/a n/a Meggs et al. [26] rat LV and RV n/a n/a n/a X Re et al. [37] dog LV, RV, LA, RA n/a n/a n/a ↓ Shi et al. [53] human LV ↓ X n/a ↓ Sjaastad et al. [55] rat LV and RV n/a n/a n/a X Steinfath et al. [46] human LV n/a n/a n/a ↑ Watanebe et al. [35] hamster whole heart n/a n/a n/a Slight↓ Zhao et al. [30] rat RA and LV n/a n/a n/a ↓ Rowley et al. [56] rat atria n/a n/a n/a X Vago et al. [47] human LV n/a n/a n/a ↑ GP: guinea pig, ↑: increased; ↓: decreased; X: unchanged; n/a: no data available; LV, left ventricle; RV, right ventricle; LA, left atrium; RA, right atrium.
However, even a lack of change in α 1 -AR protein expression may be biologically relevant for two reasons: Firstly, it is interesting to note that expression does not consistently decrease despite exposure to elevated plasma catecholamines [10,11]. However, studies outside the cardiovascular system have found that prolonged agonist exposure does not always lead to down-regulation of α 1 -AR subtypes and in some cases can even be associated with an up-regulation [16]. Secondly, an unaltered expression of α 1 -ARs in face of a consistently found down-regulation of β-ARs [1,39] implies a change in the relative role of the two adrenoceptor sub-families, which in its own right may have implications for the regulation of cardiac function [50,55].
A possible explanation for the controversial results on the regulation of cardiac α 1 -AR mRNA and protein is heterogeneity based on various etiologies of CHF. For instance, it was reported that α 1 -AR protein was increased in ischemic human cardiomyopathy but remained constant in dilated cardiomyopathy within the same study [45]. Similarly, the degree of HF may contribute to heterogeneous findings. While α 1 -AR density was not changed in a mild-to-moderate heart failure group, it was decreased in severely failing human hearts [52]. Finally, the change in α 1 -AR density may depend on the cardiac region under investigation: α 1A -and α 1B -AR protein were not changed in failing left ventricle, but α 1A -AR binding was increased in right ventricle [41].
Studies in rats mainly found that α 1 -AR density remained constant after myocardial infarction [25,26,32,55,56], which is in line with the lack of a consistent change in human CHF (Table 2). However, an increase was also reported in a rat model of myocardial infarction [54]. Noradrenaline infusion, on the other hand, resulted in a decrease in α 1 -AR density in rats [30], demonstrating that cardiac α 1 -ARs are in principle susceptible to agonist-induced down-regulation.
There is also a substantial number of studies investigating the changes in cardiac α 1 -ARs using different animal models. Cardiomyopathic hamster [35] and dog [37] hearts were shown to have decreased α 1 -AR concentrations. In contrast, an increase in cardiac α 1 -AR density was demonstrated in a guinea pig model of HF caused by aortic constriction [33]. In addition, increase α 1 -AR densities were reported in cat [36] and guinea pig [34] ischemia models following coronary ligation.

α 1 -AR Subtypes
For the interpretation of subtype expression data, there is a technical issue between α 1 -AR mRNA and protein levels. While mRNA of all α 1 -AR subtypes were detected in animal and human hearts, binding studies demonstrated the presence of only α 1A -and α 1B -subtypes, but not α 1D -AR [39,41,57]. Lack of α 1D -AR expression at the protein level has been shown in many tissues expressing corresponding mRNA [58].
Animal species do not only differ from humans by expressing more α 1 -AR in the heart but also by exhibiting a different subtype expression profile. For instance, rat [59] and mouse [58] hearts were reported to express α 1A /α 1B -AR in an approximate 20/80 ratio. Moreover, previous studies suggest that the three subtypes mediate different effects in the heart. α 1A -AR has been reported to mediate positive inotropy in right ventricles of human hearts and this response has been found to be reduced only slightly in failing hearts. Thus, α 1A -subtype-mediated inotropy is suggested to be significant source of inotropic support [60]. However, in nonfailing murine hearts, stimulation by the α 1A -AR selective agonist, A61603 or the non-subtype-selective α 1 -AR agonist, phenylephrine mediated negative inotropic effects in right ventricle which was switched to positive inotropy in failing hearts [29]. On the other hand, the inotropic response mediated by α 1B -subtype remained negative and unaltered. In the light of these findings, the authors suggested that α 1A -subtype may play a role in improving contraction in RV failure and might be a therapeutic target. Another study using transgenic mice with heart-specific overexpression of α 1A -AR displayed enhanced inotropy after myocardial infarction [61]. Moreover, α 1D -AR mRNA decreased in rats after coronary ligation; however, this study was inconclusive because there was a non-significant trend for reduced mRNA of α 1A -and α 1B -ARs [2]. Thus, the authors proposed that a decrease in α 1D -AR might be a protective response of the failing heart. However, the relevance of these data remains unclear because studies in several mouse tissues have demonstrated that many tissues express α 1D -AR at the mRNA but not at the protein level [58].
Patients who receive left ventricular assist device to unload the failing heart and improve cardiac function were shown to have a redistribution of α 1A -AR from a peri-myocytic to an intra-myocytic location along with a rise in the overall density of receptors [62]. On the other hand, α 1A -AR subtype was decreased in human dilated cardiomyopathy and this result was concluded to be consistent with the theory that α 1A -AR agonism may be cardioprotective [53]. Fischer et al. demonstrated that α 1Aand α 1B -AR mRNA levels in left ventricle of human failing hearts were increased, although α 1A -, α 1Band α 1D -AR binding sites were decreased [43]. They suggested that a decrease of α 1B -AR can have cardioprotective effects while a decrease of α 1A -AR is more likely the result of impaired heart function.

Signaling Pathways Involved
A schematic representation of α 1 -AR signal transduction is shown in Figure 2. The prototypical signaling pathway of α 1 -ARs involves coupling to G q with subsequent activation of phospholipase C (PLC), which releases Ca 2+ from intracellular stores and activates protein kinase C (PKC) [63]. However, cardiac α 1 -ARs can also couple to G i [5,7], G 12/13 [64], even G s proteins [65] and G h which is a GTP-binding protein with Transglutaminase activity [66] and to pathways such as inhibition of adenylyl cyclase [67], modulation of Ca 2+ , K + and Clchannel activity [7,68,69], ras activity [70] or activation of various mitogen activated protein kinases [71]. Most of these pathways have also been studied in human CHF or CHF animal models [72].
(PLC), which releases Ca 2+ from intracellular stores and activates protein kinase C (PKC) [63]. However, cardiac α1-ARs can also couple to Gi [5,7], G12/13 [64], even Gs proteins [65] and Gh which is a GTP-binding protein with Transglutaminase activity [66] and to pathways such as inhibition of adenylyl cyclase [67], modulation of Ca 2+ , K + and Clchannel activity [7,68,69], ras activity [70] or activation of various mitogen activated protein kinases [71]. Most of these pathways have also been studied in human CHF or CHF animal models [72]. Activation of G q -signaling pathways mediates the development of hypertrophy and the transition to HF [73]. Overexpression of G q was demonstrated to result in marked hypertrophy [74] and HF [75]. Moreover, inhibition of G q -mediated signaling was shown to inhibit cardiac dysfunction [76,77]. On the other hand, multiple studies in humans and several animal species have shown that HF is accompanied by an increased expression of G i at the mRNA and protein level, whereas a reduced expression of G s has been found less consistently [78]. Reduced GTP binding with Transglutaminase activity of Gα h were found in human HF [45]. α 1 -AR coupling to a GTP binding protein was detected only in pathological, but not in healthy human hearts [47]. α 1 -AR-mediated phosphoinositol turnover is not changed in both animal and human studies when accompanied by an unaltered receptor density in human [50] and rat [56] HF. On the other hand, in a rat model of myocardial infarction, an increase in phosphoinositol turnover was reported along with an increase in G-protein content and an upregulation of α-skeletal actin which suggests an involvement of α 1 -ARs in the reactive hypertrophic response [26]. α 1 -AR stimulation increases both the intracellular Ca 2+ concentration, which activates muscle contraction, but also the Ca 2+ sensitivity of cardiomyocytes. α 1 -AR-mediated Ca 2+ -transients and phosphorylation of cardiac myosin regulatory light chain were increased in a mouse right ventricular HF model [29]. The increase in Ca 2+ sensitivity expected from these observations was reported in a canine HF model, apparently occurring via the G q -RhoA-ROCK signaling pathway [38]. Similarly, the switch from negative to positive inotropy observed in mouse failing right ventricles (see below) was claimed to be mediated through a pathway involving increased myofilament Ca 2+ sensitivity [27].  (Table 3) [25,55,56,60]. The positive inotropic effect of α 1 -AR stimulation involves multiple subtypes, but mostly occurs via α 1B -AR in rodents; when these become inactivated in vitro, the α 1A -component becomes stronger and compensates for the loss of α 1B response resulting in overall maintained contractile responses [79]. However, depending on experimental conditions α 1 -AR stimulation can also have negative inotropic effects (Table 3) [29]. Moreover, stimulation of cardiac α 1 -AR can increase heart rate [56], but systemic stimulation can reduce heart rate secondary to Cells 2020, 9, 2412 8 of 23 an increase in total peripheral resistance and activation of the baroreflex [80], which can mask direct chronotropic effects of α 1 -AR stimulation. Table 3. Summary of changes in cardiac α 1 -adrenoceptor (AR)-mediated responses in chronic heart failure.

Species
Preparation The chronotropic and inotropic responses to the α 1 -AR agonist phenylephrine were decreased in isolated right and left atria of rats with myocardial infarction [56]. As this was not accompanied by alterations of receptor density or phosphoinositol turnover, this may reflect changes of post-receptor signaling (see above). In contrast, several studies in isolated papillary muscle from humans [49,51] or rats [25,55] reported unaltered inotropic responses in CHF, which was in line with unaltered receptor density reported in these studies. The positive inotropic response to α 1A -AR stimulation in human right ventricle also exhibited little change in HF in two studies [60,81]. In contrast, the phenylephrine-induced positive inotropic response in human CHF was abolished in another study [46]. In mouse models, negative inotropic effect observed in control animals was switched to positive inotropy in the failing right ventricles [29]. However PE-mediated positive inotropy in LV of nonfailing hearts remained unchanged in failing mouse hearts after myocardial infarction [27]. Cardiomyopathic hearts of Syrian hamsters were shown to have increased inotropy accompanied by increased Ca 2+ levels and pertussis toxin-sensitive G-proteins [82]. Thus, with few exceptions, most studies reported little change in inotropic responses in the failing heart across species and models. This may nonetheless be biologically relevant because it would indicate an α 1 /β-AR balance in light of the consistently found reduction in β-adrenergic function in CHF [50,55].
Moreover, there are opposing views on the role of α 1 -AR subtypes on hypertrophy. While some suggest that α 1A -AR were found to mediate hypertrophy development in neonatal rat cardiomyocytes in vitro [8], transgenic overexpression of α 1B -AR caused a hypertrophic response resulting in HF, while α 1A -AR overexpression did not affect hypertrophy and did not hasten development of HF [29,83].

mRNA Expression
Multiple studies have explored the expression of α 1 -AR subtype mRNA in arteries and veins of various species including rat [84][85][86][87][88], rabbit [89][90][91] and human [89,[92][93][94]. It appears from these and many other studies, that all three α 1 -AR subtypes are expressed in the vasculature of these species in a species-and vascular bed-dependent manner. This conclusion is in line with studies on the role of α 1 -AR subtypes in functional studies [3]. Our search did not identify any animal studies reporting on the regulation of vascular α 1 -AR subtype mRNA in CHF. However, it is likely that such regulation occurs because CHF is characterized by increased circulating catecholamine levels and the endogenous agonist noradrenaline has been shown to transiently down-regulate α 1 -AR mRNA in rabbit aortic smooth muscle cells whereas the down-regulation of α 1 -AR protein was longer lasting (at least 96 h) [95]. Moreover, CHF can be associated with hypoperfusion and hypoxia of tissues and vascular α 1 -AR mRNA expression can be regulated by hypoxia [96]. Vascular α 1 -AR mRNA expression is also regulated by age [97]. HF did not appear to affect the total levels of α 1 -AR mRNA in human coronary arteries [98]. On the contrary, the use of β-blockers was associated with a significant decrease in both α 1D and total α 1 -AR mRNA levels. α 1D -AR appeared to be the principal receptor in human epicardial coronary arteries, which resembles the mouse coronary arteries [99,100].

Protein Expression
Quantitative, ligand-binding studies have been performed in order to assess the α 1 -AR distribution in the various vascular beds of different species. Existing discrepancies might be due to interspecies differences, variant α 1 -AR subtype distribution in the peripheral vasculature, differences in models of HF, for example pacing-induced HF versus myocardial infarction, or even due to poorly reproducible results.
Such studies have been performed in animals and humans without coexisting diseases in order to assess the physiological α 1 -ARs at the protein level. In humans a considerable difference in α 1 -AR protein expression has been demonstrated between somatic (mammary artery) and splanchnic arteries [97]. However, this was not true for veins, in which the α 1 -AR density seemed to be independent from differences in vessel diameter. This might be explained by the importance of splanchnic circulation, especially during a situation of hypovolemic shock. Sympathetic stimulation of α 1 -ARs in those vascular territories leads to maintenance of blood flow and O 2 delivery to vital organs. On the other hand, constriction of conduit arteries and veins contribute to the preservation of blood pressure and cardiac afterload. In human epicardial coronary arteries 8.7 fmol/mg protein of total α 1 -ARs were found, of which 75% constituted of the α 1D subtype, almost twice as much as in the left ventricle [41,98].
Considering that progression of HF is associated with a decrease in β-AR protein in the heart, one could argue that similar findings would be expected for α 1 -ARs. On the contrary, α 1 -AR density has been reported to be increased, yet not statistically significant, in studies with myocardial infarction-induced CHF in rats [101,102]. However, the lack of statistical significance of this change leads to rather inconclusive results.

α1-AR Subtypes
All three subtypes have been identified in the peripheral vasculature of various species, however the subtype distribution varies between animals and humans, between different vascular beds and with age. For instance, α 1B and α 1D subtype mRNAs have been detected in human aorta [93] and are considered to mediate vascular constriction. A controversy exists in the rat aorta since some studies have identified only the subtype α 1D [103][104][105], while others report both α 1B and α 1D -ARs [101]. The α 1B subtype has been recognized to mediate vasoconstriction after sympathetic stimulation in the mouse, rabbit and dog aorta [106][107][108]. Similarly, mesenteric responses in humans have been found to be mediated by all three α 1 -AR subtypes [109], whereas in rats mesenteric vascular constriction is mediated by α 1D and α 1B subtypes in superior mesenteric artery and the mesenteric resistance arteries respectively [101,104]. In the small mesenteric arteries the subtype α 1A has also been identified [101,110]. The same receptor was found to mediate contractile responses in rat small renal arteries [111].
Similarities in vascular tone regulation between different species also exist. As it was mentioned before, α 1D is recognized as the main subtype that mediate coronary contraction in humans [98] and mice [99,100] and the same has also been reported in rats [105]. In this latter study, it was additionally shown that CHF did not modify the function of α 1D -AR, neither in the carotid artery nor in the aorta, whereas the α 1D -AR-mediated vasoconstriction was enhanced in older animals.
Furthermore, blood pressure regulation in mice is considered to be achieved via the stimulation of all three subtypes. This is confirmed in knockout mice studies in which deletion of any subtype results in a decrease in resting blood pressure [106,112,113].
The heterogeneous α 1 -AR population in the various vascular beds might be modified depending on the disease state. In support of this, antagonist studies have explicitly displayed an alteration of the antagonism against α 1 -ARs before and after the onset of CHF [114][115][116][117] and showed changes from competitive to non-competitive antagonism, or the other way around, which, however, was also related to the specific agonist and vessel studied. Distinct patterns of antagonism have been identified in different disease states and in different vessels. The specific subtype populations, however, and their involvement in the vascular responses in different disease settings are still to be elucidated.

Signaling Pathways Involved
Changes in α 1 -AR responsiveness due to chronic exposure of the peripheral vasculature to the endogenously released catecholamines might also derive from alterations in signal transduction. These changes might be due to modification of Ca 2+ sensitivity and/or other changes in the signaling pathways involved. For example, the continuous exposure of smooth muscle cells from rabbit aorta to α 1 -AR agonists led to a major loss in sensitivity, which did not seem to be due to down-regulation of α 1 -ARs, but rather to blunting of phosphatidylinositol turnover [118].
Disease progression might also modify the dependency of α 1 -AR-mediated vascular contraction on extracellular Ca 2+ which seems to increase in end-stage HF [119,120]. More precisely, the existence of a high affinity component of the α 1 -AR-mediated responses was identified which is inhibitable by a PKC inhibitor, polymyxin B, and which becomes apparent with the progression of HF. Interestingly, in this study, the dorsal pedal artery presented lower dependency on PKC activation than the saphenous vein, which underlines the existing differences between separate vascular beds [120]. Likewise, after investigating the α 1 -AR in mesenteric resistance arteries from CHF rats, it was confirmed that Ca 2+ plays a more crucial role in the vascular contractility as CHF develops [121]. Furthermore, it has been demonstrated that the diminished contractile response of rat mesenteric resistance arteries to phenylephrine and Ca 2+ is not due to changes in Ca 2+ sensitivity, nor in intracellular Ca 2+ release, but rather to changes in the α 1 -AR-triggered Ca 2+ influx [102].
Other studies, however, have identified a modification in the sensitivity of α 1 -AR vasoconstriction to Ca 2+ . For instance, in CHF rats an increased responsiveness of CHF femoral resistance arteries was identified which appeared to be the result of first, a RhoA/Rho kinase-depended increase in sensitivity to Ca 2+ and second, of the IP3-induced increase in intracellular Ca 2+ [122]. On the contrary, this arterial sensitivity to Ca 2+ was reported to be decreased in femoral resistance and muscle resistance arteries of CHF rats, which presented as a dissociation between the intracellular Ca 2+ and the generated wall tension [123].

Cell and Tissue Responses
α 1 -ARs constitute an important homeostatic mechanism that contributes to the maintenance of peripheral vascular resistance at normal conditions and also in response to shock. Nevertheless, α 1 -AR regulation and function might alter due to disease. Even though HF is characterized by enhanced vascular resistance [124], there is little, and conflicting evidence to date concerning the vascular contractile responses to α 1 -AR stimulation in HF vessels.
The chronic sympathetic stimulation during HF might result in the desensitization of vascular α 1 -AR and consequently, in the failure of the contractile apparatus to function properly. Attenuation of α 1 -AR responsiveness after continuous exposure to catecholamines has been validated in various animal studies [125,126] but also in vitro [127,128]. However, despite the fact that several studies have tried to assess the α 1 -AR mediated vascular responses in both animals and humans with HF, the findings are not unanimous (Table 4). Multiple studies have shown an increase in vascular responsiveness to α 1 -AR stimulation in animals [114][115][116][117]119,120,122,[129][130][131]. On the contrary, others have reported desensitization of α 1 -ARs in the peripheral vasculature in animals [102,121,[132][133][134][135][136] and in humans [137][138][139][140], while numerous studies failed to demonstrate any alterations of the α 1 -AR responsiveness in vessels from animals [105,123,133,[141][142][143] and humans [144]. Table 4. Summary of changes in vascular α 1 -adrenoceptor (AR)-mediated responses in chronic heart failure.

Species Preparation/Parameter Studied α 1 -AR-Mediated Responses
Goldsmith et al. [137] human In vivo hemodynamic responses ↓ Inconsistency in findings might originate from differences in the anatomy and function of distinct vascular beds. Additionally, differences in CHF model and disease state might also influence the results of different studies. For example, in some studies it was demonstrated that the saphenous vein of CHF dogs was at all times more sensitive than the dorsal pedal artery [114][115][116]120]. This might be associated to a heterogeneous smooth muscle function between the artery and the vein, but, also, to differences in receptor population in the canine vasculature [148]. In addition, differences in vascular responsiveness between different vascular beds were also identified in rat models of HF [123,130]. Moreover, differences have also been identified between the various arteries studied. For example, in similar models of pacing-induced CHF in dogs, some studies have demonstrated an increased responsiveness to α 1 -AR stimulation in the dorsal pedal artery and the contralateral saphenous vein [114][115][116]119], while others have identified a desensitization of α 1 -ARs in coronary arteries of CHF dogs [132].
Other important parameters that influence the peripheral vascular responses are the integrity and physiological function of endothelium. More specifically, the endothelium-induced vasodilation counteracts the contraction via α 1 -ARs. Therefore, the malfunction of the endothelium might lead to the absence of this inhibitory function in CHF vessels [143] and result in an enhancement of vascular contraction. Increased vascular responses to noradrenaline were reported in the intact thoracic aorta from HF rats compared to controls, however the actual vascular α 1 -AR responsiveness was diminished when the endothelium was removed [146]. On the other hand, an increase in endothelial function as was observed in other studies might oppose the α 1 -AR-mediated vascular contractions [131,132]. Others, however, did not demonstrate a difference in vascular responses between intact and denuded vessels [123].
Even though the vascular responsiveness in patients with HF has not been widely investigated, many studies have yielded data that might have important clinical implications. More precisely, Schwinn et al. [145] reported on the vascular responsiveness of HF patients that underwent coronary artery revascularization. The results from this study support that left ventricular dysfunction and more precisely, left ventricular ejection fraction equal to or lower than 40% is associated with a diminished vascular responsiveness during fentanyl anesthesia in this patient population. In contrast to more recent reports that identify cardiac surgery and cardiopulmonary bypass as predisposing factors for an attenuated vascular responsiveness [13,14,149], Schwinn et al. [145] reported that phenylephrine-induced changes in systemic vascular resistance were increased during cardiopulmonary bypass and aortic cross-clamp in both HF and control subjects. This vascular responsiveness is of interest in HF patients that undergo cardiac surgery, since this patient population exhibits an increased risk for post-operative vasoplegia [150]. This complication is characterized by low systemic vascular resistance, severe hypotension, normal or increased cardiac output and a blunted or absent response to the administration of vasopressors [151]. The diminished sensitivity of HF patients is thought to originate from the continuous, chronic exposure of these patients' vascular α 1 -ARs to catecholamines which leads to desensitization and/or downregulation of those receptors. If vasoplegia is not promptly treated, the resulted tissue malperfusion might lead to end-organ dysfunction and multiple organ failure. Therefore, this post-operative complication significantly increases the risk of morbidity and mortality [152,153]. Moreover, another study identified a maladapted sympathetic response of the skeletal muscle vasculature to exercise that was characterized by sustained vascular constriction [140]. This can have serious implications on normal blood flow during physical movement and might hamper the patient's daily life. Furthermore, the apparent up-regulation of α 1 -AR responsiveness after chronic carvedilol treatment in HF patients [147] and the lack of change after chronic angiotensin-converting enzyme inhibitors [142] might yield important consequences regarding HF therapy.

Discussion
Understanding the role of α 1 -ARs is particularly important in HF because this condition is associated with high adrenergic activation leading to down-regulation and desensitization of cardiac β-ARs. α 1 -ARs have been identified in both animal and human hearts and vasculature with a high expression in rodents but a much lower abundance in humans. Thus, caution should be applied while extrapolating data from animals to humans.
The α 1A -AR is the most abundant subtype in the human heart and α 1B -AR in the rodent heart, while α 1D -ARs are mainly found in coronary arteries ( Figure 3). However, studies on a change in the expression of the subtypes in failing hearts are inconsistent, which may indicate a relative α 1 /β-AR imbalance in light of the well-known reduction in β-adrenergic function in CHF.
Cells 2020, 9, x FOR PEER REVIEW 14 of 25 is regulated depending on which α1-AR subtypes play the predominant role. For that reason, the differences in subtype distribution might also have influenced studies that have investigated the function of α1-ARs in vivo or in vitro, leading to conflicting results. α1-ARs can couple to multiple intracellular signal transduction pathways in the heart. Under most conditions α1-AR signal transduction results in enhanced inotropy. Although data on their role in HF is limited, there is evidence indicating that they mediate enhanced positive inotropy, as well as physiological hypertrophy and decreased apoptosis [46]. α1-AR occupancy was estimated as 10% in failing human heart [155]. Thus, activation of unoccupied receptors with an exogenous ligand is suggested to be cardioprotective [46]. Clinical trials such as ALLHAT in hypertension and V-HeFT in CHF reported that treatment with α1-AR antagonists was associated with greater mortality and apoptosis than with other treatments [156]. Moreover, double knockout of α1A-and α1B-AR was shown to augment apoptosis, worsen cardiomyopathy and reduce survival [157]. Taken together, cardiac α1-ARs appear to mediate cardioprotective effects. Data on the regulation of vascular α1-ARs in HF are also conflicting. This is unfortunate because of their role in controlling total peripheral resistance and the role of afterload reductions in the treatment of patients with HF. However, clinical studies in hypertensive patients have shown that α1-AR antagonism provides less protection from HF than other classes of anti-hypertensive drugs [156]. More specifically, the non-selective α-blocker doxazosin was associated with significantly higher risk of HF compared with the other treatment arms of this trial. In the absence of a placebo arm, it remained unclear whether this reflects an adverse effect of doxazosin or less protection against hypertension-induced HF. The results from the V-HeFT trial [158] further support the detrimental effects of α-blocking therapy-, since prazosin use gave a signal for possibly increased mortality that failed to reach statistical significance. The reason for those results could possibly be explained by the findings from knock-out mice studies which confirm the importance of α1-AR signaling for the adaptational mechanisms of cardiomyocytes in disease settings [157,159].
In contrast to those studies that investigated the effect of selective α-blockade and discouraged their widespread use, the mixed β-and α1-AR antagonist carvedilol was demonstrated to be more beneficial than the β1-selective antagonist metoprolol in diminishing the risk of mortality [155]. It could, therefore, be postulated that the favorable effects of carvedilol are to some extent related to its α-blocking properties. Nevertheless, the combination of an α-with a β-blocker for long-term Another important point is that, while in heart α 1A -and α 1B -ARs dominate, all three subtypes are expressed in the vasculature and their relative expression and physiological role appear to differ between vascular beds and species (Figure 3) [3]. This varied expression of α 1 -AR has been reported to correlate with the adrenergic innervation in peripheral vessels [154]. Moreover, studies with transfected receptors suggest that the three α 1 -AR subtypes undergo differential regulation upon exposure to agonists [16]. Certainly, those discrepancies also affect the way in which blood pressure is regulated depending on which α 1 -AR subtypes play the predominant role. For that reason, the differences in subtype distribution might also have influenced studies that have investigated the function of α 1 -ARs in vivo or in vitro, leading to conflicting results. α 1 -ARs can couple to multiple intracellular signal transduction pathways in the heart. Under most conditions α 1 -AR signal transduction results in enhanced inotropy. Although data on their role in HF is limited, there is evidence indicating that they mediate enhanced positive inotropy, as well as physiological hypertrophy and decreased apoptosis [46]. α1-AR occupancy was estimated as 10% in failing human heart [155]. Thus, activation of unoccupied receptors with an exogenous ligand is suggested to be cardioprotective [46]. Clinical trials such as ALLHAT in hypertension and V-HeFT in CHF reported that treatment with α 1 -AR antagonists was associated with greater mortality and apoptosis than with other treatments [156]. Moreover, double knockout of α 1A -and α 1B -AR was shown to augment apoptosis, worsen cardiomyopathy and reduce survival [157]. Taken together, cardiac α 1 -ARs appear to mediate cardioprotective effects. Data on the regulation of vascular α 1 -ARs in HF are also conflicting. This is unfortunate because of their role in controlling total peripheral resistance and the role of afterload reductions in the treatment of patients with HF. However, clinical studies in hypertensive patients have shown that α 1 -AR antagonism provides less protection from HF than other classes of anti-hypertensive drugs [156]. More specifically, the non-selective α-blocker doxazosin was associated with significantly higher risk of HF compared with the other treatment arms of this trial. In the absence of a placebo arm, it remained unclear whether this reflects an adverse effect of doxazosin or less protection against hypertension-induced HF. The results from the V-HeFT trial [158] further support the detrimental effects of α-blocking therapy-, since prazosin use gave a signal for possibly increased mortality that failed to reach statistical significance. The reason for those results could possibly be explained by the findings from knock-out mice studies which confirm the importance of α 1 -AR signaling for the adaptational mechanisms of cardiomyocytes in disease settings [157,159].
In contrast to those studies that investigated the effect of selective α-blockade and discouraged their widespread use, the mixed βand α 1 -AR antagonist carvedilol was demonstrated to be more beneficial than the β 1 -selective antagonist metoprolol in diminishing the risk of mortality [155]. It could, therefore, be postulated that the favorable effects of carvedilol are to some extent related to its α-blocking properties. Nevertheless, the combination of an αwith a β-blocker for long-term treatment in chronic CHF patients, did not yield better clinical outcomes than β-blocker therapy alone [160], which highlights the fact that such a combined therapy does not generate satisfactory outcomes similar to that of carvedilol.
The exact reason for those discrepancies in effectiveness still remains unclear. One explanation could be that carvedilol's α-blocking properties are less potent compared to both its innate β-blocking feature and to the α-blockade produced by a specific α-AR antagonist like prazosin. This is in fact supported by the results of a small trial in which carvedilol did not antagonize phenylephrine induced responses [161]. Another hypothesis is that the existence of distinct pharmacologic properties and mechanisms of action between the various αand β-blockers could also lead to the occurrence of different side effects. Based on this assumption, one could possibly explain the contradictory findings from studies with carvedilol versus studies investigating the combination of an αwith a β-blocker. Last but not least, carvedilol use leads to pharmacologic tolerance after long term use [162]. This implies that its beneficial effects at long-term use derive mostly from its β-blocking properties rather than antagonizing α 1 -ARs.
In brief, one could state that α 1 -AR agonism could potentially be a valid treatment option for HF patients that would stimulate innate adaptive signaling procedures and would eventually illustrate its cardioprotective features. However, caution is required, since α 1 agonism could also have detrimental effects by increasing peripheral vascular resistance leading to an increased cardiac afterload and by inducing coronary vasoconstriction in those patients. Therefore, future research is necessary which should focus on compartment and/or subtype selective treatment in HF patients and which will investigate in more detail the treatment modalities of α 1 agonism.
On the other hand, carvedilol seems promising for the treatment of HF patients and might be an option for the pretreatment of HF patients that will undergo cardiac surgery. Short term α 1 antagonism with carvedilol treatment could shield α 1-ARs against continuous catecholamine exposure in the preoperative period. This could perhaps increase the sensitivity of those receptors to the administration of catecholamines for the treatment of vasoplegia during the postoperative course [147].
In conclusion, more research is required to further understand the complex mechanisms of α 1 adrenergic function and adaptation in HF. Taking into account the existing differences between animals and humans regarding the expression of α 1 -ARs, we should carefully consider our next steps and we should promote additional research which would focus more on human tissue (heart and vessels). Moreover, a cardioselective approach in the case of α 1 agonism appears to be crucial for the exploitation of only its advantageous characteristics by avoiding potential vascular side effects. A subtype approach could also be beneficial but further studies are necessary in order to gain more insight in such treatment techniques. Finally, in view of the positive outcomes from clinical studies with carvedilol, more extensive investigation is to be expected so as to better comprehend the pharmacologic properties and mechanism of action of this agent. It is, therefore, essential that we expand our knowledge, and we apprehend current inconsistencies in order to pave the way to safer and more effective therapies.